From 49c20457b3092a7af2040da6e7a2b41931e856e4 Mon Sep 17 00:00:00 2001
From: Hashim Sharif <hsharif3@miranda.cs.illinois.edu>
Date: Mon, 25 Jan 2021 14:00:59 -0600
Subject: [PATCH] Removing unused dirs from hpvm-tensor-rt
---
.../hpvm-tensor-rt/PPoPP_results/knobs.txt | 28 -
.../alexnet2_cifar10_fp16_perf_fp16_120.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_151.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_152.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_153.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_154.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_155.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_156.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_157.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_158.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_159.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_160.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_161.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_162.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_163.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_164.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_165.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_166.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_167.txt | 72 -
.../alexnet2_cifar10_fp16_perf_fp16_168.txt | 72 -
.../alexnet2_cifar10_fp16_samp_fp16_261.txt | 72 -
.../alexnet2_cifar10_fp16_samp_fp16_262.txt | 72 -
.../alexnet2_cifar10_fp16_samp_fp16_263.txt | 72 -
.../alexnet2_cifar10_fp16_samp_fp16_264.txt | 72 -
.../alexnet2_cifar10_fp16_samp_fp16_265.txt | 72 -
.../alexnet2_cifar10_fp16_samp_fp16_266.txt | 72 -
.../alexnet2_cifar10_fp16_samp_fp16_267.txt | 72 -
.../alexnet2_cifar10_fp16_samp_fp16_268.txt | 72 -
.../alexnet2_cifar10_fp16_samp_fp16_269.txt | 72 -
.../alexnet2_cifar10_fp32_perf_fp32_120.txt | 72 -
.../alexnet_canny_fp16_perf_fp16_120.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_151.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_152.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_153.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_154.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_155.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_156.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_157.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_158.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_159.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_160.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_161.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_162.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_163.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_164.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_165.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_166.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_167.txt | 108 -
.../alexnet_canny_fp16_perf_fp16_168.txt | 108 -
.../alexnet_canny_fp16_red_samp_fp16_42.txt | 108 -
.../alexnet_canny_fp16_red_samp_fp16_44.txt | 108 -
.../alexnet_canny_fp16_red_samp_fp16_46.txt | 108 -
.../alexnet_canny_fp16_samp_fp16_261.txt | 108 -
.../alexnet_canny_fp16_samp_fp16_262.txt | 108 -
.../alexnet_canny_fp16_samp_fp16_263.txt | 108 -
.../alexnet_canny_fp16_samp_fp16_264.txt | 108 -
.../alexnet_canny_fp16_samp_fp16_265.txt | 108 -
.../alexnet_canny_fp16_samp_fp16_266.txt | 108 -
.../alexnet_canny_fp16_samp_fp16_267.txt | 108 -
.../alexnet_canny_fp16_samp_fp16_268.txt | 108 -
.../alexnet_canny_fp16_samp_fp16_269.txt | 108 -
.../alexnet_canny_fp32_perf_fp32_120.txt | 108 -
.../alexnet_canny_perf_121.txt | 108 -
.../alexnet_canny_perf_122.txt | 108 -
.../alexnet_canny_perf_123.txt | 108 -
.../alexnet_canny_perf_124.txt | 108 -
.../alexnet_canny_perf_125.txt | 108 -
.../alexnet_canny_perf_126.txt | 108 -
.../alexnet_canny_perf_127.txt | 108 -
.../alexnet_canny_perf_128.txt | 108 -
.../alexnet_canny_perf_129.txt | 108 -
.../alexnet_canny_perf_130.txt | 108 -
.../alexnet_canny_perf_131.txt | 108 -
.../alexnet_canny_perf_132.txt | 108 -
.../alexnet_canny_perf_133.txt | 108 -
.../alexnet_canny_perf_134.txt | 108 -
.../alexnet_canny_perf_135.txt | 108 -
.../alexnet_canny_perf_136.txt | 108 -
.../alexnet_canny_perf_137.txt | 108 -
.../alexnet_canny_perf_138.txt | 108 -
.../alexnet_canny_red_samp_41.txt | 108 -
.../alexnet_canny_red_samp_43.txt | 108 -
.../alexnet_canny_red_samp_45.txt | 108 -
.../alexnet_canny_samp_231.txt | 108 -
.../alexnet_canny_samp_232.txt | 108 -
.../alexnet_canny_samp_233.txt | 108 -
.../alexnet_canny_samp_234.txt | 108 -
.../alexnet_canny_samp_235.txt | 108 -
.../alexnet_canny_samp_236.txt | 108 -
.../alexnet_canny_samp_237.txt | 108 -
.../alexnet_canny_samp_238.txt | 108 -
.../alexnet_canny_samp_239.txt | 108 -
.../alexnet_canny_perf_120.txt | 108 -
.../alexnet_canny_perf_121.txt | 108 -
.../alexnet_canny_perf_122.txt | 108 -
.../alexnet_canny_perf_123.txt | 108 -
.../alexnet_canny_perf_124.txt | 108 -
.../alexnet_canny_perf_125.txt | 108 -
.../alexnet_canny_perf_126.txt | 108 -
.../alexnet_canny_perf_127.txt | 108 -
.../alexnet_canny_perf_128.txt | 108 -
.../alexnet_canny_perf_129.txt | 108 -
.../alexnet_canny_perf_130.txt | 108 -
.../alexnet_canny_perf_131.txt | 108 -
.../alexnet_canny_perf_132.txt | 108 -
.../alexnet_canny_perf_133.txt | 108 -
.../alexnet_canny_perf_134.txt | 108 -
.../alexnet_canny_perf_135.txt | 108 -
.../alexnet_canny_perf_136.txt | 108 -
.../alexnet_canny_perf_137.txt | 108 -
.../alexnet_canny_perf_138.txt | 108 -
.../alexnet_canny_red_samp_41.txt | 108 -
.../alexnet_canny_red_samp_43.txt | 108 -
.../alexnet_canny_red_samp_45.txt | 108 -
.../alexnet_canny_samp_231.txt | 108 -
.../alexnet_canny_samp_232.txt | 108 -
.../alexnet_canny_samp_233.txt | 108 -
.../alexnet_canny_samp_234.txt | 108 -
.../alexnet_canny_samp_235.txt | 108 -
.../alexnet_canny_samp_236.txt | 108 -
.../alexnet_canny_samp_237.txt | 108 -
.../alexnet_canny_samp_238.txt | 108 -
.../alexnet_canny_samp_239.txt | 108 -
.../alexnet_cifar10_fp16_perf_fp16_120.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_151.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_152.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_153.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_154.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_155.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_156.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_157.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_158.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_159.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_160.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_161.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_162.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_163.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_164.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_165.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_166.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_167.txt | 63 -
.../alexnet_cifar10_fp16_perf_fp16_168.txt | 63 -
.../alexnet_cifar10_fp16_samp_fp16_261.txt | 63 -
.../alexnet_cifar10_fp16_samp_fp16_262.txt | 63 -
.../alexnet_cifar10_fp16_samp_fp16_263.txt | 63 -
.../alexnet_cifar10_fp16_samp_fp16_264.txt | 63 -
.../alexnet_cifar10_fp16_samp_fp16_265.txt | 63 -
.../alexnet_cifar10_fp16_samp_fp16_266.txt | 63 -
.../alexnet_cifar10_fp16_samp_fp16_267.txt | 63 -
.../alexnet_cifar10_fp16_samp_fp16_268.txt | 63 -
.../alexnet_cifar10_fp16_samp_fp16_269.txt | 63 -
.../alexnet_cifar10_fp32_perf_fp32_120.txt | 63 -
.../alexnet_imagenet_fp16_perf_fp16_120.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_151.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_152.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_153.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_154.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_155.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_156.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_157.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_158.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_159.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_160.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_161.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_162.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_163.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_164.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_165.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_166.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_167.txt | 81 -
.../alexnet_imagenet_fp16_perf_fp16_168.txt | 81 -
.../alexnet_imagenet_fp16_samp_fp16_261.txt | 81 -
.../alexnet_imagenet_fp16_samp_fp16_262.txt | 81 -
.../alexnet_imagenet_fp16_samp_fp16_263.txt | 81 -
.../alexnet_imagenet_fp16_samp_fp16_264.txt | 81 -
.../alexnet_imagenet_fp16_samp_fp16_265.txt | 81 -
.../alexnet_imagenet_fp16_samp_fp16_266.txt | 81 -
.../alexnet_imagenet_fp16_samp_fp16_267.txt | 81 -
.../alexnet_imagenet_fp16_samp_fp16_268.txt | 81 -
.../alexnet_imagenet_fp16_samp_fp16_269.txt | 81 -
.../alexnet_imagenet_fp32_perf_fp32_120.txt | 81 -
.../lenet_keras_fp16_perf_fp16_120.txt | 45 -
.../lenet_keras_fp16_perf_fp16_151.txt | 45 -
.../lenet_keras_fp16_perf_fp16_152.txt | 45 -
.../lenet_keras_fp16_perf_fp16_153.txt | 45 -
.../lenet_keras_fp16_perf_fp16_154.txt | 45 -
.../lenet_keras_fp16_perf_fp16_155.txt | 45 -
.../lenet_keras_fp16_perf_fp16_156.txt | 45 -
.../lenet_keras_fp16_perf_fp16_157.txt | 45 -
.../lenet_keras_fp16_perf_fp16_158.txt | 45 -
.../lenet_keras_fp16_perf_fp16_159.txt | 45 -
.../lenet_keras_fp16_perf_fp16_160.txt | 45 -
.../lenet_keras_fp16_perf_fp16_161.txt | 45 -
.../lenet_keras_fp16_perf_fp16_162.txt | 45 -
.../lenet_keras_fp16_perf_fp16_163.txt | 45 -
.../lenet_keras_fp16_perf_fp16_164.txt | 45 -
.../lenet_keras_fp16_perf_fp16_165.txt | 45 -
.../lenet_keras_fp16_perf_fp16_166.txt | 45 -
.../lenet_keras_fp16_perf_fp16_167.txt | 45 -
.../lenet_keras_fp16_perf_fp16_168.txt | 45 -
.../lenet_keras_fp16_samp_fp16_261.txt | 45 -
.../lenet_keras_fp16_samp_fp16_262.txt | 45 -
.../lenet_keras_fp16_samp_fp16_263.txt | 45 -
.../lenet_keras_fp16_samp_fp16_264.txt | 45 -
.../lenet_keras_fp16_samp_fp16_265.txt | 45 -
.../lenet_keras_fp16_samp_fp16_266.txt | 45 -
.../lenet_keras_fp16_samp_fp16_267.txt | 45 -
.../lenet_keras_fp16_samp_fp16_268.txt | 45 -
.../lenet_keras_fp16_samp_fp16_269.txt | 45 -
.../lenet_keras_fp32_perf_fp32_120.txt | 45 -
.../mobilenet_cifar10_fp16_perf_fp16_120.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_151.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_152.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_153.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_154.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_155.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_156.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_157.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_158.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_159.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_160.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_161.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_162.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_163.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_164.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_165.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_166.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_167.txt | 255 --
.../mobilenet_cifar10_fp16_perf_fp16_168.txt | 255 --
.../mobilenet_cifar10_fp16_samp_fp16_261.txt | 255 --
.../mobilenet_cifar10_fp16_samp_fp16_262.txt | 255 --
.../mobilenet_cifar10_fp16_samp_fp16_263.txt | 255 --
.../mobilenet_cifar10_fp16_samp_fp16_264.txt | 255 --
.../mobilenet_cifar10_fp16_samp_fp16_265.txt | 255 --
.../mobilenet_cifar10_fp16_samp_fp16_266.txt | 255 --
.../mobilenet_cifar10_fp16_samp_fp16_267.txt | 255 --
.../mobilenet_cifar10_fp16_samp_fp16_268.txt | 255 --
.../mobilenet_cifar10_fp16_samp_fp16_269.txt | 255 --
.../mobilenet_cifar10_fp32_perf_fp32_120.txt | 255 --
.../resnet18_cifar10_fp16_perf_fp16_120.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_151.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_152.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_153.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_154.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_155.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_156.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_157.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_158.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_159.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_160.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_161.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_162.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_163.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_164.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_165.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_166.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_167.txt | 222 --
.../resnet18_cifar10_fp16_perf_fp16_168.txt | 222 --
.../resnet18_cifar10_fp16_samp_fp16_261.txt | 222 --
.../resnet18_cifar10_fp16_samp_fp16_262.txt | 222 --
.../resnet18_cifar10_fp16_samp_fp16_263.txt | 222 --
.../resnet18_cifar10_fp16_samp_fp16_264.txt | 222 --
.../resnet18_cifar10_fp16_samp_fp16_265.txt | 222 --
.../resnet18_cifar10_fp16_samp_fp16_266.txt | 222 --
.../resnet18_cifar10_fp16_samp_fp16_267.txt | 222 --
.../resnet18_cifar10_fp16_samp_fp16_268.txt | 222 --
.../resnet18_cifar10_fp16_samp_fp16_269.txt | 222 --
.../resnet18_cifar10_fp32_perf_fp32_120.txt | 222 --
.../resnet50_imagenet_fp16_perf_fp16_120.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_151.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_152.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_153.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_154.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_155.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_156.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_157.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_158.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_159.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_160.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_161.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_162.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_163.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_164.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_165.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_166.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_167.txt | 687 ----
.../resnet50_imagenet_fp16_perf_fp16_168.txt | 687 ----
.../resnet50_imagenet_fp16_samp_fp16_261.txt | 687 ----
.../resnet50_imagenet_fp16_samp_fp16_262.txt | 687 ----
.../resnet50_imagenet_fp16_samp_fp16_263.txt | 687 ----
.../resnet50_imagenet_fp16_samp_fp16_264.txt | 687 ----
.../resnet50_imagenet_fp16_samp_fp16_265.txt | 687 ----
.../resnet50_imagenet_fp16_samp_fp16_266.txt | 687 ----
.../resnet50_imagenet_fp16_samp_fp16_267.txt | 687 ----
.../resnet50_imagenet_fp16_samp_fp16_268.txt | 687 ----
.../resnet50_imagenet_fp16_samp_fp16_269.txt | 687 ----
.../resnet50_imagenet_fp32_perf_fp32_120.txt | 687 ----
.../vgg16_cifar10_fp16_perf_fp16_120.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_151.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_152.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_153.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_154.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_155.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_156.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_157.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_158.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_159.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_160.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_161.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_162.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_163.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_164.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_165.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_166.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_167.txt | 150 -
.../vgg16_cifar10_fp16_perf_fp16_168.txt | 150 -
.../vgg16_cifar10_fp16_samp_fp16_261.txt | 150 -
.../vgg16_cifar10_fp16_samp_fp16_262.txt | 150 -
.../vgg16_cifar10_fp16_samp_fp16_263.txt | 150 -
.../vgg16_cifar10_fp16_samp_fp16_264.txt | 150 -
.../vgg16_cifar10_fp16_samp_fp16_265.txt | 150 -
.../vgg16_cifar10_fp16_samp_fp16_266.txt | 150 -
.../vgg16_cifar10_fp16_samp_fp16_267.txt | 150 -
.../vgg16_cifar10_fp16_samp_fp16_268.txt | 150 -
.../vgg16_cifar10_fp16_samp_fp16_269.txt | 150 -
.../vgg16_cifar10_fp32_perf_fp32_120.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_120.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_151.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_152.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_153.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_154.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_155.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_156.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_157.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_158.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_159.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_160.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_161.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_162.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_163.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_164.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_165.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_166.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_167.txt | 150 -
.../vgg16_cifar100_fp16_perf_fp16_168.txt | 150 -
.../vgg16_cifar100_fp16_samp_fp16_261.txt | 150 -
.../vgg16_cifar100_fp16_samp_fp16_262.txt | 150 -
.../vgg16_cifar100_fp16_samp_fp16_263.txt | 150 -
.../vgg16_cifar100_fp16_samp_fp16_264.txt | 150 -
.../vgg16_cifar100_fp16_samp_fp16_265.txt | 150 -
.../vgg16_cifar100_fp16_samp_fp16_266.txt | 150 -
.../vgg16_cifar100_fp16_samp_fp16_267.txt | 150 -
.../vgg16_cifar100_fp16_samp_fp16_268.txt | 150 -
.../vgg16_cifar100_fp16_samp_fp16_269.txt | 150 -
.../vgg16_cifar100_fp32_perf_fp32_120.txt | 150 -
.../vgg16_imagenet_fp16_perf_fp16_120.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_151.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_152.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_153.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_154.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_155.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_156.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_157.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_158.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_159.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_160.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_161.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_162.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_163.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_164.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_165.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_166.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_167.txt | 159 -
.../vgg16_imagenet_fp16_perf_fp16_168.txt | 159 -
.../vgg16_imagenet_fp16_samp_fp16_261.txt | 159 -
.../vgg16_imagenet_fp16_samp_fp16_262.txt | 159 -
.../vgg16_imagenet_fp16_samp_fp16_263.txt | 159 -
.../vgg16_imagenet_fp16_samp_fp16_264.txt | 159 -
.../vgg16_imagenet_fp16_samp_fp16_265.txt | 159 -
.../vgg16_imagenet_fp16_samp_fp16_266.txt | 159 -
.../vgg16_imagenet_fp16_samp_fp16_267.txt | 159 -
.../vgg16_imagenet_fp16_samp_fp16_268.txt | 159 -
.../vgg16_imagenet_fp16_samp_fp16_269.txt | 159 -
.../vgg16_imagenet_fp32_perf_fp32_120.txt | 159 -
.../alexnet/alexnet_valid_soc.txt | 231 --
.../alexnet2/alexnet2_valid_soc.txt | 188 -
.../resnet18/resnet18_valid_soc.txt | 1576 --------
.../vgg16_cifar10/vgg16_cifar10_valid_soc.txt | 1027 ------
.../vgg16_cifar100_valid_soc.txt | 210 --
.../PPoPP_results/soc_sim_results/README.md | 0
.../alexnet2_cifar10/alexnet2_cifar10.txt | 23 -
.../alexnet_cifar10/alexnet_cifar10.txt | 421 ---
.../alexnet_imagenet/alexnet_imagenet.txt | 289 --
.../empirical/lenet_keras/lenet_keras.txt | 409 ---
.../mobilenet_cifar10/mobilenet_cifar10.txt | 871 -----
.../resnet18_cifar10/resnet18_cifar10.txt | 91 -
.../resnet50_imagenet/resnet50_imagenet.txt | 1233 -------
.../empirical/vgg16_cifar10/vgg16_cifar10.txt | 58 -
.../vgg16_cifar100/vgg16_cifar100.txt | 77 -
.../vgg16_imagenet/vgg16_imagenet.txt | 41 -
.../alexnet2_cifar10/alexnet2_cifar10.txt | 419 ---
.../alexnet_cifar10/alexnet_cifar10.txt | 511 ---
.../alexnet_imagenet/alexnet_imagenet.txt | 229 --
.../predictive/lenet_keras/lenet_keras.txt | 409 ---
.../mobilenet_cifar10/mobilenet_cifar10.txt | 3220 -----------------
.../resnet18_cifar10/resnet18_cifar10.txt | 2296 ------------
.../resnet50_imagenet/resnet50_imagenet.txt | 1057 ------
.../vgg16_cifar10/vgg16_cifar10.txt | 913 -----
.../vgg16_cifar100/vgg16_cifar100.txt | 970 -----
.../vgg16_imagenet/vgg16_imagenet.txt | 561 ---
.../autotuner/autotuner/img_pareto_curve.py | 169 -
.../autotuner/data/alexnet/alexnet_layers.txt | 6 -
.../data/alexnet/alexnet_tensors.txt | 26 -
.../autotuner/data/alexnet/dev_knobs.txt | 6 -
.../autotuner/data/alexnet/knobs.txt | 6 -
.../autotuner/data/alexnet/knobs1.txt | 6 -
.../autotuner/data/alexnet/op_cost.txt | 6 -
.../data/alexnet2/alexnet2_layers.txt | 7 -
.../data/alexnet2/alexnet2_tensors.txt | 30 -
.../autotuner/data/alexnet2/dev_knobs.txt | 7 -
.../autotuner/data/alexnet2/knobs.txt | 7 -
.../autotuner/data/alexnet2/knobs1.txt | 7 -
.../autotuner/data/alexnet2/op_cost.txt | 7 -
.../data/alexnet2_canny_hpvm/layers.txt | 13 -
.../data/alexnet2_canny_hpvm/op_cost.txt | 13 -
.../alexnet_imagenet/layer_composition.txt | 8 -
.../data/alexnet_imagenet/op_cost.txt | 8 -
.../data/alexnet_imagenet/quant_ranges2.txt | 8 -
.../autotuner/data/blend/knobs.txt | 17 -
.../autotuner/data/blend/op_cost.txt | 17 -
.../data/blend/tuner_conf_template.txt | 20 -
.../autotuner/data/canny/knobs.txt | 9 -
.../autotuner/data/canny/op_cost.txt | 9 -
.../data/canny/tuner_conf_template.txt | 12 -
.../autotuner/data/fft/knobs.txt | 4 -
.../autotuner/data/fft/op_cost.txt | 4 -
.../data/fft/tuner_conf_template.txt | 7 -
.../autotuner/data/global_knobs.txt | 69 -
.../autotuner/data/global_knobs_dnn.txt | 25 -
.../autotuner/data/global_knobs_old.txt | 31 -
.../autotuner/data/lenet/dev_knobs.txt | 4 -
.../autotuner/data/lenet/knobs.txt | 4 -
.../autotuner/data/lenet/knobs1.txt | 4 -
.../autotuner/data/lenet/lenet_layers.txt | 4 -
.../autotuner/data/lenet/lenet_tensors.txt | 18 -
.../autotuner/data/lenet/op_cost.txt | 4 -
.../autotuner/data/mobilenet/dev_knobs.txt | 15 -
.../autotuner/data/mobilenet/knobs.txt | 15 -
.../autotuner/data/mobilenet/knobs1.txt | 15 -
.../data/mobilenet/mobilenet_layer_comp.txt | 83 -
.../data/mobilenet/mobilenet_ops.txt | 167 -
.../autotuner/data/mobilenet/op_cost.txt | 28 -
.../data/mobilenet_shallow/dev_knobs.txt | 8 -
.../data/mobilenet_shallow/knobs.txt | 8 -
.../data/mobilenet_shallow/knobs1.txt | 8 -
.../mobilenet_shallow_layer_comp.txt | 41 -
.../mobilenet_shallow_ops.txt | 47 -
.../data/mobilenet_shallow/op_cost.txt | 8 -
.../autotuner/data/mobilenet_torch/layers.txt | 81 -
.../data/mobilenet_torch/op_cost.txt | 15 -
.../autotuner/data/resnet/dev_knobs.txt | 22 -
.../autotuner/data/resnet/knobs.txt | 22 -
.../autotuner/data/resnet/knobs1.txt | 22 -
.../autotuner/data/resnet/knobs2.txt | 22 -
.../autotuner/data/resnet/op_cost.txt | 22 -
.../autotuner/data/resnet/resnet_layers.txt | 41 -
.../autotuner/data/resnet/resnet_tensors.txt | 114 -
.../autotuner/data/resnet18_torch/op_cost.txt | 21 -
.../data/resnet18_torch/resnet_layers.txt | 59 -
.../data/resnet50_imagenet/dev_knobs.txt | 54 -
.../data/resnet50_imagenet/op_cost.txt | 54 -
.../data/resnet50_imagenet/quant_ranges2.txt | 54 -
.../resnet50_imagenet/resnet50_layers.txt | 172 -
.../resnet50_imagenet/resnet50_tensors.txt | 172 -
.../data/vgg16_cifar10/dev_knobs.txt | 15 -
.../autotuner/data/vgg16_cifar10/knobs.txt | 15 -
.../autotuner/data/vgg16_cifar10/knobs1.txt | 15 -
.../autotuner/data/vgg16_cifar10/op_cost.txt | 15 -
.../data/vgg16_cifar10/vgg16_layers.txt | 15 -
.../data/vgg16_cifar10/vgg16_tensors.txt | 64 -
.../data/vgg16_cifar100/dev_knobs.txt | 15 -
.../autotuner/data/vgg16_cifar100/knobs.txt | 15 -
.../autotuner/data/vgg16_cifar100/knobs1.txt | 15 -
.../autotuner/data/vgg16_cifar100/op_cost.txt | 15 -
.../data/vgg16_cifar100/vgg16_layers.txt | 15 -
.../data/vgg16_cifar100/vgg16_tensors.txt | 64 -
.../data/vgg16_cifar10_torch/layers.txt | 46 -
.../data/vgg16_cifar10_torch/op_cost.txt | 15 -
.../data/vgg16_imagenet/dev_knobs.txt | 16 -
.../autotuner/data/vgg16_imagenet/knobs.txt | 15 -
.../autotuner/data/vgg16_imagenet/knobs1.txt | 15 -
.../autotuner/data/vgg16_imagenet/op_cost.txt | 16 -
.../data/vgg16_imagenet/quant_ranges2.txt | 16 -
.../data/vgg16_imagenet/vgg16_layers.txt | 16 -
.../data/vgg16_imagenet/vgg16_tensors.txt | 68 -
.../autotuner/opentuner/.gitignore | 51 -
.../autotuner/opentuner/AUTHORS.txt | 8 -
.../autotuner/opentuner/CHANGES.txt | 3 -
.../autotuner/opentuner/LICENSE.txt | 22 -
.../autotuner/opentuner/MANIFEST.in | 1 -
.../autotuner/opentuner/README.md | 116 -
.../autotuner/opentuner/autotuner/__init__.py | 0
.../opentuner/autotuner/accuracy_tuner.py | 198 -
.../autotuner/accuracy_tuner_piped.py | 269 --
.../autotuner/opentuner/autotuner/adddeps.py | 5 -
.../opentuner/autotuner/algo_tuner.py | 318 --
.../opentuner/autotuner/algo_tuner2.py | 339 --
.../opentuner/autotuner/approxhpvm_tuner.py | 262 --
.../autotuner/opentuner/autotuner/devtuner.py | 331 --
.../opentuner/autotuner/gettingstarted.md | 215 --
.../opentuner/autotuner/measure_confidence.py | 290 --
.../autotuner/measure_confidence2.py | 664 ----
.../opentuner/autotuner/pareto_curve.py | 264 --
.../opentuner/autotuner/promise_tuner2.py | 220 --
.../opentuner/autotuner/promise_tuner3.py | 314 --
.../autotuner/promise_tuner_piped.py | 231 --
.../opentuner/autotuner/psnr_tuner.py | 318 --
.../opentuner/autotuner/select_top_results.py | 101 -
.../autotuner/opentuner/autotuner/utils.py | 157 -
.../autotuner/opentuner/debian-packages-deps | 9 -
.../autotuner/opentuner/docs/Makefile | 177 -
.../opentuner/docs/rtd-requirements.txt | 5 -
.../autotuner/opentuner/docs/source/conf.py | 261 --
.../autotuner/opentuner/docs/source/index.rst | 27 -
.../opentuner/docs/source/params.rst | 339 --
.../opentuner/docs/source/techniques.rst | 51 -
.../autotuner/opentuner/examples/.gitignore | 4 -
.../opentuner/examples/gccflags/.gitignore | 5 -
.../opentuner/examples/gccflags/adddeps.py | 6 -
.../examples/gccflags/apps/.gitignore | 1 -
.../examples/gccflags/apps/matrixmultiply.cpp | 80 -
.../examples/gccflags/apps/raytracer.cpp | 277 --
.../examples/gccflags/apps/tsp_ga.cpp | 548 ---
.../examples/gccflags/cc_param_defaults.json | 1 -
.../opentuner/examples/gccflags/gccflags.py | 412 ---
.../examples/gccflags/gccflags_minimal.py | 92 -
.../opentuner/examples/halide/.gitignore | 2 -
.../opentuner/examples/halide/adddeps.py | 6 -
.../examples/halide/apps/bilateral_grid.cpp | 93 -
.../halide/apps/bilateral_grid.settings | 10 -
.../examples/halide/apps/halide_blur.cpp | 32 -
.../examples/halide/apps/halide_blur.settings | 4 -
.../halide/apps/interpolate-simple.cpp | 208 --
.../halide/apps/interpolate-simple.settings | 185 -
.../halide/apps/interpolate-simplest.cpp | 208 --
.../halide/apps/interpolate-simplest.settings | 124 -
.../examples/halide/apps/interpolate.cpp | 208 --
.../examples/halide/apps/interpolate.settings | 612 ----
.../examples/halide/apps/wavelet.cpp | 93 -
.../examples/halide/apps/wavelet.settings | 4 -
.../opentuner/examples/halide/halidetuner.py | 682 ----
.../opentuner/examples/halide/timing_prefix.h | 100 -
.../opentuner/examples/hpl/HPL.dat.mako | 31 -
.../opentuner/examples/hpl/adddeps.py | 6 -
.../autotuner/opentuner/examples/hpl/hpl.py | 98 -
.../opentuner/examples/mario/README.md | 33 -
.../opentuner/examples/mario/adddeps.py | 6 -
.../opentuner/examples/mario/fceux-hook.lua | 25 -
.../opentuner/examples/mario/mario.py | 341 --
.../opentuner/examples/petabricks/.gitignore | 1 -
.../opentuner/examples/petabricks/README.md | 24 -
.../opentuner/examples/petabricks/adddeps.py | 6 -
.../opentuner/examples/petabricks/deps.py | 19 -
.../petabricks/download_benchmarks.sh | 8 -
.../examples/petabricks/import_old_result.py | 116 -
.../opentuner/examples/petabricks/pbtuner.py | 188 -
.../examples/petabricks/testwrapper.sh | 11 -
.../opentuner/examples/py_api/adddeps.py | 6 -
.../opentuner/examples/py_api/api_example.py | 58 -
.../examples/py_api/multiple_tuning_runs.py | 83 -
.../opentuner/examples/rosenbrock/.gitignore | 3 -
.../opentuner/examples/rosenbrock/adddeps.py | 6 -
.../examples/rosenbrock/rosenbrock.makefile | 24 -
.../examples/rosenbrock/rosenbrock.py | 85 -
.../opentuner/examples/tsp/adddeps.py | 6 -
.../opentuner/examples/tsp/att48_d.txt | 48 -
.../opentuner/examples/tsp/p01_d.txt | 15 -
.../opentuner/examples/tsp/p01_s.txt | 16 -
.../autotuner/opentuner/examples/tsp/tsp.py | 59 -
.../examples/tutorials/#accuracy_tuner.py# | 203 --
.../opentuner/examples/tutorials/__init__.py | 0
.../examples/tutorials/accuracy_tuner.py | 198 -
.../tutorials/accuracy_tuner_piped.py | 269 --
.../opentuner/examples/tutorials/adddeps.py | 5 -
.../examples/tutorials/gettingstarted.md | 215 --
.../examples/tutorials/measure_confidence.py | 164 -
.../examples/tutorials/mmm_block.cpp | 36 -
.../opentuner/examples/tutorials/mmm_tuner.py | 116 -
.../examples/tutorials/select_top_results.py | 101 -
.../opentuner/examples/unitary/adddeps.py | 6 -
.../opentuner/examples/unitary/cla_func.py | 144 -
.../examples/unitary/input_generator.py | 100 -
.../examples/unitary/problem_description.pdf | Bin 24984 -> 0 bytes
.../opentuner/examples/unitary/testwrapper.sh | 11 -
.../opentuner/examples/unitary/unitary.py | 136 -
.../autotuner/opentuner/gen-venv-bootstrap.py | 39 -
.../autotuner/opentuner/install_reqs.sh | 3 -
.../opentuner/misc/livedisplay.gnuplot | 10 -
.../autotuner/opentuner/misc/livedisplay.py | 57 -
.../autotuner/opentuner/opentuner/__init__.py | 41 -
.../autotuner/opentuner/opentuner/api.py | 87 -
.../opentuner/opentuner/driverbase.py | 48 -
.../opentuner/measurement/#interface.py# | 359 --
.../opentuner/measurement/.#interface.py | 1 -
.../opentuner/measurement/__init__.py | 6 -
.../opentuner/opentuner/measurement/driver.py | 271 --
.../opentuner/measurement/inputmanager.py | 76 -
.../opentuner/measurement/interface.py | 366 --
.../opentuner/opentuner/resultsdb/__init__.py | 6 -
.../opentuner/opentuner/resultsdb/connect.py | 66 -
.../opentuner/opentuner/resultsdb/models.py | 319 --
.../opentuner/opentuner/search/__init__.py | 6 -
.../opentuner/search/bandittechniques.py | 316 --
.../composableevolutionarytechniques.py | 512 ---
.../opentuner/search/differentialevolution.py | 148 -
.../opentuner/opentuner/search/driver.py | 301 --
.../search/evolutionarytechniques.py | 153 -
.../opentuner/opentuner/search/globalGA.py | 125 -
.../opentuner/opentuner/search/manipulator.py | 1853 ----------
.../opentuner/search/metatechniques.py | 186 -
.../opentuner/opentuner/search/objective.py | 338 --
.../opentuner/search/patternsearch.py | 72 -
.../opentuner/opentuner/search/plugin.py | 152 -
.../opentuner/opentuner/search/pso.py | 81 -
.../opentuner/search/simplextechniques.py | 457 ---
.../opentuner/search/simulatedannealing.py | 133 -
.../opentuner/opentuner/search/technique.py | 358 --
.../opentuner/opentuner/tuningrunmain.py | 224 --
.../opentuner/opentuner/utils/__init__.py | 0
.../opentuner/opentuner/utils/adddeps.py | 13 -
.../opentuner/opentuner/utils/compactdb.py | 54 -
.../opentuner/opentuner/utils/dictconfig.py | 544 ---
.../opentuner/opentuner/utils/stats.py | 468 ---
.../opentuner/utils/stats_matplotlib.py | 290 --
.../opentuner/optional-requirements.txt | 3 -
.../autotuner/opentuner/requirements.txt | 5 -
.../autotuner/opentuner/setup.py | 33 -
.../autotuner/opentuner/stats_app/manage.py | 10 -
.../opentuner/stats_app/stats_app/__init__.py | 0
.../opentuner/stats_app/stats_app/settings.py | 162 -
.../stats_app/stats_app/static/charts.css | 11 -
.../stats_app/stats_app/templates/charts.html | 41 -
.../opentuner/stats_app/stats_app/urls.py | 20 -
.../opentuner/stats_app/stats_app/views.py | 5 -
.../stats_app/stats_app/views/__init__.py | 0
.../stats_app/stats_app/views/charts.py | 67 -
.../opentuner/stats_app/stats_app/wsgi.py | 32 -
.../opentuner/tests/test_manipulator.py | 270 --
.../opentuner/tests/test_technique.py | 77 -
.../autotuner/opentuner/venv-bootstrap.py | 2611 -------------
.../autotuner/tuner_driver_src/__init__.py | 0
.../tuner_driver_src/backup/benchmarks.py | 493 ---
.../autotuner/tuner_driver_src/benchmarks.py | 599 ---
.../tuner_driver_src/buildRtConfig.py | 583 ---
.../tuner_driver_src/compareResults.py | 66 -
.../tuner_driver_src/compute_confs.py | 56 -
.../tuner_driver_src/error_sensitivity.py | 378 --
.../autotuner/tuner_driver_src/genPlots.py | 41 -
.../tuner_driver_src/global_paths.py | 12 -
.../tuner_driver_src/knob_pruning.py | 86 -
.../autotuner/tuner_driver_src/main_driver.py | 69 -
.../tuner_driver_src/pareto_curve.py | 313 --
.../tuner_driver_src/pareto_utils.py | 257 --
.../autotuner/tuner_driver_src/profiling.py | 26 -
.../autotuner/tuner_driver_src/remap.py | 291 --
.../tuner_driver_src/run_algo_tuner.py | 102 -
.../tuner_driver_src/run_algo_tuner2.py | 186 -
.../tuner_driver_src/run_autotuner.py | 403 ---
.../tuner_driver_src/run_devtime_tuner.py | 311 --
.../tuner_driver_src/run_ha_tuner.py | 52 -
.../tuner_driver_src/run_hs_tuner.py | 185 -
.../tuner_driver_src/run_install_tuner.py | 164 -
.../autotuner/tuner_driver_src/run_psnr.py | 143 -
.../tuner_driver_src/swing_selection.py | 304 --
.../tuner_driver_src/swing_selection2.py | 289 --
.../autotuner/tuner_driver_src/utils.py | 296 --
.../autotuner/tuner_driver_src/validation.py | 802 ----
677 files changed, 115513 deletions(-)
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_151.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_152.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_153.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_154.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_155.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_156.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_157.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_158.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_159.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_160.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_161.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_162.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_163.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_164.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_165.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_166.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_167.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_168.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_261.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_262.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_263.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_264.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_265.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_266.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_267.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_268.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_269.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp32_perf_fp32_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_151.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_152.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_153.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_154.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_155.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_156.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_157.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_158.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_159.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_160.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_161.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_162.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_163.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_164.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_165.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_166.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_167.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_168.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_42.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_44.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_46.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_261.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_262.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_263.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_264.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_265.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_266.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_267.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_268.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_269.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp32_perf_fp32_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_121.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_122.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_123.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_124.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_125.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_126.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_127.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_128.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_129.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_130.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_131.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_132.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_133.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_134.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_135.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_136.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_137.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_138.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_41.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_43.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_45.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_231.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_232.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_233.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_234.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_235.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_236.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_237.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_238.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_239.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_121.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_122.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_123.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_124.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_125.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_126.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_127.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_128.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_129.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_130.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_131.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_132.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_133.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_134.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_135.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_136.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_137.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_138.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_41.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_43.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_45.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_231.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_232.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_233.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_234.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_235.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_236.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_237.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_238.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_239.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_151.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_152.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_153.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_154.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_155.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_156.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_157.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_158.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_159.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_160.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_161.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_162.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_163.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_164.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_165.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_166.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_167.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_168.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_261.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_262.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_263.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_264.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_265.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_266.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_267.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_268.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_269.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp32_perf_fp32_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_151.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_152.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_153.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_154.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_155.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_156.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_157.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_158.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_159.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_160.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_161.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_162.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_163.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_164.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_165.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_166.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_167.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_168.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_261.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_262.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_263.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_264.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_265.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_266.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_267.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_268.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_269.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp32_perf_fp32_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_151.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_152.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_153.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_154.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_155.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_156.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_157.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_158.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_159.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_160.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_161.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_162.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_163.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_164.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_165.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_166.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_167.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_168.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_261.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_262.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_263.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_264.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_265.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_266.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_267.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_268.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_269.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp32_perf_fp32_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_151.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_152.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_153.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_154.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_155.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_156.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_157.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_158.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_159.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_160.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_161.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_162.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_163.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_164.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_165.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_166.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_167.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_168.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_261.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_262.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_263.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_264.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_265.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_266.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_267.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_268.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_269.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp32_perf_fp32_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_151.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_152.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_153.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_154.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_155.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_156.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_157.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_158.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_159.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_160.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_161.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_162.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_163.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_164.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_165.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_166.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_167.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_168.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_261.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_262.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_263.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_264.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_265.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_266.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_267.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_268.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_269.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp32_perf_fp32_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_151.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_152.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_153.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_154.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_155.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_156.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_157.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_158.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_159.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_160.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_161.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_162.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_163.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_164.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_165.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_166.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_167.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_168.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_261.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_262.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_263.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_264.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_265.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_266.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_267.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_268.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_269.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp32_perf_fp32_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_151.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_152.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_153.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_154.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_155.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_156.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_157.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_158.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_159.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_160.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_161.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_162.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_163.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_164.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_165.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_166.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_167.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_168.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_261.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_262.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_263.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_264.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_265.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_266.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_267.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_268.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_269.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp32_perf_fp32_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_151.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_152.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_153.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_154.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_155.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_156.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_157.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_158.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_159.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_160.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_161.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_162.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_163.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_164.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_165.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_166.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_167.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_168.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_261.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_262.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_263.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_264.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_265.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_266.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_267.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_268.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_269.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp32_perf_fp32_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_151.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_152.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_153.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_154.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_155.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_156.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_157.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_158.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_159.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_160.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_161.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_162.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_163.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_164.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_165.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_166.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_167.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_168.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_261.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_262.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_263.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_264.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_265.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_266.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_267.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_268.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_269.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp32_perf_fp32_120.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet/alexnet_valid_soc.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet2/alexnet2_valid_soc.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/resnet18/resnet18_valid_soc.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar10/vgg16_cifar10_valid_soc.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar100/vgg16_cifar100_valid_soc.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/README.md
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/alexnet2_cifar10/alexnet2_cifar10.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/alexnet_cifar10/alexnet_cifar10.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/alexnet_imagenet/alexnet_imagenet.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/lenet_keras/lenet_keras.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/mobilenet_cifar10/mobilenet_cifar10.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/resnet18_cifar10/resnet18_cifar10.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/resnet50_imagenet/resnet50_imagenet.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/vgg16_cifar10/vgg16_cifar10.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/vgg16_cifar100/vgg16_cifar100.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/vgg16_imagenet/vgg16_imagenet.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/alexnet2_cifar10/alexnet2_cifar10.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/alexnet_cifar10/alexnet_cifar10.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/alexnet_imagenet/alexnet_imagenet.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/lenet_keras/lenet_keras.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/mobilenet_cifar10/mobilenet_cifar10.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/resnet18_cifar10/resnet18_cifar10.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/resnet50_imagenet/resnet50_imagenet.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/vgg16_cifar10/vgg16_cifar10.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/vgg16_cifar100/vgg16_cifar100.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/vgg16_imagenet/vgg16_imagenet.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/autotuner/img_pareto_curve.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/alexnet_layers.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/alexnet_tensors.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/dev_knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/knobs1.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/alexnet2_layers.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/alexnet2_tensors.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/dev_knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/knobs1.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2_canny_hpvm/layers.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2_canny_hpvm/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet_imagenet/layer_composition.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet_imagenet/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet_imagenet/quant_ranges2.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/blend/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/blend/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/blend/tuner_conf_template.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/canny/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/canny/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/canny/tuner_conf_template.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/fft/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/fft/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/fft/tuner_conf_template.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/global_knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/global_knobs_dnn.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/global_knobs_old.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/dev_knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/knobs1.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/lenet_layers.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/lenet_tensors.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/dev_knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/knobs1.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/mobilenet_layer_comp.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/mobilenet_ops.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/dev_knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/knobs1.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/mobilenet_shallow_layer_comp.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/mobilenet_shallow_ops.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_torch/layers.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_torch/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/dev_knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/knobs1.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/knobs2.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/resnet_layers.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/resnet_tensors.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet18_torch/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet18_torch/resnet_layers.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/dev_knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/quant_ranges2.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/resnet50_layers.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/resnet50_tensors.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/dev_knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/knobs1.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/vgg16_layers.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/vgg16_tensors.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/dev_knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/knobs1.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/vgg16_layers.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/vgg16_tensors.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10_torch/layers.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10_torch/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/dev_knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/knobs.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/knobs1.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/op_cost.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/quant_ranges2.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/vgg16_layers.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/vgg16_tensors.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/.gitignore
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/AUTHORS.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/CHANGES.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/LICENSE.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/MANIFEST.in
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/README.md
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/__init__.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/accuracy_tuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/accuracy_tuner_piped.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/adddeps.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/algo_tuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/algo_tuner2.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/approxhpvm_tuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/devtuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/gettingstarted.md
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/measure_confidence.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/measure_confidence2.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/pareto_curve.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/promise_tuner2.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/promise_tuner3.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/promise_tuner_piped.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/psnr_tuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/select_top_results.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/utils.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/debian-packages-deps
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/Makefile
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/rtd-requirements.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/conf.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/index.rst
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/params.rst
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/techniques.rst
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/.gitignore
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/.gitignore
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/adddeps.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/.gitignore
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/matrixmultiply.cpp
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/raytracer.cpp
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/tsp_ga.cpp
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/cc_param_defaults.json
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/gccflags.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/gccflags_minimal.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/.gitignore
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/adddeps.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/bilateral_grid.cpp
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/bilateral_grid.settings
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/halide_blur.cpp
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/halide_blur.settings
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simple.cpp
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simple.settings
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simplest.cpp
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simplest.settings
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate.cpp
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate.settings
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/wavelet.cpp
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/wavelet.settings
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/halidetuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/timing_prefix.h
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/hpl/HPL.dat.mako
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/hpl/adddeps.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/hpl/hpl.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/README.md
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/adddeps.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/fceux-hook.lua
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/mario.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/.gitignore
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/README.md
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/adddeps.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/deps.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/download_benchmarks.sh
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/import_old_result.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/pbtuner.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/testwrapper.sh
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/py_api/adddeps.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/py_api/api_example.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/py_api/multiple_tuning_runs.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/.gitignore
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/adddeps.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/rosenbrock.makefile
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/rosenbrock.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/adddeps.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/att48_d.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/p01_d.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/p01_s.txt
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/tsp.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/#accuracy_tuner.py#
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/__init__.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/accuracy_tuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/accuracy_tuner_piped.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/adddeps.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/gettingstarted.md
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/measure_confidence.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/mmm_block.cpp
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/mmm_tuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/select_top_results.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/adddeps.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/cla_func.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/input_generator.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/problem_description.pdf
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/testwrapper.sh
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/unitary.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/gen-venv-bootstrap.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/install_reqs.sh
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/misc/livedisplay.gnuplot
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/misc/livedisplay.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/__init__.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/api.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/driverbase.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/#interface.py#
delete mode 120000 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/.#interface.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/__init__.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/driver.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/inputmanager.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/interface.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/resultsdb/__init__.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/resultsdb/connect.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/resultsdb/models.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/__init__.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/bandittechniques.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/composableevolutionarytechniques.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/differentialevolution.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/driver.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/evolutionarytechniques.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/globalGA.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/manipulator.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/metatechniques.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/objective.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/patternsearch.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/plugin.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/pso.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/simplextechniques.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/simulatedannealing.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/technique.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/tuningrunmain.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/__init__.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/adddeps.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/compactdb.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/dictconfig.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/stats.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/stats_matplotlib.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/optional-requirements.txt
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/requirements.txt
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/setup.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/manage.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/__init__.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/settings.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/static/charts.css
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/templates/charts.html
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/urls.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/views.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/views/__init__.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/views/charts.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/wsgi.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/tests/test_manipulator.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/tests/test_technique.py
delete mode 100755 hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/venv-bootstrap.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/__init__.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/backup/benchmarks.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/benchmarks.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/buildRtConfig.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/compareResults.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/compute_confs.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/error_sensitivity.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/genPlots.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/global_paths.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/knob_pruning.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/main_driver.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/pareto_curve.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/pareto_utils.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/profiling.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/remap.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_algo_tuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_algo_tuner2.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_autotuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_devtime_tuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_ha_tuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_hs_tuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_install_tuner.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_psnr.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/swing_selection.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/swing_selection2.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/utils.py
delete mode 100644 hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/validation.py
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/knobs.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/knobs.txt
deleted file mode 100644
index 1be6444417..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/knobs.txt
+++ /dev/null
@@ -1,28 +0,0 @@
-perf_fp16,120,Baseline
-perf_fp16,151,Col perf 50% offset=0
-perf_fp16,152,Col perf 50% offset=1
-perf_fp16,153,Row perf 50% offset=0
-perf_fp16,154,Row perf 50% offset=1
-perf_fp16,155,Col perf 33% offset=0
-perf_fp16,156,Col perf 33% offset=1
-perf_fp16,157,Col perf 33% offset=2
-perf_fp16,158,Row perf 33% offset=0
-perf_fp16,159,Row perf 33% offset=1
-perf_fp16,160,Row perf 33% offset=2
-perf_fp16,161,Col perf 25% offset=0
-perf_fp16,162,Col perf 25% offset=1
-perf_fp16,163,Col perf 25% offset=2
-perf_fp16,164,Col perf 25% offset=3
-perf_fp16,165,Row perf 25% offset=0
-perf_fp16,166,Row perf 25% offset=1
-perf_fp16,167,Row perf 25% offset=2
-perf_fp16,168,Row perf 25% offset=3
-samp_fp16,261,Samp 50% offset=0
-samp_fp16,262,Samp 50% offset=1
-samp_fp16,263,Samp 33% offset=0
-samp_fp16,264,Samp 33% offset=1
-samp_fp16,265,Samp 33% offset=2
-samp_fp16,266,Samp 25% offset=0
-samp_fp16,267,Samp 25% offset=1
-samp_fp16,268,Samp 25% offset=2
-samp_fp16,269,Samp 25% offset=3
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_120.txt
deleted file mode 100644
index f3e1be03b6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_120.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,65.2289,635838
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.206437,13534.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.171279,13528.5
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,257.082,2.53801e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.220024,13871.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.191798,13871.6
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.78132,31392.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,79.5717,772941
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.208738,14141.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.179234,14145.7
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,132.214,1.32889e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.213516,14286.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.178101,14283.1
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,2.07068,28582.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,44.624,449041
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.199375,14430.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.155295,14431.9
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,75.4301,795035
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.205183,14622.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.157695,14624.5
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.24782,14616.9
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.20811,15614.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.164572,14641.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.996439,15121.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_151.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_151.txt
deleted file mode 100644
index 891ef46482..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_151.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,58.0027,548113
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.189544,13050.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.17103,13052
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,128.221,1.22437e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.217308,13335.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.17494,13336.5
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.54381,29480.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,48.9069,459648
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.213516,13612.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.19278,13600.7
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,66.2285,651716
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.205122,13690.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.174924,13694.9
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.83244,27400
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,27.4047,267152
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.190466,13844.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.149922,13851.3
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,39.169,396057
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.186793,13947
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.147929,13947.8
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.12431,14343.1
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.09979,14347.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.145526,13961.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.939092,13969.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_152.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_152.txt
deleted file mode 100644
index 0e1dc66146..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_152.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,57.9254,550132
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.188185,13098.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.168668,13100.2
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,128,1.22987e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.217096,13371.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.17532,13354.5
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.45412,26727.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,48.745,462457
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.20812,13623.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.176805,13623.3
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,66.1231,650723
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.204978,13775.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.185167,13766.3
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.89592,25380.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,27.2569,267804
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.187973,13877.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.146757,13869.7
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,39.3148,398486
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.18485,13995
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.147977,14000.1
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.16044,13997.9
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.14272,14765.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.154716,14008.5
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.967361,14016.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_153.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_153.txt
deleted file mode 100644
index 211011c1c8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_153.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,52.3167,486159
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.200114,12880.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.175215,12886.3
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,122.54,1.15844e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.212952,13161.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.179759,13157.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.5318,27617.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,47.8415,444086
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.203436,13433.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.170578,13433.7
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,65.082,630988
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.213602,12821.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.166886,12826.1
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.80028,25660.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,27.811,255106
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.186975,13659.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.158729,13644.4
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,39.012,391835
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.183039,13762.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.147983,13761.3
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.16698,13761.9
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.16997,14528.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.16397,13782.7
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.946711,13787
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_154.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_154.txt
deleted file mode 100644
index 4e4718b997..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_154.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,52.9971,496465
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.192665,12930.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.170856,12926.5
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,122.764,1.15342e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.215928,13205.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.175704,13204.8
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.49985,26407.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,47.9188,445564
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.204233,13461.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.177759,13459.2
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,64.6805,628375
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.210421,13606
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.173551,13591.1
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.87541,26494.5
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,27.3168,264238
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.190476,13723.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.150409,13732.8
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,38.9734,395354
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.185871,13830.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.147122,13835.5
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.13866,13834.3
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.1186,14601.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.15205,13856.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.938171,13864.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_155.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_155.txt
deleted file mode 100644
index 394d24cf90..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_155.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,64.8669,622301
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.20949,13239.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.176294,13243.4
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,162.006,1.5735e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.220239,13511.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.178239,13511.2
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.75893,29646.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,57.036,545033
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.208872,13815.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.181833,13813.6
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,84.4746,841306
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.216168,13970.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.181836,13976.3
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,2.06704,27968.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,32.4425,322697
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.199647,14092.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.160076,14084.6
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,48.8451,506372
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.196623,14181.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.159896,14172.9
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.22106,14178.1
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.21282,15029.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.16867,14202.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.963022,14204.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_156.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_156.txt
deleted file mode 100644
index 03318bf3ae..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_156.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,63.1272,608633
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.203714,13382.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.179644,13369.4
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,159.062,1.56322e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.219154,13650.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.176476,13654.5
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.6318,29500.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,58.7054,568593
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.212313,13955.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.186098,13945.6
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,88.1808,883262
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.211765,14119.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.179279,14113.7
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.87673,27502.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,31.7267,314842
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.199883,14261.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.168866,14253.9
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,47.7887,497470
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.194706,14369.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.161791,14355.9
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.18783,14357
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.19429,15205.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.174949,14370.7
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,1.00356,15533.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_157.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_157.txt
deleted file mode 100644
index 492feb3022..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_157.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,63.9795,627055
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.207468,13560.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.172213,13546.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,164.125,1.62846e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.21318,13828.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.174274,13824.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.56693,28880.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,58.2621,572924
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.205416,14092.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.182895,14103
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,88.3811,890106
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.213208,14246.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.179445,14255.8
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.93471,28531.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,34.5703,350902
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.201976,14415.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.164367,14414.3
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,53.1299,560793
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.195474,14557
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.158092,14560
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.15914,14565.3
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.14588,15448.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.158146,14582.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.947485,15022.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_158.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_158.txt
deleted file mode 100644
index 55b4439593..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_158.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,59.0766,557171
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.203615,13064.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.17004,13070.4
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,156.515,1.48722e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.242274,13321.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.179781,13329.7
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.60261,28731.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,55.4282,523293
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.211531,13627
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.179788,13630.8
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,82.562,806140
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.20909,13773.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.179433,13771.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.89329,27557.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,32.1691,313581
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.201173,13897.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.15526,13902.6
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,48.4451,495873
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.200427,14025.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.15853,14018.5
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.21556,14018.5
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.20271,14854.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.135273,14049.7
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.970621,14465.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_159.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_159.txt
deleted file mode 100644
index c1cd632397..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_159.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,59.2911,555019
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.201084,12973.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.170233,12969.6
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,156.931,1.49121e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.217797,13251.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.174917,13245.8
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.71397,29190.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,57.8418,542994
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.21652,13546.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.178658,13552.4
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,87.9539,849490
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.217259,13704.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.179077,13693.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,2.03865,27430.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,32.5739,308590
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.199641,13812
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.161615,13811.6
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,49.129,497072
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.208169,13936
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.161935,13937.9
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.28389,13935.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.25362,14764.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.163733,13937.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,1.00077,14665.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_160.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_160.txt
deleted file mode 100644
index 6ee9d4cd79..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_160.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,58.8135,556846
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.203772,13190.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.176348,13196.6
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,158.962,1.53725e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.219541,13440.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.178588,13431.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.68237,28875.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,56.3609,533809
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.210882,13724.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.176328,13726.1
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,86.1982,845559
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.214652,13882.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.173688,13888.1
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.96952,27069.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,34.5895,340452
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.200648,14023.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.159231,14025
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,52.939,546006
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.197173,14157.7
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.150853,14157.7
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.16124,14167.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.17195,15013
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.157372,14168.5
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.983837,14180.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_161.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_161.txt
deleted file mode 100644
index 3b49fae064..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_161.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,65.8243,634762
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.205215,13383
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.179957,13375.5
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,174.316,1.70456e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.219689,13640.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.182165,13619.9
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.68031,29935.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,61.4199,594851
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.215746,13926
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.185436,13916.6
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,93.149,925082
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.213054,14081
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.180697,14075.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.99828,28166.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,35.2346,352384
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.204687,14205.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.157567,14209.1
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,53.4125,557191
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.20316,14374.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.159049,14361.3
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.19671,14363.3
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.18429,15228.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.17158,14372.9
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.963239,14813.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_162.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_162.txt
deleted file mode 100644
index e1a046ef01..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_162.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,66.1421,634983
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.21741,13337.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.185688,13331.4
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,176.119,1.69722e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.227336,13574
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.188882,13570.3
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.6811,30306.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,61.7431,583447
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.219186,13881.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.181445,13877.7
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,93.914,932514
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.219256,14007.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.1794,13997.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.93367,28014.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,35.3477,350762
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.208718,14161.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.161401,14163.3
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,53.9521,560676
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.204236,14322.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.157525,14318.9
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.23967,14322.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.22098,15186.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.169871,14326.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.978561,14755.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_163.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_163.txt
deleted file mode 100644
index ad8ef25634..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_163.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,66.2346,634503
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.209573,13306.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.174533,13306.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,176.314,1.70236e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.234539,13560.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.184031,13556.8
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.71175,30262.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,61.7066,582837
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.221103,13124.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.177237,13110.9
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,94.1784,928952
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.219455,13997.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.18468,13999.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.97442,28016.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,35.4772,352075
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.209503,14157.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.161119,14161.6
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,54.1191,562078
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.207996,14315.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.160258,14309.8
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.23714,14309.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.21279,15183.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.175522,14321.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.977623,14744.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_164.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_164.txt
deleted file mode 100644
index 3461775f35..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_164.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,65.9398,633363
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.205851,13291
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.172959,13287.3
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,174.804,1.69553e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.220789,13560.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.177087,13564.6
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.81027,31205.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,61.5282,590104
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.209931,13849.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.18996,13845.8
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,93.4214,924284
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.21342,13978.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.184242,13988.7
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,2.0622,27996.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,34.9337,342861
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.202178,14110.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.160732,14112.6
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,53.8579,556438
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.199122,14251.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.157836,14257.4
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.22002,14266.9
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.23049,15130.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.172575,14268.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,1.02197,15447.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_165.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_165.txt
deleted file mode 100644
index a78c775332..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_165.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,61.5774,582474
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.210591,13093.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.16853,13084.1
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,171.588,1.6167e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.230504,13322.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.176787,13316.6
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.79893,31381.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,60.6244,573649
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.223733,13607.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.179596,13617.2
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,92.717,894994
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.221378,13754.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.177762,13744.8
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,2.00882,27528.1
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,34.7619,342294
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.204523,13901.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.15829,13901.8
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,53.3066,543043
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.203224,14037
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.161542,14042.7
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.26148,14044.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.21381,14900.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.171011,14056
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.964679,14479
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_166.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_166.txt
deleted file mode 100644
index 8c700396a9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_166.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,61.7188,579843
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.205929,13080.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.170476,13046.2
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,170.797,1.62614e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.221976,13303.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.172012,13305.2
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.66512,28575.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,60.6009,571260
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.212117,13600.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.198482,13602
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,92.7343,897664
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.214779,13754.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.180697,13735.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.93166,27486
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,35.0139,341225
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.198821,13911.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.15717,13913.2
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,53.8398,547490
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.200568,14061.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.154255,14051.6
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.22431,14057.3
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.21994,14922.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.143932,14068.7
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.950046,14491.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_167.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_167.txt
deleted file mode 100644
index ed26859f07..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_167.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,60.3008,566150
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.212953,13192.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.16919,13192.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,168.206,1.62008e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.216709,13429.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.176031,13418.5
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.62383,28235.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,59.0124,557003
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.2075,13707.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.183634,13707.1
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,90.5798,881313
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.214488,13884.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.174665,13865.5
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.87406,27761.5
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,34.5362,338970
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.197125,14010.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.156981,14014.1
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,53.1529,543496
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.199279,14147.7
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.152859,14148
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.22611,14878.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.14545,15004
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.159065,14156.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.967012,15316.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_168.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_168.txt
deleted file mode 100644
index b9ea435a19..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_168.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,60.0049,566283
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.198763,13133.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.171007,13131.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,168.986,1.60868e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.214533,13381.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.17733,13378.1
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.72244,29920.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,59.1517,556478
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.209778,13645.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.183487,13653.4
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,91.8964,875586
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.218393,13825.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.178373,13823.5
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,2.097,27656.5
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,34.1944,334924
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.199813,13966.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.157509,13945.6
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,52.856,538205
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.201836,14090.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.156297,14091.1
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.21402,14100.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.20897,14950.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.173667,14119.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.958647,14548.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_261.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_261.txt
deleted file mode 100644
index 4ceca9d3b4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_261.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,29.9901,283881
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.178492,13190.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.172648,13198.3
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,167.328,1.63105e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.213493,13541.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.182031,13537.9
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.6555,27766.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,49.9733,474729
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.197967,13766.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.176799,13753.2
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,84.2239,851536
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.197804,13933.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.173186,13933.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,2.09551,27874.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,30.1239,303528
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.185539,14072.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.148063,14057.6
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,55.3267,563272
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.177714,14210.3
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.151116,14212.2
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.13263,14214.1
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.17752,15039.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.157078,14227.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.949831,14229.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_262.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_262.txt
deleted file mode 100644
index 594b3c6cb4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_262.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,30.6343,291631
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.175893,13148.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.16885,13150.4
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,167.55,1.62919e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.208101,13511.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.181103,13511.3
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.69895,31258.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,49.6537,469423
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.20214,13730.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.183407,13074.4
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,84.4516,851268
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.197845,13902.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.175394,13893.3
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,2.03863,27061.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,30.1611,302847
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.179497,14023.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.157608,14019.6
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,55.0979,577538
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.17278,14181.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.155212,14190.7
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.15995,14196.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.15144,15014.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.170053,14202.1
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.927195,14610
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_263.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_263.txt
deleted file mode 100644
index daac5adf34..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_263.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,38.5323,380280
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.183311,13716.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.166876,13706.8
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,246.805,2.50131e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.207515,14125.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.180767,14122
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.67711,29548.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,68.6879,684023
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.201026,14390.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.175177,14388.5
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,124.545,1.31127e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.20485,14513
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.181132,14520.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.99895,29075.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,42.71,444867
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.185577,14691.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.155711,14683.8
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,79.6438,863928
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.198735,14844.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.155318,14831.4
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.21923,14833
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.23007,15807.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.179961,14844.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.944167,15336.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_264.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_264.txt
deleted file mode 100644
index 8e9d08eb04..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_264.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,37.9249,377238
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.183003,13840.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.172521,13842.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,244.866,2.50419e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.210677,14226.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.189071,14222.7
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.59858,28454.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,68.3239,682954
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.198991,14506.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.180629,14515.9
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,123.521,1.30914e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.203701,14661.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.181269,14656
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.89143,29323.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,42.2852,445845
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.187503,14802.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.154306,14798.8
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,79.0359,862979
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.188383,14951.7
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.158735,14940.1
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.20675,15447.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.07758,15450.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.140345,14971.9
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.979124,15459.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_265.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_265.txt
deleted file mode 100644
index a9a49265ef..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_265.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,38.6073,381890
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.183922,13657.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.167353,13655.2
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,247.161,2.49436e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.212076,14055.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.184153,14062.7
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.74619,30741.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,68.3499,680826
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.202505,14325.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.180396,14329.4
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,124.93,1.28676e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.204786,13799.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.175125,13782.5
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,2.03242,28261.1
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,42.9068,448281
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.191074,14629
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.153951,14617.2
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,79.7328,862210
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.191826,14783.3
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.156604,14787.1
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.19046,14789
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.23317,15763.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.1715,14808
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.950132,15288.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_266.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_266.txt
deleted file mode 100644
index 17d5963841..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_266.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,34.6388,330159
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.1851,13519.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.163324,13525.3
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,211.763,2.10457e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.208022,13853
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.172867,13851.1
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.61003,28957
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,59.4769,581575
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.20364,14117.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.186194,14113.7
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,107.233,1.09327e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.203724,14290.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.183164,14294.8
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,1.95363,28612.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,39.2333,403131
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.194492,14459.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.154262,14446.4
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,72.6672,778814
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.191938,14625.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.155673,14618.1
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.1725,14620.1
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.17531,15526
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.168101,14628
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.942702,15085.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_267.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_267.txt
deleted file mode 100644
index 2a49166de8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_267.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,35.3669,338192
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.189141,13532.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.176591,13521
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,211.87,2.10409e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.21045,13851.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.180578,13849.5
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.73284,29157.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,58.9372,579132
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.203445,14140.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.1863,14151.6
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,107.227,1.0916e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.207173,14277.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.177253,14279.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,2.05931,28583.5
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,39.4378,402404
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.197631,14460
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.161132,14444.7
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,72.8899,772755
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.191496,14641.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.155829,14628.5
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.16222,14634.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.18754,15547.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.165084,14642.1
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.931806,15103.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_268.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_268.txt
deleted file mode 100644
index 7ffc46b3fa..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_268.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,35.1044,343285
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.185285,13540.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.165138,13536.5
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,211.784,2.10554e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.216373,13853.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.180239,13853.2
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.7821,29895.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,58.7825,573791
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.211295,14119.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.17844,14119.2
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,107.332,1.09268e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.208469,14294.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.177596,14296.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,2.08082,28605.1
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,39.3551,404975
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.200089,14439.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.156082,14434
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,72.6197,776642
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.186354,14622.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.155119,14630.5
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.16887,14638.5
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.19008,15550.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.165605,14646.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.942276,14654
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_269.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_269.txt
deleted file mode 100644
index dc9228cefa..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_269.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,35.9637,343026
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.185477,13412.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.169158,13406.6
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,213.539,2.10433e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.213954,13732.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.185995,13738.6
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,2.77092,30824.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,59.257,574879
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.203871,14027.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.178715,14037.1
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,108.244,1.09504e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.211721,14189.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.178908,14195.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,2.0528,28416.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,39.7267,403494
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.193717,14337
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.160898,14335
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,73.0412,773266
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.192258,14537.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.158802,14528
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,1.22562,14522.3
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.34877,16977.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.179394,14555.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.978526,15010.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp32_perf_fp32_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp32_perf_fp32_120.txt
deleted file mode 100644
index 46c04e6d4c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp32_perf_fp32_120.txt
+++ /dev/null
@@ -1,72 +0,0 @@
-Conv1,110.867,937287
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.2524,11951.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.148963,11939.5
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,405.61,3.38379e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.52405,14207.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.314892,12284.3
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,7.49997,70147.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,146.869,1.24647e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.22959,12498.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.167686,12501
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,242.23,2.12642e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.212758,12732
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.139776,12735.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,4.07526,42078.1
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,70.2196,641128
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.201587,12969.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.157597,12973.3
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,121.098,1.12443e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.208646,13218.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,0.140093,13224.5
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,2.33548,27133.9
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,0.665967,13777.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.132666,13251.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,0.926642,14597.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_120.txt
deleted file mode 100644
index a9c2b75d06..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_120.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,361.205,887540
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.8371,25251
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.89826,25346
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1282.07,4.13212e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.66204,37762
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.98363,37781
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,17.6103,90940
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,420.625,1.73415e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.85413,41830
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.71096,41925
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,1434.84,5.80943e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.26222,43549
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.87823,43588
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,14.8755,93369
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,230.687,1.08731e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.81058,45573
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.70434,45611
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,386.122,1.88489e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.91599,48344
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.88405,48344
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,10.1237,78290
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,8.99616,67944
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.04792,48554
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,9.67766,66936
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.80403,48440
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.95634,48440
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,350.798,1.77138e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,68.4447,356768
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,29.0155,146006
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,773.778,3.92759e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,447.052,2.23187e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,422.442,2.2019e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,56.5507,316578
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.3619,182871
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.7478,75963
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,44.8441,253942
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_151.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_151.txt
deleted file mode 100644
index 0a3f9b79c8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_151.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,341.714,1.24364e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.01032,35383
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.81735,35419
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,672.798,2.67555e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.64978,40633
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.07404,40741
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,15.7876,81823
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,246.262,1.05848e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.36662,43206
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.02646,43238
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,344.014,1.55449e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.53103,45471
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.16501,45487
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,11.9697,72632
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,155.126,744177
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.07947,46898
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.29868,46819
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,222.859,1.1246e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.1875,48225
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.35925,48221
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,10.3547,86737
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.5923,58118
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.55215,48415
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,9.25993,58146
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,3.24436,48244
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.13347,48148
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,358.153,1.78208e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,70.6011,365656
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,29.3887,149498
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,380.734,1.91727e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,258.012,1.30302e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,254.065,1.31266e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,47.5991,258579
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,28.0997,160667
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,8.53453,58494
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,41.037,223342
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_152.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_152.txt
deleted file mode 100644
index 0e370a50e6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_152.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,338.653,1.22516e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.49818,35363
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.7734,35383
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,673.747,2.65805e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.72712,40710
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.994615,40797
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,16.2806,81620
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,241.163,1.03265e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.17343,43160
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.854522,43232
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,342.155,1.50794e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.54565,45654
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.22057,45727
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,11.3366,69365
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,150.324,720358
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.47724,46811
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.32594,46888
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,219.824,1.09715e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.71662,48366
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.02462,48404
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,10.3584,78167
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.2894,61905
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.67705,48499
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,9.31139,63276
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.16947,48345
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.93852,48117
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,364.322,1.77585e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,70.1004,353948
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,29.256,151166
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,370.56,1.8226e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,255.449,1.25407e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,251.415,1.29241e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,48.623,261943
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,29.175,170000
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,8.85507,63150
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,41.1791,231371
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_153.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_153.txt
deleted file mode 100644
index 2f4fe3fc94..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_153.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,313.584,1.14369e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.9961,35442
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.75755,35461
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,648.569,2.56959e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.02472,40420
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.16678,40466
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,15.6438,85388
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,240.467,1.02632e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.51647,42923
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.924186,43115
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,339.393,1.52091e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.51628,45120
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.13145,45079
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,13.4161,90533
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,149.177,687919
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.61439,46285
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.16799,46401
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,221.957,1.11058e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.56843,47735
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.70929,47772
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,10.9723,90939
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,8.98131,51771
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.62773,47906
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,11.0935,75974
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.54774,47622
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,6.20792,47603
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,403.556,2.01366e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,97.3704,495605
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.8325,213694
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,557.298,2.7772e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,362.925,1.7864e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,367.019,1.81001e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,70.1719,382270
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,41.0814,234379
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,15.901,102022
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,60.4126,330986
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_154.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_154.txt
deleted file mode 100644
index 765c4ab9e3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_154.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,315.379,1.16143e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.73517,35843
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.22104,35862
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,651.303,2.5903e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.74155,40626
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.729368,40639
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,15.1605,81426
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,236.06,1.00358e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.20156,43136
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.795643,43155
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,337.277,1.50712e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.64914,45323
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.0735,45339
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,12.5338,71911
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,152.238,733302
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.40817,50053
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.34367,46552
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,218.448,1.07146e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.36034,48079
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.2093,48098
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.3163,91399
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,8.68467,57846
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.90696,48171
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,9.91817,71589
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.82864,48004
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,6.80968,47907
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,435.426,2.16466e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,108.727,553915
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,41.8594,233214
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,616.852,3.0909e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,401.057,1.9871e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,409.951,2.06242e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,78.7003,426478
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,45.3634,257888
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,16.0649,99414
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,65.2346,361740
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_155.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_155.txt
deleted file mode 100644
index 495b8e72ce..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_155.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,367.586,1.37544e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.70513,36722
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.05311,36739
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,835.213,3.43808e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.499,42786
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.972346,42819
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,14.9356,85825
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,288.55,1.29369e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.40882,45555
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.997081,45555
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,428.832,2.03842e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.68434,48161
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.6868,48192
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,14.4762,96820
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,188.781,981227
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.77355,49647
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.3157,49647
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,292.4,1.57032e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.81982,51286
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.37624,51324
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.2112,102762
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,10.4728,81945
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.74428,51381
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,9.82457,61741
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,3.3466,51039
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.82338,51020
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,341.701,1.83633e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,65.475,355486
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,27.9864,148305
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,459.432,2.42759e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,300.123,1.56346e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,297.714,1.56658e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,48.3095,264674
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,29.424,173204
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,12.1637,86228
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,44.1133,252868
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_156.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_156.txt
deleted file mode 100644
index f41af92f3c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_156.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,365.82,1.38229e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.86135,37012
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.2317,37012
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,835.167,3.46916e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.72744,43081
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.07906,43094
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,16.1615,86240
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,300.757,1.32601e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.3894,45882
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.988184,45920
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,453.354,2.17372e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.96965,48616
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.89493,48627
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,16.5878,101588
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,194.167,1.02167e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,4.44685,53990
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.81678,50087
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,282.513,1.52215e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,3.19722,51663
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.4803,51701
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,10.7573,82207
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.65599,66535
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.10187,51739
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,9.84134,62346
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.02711,51723
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.69407,51495
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,343.65,1.85844e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,66.7983,358767
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,28.0086,154621
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,466.629,2.489e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,304.943,1.59106e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,302.615,1.58389e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,49.4705,287963
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,30.9342,189526
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,11.8809,87008
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,46.3725,260890
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_157.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_157.txt
deleted file mode 100644
index 46cb2c5848..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_157.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,370.217,1.40762e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.22554,37220
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.92616,37249
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,856.722,3.56485e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.6708,43476
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.46693,43508
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,16.5105,91535
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,305.956,1.41095e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.53477,46248
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.39539,46204
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,455.469,2.22189e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.85246,49098
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.63118,49153
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,17.2593,107081
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,206.479,1.08538e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.75124,50449
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.87596,50449
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,302.047,1.65556e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.65649,52282
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.9676,52301
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.8465,99367
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.95874,72873
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.35496,52436
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,9.91958,57749
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.3887,52113
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.9476,52113
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,345.747,1.89194e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,67.0786,367483
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,28.309,155718
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,481.635,2.59297e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,314.62,1.67524e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,314.968,1.68945e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,51.4902,300879
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.5453,196186
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,12.0044,82161
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,44.6534,251589
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_158.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_158.txt
deleted file mode 100644
index ef02807f8f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_158.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,345.457,1.28233e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.68836,35937
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.07205,35975
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,811.279,3.27855e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.72946,41898
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.27244,41949
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,16.0697,88412
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,285.874,1.26548e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.96683,44618
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.09865,44676
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,425.288,1.99292e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.96677,47315
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.17589,47370
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,15.4633,99614
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,186.796,953798
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.48859,48565
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.17202,48603
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,287.804,1.50853e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.65,50170
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.2331,50189
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.7266,89977
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.34201,65748
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.36465,50189
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,8.66339,64282
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.09815,49960
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,4.89225,49960
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,361.102,1.888e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,70.1473,377537
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,29.202,154986
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,495.785,2.58001e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,320.547,1.64075e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,320.266,1.66296e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,51.7552,286193
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,30.4092,175602
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.1634,75129
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,46.3624,254708
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_159.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_159.txt
deleted file mode 100644
index a1dfc26c51..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_159.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,339.094,1.26048e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.96416,36128
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.77992,36163
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,813.76,3.31365e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.64994,42067
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.07708,42064
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,14.8503,84343
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,293.589,1.31011e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.3999,44879
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.05596,44834
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,448.595,2.107e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.3315,47603
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.04386,47621
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,14.6064,94214
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,193.015,1.00842e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.50711,48920
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.48661,48920
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,289.684,1.54004e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.68606,50297
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.25317,50316
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.0975,100976
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.96732,70276
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.2283,50564
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,9.2103,55553
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.29655,50335
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.29321,50258
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,351.462,1.85714e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,71.227,389308
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,29.2131,161340
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,501.805,2.62186e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,324.086,1.66061e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,324.179,1.6786e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,52.8531,301753
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.8956,190928
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,12.7146,90426
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,48.6169,275573
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_160.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_160.txt
deleted file mode 100644
index 71ac477207..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_160.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,344.736,1.29476e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.74231,36242
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.35918,36299
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,835.655,3.42441e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.09224,42331
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.70002,42246
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,24.1042,124515
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,305.48,1.39635e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.33278,45027
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.73698,45038
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,448.089,2.15419e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.05134,47781
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.79851,47789
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,17.4017,110314
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,201.163,1.05746e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.5874,49107
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.32724,49126
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,298.623,1.58674e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.25227,50638
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.11102,50752
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.3346,91373
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.4071,60007
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.79112,50828
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,10.504,60122
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.05558,50600
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.80988,50603
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,352.716,1.87049e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,70.9986,377044
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,29.693,162274
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,502.505,2.64698e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,317.991,1.60963e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,319.62,1.6465e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,48.7611,272981
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.3448,186827
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,12.0792,81356
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,49.0611,267293
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_161.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_161.txt
deleted file mode 100644
index 71f39ee42f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_161.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,378.483,1.44377e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.16813,37354
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.25211,37202
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,905.01,3.78967e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.41608,43519
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.68306,43573
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,19.4582,105246
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,331.99,1.57104e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.31531,46344
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.26824,46361
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,485.924,2.40972e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.85947,49214
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.37848,49150
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,20.9687,122884
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,197.661,1.06623e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.63787,50453
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.2229,50453
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,296.445,1.62111e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.2027,52149
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.28997,52146
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.8256,104486
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.95045,78259
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.4269,52341
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,10.6427,68164
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,3.70788,52034
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.69813,51958
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,346.748,1.8922e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,68.4736,377821
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,28.9813,156405
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,536.616,2.89048e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,344.142,1.82033e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,344.951,1.83101e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,50.2737,285095
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,33.3015,206213
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,12.5197,92863
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,46.8158,273128
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_162.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_162.txt
deleted file mode 100644
index d5b585eb36..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_162.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,376.051,1.43644e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.85018,37259
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.19583,37275
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,903.953,3.79981e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.36888,43749
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.17659,43743
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,19.4205,113593
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,333.107,1.57337e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,3.22801,46684
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.21371,46628
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,481.782,2.40292e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.91195,49343
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.08034,49362
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,19.4,109107
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,197.676,1.04809e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.03656,50768
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.57624,50806
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,298.469,1.64777e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,3.01322,52469
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.26031,52393
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.8955,105111
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.20323,58051
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,3.18967,56831
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,10.956,72962
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.10368,52316
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,6.02162,52242
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,353.033,1.93595e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,67.0839,369030
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,28.8589,156817
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,529.37,2.85811e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,339.888,1.78149e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,338.989,1.80672e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,47.7108,271072
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.8336,202963
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,13.1019,88026
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,46.4985,278681
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_163.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_163.txt
deleted file mode 100644
index 51a8fab32a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_163.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,376.623,1.43041e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.2712,37335
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.29774,37259
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,906.73,3.79983e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.93634,43709
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.93493,43740
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,20.1087,113270
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,340.112,1.62416e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,3.40813,46586
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.31909,46626
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,482.654,2.42242e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.32494,49383
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.15611,49380
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,20.2263,123328
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,204.243,1.10377e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.71524,50639
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.19861,50639
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,292.261,1.60557e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.17489,52530
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.91954,52530
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.6709,105099
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.89142,72967
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.52353,52569
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,9.81087,62950
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.02868,52377
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,4.7711,52300
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,345.339,1.88692e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,65.0372,368276
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,28.1288,151188
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,514.73,2.7732e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,329.902,1.73695e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,330.174,1.75558e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,47.4751,270664
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,30.4123,187537
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.2215,83300
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,46.3845,273875
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_164.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_164.txt
deleted file mode 100644
index 29db7c0b54..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_164.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,379.306,1.42757e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.64497,37278
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.88347,37297
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,900.541,3.74501e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.1469,43521
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.54092,43559
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,18.5551,100862
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,329.675,1.53371e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.86776,46220
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.98069,46501
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,479.022,2.36315e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.10373,49269
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.92033,49278
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,18.7268,119116
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,204.671,1.0882e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.35297,50563
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.81918,50620
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,293.578,1.61435e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.22158,52268
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.11317,52301
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.4222,88604
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.56134,68623
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,3.0971,52511
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,8.69792,57728
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.90547,52226
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.07817,52155
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,353.537,1.93273e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,68.0307,372313
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,28.517,156425
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,547.356,2.93903e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,352.325,1.86501e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,352.467,1.90256e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,51.5187,301084
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,30.5184,185959
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.2667,66566
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,43.6303,263236
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_165.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_165.txt
deleted file mode 100644
index 3792531eb6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_165.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,353.773,1.32322e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.4627,36510
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.8332,36567
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,878.627,3.57603e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.05928,42794
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.75336,42851
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,17.1158,93353
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,316.068,1.43626e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.90983,45593
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.22248,45720
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,475.687,2.29564e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.32846,48366
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.68287,48382
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,19.1992,110644
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,199.562,1.04256e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.51739,49706
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,3.51719,53685
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,295.002,1.59147e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.27073,51350
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.21199,51350
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,12.6138,102966
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.33603,60989
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.8852,51521
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,10.8473,56777
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.30992,51375
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.90882,51280
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,344.155,1.85001e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,62.0464,346209
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,26.5284,148623
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,485.063,2.57151e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,317.001,1.59544e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,313.651,1.6023e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,45.7252,259752
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,27.9466,162611
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.16157,70558
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,40.4661,232799
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_166.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_166.txt
deleted file mode 100644
index 4e7dee2df5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_166.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,352.972,1.3171e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.25133,36683
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.45931,36702
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,885.443,3.58108e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.38741,42659
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.3844,42771
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,19.8063,110988
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,328.712,1.51501e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.41333,45436
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.27317,45524
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,477.967,2.32662e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.2292,48092
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.84405,48203
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,19.5186,111310
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,194.861,1.01432e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.30804,53532
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.18683,49610
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,296.149,1.59912e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.56193,51248
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.39044,51305
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.6294,97584
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.35232,61017
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.95976,51534
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,10.3614,66179
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.53209,51195
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,6.12908,51175
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,357.974,1.90505e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,70.0006,380612
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,29.5305,157813
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,544.245,2.8828e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,349.204,1.81432e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,346.818,1.81672e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,51.9549,306831
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.7948,203765
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,13.1344,91337
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,46.9918,268839
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_167.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_167.txt
deleted file mode 100644
index 8f48476be1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_167.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,350.949,1.31074e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.10317,36453
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.50366,36453
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,878.849,3.60669e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.85388,42579
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.69125,42598
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,19.9735,106383
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,325.095,1.50842e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.37345,45411
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.45422,45408
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,472.051,2.28945e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.13227,48127
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.41281,48104
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,19.9196,120940
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,194.514,1.01048e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.62532,49456
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.35806,49396
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,296.671,1.5972e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.58171,51020
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.92363,51134
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.0122,102384
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,10.6179,76199
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.00952,51154
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,10.6804,66734
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.01533,50910
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,6.19199,50682
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,349.427,1.86051e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,70.2114,389789
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,29.1434,157984
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,550.625,2.90433e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,354.709,1.85793e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,352.117,1.8427e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,48.5219,270571
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,29.9688,172978
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.4235,71395
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,46.5118,258348
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_168.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_168.txt
deleted file mode 100644
index c2573a1150..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_168.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,350.029,1.32438e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.79413,36356
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.08987,36432
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,876.486,3.58641e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.02325,42661
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.5181,42756
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,18.5093,101688
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,316.926,1.45832e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.64824,45388
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.77112,45462
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,470.364,2.27227e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.93378,48096
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.78824,48147
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,21.2716,134999
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,190.053,999012
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.09608,49460
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.13989,49533
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,295.297,1.59056e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,3.03483,51195
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.25576,51135
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.7841,97299
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,10.2158,76672
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.06478,51401
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,10.7958,72279
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.4984,56242
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,6.26824,51008
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,344.484,1.85136e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,67.526,370185
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,28.0443,147920
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,525.1,2.77217e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,338.872,1.77491e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,338.019,1.80635e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,47.0941,254613
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,27.9775,173259
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.21542,60441
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,41.7087,227963
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_42.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_42.txt
deleted file mode 100644
index e8c305acc3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_42.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,373.814,1.50342e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.2666,39127
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.80386,39069
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1320.04,5.89533e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.02645,47995
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.67969,48105
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,19.2226,119613
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,414.702,2.13905e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.43845,51060
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.92418,50984
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,679.246,3.67719e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.11397,54118
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.50782,54156
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,20.3157,136255
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,249.499,1.46504e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.95217,55715
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.73291,55734
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,403.468,2.41794e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.74872,57540
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.27275,57578
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.5194,115175
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.8221,86147
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.35854,57635
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,11.5996,103494
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.49869,57213
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.70418,57021
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,351.842,2.1004e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.4989,328936
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,28.7613,164526
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,794.891,4.65583e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,408.553,2.31782e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,409.279,2.35626e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,50.8358,322330
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,26.4958,168790
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,14.8298,101643
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,51.1146,335945
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_44.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_44.txt
deleted file mode 100644
index 3dfc825f85..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_44.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,373.04,1.49763e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.97739,38877
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.75301,38877
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1318.17,5.84816e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.5918,47981
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.52329,48175
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,17.0323,106123
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,405.754,2.06704e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.85906,50977
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.29455,51003
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,678.918,3.62642e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.84783,54143
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.59723,54251
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,21.4198,151819
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,253.501,1.49005e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.88187,55678
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.6939,55620
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,412.319,2.46737e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.74961,57593
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.98414,57593
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.5037,109358
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.03222,68299
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.92216,57574
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,11.4418,97628
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,3.94646,57346
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.55647,57120
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,359.484,2.12355e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.3003,329059
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,29.5784,175790
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,796.238,4.63702e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,407.145,2.32402e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,409.181,2.36025e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,50.0483,326751
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,30.6403,207453
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,14.9928,106593
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,52.6524,329883
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_46.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_46.txt
deleted file mode 100644
index 587a4bf63f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_46.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,372.942,1.50283e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.42906,39184
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.94965,39241
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1324.67,5.88102e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.11742,47981
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.79925,48021
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,18.3743,119504
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,410.906,2.10147e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.97375,51067
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.66037,51086
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,676.427,3.65834e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.06719,54117
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.34689,54269
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,20.9467,147949
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,251.52,1.48384e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.26901,55681
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.01198,55678
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,408.643,2.46041e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.6836,57634
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.81573,57558
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.2832,109364
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.36828,68324
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.89762,57691
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,11.4947,103479
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.2993,57538
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,6.14501,57462
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,353.979,2.13926e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,56.9928,346984
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,29.1811,175386
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,797.413,4.68361e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,407.74,2.34011e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,408.232,2.37399e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,49.4458,318489
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,28.7016,197598
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,12.5925,96118
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,48.8592,302847
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_261.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_261.txt
deleted file mode 100644
index 7cbbe3bf58..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_261.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,199.096,768777
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.87226,36317
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.31064,36355
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,862.768,3.5928e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.47999,42966
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.916697,42889
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,14.7617,85988
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,243.939,1.10794e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.35692,45410
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.885912,45430
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,432.939,2.0384e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.30457,48006
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.967639,48083
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,12.2797,91598
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,157.17,786742
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.45318,49513
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.8852,49551
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,297.318,1.59494e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.36708,51385
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.21256,51366
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.5806,97816
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,8.09633,57082
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,0.937241,51595
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.82347,51595
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,2.95617,51424
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.27187,51290
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,402.601,2.16753e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,100.073,540576
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.2652,213693
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,940.594,5.07676e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,529.699,2.81573e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,538.929,2.93758e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,68.0023,405177
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,40.2808,250158
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,11.5333,69055
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,53.7619,322533
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_262.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_262.txt
deleted file mode 100644
index b8ce558cca..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_262.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,201.491,782079
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.64602,36755
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.25925,36812
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,863.481,3.60566e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.323,43292
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.23471,43349
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,15.2389,86831
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,244.387,1.08745e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.66831,45676
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.88134,45676
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,434.161,2.03201e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.44482,48634
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.05884,48634
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,11.3014,82389
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,155.029,784704
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.46854,49989
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.37976,50063
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,294.808,1.5918e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.66305,51777
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.31285,51929
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.7435,93346
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,8.66416,62669
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.54901,52005
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.01051,52005
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,2.76116,46791
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.01021,51763
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,401.791,2.17637e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,101.101,554195
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.6083,214821
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,980.097,5.33162e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,567.423,3.03884e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,582.842,3.20481e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,70.175,419035
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,40.4284,263401
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,13.3167,91066
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,53.7532,315081
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_263.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_263.txt
deleted file mode 100644
index dec9f13da9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_263.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,239.067,970071
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.33674,38093
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.34824,38131
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1255.39,5.6059e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.37852,47241
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.771448,47298
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,14.7832,94710
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,341.94,1.68639e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.16271,50082
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.818232,50099
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,635.421,3.26296e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.23858,53688
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.90917,53652
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,17.0536,113283
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,233.522,1.32983e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.39134,55217
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.274,55138
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,411.418,2.44319e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.74261,57437
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.19678,57456
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.5902,109239
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.51897,73973
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.93151,57415
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,8.79818,63113
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.12541,57053
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.56009,56904
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,370.749,2.16788e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,73.0082,435852
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,29.2538,196008
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,719.422,4.17407e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,477.161,2.70832e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,482.106,2.80035e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,53.2209,328026
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.3862,203294
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.51689,73278
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,42.6197,273975
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_264.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_264.txt
deleted file mode 100644
index 2445a2271c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_264.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,239.767,971021
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.04048,38189
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.52667,38189
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1257.19,5.61616e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.40312,47337
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.07093,47356
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,15.0031,94899
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,342.292,1.69339e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.14918,50061
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.01116,50119
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,635.865,3.30913e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.45573,53609
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.24347,53686
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,18.5907,135377
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,233.494,1.36403e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.06065,55385
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.60712,55398
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,412.356,2.47544e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.91035,57419
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.64814,57473
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.7297,115070
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.47385,68360
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.72888,57565
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,8.59251,57568
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,2.73992,57114
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,4.57165,57041
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,352.018,2.10195e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,69.4555,427759
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,28.3901,175566
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,762.599,4.49019e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,458.711,2.59714e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,464.862,2.68914e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,51.7606,322724
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,29.6339,202986
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,11.0134,84635
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,41.902,274721
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_265.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_265.txt
deleted file mode 100644
index 62d8f0cee8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_265.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,237.945,974787
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.34541,38933
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.96817,38990
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1254.66,5.70865e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.36175,48021
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.993847,48021
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,14.85,96156
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,341.703,1.70559e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.12607,50768
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.871769,50863
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,631.892,3.30975e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.68056,54138
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.57061,54157
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,16.9613,119653
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,235.287,1.37796e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.30232,55876
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.31317,55818
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,409.034,2.46594e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.8821,57865
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.2812,57885
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,12.0105,115979
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.89378,75840
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.59963,57941
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,8.0217,57846
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,3.46285,57596
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,4.38688,57596
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,421.913,2.5351e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,113.737,672019
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,42.6905,276013
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1282.92,7.57748e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,783.575,4.51126e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,824.337,5.06865e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,76.7767,484119
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,46.8838,304945
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,13.0073,91564
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,59.7061,388204
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_266.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_266.txt
deleted file mode 100644
index ecf5ecb4c0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_266.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,219.402,872895
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.06323,37820
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.28616,37877
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1087.48,4.72668e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.45423,45690
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.889464,45710
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,14.9285,91612
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,299.698,1.42697e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.15941,48366
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.69762,48404
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,543.633,2.71115e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.6237,51587
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.23599,51603
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,13.6092,87449
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,219.523,1.22238e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.38871,53306
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.32568,53268
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,385.392,2.22364e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.79116,55394
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.15909,55391
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.5473,105219
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,8.77498,65625
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.17118,55485
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,11.9563,82410
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,3.65402,55302
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.5582,55148
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,400.445,2.30964e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,99.0877,577237
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.3662,217807
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1139.33,6.49791e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,588.399,3.32596e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,603.878,3.45864e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,68.9594,418837
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,41.7461,261641
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,14.1301,94556
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,57.1888,354516
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_267.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_267.txt
deleted file mode 100644
index 5ad85fb3de..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_267.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,220.768,858125
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.07994,37021
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.91592,37059
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1088.57,4.65225e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.56389,44988
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.27442,45124
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,15.7948,94823
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,300.073,1.40906e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.19535,47582
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.977017,47717
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,547.289,2.68416e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.61176,50906
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.42805,50864
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,15.0744,101958
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,223.919,1.23953e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.20094,52483
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.717,52617
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,385.025,2.19024e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.48081,54763
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.45169,54795
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.8127,109809
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,10.1876,70549
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.12561,54924
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,9.80338,60375
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,3.58622,54682
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,4.99952,54435
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,342.591,1.95981e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,65.2348,378357
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,27.6085,157669
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,723.164,4.09217e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,386.548,2.12536e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,385.66,2.13438e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,47.3982,289352
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,29.2509,190152
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.1746,81443
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,41.9383,254484
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_268.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_268.txt
deleted file mode 100644
index c0755261b9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_268.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,223.219,866209
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.29751,36564
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.38753,36602
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1088.38,4.59111e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.36405,44671
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.0197,44691
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,15.7704,89378
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,300.752,1.40425e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.19036,47394
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.865849,47429
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,546.413,2.67128e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.81794,50614
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.06284,50729
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,12.0299,96463
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,217.962,1.19927e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.12097,57432
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.48472,52097
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,386.475,2.17513e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.23227,54507
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.37096,54504
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,11.6492,109070
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,10.6493,87036
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.33381,48998
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.42382,59076
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,3.39623,49099
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,4.88089,49023
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,330.773,1.86958e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,51.2756,298484
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,22.9797,157104
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,570.755,3.21619e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,311.244,1.71093e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,311.758,1.73012e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,42.3526,256955
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,23.9081,160559
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,8.93398,64312
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,33.2813,213307
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_269.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_269.txt
deleted file mode 100644
index 7786d5550a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_269.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,225.308,878318
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.25354,36964
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.42312,37002
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1090.22,4.64791e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.4006,44952
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.983479,45012
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,15.8368,90062
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,301.216,1.40918e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.33717,47566
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.913721,47585
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,545.781,2.68126e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.42581,50829
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.20725,50886
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,13.1032,107069
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,231.958,1.27411e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.79127,52506
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.59179,52352
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,386.436,2.09794e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.93601,54701
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.14792,54774
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,12.0976,109698
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,9.89964,81790
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.42037,54830
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,10.0702,65802
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,3.861,54641
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,5.58297,54474
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,352.343,2.00256e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,70.9341,398430
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,30.738,178279
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,801.38,4.51729e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,443.007,2.44053e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,427.524,2.36311e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,53.5074,320543
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.6672,195257
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,12.3089,80832
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,45.586,269939
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp32_perf_fp32_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp32_perf_fp32_120.txt
deleted file mode 100644
index f00d5f1ed2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp32_perf_fp32_120.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,539.627,1.99043e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.16915,36437
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.89879,36514
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,2027.43,8.19658e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.94056,46777
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.5858,46796
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,36.2756,195821
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,950.176,4.46438e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.26696,49224
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.48219,49240
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,1414.68,7.23013e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.54139,52507
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.93207,52526
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.4147,178967
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,337.686,1.93166e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.38756,54394
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.20894,54394
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,562.317,3.29486e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.00581,57127
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.06421,57085
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.791,137046
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,4.97337,57218
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.98056,57313
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.19512,57294
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.20489,57142
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.99914,56989
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,365.507,2.19674e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.5339,340298
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,40.5251,267754
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1070.58,6.06614e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,811.869,4.21348e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,663.055,3.50782e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,76.904,435045
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.1423,199582
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.7964,86630
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,63.4149,367588
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_121.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_121.txt
deleted file mode 100644
index 507ccfb184..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_121.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,463.303,1.76918e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.80061,37657
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.71535,37676
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1139.3,4.77571e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.14229,44608
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.77874,44586
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.2747,177316
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,356.378,1.68966e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.33614,47132
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.78242,47148
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,590.817,2.9326e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.39963,49955
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.62117,49993
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,26.2844,149031
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,326.912,1.76844e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.53396,52247
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.34178,52190
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,571.489,3.25395e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,4.07786,55086
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.44709,55007
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.9743,132412
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,6.31362,59774
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.91106,55127
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.93124,55053
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.3293,55058
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.67543,54889
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,351.478,2.02206e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,52.9578,311755
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.6067,226535
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,837.679,4.63e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,551.053,2.95982e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,572.592,3.14542e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,74.4555,417885
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,28.045,166902
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,7.9946,56776
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,53.8845,306039
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_122.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_122.txt
deleted file mode 100644
index a88dd7e64b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_122.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,462.507,1.77319e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.18257,37772
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.31061,37810
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1138.87,4.79222e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.09867,44486
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.02958,44501
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,34.534,186204
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,367.376,1.75098e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.00658,47240
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.24613,47236
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,584.331,2.91639e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.4093,49994
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.35611,50087
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,27.1279,165066
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,341.834,1.85874e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.37009,56433
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.06434,52144
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,561.96,3.1899e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,3.12967,54906
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.80225,54921
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,20.1178,137144
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,4.8846,54990
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.87263,55009
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.37842,55013
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.92486,54860
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,6.98142,54764
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,359.552,2.05582e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.4482,316215
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,40.3342,252489
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,857.765,4.76229e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,559.232,2.9486e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,576.081,3.16067e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,75.9264,439079
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,29.1708,171763
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,7.90801,62038
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,56.8385,322044
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_123.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_123.txt
deleted file mode 100644
index f625708f35..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_123.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,437.953,1.66532e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.41153,37296
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.36523,37315
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1121.87,4.74012e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.36126,44302
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.44811,44322
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,35.7631,190062
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,365.679,1.77306e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,3.03326,46930
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.35995,46949
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,584.447,2.96575e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.73886,49742
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.20453,49776
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,23.884,149630
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,340.073,1.85428e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.46027,51768
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.40834,51768
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,549.789,3.10959e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.78392,54594
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.88946,54690
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.7755,126361
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.65385,59408
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.75634,54797
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.20123,54722
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.66834,54495
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.80535,54419
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,402.055,2.27725e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,73.3859,423786
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,53.3935,313453
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1248.44,6.83024e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,819.675,4.31134e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,842.22,4.67499e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,99.7207,569478
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,41.8504,250959
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,11.4542,68004
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,75.1463,426502
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_124.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_124.txt
deleted file mode 100644
index f4a26c1fd5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_124.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,433.029,1.66026e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.79847,37526
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.94894,37545
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1115.5,4.67817e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.74805,44293
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.51986,44234
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.6099,177747
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,350.715,1.66186e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.03669,46972
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.69615,47045
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,575.575,2.86038e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.9667,44899
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.67573,44937
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,23.9632,143993
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,341.352,1.82392e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.54861,51957
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.10808,51977
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,553.307,3.13107e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.85675,54603
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.11919,54638
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.0234,120596
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,6.12863,59638
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.96779,55014
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.03931,54957
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.27804,54671
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.67854,54499
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,414.334,2.34711e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,81.1861,472485
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,57.9362,345585
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1376.77,7.54063e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,905.412,4.78984e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,925.723,5.1848e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,109.2,632711
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,44.4932,264202
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.8511,84914
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,81.544,476016
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_125.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_125.txt
deleted file mode 100644
index 7804c699e5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_125.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,516.605,2.02074e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.08455,38686
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.16321,38724
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1442.45,6.28461e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.1052,47107
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.93118,47124
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.9007,173007
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,504.737,2.53152e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.44888,50485
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.63887,50523
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,818.749,4.35636e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.83912,53918
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,3.95569,59116
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.3683,183554
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,350.651,2.05332e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.03086,55901
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.18891,55960
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,581.055,3.53253e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.62206,58706
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.09442,58630
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.9037,141471
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.75583,64099
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.82994,58688
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.5515,58688
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.42665,58688
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.3056,58611
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,361.732,2.19452e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,50.9245,331948
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.0977,241075
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1033.24,6.05109e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,686.711,3.9161e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,696.342,4.00968e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,67.7964,400530
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.1881,210799
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.8773,69820
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,56.9077,350084
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_126.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_126.txt
deleted file mode 100644
index ea6400d03a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_126.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,512.889,2.01624e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.56292,38992
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.40367,39049
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1431.3,6.29537e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.60188,47503
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.86101,47420
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.7532,175943
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,516.412,2.62336e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.93544,50769
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.83394,50788
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,856.837,4.57126e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.06059,54186
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.71297,54205
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,29.5667,177722
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,354.288,2.10221e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.15538,56361
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.41349,56361
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,589.817,3.61846e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.8827,58957
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,3.13194,58976
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.2826,135265
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.32367,59049
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.12677,59106
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,5.95445,59163
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.22271,58840
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.30033,58783
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,351.258,2.19612e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,51.0832,328802
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,37.5998,249477
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1048.34,6.18942e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,697.459,4.03406e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,709.916,4.14341e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,69.5421,420211
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,29.7712,195854
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.08736,65243
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,57.317,345781
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_127.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_127.txt
deleted file mode 100644
index 11dacf8858..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_127.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,513.644,2.01811e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.59313,38800
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.09014,38819
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1459.44,6.36232e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.52418,47452
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.50447,47425
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,30.1814,170372
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,506.56,2.53914e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.60098,50870
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.42729,50870
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,864.817,4.56692e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.51477,54332
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.582,54385
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.9645,184505
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,371.805,2.18465e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.42529,56397
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.48629,56397
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,621.968,3.79972e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.37768,59376
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.3429,59258
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,19.0982,142744
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.49858,64460
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.76738,59487
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.03665,59411
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.6974,59011
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.14829,58801
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,365.736,2.25533e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,53.5463,346146
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,39.5809,265556
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1101.98,6.49844e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,732.589,4.23126e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,745.956,4.34033e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,72.0537,440311
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.4142,216953
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.3084,69893
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,57.1276,339492
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_128.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_128.txt
deleted file mode 100644
index cf76f32f4c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_128.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,490.555,1.91277e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.76923,38268
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.96098,38249
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1422.48,6.18362e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.65071,46548
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.8917,46539
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.8054,181739
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,502.785,2.53056e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.79368,49969
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.61884,50007
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,822.43,4.37771e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.58405,53441
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.36056,53498
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,27.7038,176948
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,361.565,2.11351e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.89698,55445
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.83042,55502
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,585.627,3.53768e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.72002,58036
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.98949,58112
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.8197,127234
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.31596,58169
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.9045,58284
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.18161,58149
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.77481,57881
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.7898,57805
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,359.041,2.17864e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,53.103,339081
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.7997,261123
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1104.5,6.39538e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,718.275,3.94992e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,740.786,4.25646e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,74.8935,457030
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.3358,210152
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.5654,69967
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,56.8306,347219
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_129.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_129.txt
deleted file mode 100644
index a6f28a7c60..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_129.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,488.551,1.89905e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.07681,38286
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.20149,38321
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1422.93,6.17546e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.76018,46702
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.88731,46640
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.1803,176017
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,519.02,2.61002e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.01589,50031
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.83906,50050
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,853.518,4.55277e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.62392,53482
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.36114,53516
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,27.9432,177270
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,359.421,2.09654e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.21755,55503
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.3411,55390
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,584.796,3.46279e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.82331,58295
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.24235,58232
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.4951,140724
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.48997,63602
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.69624,58402
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.33364,58325
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.03267,58021
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.19041,58021
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,357.807,2.07422e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.1799,335929
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,39.9407,245484
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1110.51,6.37089e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,724.779,4.04036e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,747.887,4.3069e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,73.0584,435819
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,33.3917,209657
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.0744,64045
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,57.0774,348729
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_130.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_130.txt
deleted file mode 100644
index 636e187079..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_130.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,489.265,1.90801e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.08052,38420
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.5533,38458
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1444.06,6.23284e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.61314,46778
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.26415,46807
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.4642,174262
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,507.373,2.5221e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.76619,49995
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.34719,50033
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,852.637,4.48079e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.97144,53488
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.44322,53431
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,29.6332,192769
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,370.72,2.17036e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.90097,55529
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.18888,55396
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,622.862,3.76753e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,3.12874,58209
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.49246,58225
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.269,128353
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,6.09938,63426
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.99973,58434
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.08824,58358
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.12934,58185
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.18916,58185
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,363.332,2.21334e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.9078,336453
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,40.3138,262016
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1166.64,6.79804e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,759.293,4.23371e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,778.059,4.47344e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,76.8648,457931
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,33.4184,215316
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.80722,64270
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,57.8009,337244
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_131.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_131.txt
deleted file mode 100644
index 50c3eb9ec4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_131.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,522.794,2.04511e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.73265,38878
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.18905,38839
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1523.97,6.65219e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.6038,47678
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.24581,47691
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.8813,166415
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,533.498,2.69408e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.74501,51110
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.20937,51126
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,942.318,4.9949e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,3.15937,54300
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.61883,54300
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,31.5194,201945
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,371.802,2.20157e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.64565,56288
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.31787,56326
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,612.351,3.74904e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.79723,59205
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.43208,59202
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.1139,124537
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,6.90027,69758
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.78789,59316
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.70946,59240
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.87174,52934
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.19189,52934
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,366.553,2.23326e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,55.254,352097
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,39.9132,271326
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1214.21,7.11476e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,801.81,4.56785e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,812.935,4.72435e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,71.2519,429776
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.3433,201451
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.90617,92328
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,57.6299,357340
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_132.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_132.txt
deleted file mode 100644
index d2c262d04f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_132.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,525.37,2.06192e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.96747,38705
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.06894,38629
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1520.58,6.66722e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.01519,47444
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.61154,47444
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.7819,170612
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,536.154,2.72908e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.54613,51080
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.84213,51191
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,932.92,4.99892e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.20914,54418
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.4404,54476
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,30.1968,197082
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,367.303,2.1788e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.35217,56422
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.98562,56365
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,610.051,3.74586e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.97608,59053
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.88159,59053
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.3418,130176
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.61065,64309
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.90408,59259
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.24766,59183
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.49644,58975
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.90756,58818
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,378.999,2.3384e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,52.0377,338169
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.9645,264915
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1200.99,7.04572e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,787.969,4.44044e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,804.791,4.6697e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,71.1626,432984
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.0567,210482
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.88435,75320
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,58.4652,354827
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_133.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_133.txt
deleted file mode 100644
index e8443470d8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_133.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,520.888,2.04579e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.47944,38877
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.12322,38972
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1514.43,6.63255e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.56204,47815
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.07541,47828
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,28.3388,163490
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,527.728,2.63761e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.54604,51022
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.10674,51096
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,957.354,5.10974e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.78369,54366
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.50894,54377
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,29.6448,202412
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,367.438,2.18595e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.01608,56323
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.23202,56400
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,616.571,3.78567e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.00635,59071
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.06529,59163
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.1188,129509
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.1254,59201
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.01387,59296
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.6084,59201
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.34773,59049
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.6952,64896
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,355.626,2.21807e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,52.0829,340629
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.1219,259621
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1158.3,6.83392e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,761.462,4.27722e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,778.415,4.54375e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,70.3622,430421
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.3293,207125
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.1576,86781
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,53.7575,330000
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_134.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_134.txt
deleted file mode 100644
index f156a272bc..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_134.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,518.225,2.03857e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.76772,38705
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.54213,38839
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1511.44,6.60475e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.64194,47527
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.6677,47581
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.5123,175937
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,531.752,2.68169e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.50665,51028
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.54812,51083
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,955.147,5.09862e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,3.28308,54426
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.75832,54345
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,30.6361,202254
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,367.801,2.18166e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.3061,56250
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.37115,56270
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,613.327,3.75759e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,3.20113,59033
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.25854,58995
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.5104,129265
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.39858,59068
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.01976,59087
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.26823,59087
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.39263,59087
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.12753,58550
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,353.751,2.19414e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,53.7501,344014
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.7621,259321
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1233.71,7.25328e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,810.47,4.59346e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,829.939,4.82735e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,73.6112,439468
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.6467,205827
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.57212,81121
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,58.215,356117
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_135.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_135.txt
deleted file mode 100644
index d93aac33b8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_135.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,497.796,1.93042e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.85409,38268
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.4189,38306
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1510.04,6.51319e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.77071,47233
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.40694,47306
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,30.0603,156603
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,532.12,2.66163e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.06754,50490
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.41295,50490
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,944.021,4.99783e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.60625,53848
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.67332,53906
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.0045,183834
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,370.626,2.18061e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.13735,55695
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.94213,55733
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,611.587,3.70482e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.46552,58631
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.97964,58419
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.2842,122698
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,6.12955,63752
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.06546,58742
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.23947,63617
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.50137,58281
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.79949,58186
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,346.122,2.12271e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,48.9698,309221
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,35.7981,229550
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1113.87,6.51033e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,721.95,4.05898e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,743.515,4.2954e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,68.1563,409251
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,29.4963,188728
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.1106,69993
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,54.0631,321801
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_136.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_136.txt
deleted file mode 100644
index 0b546e9036..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_136.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,500.05,1.94081e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.03905,38192
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.37129,38211
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1499.93,6.50211e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.79157,47249
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.8422,47278
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.4395,171501
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,527.901,2.63924e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.83384,50487
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.70271,50523
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,932.858,4.94098e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.81124,53920
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,3.28826,53958
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.963,178504
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,374.848,2.20488e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.36002,55907
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.26815,55963
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,619.314,3.75355e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.53089,58496
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.90792,58512
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.6418,134224
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.1248,58722
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.16875,58741
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.60833,63770
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,6.17848,58588
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.31312,58186
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,358.901,2.19902e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,53.4904,348181
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.9841,258346
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1228.34,7.1658e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,798.616,4.50051e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,818.663,4.72786e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,72.3452,442673
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.4503,215227
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.604,91273
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,58.0678,353514
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_137.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_137.txt
deleted file mode 100644
index 18601a156b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_137.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,494.673,1.94348e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.66638,38664
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.33449,38588
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1494.56,6.44786e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.65723,47325
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.41615,47359
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,29.287,165234
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,522.499,2.58653e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.65298,50408
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.17503,50466
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,956.173,4.99662e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,3.09796,53805
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.70279,53787
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,29.8641,194485
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,368.763,2.16565e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.92277,55739
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.94981,55758
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,618.426,3.73154e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.29106,58531
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.45537,58531
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.5694,117081
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,6.14133,63675
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.84811,58722
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.87297,58645
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.98089,58472
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.19059,58242
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,356.023,2.18003e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.9909,349138
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,41.5442,275170
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1235.38,7.14779e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,813.297,4.61604e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,816.186,4.7365e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,71.9996,429052
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.682,212127
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.3575,97760
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,58.9096,356092
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_138.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_138.txt
deleted file mode 100644
index 18e2ee76b7..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_138.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,493.702,1.93113e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.48485,38344
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.38601,38382
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1496.88,6.37549e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.5277,47233
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.27279,42369
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,30.069,166361
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,518.673,2.5721e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.63791,50474
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.15097,50529
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,951.896,4.78278e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.89698,53722
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.26968,53747
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,31.9208,199983
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,369.81,2.17044e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.24215,55713
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.09726,55636
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,622.511,3.7004e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,3.03233,58453
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.14261,58376
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.5437,133799
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.21081,58586
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.81855,58663
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.61397,58509
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.59887,58261
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.24077,58185
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,347.875,2.13694e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,52.3979,342073
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.5801,251027
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1193.2,6.96721e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,781.872,4.35658e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,800.611,4.63655e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,73.1664,436793
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.8312,199875
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.1147,75365
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,54.928,332389
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_41.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_41.txt
deleted file mode 100644
index 070e97bed1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_41.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,504.874,1.93218e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.2764,37870
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.84478,37889
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1821.15,7.86918e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.54677,48007
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.33131,48083
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,32.4474,181601
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,818.136,4.06536e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.06821,50905
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.92738,50924
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,1332.75,7.09964e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.33106,54383
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.84814,54380
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.1628,179710
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,326.011,1.91353e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.13013,56167
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.2812,56186
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,542.389,3.25438e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.66892,58802
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.12238,58840
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.9176,141070
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.40329,64347
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.65055,58916
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.26913,58916
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.00998,58802
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.06369,58631
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,359.221,2.20974e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,39.7362,254622
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.5187,230487
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1030.71,6.01365e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,594.95,3.36356e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,557.701,3.13541e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,72.117,434533
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,21.496,157302
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.77318,59332
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,60.2474,367981
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_43.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_43.txt
deleted file mode 100644
index adee4a649a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_43.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,512.584,1.97652e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.32862,38022
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,3.20465,37983
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1806.09,7.89462e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.81548,48135
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.55054,48212
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,35.4837,203280
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,808.325,4.12335e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.66647,51342
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.69652,51285
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,1131.13,6.20321e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.82613,55541
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.96615,55541
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.3146,183952
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,322.36,1.94429e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.08904,57274
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.06011,57350
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,541.034,3.34225e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.92543,59851
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.17311,59867
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.9872,149334
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,4.56032,59981
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.84616,59904
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.16059,59904
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.07836,59561
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.96989,59504
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,366.098,2.28511e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,39.3923,252201
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.8196,240015
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1017.86,6.04683e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,617.427,3.51735e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,583.057,3.35479e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,73.9926,441140
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,22.9876,163395
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.987,64947
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,62.2584,391012
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_45.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_45.txt
deleted file mode 100644
index 151d408eed..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_45.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,502.579,1.9411e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,1.98366,38272
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.89252,38310
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1802.71,7.87184e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.703,48252
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.11467,48291
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,34.2146,192488
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,780.652,3.96253e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.01461,51323
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.35058,51343
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,1158.69,6.33308e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.46347,55465
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.77848,55541
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.3453,194165
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,327.96,1.96248e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.31067,57084
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.46417,57084
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,542.253,3.3329e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.80732,59851
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.14159,59793
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.8409,143814
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.34982,65415
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.65468,59888
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.04942,59888
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.50226,59622
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.53594,65470
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,358.603,2.2467e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,45.8217,286536
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,37.9513,239505
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1017.63,6.04958e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,601.659,3.44212e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,576.172,3.33753e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,74.0101,448507
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,26.2698,164121
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.7095,60231
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,61.3979,381951
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_231.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_231.txt
deleted file mode 100644
index b1900b41c2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_231.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,312.955,1.05814e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.29102,32634
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.92555,32750
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1270.25,4.95079e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.64185,42084
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.93026,42046
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,32.9684,164910
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,375.799,1.72684e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.62242,45265
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.86757,45437
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,628.419,3.06087e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.55881,49018
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.97983,49095
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,24.1804,147478
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,236.215,1.27624e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.07611,50990
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.91714,51009
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,416.23,2.3139e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.9468,53167
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.39957,53167
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,19.4841,128208
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.49903,58078
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.9957,53450
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.69527,53298
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.90681,52994
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,6.96046,52917
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,398.579,2.22238e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,75.6271,418107
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,53.4082,301014
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,2181.08,1.18919e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,1541.07,7.72153e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,1518.65,7.74613e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,138.242,667792
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,43.985,200120
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,16.4329,76086
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,126.613,598657
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_232.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_232.txt
deleted file mode 100644
index acea102c80..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_232.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,320.765,1.0895e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.74497,32444
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.91287,32520
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1271.88,4.91642e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.52866,42009
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.32443,42066
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.5942,164724
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,374.875,1.70587e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.75916,45227
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.94523,45284
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,631.129,3.06489e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.68424,49132
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.28821,48995
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,27.4421,157819
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,231.786,1.24844e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.17838,50775
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.17326,50736
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,418.555,2.32751e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.8323,52971
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.43531,53009
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.7868,122228
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.95036,53082
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.27055,53253
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.92836,57843
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.84128,52968
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.47956,52968
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,408.421,2.2672e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,76.7764,449207
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,54.9257,305927
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,2268.42,1.23557e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,1680.26,8.35215e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,1579.97,8.0778e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,134.762,654124
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,42.9408,220383
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,14.0336,66829
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,119.188,566439
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_233.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_233.txt
deleted file mode 100644
index 47aa85aa79..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_233.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,364.281,1.42832e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.01979,37783
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.91876,37859
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1724.95,7.72746e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.61807,48656
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.72616,48711
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.8694,198585
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,485.892,2.51283e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.7669,51802
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.50969,51876
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,882.543,4.87772e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.41816,55778
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.91368,55832
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,30.075,201041
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,301.008,1.83866e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.06062,57411
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.58459,57355
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,558.599,3.47618e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.79893,59907
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.99259,59945
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,19.0353,143544
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.68991,65276
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.00786,60020
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.53729,60020
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.20431,59634
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.20196,59403
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,364.101,2.26809e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,56.8705,359483
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,41.6894,285564
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1542.04,9.30453e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,983.446,5.5907e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,1027.43,6.08045e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,76.4432,458940
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.5815,196982
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.17091,76617
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,61.1893,380972
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_234.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_234.txt
deleted file mode 100644
index 19f49e9d48..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_234.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,359.171,1.3654e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.56808,36755
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.57861,36830
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1720.43,7.51741e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.55141,48008
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.80565,47875
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,29.423,167745
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,477.2,2.42612e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.41721,51394
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.20194,51432
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,887.624,4.77616e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.74392,55249
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.76305,55283
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,30.7709,205785
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,301.656,1.81816e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.88015,57099
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.21406,57041
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,562.265,3.49027e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.89061,59568
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.32229,59603
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.4432,119295
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.97577,65011
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.85704,59733
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.36932,64953
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.80956,59428
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.87399,59198
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,354.283,2.21567e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,53.9692,353895
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,39.3985,273572
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1682.64,1.00985e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,966.772,5.42835e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,994.348,5.79107e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,82.3764,478623
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.1803,209638
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.9806,96473
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,66.7879,398824
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_235.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_235.txt
deleted file mode 100644
index f5857d3264..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_235.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,359.179,1.2212e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.8443,33172
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.75128,33192
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1725.86,7.02108e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.15976,45546
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.74376,45603
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,35.1932,192282
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,486.73,2.42326e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.32251,49345
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.11051,49344
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,874.152,4.65541e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.14318,53834
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,3.4425,53834
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,27.2372,171817
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,304.18,1.78209e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.93003,55659
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.31419,55659
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,556.548,3.35208e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.22421,58182
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.0956,58217
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.3645,116615
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.83999,63656
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.86607,58438
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,5.43334,58286
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.87446,57981
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.7042,57854
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,420.697,2.53764e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,85.3445,534738
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,60.8316,387686
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,3076.32,1.7793e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,2322.36,1.18957e+07
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,1926.57,9.90526e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,147.311,701287
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,47.9584,230534
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,14.2147,71974
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,134.215,637238
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_236.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_236.txt
deleted file mode 100644
index b38950da77..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_236.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,369.184,1.30472e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,1.83851,34353
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.51925,34373
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1814.45,7.54365e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.95391,46542
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.14664,46540
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,32.855,187012
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,507.61,2.53206e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.52722,49910
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.48972,50117
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,946.582,5.02218e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.71148,54187
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.50529,54206
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,25.3727,162878
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,339.38,1.9942e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.9773,56181
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.00066,56181
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,600.8,3.65449e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.04114,58747
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.99851,58760
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,19.3574,159457
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.00924,64151
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.56626,58838
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.17109,58838
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.10332,58765
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.09669,58536
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,416.965,2.55865e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,76.3945,481875
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,53.5784,341610
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,2759.9,1.6179e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,1799.2,9.41621e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,1565.91,8.37989e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,126.676,649481
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,41.5973,232891
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,12.366,69394
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,110.356,555019
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_237.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_237.txt
deleted file mode 100644
index 3c5e6c169a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_237.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,366.675,1.45154e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,1.87509,37821
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.80517,37879
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1807.03,8.02866e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.63839,48695
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.56049,48752
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.7969,185595
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,512.094,2.68803e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.58895,52031
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.87138,52047
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,941.746,5.20451e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.13765,55821
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,3.10807,55782
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.328,190552
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,330.901,2.01841e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.17483,57648
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.25227,57686
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,595.453,3.71972e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.89838,60127
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.00248,60085
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.9617,138566
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.71676,65494
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.72594,60254
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.70641,65417
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.9998,59796
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.73671,59723
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,358.565,2.24855e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,51.6014,332793
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,37.4462,251934
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1604.66,9.70904e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,862.655,4.92921e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,894.783,5.32491e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,69.249,427305
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.0203,213623
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,11.1213,98381
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,55.7902,348237
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_238.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_238.txt
deleted file mode 100644
index dbc1247487..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_238.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,361.396,1.42936e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.077,38038
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.85547,38057
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1801.71,8.07173e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.50293,48844
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.82636,48860
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,32.1758,175635
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,505.379,2.6346e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.64517,52142
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.96376,52219
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,943.341,5.17271e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.88607,56170
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.52344,56131
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,27.3298,184853
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,332.621,2.02922e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.81202,57654
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.94287,57596
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,597.906,3.72187e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.1228,60207
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.09502,60223
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,19.2042,151258
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.2175,65703
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.67349,60463
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.27326,60313
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.12176,60010
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.77034,59780
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,332.125,2.08543e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,40.8915,282376
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,31.3805,224408
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1229.22,7.34539e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,658.089,3.74298e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,677.838,3.98656e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,56.4301,354013
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,25.7725,165318
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.87913,71131
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,45.0084,284986
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_239.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_239.txt
deleted file mode 100644
index cc13932a92..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_239.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,365.933,1.40138e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.80164,36850
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.55448,36888
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1806.48,7.93584e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.88536,48163
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.64411,48201
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,35.6716,207285
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,512.369,2.65332e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.04005,51632
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.16469,51647
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,935.445,5.14346e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.03416,55351
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.77188,55389
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.7423,188823
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,338.056,2.03138e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.87183,57064
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.23042,57199
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,592.54,3.65101e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.68578,59431
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.24219,59489
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.0715,124954
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.97676,64936
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.76684,59600
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.85672,59600
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.97948,59450
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.37134,59450
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,354.286,2.1935e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.0783,348349
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,39.247,244872
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1792.85,1.07184e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,1012.73,5.58673e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,1045.6,5.94124e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,83.6212,481961
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.2344,206842
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,11.1349,73456
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,69.2339,398246
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_120.txt
deleted file mode 100644
index f00d5f1ed2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_120.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,539.627,1.99043e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,4.16915,36437
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.89879,36514
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,2027.43,8.19658e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.94056,46777
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.5858,46796
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,36.2756,195821
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,950.176,4.46438e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.26696,49224
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.48219,49240
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,1414.68,7.23013e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.54139,52507
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.93207,52526
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.4147,178967
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,337.686,1.93166e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.38756,54394
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.20894,54394
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,562.317,3.29486e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.00581,57127
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.06421,57085
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.791,137046
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,4.97337,57218
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.98056,57313
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.19512,57294
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.20489,57142
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.99914,56989
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,365.507,2.19674e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.5339,340298
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,40.5251,267754
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1070.58,6.06614e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,811.869,4.21348e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,663.055,3.50782e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,76.904,435045
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.1423,199582
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.7964,86630
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,63.4149,367588
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_121.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_121.txt
deleted file mode 100644
index 507ccfb184..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_121.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,463.303,1.76918e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.80061,37657
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.71535,37676
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1139.3,4.77571e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.14229,44608
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.77874,44586
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.2747,177316
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,356.378,1.68966e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.33614,47132
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.78242,47148
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,590.817,2.9326e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.39963,49955
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.62117,49993
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,26.2844,149031
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,326.912,1.76844e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.53396,52247
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.34178,52190
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,571.489,3.25395e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,4.07786,55086
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.44709,55007
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.9743,132412
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,6.31362,59774
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.91106,55127
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.93124,55053
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.3293,55058
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.67543,54889
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,351.478,2.02206e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,52.9578,311755
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.6067,226535
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,837.679,4.63e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,551.053,2.95982e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,572.592,3.14542e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,74.4555,417885
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,28.045,166902
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,7.9946,56776
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,53.8845,306039
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_122.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_122.txt
deleted file mode 100644
index a88dd7e64b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_122.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,462.507,1.77319e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.18257,37772
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.31061,37810
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1138.87,4.79222e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.09867,44486
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.02958,44501
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,34.534,186204
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,367.376,1.75098e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.00658,47240
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.24613,47236
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,584.331,2.91639e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.4093,49994
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.35611,50087
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,27.1279,165066
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,341.834,1.85874e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.37009,56433
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.06434,52144
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,561.96,3.1899e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,3.12967,54906
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.80225,54921
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,20.1178,137144
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,4.8846,54990
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.87263,55009
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.37842,55013
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.92486,54860
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,6.98142,54764
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,359.552,2.05582e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.4482,316215
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,40.3342,252489
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,857.765,4.76229e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,559.232,2.9486e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,576.081,3.16067e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,75.9264,439079
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,29.1708,171763
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,7.90801,62038
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,56.8385,322044
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_123.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_123.txt
deleted file mode 100644
index f625708f35..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_123.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,437.953,1.66532e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.41153,37296
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.36523,37315
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1121.87,4.74012e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.36126,44302
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.44811,44322
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,35.7631,190062
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,365.679,1.77306e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,3.03326,46930
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.35995,46949
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,584.447,2.96575e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.73886,49742
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.20453,49776
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,23.884,149630
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,340.073,1.85428e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.46027,51768
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.40834,51768
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,549.789,3.10959e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.78392,54594
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.88946,54690
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.7755,126361
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.65385,59408
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.75634,54797
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.20123,54722
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.66834,54495
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.80535,54419
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,402.055,2.27725e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,73.3859,423786
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,53.3935,313453
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1248.44,6.83024e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,819.675,4.31134e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,842.22,4.67499e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,99.7207,569478
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,41.8504,250959
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,11.4542,68004
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,75.1463,426502
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_124.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_124.txt
deleted file mode 100644
index f4a26c1fd5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_124.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,433.029,1.66026e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.79847,37526
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.94894,37545
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1115.5,4.67817e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.74805,44293
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.51986,44234
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.6099,177747
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,350.715,1.66186e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.03669,46972
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.69615,47045
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,575.575,2.86038e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.9667,44899
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.67573,44937
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,23.9632,143993
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,341.352,1.82392e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.54861,51957
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.10808,51977
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,553.307,3.13107e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.85675,54603
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.11919,54638
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.0234,120596
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,6.12863,59638
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.96779,55014
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.03931,54957
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.27804,54671
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.67854,54499
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,414.334,2.34711e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,81.1861,472485
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,57.9362,345585
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1376.77,7.54063e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,905.412,4.78984e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,925.723,5.1848e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,109.2,632711
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,44.4932,264202
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.8511,84914
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,81.544,476016
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_125.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_125.txt
deleted file mode 100644
index 7804c699e5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_125.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,516.605,2.02074e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.08455,38686
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.16321,38724
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1442.45,6.28461e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.1052,47107
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.93118,47124
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.9007,173007
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,504.737,2.53152e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.44888,50485
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.63887,50523
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,818.749,4.35636e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.83912,53918
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,3.95569,59116
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.3683,183554
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,350.651,2.05332e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.03086,55901
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.18891,55960
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,581.055,3.53253e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.62206,58706
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.09442,58630
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.9037,141471
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.75583,64099
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.82994,58688
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.5515,58688
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.42665,58688
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.3056,58611
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,361.732,2.19452e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,50.9245,331948
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.0977,241075
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1033.24,6.05109e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,686.711,3.9161e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,696.342,4.00968e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,67.7964,400530
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.1881,210799
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.8773,69820
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,56.9077,350084
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_126.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_126.txt
deleted file mode 100644
index ea6400d03a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_126.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,512.889,2.01624e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.56292,38992
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.40367,39049
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1431.3,6.29537e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.60188,47503
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.86101,47420
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.7532,175943
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,516.412,2.62336e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.93544,50769
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.83394,50788
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,856.837,4.57126e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.06059,54186
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.71297,54205
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,29.5667,177722
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,354.288,2.10221e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.15538,56361
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.41349,56361
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,589.817,3.61846e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.8827,58957
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,3.13194,58976
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.2826,135265
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.32367,59049
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.12677,59106
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,5.95445,59163
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.22271,58840
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.30033,58783
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,351.258,2.19612e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,51.0832,328802
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,37.5998,249477
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1048.34,6.18942e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,697.459,4.03406e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,709.916,4.14341e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,69.5421,420211
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,29.7712,195854
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.08736,65243
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,57.317,345781
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_127.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_127.txt
deleted file mode 100644
index 11dacf8858..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_127.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,513.644,2.01811e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.59313,38800
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.09014,38819
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1459.44,6.36232e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.52418,47452
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.50447,47425
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,30.1814,170372
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,506.56,2.53914e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.60098,50870
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.42729,50870
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,864.817,4.56692e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.51477,54332
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.582,54385
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.9645,184505
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,371.805,2.18465e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.42529,56397
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.48629,56397
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,621.968,3.79972e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.37768,59376
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.3429,59258
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,19.0982,142744
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.49858,64460
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.76738,59487
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.03665,59411
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.6974,59011
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.14829,58801
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,365.736,2.25533e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,53.5463,346146
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,39.5809,265556
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1101.98,6.49844e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,732.589,4.23126e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,745.956,4.34033e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,72.0537,440311
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.4142,216953
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.3084,69893
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,57.1276,339492
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_128.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_128.txt
deleted file mode 100644
index cf76f32f4c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_128.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,490.555,1.91277e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.76923,38268
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.96098,38249
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1422.48,6.18362e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.65071,46548
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.8917,46539
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.8054,181739
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,502.785,2.53056e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.79368,49969
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.61884,50007
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,822.43,4.37771e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.58405,53441
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.36056,53498
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,27.7038,176948
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,361.565,2.11351e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.89698,55445
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.83042,55502
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,585.627,3.53768e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.72002,58036
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.98949,58112
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.8197,127234
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.31596,58169
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.9045,58284
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.18161,58149
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.77481,57881
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.7898,57805
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,359.041,2.17864e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,53.103,339081
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.7997,261123
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1104.5,6.39538e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,718.275,3.94992e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,740.786,4.25646e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,74.8935,457030
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.3358,210152
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.5654,69967
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,56.8306,347219
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_129.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_129.txt
deleted file mode 100644
index a6f28a7c60..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_129.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,488.551,1.89905e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.07681,38286
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.20149,38321
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1422.93,6.17546e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.76018,46702
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.88731,46640
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.1803,176017
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,519.02,2.61002e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.01589,50031
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.83906,50050
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,853.518,4.55277e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.62392,53482
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.36114,53516
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,27.9432,177270
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,359.421,2.09654e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.21755,55503
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.3411,55390
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,584.796,3.46279e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.82331,58295
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.24235,58232
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.4951,140724
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.48997,63602
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.69624,58402
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.33364,58325
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.03267,58021
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.19041,58021
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,357.807,2.07422e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.1799,335929
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,39.9407,245484
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1110.51,6.37089e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,724.779,4.04036e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,747.887,4.3069e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,73.0584,435819
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,33.3917,209657
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.0744,64045
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,57.0774,348729
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_130.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_130.txt
deleted file mode 100644
index 636e187079..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_130.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,489.265,1.90801e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.08052,38420
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.5533,38458
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1444.06,6.23284e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.61314,46778
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.26415,46807
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.4642,174262
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,507.373,2.5221e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.76619,49995
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.34719,50033
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,852.637,4.48079e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.97144,53488
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.44322,53431
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,29.6332,192769
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,370.72,2.17036e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.90097,55529
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.18888,55396
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,622.862,3.76753e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,3.12874,58209
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.49246,58225
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.269,128353
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,6.09938,63426
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.99973,58434
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.08824,58358
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.12934,58185
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.18916,58185
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,363.332,2.21334e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.9078,336453
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,40.3138,262016
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1166.64,6.79804e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,759.293,4.23371e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,778.059,4.47344e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,76.8648,457931
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,33.4184,215316
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.80722,64270
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,57.8009,337244
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_131.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_131.txt
deleted file mode 100644
index 50c3eb9ec4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_131.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,522.794,2.04511e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.73265,38878
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.18905,38839
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1523.97,6.65219e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.6038,47678
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.24581,47691
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.8813,166415
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,533.498,2.69408e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.74501,51110
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.20937,51126
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,942.318,4.9949e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,3.15937,54300
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.61883,54300
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,31.5194,201945
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,371.802,2.20157e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.64565,56288
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.31787,56326
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,612.351,3.74904e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.79723,59205
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.43208,59202
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.1139,124537
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,6.90027,69758
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.78789,59316
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.70946,59240
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.87174,52934
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.19189,52934
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,366.553,2.23326e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,55.254,352097
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,39.9132,271326
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1214.21,7.11476e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,801.81,4.56785e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,812.935,4.72435e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,71.2519,429776
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.3433,201451
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.90617,92328
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,57.6299,357340
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_132.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_132.txt
deleted file mode 100644
index d2c262d04f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_132.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,525.37,2.06192e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.96747,38705
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.06894,38629
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1520.58,6.66722e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.01519,47444
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.61154,47444
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.7819,170612
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,536.154,2.72908e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.54613,51080
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.84213,51191
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,932.92,4.99892e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.20914,54418
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.4404,54476
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,30.1968,197082
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,367.303,2.1788e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.35217,56422
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.98562,56365
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,610.051,3.74586e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.97608,59053
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.88159,59053
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.3418,130176
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.61065,64309
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.90408,59259
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.24766,59183
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.49644,58975
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.90756,58818
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,378.999,2.3384e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,52.0377,338169
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.9645,264915
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1200.99,7.04572e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,787.969,4.44044e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,804.791,4.6697e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,71.1626,432984
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.0567,210482
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.88435,75320
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,58.4652,354827
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_133.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_133.txt
deleted file mode 100644
index e8443470d8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_133.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,520.888,2.04579e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.47944,38877
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.12322,38972
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1514.43,6.63255e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.56204,47815
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.07541,47828
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,28.3388,163490
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,527.728,2.63761e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.54604,51022
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.10674,51096
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,957.354,5.10974e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.78369,54366
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.50894,54377
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,29.6448,202412
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,367.438,2.18595e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.01608,56323
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.23202,56400
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,616.571,3.78567e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.00635,59071
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.06529,59163
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.1188,129509
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.1254,59201
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.01387,59296
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.6084,59201
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.34773,59049
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.6952,64896
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,355.626,2.21807e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,52.0829,340629
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.1219,259621
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1158.3,6.83392e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,761.462,4.27722e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,778.415,4.54375e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,70.3622,430421
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.3293,207125
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.1576,86781
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,53.7575,330000
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_134.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_134.txt
deleted file mode 100644
index f156a272bc..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_134.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,518.225,2.03857e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.76772,38705
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.54213,38839
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1511.44,6.60475e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.64194,47527
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.6677,47581
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.5123,175937
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,531.752,2.68169e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.50665,51028
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.54812,51083
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,955.147,5.09862e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,3.28308,54426
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.75832,54345
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,30.6361,202254
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,367.801,2.18166e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.3061,56250
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.37115,56270
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,613.327,3.75759e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,3.20113,59033
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.25854,58995
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.5104,129265
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.39858,59068
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.01976,59087
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.26823,59087
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.39263,59087
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.12753,58550
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,353.751,2.19414e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,53.7501,344014
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.7621,259321
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1233.71,7.25328e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,810.47,4.59346e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,829.939,4.82735e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,73.6112,439468
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.6467,205827
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.57212,81121
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,58.215,356117
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_135.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_135.txt
deleted file mode 100644
index d93aac33b8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_135.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,497.796,1.93042e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.85409,38268
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.4189,38306
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1510.04,6.51319e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.77071,47233
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.40694,47306
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,30.0603,156603
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,532.12,2.66163e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.06754,50490
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.41295,50490
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,944.021,4.99783e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.60625,53848
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.67332,53906
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.0045,183834
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,370.626,2.18061e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.13735,55695
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.94213,55733
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,611.587,3.70482e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.46552,58631
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.97964,58419
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.2842,122698
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,6.12955,63752
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.06546,58742
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.23947,63617
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.50137,58281
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.79949,58186
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,346.122,2.12271e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,48.9698,309221
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,35.7981,229550
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1113.87,6.51033e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,721.95,4.05898e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,743.515,4.2954e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,68.1563,409251
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,29.4963,188728
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.1106,69993
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,54.0631,321801
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_136.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_136.txt
deleted file mode 100644
index 0b546e9036..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_136.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,500.05,1.94081e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.03905,38192
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.37129,38211
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1499.93,6.50211e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.79157,47249
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.8422,47278
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,31.4395,171501
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,527.901,2.63924e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.83384,50487
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.70271,50523
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,932.858,4.94098e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.81124,53920
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,3.28826,53958
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.963,178504
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,374.848,2.20488e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.36002,55907
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.26815,55963
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,619.314,3.75355e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.53089,58496
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.90792,58512
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.6418,134224
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.1248,58722
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.16875,58741
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.60833,63770
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,6.17848,58588
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.31312,58186
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,358.901,2.19902e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,53.4904,348181
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.9841,258346
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1228.34,7.1658e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,798.616,4.50051e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,818.663,4.72786e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,72.3452,442673
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.4503,215227
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.604,91273
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,58.0678,353514
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_137.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_137.txt
deleted file mode 100644
index 18601a156b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_137.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,494.673,1.94348e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.66638,38664
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.33449,38588
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1494.56,6.44786e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.65723,47325
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.41615,47359
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,29.287,165234
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,522.499,2.58653e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.65298,50408
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.17503,50466
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,956.173,4.99662e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,3.09796,53805
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.70279,53787
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,29.8641,194485
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,368.763,2.16565e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.92277,55739
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.94981,55758
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,618.426,3.73154e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.29106,58531
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.45537,58531
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.5694,117081
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,6.14133,63675
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.84811,58722
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.87297,58645
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.98089,58472
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.19059,58242
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,356.023,2.18003e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.9909,349138
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,41.5442,275170
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1235.38,7.14779e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,813.297,4.61604e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,816.186,4.7365e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,71.9996,429052
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.682,212127
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.3575,97760
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,58.9096,356092
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_138.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_138.txt
deleted file mode 100644
index 18e2ee76b7..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_138.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,493.702,1.93113e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.48485,38344
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,1.38601,38382
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1496.88,6.37549e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.5277,47233
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.27279,42369
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,30.069,166361
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,518.673,2.5721e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.63791,50474
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.15097,50529
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,951.896,4.78278e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.89698,53722
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.26968,53747
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,31.9208,199983
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,369.81,2.17044e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.24215,55713
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.09726,55636
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,622.511,3.7004e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,3.03233,58453
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.14261,58376
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.5437,133799
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.21081,58586
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.81855,58663
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.61397,58509
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.59887,58261
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.24077,58185
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,347.875,2.13694e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,52.3979,342073
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.5801,251027
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1193.2,6.96721e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,781.872,4.35658e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,800.611,4.63655e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,73.1664,436793
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.8312,199875
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.1147,75365
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,54.928,332389
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_41.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_41.txt
deleted file mode 100644
index 070e97bed1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_41.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,504.874,1.93218e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.2764,37870
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.84478,37889
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1821.15,7.86918e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.54677,48007
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.33131,48083
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,32.4474,181601
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,818.136,4.06536e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.06821,50905
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.92738,50924
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,1332.75,7.09964e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.33106,54383
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.84814,54380
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.1628,179710
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,326.011,1.91353e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.13013,56167
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.2812,56186
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,542.389,3.25438e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.66892,58802
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.12238,58840
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.9176,141070
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.40329,64347
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.65055,58916
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.26913,58916
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.00998,58802
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.06369,58631
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,359.221,2.20974e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,39.7362,254622
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.5187,230487
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1030.71,6.01365e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,594.95,3.36356e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,557.701,3.13541e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,72.117,434533
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,21.496,157302
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.77318,59332
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,60.2474,367981
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_43.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_43.txt
deleted file mode 100644
index adee4a649a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_43.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,512.584,1.97652e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.32862,38022
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,3.20465,37983
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1806.09,7.89462e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.81548,48135
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.55054,48212
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,35.4837,203280
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,808.325,4.12335e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.66647,51342
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.69652,51285
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,1131.13,6.20321e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.82613,55541
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.96615,55541
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.3146,183952
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,322.36,1.94429e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.08904,57274
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.06011,57350
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,541.034,3.34225e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.92543,59851
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.17311,59867
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.9872,149334
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,4.56032,59981
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.84616,59904
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.16059,59904
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.07836,59561
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.96989,59504
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,366.098,2.28511e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,39.3923,252201
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,38.8196,240015
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1017.86,6.04683e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,617.427,3.51735e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,583.057,3.35479e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,73.9926,441140
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,22.9876,163395
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.987,64947
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,62.2584,391012
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_45.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_45.txt
deleted file mode 100644
index 151d408eed..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_45.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,502.579,1.9411e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,1.98366,38272
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.89252,38310
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1802.71,7.87184e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.703,48252
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.11467,48291
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,34.2146,192488
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,780.652,3.96253e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.01461,51323
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.35058,51343
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,1158.69,6.33308e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.46347,55465
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.77848,55541
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.3453,194165
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,327.96,1.96248e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.31067,57084
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.46417,57084
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,542.253,3.3329e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.80732,59851
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.14159,59793
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.8409,143814
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.34982,65415
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.65468,59888
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.04942,59888
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.50226,59622
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,8.53594,65470
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,358.603,2.2467e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,45.8217,286536
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,37.9513,239505
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1017.63,6.04958e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,601.659,3.44212e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,576.172,3.33753e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,74.0101,448507
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,26.2698,164121
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.7095,60231
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,61.3979,381951
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_231.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_231.txt
deleted file mode 100644
index b1900b41c2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_231.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,312.955,1.05814e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.29102,32634
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.92555,32750
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1270.25,4.95079e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.64185,42084
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.93026,42046
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,32.9684,164910
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,375.799,1.72684e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.62242,45265
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.86757,45437
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,628.419,3.06087e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.55881,49018
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,1.97983,49095
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,24.1804,147478
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,236.215,1.27624e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.07611,50990
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.91714,51009
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,416.23,2.3139e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.9468,53167
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.39957,53167
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,19.4841,128208
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.49903,58078
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.9957,53450
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.69527,53298
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.90681,52994
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,6.96046,52917
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,398.579,2.22238e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,75.6271,418107
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,53.4082,301014
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,2181.08,1.18919e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,1541.07,7.72153e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,1518.65,7.74613e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,138.242,667792
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,43.985,200120
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,16.4329,76086
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,126.613,598657
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_232.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_232.txt
deleted file mode 100644
index acea102c80..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_232.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,320.765,1.0895e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.74497,32444
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.91287,32520
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1271.88,4.91642e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.52866,42009
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.32443,42066
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.5942,164724
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,374.875,1.70587e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.75916,45227
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.94523,45284
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,631.129,3.06489e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.68424,49132
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.28821,48995
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,27.4421,157819
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,231.786,1.24844e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.17838,50775
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.17326,50736
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,418.555,2.32751e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.8323,52971
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.43531,53009
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.7868,122228
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.95036,53082
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.27055,53253
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.92836,57843
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.84128,52968
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.47956,52968
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,408.421,2.2672e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,76.7764,449207
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,54.9257,305927
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,2268.42,1.23557e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,1680.26,8.35215e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,1579.97,8.0778e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,134.762,654124
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,42.9408,220383
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,14.0336,66829
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,119.188,566439
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_233.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_233.txt
deleted file mode 100644
index 47aa85aa79..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_233.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,364.281,1.42832e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.01979,37783
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.91876,37859
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1724.95,7.72746e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.61807,48656
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.72616,48711
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.8694,198585
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,485.892,2.51283e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.7669,51802
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.50969,51876
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,882.543,4.87772e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.41816,55778
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.91368,55832
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,30.075,201041
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,301.008,1.83866e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,3.06062,57411
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.58459,57355
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,558.599,3.47618e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.79893,59907
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.99259,59945
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,19.0353,143544
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.68991,65276
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,2.00786,60020
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.53729,60020
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.20431,59634
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.20196,59403
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,364.101,2.26809e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,56.8705,359483
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,41.6894,285564
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1542.04,9.30453e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,983.446,5.5907e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,1027.43,6.08045e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,76.4432,458940
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.5815,196982
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.17091,76617
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,61.1893,380972
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_234.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_234.txt
deleted file mode 100644
index 19f49e9d48..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_234.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,359.171,1.3654e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.56808,36755
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.57861,36830
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1720.43,7.51741e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.55141,48008
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.80565,47875
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,29.423,167745
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,477.2,2.42612e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.41721,51394
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.20194,51432
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,887.624,4.77616e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.74392,55249
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.76305,55283
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,30.7709,205785
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,301.656,1.81816e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.88015,57099
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.21406,57041
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,562.265,3.49027e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.89061,59568
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.32229,59603
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.4432,119295
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.97577,65011
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.85704,59733
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.36932,64953
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.80956,59428
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.87399,59198
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,354.283,2.21567e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,53.9692,353895
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,39.3985,273572
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1682.64,1.00985e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,966.772,5.42835e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,994.348,5.79107e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,82.3764,478623
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.1803,209638
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,10.9806,96473
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,66.7879,398824
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_235.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_235.txt
deleted file mode 100644
index f5857d3264..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_235.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,359.179,1.2212e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.8443,33172
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.75128,33192
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1725.86,7.02108e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,2.15976,45546
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.74376,45603
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,35.1932,192282
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,486.73,2.42326e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.32251,49345
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.11051,49344
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,874.152,4.65541e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.14318,53834
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,3.4425,53834
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,27.2372,171817
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,304.18,1.78209e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.93003,55659
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.31419,55659
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,556.548,3.35208e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.22421,58182
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.0956,58217
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.3645,116615
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.83999,63656
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.86607,58438
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,5.43334,58286
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.87446,57981
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.7042,57854
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,420.697,2.53764e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,85.3445,534738
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,60.8316,387686
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,3076.32,1.7793e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,2322.36,1.18957e+07
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,1926.57,9.90526e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,147.311,701287
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,47.9584,230534
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,14.2147,71974
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,134.215,637238
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_236.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_236.txt
deleted file mode 100644
index b38950da77..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_236.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,369.184,1.30472e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,1.83851,34353
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.51925,34373
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1814.45,7.54365e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.95391,46542
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.14664,46540
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,32.855,187012
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,507.61,2.53206e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.52722,49910
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.48972,50117
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,946.582,5.02218e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.71148,54187
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.50529,54206
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,25.3727,162878
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,339.38,1.9942e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.9773,56181
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.00066,56181
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,600.8,3.65449e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.04114,58747
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,1.99851,58760
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,19.3574,159457
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.00924,64151
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.56626,58838
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.17109,58838
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.10332,58765
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.09669,58536
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,416.965,2.55865e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,76.3945,481875
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,53.5784,341610
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,2759.9,1.6179e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,1799.2,9.41621e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,1565.91,8.37989e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,126.676,649481
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,41.5973,232891
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,12.366,69394
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,110.356,555019
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_237.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_237.txt
deleted file mode 100644
index 3c5e6c169a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_237.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,366.675,1.45154e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,1.87509,37821
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.80517,37879
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1807.03,8.02866e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.63839,48695
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.56049,48752
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,33.7969,185595
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,512.094,2.68803e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.58895,52031
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.87138,52047
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,941.746,5.20451e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.13765,55821
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,3.10807,55782
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.328,190552
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,330.901,2.01841e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,2.17483,57648
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.25227,57686
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,595.453,3.71972e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.89838,60127
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.00248,60085
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,17.9617,138566
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.71676,65494
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.72594,60254
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.70641,65417
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.9998,59796
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.73671,59723
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,358.565,2.24855e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,51.6014,332793
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,37.4462,251934
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1604.66,9.70904e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,862.655,4.92921e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,894.783,5.32491e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,69.249,427305
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,31.0203,213623
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,11.1213,98381
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,55.7902,348237
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_238.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_238.txt
deleted file mode 100644
index dbc1247487..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_238.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,361.396,1.42936e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,3.077,38038
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.85547,38057
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1801.71,8.07173e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.50293,48844
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,1.82636,48860
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,32.1758,175635
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,505.379,2.6346e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,1.64517,52142
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,1.96376,52219
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,943.341,5.17271e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,1.88607,56170
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.52344,56131
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,27.3298,184853
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,332.621,2.02922e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.81202,57654
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,1.94287,57596
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,597.906,3.72187e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,2.1228,60207
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.09502,60223
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,19.2042,151258
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.2175,65703
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.67349,60463
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,7.27326,60313
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,5.12176,60010
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.77034,59780
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,332.125,2.08543e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,40.8915,282376
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,31.3805,224408
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1229.22,7.34539e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,658.089,3.74298e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,677.838,3.98656e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,56.4301,354013
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,25.7725,165318
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,9.87913,71131
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,45.0084,284986
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_239.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_239.txt
deleted file mode 100644
index cc13932a92..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_239.txt
+++ /dev/null
@@ -1,108 +0,0 @@
-Conv1,365.933,1.40138e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,2.80164,36850
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,2.55448,36888
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,1806.48,7.93584e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.88536,48163
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,2.64411,48201
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool1,35.6716,207285
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,512.369,2.65332e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,2.04005,51632
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,2.16469,51647
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,935.445,5.14346e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,2.03416,55351
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,2.77188,55389
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Pool2,28.7423,188823
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,338.056,2.03138e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,1.87183,57064
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,2.23042,57199
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Conv6,592.54,3.65101e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,1.68578,59431
-Add6_f2h,0,0
-Add6_h2f,0,0
-Tanh6,2.24219,59489
-Tanh6_f2h,0,0
-Tanh6_h2f,0,0
-Pool3,18.0715,124954
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,5.97676,64936
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add7,1.76684,59600
-Add7_f2h,0,0
-Add7_h2f,0,0
-Softmax1,6.85672,59600
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
-ArgMax1,4.97948,59450
-ArgMax1_f2h,0,0
-ArgMax1_h2f,0,0
-Select1,7.37134,59450
-Select1_f2h,0,0
-Select1_h2f,0,0
-Contract1,354.286,2.1935e+06
-Contract1_f2h,0,0
-Contract1_h2f,0,0
-tensorReduce1,54.0783,348349
-tensorReduce1_f2h,0,0
-tensorReduce1_h2f,0,0
-tensorMap11,39.247,244872
-tensorMap11_f2h,0,0
-tensorMap11_h2f,0,0
-Conv7,1792.85,1.07184e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Conv8,1012.73,5.58673e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Conv9,1045.6,5.94124e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-tensorMap21,83.6212,481961
-tensorMap21_f2h,0,0
-tensorMap21_h2f,0,0
-tensorReduce2,32.2344,206842
-tensorReduce2_f2h,0,0
-tensorReduce2_h2f,0,0
-tensorReduce3,11.1349,73456
-tensorReduce3_f2h,0,0
-tensorReduce3_h2f,0,0
-tensorMap22,69.2339,398246
-tensorMap22_f2h,0,0
-tensorMap22_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_120.txt
deleted file mode 100644
index ff28fccb9e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_120.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,337.4,3.68478e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.237007,15382.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.184344,15371.1
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.68,56288.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,495.317,5.5504e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.231612,16226.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.183714,16224.2
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.68191,47875.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,180.324,2.10248e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.20503,16949.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.108639,16953.5
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,262.307,3.23845e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.219605,17250.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.184172,17220.3
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,184.059,2.2692e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.218521,17340.5
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.156735,17319.5
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.96304,34615.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.5155,23479.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.161062,17309.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.16766,82622.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_151.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_151.txt
deleted file mode 100644
index bae151beca..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_151.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,190.193,2.05276e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.224536,14885.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.174028,14864.8
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.67015,56078.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,227.773,2.5145e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.215689,15782.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.167237,15790.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.83595,49012.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,105.987,1.20833e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.207775,16320.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.15142,16319.2
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,131.067,1.56614e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.212713,16587.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.152012,16572.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,96.4066,1.14964e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.208921,16684.5
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.153449,16688.5
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.84486,33369.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.49149,21018.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.174572,16683
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.27003,81690.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_152.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_152.txt
deleted file mode 100644
index a6d85b7e69..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_152.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,192.983,2.08398e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.228424,14967.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.178303,14969.6
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.74804,59087.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,229.723,2.54853e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.219602,15849.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.176313,15847
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.95143,49174.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,106.015,1.21477e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.217896,16375.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.162725,16361.7
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,132.317,1.56904e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.215007,16595.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.15861,16580.5
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,97.0358,1.15623e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.211979,16686.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.159823,16686.3
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,2.02821,33371.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.55131,20448
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.221237,17272.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.25398,80835.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_153.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_153.txt
deleted file mode 100644
index efc436221c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_153.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,178.417,1.90236e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.226863,14723.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.178841,14731.4
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.61066,54457.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,226.111,2.46688e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.218232,15622.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.175753,15628
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.81594,47730.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,105.093,1.17798e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.208955,16155.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.156338,16157.3
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,130.75,1.54768e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.206699,16418.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.155398,16405.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,96.3056,1.13727e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.212475,16512.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.153935,16485.7
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.83618,32979
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.52322,19941.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.182194,16489.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.29497,80765.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_154.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_154.txt
deleted file mode 100644
index 00f06e6d16..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_154.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,179.945,1.92348e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.225087,14961.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.180134,14929.5
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.66463,55779.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,225.428,2.49274e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.214981,15791.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.176168,15803.6
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.7666,47423.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,105.113,1.19446e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.210402,16298.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.154527,16294.4
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,130.745,1.5617e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.213,16525.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.15965,16531.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,96.5646,1.13679e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.222831,16602.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.158901,16594.5
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.87082,32332.7
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.50962,19449.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.241307,17164.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.38585,82974.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_155.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_155.txt
deleted file mode 100644
index 5fd4c898cb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_155.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,225.41,2.4281e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.22541,14922.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.176754,14929.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.75537,55816
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,319.68,3.55718e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.222315,16190
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.177609,16157.6
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.88735,49319.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,128.273,1.48664e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.210946,16714.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.155471,16712.4
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,190.818,2.30027e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.215429,16672
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.154931,16660.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,126.181,1.50897e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.211749,16716
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.153986,16695.1
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.89296,33399.7
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.53906,20586.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.149202,16712.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.26271,81001.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_156.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_156.txt
deleted file mode 100644
index bbc1088bcd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_156.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,224.225,2.43331e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.221042,15080.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.171714,15086.5
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.85658,58264
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,332.686,3.76477e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.219249,16390.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.177142,16384.5
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.94905,52366.1
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,126.868,1.48313e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.209633,16919.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.158117,16915.9
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,188.538,2.29799e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.212866,16876
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.159141,16860.8
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,124.236,1.50017e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.205157,16906.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.1519,16899
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.88824,32924.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.47725,20776.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.158312,16901
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.36032,83468
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_157.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_157.txt
deleted file mode 100644
index cec1b02990..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_157.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,228.935,2.55515e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.236312,15404.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.198546,15397.2
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.53177,56761.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,333.316,3.83079e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.228549,16594.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.173087,16566.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.9573,51313.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,140.057,1.64818e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.220965,17157
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.158402,17149.4
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,185.221,2.31232e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.220901,17351.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.16015,17353.1
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,132.938,1.64365e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.224849,17416.5
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.156607,17412.7
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.90251,34808.7
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.50488,21410.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.184002,17410.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.37759,85043.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_158.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_158.txt
deleted file mode 100644
index 65dd443832..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_158.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,215.378,2.27857e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.217918,14646.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.178741,14652.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.9282,57868.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,316.597,3.45691e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.214648,15928.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.172655,15926.2
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,4.01175,49420.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,127.865,1.46021e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.213186,16448.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.15798,16429.3
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,192.923,2.28677e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.210629,16387.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.157673,16397.2
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,126.591,1.46955e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.212421,16428.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.1622,16430
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.95539,32026.5
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.56776,20221.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.179224,16435.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.40989,81173.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_159.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_159.txt
deleted file mode 100644
index bcad050061..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_159.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,214.73,2.27431e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.234056,14695.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.180958,14716.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.59181,58061.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,328.568,3.60365e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.218437,16028.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.181429,16020.8
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.89355,48878.5
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,128.364,1.46049e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.209243,16527
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.168834,16513.5
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,192.55,2.29348e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.213163,16511.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.158575,16488.5
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,126.548,1.47275e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.211439,16494.5
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.170008,16481.2
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.8865,32958.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.52285,20251.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.189496,16481.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.2397,78864.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_160.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_160.txt
deleted file mode 100644
index fa9616a04c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_160.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,225.508,2.43633e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.228946,15126.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.208084,15116.8
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.87676,58101
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,329.712,3.7038e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.22812,16265.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.171499,16277.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,4.1347,53740.5
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,140.863,1.64544e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.215406,16803.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.156818,16782.7
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,186.483,2.28163e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.217285,17000.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.157375,16977.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,133.798,1.62205e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.210635,17017.5
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.156533,17009.9
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,2.22091,33981.5
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.59925,20940.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.164171,17017.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.21735,81657.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_161.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_161.txt
deleted file mode 100644
index ee0b75a4e8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_161.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,235.923,2.61302e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.230625,15330.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.18556,15318.8
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.80264,59670.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,342.963,3.89751e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.22461,16497.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.169845,16488.6
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.94556,50348
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,140.457,1.66547e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.21917,17071.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.157157,17065.7
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,185.744,2.3113e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.221905,17271.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.161519,17248.5
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,133.716,1.64908e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.217829,17326.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.157448,17322.5
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,2.01388,34639.3
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.54772,21361.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.159468,17303.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.33729,84475.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_162.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_162.txt
deleted file mode 100644
index 37bce90b93..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_162.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,240.706,2.69778e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.230961,15507
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.17644,15512.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.5114,57017.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,342.453,3.94576e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.214651,16661.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.174081,16650.3
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.93383,52551.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,139.38,1.65061e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.211301,17204.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.15572,17200.5
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,184.095,2.30956e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.217528,17414
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.161151,17416.1
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,132.362,1.64178e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.21267,17474
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.156981,17466.4
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.85461,34934.7
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.48784,21484.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.17036,17458.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.28267,84640.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_163.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_163.txt
deleted file mode 100644
index 33ad0446b5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_163.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,235.701,2.59326e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.231121,15368.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.176635,15372.3
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.83085,59848.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,342.133,3.92817e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.219285,16562.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.173977,16559.1
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,4.01668,52238.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,140.035,1.66625e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.211371,17128
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.153938,17116.5
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,185.731,2.32171e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.21819,17312.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.166744,17316.2
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,133.464,1.65251e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.211656,17390.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.155691,17375.1
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.92438,34744.5
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.53433,20733.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.207566,17367.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.29477,85021.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_164.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_164.txt
deleted file mode 100644
index 5f6fe09a65..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_164.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,236.787,2.61292e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.231442,15246.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.179746,15238.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.85249,60955.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,344.584,3.88198e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.22509,16425.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.171592,16414.1
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,4.00496,51666.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,140.511,1.65112e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.222171,16977.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.154984,16981.2
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,185.223,2.29166e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.217253,17197.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.155346,17178.1
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,133.207,1.62981e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.21491,17272.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.154895,17265.3
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,2.02137,34521.1
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.5765,21294.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.158268,17267.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.23312,83639.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_165.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_165.txt
deleted file mode 100644
index 2de11a32b2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_165.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,225.813,2.45634e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.233342,15120.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.180479,15120.5
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.91967,59642.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,341.688,3.82968e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.224363,16277.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.18157,16264
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,4.14589,52613.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,139.652,1.62184e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.21387,16796.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.157282,16792.4
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,185.498,2.27214e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.217282,17019.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.159055,17003.8
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,132.844,1.61159e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.213957,17076.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.164101,17055.2
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,2.00697,34112.3
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.59257,21871.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.187244,17034.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.17514,80647.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_166.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_166.txt
deleted file mode 100644
index a3e6094c95..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_166.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,227.965,2.49644e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.233832,15110.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.183643,15089.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.71041,58809.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,338.785,3.79768e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.224769,16292.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.177023,16275.6
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.87674,49693.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,139.793,1.63383e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.216245,16849.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.153941,16843.8
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,185.055,2.27607e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.221573,17061.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.161151,17063.1
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,132.606,1.61497e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.216773,17149.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.156645,17137.8
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.88502,34245.1
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.53558,21072.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.190939,17103.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.29871,82918.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_167.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_167.txt
deleted file mode 100644
index 6352338301..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_167.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,229.121,2.46673e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.237256,14986.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.180729,14979.2
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.91751,59875
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,338.076,3.76931e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.226971,16207.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.178975,16190.2
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,4.16684,53615.7
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,140.639,1.62088e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.216751,16727.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.158767,16737.2
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,185.833,2.2697e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.423038,18195.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.164844,16940.8
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,133.264,1.61332e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.218312,16998.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.160332,16998.2
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,2.05673,34000.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.54658,20967.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.178684,17000
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.39485,83929.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_168.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_168.txt
deleted file mode 100644
index 4939acc2b9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_168.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,221.258,2.3963e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.231151,15147.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.180821,15134.3
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.89041,59736.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,339.061,3.79971e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.221039,16271.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.177874,16271.6
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,4.05763,51343.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,137.873,1.60014e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.215893,16805.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.158648,16800
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,184.004,2.25579e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.21884,17009.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.159909,17011.3
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,131.748,1.60159e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.216965,17066.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.160671,17051.4
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.9999,34102.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.55789,20987.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.165359,17047.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.33899,83291.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_261.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_261.txt
deleted file mode 100644
index d6b5d9b3b2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_261.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,193.96,2.08465e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.223755,14883.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.181375,14887.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,5.19233,57181.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,263.654,2.89224e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.206002,15862
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.181723,15858.1
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.80138,47581.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,125.379,1.44447e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.201644,16346.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.156815,16342.9
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,180.376,2.12976e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.209685,16576.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.158718,16549.7
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,127.509,1.53368e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.206152,16654.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.155881,16631.7
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.80793,33276.7
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.49461,20365.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.188994,16641.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.29749,80568.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_262.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_262.txt
deleted file mode 100644
index 114fd5b401..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_262.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,199.126,2.1474e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.224386,14893.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.182623,14902.8
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.89175,58873.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,267.198,2.93259e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.217196,15841.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.177897,15843
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.99445,49056.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,126.885,1.46094e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.207109,16293.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.160601,16303
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,182.224,2.17613e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.210142,16508.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.158648,16504.5
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,129.398,1.54729e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.205438,16578.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.156853,16565.3
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.98616,33117.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.55359,20260.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.191097,16559.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.27832,80997.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_263.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_263.txt
deleted file mode 100644
index 38f5c465ae..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_263.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,237.756,2.58801e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.211435,15177.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.177055,15162
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.77029,55082.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,325.669,3.62553e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.210341,16274.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.174693,16276.3
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.98304,50490.5
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,166.14,1.95866e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.203864,16771.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.156665,16773.5
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,253.933,3.09534e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.209547,16969.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.160223,16948.5
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,176.456,2.16171e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.207643,16997.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.155656,16996.9
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.97852,33971.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.53225,21108.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.145474,16990
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.29311,81391.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_264.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_264.txt
deleted file mode 100644
index dce8d97e08..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_264.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,235.998,2.58229e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.220895,15223
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.182578,15202.1
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.71929,59216.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,325.495,3.59715e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.213483,16263.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.175394,16267.2
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.89997,48840.1
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,166.348,1.94807e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.202837,16815.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.151289,16789
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,253.407,3.0948e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.205586,16987.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.156341,16989.7
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,176.472,2.15066e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.205435,16995.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.159058,16989.5
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.90611,33983
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.55878,21071.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.197848,16995.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.32883,82994.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_265.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_265.txt
deleted file mode 100644
index b552df3922..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_265.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,236.262,2.55492e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.213294,15194.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.172805,15192.6
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.63543,58406.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,323.781,3.60521e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.209067,16285.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.172606,16264.6
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.87703,49669.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,165.715,1.94638e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.203263,16813.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.156082,16817.6
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,253.168,3.09459e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.210456,16974.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.151186,16970.8
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,175.698,2.1386e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.203525,17033.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.153666,17016.4
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.89708,34036.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.58343,21124.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.286414,17699.3
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.32412,82374.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_266.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_266.txt
deleted file mode 100644
index 22b32b349f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_266.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,260.084,2.82767e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.21558,15131.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.173804,15140.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,5.15458,60615.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,360.348,4.0013e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.211848,16307.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.176095,16302.2
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,4.2723,53531.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,162.996,1.91715e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.206091,16838.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.158556,16827.2
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,239.795,2.92844e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.209947,16970
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.159752,16962.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,167.049,2.0305e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.206168,17022.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.158207,17011.3
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,2.10691,34018.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.61344,21134
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.149477,17000.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.2582,82237.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_267.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_267.txt
deleted file mode 100644
index 078360ac09..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_267.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,252.535,2.80306e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.217663,15443.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.173861,15437.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,4.64457,58441
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,353.487,3.9959e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.211019,16588.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.174386,16588.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,3.83916,50282
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,160.342,1.90628e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.207455,17071.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.159116,17073.6
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,235.642,2.91771e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.207208,17231.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.15366,17212.3
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,163.899,2.026e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.2098,17251.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.159298,17246.2
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.88925,34482.7
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.50502,21389.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.187544,17253.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.44141,85096.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_268.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_268.txt
deleted file mode 100644
index 7ef17faf3c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_268.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,260.445,2.83985e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.216651,15159.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.179538,15152.1
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,5.12406,59811.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,359.9,3.99244e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.211919,16330.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.184728,16334.9
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,4.28685,56435.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,163.118,1.91288e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.208735,16854.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.165135,16848.4
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,239.929,2.92493e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.209388,16985.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.162031,16976
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,166.986,2.0313e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.205096,17001.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.160009,16994.1
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,2.10721,33995.9
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.58976,21106.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.155829,16999.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.24343,82294.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_269.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_269.txt
deleted file mode 100644
index d149a94dd8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_269.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,257.784,2.8414e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.218043,15340.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.180108,15340.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,5.07471,61314.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,357.124,4.0032e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.205896,16518.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.173925,16516.7
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,4.18427,52019.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,161.585,1.88286e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.210111,17031.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.158485,17026
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,237.55,2.92622e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.206741,17144.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.157791,17134.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,165.515,2.03034e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.209352,17161.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.159407,17163.9
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,1.97624,34313.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,1.54299,21315.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.159292,17162.7
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,6.30759,83087.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp32_perf_fp32_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp32_perf_fp32_120.txt
deleted file mode 100644
index 25cd9de0d0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp32_perf_fp32_120.txt
+++ /dev/null
@@ -1,63 +0,0 @@
-Conv1,575.774,5.82108e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.232444,14477.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Tanh1,0.207302,14456.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Pool1,12.7919,128991
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,1017.01,1.00097e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.223817,14906
-Add2_f2h,0,0
-Add2_h2f,0,0
-Tanh2,0.132729,14921.2
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Pool2,8.00501,91648.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,287.023,3.18789e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.21968,16490.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.14631,16490.1
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Conv4,396.7,4.79884e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.22079,17588.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.176054,17577
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Conv5,289.129,3.55552e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.211507,17881.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Tanh5,0.130201,17874.8
-Tanh5_f2h,0,0
-Tanh5_h2f,0,0
-Pool3,2.80131,41084.1
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,0.909982,22973.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,0.142185,17870.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Softmax1,7.7959,102068
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_120.txt
deleted file mode 100644
index 391282ad11..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_120.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,389.264,3.32015e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.483012,24306.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.370173,24322.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.00216,49953.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,519.829,4.84468e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.441751,27734.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.368817,27753.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.49479,55560
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,187.256,1.86581e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.429559,29112.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.341534,29112.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,260.793,2.80254e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.428932,29992.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.339095,29950.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,181.954,1.97205e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.417661,30502.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.337995,30509
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.74986,32041.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.70647,83156.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.92935,44395.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.270181,30508.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.68727,51425.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.0127,36299.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.277624,30641.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.41295,35288.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.43176,31835.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,38.5284,446826
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_151.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_151.txt
deleted file mode 100644
index 7b36fde138..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_151.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,249.629,1.96776e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.492637,21956.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.396279,21952.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.75587,45092.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,257.644,2.14873e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.439127,24433.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.373354,24445.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.29218,48932.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,136.118,1.22967e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.409822,25675.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.33148,25663.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,149.576,1.44381e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.394558,26700.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.337138,26719.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,110.543,1.09229e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.385643,27381.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.335415,27397
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.60113,27404.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.17822,69631
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.86901,38001
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.276433,27424
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.59152,43516.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.06847,30872
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.249387,26073.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.32902,30994.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.274226,27134.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.8362,409193
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_152.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_152.txt
deleted file mode 100644
index dc32d18e93..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_152.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,249.51,1.99001e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.515593,22258
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.383677,22246.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.72443,45690.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,267.601,2.27311e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.453923,24781.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.381284,24816
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.27455,49654.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,139.903,1.27269e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.416875,26016.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.341041,26020.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,158.912,1.54462e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.428375,27092.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.34789,27065.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,117.123,1.16661e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.405021,27677
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.339722,27677
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.80161,30316.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.25399,70292.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.82906,38380.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.283857,27738
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.73987,46457
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,0.978144,32720.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.273834,27910.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.41574,31870.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.456221,28963.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,36.5775,406615
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_153.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_153.txt
deleted file mode 100644
index 5838b4148e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_153.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,238.617,1.86414e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.656252,21677.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.429636,21685.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.13473,46761.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,251.066,2.09823e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.480643,24174.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.39256,24189.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.665,48410.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,135.146,1.2094e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.407645,25429.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.33402,25418
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,150.325,1.43901e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.412893,26413.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.335717,26432.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,111.506,1.0936e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.394468,27043.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.334455,27054.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.84151,32371.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.15433,68457.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.8628,37471.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.275672,27115.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.7518,44288.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.01497,31971.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.218929,27275.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.23204,31135.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.299947,28339.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,36.6818,395989
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_154.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_154.txt
deleted file mode 100644
index e285249a34..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_154.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,239.504,1.9089e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.650818,22087
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.299569,22075.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.39131,49930.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,255.645,2.17619e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.475823,24601.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.397444,24620.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.71863,49305.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,138.815,1.25627e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.411556,25868.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.325137,25884
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,158.938,1.53668e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.433379,26864.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.340388,26876
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,116.678,1.14687e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.417168,27485.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.334533,27493.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.78091,31746.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.30965,69865
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.90527,39089.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.283275,27570
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.8039,45151
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.0954,33544
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.285861,27719.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.41655,31690.8
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.380522,28765.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,36.7626,404634
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_155.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_155.txt
deleted file mode 100644
index 5e943f57f9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_155.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,276.643,2.24322e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.479946,22664.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.394717,22641.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.99556,47675.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,315.117,2.73945e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.431011,25580.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.386417,25577
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.63967,52256.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,143.233,1.33647e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.419882,26834.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.341028,26838.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,166.044,1.65032e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.415991,27803.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.343544,27800.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,122.109,1.20788e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.408822,28445.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.342884,28456.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.75296,29929.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.25413,73012.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.87899,39619.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.285656,28488
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.66321,46903
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.06091,33666.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.270271,28558.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.45742,32708
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.432266,29718
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.0684,418536
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_156.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_156.txt
deleted file mode 100644
index 05d5be47e5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_156.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,288.752,2.36707e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.48253,22860.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.400331,22856.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.66718,45729
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,321.833,2.80938e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.434877,25743
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.367383,25746.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.47794,51540.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,147.136,1.38368e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.41173,27110.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.343978,27107.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,175.018,1.76817e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.412656,28152.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.335921,28167.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,129.575,1.32628e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.41183,28804
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.334725,28788.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.76962,33243.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.32825,74271.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.94442,41180.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.278149,28848.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.75426,47543.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.02963,34074.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.276062,28937
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.4788,33152
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.453994,30107.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.0387,421197
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_157.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_157.txt
deleted file mode 100644
index 64684f2413..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_157.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,287.849,2.34643e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.49487,22794.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.395504,22790.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.06029,46818.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,318.686,2.78154e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.436592,25620.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.371281,25635
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.65472,51293.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,147.071,1.37849e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.433847,26983.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.339428,26995.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,174.831,1.76004e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.430601,28068
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.342129,28068.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,129.317,1.32483e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.419414,28668.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.338391,28691.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.90188,31434.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.3576,75277
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.83406,40071.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.283825,28717
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.75734,47488.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.01963,33994.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.274066,28845
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.44692,33041.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.424464,29997
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.4332,418065
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_158.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_158.txt
deleted file mode 100644
index 905a3f5e14..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_158.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,270.088,2.18654e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.490774,22505.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.399998,22517.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.98629,45042.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,314.854,2.71404e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.451389,25364.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.386513,25341.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.99894,53387.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,142.604,1.31369e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.433437,26636
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.342642,25282.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,165.705,1.63866e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.413342,27617
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.339486,27605.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,121.446,1.22417e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.413795,28278.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.339543,28294
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.78271,31197.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.3198,72896.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.91834,40430.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.293348,28305.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.69246,46658.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.04285,33435.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.284421,28385.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.48066,32561.8
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.443658,29560.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.0832,414872
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_159.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_159.txt
deleted file mode 100644
index 32c740c4ce..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_159.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,279.982,2.26971e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.481142,22551
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.373597,22543.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.8545,46030.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,313.554,2.70775e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.437494,25421.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.394635,25394
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.40226,50879.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,145.959,1.3586e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.415294,26721.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.33779,26740.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,174.225,1.73974e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.425034,27794.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.33946,27809.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,127.499,1.29883e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.422186,28431.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.340958,28431.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.84849,34241.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.30764,73589.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.94725,40734.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.285675,28477.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.72433,47070.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.02435,33728.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.278507,28597
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.45965,32815.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.423805,29740.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.1753,416022
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_160.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_160.txt
deleted file mode 100644
index 87094f1c2c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_160.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,278.582,2.26112e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.478787,22581.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.385738,22528.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.6595,45160
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,314.579,2.71203e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.437591,25478.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.384951,25501
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.46229,51024.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,145.181,1.35071e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.450039,26805.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.341073,26817.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,173.306,1.72907e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.42312,27889.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.345725,27897.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,127.616,1.29542e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.413476,28473.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.33898,28473.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.84259,32887.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.32521,73432.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.90936,40742.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.282967,28515.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.73178,47054.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.01908,33731
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.279268,28668
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.44412,32770.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.432491,29714.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.1026,416622
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_161.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_161.txt
deleted file mode 100644
index 3a56a4bf67..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_161.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,304.167,2.52881e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.493008,23243
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.369675,23215.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.0829,48876.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,356.731,3.18077e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.433879,26472.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.391038,26438.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.9814,54261
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,146.87,1.40829e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.410423,27754.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.33619,27758.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,174.111,1.79333e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.412714,28703.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.339773,28695.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,128.736,1.34597e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.406551,29302
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.34323,29298
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.83054,33775.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.32489,75808.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.95344,41951.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.27694,27830
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.66519,48503.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.02264,34706.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.28323,29449.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.47741,33744.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.427056,30620.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.3337,429606
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_162.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_162.txt
deleted file mode 100644
index da47c26710..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_162.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,304.004,2.6168e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.539228,23990.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.401866,23998.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.68873,47069.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,359.454,3.24672e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.442576,27015.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.400215,27000.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.44952,54042.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,207.715,2.04096e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.428535,28615.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.339985,28602.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,258.236,2.76714e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.431414,29939.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.345457,29932.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,184.405,2.0265e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.421232,30677.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.339416,30684.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.81525,36810.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,8.01571,86814.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.88103,44284.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.288837,30681
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,5.04182,53482.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,0.999456,36414.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.280305,30817.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.52861,36528.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.379134,30897.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.1339,443636
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_163.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_163.txt
deleted file mode 100644
index 5ceaa1add1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_163.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,304.407,2.62039e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.49064,23937
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.38202,23940.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.76337,47931.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,356.786,3.26015e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.434896,26972.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.376894,26999.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.44882,52672.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,207.584,2.07129e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.41841,28550.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.334371,28542
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,257.275,2.75946e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.417578,29920.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.333657,29916.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,183.961,1.97748e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.418078,30695.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.34346,30676.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.73845,36752.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.40277,82432
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.88325,44330.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.273969,30680.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.65854,51315.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.02292,36448.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.269804,30768
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.41758,35334.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.413661,31958.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.1711,444967
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_164.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_164.txt
deleted file mode 100644
index d5d4542dee..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_164.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,308.637,2.64942e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.474793,24067.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.377118,24075.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.90753,48177.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,378.192,3.47264e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.435856,27087.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.377188,27084
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.78108,56999
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,207.571,2.06591e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.428707,28731
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.339512,28734.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,258.155,2.77741e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.428867,29970.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.346398,29981.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,183.849,2.02864e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.429629,30730.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.345482,30737.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.71801,33888.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.37724,82581
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.89893,44468.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.282481,30734.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.70085,51423
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,0.946164,36590
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.281477,30848.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.4291,35522
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.424144,32133.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.4098,447703
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_165.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_165.txt
deleted file mode 100644
index adeba7a2fe..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_165.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,297.534,2.43996e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.481245,22898.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.397425,22875.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.61146,46951.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,355.533,3.1295e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.438102,26065.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.37743,26088.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.47939,52203
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,145.64,1.37532e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.419159,27365.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.338443,27381.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,173.563,1.76793e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.453616,28396.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.345111,28403.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,127.973,1.3209e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.429303,28982.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.353804,28985.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.80588,33447.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.96746,80240.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.94574,41473.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.28414,28989.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.94779,51072
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,0.976051,34345.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.277368,29146
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.51731,34456
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.363063,29203.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.1203,424598
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_166.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_166.txt
deleted file mode 100644
index 90a5a9bb27..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_166.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,297.484,2.52426e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.476068,23653.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.36666,23665.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.78486,48609.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,353.99,3.20521e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.449981,26710.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.364733,26714
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.53371,53478
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,206.6,2.04905e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.419152,28294.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.330603,28286.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,257.577,2.73938e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.419702,29649.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.336337,29603.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,184.479,2.01094e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.419377,30448.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.336491,30425
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.82156,38140.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.39604,81535
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.87415,42816
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.274495,30436.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.78339,50877
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.06535,36140.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.272036,30501.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.51078,35071.8
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.42547,31745.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.1993,440895
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_167.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_167.txt
deleted file mode 100644
index c3e83a0ac7..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_167.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,296.562,2.54573e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.483562,23826.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.389066,23837.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.87391,46458.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,352.899,3.1982e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.458154,26897
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.379659,26893.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.77367,53820.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,206.292,2.04068e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.430826,28503.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.346884,28484.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,256.01,2.73599e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.432797,29775.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.335576,29790.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,184.389,2.0146e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.428176,30532
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.335723,30501.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.66343,32065.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.37068,81886.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.88047,44132.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.293502,30505.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.62939,51094.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.06439,36277.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.276958,30669.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.42183,35266.2
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.508649,31890.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.754,450476
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_168.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_168.txt
deleted file mode 100644
index 87a8b5b753..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_168.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,298.737,2.56173e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.487766,23787.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.394545,23772.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.6904,47590.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,369.368,3.34402e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.443563,26817.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.376375,26837
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.54972,53705
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,205.062,2.02188e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.423229,28422.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.353867,28418.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,256.725,2.72743e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.440644,29779.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.349207,29725.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,182.983,1.99774e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.422934,30528.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.343504,30547.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.81525,36790.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.38699,81530.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.90984,44390.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.30305,30577.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.77261,51388
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.08207,36369
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.282924,30616.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.45893,35281.8
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.450915,31905.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.0269,442478
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_261.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_261.txt
deleted file mode 100644
index 93838593b0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_261.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,239.159,1.8778e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.500221,21907
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.425374,21899.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.93152,46327.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,319.478,2.69472e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.409328,24677.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.389622,24673.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.46425,49392.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,141.826,1.27388e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.383255,25669.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.337892,25673.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,189.207,1.81101e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.403869,26546.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.344977,26527.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,133.674,1.30657e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.385706,27047.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.341623,27047.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.67763,28436
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.2637,69883.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.90429,38818.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.281375,27105.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.6535,44900.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.06299,32098.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.280145,27204.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.46036,31217.8
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.462704,28331
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.1553,401965
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_262.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_262.txt
deleted file mode 100644
index 14eb5aebdc..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_262.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,240.369,1.8844e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.47601,21922.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.407965,21938.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.92321,44451.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,319.346,2.68915e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.40577,24650.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.367959,24658.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.37474,49381.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,141.115,1.27234e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.383088,25688.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.340407,25688.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,188.159,1.79488e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.399665,26542.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.347332,26550.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,133.456,1.30303e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.383908,27082.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.331986,27059.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.67921,27059.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.26349,69737.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.86217,37769.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.285144,27090.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.73779,44848
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.04095,32091.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.271524,27231.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.4478,31233.2
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.396132,28284.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.4397,403698
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_263.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_263.txt
deleted file mode 100644
index 4d7d8bd066..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_263.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,283.479,2.34161e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.461636,23092
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.373643,23084.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.93579,47439.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,379.707,3.37608e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.417068,26186.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.394808,26185.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.92988,53769.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,177.266,1.69806e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.403147,27381.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.343039,27404.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,256.526,2.61356e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.411801,28373.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.339218,28351
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,178.12,1.84321e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.404037,28901
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.347307,28877.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.69885,30353.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.35126,76658
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.9629,41851
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.277427,28908.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.6767,48488.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,0.96418,34421
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.277983,29073.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.45029,33452
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.423179,30259
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.5578,426495
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_264.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_264.txt
deleted file mode 100644
index ccb144aca9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_264.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,283.059,2.33068e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.461873,23114.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.381311,23118.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.78863,47515.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,379.461,3.38573e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.430194,26185.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.383231,26212.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.5791,53831.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,177.436,1.69892e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.402565,27384.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.339647,27381.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,257.399,2.62307e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.43002,28343.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.34787,28331.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,178.481,1.85985e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.408601,28912.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.33639,28893.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.79199,31830.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.87687,80616.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.9147,41782.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.28384,28931.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,5.02516,50517.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.0002,34359.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.282002,29023.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.55902,34428.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.365131,29122.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.2455,423237
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_265.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_265.txt
deleted file mode 100644
index 8af78c59c5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_265.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,283.607,2.34483e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.452094,23164.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.377304,23179.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.8533,46390
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,380.062,3.38763e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.430624,26235
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.378015,26253.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.61293,52555.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,177.731,1.70178e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.400748,27396.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.357381,27412
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,257.179,2.61661e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.412037,28351
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.345541,28351
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,178.761,1.86393e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.398309,28962.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.35326,28966.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.75246,33364.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.30033,77764.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.86243,40825.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.292518,28973.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.82387,48493.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.0437,34459.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.280441,29096.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.47374,33594
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.437816,30351.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.0961,422772
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_266.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_266.txt
deleted file mode 100644
index d0e6e9a17e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_266.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,309.155,2.57416e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.442635,23297.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.388606,23309
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.81827,47869.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,413.862,3.71366e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.427103,26663
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.389618,26678.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.86912,53372.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,171.531,1.64968e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.416357,27856.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.353023,27849
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,240.222,2.46913e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.418022,28696.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.355481,28696.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,167.655,1.75969e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.413458,29210.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.352472,29165.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.78953,35131.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.97524,82807.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.90814,42220.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.304992,29205.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.97367,51161.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.00421,34737.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.286917,29374.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.54951,34882
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.376614,29461.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.423,424489
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_267.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_267.txt
deleted file mode 100644
index 78ebfbbbb1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_267.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,309.406,2.5672e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.434609,23278.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.365573,23290.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.04474,46626
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,414.939,3.71638e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.430028,26594.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.382379,26606.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.81203,53236.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,171.216,1.65335e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.415865,27803.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.343807,27814.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,240.439,2.47009e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.418278,28666.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.337977,28659
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,168.62,1.74734e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.412798,29188
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.347718,29165
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.77458,30668
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.43128,78855.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.89549,42407.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.288448,29208.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.80043,49215.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.07115,34829
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.290028,29376
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.47451,33809.2
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.4427,30493.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.789,429834
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_268.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_268.txt
deleted file mode 100644
index 0c2c340b50..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_268.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,309.613,2.57011e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.447346,23316.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.376549,23313
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.80858,46652.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,414.747,3.7171e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.43125,26606
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.385976,26629.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.46542,53344
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,171.413,1.66018e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.427141,27837.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.34727,27810.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,240.804,2.47861e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.421176,28731
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.348403,28734.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,167.967,1.75774e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.411128,29194.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.349682,29206.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.73172,32208
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.55088,79804.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.95532,42358.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.29102,29216.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.78787,50123.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,0.972817,34818.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.293918,29348.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.47703,33829
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.459557,30543.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.5514,428600
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_269.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_269.txt
deleted file mode 100644
index 16cdb9e385..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_269.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,312.658,2.60353e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.447026,23362
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.37452,23369.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.84804,46754.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,414.875,3.72555e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.429964,26648.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.39534,26644.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,4.53306,53328.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,171.436,1.64858e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.414898,27780.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.348965,27795.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,241.172,2.47577e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.426834,28749.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.359078,28726.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,167.408,1.72382e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.410566,29215
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.343347,29180.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,2.7337,33650
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,7.39321,78142.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,1.86438,41160.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.283916,29246.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,4.7498,48955.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.05844,34782
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.278342,29430.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.46566,33851.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.461394,30573.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,37.4792,430827
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp32_perf_fp32_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp32_perf_fp32_120.txt
deleted file mode 100644
index 649a1f51be..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp32_perf_fp32_120.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-Conv1,650.457,5.73552e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.474304,25907.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.332364,25911.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,8.92677,90320.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv2,1057.96,9.77033e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.453234,28497
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.372524,28497.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool2,8.27002,90575.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv3,286.088,2.90251e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.456562,30895.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.377951,30887.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,414.867,4.62072e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.430777,32026.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.258464,32044.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,274.551,3.07434e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.429369,32951.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.375225,32927.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Pool3,4.34963,65850.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Mul1,16.8142,166920
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add6,3.2153,60129.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.388544,32831.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Mul2,5.96569,71879
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add7,1.62928,45759.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.276691,32854.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Mul3,1.64413,41948.8
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add8,0.586707,35509.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Softmax1,76.6594,908040
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_120.txt
deleted file mode 100644
index ab5123cc67..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_120.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,53.2416,410603
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.200686,10812.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.35149,30661.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.16412,10836
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,141.815,1.1076e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.212245,11246.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.24734,22510.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.154683,11261.9
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.798899,12588.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.399194,12067
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.12763,11294.3
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.19088,11302
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.159733,11343.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.132791,11366.9
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.12994,12188.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_151.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_151.txt
deleted file mode 100644
index f615120716..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_151.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,48.1242,345225
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.194072,9955.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.02115,23780.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.170353,10001.6
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,76.2011,563364
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.208235,10230.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.09816,20467.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.151428,10246.1
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.724709,11214.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.400649,10995.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.127191,10307.3
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.15107,10324.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.120308,10366.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.134987,10383.8
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.00837,10992.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_152.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_152.txt
deleted file mode 100644
index a3c51be3cf..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_152.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,48.9399,350989
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.197227,9884.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.11865,27001.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.181156,9936.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,77.7687,570069
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.22133,10144.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.04445,20337.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.161111,10183.3
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.71334,11153.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.396691,10940.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.128478,10231.1
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.25817,10834
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.159803,10269.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.134676,10292.2
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.0811,10930.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_153.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_153.txt
deleted file mode 100644
index 2dd6e9a88a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_153.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,46.6348,325536
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.195595,9778.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.12681,24863
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.176449,9806.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,77.1562,559152
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.210802,10053.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.14929,20128.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.158446,10082.3
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.692079,11008
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.388636,10829.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.129313,10124.2
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.18317,10147.1
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.158939,10166
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.133499,10167.9
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.05805,11298.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_154.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_154.txt
deleted file mode 100644
index 389e60070d..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_154.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,48.0157,339125
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.194162,9707.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.24845,26882.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.16954,9740.2
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,78.0062,562935
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.207128,9963.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.24529,19938.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.153448,9986.5
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.719999,10933.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.409574,10722.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.132369,10066.8
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.23904,10084
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.157931,10093.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.140488,10118.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.10729,11706.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_155.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_155.txt
deleted file mode 100644
index ca03841e4a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_155.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,51.1264,364502
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.204514,9953.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.1248,26736.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.181972,9987.8
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,89.6018,660985
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.219131,10209.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.19727,20441
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.159291,10238.7
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.715628,11189.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.407411,11204.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.130753,10273.1
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.20173,10298
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.160679,10311.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.140894,10336.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.10275,11773.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_156.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_156.txt
deleted file mode 100644
index 1f4c7f6825..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_156.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,52.9157,380778
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.203973,10001.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.24634,29547.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.164356,10036.1
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,89.8067,667021
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.224731,10238.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.22599,20477.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.159969,10246.5
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.719423,11220.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.415769,11277.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.143569,10311.4
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.21795,10340.1
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.17548,10341.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.137815,10366.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.09577,12943.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_157.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_157.txt
deleted file mode 100644
index 4abd2ff1d6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_157.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,52.7686,392792
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.19764,10428.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.05472,27189.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.169716,10457.4
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,107.438,830153
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.214123,10726.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.00577,21072.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.155681,10755.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.798201,12047.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.402019,11556.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.123946,10812.3
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.19996,10820
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.164612,10846.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.127936,10856.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.0511,12358.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_158.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_158.txt
deleted file mode 100644
index ae604d7bdd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_158.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,49.0993,342331
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.203544,9711
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.05407,24425.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.179067,9764.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,90.8712,654519
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.221912,9948.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.2273,19913.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.159806,9969.1
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.716977,10902.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.400351,10695.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.134183,10038
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.23569,10095.5
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.168154,10116.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.143252,10126.1
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.07097,11713.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_159.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_159.txt
deleted file mode 100644
index 955744472e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_159.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,50.8444,350982
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.209797,9758.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.27614,28217.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.164961,9789.6
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,90.4073,655332
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.223419,10003.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.27863,20033.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.158462,10049.5
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.702981,11026.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.410073,10794.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.136487,10114.5
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.21371,10122.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.133022,10131.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.138929,10133.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.22311,12791.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_160.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_160.txt
deleted file mode 100644
index 5f142c7920..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_160.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,50.7658,368438
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.195457,10169.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.35397,30084.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.185678,10215.8
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,106.684,801549
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.213777,10460.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.24265,20941.7
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.155121,10488.9
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.849846,11984.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.372124,11263.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.132897,10551.8
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.22824,10855.9
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.133111,10606.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.133147,10626.9
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.15523,11644.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_161.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_161.txt
deleted file mode 100644
index 092b6f5d79..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_161.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,54.4765,397281
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.198766,10384.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.2986,30571.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.157898,10413.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,105.108,811439
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.206196,10702.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.20963,21433.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.153771,10728.9
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.821688,12009
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.417266,11780.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.128219,10780.3
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.21484,10801.1
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.168548,10802.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.128519,10837.3
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.06899,11700.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_162.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_162.txt
deleted file mode 100644
index 3ca4277de1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_162.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,55.2578,412629
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.201061,10337.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.26118,28870.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.166932,10368.3
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,107.154,823003
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.211173,10624.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.21668,21291.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.156635,10666.6
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.828248,11977.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.415174,11726
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.131767,10721.4
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.2329,10746.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.169524,10763.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.134702,10786.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.0436,12095.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_163.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_163.txt
deleted file mode 100644
index dbfbf37ebc..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_163.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,55.6706,411214
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.199019,10259.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.40361,29808.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.162872,10296
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,117.265,891235
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.209499,10559.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.27431,21129.1
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.155639,10569.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.867055,12131.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.355961,11384.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.129613,10613.5
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.25566,10645.9
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.139422,10668.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.135556,10685.9
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.03205,12631
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_164.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_164.txt
deleted file mode 100644
index 99b2446d3c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_164.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,54.555,404489
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.195633,10248
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.32372,29348.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.158491,10299.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,116.045,885287
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.212567,10548.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.25185,20535.7
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.154827,10586.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.813368,11876.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.421801,11632.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.132619,10628.5
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.22149,10645.7
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.168667,10685.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.140765,10703
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.10036,12345.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_165.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_165.txt
deleted file mode 100644
index e1c00548bb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_165.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,51.8461,378176
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.205409,10146.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.29647,27806
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.175278,10179.2
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,107.034,806291
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.220113,10469.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.22665,20951
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.156977,10479.3
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.81621,11727.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.401843,11518.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.128913,10538.5
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.19511,10555.5
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.174436,10587.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.131604,10599.2
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,0.973702,11235.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_166.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_166.txt
deleted file mode 100644
index d0c78df48e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_166.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,51.845,377220
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.202846,10185.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.31109,28424.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.167067,10213.5
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,106.794,808042
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.210497,10525.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.26955,20816.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.157351,10544.1
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.762453,11546.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.423516,11571.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.135339,10587.9
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.23369,11207.7
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.14531,10629.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.139399,10631.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.18922,11353.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_167.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_167.txt
deleted file mode 100644
index 4582f6b782..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_167.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,51.2702,371429
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.199473,10120.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.31699,29873.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.158638,10168.4
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,113.259,848992
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.2082,10447.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.21405,20910.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.154529,10462.8
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.818962,11718.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.414665,11511.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.132638,10531.9
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.17942,10533.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.168603,10544.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.13177,10573
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.13253,11970.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_168.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_168.txt
deleted file mode 100644
index 55d8d4862b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_168.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,51.2411,369185
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.195915,10109.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.21625,29217.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.164785,10137.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,113.493,848930
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.211281,10412.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.27259,20860
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.160157,10443.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.865858,11955.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.401084,11221.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.132119,9920.8
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.20541,10189.3
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.125783,9991.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.134711,10000.9
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.26711,10830.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_261.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_261.txt
deleted file mode 100644
index ceeccbe71a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_261.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,24.6817,175003
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.17307,9823.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.04969,26588.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.122599,9861.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,92.441,682924
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.198039,10233.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.14033,20486.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.161636,10260.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.715854,11188.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.393919,10990.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.127492,10331.1
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.26433,10350.3
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.163857,10375
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.137178,10390.3
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,0.975836,10632
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_262.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_262.txt
deleted file mode 100644
index ae7893234b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_262.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,24.6179,176890
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.166737,9825.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,2.9676,24858.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.11938,9854
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,91.1076,673560
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.191941,10287.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.09591,20588.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.164788,10321.9
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.67771,11241.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.387586,11046.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.13265,10342.8
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.2241,10365.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.165851,10396.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.145143,10421.5
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.00152,10659.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_263.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_263.txt
deleted file mode 100644
index bd428223fa..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_263.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,26.2544,191225
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.170372,10033.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.13531,28341.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.156977,10071.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,110.735,835247
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.196139,10494.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.22919,20410.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.15539,10517.1
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.846818,11995.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.367052,11283.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.127601,10572.1
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.26546,10844.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.124983,10625.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.136359,10627.5
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.03664,11336.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_264.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_264.txt
deleted file mode 100644
index f680987458..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_264.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,27.192,198660
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.170811,10189.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.11118,28018.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.161054,10229.6
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,109.608,835120
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.194372,10661.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.27415,21780.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.159457,10692
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.815076,12175.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.38781,11445.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.128907,10711.1
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.19239,10745.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.173306,10782.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.134577,10787.7
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.073,11933.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_265.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_265.txt
deleted file mode 100644
index 525e1c1bce..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_265.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,27.1287,199754
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.177319,10183.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.09268,27524.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.15875,10220.3
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,109.718,837972
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.195335,10667.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.10515,21350
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.15731,10684.5
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.763332,11923
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.396937,11441.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.12577,10722.5
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.21391,10741.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.169403,10758.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.131419,10768.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,0.991177,11341.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_266.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_266.txt
deleted file mode 100644
index a7e4b208e3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_266.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,28.1301,210769
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.175569,10256.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.25761,30029
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.153194,10300.6
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,119.738,911657
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.192788,10749.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.21573,21510.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.145374,10768.5
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.816036,11993.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.415413,11776.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.128081,10810.5
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.26205,10829.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.161025,10856.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.13923,10894.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.00547,11532.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_267.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_267.txt
deleted file mode 100644
index 3948a14350..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_267.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,28.2398,208264
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.175534,10338.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.12943,27848.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.163847,10395.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,118.691,913277
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.197476,10825.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.11307,21657.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.160868,10845
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.809361,12336.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.360213,11630.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.133012,10923.2
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.22352,10932.7
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.143597,10957.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.133201,10972.6
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.0499,13248.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_268.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_268.txt
deleted file mode 100644
index 2e3cfc4675..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_268.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,28.1788,210858
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.176807,10342.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.16951,27912.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.165633,10369.2
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,118.568,907059
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.195265,10820.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.1375,21644
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.155005,10825.8
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.78339,12064.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.405327,11619
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.131358,10881.1
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.21585,10925
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.144199,10955.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.13563,10963.1
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.16035,11521.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_269.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_269.txt
deleted file mode 100644
index 1910111015..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_269.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,28.5102,202387
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.180833,10300.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,3.22101,28427.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.169076,10328.9
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,118.634,917354
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.198001,10782
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.16212,21577.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.157057,10795.4
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,0.787105,12035.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.407577,11583
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.133399,10844.9
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,1.22979,10879.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.165399,10896.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.142606,10913.4
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,1.13882,11704.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp32_perf_fp32_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp32_perf_fp32_120.txt
deleted file mode 100644
index df787165b7..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp32_perf_fp32_120.txt
+++ /dev/null
@@ -1,45 +0,0 @@
-Conv1,67.8081,567302
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.213878,12094.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Pool1,4.51207,41956.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Tanh1,0.210908,12134.7
-Tanh1_f2h,0,0
-Tanh1_h2f,0,0
-Conv2,203.117,1.77821e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.207398,12651.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Pool2,2.85643,30378
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Tanh2,0.178364,12683.6
-Tanh2_f2h,0,0
-Tanh2_h2f,0,0
-Mul1,1.98613,17351.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add3,0.617506,14584.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Tanh3,0.141305,12708.4
-Tanh3_f2h,0,0
-Tanh3_h2f,0,0
-Mul2,0.396533,12723.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add4,0.136624,12725.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Tanh4,0.131891,12770.7
-Tanh4_f2h,0,0
-Tanh4_h2f,0,0
-Softmax1,4.99293,52896.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_120.txt
deleted file mode 100644
index 5d79cb1b4b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_120.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,73.9502,645096
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.247835,12623.1
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.164623,12625
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,294.637,2.65058e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.258104,12842.2
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.169119,12853.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,132.495,1.19912e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.253916,12988.5
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.161241,12980.9
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,151.749,1.43286e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.255339,13213.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.176837,13213.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,70.4482,661533
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.239563,13236.8
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.154121,13223.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,324.851,3.12101e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.253886,13583.1
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.160607,13577.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,96.3067,924399
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.243679,13037.4
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.155141,13024.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,87.4979,863312
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.243333,13854.2
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.152028,13848.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,37.7534,377593
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.181286,13825.4
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.146502,13817.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,191.548,1.93282e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.250984,13866.8
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.154498,13859.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,52.3825,519540
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.19949,13929.8
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.150866,13924.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,51.9763,533162
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.187788,13946.9
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.145609,13939.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,22.4328,231407
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.162892,13906.8
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.14126,13907.1
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,132.226,1.36954e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.236005,14206.9
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.1487,14183.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,32.2877,327249
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.179397,14208.8
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.143974,14193.5
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,132.291,1.39504e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.234907,14439.9
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.145369,14432.2
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,32.1337,336438
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.185467,14453.3
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.143277,14453.3
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,132.577,1.42292e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.239714,14618.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.147318,14611.2
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,32.3695,341956
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.184789,14667
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.146847,14651.6
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,132.132,1.43444e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.236386,14816.4
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.146639,14801.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,32.1764,345701
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.183279,14812.9
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.151174,14805.3
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,131.917,1.44904e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.236888,14948.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.150729,14933
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,32.167,345983
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.182521,14187.1
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.144009,14940.5
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,65.8819,728654
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.162079,14775.8
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.135289,14766.2
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.9825,193420
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.406036,16917.4
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.103993,14697.4
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,220.662,2.32068e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.22685,14113.2
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.14111,14113.2
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,32.4289,340589
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.186716,14063.5
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.124931,14048.1
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.262235,14036.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.00585,21818.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0685114,14032.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.668,147250
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_151.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_151.txt
deleted file mode 100644
index dc27bd9b0c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_151.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,65.389,551425
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.242395,11938.4
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.162293,11947.9
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,150.895,1.29446e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.2506,12167.6
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.155247,12171.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,123.348,1.09867e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.236335,12533.1
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.151896,12540.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,78.0618,719279
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.237803,12829.5
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.154418,12818.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,63.5923,594937
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.219042,12945.5
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.14911,12938
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,164.157,1.55489e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.249452,13138.3
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.156713,13134.5
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,76.2518,725387
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.236447,13396.3
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.158629,13400.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,44.9147,443219
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.217602,13519.2
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.149452,13505.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,34.1333,341062
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.191435,13526.7
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.142422,13526.5
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,99.2134,991306
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.242485,13607.4
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.147484,13591.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,41.2043,408958
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.216159,13685.6
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.143964,13687.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,32.211,331089
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.196456,13729.5
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.134588,13722
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,18.5924,193537
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.181948,13723.9
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.106428,13716.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,109.057,1.0975e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.220994,13614.9
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.146076,13607.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,26.7251,268025
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.185087,13633.5
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.131715,13624.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,109.704,1.10027e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.224808,13509
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.140105,13499.5
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,26.6311,267766
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.186693,13562.2
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.122294,13562.2
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,109.839,1.09124e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.216386,13469.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.141992,13462.1
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,26.6122,264636
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.190396,13481.1
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.119471,13479.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,108.943,1.07929e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.216264,13396.9
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.138821,13395
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,26.6761,266045
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.165205,13427.3
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.125657,13419.7
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,109.453,1.0827e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.221314,13372.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.142541,13349.4
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,26.696,263889
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.159327,13373.8
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.129939,13373.8
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,43.2144,426715
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.139647,13281.1
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.128982,13273.5
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,13.3333,132384
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.128908,13269.6
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0828864,13263.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,141.034,1.34819e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.207995,13055.3
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.135052,13055.3
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,26.1664,254805
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.14173,13030.6
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.094348,13026.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.245727,13024.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.99098,19884.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0685148,13002
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.2567,137044
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_152.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_152.txt
deleted file mode 100644
index 8964f1786a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_152.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,65.3376,550172
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.236098,11944.7
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.159887,11946.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,150.596,1.2911e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.259371,12169.5
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.161516,12169.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,123.092,1.08442e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.240299,12538.9
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.152095,12533.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,78.451,723292
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.234958,12825.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.15686,12829.5
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,63.581,596887
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.215909,12951
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.152786,12949.1
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,164.393,1.56285e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.251448,13130.5
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.163087,13132.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,76.3737,720128
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.233675,13405.9
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.157801,13405.9
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,45.5004,448326
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.218526,13506.3
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.146137,13508.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,34.1937,339642
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.194184,13506.1
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.14062,13506.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,99.0062,986539
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.23893,13607.7
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.15061,13615.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,41.238,409989
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.211727,13689.8
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.15165,13697.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,33.4938,341257
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.200818,13716.2
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.136246,13716.5
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,18.7081,194156
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.183381,13707.1
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.106198,13694
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,109.543,1.10039e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.218457,13575.3
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.142188,13575.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,26.555,267700
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.184959,13628.5
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.124883,13621.1
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,109.565,1.09646e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.212393,13512.9
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.140883,13505.3
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,26.5678,268497
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.185747,13548.9
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.122438,13533.7
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,109.932,1.09239e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.232188,13454
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.143193,13446.3
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,26.6109,265493
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.155186,13482.7
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.126921,13482.7
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,108.969,1.08043e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.217931,13385
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.141852,13377.3
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,26.5757,264539
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.155113,13423.2
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.122972,13421.2
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,109.488,1.08123e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.217707,13357
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.145151,13347.5
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,26.63,262356
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.153769,13360.3
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.127503,13360.3
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,42.3842,426475
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.141138,13286.8
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.127612,13286.8
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,13.1488,133664
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.123285,13246.7
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0761851,13246.7
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,139.682,1.35442e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.199247,13059.1
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.139247,13059.1
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,25.6908,251420
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.145958,13032.5
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.0989915,13015.4
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.234261,13015.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.983277,19896.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0692538,12988.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.3946,136283
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_153.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_153.txt
deleted file mode 100644
index 39c3c4bfc3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_153.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,59.192,492599
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.230488,11765.4
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.162552,11765.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,146.043,1.23587e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.248728,11990.4
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.160364,11984.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,113.012,984004
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.23555,12316.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.171029,12318.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,77.4188,700262
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.239547,12557.5
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.158214,12553.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,61.8709,566799
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.215836,12673.7
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.152092,12675.7
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,162.923,1.51363e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.245311,12839.9
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.158648,12839.9
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,74.5823,688044
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.229919,12388.6
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.152201,12396.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,45.2625,433952
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.217247,13213.5
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.151206,13213.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,34.2165,334105
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.191567,13197.6
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.143942,13197.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,99.0185,964889
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.239947,13256.6
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.149024,13251
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,41.4294,388095
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.213621,13344.9
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.148351,13346.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,32.7904,326992
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.198322,13398.5
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.141132,13390.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,19.1902,193601
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.180341,13379.4
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.114479,13369.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,109.805,1.07841e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.21853,13285.4
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.143682,13277.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,26.9344,265827
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.192278,13317.8
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.125276,13310.1
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,110.593,1.08168e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.213032,13214.9
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.142434,13207.1
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,26.9675,267304
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.187317,13279.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.118931,13270
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,110.601,1.07806e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.216351,13167.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.147916,13167.8
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,26.9098,265035
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.154441,13202.1
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.121635,13190.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,109.682,1.06469e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.212578,13116
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.140905,13108.3
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,26.9461,262889
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.15766,13169.7
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.121433,13167.7
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,110.175,1.06465e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.213787,13087.3
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.142697,13087.3
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,26.9743,263452
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.160233,13110.2
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.12678,13108.3
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,43.3574,425833
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.136303,13040
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.128598,13038.1
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,13.3003,131934
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.12925,13047.5
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0788378,13032.3
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,140.09,1.3326e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.209426,12877.9
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.13477,12877.9
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,26.0744,250139
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.137375,12849.4
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.0982747,12841.7
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.237528,12830.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.966852,19076.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.073513,12807.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.4019,134757
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_154.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_154.txt
deleted file mode 100644
index af20d986d0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_154.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,59.7437,495554
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.234494,11713.7
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.158329,11715.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,147.27,1.22824e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.251391,11941.4
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.158031,11943.3
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,113.406,985062
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.232309,11610
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.150818,12241.5
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,77.1911,695760
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.239986,12551.8
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.159061,12549.9
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,61.5726,562432
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.217922,12654.6
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.154696,12656.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,162.169,1.49227e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.253391,12843.6
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.159193,12845.5
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,74.2893,687068
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.234261,13046.6
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.150271,13046.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,45.0143,431818
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.226866,13199.5
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.149919,13199.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,34.0262,328298
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.194085,13204.8
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.144054,13199.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,98.5223,958944
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.235368,13271.9
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.149695,13266.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,40.8996,394590
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.210469,13381.2
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.151861,13383.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,33.4273,331360
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.201967,13387
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.137839,13387
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,18.9907,190974
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.179532,13398.7
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.112198,13398.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,109.752,1.07808e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.216092,13300.5
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.143657,13292.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,26.9502,268056
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.18933,13323.4
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.118691,13321.5
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,110.201,1.07843e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.217835,13226.4
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.139231,13218.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,26.902,265461
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.194255,13264
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.120886,13262.1
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,110.725,1.0789e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.217611,13194.6
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.138386,13194.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,26.928,265094
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.188508,13192.2
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.117836,13190.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,109.97,1.06774e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.220907,13127.6
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.141388,13127.6
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,26.9667,264174
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.188917,13177.2
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.118879,13175.3
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,110.078,1.06532e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.214929,13077.8
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.141698,13077.8
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,27.0638,264730
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.18982,13115.9
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.118863,13115.9
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,42.8797,419154
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.178261,13039.9
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.131683,13039.9
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,13.0624,130440
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.153228,13019
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0731229,13019
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,140.751,1.33412e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.204744,12845.3
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.138002,12828
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,25.6298,245932
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.171794,12824.5
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.09951,12807.3
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.232002,12799.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.00246,19515.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.067734,12790
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.3304,135313
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_155.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_155.txt
deleted file mode 100644
index 3cd198c77e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_155.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,71.1138,599921
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.237858,11925.6
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.156825,11929.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,187.397,1.62341e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.252718,12186.7
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.159817,12173.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,135.388,1.17338e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.242434,12581.7
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.154873,12576.1
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,98.0578,895295
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.241234,12892.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.158828,12869.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,70.626,663698
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.227765,13028
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.147266,13029.9
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,218.255,2.09237e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.250332,13410
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.156972,13411.9
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,88.0488,844329
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.233976,13637.7
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.154652,13630.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,55.1219,553498
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.229189,13733.6
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.147564,13735.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,36.8881,373914
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.205544,13750.4
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.1411,13758.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,128.082,1.29008e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.245151,13759.7
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.156329,13761.7
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,46.1832,445284
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.222834,13870.6
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.148393,13872.3
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,32.5254,333762
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.196988,13884.1
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.137001,13874.5
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,18.3932,190258
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.173331,13878
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.117833,13862.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,106.178,1.0807e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.217778,13743.6
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.145286,13728.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,26.2919,265060
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.18885,13060.4
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.124175,13741.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,106.833,1.07931e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.21707,13623.6
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.143596,13623.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,26.2662,265123
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.180648,13648
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.124159,13638.3
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,106.912,1.07212e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.216444,13541.2
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.144072,13539.3
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,26.4088,264220
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.194187,13542.9
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.1283,13535.3
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,106.146,1.05831e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.211371,13462.3
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.146716,13454.7
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,26.315,261815
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.184258,13495
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.127935,13470.1
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,106.704,1.05782e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.221704,13406.8
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.142434,13391.6
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,26.3237,260152
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.188457,13414.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.131734,13406.8
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,43.3881,435180
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.170901,13338.3
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.127836,13315.4
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,13.6722,134732
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.158857,13315.4
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0751612,13288.6
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,141.205,1.3677e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.198412,13051.5
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.134025,13049.6
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,26.3303,255393
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.167004,13024.9
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.0932922,13013.5
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.238219,13011.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.00843,20004.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0678684,12990.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.4966,136714
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_156.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_156.txt
deleted file mode 100644
index 8019eb1b69..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_156.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,71.2761,608731
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.244776,12084.3
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.158962,12088
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,188.912,1.64896e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.255058,12293.5
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.162853,12295.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,136.394,1.19324e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.238479,12658.1
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.158102,12652.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,104.663,967511
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.243407,12949.9
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.159599,12955.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,72.7473,673075
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.230959,13060.8
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.156332,13053.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,231.279,2.19977e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.253381,13464.3
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.17527,13470
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,91.4477,882583
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.232638,13666
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.154396,13669.9
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,55.869,562471
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.223273,13816.7
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.15093,13790
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,37.16,376760
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.20005,13791.8
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.143135,13791.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,126.416,1.26317e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.242568,13828.7
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.15438,13828.7
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,46.471,461018
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.221775,13910.5
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.148053,13902.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,46.0725,473517
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.203042,13897.1
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.143273,13889.5
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,22.7659,234615
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.185301,13891.7
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.133001,13884
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,150.933,1.52903e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.226543,13816.2
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.14726,13816.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,31.4032,317298
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.201442,13834.8
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.14229,13835.1
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,151.7,1.5283e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.236744,13765.1
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.145177,13757.5
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,31.5003,318316
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.204392,13814.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.147897,13784.2
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,151.53,1.52452e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.230943,13756.2
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.144997,13748.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,31.5601,316690
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.199858,13787.9
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.140754,13772.7
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,151.3,1.52384e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.235752,13709.1
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.143299,13701.7
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,31.5201,318085
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.201775,13786.5
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.138499,13771.3
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,151.321,1.52065e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.230063,13711.6
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.145682,13711.6
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,31.5567,316267
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.204805,13747.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.14038,13732.4
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,42.7856,438524
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.176604,13635.4
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.131766,13620
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,13.6054,139742
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.150604,13615
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0857691,13588.2
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,140.048,1.38348e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.201449,13321.5
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.13701,13313.8
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,26.1387,260393
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.174642,13276.5
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.0936986,13264.9
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.242418,13255.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.997105,20473.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0684222,13234.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.7298,139782
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_157.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_157.txt
deleted file mode 100644
index d8489f200d..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_157.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,72.46,609354
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.240213,12032.8
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.162009,12032.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,193.444,1.66746e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.254498,12239.9
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.157048,12241.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,137.521,1.22085e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.248389,12658.1
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.164284,12644.9
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,103.363,961789
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.245477,12934.7
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.160453,12928.9
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,72.2451,654769
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.234348,13047.4
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.165647,13047.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,229.582,2.2071e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.251861,13502.8
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.160018,13504.7
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,91.5788,884955
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.237893,13725
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.15511,13728.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,62.2835,627100
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.230226,13832.7
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.155564,13830.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,39.4552,397671
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.202898,13832.7
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.146067,13832.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,137.42,1.38913e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.242076,13900.6
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.157247,13900.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,50.6253,504818
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.223986,13993.6
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.149266,13978.3
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.7706,467971
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.214168,13992.5
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.146117,13984.9
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,21.7253,226389
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.193132,13951.8
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.13926,13936.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,150.085,1.52527e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.233119,13863.6
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.150428,13852.1
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,30.5291,308692
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.204037,13907.8
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.142162,13892.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,150.359,1.52327e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.231486,13824
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.146287,13808.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,30.7597,309798
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.206306,13867.2
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.148242,13859.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,150.606,1.52351e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.230056,13759.7
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.146265,13759.7
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,30.5411,306875
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.200361,13818.4
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.140364,13818.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,150.637,1.51677e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.230936,13747.9
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.144723,13732.6
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,30.7032,304003
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.205586,13784.3
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.144537,13784.7
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,150.371,1.51111e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.225404,13721.5
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.147328,13721.5
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,30.6697,303778
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.202386,13765
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.141798,13750
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,66.0462,671959
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.198402,13616.1
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.137314,13614.2
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.6425,169765
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.175077,13597.3
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.113548,13589.6
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,221.249,2.15435e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.21949,13160.3
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.142856,13152.7
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,31.0835,301232
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.190239,13165.6
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.127974,13150.4
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.280715,13140.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.999888,20340.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.075977,13120
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.6519,138176
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_158.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_158.txt
deleted file mode 100644
index 90ffe2f12a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_158.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,67.3077,556049
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.238328,11681
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.163951,11682.9
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,184.143,1.55929e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.247538,11908.5
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.162972,11910.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,126.709,1.09389e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.234011,12236.4
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.157218,12230.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,96.8968,872072
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.247083,12522.7
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.157017,12536
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,68.2742,621439
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.22669,12633.8
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.152831,12639.7
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,216.566,2.01758e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.245538,13050.4
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.158569,13052.3
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,85.8555,805974
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.232828,13260.7
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.159292,13268.3
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,55.0325,538310
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.226913,13382.1
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.149215,13374.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,36.76,363572
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.196174,13383.6
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.142963,13383.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,128.515,1.26005e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.239979,13383.1
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.148604,13383.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,45.7883,443150
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.216943,13459.6
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.145593,13471.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,33.0725,330537
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.197926,13499.7
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.137593,13492
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,18.5424,187872
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.175276,13494.1
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.114034,13482.5
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,108.862,1.07927e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.212722,13372.3
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.139647,13362.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,26.5576,259574
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.187423,13395
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.124197,13387.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,109.336,1.07399e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.214965,13263.7
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.141286,13263.7
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,26.4503,261680
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.186751,13309.5
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.127776,13299.9
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,109.411,1.07093e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.214437,13206.4
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.140688,13206.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,26.5886,259645
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.188684,13219.9
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.128895,13219.9
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,108.548,1.0561e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.212373,13123.7
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.14294,13114.1
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,26.4395,256976
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.190444,13164.2
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.12574,13154.6
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,109.135,1.05882e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.211669,13070.5
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.142716,13070.5
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,26.4033,255198
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.1842,13097.1
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.125782,13087.5
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,43.8395,430786
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.174146,13004
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.157049,12994.5
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,13.6335,134842
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.16207,12988.8
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0832412,12983.1
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,139.889,1.32864e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.22301,12835.3
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.138044,12804.7
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,26.3868,253192
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.169272,12810.7
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.0911609,12795.3
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.236268,12778.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.00093,19650.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.064918,12753.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.5151,134194
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_159.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_159.txt
deleted file mode 100644
index f516e0ae4c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_159.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,65.3362,550872
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.23892,11881.8
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.163554,11881.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,183.873,1.57523e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.249694,12067.6
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.15319,12063.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,126.067,1.09797e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.238168,12364.1
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.15277,12358.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,102.987,934710
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.251384,12637.4
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.156582,12641.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,70.0595,632959
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.228197,12755.7
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.162607,12742.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,228.783,2.12873e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.248139,13146.3
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.1567,13140.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,88.9911,839060
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.236088,13360.8
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.15157,13353.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,55.374,543389
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.228302,13461.7
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.146223,13463.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,36.8056,364843
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.205036,13448.7
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.143772,13448.9
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,128.391,1.26181e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.241208,13441.1
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.148406,13441.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,46.0282,448468
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.213861,13495.4
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.156479,13495.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.5721,458472
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.204981,13505.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.140918,13506
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,22.7383,231087
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.1858,13490.9
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.133187,13490.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,149.135,1.47567e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.229406,13452.5
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.139452,13441.1
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,31.8132,312152
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.196991,13491.1
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.136844,13491.1
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,149.39,1.47069e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.233442,13433.4
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.142633,13433.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,31.8094,312081
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.201141,13486.9
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.136691,13464
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,149.274,1.46842e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.226869,13431.5
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.14517,13416.3
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,31.806,311953
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.202188,13467.8
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.136435,13465.9
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,149.051,1.46773e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.225705,13406.8
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.144131,13399.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,31.8468,312188
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.206552,13448.8
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.142217,13448.8
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,149.405,1.46552e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.228936,13410.7
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.144601,13408.8
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,31.9564,313028
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.200817,13441.2
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.142169,13433.6
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,43.1695,434683
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.139756,13314.3
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.132351,13308.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,13.6361,140148
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.146761,13308.8
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0831129,13293.5
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,141.342,1.36962e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.203528,13087.8
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.135596,13072.6
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,26.1524,257254
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.163311,13041.5
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.0935357,13035.5
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.234783,13033.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.00197,20208.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0691773,13017.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.6763,136774
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_160.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_160.txt
deleted file mode 100644
index c5475d1abb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_160.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,67.2762,561170
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.244971,11778
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.159462,11776
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,186.843,1.58778e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.25467,12003.4
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.155311,11995.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,126.062,1.07634e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.236984,12320.4
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.155788,12322.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,100.552,898784
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.245582,12602.9
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.155635,12604.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,67.8345,619418
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.223733,12736.7
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.152607,12723.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,225.203,2.10712e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.254111,13166.3
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.156988,13166.3
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,86.4242,812090
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.234776,13374.4
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.152287,13366.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,61.3563,604222
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.230159,13505.7
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.15541,12809.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,39.2153,379525
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.205781,13465.7
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.14428,13459.9
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,136.112,1.34823e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.248104,13534.9
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.157125,13534.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,50.3802,490971
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.222866,13590.5
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.15254,13583.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.5792,442030
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.207859,13594.9
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.144908,13572
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,21.8041,220761
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.183919,13585.6
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.137727,13559.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,147.762,1.46669e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.234757,13515.9
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.143222,13500.5
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,30.8424,301925
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.200748,13561.8
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.138672,13529.3
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,148.326,1.46578e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.229147,13479.1
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.145971,13471.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,30.8856,302761
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.166258,13498.3
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.142009,13490.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,148.208,1.46201e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.225336,13465.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.144601,13450.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,31.0143,303698
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.166949,13500
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.136172,13490.5
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,148.247,1.46123e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.226216,13410.6
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.143381,13408.7
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,30.8179,300303
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.199547,13456.3
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.138137,13441.1
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,148.226,1.45887e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.229896,13427.7
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.147097,13420.1
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,30.869,300279
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.166249,13443.1
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.141752,13431.7
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,65.8585,656518
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.198364,13322.2
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.135327,13314.5
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.807,170841
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.146767,13308.4
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.105954,13306.5
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,221.517,2.11754e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.220639,12969.1
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.138908,12951.7
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,31.8758,302433
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.154178,12940.8
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.129292,12931.1
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.274024,12931.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.982346,19414.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0827833,12923.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.7055,136431
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_161.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_161.txt
deleted file mode 100644
index ec4bf136fd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_161.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,73.8414,631151
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.242539,12072.7
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.16669,12064.9
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,204.014,1.77199e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.255566,12266.7
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.155884,12259
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,141.638,1.24598e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.244312,12597.7
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.154118,12605.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,109.314,1.01021e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.244891,12856.5
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.154319,12843.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,74.6864,695116
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.229676,12984.4
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.15126,12980.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,241.463,2.28995e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.253054,13320
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.153273,13322
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,95.1088,896183
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.239807,13546.3
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.156552,13534.9
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,62.5937,621998
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.228834,13629.2
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.149055,13623.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,39.9155,389179
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.21284,13627.4
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.148124,13629.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,137.782,1.37559e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.365959,14374.3
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.155062,13697.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,51.2204,504730
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.228258,13752.2
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.147311,13767.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,44.3395,442975
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.211893,13765.7
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.13869,13765.9
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,22.061,224984
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.182456,13764
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.131964,13756.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,146.744,1.474e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.266543,14269.2
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.139257,13677.9
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,30.8873,310956
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.198693,13723.5
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.137286,13715.9
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,146.728,1.46661e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.228526,13677.1
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.140067,13677.1
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,31.1362,304718
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.203583,13720.3
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.137717,13720.3
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,146.902,1.46867e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.229548,13669.6
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.139993,13662.1
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,30.9505,308310
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.201471,13709
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.138271,13701.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,146.573,1.46703e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.227573,13643
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.142895,13635.3
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,30.9968,309732
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.201286,13709.4
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.134083,13701.8
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,146.49,1.46573e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.246264,13661.8
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.145276,13638.9
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,31.0376,311391
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.199903,13678.7
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.135753,13671.2
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,43.7408,447640
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.171039,13570.5
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.126866,13555.3
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,13.8324,147999
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.149698,13526.8
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0732701,13526.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,141.664,1.3947e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.198121,13262.6
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.133129,13260.7
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,26.505,266031
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.172438,13224.3
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.0853532,13218.3
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.235253,13206.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.994403,20461.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0672859,13187.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.8631,139127
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_162.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_162.txt
deleted file mode 100644
index 02527d21f3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_162.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,73.8135,630460
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.239781,12074.8
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.16229,12076.7
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,205.848,1.78853e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.25524,12216.8
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.164927,12211
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,142.182,1.24893e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.242482,12585.8
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.159045,12599.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,110.367,1.01537e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.248456,12852.7
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.162662,12852.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,74.9507,700660
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.232584,12970.6
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.155391,12974.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,242.41,2.29846e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.259675,13325.8
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.160447,13318.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,95.6071,907276
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.271394,14051.3
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.162002,13534.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,62.8554,626911
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.230712,13619.3
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.153519,13621.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,40.1023,398753
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.203317,13623
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.149542,13617.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,138.461,1.38418e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.247029,13726
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.153794,13714.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,51.1921,505798
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.225032,13815.4
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.155001,13811.3
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,46.7371,478281
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.211394,13801.6
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.143395,13786.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,22.6998,233105
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.182991,13784.7
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.129795,13784.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,151.464,1.52713e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.228754,13738.4
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.150195,13738.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,31.5123,318125
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.196885,13751.4
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.141497,13751.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,152.001,1.52534e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.232968,13697.6
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.145913,13674.9
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,31.51,317531
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.199567,13742.8
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.140191,13735.4
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,151.76,1.52232e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.237976,13698.3
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.144409,13698.3
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,31.5362,318893
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.204191,13709.4
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.143964,13709.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,151.746,1.52187e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.228536,13694.8
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.142022,13685.3
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,31.4799,317124
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.200994,13713.6
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.138153,13705.9
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,151.636,1.51939e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.229628,13664.1
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.144275,13633.8
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,31.6331,318003
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.200764,13721.3
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.139787,13713.8
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,42.7572,439772
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.174015,13605.1
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.134892,13605.1
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,13.5968,140312
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.155187,13567.1
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0838394,13561.4
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,140.21,1.38343e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.202767,13279.5
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.137423,13271.9
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,26.1867,260375
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.16566,13222.8
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.0968313,13220.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.23684,13213.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.997553,20469.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0675742,13197.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.7524,138858
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_163.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_163.txt
deleted file mode 100644
index bfaefe1f93..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_163.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,73.5306,618720
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.23948,11995.2
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.160278,12002.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,204.09,1.76839e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.258773,12169.2
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.156037,12169.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,141.132,1.24245e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.245346,12557.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.154549,12553.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,109.832,1.00734e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.248127,12826.1
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.156757,12826.1
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,74.6003,692660
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.228661,12963.1
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.15245,12965
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,242.491,2.28196e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.253103,13331.6
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.155417,13325.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,95.0336,907107
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.237554,13533.5
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.155826,13535.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,62.5935,621315
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.228398,13659.2
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.146105,13659.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,39.6135,397543
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.20564,13661.1
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.146252,13663.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,137.839,1.38205e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.239839,13740.6
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.149785,13732.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,50.9582,493150
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.223737,13832.2
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.146992,13832.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,46.001,471056
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.205928,13121.2
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.137394,13113.7
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,21.6703,223766
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.184319,13805.5
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.130399,13790.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,150.402,1.51399e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.227282,13742
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.141,13702
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,30.6438,306393
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.199957,13793.3
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.138226,13785.7
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,150.584,1.51175e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.231679,13712.1
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.143673,13710.2
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,30.5536,305402
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.198997,13731.1
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.138623,13719.7
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,150.714,1.51273e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.229429,13673
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.141375,13671.1
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,30.5783,305575
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.20109,13721.9
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.14039,13712.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,150.393,1.50487e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.230805,13674
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.139986,13658.8
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,30.5596,304592
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.197362,13712.6
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.135807,13713
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,150.74,1.49737e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.232216,13677.4
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.14262,13662.4
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,30.5872,303610
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.199298,13698.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.141807,13696.5
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,65.9365,667846
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.198719,13575.9
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.132825,13575.9
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.6592,172608
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.174741,13551.2
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.102927,13549.3
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,221.694,2.14657e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.227048,13133.9
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.141084,13133.9
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,31.8577,309815
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.189375,13121.9
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.124553,13120
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.271298,13108.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.01795,20355.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0693854,13087.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.795,138076
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_164.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_164.txt
deleted file mode 100644
index 82957f7e57..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_164.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,73.3237,616980
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.252975,11904.7
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.156789,11904.7
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,203.71,1.75182e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.257183,12094.3
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.160783,12099.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,141.917,1.24503e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.238699,12459.5
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.158214,12451.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,109.804,1.00448e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.243948,12738.5
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.15534,12740.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,74.5378,683595
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.226549,12869.7
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.160021,12869.7
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,240.355,2.24986e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.253925,13206.9
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.158899,13210.7
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,94.6978,899986
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.241532,13462.9
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.162335,13455.3
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,62.8486,617053
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.239682,13570.7
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.159285,13567.3
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,39.7212,396716
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.20213,13559.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.150242,13566.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,137.215,1.36729e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.252351,13621.6
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.155605,13614
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,51.2211,499446
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.229266,13702.6
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.149071,13704.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.0727,457483
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.212652,13685.6
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.139596,13676
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,19.1047,196344
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.184997,13651.9
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.131113,13652.1
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,144.047,1.44141e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.22854,13589.9
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.154277,13582.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,27.3837,274863
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.17157,13646.6
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.139769,13639
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,144.48,1.44028e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.229426,13571.2
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.142994,13563.5
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,27.2863,271498
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.209167,13595.7
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.1415,13571
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,145.157,1.44271e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.232255,13521
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.143574,13503.9
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,27.4124,273426
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.169881,13565.2
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.144831,13565.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,144.194,1.43332e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.226523,13519.1
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.144755,13496.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,27.5437,272377
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.171343,13528.6
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.139577,13521
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,144.397,1.43233e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.238267,13517.5
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.145174,13502.2
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,27.3829,272840
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.168835,13523.2
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.140742,13521.3
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,65.9261,660511
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.197484,13423.2
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.135283,13415.5
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.738,173898
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.147356,13393.8
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0986681,13382.1
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,221.58,2.12602e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.217656,12998
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.140076,12988.4
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,31.9016,306467
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.15629,12985.9
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.129427,12985.9
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.283336,12970.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.980369,19534.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0798522,12962.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.7891,136997
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_165.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_165.txt
deleted file mode 100644
index cd027ffd1d..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_165.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,69.3191,574626
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.236443,11743.2
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.161801,11743.3
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,201.037,1.70432e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.252194,11916.6
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.166258,11918.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,132.925,1.12746e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.240975,12253.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.1583,12262.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,108.741,976297
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.24661,12497.8
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.159158,12501.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,72.0484,646729
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.234565,12614.2
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.160024,12616.1
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,238.309,2.20005e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.25172,12985.9
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.166444,12980.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,92.4312,855102
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.239563,13230.6
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.155676,13224.9
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,61.9394,600469
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.23356,13330.3
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.150198,13332.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,38.9737,379692
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.207391,13326.5
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.146089,13326.5
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,136.564,1.335e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.248737,13370.1
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.155695,13372.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,50.3563,484603
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.231282,13480.9
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.151647,13458
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.4751,452573
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.210892,13458.1
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.148863,13458.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,22.5207,223934
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.184828,13446.4
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.135679,13431
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,148.642,1.45985e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.230367,12695.4
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.149679,12680
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,31.6327,306081
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.179061,13399.2
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.146617,13382
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,149.195,1.45513e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.237883,13288.4
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.144367,13288.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,31.1472,304086
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.208437,13340
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.140396,13332.3
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,149.03,1.45434e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.234428,13272.9
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.145771,13253.7
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,31.1364,302884
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.168044,13317.1
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.142767,13315.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,148.752,1.44963e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.228073,13249.6
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.145705,13203.6
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,31.1847,302886
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.211768,13309.2
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.138828,13293.9
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,148.805,1.44547e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.22502,13232.1
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.147206,13216.9
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,31.281,304041
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.170371,13285.9
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.141096,13278.2
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,44.15,437820
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.151567,13188.4
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.134178,13177
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,13.7904,143243
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.114892,13154.2
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0781433,13152.3
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,140.601,1.35055e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.167247,12963.2
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.136684,12947.8
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,26.487,259063
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.137045,12926.3
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.0862073,12908.9
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.238485,12897.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.982116,19970
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0672891,12869.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.7986,135293
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_166.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_166.txt
deleted file mode 100644
index 0a1f2f2a19..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_166.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,68.7667,574325
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.240619,11857.4
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.16045,11859.3
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,199.437,1.692e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.254156,12003.4
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.164009,12003.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,132.105,1.14462e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.235595,12305.1
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.159211,12307
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,107.663,973206
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.244917,12572.4
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.167698,12574.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,71.6641,652760
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.227627,12679.4
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.159772,12696.7
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,238.694,2.20659e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.252536,13073.7
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.15734,13075.7
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,91.8926,858257
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.236524,13272.5
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.159874,13270.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,62.2652,605043
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.234482,13418.6
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.154556,13405.1
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,39.6306,388095
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.20694,13403.4
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.153762,13395.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,137.151,1.34841e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.244786,13454.4
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.152815,13448.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,50.6778,491265
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.222187,13531.5
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.154853,13525.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,46.1488,462071
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.209093,13536.1
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.147392,13536.1
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,22.9519,231737
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.184479,13515.2
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.137759,13515.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,149.264,1.47922e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.228498,13450.9
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.14774,13443.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,31.8358,313201
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.200178,13523.4
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.142124,13515.8
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,149.572,1.47623e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.229621,13439.2
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.152786,13427.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,31.8346,312797
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.200021,13473.7
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.144441,13466
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,149.62,1.47458e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.228017,13448.7
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.152805,13441.1
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,31.909,312018
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.201557,13454.5
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.140764,13454.5
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,149.771,1.47547e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.243183,13427.7
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.149849,13420.1
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,32.009,310425
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.200818,13469.7
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.142956,13462.1
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,149.492,1.47339e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.227759,13403
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.148809,13387.7
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,31.9162,312094
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.199403,13446.8
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.140626,13431.6
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,43.3567,434003
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.176009,13349
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.135337,13345.2
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,13.6609,138922
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.119318,13324.3
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0870107,13316.6
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,141.172,1.37186e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.198732,13080.2
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.137462,13070.7
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,26.3022,259044
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.172399,13049.1
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.100758,13039.5
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.238927,13014.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.01879,20127.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0675326,12985
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.7069,137292
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_167.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_167.txt
deleted file mode 100644
index 1b522abbf0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_167.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,68.6673,552231
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.241208,11726.2
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.155058,11720.5
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,200.192,1.70167e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.257295,11921.1
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.161327,11923
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,131.542,1.12377e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.239183,12239.9
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.162457,12232.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,107.076,962431
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.244555,12508.8
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.162063,12501.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,69.6903,626210
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.223259,12618.4
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.160703,12620.3
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,236.566,2.18903e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.248223,12991.4
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.160949,12985.7
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,89.7156,835641
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.240424,13234.7
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.157826,13240.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,62.0907,603424
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.227704,13335.9
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.151317,13337.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,39.5709,386957
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.206626,13336
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.151522,13336
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,137.07,1.34385e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.246549,13408.6
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.153139,13408.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,50.8769,491270
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.219637,13464.3
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.156767,13464.3
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.8861,459555
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.206898,13452.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.151388,13437.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,22.101,217288
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.184425,13429.7
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.136905,13429.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,148.381,1.45915e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.229173,13397.2
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.146972,13389.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,30.8738,301642
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.193752,13426.1
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.139452,13403
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,149.036,1.46068e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.235433,13366.7
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.144086,13366.7
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,30.858,297086
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.196265,13435.3
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.143116,13435.3
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,148.774,1.45894e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.22797,13347.6
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.146252,13347.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,30.786,299719
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.207135,13408.7
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.141891,13378.1
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,148.546,1.45724e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.226053,13353.2
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.145618,13353.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,30.9501,300799
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.196591,13372.5
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.140838,13362.9
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,148.89,1.45648e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.233234,13362.6
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.145974,13354.9
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,30.9092,300797
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.193576,13380
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.140057,13372.3
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,65.843,653244
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.192511,13268.3
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.134649,13268.3
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.6744,171486
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.188306,13254.9
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0999609,13253
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,221.692,2.10795e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.221845,12889.8
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.163077,12882.1
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,31.7774,300414
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.183574,12255.7
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.125654,12240.3
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.277272,12236.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.01792,19313.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0706011,12192.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,14.0469,136948
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_168.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_168.txt
deleted file mode 100644
index feaab7025a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_168.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,68.0346,567023
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.242926,11653.9
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.157193,11633
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,198.122,1.67044e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.250604,11865.9
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.15671,11867.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,130.013,1.11358e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.239193,12203.1
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.155884,12205
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,107.084,959899
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.244674,12488
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.15702,12474.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,69.8678,621551
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.225858,12572.3
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.156917,12572.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,236.068,2.17194e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.253093,12960.8
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.154319,12953.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,89.75,832064
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.237247,13158.6
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.154725,13150.9
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,61.5093,591695
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.236501,13313.2
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.15198,13317
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,37.9497,367549
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.205176,13299.7
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.143887,13294
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,135.42,1.31761e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.248952,13345.5
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.153813,13345.5
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,49.1257,471004
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.224338,13416.1
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.144508,13408.7
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,44.7035,441269
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.20708,13404.7
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.144809,13387.5
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,19.0943,190322
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.183471,13400.9
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.129663,13385.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,146.109,1.42987e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.232396,13298.1
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.14292,13288.5
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,27.4608,269446
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.202956,13334.3
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.139001,13334.3
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,146.52,1.4281e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.229522,13257.5
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.137132,13249.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,27.5975,268547
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.204408,13293.9
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.142655,13286.2
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,146.527,1.42242e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.231272,13205.4
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.14526,13190.2
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,27.611,258973
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.203848,13226.3
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.142873,13226.3
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,146.34,1.41504e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.233778,13184.5
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.144911,13161.7
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,27.5379,263076
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.200987,13235.8
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.13606,13235.8
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,146.213,1.41418e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.229749,13141.5
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.143324,13141.5
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,27.4285,266700
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.203,13192.8
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.135372,13177.6
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,65.9098,644504
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.19382,13116.1
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.132486,13104.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.6885,169061
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.170767,13104.2
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0958996,13098.2
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,221.829,2.08623e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.223932,12736.2
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.139416,12736.2
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,31.7754,300848
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.186473,12762.3
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.121526,12762.3
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.272642,12743.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.01082,19715.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0705436,12721.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.7132,133774
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_261.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_261.txt
deleted file mode 100644
index 5764c923a5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_261.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,34.7468,335268
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.235557,13200.9
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.159352,13200.9
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,188.164,1.78043e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.241419,13365.6
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.158783,13354.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,61.9979,590333
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.213993,13439.7
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.151324,13424.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,95.3444,954185
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.235346,13743.4
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.156463,13737.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,32.7293,318875
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.200117,13729.8
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.145378,13724
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,197.554,1.98905e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.240514,13917.1
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.160147,13921
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,52.1635,514375
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.215471,14052.2
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.17894,14056
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,59.7311,622031
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.207858,14206.9
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.146476,14208.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,20.8594,219771
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.179158,14168.7
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.101334,14168.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,131.122,1.36675e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.233304,14193.9
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.150223,14186.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,32.7903,336046
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.212814,14293
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.142086,14293
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.1962,432834
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.199724,14300.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.137625,14283.5
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,25.6303,273951
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.165756,14352.6
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.0852767,14337.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,95.0072,1.01095e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.220683,14426.8
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.139564,14426.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,49.6761,531193
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.193132,14548.9
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.103925,14548.9
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,95.1137,1.02728e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.217877,14622.9
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.142115,14615.2
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,49.5112,539310
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.181189,14729.1
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.0981597,14713.9
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,95.2238,1.03906e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.220635,14755.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.139324,14750
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,49.51,542456
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.183189,14866.4
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.10278,14866.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,95.1583,1.04659e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.217365,14870.2
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.142703,14872.1
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,49.5449,546845
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.146153,14966
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.102185,14950.6
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,95.1242,1.05274e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.215711,14988.9
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.141375,14981.3
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,49.6454,551692
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.189759,15059.8
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.0977175,15059.8
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,30.3868,347903
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.180328,15052.1
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.13094,15044.5
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,15.1535,174602
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.168495,15004.3
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0925081,14996.6
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,103.346,1.14545e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.20974,14898.6
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.13838,14883.3
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,29.2595,326863
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.184962,14793.9
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.10583,14784.4
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.249224,14761.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.00365,22521
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0667038,14748.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.3432,156685
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_262.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_262.txt
deleted file mode 100644
index 9d6d32c0c2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_262.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,35.2933,336232
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.215829,13182
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.16686,13174.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,187.044,1.77672e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.238293,13350.7
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.161567,13354.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,61.4642,584495
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.216792,13406
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.152444,13400.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,95.044,944613
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.229416,13652.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.164559,13652.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,32.8006,315961
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.199932,13605
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.151244,13608.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,197.839,1.97877e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.241247,13803.4
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.168761,13790.3
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,52.2454,516670
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.216604,13924.7
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.157567,13924.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,60.1604,617242
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.205295,14096.4
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.154457,14090.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,21.2047,220074
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.18558,14069.5
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.111026,14077.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,131.422,1.35975e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.233743,14117.6
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.159439,14109.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,33.0462,335225
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.211237,14205.5
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.152405,14205.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.3616,430000
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.200959,14211.3
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.142543,14211.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,25.7155,273098
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.168764,14264.6
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.0937949,14234.1
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,95.5801,1.00776e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.219084,14354.2
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.152822,14331.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,49.4638,526699
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.186341,14473.5
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.105312,14450.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,95.3545,1.01838e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.222056,14553.6
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.147052,14538.3
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,49.4709,534382
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.154511,14660.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.100972,14645.3
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,95.2174,1.03527e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.215919,14685.3
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.144258,14670.1
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,49.5222,541826
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.151487,14773
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.102262,14771.1
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,95.4371,1.04533e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.221906,14796.1
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.147052,14788.4
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,49.4384,545041
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.152898,14903.1
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.103965,14885.9
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,95.3308,1.05056e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.222447,14920.3
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.150613,14907
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,49.392,548923
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.147727,15015.7
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.103875,15008.1
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,30.3057,345984
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.186344,15013.9
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.132668,14998.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,15.2037,175408
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.136431,14943.3
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.0966875,14933.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,103.299,1.14147e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.212952,14843.4
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.142169,14835.8
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,29.2995,324166
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.151148,14721.6
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.116937,14714
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.254367,14714
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.98345,22492.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0701629,14696.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.3137,154797
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_263.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_263.txt
deleted file mode 100644
index 5bd59d30a0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_263.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,43.9947,429024
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.238895,13535.3
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.173062,13533.7
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,271.997,2.63297e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.238738,13720.9
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.161378,13719
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,66.9047,646028
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.218233,13730.5
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.15608,13722.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,136.932,1.38122e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.229388,13962.5
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.160652,13966.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,35.9886,354647
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.209474,13974.1
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.151762,13960.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,287.028,2.93179e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.23454,14219.6
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.160588,14208
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,57.1948,573312
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.22621,14302.2
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.16343,14306.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,85.8868,899638
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.219551,14474
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.15821,14460.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,23.214,242716
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.188582,14437.7
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.121401,14447.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,180.243,1.91073e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.238383,14465.8
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.159004,14458.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,37.0562,386566
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.214591,14549.9
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.147427,14551.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,48.1684,523008
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.200982,14576.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.13942,14576.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,31.728,346355
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.168511,14658.7
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.0972408,14658.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,116.281,1.26585e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.224591,14808.9
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.143794,14801.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,59.4537,654842
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.190713,14971.6
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.112208,14971.6
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,116.333,1.28943e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.219241,15069.1
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.145331,15046.1
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,59.4544,663491
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.157176,15229.1
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.116576,15221.5
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,116.33,1.3099e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.220723,15291.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.140172,15293.7
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,59.5804,676158
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.163391,15422.3
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.115487,15415
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,116.449,1.32912e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.227142,15469.2
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.143212,15454.1
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,59.6298,678938
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.166425,15587.1
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.118591,15579.4
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,116.215,1.33947e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.221916,15608.1
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.143376,15610
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,59.5738,689721
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.159478,15717.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.115145,15713.6
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,39.2803,462516
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.157212,15690.8
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.135881,15683.1
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.2666,191962
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.14735,15607.3
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.109309,15605.4
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,149.547,1.70744e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.215286,15291.7
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.139692,15276.5
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,31.2073,352971
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.154032,15169.4
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.131046,15163.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.258079,15152.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.977345,22692.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0791673,15126.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.5271,159046
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_264.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_264.txt
deleted file mode 100644
index 6ae91b073a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_264.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,43.0578,421884
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.214812,13509
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.156427,13501.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,272.508,2.63415e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.241632,13681.8
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.166139,13687.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,66.4178,642908
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.218409,13740.4
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.155096,13734.7
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,136.769,1.38371e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.23343,13985.5
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.158338,13989.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,35.762,358862
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.203519,13957.2
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.151323,13957.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,288.131,2.93659e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.238605,14236.5
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.159807,14236.5
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,57.6279,581751
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.221084,14321.8
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.15844,14323.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,86.2336,907284
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.221942,14489.6
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.15181,14493.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,23.4802,248247
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.187858,14471
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.124622,14465.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,180.724,1.91635e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.23222,14490.6
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.155435,14492.5
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,37.3881,388915
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.209916,14576.9
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.151451,14571.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,48.6641,527750
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.206597,14586.5
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.144513,14582.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,31.694,345698
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.167983,14692.8
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.0933739,14668
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,116.311,1.26531e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.22319,14795.6
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.144773,14778.5
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,59.5977,652013
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.195503,14971.3
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.120049,14956.1
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,116.306,1.28727e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.221833,15049.8
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.142513,15053.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,59.7765,667706
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.166709,15196.8
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.123617,15181.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,116.405,1.31163e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.228537,15306.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.141649,15284
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,59.6347,674600
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.168309,15419.3
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.117682,15412.5
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,116.76,1.33236e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.222748,15445.9
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.150523,15440.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,59.5797,682775
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.192722,15566.1
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.115624,15559.2
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,116.612,1.34262e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.225119,15590.3
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.144351,15575.1
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,59.569,690163
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.195113,15690.7
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.110791,15689
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,39.4258,463999
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.189731,15675.4
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.130619,15660.5
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.2577,193112
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.177509,15590.2
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.108363,15582.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,149.611,1.70258e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.220252,15255.5
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.136232,15247.9
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,31.1682,353272
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.193814,15148.3
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.130507,15142.7
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.265622,15131
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.00494,23412.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0722314,15092.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.6637,160822
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_265.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_265.txt
deleted file mode 100644
index c74cb8bdba..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_265.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,43.0297,421590
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.216623,13512.4
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.161039,13504.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,271.259,2.65759e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.250704,13709.9
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.158408,13706.1
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,66.9953,631417
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.220338,13734.5
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.15861,13721.1
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,136.047,1.37755e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.233353,14008.4
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.161205,14010.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,35.6025,356137
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.210802,13974.2
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.15131,13266.7
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,287.591,2.92254e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.239705,14221.4
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.16699,14211.9
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,57.183,573760
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.221913,14350.6
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.160927,14331.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,85.8869,901906
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.217967,14530.2
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.150658,14522.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,23.1587,246421
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.191759,14495.6
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.118612,14497.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,180.139,1.91246e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.241638,14523.1
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.159205,14515.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,37.1515,386630
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.21773,14594.3
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.151064,14600.3
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,48.151,521378
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.204197,14636.4
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.139547,14621.1
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,31.8806,346492
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.178757,14691
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.096356,14691
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,116.125,1.26545e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.24807,14816.4
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.153675,14808.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,59.5723,655922
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.159397,15021.4
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.115953,15009.9
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,116.411,1.29243e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.226863,15101.4
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.144568,15093.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,59.5557,666248
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.162203,15236.7
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.117784,15229.1
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,116.319,1.30925e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.222786,15306.9
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.144926,15286.2
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,59.6213,676288
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.156776,15417.6
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.115205,15416
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,116.178,1.32115e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.230575,15465.3
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.141096,15457.9
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,59.6443,682331
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.195749,15587.2
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.11828,15579.5
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,116.42,1.33942e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.224425,15607.5
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.147016,15586.5
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,59.684,689434
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.190479,15713.6
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.115559,15694.6
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,39.28,461424
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.193173,15673.5
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.13644,14878.5
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.4487,189484
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.17452,15603.2
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.113435,15603.2
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,149.442,1.70307e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.22726,15270.9
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.144421,15263.3
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,31.278,353649
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.18965,15169.9
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.131013,15160.2
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.267964,15140.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.00304,23360.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0676843,15117.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.452,158737
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_266.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_266.txt
deleted file mode 100644
index 6c3baca148..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_266.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,39.7146,378068
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.218108,13238.4
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.159422,13236.5
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,239.56,2.26178e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.241266,13340
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.164277,13341.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,67.4237,627026
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.225317,13389.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.159643,13389.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,121.455,1.18506e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.236952,13623.4
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.157358,13614.1
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,36.5663,356874
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.211567,13595.9
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.160708,13597.9
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,258.364,2.54603e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.240079,13820.3
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.161755,13812.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,59.1213,582336
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.227391,13945.9
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.161138,13940.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,80.4979,818867
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.226783,14145.9
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.156463,14153.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,24.1107,247697
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.200082,14140.3
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.133041,14132.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,169.174,1.74982e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.241122,14159
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.157157,14145.7
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,39.0231,395088
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.212047,14271.4
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.152168,14261.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,48.6749,514228
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.209736,14279
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.145368,14279
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,33.6214,356908
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.171435,14365.5
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.0998312,14350.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,118.443,1.26131e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.230053,14506.8
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.147041,14499.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,63.9063,683456
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.196482,14700.1
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.11795,14684.9
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,118.488,1.28619e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.229698,14816.6
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.145739,14820.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,63.8902,700698
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.162098,14979.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.134283,14972.2
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,118.587,1.31137e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.22653,15071.9
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.142443,15041.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,63.9643,713251
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.164258,15223.1
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.124923,15223.1
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,118.392,1.33126e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.22989,15285.2
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.145662,15270
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,64.2011,724774
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.165378,15399
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.133384,15399.4
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,118.576,1.34731e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.225346,15421.9
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.144219,15422.3
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,64.1962,731146
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.174882,15539.3
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.131089,15524.1
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,46.0521,531288
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.198956,15437.5
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.139137,15438.1
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.8328,200874
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.151061,15390
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.103898,15373.1
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,167.132,1.86052e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.227692,14944.2
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.143841,14936.6
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,32.1129,354746
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.162549,14835.7
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.125841,14820.5
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.272499,14803.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.995802,22882.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0691849,14793.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.8295,157123
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_267.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_267.txt
deleted file mode 100644
index e9b0012ca0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_267.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,40.807,387831
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.220335,13226.9
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.16691,13219.3
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,241.286,2.28873e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.242402,13325
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.16435,13321.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,68.1855,633617
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.222309,13363.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.161061,13367
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,122.519,1.1961e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.245535,13597.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.164098,13592.1
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,36.4864,348271
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.209871,13575.2
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.15557,13552.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,258.982,2.55435e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.240303,13783.1
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.163179,13775.5
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,59.5061,579185
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.223708,13926.8
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.155755,13917.3
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,80.6501,819398
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.223701,14094.4
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.156642,14102.1
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,23.992,246036
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.193826,14096.3
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.13169,14100.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,169.762,1.75316e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.240492,14117.2
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.156264,14119.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,39.0104,395340
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.212741,14260
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.148468,14252.3
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,48.3044,509708
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.202747,14254.4
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.13979,14254.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,33.4185,355708
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.174098,14332.9
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.0953958,14332.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,118.633,1.25677e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.231973,14484
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.14668,14476.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,63.8179,687296
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.16187,14692.6
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.116843,14685
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,118.798,1.2919e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.227439,14812.7
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.146049,14782.2
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,63.8659,698343
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.159806,14989.2
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.126929,14989.2
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,118.613,1.31388e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.229301,15041.3
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.476235,18046.9
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,63.6028,710556
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.159915,15238.6
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.116596,15223.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,118.725,1.33224e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.234994,15237.1
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.152865,15231.5
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,63.8327,720263
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.163807,15400.7
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.122698,15378.1
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,118.892,1.34722e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.227666,15421.5
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.146443,15406.3
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,63.9231,730130
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.161755,15534.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.120267,15534.4
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,45.9485,530574
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.19564,15451.1
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.135684,15428.5
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.7976,199227
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.14458,15389.8
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.104593,15372.7
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,166.909,1.86195e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.221455,14938.5
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.142772,14930.9
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,32.1133,354758
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.15667,14849.1
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.125627,14818.6
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.257753,14814.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.983078,22917.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0701449,14814.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.603,155417
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_268.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_268.txt
deleted file mode 100644
index 805f15081a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_268.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,41.0609,389058
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.248735,13243.9
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.164907,13230.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,240.56,2.2744e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.239839,13317.2
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.163666,13321
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,68.2105,640397
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.220434,13378
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.160491,13379.9
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,122.194,1.19443e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.231138,13605.1
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.16555,13599.5
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,36.4448,348862
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.206357,13575.4
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.15819,13571.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,258.714,2.55796e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.234524,13822.8
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.162609,13809.3
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,59.2753,577422
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.21676,13918.9
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.159845,13920.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,80.5336,819867
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.222091,14096.1
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.159339,14090.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,23.9475,248301
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.194347,14098
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.122091,14105.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,169.449,1.75265e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.234729,14145.6
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.153848,14139.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,39.1202,396021
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.216245,14237.3
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.165019,14241.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,48.253,511164
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.201928,14256.6
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.142801,14233.5
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,33.4833,357950
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.138616,14338.5
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.0883655,14329
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,118.579,1.26092e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.229301,14493.5
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.148731,14485.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,63.8307,685593
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.164107,14680.7
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.126456,14680.7
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,118.899,1.29456e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.224338,14784.1
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.148226,14776.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,63.7843,700211
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.162516,14978
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.119371,14964.7
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,118.824,1.31562e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.231487,15049.1
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.147883,15033.8
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,63.9343,712596
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.165422,15192.9
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.118538,15192.9
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,118.816,1.33376e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.230863,15231
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.148936,15223.3
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,64.0385,721620
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.15962,15370.8
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.121418,15355.7
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,118.945,1.35219e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.224491,14615.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.141688,15376.1
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,64.0118,729402
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.158462,15527.1
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.118494,15512.3
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,46.0323,531934
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.195115,15447.7
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.134382,15440.1
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.8233,196867
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.149906,15365.1
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.10114,15340.9
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,166.984,1.85964e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.22461,14953.7
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.144459,14946.1
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,32.2532,357170
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.165377,14841.6
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.121608,14841.6
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.266719,14826.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.987314,22923
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0666438,14818.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.6465,155554
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_269.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_269.txt
deleted file mode 100644
index f64ae15dfc..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_269.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,41.0589,389526
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,0.239077,13229
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.165512,13213.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,241.763,2.27007e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.237893,13285.2
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.161691,13285.1
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,68.0729,640085
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.224187,13353.6
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.156254,13353.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,122.153,1.1931e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.229454,13597.7
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.165598,13592.1
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,36.4707,343294
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.21012,13567.9
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.157995,13546.9
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,258.589,2.55164e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.240277,13782.8
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.166158,13784.7
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,59.2977,569658
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.219281,13917.2
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.158853,13894.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,80.386,817142
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.216875,14098.1
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.157915,14090.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,23.9888,244454
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.203624,14098.1
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.138408,14090.5
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,169.163,1.75117e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.234482,14122.7
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.1621,14099.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,39.3339,400570
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.215282,14225.5
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.151787,14227.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,48.4712,512325
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.206414,14233.1
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.146151,14235
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,33.6261,357581
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.170184,14355.9
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.0956167,14354
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,118.948,1.24848e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.224434,14495.1
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.143509,14479.9
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,63.8797,686364
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.196584,14665.7
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.121214,14652.3
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,119.21,1.29817e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.223486,14791.7
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.148053,14778.5
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,64.0499,700906
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.16453,14947.1
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.129873,14924.2
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,118.957,1.31703e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.231995,15031.9
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.152062,14993.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,64.0246,712880
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.158401,15179.6
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.124209,15172
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,118.943,1.33614e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.225058,15206.6
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.15124,15201
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,63.9481,716213
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.194782,15367.5
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.126907,15360.1
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,119.141,1.34272e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.22451,15388.5
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.147445,15380.9
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,63.9854,709064
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.195573,15498
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.116977,15498
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,46.0994,530583
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.190437,15438.4
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.136027,15419.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,16.8004,201115
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.178411,15357.9
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.100967,15331.3
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,167.193,1.86327e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.225605,14921.4
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.143633,14919.5
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,32.2525,357214
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.193957,14818.8
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.120766,14811.2
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.25774,14799.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.01096,22894.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0683877,14774.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,13.7291,155959
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp32_perf_fp32_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp32_perf_fp32_120.txt
deleted file mode 100644
index a32a995a82..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp32_perf_fp32_120.txt
+++ /dev/null
@@ -1,255 +0,0 @@
-Conv1,96.9252,865993
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-BatchNorm1,1.44629,18053.8
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Relu1,0.468751,12163.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,380.5,3.34989e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-BatchNorm2,0.320127,12396.4
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.158969,12386.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,224.11,1.88066e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-BatchNorm3,0.258617,12404.1
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.133142,12409.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,208.316,1.9126e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-BatchNorm4,0.245676,13072.1
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Relu4,0.141856,13068.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,81.0615,730270
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-BatchNorm5,0.244313,13234.1
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Relu5,0.189974,13243.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,593.95,5.13284e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-BatchNorm6,0.27095,12120.5
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu6,0.191753,12120.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,146.124,1.21007e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-BatchNorm7,0.263548,12267.4
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu7,0.192726,12279.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,123.276,1.05377e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-BatchNorm8,0.245183,12473.2
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Relu8,0.123152,12465.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,43.0723,393906
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-BatchNorm9,0.177062,12631.5
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu9,0.107667,12633.5
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,266.287,2.49353e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-BatchNorm10,0.242575,13217.8
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu10,0.132793,13223.5
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv11,67.4417,625092
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-BatchNorm11,0.239804,13424.4
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Relu11,0.175203,13418.7
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,75.8047,715723
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-BatchNorm12,0.241129,13595.4
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu12,0.151622,13585.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,27.137,260044
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-BatchNorm13,0.217503,13681.6
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu13,0.14694,13672.1
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv14,199.051,2.00991e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-BatchNorm14,0.2499,14265.6
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Relu14,0.156185,14265.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv15,41.0208,407067
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-BatchNorm15,0.23191,14400.4
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Relu15,0.15607,14400.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv16,198.945,2.13558e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-BatchNorm16,0.253142,15090
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu16,0.160892,15065
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv17,40.8306,435387
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-BatchNorm17,0.232492,15161
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu17,0.153104,15151.3
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv18,205.78,2.29436e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-BatchNorm18,0.257356,15630.6
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Relu18,0.154368,15621.1
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv19,40.2938,441269
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-BatchNorm19,0.235654,15674.7
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu19,0.156019,15655.5
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv20,236.123,2.63379e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-BatchNorm20,0.260556,15668.4
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu20,0.155836,15660.8
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv21,40.2208,434870
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-BatchNorm21,0.231612,15753
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Relu21,0.159161,15730.1
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv22,234.944,2.64206e+06
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-BatchNorm22,0.249119,15735.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu22,0.160815,15717.9
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv23,39.8626,434106
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-BatchNorm23,0.239164,15796.7
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu23,0.156585,15777.4
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv24,84.0255,958156
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-BatchNorm24,0.228284,15628.5
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Relu24,0.146156,15613.1
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv25,21.2845,239748
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-BatchNorm25,0.186892,15613.1
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu25,0.110233,15607.2
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv26,305.989,3.38538e+06
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-BatchNorm26,0.246278,15220.6
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu26,0.151472,15203.2
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv27,40.6724,447027
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-BatchNorm27,0.207692,15250.9
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Relu27,0.132445,15241.2
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Pool1,0.908599,15233.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,1.12393,23627.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add1,0.0794172,15208.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Softmax1,14.3291,173664
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_120.txt
deleted file mode 100644
index e5768fe5ea..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_120.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,62.6208,504839
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.204305,11274.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.173355,11272.9
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,177.564,1.46427e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.220104,11391.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.171121,11382.1
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,179.212,1.47709e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.220852,11464.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.148129,11452.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.159691,11460.5
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,180.398,1.5005e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.220692,11556
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.175614,11554
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,179.578,1.50176e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.21883,11555.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.155486,11550
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.161313,11559.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,180.753,1.51802e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.219291,11666.9
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.172926,11647.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,180.126,1.51393e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.22139,11636
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.147553,11632.1
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.164721,11639.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,47.5749,393350
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.197073,11783.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.168151,11780
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,76.0434,647866
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.208026,11838.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,19.6613,166734
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.174935,11869.3
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.14193,11873.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.165761,11883.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,75.0139,654553
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.200519,11928.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.165636,11920.5
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,75.9454,655681
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.20459,11960.9
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.148145,11949.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.170452,11955.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,76.3321,664495
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.202321,12030.2
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.169706,12028.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,75.7521,659866
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.202228,12051.3
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.146369,12051.1
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.161882,12060.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,24.4298,210793
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.181179,12086.7
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.146884,12092.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,38.8564,344124
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.180357,12125.2
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,10.888,96816.9
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.155514,12134.9
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.119579,12113.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.122762,12125
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,37.5035,339187
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.169486,12146.3
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.141293,12138.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,38.8939,342824
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.187201,12153.3
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.126532,12149.4
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.148068,12160.9
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,39.1741,348960
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.183642,12157.2
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.148407,12149.3
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,39.0636,351543
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.182654,12185.6
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.127377,12181.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.145982,12198.7
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.528313,12200.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.30332,12202.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.12313,12172.3
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.3872,13996.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_151.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_151.txt
deleted file mode 100644
index 786d8dbd8e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_151.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,42.1025,331178
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.203914,11070.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.167274,11072.1
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,77.2909,624173
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.20379,11185
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.165642,11171.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,78.8872,641882
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.21258,11297.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.156017,11288.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.156026,11303.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,79.988,655956
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.202046,11393.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.165655,11393.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,79.2363,654772
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.202391,11454.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.149879,11456.7
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.157962,11460.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,79.9301,667196
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.203089,11592.3
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.168244,11590.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,79.0317,663786
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.198548,11645.7
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.149294,11636
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.156958,11643.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,29.7754,250662
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.201719,11739
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.160871,11744.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,39.3569,342497
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.190558,11758.1
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,18.1525,156927
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.152638,11794.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.116695,11799.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.0913478,11803.5
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,38.6218,342251
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.175514,11823
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.164484,11822.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,39.8432,345502
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.182158,11813.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.141978,11823.1
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.160209,11834.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,40.0361,346066
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.182388,11845.9
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.165822,11847.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,39.9824,345056
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.186641,11834.9
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.146154,11842.5
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.164461,11848
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,15.6067,135479
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.15931,11868.7
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.138327,11864.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,20.4178,180705
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.154503,11878.3
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,10.4109,95442.7
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.138993,11876.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0791971,11886
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0786115,11887.9
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,20.4469,182135
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.154919,11876.5
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.138039,11880.3
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,21.0178,188571
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.163204,11871
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.129565,11878.7
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.143633,11882.5
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,20.9779,187402
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.160859,11890.4
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.142465,11875.1
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,21.0379,189157
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.164174,11869.3
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.122781,11874.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.151527,11874.8
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.46661,11886.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.24731,11886.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.0726629,11878.8
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.23727,12451.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_152.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_152.txt
deleted file mode 100644
index 600dc88cbe..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_152.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,41.7255,328354
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.195306,11043.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.172791,11030.3
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,76.8293,618767
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.202491,11167.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.164016,11162
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,77.9585,629627
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.204298,11253.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.147258,11253.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.153575,11259.5
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,79.4552,645443
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.204382,11412.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.166052,11404.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,78.5592,650498
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.202612,11458.7
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.152455,11468.1
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.157041,11471.9
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,79.5538,662985
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.202208,11578.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.170609,11576.9
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,78.6036,661670
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.201694,11639.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.147361,11639.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.15827,11645.3
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,29.5924,247685
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.183873,11734.9
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.161406,11744.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,39.4092,342559
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.18418,11740.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,18.2101,157182
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.154878,11763.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.119354,11763.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.0931747,11775
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,38.7238,337313
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.17578,11808
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.159248,11794.3
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,40.0541,345549
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.182059,11786.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.141815,11800.2
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.165885,11804
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,40.1396,346722
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.17819,11807.9
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.161015,11804
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,39.9928,347406
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.179018,11805.8
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.143879,11803.8
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.162426,11817.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,15.7216,137456
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.157719,11830.4
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.138039,11834.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,20.479,181242
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.153914,11838.3
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,10.4474,95123.7
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.136001,11851.7
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0753923,11838.2
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0775427,11849.7
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,20.3511,181402
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.150442,11850
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.135665,11851.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,20.9284,188101
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.155991,11851.9
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.12353,11855.7
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.138263,11859.4
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,20.6996,183975
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.155792,11863.4
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.135159,11859.5
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,21.0454,186307
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.156814,11871.1
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.123629,11863.4
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.1473,11865.3
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.470191,11878.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.250315,11874.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.0783232,11867.2
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.25459,12876
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_153.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_153.txt
deleted file mode 100644
index a8c00d4576..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_153.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,39.2902,302746
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.194282,10753
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.171389,10772
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,74.5737,591991
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.208363,10884.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.164769,10888.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,76.563,599496
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.206788,10997.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.14889,10991.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.158644,10995.5
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,77.5378,622350
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.204366,11152.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.182657,11158.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,76.5488,616585
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.203781,11171.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.150526,11171.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.157367,11179.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,77.5713,634131
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.206548,11309.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.166986,11313.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,76.8136,623860
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.202513,11364.5
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.147665,11360.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.157863,11374.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,29.2647,238370
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.181006,11441.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.161915,11447.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,38.9844,330356
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.18122,11456.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,17.8395,151609
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.157044,11467.9
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.109652,11464.1
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.0951045,11489.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,38.4034,327927
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.174897,11525.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.167889,11519.5
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,39.8942,337672
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.182264,11498.3
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.143038,11494.6
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.165034,11509.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,39.8426,338902
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.179719,11530.9
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.161051,11536.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,39.807,336342
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.190631,11534.4
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.143041,11544
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.160606,11547.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,15.6155,133057
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.153764,11559.3
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.144916,11561.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,20.6401,179060
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.156493,11559.3
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,10.4302,92865.8
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.140883,11551.7
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0839937,11544.1
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0774945,11557.5
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,20.341,178133
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.148689,11549.9
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.135156,11555.7
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,21.0171,182230
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.158154,11553.8
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.125069,11553.8
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.141309,11559.5
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,20.8662,179452
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.150285,11555.7
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.139575,11559.5
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,21.0884,182911
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.166781,11557.6
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.126864,11559.5
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.142695,11567.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.477797,11576.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.254052,11574.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.0714568,11546
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.26098,12133.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_154.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_154.txt
deleted file mode 100644
index 6480cd1c98..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_154.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,38.7211,300352
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.196947,10836.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.165751,10838.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,74.523,594803
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.202011,10941.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.166049,10957.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,76.0639,606974
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.206692,11056.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.147703,11049
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.158186,11052.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,77.679,626789
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.200996,11183
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.171268,11186.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,76.4922,619582
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.20427,11230.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.150676,11228.9
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.154174,11236.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,76.9268,630270
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.201924,11351.1
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.168199,11347.3
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,76.2617,627009
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.208318,11391.3
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.153293,11387.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.157697,11391.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,28.9703,236084
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.176138,10897.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.159335,10903.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,38.7976,324440
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.175629,11521.3
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,17.5909,149617
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.152413,11534.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.113466,11540.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.0900132,11554.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,38.0111,326607
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.171316,11555.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.162138,11557.7
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,39.5475,335374
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.184955,11547.9
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.152225,10964.3
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.156791,10972
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,39.399,334126
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.184081,11584.3
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.168109,11603.5
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,39.2594,331374
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.191911,11584.4
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.145463,11603.5
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.163089,11616.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,15.4498,130494
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.156122,11636.1
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.144304,11636.1
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,20.2917,176364
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.154055,11641.8
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,10.3201,93084.5
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.136158,11622.7
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0715077,11632.3
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0745285,11645.7
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,20.129,176475
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.148567,11626.7
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.138865,11632.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,20.7945,182563
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.155272,11615.2
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.125051,11622.9
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.14018,11634.4
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,20.7104,179669
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.152506,11619.1
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.139943,11621
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,20.8592,183779
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.15529,11626.6
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.124215,11618.9
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.139051,11618.9
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.455749,11628.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.236337,11630.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.0688514,11618.9
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.27419,12648.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_155.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_155.txt
deleted file mode 100644
index 516dfb5b2e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_155.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,46.4951,363735
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.196449,11194.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.166686,11194
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,95.1914,775570
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.208955,11343.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.169508,11355.3
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,96.8453,795969
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.214564,11434
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.158301,11439.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.162797,11449.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,98.0776,814828
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.208065,11620.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.165489,11616.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,97.0066,811310
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.202301,11643.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.144353,11645.5
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.158458,11647.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,97.9465,823840
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.204615,11737
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.172193,11742.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,97.1248,825100
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.207636,11784.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.151512,11771.5
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.156084,11779.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,35.1464,299384
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.188359,11908.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.173533,11936.7
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,48.9689,428331
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.185953,11953.9
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,19.9562,173532
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.158212,11960.9
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.12786,11966.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.115277,11970.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,47.902,423740
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.185159,11995.5
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.161834,11993.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,49.5207,432998
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.184455,11988.5
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.143089,11984.6
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.165818,11991.7
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,49.5391,434498
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.183556,12037.6
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.171566,12043.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,49.448,433688
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.187249,12032.6
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.142932,12028.6
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.163277,12040
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,17.7985,157549
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.165853,12054.3
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.145255,12048.7
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,24.4838,219883
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.160673,12032.2
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.0361,99109.8
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.132864,12024.5
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0875301,12022.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.081645,12031.9
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,24.3331,219919
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.155952,12019
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.142647,12015
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,25.0519,225088
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.161057,12022.8
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.123367,12024.7
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.147978,12034.2
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,25.069,224657
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.158282,12017.1
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.140254,12028.5
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,25.0504,223780
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.162653,12019.4
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.127204,12024.9
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.139469,12026.8
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.496319,12032.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.290376,12030.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.103578,12030.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.28758,12612.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_156.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_156.txt
deleted file mode 100644
index 1f5bb7ca44..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_156.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,46.3793,360478
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.195889,11276.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.170801,11274.7
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,94.6525,782990
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.206669,11357.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.164996,11364.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,96.5177,794718
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.214778,11485.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.15064,11489.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.15745,11491.1
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,97.7861,812718
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.207312,11624.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.168129,11622.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,96.797,812690
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.209047,11649.3
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.147086,11657
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.155418,11660.7
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,97.7592,828084
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.203898,11772.9
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.168977,11782.5
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,96.6002,821463
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.202871,11824.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.148881,11828.3
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.155297,11828.3
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,36.5045,309108
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.188538,11917.5
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.16497,11915.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,51.893,453452
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.193611,11943.9
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,20.3151,177282
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.164496,11978.1
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.128493,11976.1
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.109293,11983.9
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,50.5423,448585
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.184442,12021.5
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.162596,12031.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,52.2169,457988
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.189431,12033.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.141422,12035.5
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.161936,12048.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,52.6134,461346
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.189623,12077
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.165716,12065.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,52.2984,459593
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.191367,12054.6
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.146541,12064.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.501346,13860.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,17.9254,157814
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.162205,12078.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.140787,12082.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,24.0885,217926
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.15657,12086.5
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,10.9532,97077.8
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.133943,12069.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0983812,12080.8
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.081053,12084.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,23.9958,216954
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.153847,12098
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.140803,12094.2
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,24.8268,219949
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.162945,12080.9
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.125815,12088.5
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.139783,12092.3
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,24.6496,219469
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.156685,12080.8
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.139712,12077
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,24.6868,219414
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.159601,12081.1
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.123994,12080.9
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.141802,12080.9
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.482347,12082.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.288257,12080.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.10329,12080.9
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.32124,14976.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_157.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_157.txt
deleted file mode 100644
index 17b884ece4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_157.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,47.1371,365002
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.200926,11232.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.168077,11238.3
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,97.9776,806133
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.210599,11360.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.17601,11366.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,100.175,818140
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.211361,11449.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.148887,11449.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.155687,11456.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,101.313,837607
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.211018,11595.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.170058,11597.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,100.235,837138
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.208558,11626.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.147578,11628.5
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.152945,11634.1
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,101.425,852459
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.217012,11771.1
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.173047,11765.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,100.636,845628
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.209892,11780.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.149863,11765.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.155021,11771
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,36.4794,307534
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.188935,11887.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.16512,11892.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,52.208,452912
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.195443,11894.1
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,20.4608,177813
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.167232,11907.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.133386,11921.1
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.118429,11932.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,50.7088,447510
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.185597,11964.7
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.158593,11966.5
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,52.4322,458226
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.192458,11970.3
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.144791,11964.6
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.161889,11985.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,52.4391,460381
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.193543,12006.7
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.169569,12018.1
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,52.3099,457756
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.191437,12012.3
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.144557,12014.3
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.154932,12016.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,19.359,168131
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.170564,12050.5
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.14258,12050.5
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,27.2208,241963
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.163844,12020.1
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.4742,103374
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.147642,12031.5
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0893282,12033.4
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0800389,12035.3
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,26.8358,242244
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.167018,12031.5
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.146839,12025.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,28.0046,246969
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.169105,12048.8
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.125393,12037.3
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.143005,12039
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,27.8585,246907
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.16721,12047
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.140291,12023.9
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,27.9707,246224
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.17083,12027.7
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.123367,12029.6
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.150414,12044.8
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.500584,12061.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.323294,12046.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.120858,12043.1
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.38314,14412
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_158.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_158.txt
deleted file mode 100644
index e98ad1dac3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_158.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,45.3649,351246
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.198077,10842.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.167095,10848.1
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,93.4682,742975
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.212638,10985.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.16466,10989.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,95.4044,758756
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.21259,11089.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.152186,11089.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.156686,11100.7
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,97.06,778990
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.207981,11202.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.169482,11213.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,95.9929,776887
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.207693,11238.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.145626,11234.7
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.157671,11248.1
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,96.7122,786602
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.208877,11416.3
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.171114,11410.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,96.0349,778312
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.208868,11437
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.151575,11444.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.15857,11450.5
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,34.5979,281611
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.189521,11522
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.168845,11529.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,48.8281,413618
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.185639,10966.5
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,19.5382,161047
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.162221,11590.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.127472,11595.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.112388,11599.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,47.7494,411233
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.179268,11640.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.164192,11644
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,49.3253,412576
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.195249,11630.5
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.143933,11640.1
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.158637,11653.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,49.4295,410691
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.189815,11071.5
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.157168,11077.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,49.9413,408263
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.189271,11052.2
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.143197,11063.7
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.162183,11069.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,18.1605,150442
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.16536,11685.7
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.142289,11678
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,24.9233,214438
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.16411,11672.1
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.1513,97298.5
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.164996,11071.3
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0814528,11065.5
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0784673,11065.5
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,24.6982,214926
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.167987,11666.3
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.143037,11660.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,25.4734,221295
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.173341,11668.2
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.137431,11670.1
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.149933,11673.9
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,25.3427,218774
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.16626,11658.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.148352,11670.1
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,25.3068,217734
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.16464,11670.1
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.129645,11660.5
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.146871,11671.9
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.512827,11679.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.299771,11679.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.110944,11664.3
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.33906,14567.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_159.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_159.txt
deleted file mode 100644
index 84b449752c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_159.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,44.4202,346297
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.196084,10939.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.172138,10943.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,92.4501,740633
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.205501,11047.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.165642,11050.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,94.4691,750930
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.204541,11179.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.149248,11183.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.161351,11187
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,95.6935,771756
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.515864,13069.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.167431,10736.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,95.9264,767258
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.202272,11335.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.157895,11339.7
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.158212,11343.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,95.6623,784604
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.204113,11469.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.17234,11468.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,94.4249,774456
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.205649,11513.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.148695,11509.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.154689,11519.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,35.5736,295192
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.186474,11631.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.166993,11629.3
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,51.0173,432711
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.185281,11671.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,19.5907,165968
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.159095,11677.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.134301,11696.3
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.116272,11698
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,49.8657,433978
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.180228,11723.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.160599,11723.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,51.4392,439480
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.190029,11707.1
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.13889,11707.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.158903,11709.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,51.5985,443420
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.188548,11756.9
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.160647,11759.1
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,51.7089,441483
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.191761,11762.4
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.146625,11768.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.165668,11777.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,17.7843,154158
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.162772,11789.7
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.146276,11797.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,24.4869,215315
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.161156,11785.5
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.012,93090.3
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.141012,11764.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0942946,11768.2
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0804673,11779.9
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,24.2635,213766
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.155303,11783.5
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.138612,11775.9
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,25.1662,222125
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.163773,11760.4
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.12633,11764.3
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.140624,11781.5
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,25.0422,218318
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.160689,11771.8
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.139709,11768.1
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,25.1088,220040
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.159495,11760.4
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.128381,11762.3
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.142407,11766.1
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.489137,11768
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.280977,11762.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.0977411,11743.1
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.30064,13360.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_160.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_160.txt
deleted file mode 100644
index ba5a39afb5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_160.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,44.8557,349989
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.201735,10935.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.161296,10937.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,95.4328,764135
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.209043,11020.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.164733,11026
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,97.4355,765626
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.210622,11144.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.152538,11133.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.155821,11148.9
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,98.9292,797493
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.205998,11291.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.167424,11280.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,98.3052,780351
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.211626,11320.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.152692,11322.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.154954,11330.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,98.8454,809359
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.207902,11445.1
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.166164,11443.3
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,97.9064,789390
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.215041,11473.7
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.154052,11477.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.153735,11477.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,35.7972,295329
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.192909,11571.9
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.166784,11570
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,51.1848,432395
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.186804,11598.7
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,19.5652,163840
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.163956,11619.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.133972,11623.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.112468,11644.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,50.2553,433771
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.187722,11669.5
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.164503,11679.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,51.7372,440714
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.19282,11665.4
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.144609,11673.3
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.165991,11677.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,51.8961,438430
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.194461,11682.6
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.168932,11694.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,51.7607,439705
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.189213,11693.8
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.152148,11699.7
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.16537,11709.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,19.5476,167143
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.169943,11721
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.148192,11721
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,27.3335,237228
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.163809,11695.9
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.638,104706
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.138474,11688.3
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0751105,11690.2
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0770753,11699.8
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,26.7729,236639
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.16017,11726.2
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.140992,11720.5
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,27.803,238869
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.17377,11720.3
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.128241,11720.5
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.149443,11720.5
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,27.7809,238580
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.169303,11731.8
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.142506,11728
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,27.8468,239035
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.166442,11716.5
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.125236,11727.9
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.140637,11731.8
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.507262,11737.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.287022,11733.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.118522,11714.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.35282,14060.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_161.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_161.txt
deleted file mode 100644
index 84970f1fe4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_161.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,48.6326,382441
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.202526,11056.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.170983,11054.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,103.385,825432
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.21745,11179
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.164506,11184.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,105.014,844958
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.217178,11311
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.148206,11312.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.151754,11316.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,107.129,873656
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.2093,11441
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.174301,11442.9
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,106.188,861404
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.208752,11481.3
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.156423,11486.9
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.157261,11496.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,107.143,878038
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.207716,11599.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.169047,11605.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,106.132,879328
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.20818,11635.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.151335,11641.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.155604,11647.3
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,38.1243,314346
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.190292,11735.1
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.168061,11737.1
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,55.4229,473356
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.195706,11744.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,21.316,182671
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.165684,11793.9
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.13834,11791.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.129319,11803.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,53.8285,470884
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.188733,11827.9
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.167789,11827.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,55.8302,479591
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.190257,11810.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.150512,11818.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.163104,11833.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,55.5428,478772
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.188622,11847.3
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.164317,11854.9
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,55.4859,475940
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.19601,11254
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.143582,11248.5
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.158666,11261.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,19.477,164523
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.173245,11868.2
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.142557,11881.5
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,27.3618,236204
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.167892,11286.6
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.5708,104259
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.140826,11866.6
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0799075,11866.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0791585,11870.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,27.0259,240364
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.163123,11881.9
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.139795,11881.7
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,27.9777,245697
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.167933,11880
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.127971,11883.8
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.14249,11889.3
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,27.9434,244459
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.163664,11870.3
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.141876,11881.7
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,28.0524,244842
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.168151,11870.8
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.122461,11880.4
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.140733,11889.6
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.508731,11895.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.29009,11895.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.120023,11887.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.34269,12462
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_162.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_162.txt
deleted file mode 100644
index 2f1f3b1e54..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_162.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,49.2141,390239
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.205447,11101
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.172835,11087.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,103.701,844682
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.21776,11211.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.170279,11209.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,106.573,862929
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.216039,11318.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.151764,11316.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.1577,11318.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,107.64,882594
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.217053,11454.5
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.170574,11456.5
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,106.41,874664
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.217415,11485.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.158746,11488.9
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.168378,10907.1
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,107.563,890736
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.252208,11601.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.171156,11594.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,106.21,878230
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.214292,11645.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.160173,11640.1
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.15826,11642
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,38.645,322631
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.195962,11760
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.168909,11768
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,55.2585,475406
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.194413,11767.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,21.1175,183638
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.169248,11809.5
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.129863,11809.5
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.120483,11813.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,53.7737,472739
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.187588,11849.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.164023,11858.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,55.7519,480424
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.195216,11850
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.145911,11817.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.166791,11842.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,56.1452,484693
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.194375,11876.6
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.172128,11891.9
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,55.7553,474578
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.202647,11863.4
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.149229,11874.8
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.168653,11878.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,19.7031,170962
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.174887,11891.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.144996,11891.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,27.6268,243927
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.171047,11914.6
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.6937,106426
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.14586,11893.7
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0834371,11893.7
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0801792,11893.7
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,27.1345,242994
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.168564,11888
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.145725,11895.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,28.2077,246745
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.17568,11876.7
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.12961,11876.7
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.145057,11891.9
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,28.2379,249840
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.169825,11897.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.147549,11897.7
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,28.207,239408
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.176106,11895.8
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.128208,11895.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.148234,11901.6
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.514369,11903.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.289118,11905.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.120292,11895.8
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.39271,13557.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_163.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_163.txt
deleted file mode 100644
index cd96db7497..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_163.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,48.917,387530
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.201447,11074.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.164775,11077.9
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,103.775,838117
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.21514,11215.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.163821,11219.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,106.357,854075
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.21345,11305.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.149136,11299.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.155927,11315
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,107.042,875508
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.206573,11445
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.183095,11450.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,106.307,872832
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.21002,11475.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.152874,11481.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.155399,11492.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,107.039,887648
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.206352,11595.9
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.169418,11603.5
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,106.703,888928
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.211642,11640
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.158547,11643.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.15465,11651.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,38.4583,320287
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.194628,11740.9
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.170791,11746.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,55.2686,473661
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.194435,11782.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,21.2744,182350
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.167949,11824.5
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.13272,11822.7
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.122724,11834.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,54.1731,472155
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.186141,11840.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.16338,11845.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,55.9995,480723
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.196103,11834.5
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.145139,11826.7
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.164205,11832.3
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,55.9716,481882
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.196506,11864.9
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.171933,11862.7
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,55.5609,476963
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.196423,11840.5
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.139981,11851.7
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.158179,11855.5
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,19.7503,171185
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.168394,11868.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.144119,11879.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,27.5084,242291
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.163031,11878.3
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.6804,106287
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.145466,11889.8
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0808514,11884
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0774689,11887.8
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,27.0126,241257
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.162461,11884
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.141376,11889.9
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,28.0699,244691
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.165706,11872.9
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.130356,11880.5
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.151485,11884.1
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,28.4861,249320
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.175658,11895.7
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.140631,11880.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,28.2804,248130
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.166903,11886.1
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.131748,11880.5
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.147271,11880.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.510679,11880.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.290846,11880.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.123501,11865.1
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.35638,13059.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_164.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_164.txt
deleted file mode 100644
index 7f3da6f1d9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_164.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,48.8895,384754
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.198016,11106.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.174464,11106.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,102.8,833768
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.223373,11230.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.169821,11242.1
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,104.198,843830
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.211972,11341.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.146963,11347.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.153901,11351.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,105.549,860690
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.213869,11489
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.170711,11492.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,104.681,855337
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.210925,11517.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.146103,11523.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.154807,11529
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,105.782,880310
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.210506,11636.1
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.167152,11639.9
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,105.245,881034
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.210071,11663
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.152551,11674.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.163059,11678.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,38.0209,319075
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.193184,11774.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.165699,11774.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,54.6525,467069
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.193767,11789.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,20.9009,179117
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.167319,11839.1
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.141549,11848.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.135396,11850.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,53.2668,464535
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.187693,11847.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.163575,11861.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,54.9499,472197
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.195971,11851.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.154116,11851.5
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.165395,11858.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,55.38,476665
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.196314,11866.8
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.165891,11872.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,54.8451,470160
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.191667,11857.4
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.143427,11863
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.162717,11876.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,19.4799,168624
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.170973,11899.1
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.147757,11895.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,27.4145,241547
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.167955,11889.7
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.5731,105172
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.146445,11891.6
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0832931,11912.4
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0781089,11914.3
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,26.8719,239354
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.16264,11916.6
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.143469,11912.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,27.7771,242174
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.16807,11914.8
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.126004,11922.4
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.13968,11924.2
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,27.8599,244873
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.174852,11924.4
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.145005,11928
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,27.8445,244668
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.167303,11897.9
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.124483,11892.2
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.14384,11903.5
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.509579,11926.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.290983,11913.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.121802,11894
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.35496,14169.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_165.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_165.txt
deleted file mode 100644
index f535f820c9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_165.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,46.7706,366332
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.196864,10897.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.165715,10884.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,101.034,805442
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.214007,11024.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.166954,11027.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,103.116,814515
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.209031,11118
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.155533,11123.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.155629,11129.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,104.52,841564
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.20441,11250
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.167997,11257.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,103.536,830087
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.207524,11273
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.149543,11278.7
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.156848,11278.7
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,104.167,851167
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.207072,11441.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.168394,11434
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,103.355,848740
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.206832,11443
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.157267,11452.9
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.154877,10875.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,37.4778,304343
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.189639,11571.9
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.164474,11564.3
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,54.1309,457152
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.193898,11579.5
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,20.3232,171197
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.162643,11587.3
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.136397,11598.7
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.127139,11604.5
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,53.0287,456900
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.185508,11644.7
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.165882,11646.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,54.6935,462937
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.191965,11628.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.147143,11636.7
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.164816,11660
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,55.0262,469640
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.189905,11670.8
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.165431,11676.9
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,55.213,469002
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.191364,11658.9
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.145229,11672.5
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.162064,11680.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,19.2548,166311
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.166749,11698.1
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.142829,11698.1
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,27.3144,239688
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.16433,11689.8
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.4811,99846.8
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.142627,11691.9
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0968995,11688.1
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0832864,11697.7
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,26.8172,235710
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.15953,11691.6
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.142778,11683.9
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,28.0084,239588
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.167965,11680
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.126631,11681.9
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.143629,11681.9
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,28.4503,241887
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.168442,11685.7
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.150135,11695.3
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,28.0136,240097
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.16489,11695.1
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.129965,11689.4
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.146049,11693.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.512756,11702.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.28954,11704.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.123575,11680
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.34312,12874.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_166.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_166.txt
deleted file mode 100644
index 4568dca5cc..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_166.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,46.5866,363696
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.196045,10861.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.167866,10869.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,101.518,809381
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.208125,10968.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.166212,10960.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,103.489,815567
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.215581,10515.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.149162,10524.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.155725,11108.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,104.95,841638
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.204362,11215.5
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.164708,11217.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,103.941,825927
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.207079,11248
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.148653,11248
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.159725,11253.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,105.078,854052
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.208765,11374.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.168138,11380
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,103.767,849143
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.210032,11404.7
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.148023,11391.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.165571,11399.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,37.4473,304979
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.19322,11518.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.175434,11541.7
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,54.5451,451091
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.195261,11543.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,20.2246,169907
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.162151,11549
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.13777,11552.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.12792,11566.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,53.0257,453740
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.183991,11579.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.166506,11592.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,54.6933,460654
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.192186,11586.5
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.153712,11577
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.162308,11594.3
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,55.0575,465877
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.188503,11620.8
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.163811,11634.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,54.849,462686
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.204007,11616.7
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.146029,11630.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.161895,11634
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,19.4899,166249
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.1684,11663.2
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.146861,11670.9
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,27.4326,237190
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.162688,11632.1
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.5929,102938
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.139847,11645.6
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0863104,11641.8
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0836291,11653.3
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,27.0489,236606
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.164016,11658.8
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.142861,11664.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,27.9672,237033
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.162813,11660.6
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.125462,11660.6
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.141907,11674.1
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,28.0632,241426
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.165491,11679.8
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.13993,11666.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,28.1957,242836
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.167428,11658.8
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.122228,11651.2
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.143012,11662.6
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.508704,11685.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.326176,11672.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.122052,11641.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.34069,12221.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_167.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_167.txt
deleted file mode 100644
index 3a5e9b50bc..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_167.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,46.0976,356388
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.196816,10880.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.17264,10882.9
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,100.932,797135
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.206067,10978.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.164071,10980.3
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,102.59,802668
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.207629,11096.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.149181,11096.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.149274,11096.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,104.309,830819
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.209415,11206.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.180269,11211.9
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,102.598,826477
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.211831,11248.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.152788,11250.1
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.156989,11259.7
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,103.745,844893
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.207379,11399.7
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.169136,11405.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,103.469,840953
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.209504,11410.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.155722,11416.9
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.152909,11424.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,36.7554,300738
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.1845,11522.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.162394,11529.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,53.4419,449755
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.19576,11566.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,19.8333,167843
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.163091,11575.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.137405,11591.1
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.13127,11591.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,52.351,443942
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.191978,11616.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.163703,11612.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,54.1502,455838
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.191738,11604.1
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.141879,11606.3
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.163155,11631.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,54.4963,462300
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.189389,11629
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.16391,11636.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,54.407,460769
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.194861,11607.6
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.151738,11607.7
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.156486,11629
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,19.6738,166906
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.171543,11656
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.149437,11661.9
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,27.475,237689
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.166791,11661.4
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.588,101903
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.146013,11651.8
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.089053,11651.8
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0794466,11663.3
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,26.8633,236085
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.158327,11659.1
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.139853,11666.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,27.7566,236799
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.165412,11661
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.12457,11664.9
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.140538,11666.9
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,27.8764,239796
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.166035,11674.3
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.139789,11668.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,28.0507,241308
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.187427,11651.1
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.12649,11658.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.146099,11672.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.508394,11676.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.287876,11668.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.123875,11651.5
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.35771,13408.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_168.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_168.txt
deleted file mode 100644
index 80dc433310..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_168.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,45.5094,352968
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.200602,10879
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.162103,10875.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,100.172,798998
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.212454,11014.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.168288,11022
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,102.026,811658
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.218547,11117.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.152218,11112.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.158416,11119.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,103.418,826858
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.211335,11286.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.173642,11292.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,102.951,828657
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.214737,11309.7
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.156819,11315.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.152829,11317.3
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,103.667,834186
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.211239,11415.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.178768,11419.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,102.79,843458
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.209095,11456.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.151837,11447.7
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.155104,11449.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,36.8962,303297
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.188413,11556.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.16584,11566.3
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,53.9649,453644
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.191456,11568.3
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,19.9096,167004
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.164685,11595
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.138442,11600.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.131709,11614.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,52.4399,447907
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.188679,11637
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.163651,11648.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,54.0805,454884
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.192717,11621.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.147584,11633.1
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.166903,11654.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,54.4074,445952
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.191408,11679.1
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.16754,11681
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,54.3184,458929
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.195213,11667.1
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.144723,11671.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.160759,11688.5
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,18.9063,159096
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.167501,11700
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.142128,11698.1
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,26.87,234233
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.173047,11690.5
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,11.0846,96393
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.140516,11675.3
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0902083,11675.3
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0801635,11675.3
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,26.5137,232066
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.163696,11701.7
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.137037,11703.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,27.3732,237338
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.167344,11685.9
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.125175,11682.3
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.148179,11699.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,27.2751,235005
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.164455,11705.1
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.146282,11689.8
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,27.4971,235337
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.170714,11693.4
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.12768,11695.4
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.144275,11697.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.499677,11699.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.290938,11699.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.124067,11691.7
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.37835,14053.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_261.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_261.txt
deleted file mode 100644
index 6caeadddfd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_261.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,30.4317,249183
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.18295,11714.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.171367,11693.1
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,95.1785,823199
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.193626,11863.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.169674,11865.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,96.08,823526
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.20386,11990.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.14537,12002.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.161354,12013.9
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,97.1222,856045
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.197712,12142.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.167699,12152.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,96.027,851055
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.202531,12241.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.154144,12243.9
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.157949,12242.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,97.1949,870765
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.196445,12383.3
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.1645,12394.7
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,95.9803,870248
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.194339,12472.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.149552,12467.1
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.167159,12465.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,28.1835,256585
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.176781,12536
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.16575,12547.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,47.2133,439407
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.171853,12570.7
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,9.97099,92791
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.149904,12549.5
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.116605,12551.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.124522,12574.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,46.3887,437112
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.165469,12587.9
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.167197,12584
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,47.7867,439935
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.172295,12593.3
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.142474,12601.1
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.16168,12612.7
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,47.976,445621
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.176023,12635.4
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.160746,12641.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,47.7349,437372
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.173604,12637.1
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.144672,12650.6
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.168858,12662.1
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,15.8447,148487
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.148803,12668
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.13809,12669.9
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,25.8494,247000
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.143344,12671.5
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,6.35713,63182.7
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.120749,12669.8
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0672673,12662.2
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0744546,12675.6
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,25.6651,247505
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.146397,12652.3
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.137776,12656.2
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,26.3412,249413
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.15121,12658.1
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.123315,12660
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.144339,12673.3
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,26.2191,247889
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.146538,12677
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.139453,12673.3
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,26.6016,243341
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.151043,12039.8
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.121338,12047.4
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.138048,12049.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.418199,12043.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.251808,12032.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.0764609,12018.9
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.24904,12640.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_262.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_262.txt
deleted file mode 100644
index 9194ee12f9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_262.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,31.084,264062
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.18576,11814.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.166448,11808.3
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,94.7327,823938
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.189786,11928.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.168368,11931.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,95.6342,834776
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.187223,12083.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.141821,12091.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.154775,12097.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,96.4715,854397
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.18544,12251.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.166947,12230.9
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,95.5344,853289
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.186301,12325.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.143562,12323.7
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.155834,12335.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,96.6132,871318
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.186845,12457.3
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.168592,12465.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,95.5462,858667
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.184093,12558.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.146013,12554.9
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.157898,12560.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,27.9198,253552
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.168058,12627.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.163808,12623.7
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,46.8795,437433
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.163354,12648.9
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,9.84177,92534.6
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.143799,12629.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.120464,12642.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.128582,12660.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,46.1099,434555
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.159117,12666.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.163632,12668
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,47.4097,442133
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.168586,12673.7
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.143165,12681.3
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.164135,12690.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,47.5063,444863
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.165943,12706.1
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.160487,12711.9
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,47.4321,444154
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.166349,12700.2
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.141156,12707.9
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.159408,12725.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,15.5918,147182
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.143619,12715.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.141895,12723.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,25.6326,245643
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.140343,12723.3
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,6.12304,64077.9
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.117616,12711.8
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0659712,12711.8
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0825504,12719.5
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,25.4628,244787
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.137773,12717.4
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.137437,12719.3
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,26.1592,247632
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.14576,12757.5
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.120461,12749.8
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.138835,12761.3
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,26.1382,248720
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.139706,12764.8
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.137284,12768.8
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,26.3031,252754
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.150208,12779.7
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.124381,12785.6
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.141296,12787.6
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.386602,12791.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.245216,12772.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.0738144,12757
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,0.821626,12757
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_263.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_263.txt
deleted file mode 100644
index c5025fd869..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_263.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,39.5327,347153
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.183283,12189.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.165594,12191.1
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,137.109,1.23506e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.19592,12410.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.165744,12391.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,138.873,1.25827e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.201104,12559.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.152167,12555.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.154701,12565.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,139.762,1.27147e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.202093,12732.5
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.165715,12728.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,138.636,1.28659e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.194122,12806.3
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.144246,12787.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.156218,12791.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,140.121,1.29954e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.196861,12945.9
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.164419,12951.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,138.973,1.29785e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.196806,12991.5
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.147357,12995.5
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.155949,13005.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,39.2133,367153
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.172595,13077.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.158445,13085
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,68.5708,663113
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.180727,13114
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,11.2781,109844
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.150797,13071.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.135869,13079.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.16025,13083.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,68.0671,658180
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.181405,13125.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.161194,13127.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,69.0828,665526
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.177923,13125.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.147206,13123.6
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.164227,13140.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,69.3119,669416
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.177683,13158
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.1604,13152.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,69.2262,669360
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.178531,13183
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.142541,13184.9
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.157946,13188.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,21.6755,211362
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.15368,13209.8
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.14431,13204
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,37.4536,361340
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.150858,13221.2
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,7.03754,70462.1
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.126621,13179.2
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0792961,13171.5
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0759552,13177.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,37.1024,369906
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.148682,13207.8
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.137706,13209.7
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,38.1749,373664
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.15682,13213.5
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.128525,13217.4
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.146595,13219.3
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,38.2556,375394
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.168535,13217.2
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.1396,13221.2
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,38.1804,375556
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.152038,13222.5
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.13,13211.1
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.149703,13217.1
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.477028,13217.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.304611,13219.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.112051,13219.1
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.30261,14525.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_264.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_264.txt
deleted file mode 100644
index 951b8c465b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_264.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,39.323,345738
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.182237,12241.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.167891,12237.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,137.261,1.23798e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.197783,12424
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.166653,12432
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,138.854,1.25972e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.197824,12569.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.157556,12557.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.155968,12569.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,139.918,1.28635e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.19745,12736.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.162301,12743.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,139.255,1.29224e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.199748,12852.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.146711,12846.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.155687,12854.3
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,140.069,1.31035e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.194759,12990.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.165734,12980.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,138.89,1.31491e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.194778,13026.3
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.152538,13005.3
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.152458,13016.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,39.0674,366522
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.173783,13125.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.165751,13119.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,68.6855,665142
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.173549,13175.5
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,11.2636,109301
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.151315,13110
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.13929,13119.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.156579,13127.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,68.1344,662735
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.171584,13175.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.160071,13160.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,69.2405,667781
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.183485,13162
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.143507,13156.2
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.158371,13160
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,69.5357,671279
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.178115,13188.6
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.167236,13179.1
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,69.1784,667827
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.179719,13190.6
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.145462,13190.6
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.163782,12523
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,21.7289,208856
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.154314,13200.1
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.139331,13223.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,37.6332,372906
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.157072,13246.2
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,7.01964,68656
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.134282,13204
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0863424,13196.3
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0791584,13200.1
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,37.1771,370294
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.149443,13242.4
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.136618,13229
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,38.075,372443
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.15465,13246.1
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.133415,13246.1
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.144989,13253.9
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,38.3939,374949
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.161687,13234.5
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.143827,13232.7
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,38.1879,373605
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.153891,13264.9
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.124851,13270.9
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.143171,13272.8
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.457245,13272.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.287354,13267
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.114048,13236.4
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.31037,14472.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_265.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_265.txt
deleted file mode 100644
index 49017013dc..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_265.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,38.7707,340586
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.17944,12187.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.162285,12192.9
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,136.423,1.22811e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.197392,12401.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.168198,12414.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,138.429,1.22523e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.201869,12598.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.153661,12603.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.157661,12592.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,139.723,1.28525e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.196979,12776.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.177027,12778.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,138.335,1.28467e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.194768,12842.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.15399,12846.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.165734,12848.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,139.253,1.30786e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.196186,13013.1
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.170624,13005.5
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,138.422,1.30939e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.197789,13048.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.151229,13052.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.154845,13058.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,39.2862,367799
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.175018,13146.7
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.167898,13137
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,68.7495,664342
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.179974,13171.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,11.558,110251
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.148966,13139
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.143213,13125.5
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.152838,13138.9
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,68.1173,661585
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.174605,13156.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.161184,13160.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,69.2829,669275
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.184016,13167.9
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.154615,13166
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.164211,13173.7
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,69.504,671243
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.180176,13192.8
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.167926,12536.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,69.2877,663966
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.181718,13198.5
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.146775,13200.4
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.161718,13208
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,21.7855,210477
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.155574,13217.4
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.143063,13221.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,37.5259,372285
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.157907,13231
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,7.10395,69331.8
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.130576,13206
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0816704,13207.9
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0799008,13211.7
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,37.0512,370145
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.151184,13215.6
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.143306,13217.5
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,37.9731,370874
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.158128,13209.8
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.125568,13213.6
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.146739,13217.5
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,37.9281,371303
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.155536,13209.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.146294,13215.5
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,37.974,372004
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.15208,13215
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.129706,13207.3
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.14801,13211.3
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.446621,13221.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.288829,13225.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.118842,13217.4
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.30021,13879.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_266.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_266.txt
deleted file mode 100644
index 595da5c025..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_266.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,35.8886,295686
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.189232,11672.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.173853,11670.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,122.434,1.04701e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.202077,11838.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.176003,11839.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,125.275,1.04456e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.203037,11957.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.152624,11953.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.163437,11959.7
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,124.858,1.08686e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.204131,12140
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.180992,12147.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,123.823,1.08755e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.201479,12168.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.149837,12174.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.158503,12184
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,124.805,1.10516e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.202618,12362.9
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.169341,12364.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,123.764,1.10364e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.200304,12399.1
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.162019,12399.1
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.162538,12408.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,35.714,315738
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.179491,12482.7
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.168704,12471.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,61.0415,557179
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.179107,12488.5
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,11.4739,104720
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.151011,12477.1
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.141635,12480.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.159606,12494.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,60.0915,551111
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.174803,12528.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.174106,12526.7
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,61.07,556124
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.185459,12513.4
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.147242,12519.1
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.166086,12526.7
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,61.3889,559492
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.183411,12547.7
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.166381,12553.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,61.3062,561112
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.180656,12540.1
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.151719,12551.6
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.164122,12561
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,19.9997,183106
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.159094,12578
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.14415,12579.9
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,33.9728,317208
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.154083,12589.4
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,7.08242,65922.6
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.132074,12570.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0984769,12566.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0969025,12568.5
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,33.3788,316248
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.153856,12602.7
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.140653,12606.5
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,34.5072,318444
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.159126,12606.3
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.128755,12606.4
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.146691,12623.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,34.5515,321386
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.157776,12638.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.152224,12642.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,34.5678,322899
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.15791,12657.6
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.133059,12650
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.147149,12650
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.471735,12653.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.294282,12653.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.120842,12638.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.34878,14425.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_267.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_267.txt
deleted file mode 100644
index df31c8040e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_267.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,36.5494,304021
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.185594,11614.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.161011,11616.7
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,122.686,1.05018e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.204803,11767.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.172592,11168.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,124.667,1.0649e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.205824,11915.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.151706,11914.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.161552,11905.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,125.785,1.09457e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.20232,12079.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.1696,12083.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,125.033,1.08743e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.201053,12131.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.153542,12120
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.159015,12125.9
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,125.773,1.10955e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.209597,11642.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.164653,11651.7
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,124.837,1.10726e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.206019,12300.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.15695,12298.5
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.160086,12306
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,36.0871,318480
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.184445,12393.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.162842,12389.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,61.4807,560012
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.181046,12424
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,11.7315,107250
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.159431,11771.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.137757,11775.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.160611,11783
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,60.877,553995
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.181501,12424.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.168326,12429.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,61.4704,552925
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.184474,12426.1
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.147853,12433.9
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.175779,12439.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,61.8327,561909
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.183888,12466.1
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.169379,12469.9
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,61.4219,560322
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.18993,12466
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.142829,12464.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.166166,12466.1
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,20.0654,180829
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.160458,12494.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.143213,12496.7
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,33.8782,316751
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.15537,12521.3
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,7.02251,66236.9
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.130816,12485.3
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0892032,12494.8
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0848225,12511.9
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,33.3996,315344
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.150147,12515.5
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.143674,12517.5
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,34.5209,321187
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.158765,12532.5
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.127437,12538.2
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.143302,12560.9
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,34.544,321689
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.159181,12587.1
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.143968,12571.9
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,34.5306,320699
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.157373,12584.7
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.128682,12581.1
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.145571,12585.1
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.469223,12584.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.299597,12577.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.119949,12554.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.36106,13735.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_268.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_268.txt
deleted file mode 100644
index 0fdf2c3e3f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_268.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,35.1239,295538
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.182758,11727.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.164346,11729.5
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,121.773,1.04894e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.197773,11912
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.169786,11917.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,123.171,1.06445e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.20489,12037.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.150778,12039
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.153834,12042.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,124.012,1.08809e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.206176,12212
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.165619,12213.9
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,123.145,1.078e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.198598,12259.7
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.148278,12257.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.161981,12263.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,124.316,1.10887e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.201136,12431.1
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.168058,12419.7
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,123.044,1.1028e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.201034,12454.1
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.14545,12465.5
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.152358,12469.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,35.5114,317122
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.180973,12530.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.165341,12534
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,60.6231,554642
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.185914,12562.7
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,11.3358,102844
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.149994,12555.1
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.138832,12558.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.157642,12562.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,60.2884,557059
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.177584,12545.7
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.166579,12560.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,61.0292,548165
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.184675,12547.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.14408,12549.5
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.167789,12555.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,61.1004,562212
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.181363,12623.5
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.161357,12610.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,60.7832,558608
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.179795,12610.3
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.138381,12606.5
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.165776,12614.1
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,19.9309,182130
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.162442,12635
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.145888,12636.9
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,33.8825,317811
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.156323,12659.7
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,6.96137,64378.3
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.139034,12638.8
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.114899,12642.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0988992,12646.5
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,33.1405,315642
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.153949,12701.5
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.141235,12680.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,34.2446,322151
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.160592,12690.1
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.124125,12684.4
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.143011,12688.2
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,34.4339,323293
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.162896,12686.2
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.141651,12695.8
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,34.2692,322060
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.16088,12697.6
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.127555,12697.7
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.146525,12697.7
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.465504,12699.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.288192,12699.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.120883,12693.9
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.39848,15783.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_269.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_269.txt
deleted file mode 100644
index b7d2ecb6a5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_269.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,35.9376,303218
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.185997,11713.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.16551,11711.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,121.801,1.04827e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.198381,11889.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.168195,11893.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,123.312,1.06282e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.206211,12016.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.15176,12014.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.158547,12020
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,124.478,1.09033e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.198874,12203.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.176634,12189.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,123.559,1.09049e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.200608,12241.3
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.148755,12228
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.157936,12237.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,124.433,1.09011e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.19607,12410.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.170048,12406.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,123.773,1.08023e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.199779,12456
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.151613,12450.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.159293,12454.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,35.6177,317720
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.180666,12516.9
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.16599,12516.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,61.1719,557044
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.180112,12547.5
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,11.3974,104551
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.152074,12532.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.139946,12541.7
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.161942,12551.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,60.5135,550811
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.179066,12568.5
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.167898,12562.7
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,61.1387,556899
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.182877,11941
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.137699,11929.6
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.161021,12576.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,61.2741,562490
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.180557,12585.7
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.167786,12599
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,61.062,561749
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.181517,12598.9
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.144045,12597.1
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.161789,12600.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,20.1235,182024
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.163299,12627.7
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.144845,12631.5
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,33.9154,319167
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.158173,12642.8
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,7.03865,67488.9
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.131427,12608.7
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.0977376,12620.1
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.0964256,12622
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,33.2802,315729
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.155763,12637.2
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.138186,12631.5
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,34.4401,320391
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.160784,12663.7
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.12617,12650.5
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.147306,12658.1
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,34.4106,323358
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.156493,12675
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.143901,12667.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,34.4259,322235
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.158794,12695.8
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.126093,12693.9
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.143821,12703.5
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.459843,12711.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.290442,12686.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.121517,12671.2
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.37308,15102.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp32_perf_fp32_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp32_perf_fp32_120.txt
deleted file mode 100644
index 2323ecc3da..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp32_perf_fp32_120.txt
+++ /dev/null
@@ -1,222 +0,0 @@
-Conv1,88.955,715067
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.224621,11381.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.16934,11387.1
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,241.657,2.03478e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.209395,11869.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.14703,11871.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,244.064,2.05992e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.219187,12109.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Add4,0.188809,12120.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu3,0.188102,12127.5
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,245.988,2.13135e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add5,0.215542,12382.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu4,0.212092,12385.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv5,246.197,2.13653e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add6,0.217254,12354.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Add7,0.202937,12358.1
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu5,0.196607,12358.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,248.37,2.18167e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add8,0.223836,12554.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu6,0.216799,12557
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,254.441,2.24353e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add9,0.224111,12452.1
-Add9_f2h,0,0
-Add9_h2f,0,0
-Add10,0.20151,12453.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.195772,12471.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv8,63.85,549473
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add11,0.205199,12553.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu8,0.198284,12549.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,99.6746,865510
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add12,0.216694,12560.9
-Add12_f2h,0,0
-Add12_h2f,0,0
-Conv10,24.7638,221065
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add13,0.183145,12567.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Add14,0.167753,12568.3
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu9,0.169199,12571.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,99.8741,901134
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add15,0.206124,12536.7
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu10,0.166467,12526.7
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,97.7606,859026
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add16,0.211369,12457.7
-Add16_f2h,0,0
-Add16_h2f,0,0
-Add17,0.175625,12461.5
-Add17_f2h,0,0
-Add17_h2f,0,0
-Relu11,0.173411,12461.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,100.178,878176
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add18,0.211887,12489.9
-Add18_f2h,0,0
-Add18_h2f,0,0
-Relu12,0.189865,12482.5
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,97.2912,856454
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add19,0.21182,12419.1
-Add19_f2h,0,0
-Add19_h2f,0,0
-Add20,0.177296,12421
-Add20_f2h,0,0
-Add20_h2f,0,0
-Relu13,0.177728,12428.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv15,35.0712,307328
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add21,0.194115,12497.7
-Add21_f2h,0,0
-Add21_h2f,0,0
-Relu14,0.168073,12507.1
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv16,55.057,488891
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add22,0.206614,12537.9
-Add22_f2h,0,0
-Add22_h2f,0,0
-Conv17,14.4345,130102
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add23,0.174803,12513.1
-Add23_f2h,0,0
-Add23_h2f,0,0
-Add24,0.146784,12518.8
-Add24_f2h,0,0
-Add24_h2f,0,0
-Relu15,0.173859,12520.7
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,53.9116,481649
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add25,0.197971,12575.8
-Add25_f2h,0,0
-Add25_h2f,0,0
-Relu16,0.161423,12585.3
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,54.2887,491768
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add26,0.200943,12613.9
-Add26_f2h,0,0
-Add26_h2f,0,0
-Add27,0.146553,12615.9
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu17,0.173961,12617.8
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,54.2638,481821
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add28,0.198764,12663.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-Relu18,0.16174,12649.8
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,54.2923,489881
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add29,0.202617,12678.6
-Add29_f2h,0,0
-Add29_h2f,0,0
-Add30,0.14694,12674.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-Relu19,0.192633,12682.1
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Pool1,0.593189,12686
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Mul1,0.537061,12688
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add31,0.124991,12688
-Add31_f2h,0,0
-Add31_h2f,0,0
-Softmax1,1.10314,12678.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_120.txt
deleted file mode 100644
index 76d15fea32..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_120.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,222.258,2.0749e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.442448,26847.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.275467,26862.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.74474,55144.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.81639,35081.2
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,91.0332,928887
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.280939,27160
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.381214,27666
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.193611,27156
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,143.725,1.4505e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.412452,27500.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.485623,28109.8
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.308447,27492.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,115.859,1.15275e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.395543,27672.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.11319,31018.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,121.961,1.24427e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.388675,28054
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.1069,31686.6
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.285963,28035.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.318866,28046.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,160.515,1.67448e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.39484,28514.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.47889,29346.4
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.311915,28534.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,147.854,1.54288e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.415806,28745.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.489366,29611.4
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.304626,28749.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,115.701,1.20405e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.40035,28830.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.12104,33492.2
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.296953,28853.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.335487,28857.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,160.436,1.71285e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.388203,29180
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.467575,30156.4
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.302168,29161
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,147.717,1.58411e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.420554,29365.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.486832,30396.2
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.313963,29362.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,116.062,1.23173e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.41594,29403.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.10304,34761
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.288517,29400
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.308248,29400.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,92.3545,1.00728e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.332414,29511.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.271608,29519.4
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.274206,29515.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,89.6765,985036
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.385354,29629
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.315198,29632.4
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.288933,29632.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,65.2853,714573
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.361253,29618.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.703759,31895.8
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,85.8508,952153
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.343018,29915
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.70612,32257.2
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.27141,29930
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.31845,29918.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,83.7199,938025
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.331915,30140.4
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.31132,30125
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.281701,30128.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,90.0234,1.00289e+06
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.376228,30205.8
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.311915,30190.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.294481,30194.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,64.8775,715057
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.36378,30160.2
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.705289,32583
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.2774,30152.2
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.29829,30148.2
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,83.082,941347
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.325528,30366.4
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.305522,30355
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.295685,30343.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,89.579,1.00933e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.383781,30462
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.312107,30462
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.289656,30450.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,65.679,735392
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.352721,30423.6
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.700783,32971.6
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.266955,30442.6
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.311871,30427.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,83.475,950863
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.322143,30556.8
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.298949,30530.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.284524,30522.4
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,89.5125,1.01373e+06
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.379492,30576
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.312849,30568.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.289771,30579.8
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,65.02,732056
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.357079,30545.2
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.701187,33158.8
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.268953,30526.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.308959,30538
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,44.9954,522675
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.294277,30595.2
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.233733,30580
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.163084,30580
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,55.8593,644514
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.334571,30663.8
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.314469,30622
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.262501,30595.2
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,39.7049,459455
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.316766,30583.6
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.627164,33178
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,54.4462,628014
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.305176,30754
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.640425,33401.8
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.236959,30761.8
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.210693,30765.6
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,47.9702,560717
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.287179,30911.6
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.262834,30896.2
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.207717,30854
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,55.6275,648445
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.334936,30934.2
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.313138,30911.4
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.279863,30914.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,39.3969,458812
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.302949,30817.4
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.596573,33480.8
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.205784,30794.6
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.175743,30767.4
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,48.1546,559611
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.291487,30967.6
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.261042,30956.2
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.217541,30929
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,55.687,650075
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.329124,30924
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.312491,30908.6
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.268952,30908.6
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,39.6102,465238
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.305777,30855
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.59313,33548.2
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.202629,30843.6
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.169305,30851.2
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,48.1046,561871
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.295256,30943.4
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.260997,30920.4
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.215461,30885.8
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,55.5951,647114
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.336408,30936
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.302699,30905.4
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.26341,30897.6
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,39.7209,464708
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.321328,30801.8
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.629953,33487.2
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.209598,30797.8
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.169573,30797.8
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,48.1338,561238
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.292305,30943.6
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.264645,30932
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.216831,30882.4
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,55.6049,650375
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.327671,30936.2
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.272312,30924.8
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.263237,30905.8
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,39.8099,461974
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.309188,30825.4
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.607407,33495.4
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.216894,30802.2
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.177676,30806
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,47.9678,560014
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.28709,30936.2
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.265036,30905.4
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.218297,30913
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,55.7986,645697
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.327218,30932.8
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.272556,30921.2
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.257401,30909.6
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,39.8499,459340
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.311985,30879.2
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.587318,33603.2
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.214584,30852.4
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.174379,30837
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,27.0353,319367
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.249739,30909.6
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.187717,30886.6
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.138041,30898.2
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,43.8386,509940
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.301963,30963.2
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.360465,30971
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.232357,30906
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.2759,311413
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.272785,30883.2
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.624981,34934.6
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,41.4418,496244
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.282853,31000.8
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.612125,35085.6
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.12702,31000.8
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.142144,30997
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,36.6876,435790
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.273567,31100
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.232082,31050.4
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.147743,31050.4
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,43.5691,509513
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.307409,31149.8
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.285048,31149.8
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.237906,31138.2
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.4615,316125
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.272042,31070
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.627721,35140.8
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.140921,31054.8
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.148716,31039.4
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,36.6846,436449
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.282334,31161.6
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.217605,31146.4
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.148044,31127.4
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,43.3849,483902
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.300376,31165.4
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.282494,31169.2
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.234245,31169.2
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.4158,316772
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.278558,31097
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.615042,35152.6
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.124242,31078
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.135564,31078
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.614575,31074.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.25804,37872.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.135737,31074.2
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.86336,121794
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_151.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_151.txt
deleted file mode 100644
index 17423a2571..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_151.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,161.277,1.49722e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.437943,26583
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.245022,26590.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.93526,55943.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.79932,35419.4
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,86.4092,867761
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.221208,26870.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.31276,27399.6
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.153382,26862.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,71.9345,727264
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.384932,27068.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.480458,27639.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.291487,27071.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,116.321,1.16146e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.366949,27352.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.08605,30628.4
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,127.294,1.29861e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.357188,27822
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.09975,31416.6
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.274488,27822
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.303947,27810.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,121.623,1.27096e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.340593,28135.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.460567,28939.6
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.289874,28150.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,74.843,780984
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.389528,28204
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.485681,28230.8
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.30003,28200
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,117.018,1.21288e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.362577,28472.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.11612,32858
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.282622,28468.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.313067,28472.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,121.697,1.29861e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.338628,28756.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.450673,29649
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.294111,28737.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,75.0919,797139
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.388612,28786.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.489814,29752.2
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.306109,28779.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,116.416,1.23036e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.382213,28989.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.1032,33986.8
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.291582,28955.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.299326,28963.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,82.2068,888263
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.299499,29050.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.272868,29054.4
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.17011,29054.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,49.1144,537753
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.337585,29130.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.313495,29126.8
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.290679,29130.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,67.2893,729534
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.322308,29156.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.658723,31266.6
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,80.8107,884974
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.325784,29418.8
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.709072,31605.2
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.26444,29419.4
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.236965,29423.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,63.4299,708703
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.303756,29593.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.297727,29562.4
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.214969,29577.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,48.5833,535714
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.332677,29573.8
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.313304,29573.8
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.275672,29562.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,66.6726,738144
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.322942,29579.2
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.683464,31833.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.234987,29580.2
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.185554,29580.2
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,63.3907,713946
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.318301,29726.8
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.27434,29715.6
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.211729,29715.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,48.7588,541697
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.334642,29715
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.276453,29715
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.290021,29726.8
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,66.9411,746132
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.325509,29750.6
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.619472,32089.4
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.235423,29724.4
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.183282,29728.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,62.9943,713105
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.31013,29867
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.273766,29855.8
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.220639,29848.6
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,48.6576,542380
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.335166,29866.2
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.273124,29855
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.291678,29859
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,66.7553,745351
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.31619,29880.2
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.632329,32280
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.23445,29876.4
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.183461,29872.6
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,39.8456,450007
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.277516,29947.4
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.201535,29909
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.146265,29917
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,33.1384,385340
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.298629,29915
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.266763,29923
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.20604,29919.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,38.2738,443559
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.287417,29918.4
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.511338,32341.4
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,50.0473,571299
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.295723,30120.4
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.535844,32600
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.13381,30120.4
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.154309,30120.4
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,37.8943,432228
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.275147,30148.8
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.216997,30122.2
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.145849,30102.8
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,33.1145,379146
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.305566,30195
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.27557,30164.4
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.23877,30103.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,38.3963,443739
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.293425,30085.6
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.523005,32599.4
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.133797,30089.4
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.145548,30089.4
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,37.005,423317
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.280523,30206.8
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.21717,30195.8
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.149612,30145.8
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,33.089,383848
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.300561,30184.2
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.269573,30177
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.219045,30158
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,38.3927,443904
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.282028,30143.4
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.522372,32688
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.136665,30132
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.140575,30120.4
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,36.8716,423474
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.272529,30203
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.208875,30188.2
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.146035,30161.2
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,33.0235,384188
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.310783,30219.6
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.271762,28690.4
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.209067,30193.4
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,38.615,446069
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.287557,30148.2
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.524836,32719.4
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.134003,30148.2
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.142399,30148.2
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,36.9047,424024
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.279595,30216
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.212831,30223.6
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.147193,30185.8
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,33.0697,384299
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.306758,30224.4
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.275717,30224.4
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.219756,30212.8
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,38.6799,452297
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.298751,30167.6
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.515383,32769.2
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.134469,30167.6
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.142284,30171.4
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,36.7945,424393
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.275973,30277.4
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.208985,30254.6
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.147001,30220.6
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,33.0679,381838
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.307762,30277.4
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.265765,30243
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.228114,30212.4
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,38.4138,444998
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.287371,30159.6
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.52682,32768.8
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.134489,30174.8
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.138335,30174.8
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,22.589,271869
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.205234,30181.2
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.145875,30166.4
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.130707,30152
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,26.7038,309479
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.26252,30133.4
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.196485,30133.4
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.14885,30098.8
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,22.1396,265728
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.217355,30072.2
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.547548,33921
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,34.6527,393681
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.273413,30160.8
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.594492,34032
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.136147,30103
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.147941,30103
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,27.3005,330137
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.202751,30107
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.152057,30107
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.135743,30103.2
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,26.5471,308632
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.257349,30068.8
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.199666,30049.6
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.143832,30022.8
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,22.234,269683
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.219601,29969.4
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.548221,33871.8
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.110355,29965.6
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.142246,29958
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,27.2951,326937
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.213369,29969.4
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.160793,29965.6
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.134348,29954.2
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,26.3678,304418
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.270353,29923.4
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.204568,29908
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.152408,29889
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,22.3208,266934
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.22003,29850.4
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.547914,33695.2
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.110444,29835
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.132806,29823.2
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.592413,29823.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.23316,36239.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.126418,29835
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.87424,118486
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_152.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_152.txt
deleted file mode 100644
index 65b63cbff6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_152.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,162.322,1.52781e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.436158,26779.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.234085,26783.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.78165,53550.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.80536,35479.4
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,86.5925,881289
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.220376,27068.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.310834,27593.6
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.158777,27075.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,74.8456,755086
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.376696,27221.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.486884,27800.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.304804,27236.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,117.01,1.17805e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.36929,27490
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.17302,30754.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,128.337,1.31563e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.360267,27979.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.08126,31593.4
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.261458,27971.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.298366,27956.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,123.913,1.29701e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.342955,28334.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.461482,29089
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.310303,28319
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,77.3257,805207
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.38938,28380.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.466839,29207.6
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.302655,28372.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,117.397,1.22998e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.363492,28601.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.07876,32926.4
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.272082,28617
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.309304,28613.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,123.748,1.32248e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.337784,28885.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.454244,29793.4
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.297989,28874.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,77.7857,827634
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.389624,28950.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.477482,29877.6
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.29832,28939.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,117.627,1.25466e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.365746,29100
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.08466,34027.6
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.264901,29119.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.303711,29100.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,82.4163,905696
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.294437,29248.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.22012,29252
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.159391,29255.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,49.064,541506
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.328504,29274.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.30483,29252
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.279295,29255.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,67.2289,734384
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.325285,29240.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.66744,31377
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,80.3761,885858
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.324055,29542.6
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.667843,31702
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.278809,29547
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.238066,29547
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,63.4284,710517
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.303518,29689
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.295275,29696.6
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.223358,29700.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,48.8434,542391
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.33578,29685
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.310603,29685
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.286219,29696
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,67.1127,747308
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.31882,29655.4
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.679171,31914
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.243461,29694.8
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.181772,29702.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,63.2483,715200
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.307575,29882.8
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.299697,29837
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.216704,29840.8
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,49.5055,552676
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.33708,29854.6
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.30684,28283
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.284548,28287
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,67.5103,754609
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.320107,29856.8
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.657501,32191.6
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.233132,29841.6
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.197964,29834.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,63.3259,718270
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.297201,29988
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.293892,29961.4
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.214258,29969.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,48.991,554526
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.334539,29961
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.312036,29949.6
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.284133,29957.4
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,66.955,750915
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.325503,29945.2
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.665117,32341.4
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.229772,29941.2
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.191116,29933.4
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,39.8798,452314
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.265106,29992
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.226661,29976.8
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.142649,29965.2
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,33.1928,382874
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.295128,29998.8
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.297317,29964.8
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.217862,29949.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,38.4575,446531
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.283333,29903.2
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.544074,32344.8
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,50.2538,574126
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.299044,30069.2
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.570378,32517.6
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.129746,30046.6
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.147564,30047
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,37.1679,424322
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.277617,30164.8
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.253343,30161
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.144773,30126.2
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,33.0095,382176
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.292946,30164.8
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.294923,30153.2
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.224139,30107.4
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,38.5521,446750
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.290744,30104.2
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.571491,32606.2
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.129016,30104.2
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.141541,30104.2
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,36.9386,423284
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.269515,30188
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.233753,30172.6
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.14174,30134.2
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,33.1224,379318
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.299353,30173
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.29516,30165.2
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.226777,30153.8
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,38.3644,441241
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.287377,30162.2
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.592399,32668.6
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.134304,30121.2
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.140185,30132.6
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,37.0016,423906
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.270276,30226.8
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.243237,30230.6
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.145586,30192.2
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,32.9933,379818
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.293957,30231.2
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.294686,30204.4
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.223903,30177.6
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,38.4187,444392
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.281348,30143.8
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.523965,32703.6
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.130419,30140.4
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.140556,30144.2
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,36.7462,423939
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.268798,30231.6
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.201407,30232
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.141273,30220.6
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,32.9878,378675
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.2974,30209.2
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.266462,30197.6
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.230828,30198.2
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,38.3983,447953
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.284005,30198.2
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.513821,32799.4
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.129734,30205.8
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.15061,30190.6
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,36.7534,423099
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.270129,30247.8
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.209489,30251.6
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.143627,30232.4
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,33.2255,389628
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.297746,30244.2
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.264063,30228.8
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.217017,30233
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,38.3565,447951
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.285015,30206.6
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.506621,32819.4
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.128345,30179.8
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.138111,30187.4
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,23.4658,271850
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.241042,30202.4
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.158636,30175.6
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.130854,30164
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,29.5053,339899
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.261996,30156.8
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.228486,30141.4
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.152249,30118.4
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,22.889,272128
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.219589,30103.2
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.54353,33974.4
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,35.5487,403543
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.256638,30222
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.586621,34154.6
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.130585,30218
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.153317,30206.8
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,29.1081,332480
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.243006,30283
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.191417,30283.4
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.134598,30260.6
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,29.3091,337351
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.268472,30230
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.232293,30230
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.153567,30226.2
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,23.0443,274171
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.22309,30165
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.538141,34101.6
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.107986,30165
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.133298,30149.8
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,29.3233,333412
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.242002,30215
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.188844,30199.6
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.140057,30195.6
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,29.7913,342956
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.277387,30169
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.241522,30157.4
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.155206,30123
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,23.0371,272128
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.23109,30115.2
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.546096,34048.6
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.110629,30100
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.129702,30100
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.576169,30100
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.20041,35364.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.195615,30084.6
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.72265,116436
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_153.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_153.txt
deleted file mode 100644
index 5b0c6da1e9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_153.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,154.881,1.39426e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.466174,25471.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.294892,25483.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.66466,61522.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.8784,33553.2
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,81.8855,800623
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.307423,25782
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.325067,25789.6
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.171865,25774.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,72.5312,695061
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.386955,25939
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.462193,26491.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.30316,25938.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,112.118,1.06904e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.365476,26266.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.10847,29393.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,122.337,1.19766e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.364977,26769.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.08946,30267.6
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.276382,26811.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.299351,26788
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,121.033,1.21575e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.338123,27131.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.467229,27878.6
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.29541,27127.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,74.4617,744821
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.398219,27207.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.47084,28000.2
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.302668,27211
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,112.217,1.12816e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.361803,27420
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.09068,31652
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.283467,27423.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.324094,27442.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,121.17,1.24918e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.338988,27749.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.443281,28650.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.291103,27761.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,74.5518,767468
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.383985,27826.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.475492,28760.8
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.307218,27842
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,112.049,1.14599e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.363665,28059
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.08552,32891.6
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.268242,28039.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.30931,28040
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,82.089,867210
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.295851,28165.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.224261,28172.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.159992,28176.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,48.7043,510464
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.331205,28188.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.317279,28203.4
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.2814,28211
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,64.7854,679390
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.31685,28236.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.647459,30292.8
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,78.1458,826825
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.317604,28510
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.644643,30631.6
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.271288,28538
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.245285,28545.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,62.8648,683767
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.294712,28746.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.256907,28738.4
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.221835,28738.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,48.6925,523576
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.330628,28791.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.309124,28799.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.282507,28788.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,64.5534,692488
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.323869,28781.8
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.641878,30986.6
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.254795,28790
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.191698,28790
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,62.7471,687812
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.292856,28890.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.261221,28894.4
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.218284,28894.4
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,48.5235,522385
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.330103,28905
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.299639,28913
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.278494,28924.8
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,64.9505,698029
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.330744,28960.4
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.640349,31245.6
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.254686,28968.4
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.191116,28972.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,62.7637,689156
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.297995,29126.2
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.264517,29134.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.218284,29123
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,48.6746,530442
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.332523,29163.2
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.304709,29156
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.280882,29137.6
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,64.5648,700755
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.320146,29124
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.629795,31516.4
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.253054,29166.2
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.197145,29170
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,39.3902,437784
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.282847,29232.8
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.195135,29217.6
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.151717,29191.2
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,33.2572,374470
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.293092,29228.2
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.263659,29182.6
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.212978,29156
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,37.8866,416740
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.278962,29189
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.548023,31588.8
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,49.2656,548750
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.292216,29341.2
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.551229,31768.2
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.146623,29352.8
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.150764,29352.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,37.4703,419227
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.274719,29436.8
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.212888,29436.8
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.148544,29440.6
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,32.9521,373111
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.300337,29466.8
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.267301,29451.4
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.221157,29409.2
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,38.0559,421559
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.277624,29424.6
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.630025,31889
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.158854,29378.6
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.153465,29382.4
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,36.9282,413650
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.282052,29485.8
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.215326,29489.6
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.148972,29493.6
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,33.0634,375573
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.30092,29563
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.264696,29516.8
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.227653,29466.6
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,37.8484,423841
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.28476,29513.4
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.552183,31966
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.149324,29471
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.152908,29471
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,37.1316,416315
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.283986,29609.2
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.219275,29590
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.147315,29555.6
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,33.253,380507
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.298034,29574.8
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.264511,29563.2
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.216824,29532.4
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,37.9441,425753
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.281144,29521.4
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.544509,32023.6
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.145132,29513.6
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.148178,29498.2
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,36.8519,415022
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.267993,29617.4
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.208248,29602
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.147468,29586.6
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,33.113,376414
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.301438,29579.2
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.328049,29552.2
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.219006,29517.6
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,38.2189,428933
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.282444,29502.8
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.532042,32028.2
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.147807,29495
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.144696,29498.8
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,36.9107,413669
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.284325,29602.2
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.215685,29594.4
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.149811,29571.4
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,33.1354,377887
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.304427,29583.4
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.267602,29541
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.215435,29510.2
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,38.1229,426276
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.277861,29560.4
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.542455,32131.6
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.141919,29533.6
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.14318,29522
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,22.5776,265834
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.205087,29556.2
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.152338,29556.2
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.138809,29533.2
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,27.414,308877
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.252191,29471.8
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.210462,29471.8
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.147788,29456.6
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,22.3812,264550
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.216171,29414.6
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.535364,33256
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,34.752,385615
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.268031,29533.4
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.591761,33382.4
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.127071,29525.6
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.140767,29521.8
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,27.3634,322090
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.210488,29556.4
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.140217,29548.8
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.132031,29533.4
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,27.1453,302907
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.256594,29487.6
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.203961,29487.6
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.145919,29453.2
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,22.5727,266023
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.223788,29395.8
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.540617,33248.8
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.114143,29392
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.138911,29388.2
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,27.3115,320898
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.212273,29434
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.143405,29418.8
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.131839,29388
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,27.2872,302823
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.257631,29357.6
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.211833,29342.2
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.151666,29323
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,22.5029,264889
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.216281,29273
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.536426,31828.8
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.106489,29257.6
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.131033,29238.2
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.571792,29234.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.31415,35662.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.18076,29234.4
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.81389,113120
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_154.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_154.txt
deleted file mode 100644
index 067c920d31..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_154.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,157.412,1.44002e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.426526,24981
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.238553,24992.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,4.97158,53004.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.63666,33152.2
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,85.7847,864414
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.251442,26673.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.312069,27214
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.145926,26688.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,74.0809,731025
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.380926,26802.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.472272,27400.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.304971,26829.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,113.196,1.11623e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.36746,27088.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.09909,30296
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,123.136,1.24187e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.364293,27520.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.09994,31054
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.272811,27520.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.308165,27509
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,122.748,1.26734e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.345758,27815.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.459882,28574.6
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.286898,27820.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,76.3162,785326
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.413291,27907
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.49345,28723.6
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.298609,27931
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,112.878,1.15786e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.379838,28077.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.09535,32336.6
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.281636,28100.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.306954,28081.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,123.218,1.29815e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.345298,28352
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.448587,29271
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.285061,28386.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,76.3109,799976
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.386641,28459.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.467153,29374.6
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.299596,28432.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,113.042,1.17779e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.365445,28576.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.09179,33428.8
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.269726,28585.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.309393,28578.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,82.3655,887395
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.317713,28696.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.211417,28696.2
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.158912,28700
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,48.9766,528817
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.329003,28685.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.316798,28692.8
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.280562,28692.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,64.8027,694985
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.31427,28712.2
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.64056,30779.8
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,78.3311,843330
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.313381,28993.8
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.626512,31110.6
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.276158,29016.4
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.251454,29016.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,62.9651,693034
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.295518,29153.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.256543,29138
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.221964,29153.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,48.4493,529819
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.340977,29168.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.30883,29157.2
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.275044,29153.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,64.5001,703081
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.313451,29183.2
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.635261,31407.4
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.240779,29198.8
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.192492,29187.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,62.9989,700218
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.304178,29316.8
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.26213,29294
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.211353,29301.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,49.2558,539968
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.336613,29317
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.309387,29305.6
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.284017,29317
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,65.2136,715036
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.331486,29307.6
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.673769,31604.6
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.232555,29289.6
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.179698,29297.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,63.0713,702641
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.300715,29427.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.266252,29423.6
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.222828,29427.6
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,48.9043,538881
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.334436,29438.8
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.300997,29454.6
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.278214,29458.4
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,64.7474,710926
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.313753,29398.6
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.644713,31798.8
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.233675,29414.4
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.194936,29426.2
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,39.5635,441931
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.283596,29477.4
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.186444,29481.4
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.140466,29450.6
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,33.2257,382372
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.30698,29484.6
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.255805,29469.4
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.203441,29473.4
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,37.8014,420480
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.277144,29445.2
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.557737,31872.2
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,49.3701,556516
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.292926,29663.4
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.519274,32112.2
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.14069,29632.8
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.148063,29606
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,36.8472,408116
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.277093,28481.8
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.357189,28489.6
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.147686,28466.2
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,33.0947,368114
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.294905,29740
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.329503,29724.6
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.225835,29709.4
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,37.9501,423559
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.28419,29674
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.557891,32122.8
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.160114,29658.8
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.151301,29662.8
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,37.0226,417085
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.26883,29724.2
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.210149,29709
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.147206,29705
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,33.1235,376699
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.304235,29705
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.268101,29708.8
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.2283,29686
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,37.8752,420981
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.28149,29647.6
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.561463,32134.8
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.157068,29655.2
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.151455,29655.2
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,36.9585,408188
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.277784,29765.6
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.204543,29765.6
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.144562,29731.2
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,33.0791,377044
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.297194,29754.2
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.259218,29727.4
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.21715,29708.4
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,37.9043,424006
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.282616,29693.2
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.548387,32211
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.141273,29685.6
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.138105,29685.6
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,37.064,418788
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.27884,29743
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.20677,29735.4
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.143935,29716.4
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,33.1562,376862
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.299825,29693.6
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.269265,29701.2
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.222162,29701.2
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,37.8704,424139
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.275883,29689.6
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.561386,32238.2
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.154866,29701.2
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.148389,29705
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,37.0441,414478
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.270879,29747
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.210706,29750.8
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.149049,29731.6
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,33.4192,378709
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.304953,29747
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.267794,29724.2
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.22981,29701.4
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,38.1146,427383
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.298494,29739.6
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.553098,32253.6
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.150943,29712.8
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.16142,29697.6
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,23.2619,262910
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.243991,29751
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.172358,29751
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.139942,29701.4
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,29.4479,325576
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.313118,29697.6
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.235096,29697.6
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.157829,29682.4
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,23.4012,269476
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.247013,29648.2
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.539818,32215.6
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,35.8414,412556
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.256472,29755
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.546928,32318.6
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.117901,29747.4
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.14503,29743.6
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,29.3886,330189
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.240606,29812
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.186482,29808.2
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.155845,29755
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,29.1706,330761
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.275525,29747.4
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.213388,29728.4
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.152934,29693.8
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,23.5147,269612
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.256504,29678.4
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.555767,33523.2
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.10871,29678.4
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.139443,29678.4
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,29.3091,327858
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.235557,29739.4
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.203141,29724.2
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.135724,29693.8
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,29.4449,337069
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.273938,29705
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.245151,29705
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.148031,29689.8
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,23.349,269193
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.244395,29659.4
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.538243,33489
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.113638,29640.4
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.133036,29640.4
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.566122,29625.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.21819,36033.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.129061,29625.2
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.76049,114679
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_155.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_155.txt
deleted file mode 100644
index 37a7257c20..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_155.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,183.655,1.70499e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.456855,26896.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.24339,26873.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.25813,61030.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.67416,35320.2
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,89.6427,920007
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.234719,27194
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.324184,27723
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.152524,27182.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,89.23,906289
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.406942,27378
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.479761,27948.8
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.310475,27354.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,126.38,1.27529e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.37779,27665.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.14248,30987.8
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,134.546,1.3891e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.373188,28139.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.11864,31847.8
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.271109,28139.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.308266,28127.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,137.138,1.44182e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.375184,28505.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.484829,29325
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.3115,28498
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,92.5404,975203
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.404177,28539.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.489238,29379
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.301674,28532.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,126.653,1.33032e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.378743,28758
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.10453,33265.6
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.277784,28784.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.311582,28754.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,137.615,1.48066e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.370686,27588.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.456746,30026.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.291621,29072.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,92.7001,993818
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.406884,29084.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.486877,30083.6
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.301214,29099.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,126.857,1.35328e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.379511,29238.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.10819,34343.6
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.27477,29251.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.30197,29263.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,85.9931,937869
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.307563,29403.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.262501,29407.6
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.209445,29407.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,56.3445,617641
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.351352,29411.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.309567,29415.4
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.29502,29419.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,71.4179,779676
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.335896,29435.2
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.672188,31613.2
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,85.2213,936301
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.322757,29714
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.684764,31941.6
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.26709,29729.2
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.285112,29728.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,69.5055,777360
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.307467,29888.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.295346,29865.2
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.273612,29838.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,55.6456,622355
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.354283,29855.2
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.307979,29859.4
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.298238,29851.8
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,71.1005,788790
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.344088,29886.4
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.67972,32201.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.257502,29890.2
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.268862,29875
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,69.5758,781009
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.309886,30051.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.291429,30052
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.278302,30047.8
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,55.7293,630570
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.358417,30044.8
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.302398,30037.2
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.280664,30048.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,70.8152,789570
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.331947,30025.4
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.682492,32437.2
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.257657,30029.6
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.260459,30037.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,69.301,786044
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.304894,30165.4
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.292907,30158.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.271474,30165.8
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,55.4304,626396
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.343966,30192.6
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.310162,30185
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.279954,30169.8
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,71.2128,794468
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.337311,30157.6
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.679983,32637.4
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.258974,30166.2
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.26563,30132
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,41.3325,482609
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.276421,30189
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.173746,30143.2
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.140543,30112.8
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,37.025,429270
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.303179,30174.2
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.291083,30158.8
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.25411,30147.4
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,40.5378,462081
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.279902,30127.4
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.59878,32637.4
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,52.3163,606454
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.303967,30369.8
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.59624,32910.4
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.149779,30354.4
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.151398,30370.6
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,39.9549,457407
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.267909,30432
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.29644,30432
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.167033,30432.6
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,36.6152,426784
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.305445,30411.8
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.265861,30411.8
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.253298,30404.4
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,40.3344,461444
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.286495,30407.2
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.558589,32982.8
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.168377,30396.4
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.153906,30381
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,39.9332,459379
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.275295,30502.4
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.238168,30506.4
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.164415,30471.8
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,36.4597,427336
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.306392,30482.8
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.259096,30483.4
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.256478,30472.2
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,40.135,460351
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.274635,30451
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.539401,33064
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.165599,30458.6
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.145452,30445
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,39.8381,458724
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.276383,30538.6
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.240191,30550.4
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.158464,30512.4
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,36.4469,427860
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.310207,30493.8
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.273945,30501.4
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.253772,30501.4
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,42.2629,466832
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.285644,30449.8
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.534654,33089.8
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.156652,30449.8
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.142707,30453.6
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,40.058,458736
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.27996,30544
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.243788,30529
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.179628,30472
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,36.7695,429140
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.315858,30475.4
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.263141,30483
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.257029,30468.4
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,40.2568,460611
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.279065,30434.8
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.558212,33082.4
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.171039,30431
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.147628,30423.4
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,39.7867,457188
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.275621,30567.6
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.240747,30536.8
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.164396,30506.2
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,36.5835,426645
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.3065,30533.8
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.269969,30506.8
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.253131,30480.2
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,40.4429,464271
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.279346,30488.4
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.559511,33128.2
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.159743,30442.6
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.140792,30450.2
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,23.5917,274302
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.23493,30522.2
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.165426,30506.8
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.133791,30468.6
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,28.4656,337130
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.279794,30480.8
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.203634,30465.4
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.144748,30415.4
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,23.2664,274717
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.237093,30412.8
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.546628,34418.4
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,35.5477,414063
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.251928,30522.8
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.56863,33220.6
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.125317,30519.4
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.140825,30519.8
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,29.5255,340010
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.23292,30541.8
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.180313,30534.2
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.146803,30534.6
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,28.3773,339285
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.278532,30481.6
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.203046,30462.4
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.152735,30467
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,23.251,276655
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.236005,30448.6
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.552611,34492.8
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.124115,30437.2
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.14535,30429.6
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,29.257,334906
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.238271,30482.4
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.192607,30482.4
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.138751,30432.8
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,28.0083,338484
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.277989,30411
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.199072,30407.2
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.147059,30384.6
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,23.2564,274690
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.254187,30358.8
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.532086,33021.6
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.11239,30355
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.134969,30355
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.580304,30339.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.31172,37012
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.165446,30355
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.76924,117418
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_156.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_156.txt
deleted file mode 100644
index f9f71a75a2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_156.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,183.841,1.71106e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.447505,27072.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.23525,27075.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.84951,65216.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.83568,35526.8
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,89.8907,926980
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.244222,27367
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.330532,27892.2
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.153331,27351.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,91.3554,928589
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.397432,27500.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.481201,28094.8
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.293297,27500.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,127.227,1.28517e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.380126,27852.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.11322,31224.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,135.209,1.40165e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.367435,28219.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.11903,32012.6
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.272037,28246.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.312255,28265.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,138.705,1.46232e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.375307,28600.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.472489,29416.2
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.296324,28600.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,94.4063,989034
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.401322,28695.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.471633,29519.2
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.305631,28676.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,127.311,1.33592e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.382903,28864.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.12661,33387.4
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.264179,28868.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.306239,28873.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,139.275,1.49541e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.386826,29186.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.466039,30159.4
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.304684,29186.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,94.6285,1.01322e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.41301,29183.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.485841,30186.2
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.300637,29187
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,127.613,1.36217e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.381278,29386
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.12636,34566.8
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.275724,29362.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.336292,29354.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,87.8895,963427
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.312837,29488.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.282719,29438.4
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.189074,29442.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,60.8412,672193
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.357085,29472.4
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.310352,29468.6
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.28499,29472.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,73.548,805069
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.327551,29504.6
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.672323,31698
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,89.3331,990462
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.336407,29853.4
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.681084,32103.8
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.254277,29822.2
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.297612,29833.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,72.762,812777
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.310539,30015
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.298385,30000.8
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.278111,30016.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,61.0433,683127
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.355538,30011.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.31283,30004.4
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.288356,30012
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,73.3172,817841
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.338277,30028.2
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.683209,32359.4
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.265445,30020.8
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.292492,30005.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,72.7418,819128
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.322942,30167.2
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.303774,30175.2
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.282187,30182.8
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,60.631,678941
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.364017,28657.2
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.30645,28645.8
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.278111,28653.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,73.3072,820758
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.337003,30144.6
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.678677,32608.2
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.261753,30163.2
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.26421,30163.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,72.3336,817532
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.303852,30409.2
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.263781,30382.6
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.279262,30371.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,60.9911,690520
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.356503,30398.2
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.282616,30360
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.284019,30352.4
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,73.3651,825208
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.336298,30372.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.64561,32890.2
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.256133,30380.2
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.27276,30365
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,41.2758,486233
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.271634,30402.2
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.173995,30387.4
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.145017,30357.4
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,36.6619,426979
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.309931,30407.2
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.272978,30369.8
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.248748,30358.4
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,40.1366,460069
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.280254,30320.6
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.562332,32888.2
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,52.0583,600905
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.301892,30558.8
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.566268,33126
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.165676,30524.2
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.153771,30505.2
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,39.6696,457791
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.277062,30555.4
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.244108,30543.8
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.174348,30540
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,36.6356,430399
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.314059,30563.6
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.269349,30528.8
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.251333,30471.2
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,40.3201,465937
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.284267,30498.2
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.542653,33088
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.154604,30471.8
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.142611,30441
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,39.7402,459177
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.275288,30613.2
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.242481,30620.8
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.16108,30560.2
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,36.5743,428505
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.306814,30579
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.263998,30579
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.257918,30532.8
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,40.3354,464741
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.296484,30505
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.551325,33157
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.164127,30509.4
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.154897,30513.4
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,39.9405,461249
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.281093,30612
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.242578,30616.2
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.171832,30601.2
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,36.7744,428966
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.310782,29056.2
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.25765,29041
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.247243,29010.4
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,40.3897,464965
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.290585,30532.8
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.539595,33199.2
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.152684,30521.2
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.144799,30521.2
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,39.7905,459768
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.274194,30628
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.240805,30589.8
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.164262,30574.4
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,36.4401,427149
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.310365,30578.2
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.265662,30585.8
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.255384,30559.2
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,40.2121,461638
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.288658,30536.8
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.561917,33241.6
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.164185,30532.8
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.144428,30532.8
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,40.0179,461323
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.286577,30570
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.245457,30562.4
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.172946,30555.6
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,36.7732,430440
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.312637,30575.2
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.26755,30548.6
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.252383,30533.6
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,40.329,464807
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.292203,30492
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.548919,33212.2
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.157554,30511
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.14229,30511
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,25.0052,303830
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.207231,30502.6
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.145625,30499.2
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.1291,30499.8
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,35.0375,405627
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.267736,30515.4
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.25946,30500.2
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.172242,30466.4
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.0376,305400
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.261036,30417
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.566115,34476.4
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,40.0182,459987
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.286974,30573
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.602563,34681.8
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.127686,30573.4
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.144646,30573.4
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,32.5317,373056
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.247097,30618
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.202796,30618.4
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.141715,30596.4
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,35.312,410153
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.275121,30577.4
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.25475,30581.2
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.187147,30558.2
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.039,306435
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.258117,30535.2
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.574249,34644
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.115705,30535.2
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.135147,30531.4
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,32.2592,371059
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.236882,30599.6
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.190277,30580
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.142733,30554.2
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,35.026,406656
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.267121,30550
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.243339,30546.2
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.173516,30511.8
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.0226,306052
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.265285,30443.4
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.565577,34549.2
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.112774,30440.4
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.129964,30440.4
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.559332,30425.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.19279,35903.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.183986,30410
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.72707,117603
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_157.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_157.txt
deleted file mode 100644
index d83b30ed36..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_157.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,184.037,1.74891e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.443012,27344.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.243077,27351.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.1934,60736.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.65401,35365.2
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,90.7489,929264
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.23653,27619
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.289701,27638
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.170201,27622.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,91.7284,947562
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.406648,27802.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.484952,28411.8
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.308108,27794.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,127.872,1.29485e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.384727,28082.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.13501,31447
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,135.453,1.41071e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.374609,28536.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.12737,32314
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.283288,28521
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.308958,28528.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,138.84,1.47588e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.366968,28854.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.469406,29658.8
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.301509,28847
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,95.2008,1.01004e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.416651,28904.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.471543,29724.2
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.300472,28881.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,127.64,1.35579e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.37987,29020.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.11009,33543.2
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.273061,29023.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.325067,29027.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,139.131,1.50154e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.370149,29397.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.468132,30350.6
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.30277,29362.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,94.9011,1.02157e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.414647,29393.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.472419,30342.8
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.306379,29335.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,128.31,1.38079e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.395966,29507.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.11566,34712.2
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.27116,29523.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.324018,29515
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,87.7195,973511
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.318282,29590
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.266244,29574.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.193772,29559.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,61.1538,676838
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.367569,29678
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.318193,29651.2
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.291051,29662.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,73.4264,808655
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.341751,29682.6
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.686966,31872.8
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,89.0127,990530
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.341112,29980.8
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.719394,32223.6
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.927963,36037
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.291153,29973
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,71.9851,811042
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.306795,30089.8
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.300305,30097.4
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.273918,30101.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,60.922,686665
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.371102,30097.4
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.319826,30105
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.282437,30105
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,73.4337,820665
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.349451,30063
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.673603,32424.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.273355,30078.6
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.350481,30090
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,72.3126,817258
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.312568,30299.4
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.26147,30299.4
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.279896,30288
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,60.7685,685817
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.358328,30295.4
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.309534,30303.4
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.28666,30299.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,73.5745,826331
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.339096,30348.8
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.679562,32809.4
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.273196,30329.2
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.293163,30317.6
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,72.5525,824476
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.312811,30427.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.270135,30435.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.274104,30454.4
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,61.2862,697751
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.358526,30439.8
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.3166,30417
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.285343,30409.2
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,73.6448,831108
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.348503,30416
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.778728,32949
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.295646,30431.8
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.27875,30424.2
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,44.9516,517351
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.273887,30497
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.206098,30485.4
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.15189,30466.4
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,50.8041,585431
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.315,30500.8
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.267314,30504.6
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.261637,30504.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,47.2666,549398
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.305771,30480.8
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.611209,33044.8
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,63.2374,736038
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.31301,30702.6
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.585501,33293
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.170975,30706.6
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.155448,30710.6
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,47.0784,554991
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.285502,30790.6
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.230347,30779.2
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.156792,30783.2
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,51.0753,592644
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.317253,30911.8
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.270072,30881.4
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.259582,30851
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,47.033,554408
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.317163,30842.4
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.551888,33463.4
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.158744,30819.6
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.15255,30831.2
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,47.1293,558905
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.290303,30960.6
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.265464,30968.4
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.147993,30915
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,51.0673,592923
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.319966,30987
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.266552,30979.4
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.262852,30949
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,47.0602,557219
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.289369,30937.6
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.605668,33585.2
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.144992,30914.8
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.14478,30922.4
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,47.3598,560137
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.287717,30987
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.227545,30971.2
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.148748,30968
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,51.1172,594484
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.325733,31055
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.275045,31005.6
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.256914,31009.4
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,47.2048,558233
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.287537,30945.4
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.548771,33642.8
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.146091,30937.8
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.142482,30937.8
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,47.2641,559373
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.286123,31100.8
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.229521,31100.8
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.153119,31024.8
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,51.5505,596775
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.334679,31116
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.317009,31127.4
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.260927,31097.4
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,47.327,559971
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.29964,31070.8
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.615836,33802.4
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.149376,31070.8
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.142489,31063.2
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,47.1668,562303
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.289791,31146.6
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.272671,31131.2
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.153344,31131.2
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,51.1954,597754
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.333776,31120
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.316363,31131.4
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.265713,31135.2
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,47.176,558821
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.295934,31078.2
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.559991,33840.6
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.146329,31078.2
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.140812,31082
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,25.0182,310885
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.213477,31093.4
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.147782,31089.6
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.130559,31089.6
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,35.141,413296
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.272318,31070.4
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.264689,31059
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.183947,31036.2
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.0937,312512
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.26469,30994.4
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.584111,35088
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,40.3002,471372
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.285515,31112.4
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.609066,35256
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.130501,31108.6
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.137804,31104.8
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,32.3058,377751
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.229446,31162.4
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.203078,31127.8
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.139257,31101
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,35.0721,413987
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.264562,31108.6
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.260287,31108.6
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.179148,31070.6
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.5144,317268
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.268465,30994.4
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.579894,35088
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.119564,30971.6
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.137945,30967.8
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,32.1678,378250
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.22988,31044.2
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.192691,31040.4
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.136056,31021.4
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,34.988,412461
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.270271,30998.4
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.253483,30986.8
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.172108,30926.4
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.1216,311619
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.273752,30892.6
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.568605,34963.2
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.119564,30862
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.128972,30862
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.590134,30862
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.28456,37672
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.128435,30862
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.84476,119309
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_158.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_158.txt
deleted file mode 100644
index e2b008c239..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_158.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,177.887,1.63634e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.433195,26363.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.295205,26363.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.33953,58518.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.66843,34145.6
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,88.8278,893770
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.228069,26669.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.318456,27214.4
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.148338,26670.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,88.9684,884484
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.416318,26825.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.483959,27427.8
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.31468,26819
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,121.7,1.19676e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.382007,27099.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.12918,30394.4
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,128.357,1.28909e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.372445,27550.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.11873,31183.8
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.278878,27554.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.318617,27562
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,136.613,1.41016e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.359582,27908.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.470512,28686
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.295352,27939.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,91.951,938582
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.411972,27973.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.480304,28801.2
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.312747,27989
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,121.65,1.23837e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.405841,28158.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.1231,32583.8
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.2766,28175.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.335831,28183.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,137.041,1.43943e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.375626,28462.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.463965,29389.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.303831,28447.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,92.4919,953552
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.407895,28512.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.502417,29454.2
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.318731,28470.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,121.354,1.26669e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.396478,28616
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.11292,33598.2
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.274649,28643
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.316722,28650.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,85.7043,917592
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.309611,28787.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.306098,28779.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.204434,28772
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,55.356,598359
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.351076,28783.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.323435,28791.4
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.284332,28799
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,67.7671,728263
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.327902,28776.2
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.681545,30912.2
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,81.8561,878304
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.333324,29065.8
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.67736,31217
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.264946,29069.6
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.290443,29084.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,68.9511,759009
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.303659,29240.8
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.302053,29233.2
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.274181,29222
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,55.2878,606947
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.362929,29221.8
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.314289,29206.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.289605,29195.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,67.6555,734937
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.329559,29202.8
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.685903,31468.4
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.260734,29210.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.262245,29191.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,69.3951,766491
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.312222,29420
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.270744,29378.2
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.287359,29336.4
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,56.0401,617063
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.364241,29366.8
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.279359,29370.6
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.286239,29344
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,67.7171,742335
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.329911,29325.2
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.660981,31690.8
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.261893,29325.2
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.243436,29325.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,68.7807,765743
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.307409,29488.8
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.260831,29458.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.276344,29466
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,55.5994,617697
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.35715,29469.8
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.274527,29439.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.286756,29450.6
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,67.9684,744086
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.327621,29466
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.643491,31895.8
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.267435,29477.4
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.25788,29469.6
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,41.0915,471844
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.269349,29523.2
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.163085,29481
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.140473,29462
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,36.564,413404
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.312094,29542.2
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.268018,29527
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.262546,29496.4
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,39.5747,442848
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.285265,29496.6
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.580727,31934
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,51.1295,565950
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.293726,29653.6
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.580413,32152.2
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.187782,29649.6
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.153835,29641.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,39.8871,446857
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.283499,29787.4
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.248978,29775.8
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.165996,29756.6
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,36.524,417383
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.30867,29768.4
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.266059,29779.8
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.257279,29749.2
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,39.5029,445415
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.280012,29741.2
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.56408,32266.4
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.159545,29729.6
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.151756,29733.6
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,40.4872,456874
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.29162,29822.2
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.24579,29826
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.167641,29826
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,36.4926,416681
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.317535,29814.2
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.272741,29806.6
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.262469,29772.2
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,40.1935,453141
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.324958,29735.4
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.566973,32295.4
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.163078,29739.4
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.155326,29739.4
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,39.9648,449639
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.290092,29827.8
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.24638,29820.6
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.168294,29790
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,36.5355,418233
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.317527,29796.6
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.278052,29781.4
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.265791,29781.4
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,39.537,447472
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.294897,29765.6
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.565488,32352
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.155013,29750.4
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.154271,29762
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,39.8897,449917
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.295204,29853.8
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.23029,29808
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.158713,29792.8
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,36.6149,418511
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.321541,29807.4
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.26588,29780.8
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.263481,29769.4
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,39.5218,447389
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.288312,29750.4
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.563862,32314
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.148448,29735.2
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.140882,29739
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,39.8227,448320
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.276875,29891.6
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.246776,29880.2
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.162086,29884
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,36.352,418849
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.316415,29872.6
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.269099,29876.4
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.256485,29849.8
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,39.4144,448397
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.290315,29823
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.565086,32439.8
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.149798,29834.4
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.140882,29830.6
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,23.5226,269978
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.236965,29895.4
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.156082,29880.2
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.137004,29876.4
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,29.2506,334447
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.266174,29838.2
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.223263,29838.2
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.143045,29800
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,23.1934,269074
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.248857,29742.8
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.5436,33656.6
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,35.4628,397625
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.258136,29868.8
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.585308,33767.4
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.12638,29853.6
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.137017,29853.6
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,29.1434,328638
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.237393,29960.6
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.186795,29960.6
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.133214,29907
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,29.303,335380
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.27973,29872.8
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.220441,29872.8
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.144159,29865.2
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,23.3877,271183
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.247499,29861.6
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.546614,33821.6
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.112639,29846.4
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.140646,29838.8
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,29.4813,329859
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.252971,29876.8
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.201963,29861.6
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.154258,29861.6
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,29.6998,336044
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.279595,29819.6
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.229739,29815.8
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.152671,29770
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,23.3263,268863
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.249842,29728
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.545891,33665.2
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.109421,29709
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.129702,29705.2
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.589693,29709
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.17853,34981.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.183544,29705.2
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.8411,114887
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_159.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_159.txt
deleted file mode 100644
index 71d4f29c81..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_159.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,178.95,1.64736e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.445469,26509.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.249343,26486.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.23551,56963
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.80194,34749.6
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,89.6136,903481
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.237195,26819.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.336881,27348.8
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.150284,26838.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,90.2661,891722
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.39884,26975.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.466648,27558.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.31276,26956.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,121.864,1.20135e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.378122,27213.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.11757,30508.8
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,129.262,1.31027e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.37068,27661.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.11818,31301.2
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.282495,27650.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.299224,27654.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,138.12,1.42603e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.38099,28004.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.472016,28816.4
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.30602,28023.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,93.9075,952876
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.408291,28084.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.48696,28916
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.314092,28065.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,122.491,1.26097e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.379704,28251.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.12823,32652.6
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.271026,28225.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.31299,28218.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,138.171,1.45993e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.367133,28565.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.462378,29496.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.295204,28558
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,93.9862,982531
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.406078,28611.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.475434,29588.2
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.313521,28615.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,122.728,1.2801e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.379076,28704.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.12064,33732.4
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.271684,28700.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.298533,28696.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,87.5864,943336
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.324734,28787
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.293387,28794.6
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.18544,28802.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,60.394,652253
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.363378,28848
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.31731,28851.8
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.286635,28855.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,70.8305,760619
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.336478,28844.8
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.676765,30977
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,86.0779,932212
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.336766,29172.4
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.699401,31354.4
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.271487,29168.6
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.307615,29168.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,71.9579,791814
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.315326,29332.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.30165,29343.6
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.276536,29336
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,60.8314,667645
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.360236,29328.2
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.309119,29320.8
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.282699,29336.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,70.6395,769970
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.336945,29343.8
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.673769,31617.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.26924,29340.2
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.279647,29344
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,72.1388,797935
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.317208,29507.8
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.303833,29477.2
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.283333,29492.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,60.3464,665570
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.36117,29538.2
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.319621,29492.6
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.290712,29496.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,70.671,771160
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.34435,29481
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.665565,31865.6
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.256589,29484.8
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.277196,29492.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,72.1152,803586
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.314015,29618.4
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.30092,29614.6
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.278603,29626.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,60.7743,673067
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.358865,29641.4
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.304389,29641.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.300766,29630
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,70.5842,781178
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.340011,29606.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.706326,32074.8
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.270098,29618.2
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.261983,29622
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,41.195,474310
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.268132,29667.6
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.188249,29633.4
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.139129,29625.6
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,36.5166,416155
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.30549,29668
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.292786,29675.6
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.253151,29675.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,39.5776,447119
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.29996,29634
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.597469,32113.4
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,51.2933,570957
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.289547,29870.8
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.645821,32403.6
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.191001,29878.4
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.162963,29867
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,40.1829,452606
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.278526,29969.6
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.254956,29973.4
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.170412,29970.2
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,36.3363,420228
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.311231,29985.6
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.26821,29993.2
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.258386,29981.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,39.6441,450149
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.291339,29958.8
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.573239,32507.2
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.154604,29947.2
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.144831,29920.4
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,40.5696,456602
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.295634,29973.2
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.245874,29974
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.16439,29958.6
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,36.2945,420522
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.319058,29962.6
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.263538,29921
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.253163,29909.8
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,39.7472,452460
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.341124,29870
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.567389,32456.8
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.153478,29862.8
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.15957,29866.8
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,39.587,448730
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.281087,29996.4
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.246149,30004.2
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.169196,29992.6
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,36.5315,418876
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.321093,29991.6
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.26629,29961.2
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.257361,29946
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,39.5941,448639
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.281567,29933.6
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.574582,32542.8
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.154501,29937.4
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.146572,29941.2
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,39.6379,452452
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.278629,29998.8
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.239449,29998.8
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.157145,29998.8
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,36.4814,422310
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.31004,29998
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.267583,29986.6
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.254885,29990.6
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,39.5472,448826
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.29164,29963.6
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.568028,32588.4
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.150584,29952.2
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.142399,29952.2
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,39.8558,451137
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.272004,30039.6
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.244734,30020.8
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.165535,29997.8
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,36.6778,424021
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.310711,30023.8
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.26261,30024.4
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.25989,29997.8
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,39.7759,452455
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.287717,29994.6
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.565059,32642.2
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.152563,30006
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.142476,29990.8
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,25.1332,300041
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.207813,30013
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.144984,30009.2
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.128691,30006
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,34.9348,399055
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.272933,28503
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.255666,30021.2
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.180089,29983.2
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.5154,303522
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.263698,29952.6
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.568029,32596.4
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,40.35,455296
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.277919,30044.6
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.58321,32684.2
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.13006,30013.8
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.143826,30013.8
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,32.5606,371085
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.239788,30120.6
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.183615,30101.6
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.135814,30074.8
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,34.6022,394215
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.275653,30082
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.264139,30047.4
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.175404,30032.6
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.3439,301169
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.253778,30032.4
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.556323,32664.4
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.117516,29986.4
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.140172,29987.6
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,32.433,369142
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.239486,30086.2
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.201214,30082.4
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.144537,30048
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,35.0333,402416
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.276101,29978.8
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.255513,29978.8
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.17397,29978.8
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.3247,300612
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.25468,29960
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.564028,32631
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.118381,29960.6
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.132915,29960.6
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.576157,29945.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.37604,37941.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.131596,29945.4
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.81582,115814
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_160.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_160.txt
deleted file mode 100644
index bd61005808..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_160.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,180.811,1.66495e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.434461,26820.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.242584,26832.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.39652,61834.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.81126,35642.4
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,88.8753,907489
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.234514,27125.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.33219,27658.4
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.153164,27121.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,91.5815,923378
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.407933,27254.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.470582,27852.8
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.306903,27251.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,122.833,1.22597e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.403633,27420.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.12989,30722.8
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,129.745,1.32135e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.379076,27837
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.12334,31477
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.268107,27848.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.300171,27863.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,137.759,1.42956e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.372881,28160.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.465329,28918.6
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.302916,28133.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,94.013,968078
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.41544,28183.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.469956,28991.8
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.309413,28191.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,122.4,1.25462e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.381361,28380.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.11527,32713
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.278411,28389.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.304254,28401
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,138.469,1.45279e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.377572,28665.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.473507,29592.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.287282,28638.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,93.8208,981679
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.409526,28669.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.471421,29626.6
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.300747,28673.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,122.117,1.27401e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.381886,28814.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.12017,33743.2
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.279212,28773
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.309137,28765.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,86.9685,938565
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.31651,28860.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.295096,28852.6
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.189177,28860.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,60.3326,650886
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.355115,28917.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.314309,28902.2
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.286417,28909.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,70.475,756598
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.342878,28914
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.671484,31050
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,85.3623,920484
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.340875,29218.2
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.640086,31411.6
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.260504,29244.8
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.291947,29248.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,71.6881,790987
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.304843,29450.4
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.257925,29412.4
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.276664,29397
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,59.958,662105
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.353873,29427.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.271909,29412.4
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.286943,29427.6
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,69.4054,763271
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.332958,29420.2
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.631779,31713.2
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.269374,29405
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.258724,29405
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,71.4036,797477
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.303051,29583.8
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.262584,29568.6
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.273131,29549.4
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,59.7342,658091
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.359044,29584
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.269835,29553.4
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.282751,29564.8
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,70.346,771867
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.341105,29511.6
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.651395,30362.8
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.260607,28004.8
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.267992,28012.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,72.0731,800823
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.309176,29729.8
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.271172,29718.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.279179,29725.8
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,60.274,668407
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.351018,29737.2
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.269176,29729.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.293016,29733.2
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,69.6733,767808
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.336714,29710.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.644956,32182.4
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.267454,29729.8
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.256735,29729.8
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,44.7788,505807
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.275563,29771.4
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.204639,29767.6
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.142022,29775.2
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,51.0445,562260
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.316734,29863.4
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.285528,29848.2
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.258707,29806
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,45.6316,510400
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.286731,29799.2
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.580278,32294
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,61.9518,698222
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.300747,30034.4
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.599926,32563.4
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.191467,30023
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.155768,30019.4
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,47.5582,539705
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.292823,30171.6
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.227557,30160.2
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.161867,30156.8
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,50.709,575155
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.319249,30184.6
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.267609,30169.6
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.259512,30154.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,45.7716,517697
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.279096,30128.2
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.574224,32699.4
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.174585,30113
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.145433,30113
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,46.8376,544443
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.283781,30240
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.214412,30225.2
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.147859,30203
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,50.6623,577137
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.31436,30303.8
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.266424,30311.6
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.251577,30296.8
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,45.9228,519230
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.278909,30255
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.56934,32852.8
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.179666,30228.2
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.146393,30216.8
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,47.0784,544933
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.285554,30374.6
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.222854,30359.2
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.152709,30330.6
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,50.6732,578679
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.315857,30373.6
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.273548,30369.8
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.256569,30344.2
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,45.7746,521148
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.27715,30323.2
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.569488,32993.6
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.173811,30330.8
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.150284,30330.8
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,47.6709,552752
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.296427,30398.6
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.230423,30398.6
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.159129,30383.2
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,50.7314,581584
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.325162,30439.4
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.267608,30439.8
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.258353,30428.2
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,46.0305,525439
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.291218,30391.8
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.575217,33093
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.173823,30376.4
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.149996,30387.8
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,47.0194,544943
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.288383,30501.4
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.223448,30490
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.155475,30474.6
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,50.6882,584794
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.321521,30546.2
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.270443,30508.2
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.257157,30497.8
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,45.7568,523504
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.281439,30480.2
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.546327,33211.4
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.175622,30468.4
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.140556,30468.4
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,24.9118,304940
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.210175,30509.6
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.144536,30487.4
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.128287,30480.4
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,34.5936,400818
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.266424,30459
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.266981,30459
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.17564,30413.8
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.3795,309588
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.265292,30406.6
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.570391,34488.6
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,40.2506,461108
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.28067,30533.8
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.614173,33269.2
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.138079,29165
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.140287,29153.6
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,32.8367,376914
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.239909,30568
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.201535,30537.6
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.152203,30530.4
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,35.2247,406414
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.287941,30515.6
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.266418,30500.6
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.17854,30462.6
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.3405,305257
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.26234,30414
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.580495,34511.4
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.119391,30398.8
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.135698,30383.6
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,32.2078,367891
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.229311,30459.2
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.196408,30444
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.138092,30409.8
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,34.6151,400496
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.268049,30432.6
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.26853,30402.2
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.175622,30387.4
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.2198,306581
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.260862,30383.6
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.572611,34481
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.116781,30376
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.127788,30376.4
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.556406,30376.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.19609,37159.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.126393,30361.2
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.81717,118923
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_161.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_161.txt
deleted file mode 100644
index 393a1d43f9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_161.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,195.108,1.81519e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.433956,27253
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.248447,27256.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.56069,60508.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.68245,35757.4
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,92.8657,961447
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.246341,27557.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.353304,28102.4
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.16517,27550.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,101.276,1.03164e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.41187,27668.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.490878,28281.8
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.303582,27683.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,135.196,1.36598e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.407741,27882.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.13266,31293.8
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,139.646,1.44712e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.38332,28414.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.14604,32184.8
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.287864,28391.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.313713,28384.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,146.114,1.54317e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.372536,28743.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.474852,29595
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.29505,28767.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,104.885,1.10969e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.417706,28755
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.488695,29644.6
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.303038,28771.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,134.066,1.40621e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.378788,28913.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.12182,33515.8
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.295146,28905.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.328184,28894.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,146.041,1.572e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.376836,29232.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.47919,30201.4
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.296119,29240.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,104.732,1.11895e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.418819,29202.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.487286,30217.4
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.309464,29221.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,133.635,1.42558e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.387653,29343
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.10995,34547
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.278129,29366.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.314615,29335.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,89.8875,985723
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.329233,29408.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.274905,29415.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.205906,29419.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,64.8379,712536
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.364599,29438.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.320529,29446.2
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.286571,29435
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,76.2495,821583
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.343204,29527.2
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.692476,31716.4
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,92.6438,1.02244e+06
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.336126,29839.4
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.702646,32089.6
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.275473,29789.6
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.292702,29774.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,76.3134,851660
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.320133,29960.6
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.315825,29975.8
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.278495,29971.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,65.4582,729145
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.365451,29975
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.320875,29944.4
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.282949,29959.8
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,76.2076,848956
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.343403,29942.2
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.694325,32345.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.271698,29968.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.283691,29952.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,75.777,848897
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.307627,30152.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.306398,30145.2
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.279672,30088
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,65.7397,738812
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.369592,30121.6
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.32835,30129.8
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.287646,30121.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,76.5729,857258
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.342731,30072.4
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.707074,32544.2
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.264498,30091.8
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.287961,30095.6
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,75.7489,852577
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.320126,30201.2
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.326148,30189.8
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.279717,30205
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,65.3377,734122
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.356043,30238.4
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.31676,30200.8
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.287429,30200.8
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,76.4302,856344
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.340702,30193.2
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.701481,32737.6
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.264075,30182
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.298616,30205.6
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,45.2417,514760
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.289854,30253.8
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.204082,30250.4
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.146118,30228
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,51.639,589492
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.312382,30387.6
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.306705,30372.4
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.257554,30341.8
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,47.3801,543185
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.296766,30340.6
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.567389,32927.2
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,63.5269,733494
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.311218,30584.2
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.582103,33185.6
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.18062,30522.8
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.155372,30530.6
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,47.296,553746
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.286641,30668.6
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.237893,30661.2
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.152697,30641.8
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,51.5303,590393
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.318053,30694.8
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.320356,30671.6
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.259601,30664
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,47.4136,552558
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.2918,30659.8
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.558128,33295.8
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.158073,30648.2
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.155602,30632.8
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,47.0885,553691
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.285164,30758.8
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.229561,30728.4
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.159206,30717
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,51.1918,575795
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.317476,30789.2
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.27125,30796.8
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.25701,30770.2
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,47.3481,551140
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.295378,29172.4
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.556476,31850.4
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.169285,29172.4
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.154431,29172.4
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,47.8917,552417
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.290713,30861.8
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.235839,30861.8
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.164613,30846.6
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,51.3383,595616
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.318104,30839
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.270501,30842.8
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.267051,30835.2
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,47.4037,549551
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.291416,30804.8
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.580829,33498.4
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.161695,30801
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.146879,30808.6
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,47.2533,558207
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.290175,30872.6
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.230987,30872.6
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.154092,30850.4
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,51.1356,596563
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.312402,30922
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.30723,30895.4
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.25859,30872.6
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,47.2816,556614
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.291915,30842.8
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.555196,33551.6
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.154124,30823.8
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.143673,30808.6
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,47.5936,558954
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.29347,30979.4
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.231653,30983.2
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.158578,30949
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,51.1534,596072
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.31096,30967.4
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.276938,30967.4
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.260133,30956.6
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,47.4954,557616
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.285061,30945.2
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.559511,33691.2
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.162923,30918.6
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.148645,30922.4
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,25.147,307792
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.212024,30937
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.152408,30937
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.131871,30906.6
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,34.4724,405121
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.269125,30877.2
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.266245,30873.4
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.177253,30858.2
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.3295,311095
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.25173,30843.4
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.57754,34968
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,40.3556,469204
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.283448,30971.6
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.608348,35095
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.132089,30952.6
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.147173,30937.4
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,32.1849,374702
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.235044,30990.6
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.20373,30975.4
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.145592,30971.6
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,34.5602,407624
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.266455,30956.8
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.251327,30941.6
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.172813,30911.2
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.3661,311313
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.266814,30889
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.587753,35013.6
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.124921,30870
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.142643,30862.4
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,32.379,373741
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.253144,30908
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.205752,30904.2
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.152984,30881.4
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,34.4038,409689
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.275249,30874.2
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.253471,30859
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.169976,30855.2
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.177,309213
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.255499,30813.2
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.582589,34945.4
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.115347,30798
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.131411,30782.8
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.581072,30782.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.22635,36307.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.190278,30786.6
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.80082,118999
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_162.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_162.txt
deleted file mode 100644
index fb4626bdf9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_162.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,192.851,1.77804e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.479011,26612.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.291083,26605.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.81974,59212.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.68096,35310
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,90.9513,911523
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.291019,25482.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.386129,26038.6
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.204562,25494
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,102.102,1.0085e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.414692,27095.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.480548,27689.6
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.308171,27068.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,134.455,1.33407e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.399825,27398.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.34557,30832.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,138.664,1.41189e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.378116,27868
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.13367,31649.2
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.265528,27856.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.322885,27863.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,146.3,1.52113e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.383197,28269.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.464099,29123.8
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.291512,28304.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,105.171,1.09026e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.626102,29757.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.466179,29189
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.306781,28330.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,134.101,1.38829e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.386884,28544.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.08735,33112.8
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.271897,28540.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.335992,28544.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,145.958,1.54848e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.378865,28862.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.451722,29831.8
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.29027,28867
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,104.901,1.10299e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.414103,28859.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.473232,29874.2
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.312261,28875.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,133.38,1.3813e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.383249,29075.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.10488,34363
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.275263,29071.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.30286,29048.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,90.463,982056
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.326122,29203
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.260011,29191.6
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.199768,29164.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,65.2848,708388
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.363716,29233.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.317847,29207
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.284055,29210.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,76.133,824530
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.342418,29260
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.636086,31491.4
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,92.9708,1.01326e+06
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.337374,29599.6
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.636726,31900
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.271608,29615
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.3094,29607.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,75.9321,839184
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.304394,29759.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.260695,29763
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.272869,29778.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,65.6711,727721
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.365412,29772
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.314136,29779.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.283013,29764.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,76.3715,843024
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.347626,29736.6
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.685833,32136.6
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.269579,29756
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.280543,29748.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,75.8887,843630
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.310846,29977.2
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.266642,29980.4
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.275602,29991.8
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,65.5386,730208
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.361694,29980.8
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.315889,29962
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.280895,29969.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,76.5441,848252
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.343064,29923
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.64904,32371.4
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.259704,29919
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.278917,29938.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,75.6751,848366
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.31491,30110.8
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.291333,30118.4
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.273503,30118.4
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,65.5927,738506
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.374788,30095.6
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.315845,30107.2
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.279487,30118.8
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,76.3577,849457
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.368241,30080
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.654159,32636
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.268242,30091.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.289842,30103.2
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,45.0053,513316
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.279192,30152.8
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.192421,30152.8
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.141931,30133.8
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,51.3991,584398
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.316184,30243.2
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.296075,30228
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.252133,30212.8
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,47.2322,544416
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.292427,30193
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.567843,32787.2
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,63.3416,729243
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.306552,30450.8
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.568925,33071.6
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.164837,30444
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.153945,30451.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,47.4376,551935
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.28547,30554.6
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.225464,30543.2
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.153439,30543
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,51.2885,587265
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.344139,30622.4
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.316299,30607
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.261413,30592.2
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,47.3575,548002
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.295147,30580
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.5724,33250.4
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.157925,30572.4
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.147264,30568.6
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,47.308,555436
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.284984,30652.6
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.281573,30660
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.162776,30656.4
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,51.3185,595613
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.317982,30746.8
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.306072,30708.6
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.264127,30701
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,47.7061,555775
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.294418,30672.4
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.573706,33365.8
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.162495,30668.4
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.144012,30672.2
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,47.1191,551374
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.282993,30777.6
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.225099,30785.2
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.154136,30751
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,51.2538,593829
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.315755,30789
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.298027,30792.8
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.258776,30797
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,47.584,556042
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.287986,30752
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.550947,33495
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.151531,30755.8
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.146118,30744.4
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,47.0617,554070
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.280338,30777.6
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.223717,30785.2
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.153798,30785.2
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,51.4819,595343
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.322699,30831
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.307492,30808.2
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.263103,30796.8
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,47.4117,555088
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.289918,30774
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.558429,33517.4
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.145516,30751.2
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.140901,30736
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,47.0976,556246
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.283743,30865.4
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.225432,30865.4
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.150361,30865.4
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,51.5105,596858
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.33866,30967.4
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.301816,30914.2
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.263673,30895.2
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,47.3273,556680
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.299359,30872.4
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.616617,33633.6
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.144857,30854
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.149798,30838.8
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,25.0516,308726
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.211608,30880
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.155365,30880
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.138572,30827.4
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,35.1594,408993
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.260689,30834.4
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.266424,30823
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.187264,30804
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.1473,309016
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.253355,30774
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.570858,34923.4
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,40.0525,466727
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.290341,30887.8
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.605174,35083.4
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.133535,30884
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.148998,30884
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,32.4357,375446
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.23738,30929.6
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.195602,30925.8
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.138841,30891.6
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,35.2437,409419
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.255954,30857.2
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.253631,30857.2
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.186636,30842
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.203,309345
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.25861,30811.6
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.580131,34964.8
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.115737,30811.6
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.136569,30811.6
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,32.1955,374537
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.236248,30861.6
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.193721,30826.8
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.1371,30827.4
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,35.0225,405808
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.258142,30828.6
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.255333,30813.4
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.187666,30798.6
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.1779,308798
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.255858,30756.4
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.576118,34903.6
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.118701,30748.6
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.130361,30725.8
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.578742,30729.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.23008,36273.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.201163,30733.4
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.72888,118760
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_163.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_163.txt
deleted file mode 100644
index aeaff59555..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_163.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,194.17,1.81677e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.471095,26662.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.268471,26665.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.69155,58763
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.91877,35768.6
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,90.7652,914512
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.298795,27000.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.394622,27537.4
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.205254,26970
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,101.487,1.01196e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.409668,27141.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.502915,27731.6
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.317176,27122.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,134.518,1.336e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.387409,27421
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.13185,30839.8
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,138.94,1.40372e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.37715,27898.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.10309,31718
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.282961,27917.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.317867,27928.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,145.963,1.51736e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.383461,28276.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.49185,29119
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.318853,28260.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,105.218,1.08832e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.427901,28303
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.485443,29176.2
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.321003,28291.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,134.832,1.35957e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.39722,28491.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.12655,33097.8
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.289867,28491.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.335992,28518
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,146.445,1.55361e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.379326,28920.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.470186,29855.4
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.30901,28879.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,104.74,1.10753e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.413009,28878.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.493054,29893.8
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.319953,28887.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,134.504,1.4184e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.388535,29002
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.12097,34281.2
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.280741,29013.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.312927,29013.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,89.6769,970715
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.318693,29168.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.266552,29176.4
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.222482,29187.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,65.4874,708922
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.37139,29184.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.324145,29191.8
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.288875,29203.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,76.1659,822389
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.340567,29199.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.694005,31438.4
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,92.829,1.01024e+06
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.343326,29558.2
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.686396,31858.4
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.272716,29573
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.297176,29546
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,75.9821,838176
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.309234,29787.6
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.302642,29780
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.281067,29780
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,65.6662,724643
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.362333,29791.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.323665,29799.2
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.298039,29791.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,76.1126,839439
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.3462,29799.4
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.683401,32141.6
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.275929,29772.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.28997,29776.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,75.846,849095
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.31347,29984
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.306104,29957.8
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.283838,29965.4
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,66.0305,737580
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.364945,30023.6
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.314545,29985.4
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.290897,29974
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,76.1082,846427
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.351223,29992.6
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.706172,32487.4
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.276217,30000.8
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.300684,30008.6
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,75.8901,854174
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.317476,30127
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.306673,30135
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.289739,30131.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,65.2449,730228
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.364068,30108.6
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.305515,30097.6
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.288747,30101.4
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,76.3106,848727
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.346116,30078.6
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.709289,32642.6
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.274649,30079.2
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.300811,30083.2
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,45.8197,519414
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.282437,30124.8
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.202962,30114.4
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.143506,30110.6
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,53.1664,603932
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.321189,30247.2
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.302674,30216.8
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.263454,30202
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,47.9019,544268
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.291192,30201
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.57128,32795.4
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,64.467,735029
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.301604,28901.4
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.587612,31502.8
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.203679,28870.8
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.164018,28874.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,49.2696,557497
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.29077,30588.8
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.298527,30573.4
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.191788,30558
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,53.0797,614045
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.328044,30683.2
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.319473,30683.2
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.262686,30675.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,48.1063,553127
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.305259,30651.8
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.603971,33337.6
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.203954,30663.4
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.152531,30663.4
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,48.6753,564275
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.280581,30724.2
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.246891,30709
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.183666,30693.8
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,53.0236,616878
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.319953,30778
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.333387,30751.4
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.263384,30705.6
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,48.3071,561323
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.288996,30709.6
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.572592,33418.4
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.187225,30686.8
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.154489,30686.8
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,48.587,565833
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.277841,30830.8
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.245004,30834.6
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.181784,30777.4
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,53.0612,618927
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.318367,30910.2
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.305618,30914
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.262636,30906.4
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,48.1948,560432
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.284606,30850
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.576484,33584.6
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.185843,30857.6
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.150163,30846.2
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,48.7421,569459
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.282002,30933
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.291896,30917.8
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.191571,30917.8
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,53.1103,620979
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.323359,30925.4
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.291473,30925.4
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.270705,30929.2
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,48.2103,560372
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.290801,30887.4
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.587753,33652.4
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.18206,30891.2
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.143545,30891.2
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,48.5894,566127
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.284504,31066.4
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.251953,31043.6
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.193886,31013.2
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,53.0382,619712
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.331025,31081.6
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.335909,31085.4
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.267301,31009.4
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,48.5999,566878
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.291281,30994.2
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.577681,33786.2
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.175231,30967.6
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.144006,30956.2
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,25.0729,306800
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.210501,30990.4
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.152184,30990.4
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.131948,30986.6
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,34.2558,406249
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.267992,30971.4
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.250961,30975.2
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.171935,30956
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.2583,308937
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.257356,30929.6
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.576317,35113.8
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,40.2161,467481
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.278571,31036
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.60317,35208.8
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.133311,31001.8
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.137945,30994.2
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,32.3775,377089
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.249599,31112.2
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.199616,31062.8
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.148269,31017.2
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,34.3474,409987
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.27964,30998.2
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.253714,30998.2
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.17093,30986.8
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.3062,310852
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.261752,30949.2
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.583273,35129.6
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.117695,30941.6
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.140761,30911
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,32.2168,375931
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.232209,30953
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.196683,30953
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.139289,30953
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,34.3335,407522
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.278418,30945.4
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.262277,30930.2
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.166648,30915.4
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.3952,311507
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.267774,30866.2
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.583516,35061.8
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.116473,30858.6
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.134681,30858.6
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.593769,30851
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.22655,37850
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.13328,30854.8
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.72798,119250
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_164.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_164.txt
deleted file mode 100644
index 2d2d833292..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_164.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,196.677,1.87024e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.476209,26619.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.277675,26608.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,6.19781,72892.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.76525,35157.8
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,89.5278,900211
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.294744,26920.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.390743,27472.2
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.205452,26916.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,106.154,1.05704e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.422347,27110.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.482698,27723.8
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.31603,27118.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,136.814,1.35626e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.404299,27421.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.13557,30817
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,140.182,1.42243e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.379813,27902.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.12058,31683.8
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.285457,27879.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.320971,27887
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,148.429,1.54201e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.382916,28299.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.474064,29142.4
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.296631,28315
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,109.015,1.12637e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.425053,28342
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.500887,29230.8
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.30426,28341.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,136.448,1.40933e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.399543,28513.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.12568,33105
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.28876,28532.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.316375,28506.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,148.642,1.57655e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.390923,28846.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.469572,29843.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.289662,28863.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,108.169,1.1371e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.418371,28862
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.488554,29885.2
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.302508,28886
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,136.301,1.43639e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.399812,29028.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.12726,34232.8
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.288171,29040.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.316862,29044
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,90.0136,973729
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.33123,29142.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.31162,29130.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.213529,29138.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,65.3819,706628
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.372875,29187.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.320721,29195.4
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.293732,29206.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,76.56,827587
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.343786,29233
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.695081,31468.4
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,93.1894,1.02024e+06
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.338359,29538.6
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.703132,31835.2
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.264083,29538.6
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.284715,29561.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,76.0143,842060
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.318629,29715.6
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.268901,29704.2
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.278788,29708
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,65.302,719244
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.374111,29751.2
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.315211,29739.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.290341,29712.6
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,76.3842,841344
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.349937,29732
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.700502,32109.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.26583,29755.8
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.280402,29747.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,75.7709,845864
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.313995,29936.4
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.290219,29902.2
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.272855,29913.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,65.287,728406
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.380542,29932.4
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.317131,29905.6
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.28444,29889.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,76.2274,844914
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.340734,29916.8
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.68397,32365.8
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.261394,29913.4
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.284312,29913.8
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,75.998,849458
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.326501,30103.4
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.303333,30107.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.269528,30111
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,65.0639,729206
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.369458,30083.4
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.31539,30095.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.307159,30084
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,76.3686,849666
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.341623,30044.2
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.701186,32547
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.256984,30025.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.296043,30029.6
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,46.024,520112
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.276991,30126.2
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.219787,30080.6
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.145325,30065.6
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,52.8361,604711
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.321949,30205
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.307429,30209
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.252895,30170.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,48.072,546347
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.290636,30174
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.564842,32753
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,64.5646,735721
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.310693,30403.4
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.635202,33032
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.231326,30415.2
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.175717,30392.2
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,48.6423,558770
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.285221,30568.4
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.247161,30553
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.178642,30491.6
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,53.025,611575
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.324209,30625.6
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.31955,30591
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.260997,30579.6
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,48.2043,552201
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.299077,30579.4
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.58257,33234.2
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.18981,30544.8
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.144927,30560
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,48.6939,560116
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.283659,30678.6
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.247857,30678.6
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.182604,30663.2
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,53.0421,614726
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.33557,30740
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.30391,30732.4
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.272332,30736.2
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,48.1014,553523
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.294008,30698.2
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.577411,33395.6
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.183513,30690.6
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.14597,30690.6
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,48.4972,559746
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.283173,30858
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.253855,30827.6
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.186899,30812
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,53.1383,619016
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.332178,30880.4
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.321874,30888
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.258212,30861.4
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,48.2211,555602
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.294341,30834.8
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.580708,33566.4
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.190424,30827.2
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.154386,30834.8
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,48.5539,560401
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.275449,30971.8
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.263436,30956.6
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.191954,30926
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,53.1818,619586
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.332549,30986.4
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.268671,30986.4
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.255704,30971.2
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,48.3366,560682
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.300018,30956
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.591779,33713.4
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.181241,30929.4
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.154585,30914.2
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,48.6633,568122
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.27875,31066.4
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.251985,31020.6
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.193195,31024.4
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,53.1455,624663
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.336222,31081.6
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.267991,31093
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.25706,31036
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,48.1802,562131
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.293195,31009.4
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.585968,33816.6
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.183603,31017
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.147794,30990.4
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,27.1268,330154
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.235679,31024.6
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.166124,30994.2
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.132703,30960
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,42.6636,499208
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.299429,30956.2
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.284715,30956.2
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.209054,30941
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,30.7526,369228
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.287576,30922
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.599606,35118
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,44.7328,529385
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.283288,31043.6
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.603421,35243
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.139148,31028.4
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.141874,31024.6
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,35.9803,434790
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.279314,31078
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.198214,31059
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.139136,31055.2
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,42.597,499959
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.292939,31024.6
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.265554,31024.6
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.20087,30994.2
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,30.6962,369654
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.2883,30941
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.618647,35136.6
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.126662,30933.4
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.141925,30906.8
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,36.0131,431839
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.273458,31005.6
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.199551,30982.8
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.142956,30948.6
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,42.5826,498462
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.313637,30930
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.275902,30930
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.204204,30895.8
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,30.5259,371689
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.286974,30823.4
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.588553,35019
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.123884,30815.6
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.137881,30811.8
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.663177,30811.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.25459,36407.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.201714,30811.8
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,10.1198,123736
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_165.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_165.txt
deleted file mode 100644
index 1b536588b3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_165.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,188.372,1.71448e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.473155,26155.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.304273,26147.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.99152,63538.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.68727,34007.6
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,89.1236,881840
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.326973,26468.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.422308,26993.8
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.240767,26453.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,101.491,994426
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.4217,26629
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.475229,27211.6
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.303339,26629
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,128.452,1.25271e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.397739,26876.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.13777,30203
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,132.541,1.31537e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.380861,27340.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.13031,31002.8
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.284056,27347.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.317233,27332.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,145.481,1.48434e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.387735,27695.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.499287,28504.2
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.294975,27711.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,103.823,1.04714e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.410494,27749.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.476228,28592.4
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.320049,27738
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,128.284,1.29549e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.39569,27859.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.13118,32328.4
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.281304,27909.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.323614,27898
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,145.419,1.51675e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.385681,28263.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.467338,29216.8
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.292882,28282.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,103.807,1.0727e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.414878,28294
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.48474,29282
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.314219,28313.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,128.352,1.32989e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.385592,28414.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.1282,33438.8
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.274436,28392
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.308824,28400
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,89.2635,948562
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.33171,28584
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.295217,28576.2
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.209253,28557
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,64.5101,686757
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.366999,28656.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.318366,28645.2
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.291032,28645.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,72.3554,768390
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.343646,28615
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.681039,30789
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,89.1322,959610
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.343556,28929
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.682837,31141.8
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.272517,28948.2
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.317374,28959.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,75.5839,821413
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.321937,29180.6
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.305092,29150.2
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.276599,29165.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,64.5872,703058
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.370341,29184.4
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.305062,29188.2
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.281272,29180.6
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,72.6249,778701
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.339358,29192.2
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.689993,31454.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.252185,29165.6
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.290046,29169.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,75.2174,827532
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.3145,29374.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.306948,29363.2
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.281905,29355.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,64.7671,710500
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.37162,29390
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.323448,29367.2
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.284018,29359.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,72.5368,789943
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.34243,29367
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.69204,31767
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.269125,29374.8
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.286622,29352
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,75.356,827700
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.32291,29565
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.298994,29538.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.278015,29503.8
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,64.6655,710219
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.369829,29523.2
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.319186,29530.8
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.28476,29538.4
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,72.8163,798187
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.353181,29488.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.705097,31972
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.265644,29515.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.298751,29523.2
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,45.0911,503591
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.281752,29576.6
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.249989,29576.6
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.145644,29534.6
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,51.5279,573066
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.315909,29634
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.267244,29634
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.255909,29603.4
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,46.1579,513194
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.291653,29664.4
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.574384,32140.2
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,62.2278,696063
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.301413,29844.6
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.577495,32385.2
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.19797,29840.8
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.161215,29844.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,47.0515,539890
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.2881,29997.4
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.230546,29982.2
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.156601,29986
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,51.2828,576633
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.320779,30053.2
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.266961,30061
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.26389,30053.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,46.3387,527558
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.291461,30025
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.565687,32611.4
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.161196,30028.8
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.149298,30028.8
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,47.1128,544105
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.287774,30127.8
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.225566,30135.8
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.153247,30113
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,51.2804,582201
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.317784,30160.6
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.267724,30133.8
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.259391,30137.6
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,46.2918,529324
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.283877,30082
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.540048,32691
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.153772,30085.6
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.149535,30089.4
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,46.8508,541364
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.294667,30269.4
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.230718,30253.4
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.158949,30222.6
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,51.5104,580771
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.332625,30299.8
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.26924,30299.8
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.261963,30285.4
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,46.3933,529648
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.288082,30235.8
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.557968,32921.4
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.159653,30243.4
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.149817,30243.4
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,47.1085,546242
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.288529,30283.8
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.221574,30283.8
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.149664,30253.6
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,51.2349,585475
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.316498,30368.6
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.27082,30372.4
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.256293,30314.8
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,46.4799,537067
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.300236,30300.4
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.560516,33009.2
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.146003,30269.6
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.148434,30255
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,46.9077,546948
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.28963,30367.6
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.212139,30371.4
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.146431,30352.8
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,51.3607,585525
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.318878,30435.2
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.263941,30420.4
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.256395,30424.2
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,46.4972,533782
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.318559,30361.6
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.567991,33085.4
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.165996,30346.2
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.160313,30350.6
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,25.2471,303601
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.210079,30360
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.14661,30344.6
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.131001,30337.6
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,34.9345,400778
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.268818,30346.4
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.259998,30285
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.18213,30285.6
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.4139,306926
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.256146,30274.2
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.562986,33013.4
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,40.2054,459381
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.286482,30421.2
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.619344,34530
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.130047,30421.6
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.144262,30421.6
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,32.2646,368114
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.244337,30474
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.195404,30473.6
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.136678,30455.4
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,35.0359,400855
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.273394,30447.6
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.268907,30451.4
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.186054,30428.4
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.346,306268
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.254462,30410.6
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.581475,34500.2
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.114943,30391.4
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.139059,30376.2
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,32.1895,367568
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.234072,30429.4
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.20348,30429
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.137522,30395.2
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,34.8227,401303
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.269688,30375.2
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.273508,30344.6
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.188971,30310.6
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.4224,308667
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.256357,30299.2
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.574813,34362.8
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.114502,30300.2
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.13255,30300.2
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.575638,30300.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.22007,35714.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.206911,30300.6
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.78289,117079
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_166.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_166.txt
deleted file mode 100644
index 36738891fb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_166.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,186.531,1.73641e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.472995,26172.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.296203,26176
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,6.19001,73155
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.73582,34498.6
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,87.1731,857608
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.317855,26521
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.407838,27023
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.216542,26482.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,102.118,1.00207e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.422244,26643
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.468829,27240.4
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.313009,26654
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,129.491,1.26462e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.388964,26871.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.13265,29518.8
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,132.837,1.32714e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.38195,27347.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.10837,31048.4
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.288165,27347.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.32444,27355.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,145.465,1.48396e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.380254,27714.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.457232,28527.4
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.299646,27738.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,104.739,1.06261e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.422225,27753.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.48312,28592.4
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.318654,27788.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,128.754,1.31054e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.38698,27925.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.10193,32413.4
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.275922,27944.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.3259,27936.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,145.581,1.51079e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.380458,28271.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.464489,29198.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.309118,28275.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,104.308,1.07686e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.411249,28306
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.472208,29267.4
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.303307,28290.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,128.739,1.32451e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.385213,28399.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.10737,33443.6
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.285868,28392.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.317791,28393
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,89.453,943521
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.316389,28522.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.325982,28519
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.226008,28519
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,64.9095,691356
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.361393,28583.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.31939,28591.4
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.292069,28614.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,73.07,776023
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.339998,28626.6
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.687497,30804.4
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,89.8433,962539
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.332779,28929
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.700239,31149
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.287896,28929
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.318654,28932.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,75.8146,823152
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.315422,29138.4
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.298263,29127
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.31722,29134.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,65.0815,706067
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.369649,29119.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.313534,29123.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.293265,29127.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,73.0975,788785
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.347544,29104.4
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.702179,31431.6
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.272088,29112
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.303992,29100.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,75.1378,818091
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.312831,29313.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.298264,29283.4
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.276459,29283.4
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,65.2638,710267
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.370187,29333
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.307717,29279.8
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.284728,29295
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,72.9903,789893
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.336656,29336.4
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.685218,31740.2
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.280985,29329
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.306988,29298.6
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,75.3961,826425
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.321169,29477.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.304811,29485.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.283812,29492.8
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,65.0191,714935
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.362769,29492.2
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.311806,29496.2
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.288383,29488.6
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,73.2684,799801
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.346448,29461.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.693301,31910.6
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.268588,29446.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.303103,29435.2
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,44.9567,502652
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.279147,29587.6
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.220191,29591.4
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.143871,29557.2
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,51.3347,569372
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.306546,29587.6
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.293592,29591.4
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.262104,29576.2
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,46.1498,509793
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.2878,29603.2
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.613162,32086
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,62.5655,696593
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.30636,29840.2
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.631151,32408
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.196825,29829
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.160723,29829
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,47.1387,539387
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.28378,29979.8
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.259077,29945.8
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.15605,29912
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,51.2383,577583
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.315172,30036
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.302309,30024.8
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.262629,30013.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,46.2642,519585
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.275941,29959.4
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.607644,32549.6
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.177874,29974.6
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.152684,29959.6
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,46.981,541843
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.275115,30103.4
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.255705,30092
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.148959,30073.2
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,51.1828,581314
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.316741,30136.4
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.30243,30140.2
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.267153,30106.4
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,46.6132,524526
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.289227,30084.8
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.6156,32724.8
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.174668,30096.2
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.154962,30096.2
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,47.0275,544213
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.285291,30235.6
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.282213,30228.4
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.156492,30198.6
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,51.4135,582294
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.315224,30258.8
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.303787,30247.2
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.261234,30236.6
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,46.4337,526768
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.291122,30218
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.604528,32896
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.168082,30210.2
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.147993,30210.2
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,47.1397,547582
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.295639,30299.6
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.259371,30296.4
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.152716,30300.8
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,51.1781,582560
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.311326,30349.8
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.299928,30342.2
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.272689,30331.2
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,46.4606,526539
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.282366,30263.8
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.610249,32968.2
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.173509,30270.8
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.149471,30275.2
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,47.0671,544470
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.291358,30406.6
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.285246,30364.6
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.170757,30338.4
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,51.1047,582816
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.309451,30432
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.301649,30417.2
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.258238,30402
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,46.442,521959
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.296433,30369.6
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.623152,33112.6
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.182432,30388.6
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.157324,30396.2
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,25.1297,305493
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.215115,30418.4
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.19477,30403
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.131859,30373.4
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,35.1078,402907
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.266361,30362.8
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.25276,30362.8
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.192268,30337.6
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.5895,307234
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.25669,30304.6
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.618455,34386.6
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,40.2912,459228
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.269497,30377.2
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.644354,34463.2
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.130911,30373.4
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.143628,30361.8
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,32.3754,373876
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.2286,30469
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.183007,30469
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.136134,30446
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,35.0472,403166
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.279384,30442.4
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.313585,30427.2
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.185727,30385.8
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.4644,306032
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.252465,30359.2
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.603957,34399.6
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.118143,30355.4
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.139922,30356
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,32.2744,371834
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.233022,30389.6
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.251192,30389.6
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.14085,30375
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,34.7799,401179
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.266322,30386.4
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.296485,30386.4
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.184664,30363.4
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.4061,305656
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.257669,30332.8
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.612023,34342.6
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.117856,30302.2
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.133714,30298.4
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.579619,30299
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.22218,37033
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.129746,30299
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,10.0076,118690
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_167.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_167.txt
deleted file mode 100644
index 023a48c70f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_167.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,185.747,1.72403e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.471019,26141.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.26005,26149.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,6.07212,68952.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.69734,34508
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,88.6917,884746
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.265528,26457
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.370738,27031.4
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.202348,26498.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,101.997,1.00071e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.41624,26636
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.479319,27191.6
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.3089,26597.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,128.724,1.25154e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.392318,26822.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.1251,30137
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,132.441,1.31389e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.376164,27260
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.12227,30923.4
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.271704,27290.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.310385,27286.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,144.94,1.4778e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.377586,27684.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.46849,28489
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.298512,27692.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,105.018,1.0657e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.430609,27711.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.473853,28550.4
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.310654,27726.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,128.544,1.29726e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.395645,27902.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.12197,32353.2
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.27923,27921.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.313591,27902.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,145.157,1.51043e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.379575,28217.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.459991,29167.4
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.312914,28202.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,104.785,1.07461e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.418263,28195
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.479626,29171.4
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.316997,28221.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,128.35,1.32167e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.385144,28369.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.10238,33424.4
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.27699,28373.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.306655,28373.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,89.3076,940678
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.363455,28549.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.315364,28523
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.231301,28507.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,64.8025,691325
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.366897,28569
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.310392,28569
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.283762,28572.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,72.495,768189
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.345156,28580.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.689026,30720.2
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,89.2446,942883
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.337899,28868.2
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.704893,31069.2
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.271442,28879.8
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.297202,28887.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,75.3221,816187
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.307319,29070.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.29477,29085.4
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.277426,29077.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,64.9142,699353
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.365962,29100.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.317477,29104.4
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.285631,29104.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,72.4936,785355
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.337707,29097
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.694115,31393.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.273796,29112.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.283633,29120
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,75.1319,822406
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.325278,29253
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.308408,29256.8
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.279467,29268.2
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,65.1364,708002
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.363249,29268.2
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.305189,29241.6
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.297003,29245.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,72.6949,789732
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.334917,29234
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.701423,31611.2
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.266379,29253.2
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.272082,29238
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,74.8667,824538
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.315851,29412.8
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.288152,29420.4
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.274187,29420.4
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,64.6983,710583
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.375128,29424.4
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.309842,29428.2
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.287973,29428.2
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,73.0753,796689
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.345035,29378.2
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.688413,31815.6
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.269138,29397.4
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.305291,29401.2
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,45.7878,504237
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.277157,29462
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.214104,29465.8
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.151519,27943.8
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,52.7513,592082
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.31692,29588
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.300306,29550
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.256153,29534.8
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,46.5979,529228
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.297086,29584.2
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.591689,32067.4
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,63.8081,704183
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.307607,29783.6
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.629724,32308.8
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.184102,29779.8
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.152985,29783.6
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,48.4794,540939
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.275365,29909.8
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.291256,29894
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.198367,29879
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,53.0634,600325
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.32613,29984.4
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.296703,29961.4
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.261643,29965.2
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,46.6864,534998
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.291966,29865.8
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.602704,32421.8
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.156888,29874.2
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.151013,29874.2
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,48.6048,547595
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.276516,30055.6
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.291294,30063.2
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.181881,30029.6
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,52.6883,601634
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.318929,30081
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.304158,30081
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.263404,30073.6
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,46.7287,538758
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.300791,30031.6
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.611485,32648.4
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.153561,30031.6
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.149715,30039.2
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,49.0322,554821
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.277599,30148.8
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.280415,30126.4
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.181196,30107.2
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,52.8603,603790
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.320542,30233
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.293023,30206.2
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.269215,30175.6
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,46.925,540495
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.287019,30115.8
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.634186,32748
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.155859,30085
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.145651,30088.8
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,48.6358,547277
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.279313,30271.2
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.279646,30275.4
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.188864,30255.6
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,53.0178,606910
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.333694,30318
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.306481,30278.8
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.264049,30283.2
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,46.947,545247
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.294232,30212.6
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.616201,32890.6
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.156313,30216.4
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.151301,30216.4
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,48.6415,555286
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.27646,30347
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.303135,30335.4
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.18398,30317.4
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,52.8918,608811
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.319307,30351.6
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.296177,30359.2
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.264017,30340.8
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,46.8842,543660
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.292664,30303.4
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.596572,33016.6
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.155877,30307.8
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.146079,30292.4
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,25.1613,304601
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.201387,30342
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.200901,30338.2
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.132159,30319
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,34.9671,400138
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.261586,30318.4
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.253177,30319
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.18231,30300
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.4331,306582
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.261829,30242.4
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.617053,34244
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,40.356,459774
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.275997,30310.4
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.642492,34312
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.135647,30295
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.139512,30310.4
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,32.2942,366911
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.239615,30425.6
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.254335,30410.2
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.13607,30375.6
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,34.9716,401133
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.281458,30371.8
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.332223,30357.2
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.181638,30338.4
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.4197,305544
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.259352,30284.4
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.611081,34286
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.117739,30265.2
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.145567,30261.4
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,32.298,372522
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.234175,30334
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.227308,30334
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.135481,30281.2
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,34.8162,400368
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.266527,30307
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.261023,30276.4
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.183808,30261
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.5108,308139
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.260242,30250.2
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.614461,34255.8
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.112313,30238.8
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.1288,30224
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.57665,30224
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.23413,36922.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.128595,30227.8
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.89324,118435
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_168.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_168.txt
deleted file mode 100644
index 22ccbe9e11..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_168.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,188.566,1.72179e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.476157,26048.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.277209,26067.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,6.07835,71442
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.74726,34453.4
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,87.8974,858826
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.293349,26414.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.393579,26950.8
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.21299,26414
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,103.891,1.01659e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.422737,26536
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.491523,27122.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.303857,26535.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,129.64,1.26051e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.395896,26758.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.12807,30100
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,132.752,1.3201e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.389463,27282.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.12683,30922.6
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.282873,27271.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.308677,27259.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,146.611,1.48933e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.391928,27645.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.491434,28458
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.291953,27657.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,106.788,1.08451e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.412445,27661
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.485911,28480.8
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.302098,27661
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,129.827,1.31423e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.406795,27925.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.13064,32406.2
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.276991,27887
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.320983,27894.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,146.885,1.51999e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.387735,28206.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.465821,29141.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.318264,28202.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,107.088,1.08664e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.411025,28210.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.477092,29195
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.302456,28226
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,129.694,1.33681e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.386973,28368.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.11293,33424
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.278412,28342.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.309688,28354.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,88.672,937616
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.338763,28507.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.306687,28504
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.211519,28511.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,64.9858,691893
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.368126,28557
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.330405,28545.4
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.286629,28553
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,72.428,760826
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.354264,28534.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.691458,30704.6
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,88.6422,948825
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.338571,28833.8
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.697097,31034.8
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.274552,28833.6
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.297246,28841.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,74.6706,814054
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.31804,29070.4
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.296645,29059
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.287986,29043.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,65.3855,705922
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.374193,29104.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.314571,29059
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.284408,29051.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,71.7419,768530
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.335006,29070.6
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.681648,31363.6
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.27397,27590.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.298001,27579
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,74.6253,821510
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.321175,29253
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.269944,29256.8
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.276108,29260.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,65.3408,710892
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.372292,29264.6
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.281131,29253.2
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.287416,29264.6
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,72.0366,778906
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.341937,29215.2
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.674518,31622.6
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.260031,29230.4
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.286456,29238
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,74.6093,821774
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.317183,29386
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.269707,29393.6
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.285002,29397.4
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,65.2614,713076
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.368433,29389.8
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.277072,29386
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.28531,29378.4
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,71.9477,781421
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.335922,29371
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.651632,31808.4
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.267973,29329.2
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.289764,29336.8
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,45.4726,481415
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.287691,29447
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.200985,29420.4
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.150092,29405.2
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,53.2107,592672
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.32549,29576.4
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.279749,29580.2
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.269662,29553.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,46.6923,522682
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.290303,29519.2
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.579369,32025.4
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,63.1366,704768
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.308849,29714.2
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.593833,32231.6
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.199467,29698.8
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.174399,29702.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,48.305,541769
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.290411,29875.8
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.256363,29875.8
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.181797,29860.4
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,52.7721,584470
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.322449,29882
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.279621,29890
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.253496,29870.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,46.7327,527628
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.290014,29820.8
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.573135,32422.4
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.169765,29836.2
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.144825,29836.2
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,48.1901,540975
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.277035,29992.2
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.255377,29954.4
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.188799,29954.2
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,52.7753,600403
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.322046,30014.6
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.265637,30018.6
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.264363,30011.2
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,46.8435,526901
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.309361,30002.2
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.605398,32630.4
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.209478,29990.6
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.156985,29990.6
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,48.1863,546523
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.283237,30112.8
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.260235,30097.6
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.181688,30066.8
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,52.7346,603205
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.327588,30191
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.269496,30191
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.256793,30175.8
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,46.5578,526391
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.328593,30146.6
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.580381,32817
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.187378,30139.6
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.15941,30139.6
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,48.2664,547531
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.282457,30240.2
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.253528,30209.8
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.193323,30214
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,52.7063,606921
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.33267,30259.6
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.27267,30262.8
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.258763,30252.6
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,46.5046,534309
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.287179,30203.6
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.571075,32904.6
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.163417,30199.6
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.151884,30199.6
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,48.3282,550677
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.28186,30335.8
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.255992,30305.2
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.186604,30298
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,52.853,611652
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.325215,30397.4
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.273682,30396.8
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.253023,30351.2
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,46.492,539149
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.292088,30322.6
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.559888,33050.4
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.161836,30311.2
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.144857,30299.6
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,27.3874,327502
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.235115,30345
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.171104,30329.6
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.134655,30299.8
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,43.021,489304
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.297919,30288.4
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.27731,30292.2
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.202142,30288.4
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,30.6324,364050
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.27772,30235.4
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.588975,34340.2
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,44.8337,517828
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.286155,30387.8
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.614268,34504.8
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.145157,30384.6
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.14967,30377
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,36.8878,431836
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.284741,30452.2
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.207519,30433.2
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.147077,30395
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,42.6788,489156
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.291198,30395.6
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.281867,30365
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.207058,30350.2
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,30.8938,363335
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.280722,30289.8
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.611113,34406.2
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.122841,30278.4
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.144153,30278.4
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,36.2201,424233
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.285489,30330.8
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.207006,30316
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.146872,30281.6
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,43.0324,492689
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.291787,30294.2
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.268107,30294.2
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.203973,30286.6
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,30.8337,356518
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.288017,30226
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.597366,34315.8
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.120351,30191.6
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.134489,30187.8
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.635638,30187.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.21791,37035.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.12871,30187.8
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.91135,116665
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_261.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_261.txt
deleted file mode 100644
index 45af9a2477..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_261.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,127.839,1.1273e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.390557,25144
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.309445,25132.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.74866,58302.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.72059,32929
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,74.46,711929
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.237394,25502
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.324882,25505.8
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.189132,25505.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,99.7869,948743
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.326603,25778
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.361144,26318.6
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.200486,25739.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,47.8934,455092
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.27861,25800.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,0.887042,28675.4
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,59.7064,572686
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.288888,25987.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,0.912929,29057.4
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.141235,26002.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.161388,26022.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,135.055,1.32036e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.307807,26435
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.405905,26442.6
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.28835,26481.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,101.363,992393
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.355832,26630.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.468689,27373.6
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.305099,26618.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,48.2533,468940
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.295826,26660.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,0.899457,30410.4
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.133382,26645
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.165068,26653
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,134.672,1.30908e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.315544,27066.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.398058,27043.6
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.277278,27043.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,101.475,1.01798e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.356894,27158
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.459325,27993.4
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.303762,27127.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,48.2293,478325
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.31059,27147
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,0.898324,31506.8
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.132204,27140.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.16446,27152.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,79.656,814898
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.27118,27297.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.198412,27297.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.151481,27301.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,60.0469,613749
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.310449,27427.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.265694,27400.6
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.295269,27412.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,28.0435,290475
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.25594,27328.2
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.482314,29181.4
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,46.4545,477555
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.279032,27568.4
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.545572,29468
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.131814,27587.6
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.159692,27591.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,68.7911,720711
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.281758,27808.6
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.253868,27778.2
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.208249,27782
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,59.9206,624912
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.309457,27885
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.300414,27873.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.280287,27889
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,27.7967,291597
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.266194,27843.2
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.434398,28858
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.122822,27858.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.15829,27846.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,68.8088,726603
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.282027,28046.2
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.255922,28034.6
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.204447,28019.2
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,60.0078,631663
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.313067,28084.4
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.296215,28092.2
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.281343,28100
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,27.8042,287104
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.278226,28065.4
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.496516,30160
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.125311,28077
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.154796,28084.6
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,68.9851,734403
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.281355,28252.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.256722,28252.6
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.210597,28267.8
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,60.0106,636585
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.301823,28316.8
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.304267,28286.2
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.28035,28298
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,27.5868,290160
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.265406,28240
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.503703,30441.6
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.12485,28259.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.154482,28275.4
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,38.9478,425707
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.221689,26868
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.141222,26852.8
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.133574,26849
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,42.3895,454461
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.296184,28305.8
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.260588,28305.8
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.227627,28305.8
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,19.4529,221173
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.222508,28255.4
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.457092,30464
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,45.522,483025
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.298673,28366.4
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.639824,30616.8
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.217945,28370.4
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.178559,28374.4
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,40.6044,438384
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.254968,28484.8
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.189906,28469.4
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.142437,28450.2
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,42.1817,457021
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.298104,28517.6
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.29011,28502.4
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.22421,28452.6
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,19.2937,218164
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.212351,28398.4
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.453104,30687.2
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.112415,28402.2
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.134616,28402.2
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,40.2507,430708
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.241778,28569.8
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.206706,28569.8
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.143916,28554.6
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,42.0704,458056
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.292549,28539.4
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.307762,28543.2
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.224198,28543.2
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,19.3591,223505
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.208376,28452
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.441463,30775
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.122802,28459.6
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.131224,28459.6
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,40.2674,434921
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.252927,28551.2
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.2022,28535.8
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.143827,28535.8
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,42.7277,461822
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.305886,28505.4
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.258526,28505.4
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.231718,28490.2
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,19.3671,224635
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.21292,28463.4
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.440369,30843.8
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.116537,28463.4
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.137433,28463.4
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,40.3751,433192
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.243282,28559.2
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.206021,28544
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.142706,28536.2
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,42.1652,455575
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.301412,28528.8
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.255857,28536.4
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.219116,28471.4
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,19.3949,221535
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.220082,28410.6
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.448932,30821.8
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.132511,28410.6
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.149229,28414.4
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,40.4039,431447
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.251909,28559.4
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.202962,28555.6
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.145484,28540.4
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,42.3478,458456
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.295544,28536.6
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.272056,28517.4
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.232939,28506
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,19.4228,224641
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.209586,28460.2
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.442936,30901.8
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.112243,28460.2
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.132518,28467.8
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,25.5804,284918
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.251302,28475.4
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.163615,28475.4
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.138957,28460
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,34.0129,372945
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.288134,28479.2
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.244435,28475.4
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.156217,28425.8
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,22.5114,257579
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.260402,28341.8
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.565206,31970
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,35.341,372860
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.278604,28464.2
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.65702,32126.6
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.168146,28464.2
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.147762,28456.6
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,33.0313,367602
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.236991,28514
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.16494,28514
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.136096,28487.2
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,34.0946,373873
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.281483,28521.8
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.242777,28506.6
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.160569,28491.2
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,22.38,257012
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.262552,28430.2
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.571095,32084.8
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.11632,28422.4
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.131929,28422.4
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,33.208,370059
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.234508,28545
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.156902,28545
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.138726,28472.2
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,34.2549,374759
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.287737,28472.2
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.242924,28472.2
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.173835,28468.4
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,23.2326,262480
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.259678,28422
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.576547,32069
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.120069,28391.4
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.133849,28376
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.577596,28391.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.18945,33290
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.179046,28406.8
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.66568,110034
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_262.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_262.txt
deleted file mode 100644
index afd86204a8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_262.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,130.974,1.14388e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.422128,25056.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.337956,25041.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.73444,55571.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.73878,32802.6
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,74.7484,714104
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.223724,25414.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.294456,25410.4
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.169944,25429.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,99.5155,942655
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.313055,25632.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.361784,26169.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.182188,25659
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,47.4843,447264
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.287033,25689.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,0.934472,28579.8
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,59.9335,548541
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.314692,25938
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,0.982721,29015
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.150764,25945.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.180946,25933.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,135.028,1.30745e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.313912,26396
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.447402,27066.4
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.298123,26376.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,100.959,984439
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.3537,26570.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.458276,27294.4
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.296287,26570.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,48.0152,459941
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.310622,26535.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,0.960801,30278
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.132275,26527.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.157625,26547
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,134.782,1.34262e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.308229,26962.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.396177,26970.6
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.28124,26963.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,100.934,1.00789e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.348401,27108.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.458417,27951.4
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.2902,27104.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,48.288,476754
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.315006,27116
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,0.901346,31483.2
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.134297,27120.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.160998,27106
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,79.7201,815316
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.261189,27267.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.212888,27256.4
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.154655,27256.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,59.7785,609881
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.302864,27382
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.263519,27374.4
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.271416,27359.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,28.0965,286179
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.26293,27305.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.494506,29170.2
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,46.4392,475503
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.278738,27541.2
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.525149,29464.4
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.132838,27545.8
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.157496,27572.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,69.0729,721180
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.284684,27732.4
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.259467,27724.6
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.217272,27728.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,61.6447,620784
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.307454,27800.2
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.264792,27777.4
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.275378,27781.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,27.893,287881
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.273906,27742.8
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.499799,29764.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.126259,27746.8
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.148768,27750.8
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,68.6433,724543
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.284261,27931
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.257458,27931
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.209068,27931
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,60.311,638600
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.310737,27983.4
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.264466,27995.2
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.288504,27964.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,27.7523,289975
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.265215,27929.4
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.49457,30031.2
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.12334,27933.6
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.147026,27933.6
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,68.6011,729542
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.28074,28133.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.263902,28133.6
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.21548,28122.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,60.1866,639645
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.311613,28151
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.263928,28155
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.283538,28174.2
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,27.8113,293495
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.266328,28112.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.489245,30268
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.122034,28116.8
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.149829,28120.8
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,38.5317,421434
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.217446,28191.2
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.147967,28191.2
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.129183,28175.8
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,42.097,451424
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.305055,28197.8
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.259896,28197.8
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.230398,28197.8
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,19.3353,222079
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.217471,28124.6
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.436817,30356
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,45.8446,486902
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.306315,28269.6
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.627261,30497
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.219947,28247
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.184434,28247
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,40.6316,435687
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.26211,28414.6
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.199538,28399.2
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.154303,28315.2
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,42.3889,455434
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.301105,28398.2
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.262699,28375.6
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.231858,28348.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,19.5206,220360
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.214131,28272.6
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.436779,30561.4
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.113344,28280.2
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.131218,28280.2
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,40.2284,430595
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.241055,28413.4
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.206924,28413.4
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.145798,28394.2
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,42.3943,456062
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.318206,28409.4
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.265957,28394
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.235448,28375
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,19.4523,224902
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.211768,28303
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.421264,30622.2
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.107852,28303
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.130757,28303
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,40.3351,432197
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.255058,28447.8
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.208414,28436.2
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.145253,28440
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,42.2895,455228
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.294494,28463.4
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.260152,28467.2
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.232075,28406.4
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,19.5995,226839
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.214444,28314.8
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.442001,30664.8
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.127622,28318.6
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.151276,28322.4
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,40.4216,432729
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.251001,28475
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.199999,28475
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.140882,28440.6
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,42.4534,458259
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.306072,28418
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.257426,28414
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.231122,28410.2
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,19.3812,223851
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.210143,28334.2
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.439582,30714.6
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.11015,28334.2
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.133849,28318.8
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,40.2117,426849
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.24325,28471.6
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.204114,28425.6
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.136326,28394.8
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,42.3894,456951
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.299633,28471.8
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.264127,28468
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.253753,28441
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,19.6324,225486
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.220683,28384.2
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.439492,30799.2
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.110009,28388
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.133113,28391.8
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,25.5856,285321
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.250406,28414
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.160038,28414
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.133542,28395.6
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,34.1475,373680
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.287411,28391
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.306469,28394.8
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.15637,28383.4
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,22.6132,257441
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.271735,28334.6
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.572727,31947.6
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,35.4462,373580
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.277727,28399
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.656117,32080.4
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.168613,28414.2
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.143622,28414.2
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,33.0397,369920
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.24115,28499
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.161759,28495.2
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.130751,28472.2
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,34.2683,374324
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.284798,28445.2
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.299531,28445.2
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.163064,28426.2
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,22.5977,257851
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.264248,28399.4
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.56959,32031
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.116518,28368.6
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.132921,28368.6
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,33.1562,369473
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.239724,28476.2
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.159667,28472.4
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.128787,28445.6
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,34.2024,373938
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.291109,28426.4
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.299448,28411.2
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.160889,28399.6
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,22.6335,257597
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.270763,28353.6
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.574755,32012
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.116223,28345.6
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.137164,28322.4
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.582128,28318.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.233,34445.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.130976,28345.6
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.97631,111206
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_263.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_263.txt
deleted file mode 100644
index 49071a42cb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_263.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,149.5,1.34458e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.40076,26003.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.24693,25988.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.77476,64404.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.73732,33446.4
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,77.2371,765319
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.198579,26286.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.294085,26800.2
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.146694,26297.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,145.879,1.42331e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.343147,26578.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.364849,27180.4
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.207563,26601.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,51.4426,490073
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.324177,26643.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,0.978592,29743.4
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,62.5683,622236
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.298718,26829
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,0.979239,30056.2
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.166489,26840.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.160242,26833
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,141.113,1.42576e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.331185,27225.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.454456,27964.6
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.2894,27255.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,147.068,1.48146e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.379978,27486.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.474256,28279.4
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.304568,27501.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,51.8237,510629
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.324082,27512.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,0.976686,31587.8
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.165004,27535.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.168773,27527.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,140.914,1.45471e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.336215,27860.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.447492,28734
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.301944,27837.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,147.121,1.51427e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.375141,28005.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.474769,28920.4
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.30844,28016.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,51.9479,520379
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.327582,27978.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,0.988404,32647
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.164019,28001.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.161324,28005.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,81.8219,866675
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.28908,28127.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.242873,28105.4
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.153894,28113.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,86.0812,904262
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.324715,28245.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.312453,28260.8
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.275416,28268.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,31.47,332524
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.275685,28242
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.537712,30263.6
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,51.624,544118
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.280056,28479.8
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.64241,30513
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.169286,28456.6
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.160383,28460.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,73.1071,779919
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.287967,28678.2
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.296549,28670.8
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.269297,28659.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,86.6251,921348
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.329214,28742.8
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.317463,28742.8
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.278674,28754.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,31.3964,339929
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.277074,28731.6
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.529405,30898.4
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.12142,28727.8
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.147974,28731.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,73.1379,786932
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.288555,28959.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.290929,28944.4
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.263877,28921.6
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,87.0549,938136
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.332466,29074
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.268152,29016.8
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.288913,29013
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,31.2147,342825
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.268626,28940.6
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.486468,31191
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.12398,28952
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.150303,28955.8
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,73.0711,793528
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.283372,29146.4
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.261912,29146.4
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.258021,29142.6
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,86.7318,942908
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.325195,29192.6
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.266936,29196.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.283819,29204
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,31.0783,343405
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.276383,29139.2
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.495326,31458.6
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.124857,29127.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.148978,29139
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,40.1748,440482
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.252446,29196
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.183308,29196
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.137574,29177
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,56.259,614735
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.311775,29181
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.269336,29184.8
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.267,29188.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,22.2687,257834
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.234181,29142.8
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.471145,31515.6
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,50.5911,557605
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.317796,27888.4
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.583344,30249.6
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.201195,29303.4
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.173586,29276.4
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,43.7502,492967
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.262661,29430
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.214886,29414.6
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.148608,29395.6
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,56.8286,623161
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.310085,29457
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.268651,29411.2
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.267301,29395.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,22.2759,263969
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.224165,29327
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.474507,31749.6
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.126534,29304
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.138168,29304
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,43.2961,488922
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.271308,29441.6
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.213631,29437.8
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.147737,29403.4
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,56.4881,619043
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.31285,29476.4
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.263096,29461.2
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.256626,29445.8
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,22.2316,261121
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.220473,29369.2
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.451178,31818.8
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.115801,29357.8
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.133772,29357.8
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,43.6382,494029
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.288281,29491
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.214949,29495
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.147142,29487.2
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,56.7421,626686
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.314667,29437.6
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.266501,29445.2
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.258879,29426
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,22.2647,262846
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.226443,29384
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.445085,31863.2
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.114329,29357.2
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.132531,29361
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,43.6635,495499
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.273048,29464.2
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.207397,29464.2
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.155647,29449
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,56.4914,623114
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.305144,29476.2
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.273835,29483.8
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.258853,29437.8
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,22.3705,264427
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.220735,29422.4
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.444414,31932.4
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.119622,29407.2
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.134738,29411.2
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,43.5052,488218
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.265599,29476.4
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.213849,29480.2
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.153183,29472.6
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,56.6067,626227
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.31043,29482.4
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.261662,29475.2
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.255742,29475
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,22.3768,262976
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.225106,29424.6
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.439358,31969
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.11017,29413.4
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.13687,29413.4
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,26.8292,297098
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.260844,29460.8
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.197938,29457
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.138476,29433.6
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,43.6757,482681
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.296977,29452.2
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.274373,29448.4
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.193957,29448.4
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.4725,296284
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.279531,29371.4
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.624809,33200.8
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,39.504,445501
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.305003,29551.4
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.638378,33380.8
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.142706,29547.4
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.155276,29528
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,35.9164,413756
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.283415,29613.2
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.199179,29609.4
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.13806,29570.8
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,43.6754,487696
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.299723,29596.6
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.267812,29600.4
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.197784,29566
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.3593,297445
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.276024,29485
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.614115,33318.4
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.125299,29485
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.141888,29485
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,35.9393,414572
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.278231,29573.4
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.197785,29565.4
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.138502,29500.2
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,43.6676,482678
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.293476,29592
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.270846,29592
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.204011,29530.8
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.5382,298627
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.283295,29458
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.622243,33276.2
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.135679,29442.8
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.134899,29442.8
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.616099,29439
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.27162,35828
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.129593,29442.8
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.76313,113945
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_264.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_264.txt
deleted file mode 100644
index ed370c72a3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_264.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,148.994,1.32122e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.395691,26011.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.32693,26026.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.7306,59843.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.84757,34416.6
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,77.9006,775224
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.225375,26349.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.302373,26874.4
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.15287,26368.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,145.387,1.40239e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.343045,26638.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.357355,27232.8
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.214271,26657.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,52.2534,510934
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.324337,26680.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,0.93906,29819
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,62.7484,618697
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.296255,26932.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,0.962548,30235.2
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.184032,26932
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.174713,26916.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,141.241,1.41926e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.338661,27377.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.454379,28094
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.309047,27362.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,147.711,1.48472e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.378155,27559
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.47848,28321.2
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.350666,27535.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,52.4503,506217
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.33299,27504.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,0.944763,31548.8
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.160838,27477.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.16903,27489
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,141.727,1.45782e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.334065,27913.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.458654,28802.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.298713,27917.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,147.67,1.5201e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.383998,28128
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.465149,29024
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.305655,28097
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,52.2856,523803
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.324471,28055
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,0.974343,30865.6
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.176921,28032
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.168172,28016.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,81.8246,869741
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.285867,28171.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.205734,28171.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.15872,26745.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,86.2508,906218
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.336004,28356.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.311959,28356.6
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.294302,28356.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,31.6299,334714
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.277438,28337.4
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.49089,30351.4
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,51.4024,542566
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.274891,28564.2
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.58534,30578.2
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.174252,28510.8
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.170354,28514.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,73.1819,781119
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.281784,28790.6
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.261529,28760.2
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.278866,28784
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,86.5602,928592
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.33219,28898.8
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.312721,28876
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.302034,28872
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,31.6048,339298
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.276332,28807.6
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.507127,30989.6
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.12759,28826.6
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.161535,28819
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,73.447,793508
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.299147,29043.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.265803,28982.8
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.2763,28986.8
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,86.6137,937295
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.328471,29074
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.303256,29062.6
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.29006,29074
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,31.5062,341941
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.280708,29005.6
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.497968,31283
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.131948,29017
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.156505,29024.6
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,72.909,795243
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.284844,29249
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.261714,29241.4
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.262513,29233.8
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,86.6832,942072
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.320683,29265.2
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.3011,29272.8
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.288933,29280.4
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,31.3662,347207
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.27564,29203.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.492195,31523.2
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.123621,29180.8
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.161791,29180.8
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,40.2632,444370
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.235896,29291
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.183474,29264.4
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.142968,29249.2
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,56.4478,620816
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.309157,29341.4
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.269457,29303
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.26563,29291.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,22.1958,260134
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.22718,29265
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.451729,30451.4
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,50.4672,560744
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.308011,29391.4
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.599286,31787.2
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.234744,29387.4
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.187577,29375.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,43.355,490122
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.26444,29525.2
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.221817,29525.2
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.158886,29525.2
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,56.7003,627833
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.311109,29537
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.338418,29540.8
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.261137,29514
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,22.3318,258914
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.233419,29476
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.447773,31948.4
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.120096,29479.8
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.13735,29491.2
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,43.5133,489889
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.264575,29571
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.226654,29555.8
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.160472,29555.8
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,56.607,626551
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.314565,29552
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.276517,29544.4
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.276043,29548.2
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,22.3335,263821
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.223103,29502.6
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.444132,31958.8
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.115487,29487.2
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.138361,29487.2
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,43.4609,491459
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.26218,29575
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.222424,29575
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.159134,29575
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,56.3371,626651
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.318962,29579.2
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.296062,29567.8
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.266821,29529.6
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,22.3506,262243
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.223583,29449.2
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.432414,31951.2
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.116966,29433.8
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.142668,29437.6
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,43.3827,489727
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.264158,29602.4
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.218584,29591
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.1622,29575.6
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,56.4339,622509
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.308638,29606.4
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.268472,29610.2
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.270987,29579.6
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,22.3047,263895
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.220582,29518.4
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.440913,32070.2
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.116639,29510.8
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.140204,29495.4
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,43.3462,490552
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.261016,29625.4
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.2171,29610.2
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.156274,29598.8
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,56.4875,627562
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.317047,29591
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.31402,29587.2
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.26974,29575.6
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,22.2393,264178
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.223096,29537
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.439633,32115.2
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.112326,29525.6
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.138578,29533.4
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,26.5744,298423
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.25212,29579.6
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.18508,29579.6
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.146546,29541.6
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,43.4373,481503
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.307896,29595
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.352932,29579.8
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.216543,29564.4
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.8001,301597
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.289804,29503
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.617744,33362.6
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,39.2018,446936
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.308101,29610
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.628739,33504.4
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.137862,29610
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.146188,29610
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,35.9158,414493
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.281253,29644.6
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.197976,29644.6
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.14958,29629.2
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,43.4436,480200
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.287646,29667.2
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.274674,29652
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.219749,29632.8
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.2219,298626
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.277842,29590.2
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.608412,33480.6
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.126661,29590.2
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.144082,29574.8
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,35.8793,415777
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.293189,29640
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.205157,29640
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.158162,29616.6
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,43.7266,480090
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.306955,29650.4
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.273541,29635.2
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.224754,29616.2
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.3017,298301
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.282514,29577.8
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.61521,33452.8
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.128326,29562.4
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.142233,29558.4
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.645008,29558.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.2358,34745.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.184837,29562.4
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.68554,114383
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_265.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_265.txt
deleted file mode 100644
index 50304795f4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_265.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,150.552,1.35888e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.382418,25938.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.257689,25958
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.86179,62845.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.84681,34340.4
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,77.6417,765948
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.207615,26247.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.325778,26757.6
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.159058,26236
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,145.852,1.42986e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.356312,26569.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.358014,27121.4
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.194943,26538.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,52.3246,501011
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.321899,26609.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,0.972392,29725
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,62.9844,626051
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.300069,26810
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,0.983777,30124.6
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.17022,26810.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.162905,26802.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,141.377,1.42327e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.329854,27236.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.463089,27972.6
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.289489,27259.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,147.681,1.48593e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.372913,27490
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.475754,28271.4
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.306226,27493.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,52.6049,521657
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.329522,27451
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,0.981095,31530.2
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.162117,27477.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.162764,27470
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,141.39,1.46028e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.327148,27868.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.451159,28726.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.289183,27841.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,147.292,1.51208e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.471459,28066.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.465469,28989.4
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.301272,28055.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,52.168,528583
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.327805,28055
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,0.981364,32803
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.155705,28062.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.160735,28070.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,81.8565,870716
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.287596,28186.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.239371,28171.2
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.157522,28175
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,86.1486,908432
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.323148,28253.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.301502,28257
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.287108,28272.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,31.5869,338316
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.283557,28238
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.540272,30260
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,51.5673,545026
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.28437,28518.2
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.616055,30628.2
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.169004,28541
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.154969,28544.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,73.1859,779608
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.27756,28679.4
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.29299,28653
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.26851,28657
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,86.4539,926669
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.320779,28830.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.303576,28834.4
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.283192,28846.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,31.4412,343147
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.274776,28735.4
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.524304,30917.4
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.120479,28746.8
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.150585,28758.2
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,73.1047,784667
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.281611,28944.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.300356,28944.6
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.273567,28948.4
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,86.5005,934084
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.320491,29005.6
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.309111,28975.2
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.274513,28982.8
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,31.8439,335448
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.320767,28925.8
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.520132,31214.6
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.131027,28937.2
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.162879,28941
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,72.7452,785752
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.280894,29157.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.299032,29165.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.270738,27685.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,87.9394,946851
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.328971,29173.2
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.311525,29169.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.284369,29177
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,31.4698,347643
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.27571,29127.6
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.538602,31485.2
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.128217,29154.2
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.155928,29161.8
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,40.3491,443430
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.243557,29207.4
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.185292,29215
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.138815,29199.8
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,56.8501,621376
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.312875,29211.2
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.27173,29211.2
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.260895,29203.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,22.2888,259253
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.226175,29158
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.446026,31538.4
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,50.6775,560423
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.307133,29261
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.595459,31679.8
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.226328,29253.4
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.180761,29253.4
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,43.6579,490840
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.26915,29402.6
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.216011,29391.2
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.156952,29326
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,56.8946,626642
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.305817,29410.8
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.262078,29410.8
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.270603,27884.6
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,22.2916,259351
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.232062,29330.4
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.452125,31787.4
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.122297,29315
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.133773,29318.4
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,43.5641,496343
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.266226,29502.4
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.20839,29490.8
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.144172,29444.8
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,56.909,629275
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.313323,29479.6
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.267307,29483.4
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.258059,29441.2
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,22.3601,261086
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.229388,29346
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.444542,31802.2
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.114572,29315.2
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.136979,29315
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,43.4579,489278
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.281528,29472
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.21589,29464.4
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.158578,29445.2
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,57.0732,630778
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.304421,29460.8
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.263667,29453.2
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.25927,29453.2
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,22.4552,261117
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.230489,29349.6
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.451255,31821
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.118591,29334.2
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.136082,29326.4
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,43.5808,495239
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.271641,29468.4
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.212319,29468.4
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.151634,29460.6
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,57.1736,634202
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.323461,29438.2
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.324248,29445.8
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.265074,29430.4
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,22.5101,264327
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.230584,29407
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.448759,31954.8
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.115992,29395.6
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.132357,29395.6
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,43.5573,494485
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.260505,29453.2
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.214732,29453.2
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.151001,29453.4
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,56.8903,625869
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.312209,29432.6
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.275122,29429.2
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.264888,29414
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,22.5955,264234
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.23799,29374.6
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.44554,31941.8
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.116902,29371.2
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.139641,29371.2
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,26.776,298332
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.259723,29422.2
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.193643,29422.2
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.139845,29406.8
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,43.4762,479453
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.296286,29445
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.280638,29429.6
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.202527,29410.2
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,26.5241,297514
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.280017,29340.2
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.631376,33193
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,39.2177,444915
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.295845,29463.6
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.633552,33293
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.129983,29448.2
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.143276,29448.2
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,35.9008,414000
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.294046,29494.4
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.196543,29490.4
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.137887,29475
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,43.5729,485423
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.288203,29555.2
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.26972,29524.4
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.195091,29497.6
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,26.3235,296875
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.275641,29431.6
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.618121,33276.4
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.123264,29431.6
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.134566,29427.6
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,36.1377,415367
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.279871,29516.4
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.201458,29481.8
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.139244,29470.2
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,43.5938,481908
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.295012,29504
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.271332,29488.8
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.215391,29488.8
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,26.5262,298122
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.273356,29404.4
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.620394,33249.2
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.127449,29400.6
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.132012,29400.6
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.614307,29400.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.25512,35808.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.129037,29400.6
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.76575,113784
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_266.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_266.txt
deleted file mode 100644
index 2fec11a037..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_266.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,157.049,1.42437e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.408983,26264
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.282687,26279.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.91682,60773.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.68663,34270.2
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,78.2223,786958
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.210668,26541.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.291756,27063
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.150566,26553
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,126.993,1.24914e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.343666,26756.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.357938,27350.6
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.195104,26783.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,52.8028,515675
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.330744,25608.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,0.931586,28666.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,63.1143,630945
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.301374,26999
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,0.933979,30261.2
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.18012,27034.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.166655,27038
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,143.405,1.45185e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.329886,27453
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.455786,28181
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.292895,27460.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,129.128,1.30214e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.380113,27567
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.461597,28344.8
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.303186,27571.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,53.3772,531164
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.338232,27567.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,0.93675,31602.6
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.150649,27571.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.178693,27571.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,144.234,1.48584e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.33045,27966.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.443537,28832
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.292471,27943.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,129.334,1.32744e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.377597,28114.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.453579,28992.2
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.290258,28092.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,53.0468,537122
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.358981,28058
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,0.960603,32702.4
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.16085,28058
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.166931,28050.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,82.759,871886
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.285643,28191.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.249669,28191.4
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.185452,28199
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,76.9195,815515
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.334098,28355.4
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.295704,28374.4
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.280734,28374.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,32.5495,343613
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.291799,28332.6
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.481424,30320
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,53.1383,571387
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.285579,28618.8
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.579932,30637
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.155531,28561.4
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.161464,28569
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,74.8493,805997
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.287793,28825.8
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.258124,28798.8
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.270962,28810.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,77.4678,838618
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.329745,28868.2
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.30142,28879.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.274936,28883.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,32.5414,349991
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.282558,28837.6
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.498608,30962
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.123052,28833.6
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.151033,28822
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,74.917,814945
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.296345,29035.6
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.264639,29039.6
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.273247,29047.2
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,77.696,847097
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.338833,29107.8
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.305822,29115.6
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.298084,27689.2
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,32.4144,351142
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.291525,29032
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.486352,31305.6
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.121906,29043.4
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.158892,29047.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,74.6395,816671
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.292882,29256.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.258073,29256.6
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.27644,29264.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,77.7032,850474
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.335512,29287
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.290411,29298.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.276901,29287
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,32.2584,353586
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.289056,29226.2
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.491646,31568.6
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.124115,29241.4
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.156268,29245.2
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,41.0142,464333
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.240817,29314
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.156594,29317.8
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.142086,29294.8
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,53.1458,589389
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.314181,29310
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.304126,29302.4
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.256216,29306.2
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,22.8487,264330
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.215628,29241.6
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.435908,31614.8
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,52.4881,589318
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.304498,29462
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.584496,31858
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.203153,29443
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.166846,29443
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,44.7881,501309
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.266552,29549.8
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.236479,29542
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.153299,29534.4
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,53.2674,595662
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.310142,29580.4
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.27061,29549.6
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.265196,29519
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,22.9301,264936
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.222962,29469.4
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.443895,31896.2
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.119084,29488.4
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.134879,29496.2
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,44.7429,502162
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.268177,29626.4
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.240844,29611
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.164883,29611
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,53.1094,592137
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.309982,29640.8
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.29708,29640.8
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.26332,29625.4
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,22.8085,265792
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.229713,29536.2
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.426531,32031
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.111513,29544.2
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.143206,29513.4
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,44.6512,501649
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.270193,29736.6
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.229976,29740.6
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.160383,29706
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,53.1013,593487
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.322487,29624.2
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.306865,29639.6
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.25237,29632
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,23.0185,265912
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.228312,29616.6
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.440854,32141.8
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.112773,29620.4
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.141586,29605
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,45.1498,505203
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.265086,29666.6
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.232164,29666.6
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.164978,29670.4
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,53.4391,599384
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.309751,29640
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.271525,29636.2
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.254674,29628.6
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,22.9036,267502
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.219288,29590.4
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.439069,32134.8
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.117388,29582.6
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.135366,29586.4
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,44.6942,503734
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.269465,29690
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.228344,29674.6
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.1635,29655.4
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,53.4426,599226
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.314264,29674.6
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.309079,29659.2
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.260229,29651.6
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,23.0238,267446
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.230955,29602
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.443441,32180.8
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.116243,29605.8
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.141785,29605.8
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,27.1801,306636
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.255339,29632.4
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.202194,29632.4
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.142521,29632.4
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,43.0601,480255
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.301515,29697.6
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.281208,29705.2
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.224875,29670.8
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,27.269,305710
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.267052,29602
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.625072,33466.2
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,41.0977,460679
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.285746,29743.4
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.647145,32329.8
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.159551,29743.4
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.147174,29724.4
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,37.0261,419154
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.269951,29800.4
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.214911,29800.4
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.141933,29781.4
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,43.0102,481247
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.294341,29819.6
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.266744,29819.6
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.224812,29789.2
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,27.233,305554
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.275448,29694
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.646633,33607.8
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.132243,29694
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.140032,29694
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,36.8916,417611
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.272523,29812
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.219864,29812
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.14917,29789.2
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,42.9744,482386
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.30058,29796.8
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.268837,29785.4
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.216754,29766.4
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,27.1978,302739
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.26556,29716.8
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.642179,33630.6
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.14204,29701.6
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.134662,29694
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.62719,29694
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.222,36217
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.129823,29697.8
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.67455,114885
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_267.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_267.txt
deleted file mode 100644
index 93685543e1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_267.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,159.671,1.45097e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.414032,26199
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.277419,26206.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.93183,63592.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.73804,34308.8
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,77.5136,776185
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.204089,26517.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.285169,26521
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.155007,26517.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,127.27,1.25111e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.348664,26680.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.351409,27259.2
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.19509,26680.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,53.0549,514303
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.322122,26730.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,0.960058,29834.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,62.9681,628946
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.290238,26930.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,0.951002,30196.4
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.185381,26958
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.165177,26958
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,143.662,1.44917e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.335621,27403.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.46296,28143.2
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.285438,27388.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,130.149,1.30797e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.378846,27554.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.461866,28340.2
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.294385,27532.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,53.593,528992
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.338545,27582.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,0.943111,31579.4
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.179487,27529.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.169733,27529.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,144.141,1.48523e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.329348,27935.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.459088,28816.4
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.288075,27912.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,129.962,1.33662e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.38154,28042.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.467312,28969
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.304459,28050.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,53.3228,541727
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.333674,28031
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,0.951296,32722
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.166136,28019.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.161759,28027.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,82.4119,864320
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.269355,28141.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.251513,28129.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.182246,28133.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,77.2159,815017
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.337451,28298.4
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.304139,28264.2
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.280632,28279.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,32.5272,342084
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.304466,28229.8
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.511754,30240
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,53.1151,569672
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.291012,28527.2
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.590615,30587.4
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.163993,28519.8
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.168287,28523.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,75.0629,805863
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.290321,28726
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.261106,28733.6
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.276549,28741.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,77.6741,836652
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.337425,28783.4
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.30309,28768
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.289862,28775.6
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,32.469,344041
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.291384,28764
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.504125,30919
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.124108,28756.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.15541,28756.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,74.762,808017
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.293624,28967.2
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.258335,28952
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.285252,28940.4
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,77.604,841181
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.33068,29016.6
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.298783,29020.4
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.283442,29020.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,32.2171,347243
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.296856,28963.4
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.511684,31240.8
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.135359,28982.4
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.166476,28986.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,75.0394,815546
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.29859,29142.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.26028,29112.2
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.279192,29119.8
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,77.7645,849580
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.335723,29199.6
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.298911,29188.2
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.281572,29192
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,32.2133,351783
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.291435,29142.6
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.508861,31466
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.124991,29135
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.152646,29146.6
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,40.9805,463475
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.232594,29211.4
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.152019,29184.8
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.135743,29139
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,53.2529,588193
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.306066,29200
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.3254,29200
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.257592,29162
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,22.9867,263282
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.225362,29146.6
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.455664,31500.8
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,52.8654,586985
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.307268,29313.8
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.594301,31713.6
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.210764,29310
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.17221,29298.6
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,45.3646,502108
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.277368,29447.2
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.25107,29447.2
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.166514,29443.4
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,53.3835,591846
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.312357,29420.4
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.314724,29405.2
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.255122,29409
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,23.0297,265575
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.223673,29366.2
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.444933,31842.6
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.119622,29393.2
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.134604,29397.2
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,44.8916,501681
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.25843,29496.6
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.2355,29477.4
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.168998,29481.2
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,53.374,591221
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.31219,29475.4
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.262372,29448.8
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.264952,29437.4
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,23.1118,265844
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.224805,29410
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.435806,31867.2
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.108589,29410
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.13769,29410
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,44.9201,502243
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.26563,29551.8
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.248153,29555.6
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.178182,29528.8
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,53.8614,599626
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.310801,29562.6
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.28076,29501.4
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.257471,29486
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,23.0662,264503
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.233144,29436.4
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.438775,31938.6
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.120959,29421
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.147602,29424.8
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,45.0612,504327
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.275282,29585.4
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.243153,29593.2
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.172024,29573.8
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,53.3727,592635
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.308933,29582
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.262495,29582
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.268223,29566.6
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,23.3319,268437
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.239307,29536
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.437143,32088.2
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.119718,29547.4
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.136569,29532
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,44.8987,502620
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.263097,29624.4
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.242533,29593.6
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.168633,29593.6
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,53.3799,595506
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.306674,29601.6
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.259871,29609.2
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.259,29597.8
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,23.0855,265645
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.230194,29582.6
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.432292,32161.4
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.109715,29586.4
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.137657,29571
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,27.1782,306141
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.245106,29632
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.186232,29628.2
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.142226,29597.6
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,42.9186,476528
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.299729,29643.6
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.271442,29643.6
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.229536,29597.6
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,27.2244,305476
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.273778,29563.2
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.64797,33435.2
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,41.1902,458662
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.286846,29689.6
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.642614,33542.6
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.163525,29670.6
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.143084,29666.8
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,36.9691,417182
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.271416,29770
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.213944,29762.4
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.147404,29739.6
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,43.1152,483409
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.294827,29808.2
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.314462,29762.4
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.224299,29743.4
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,27.2069,305348
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.267768,29743.6
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.639382,33630.6
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.140268,29713
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.14254,29693.8
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,36.8511,417525
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.26949,29800.6
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.212146,29770.2
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.141663,29755
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,43.1946,480936
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.300548,29796.8
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.267365,29796.8
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.232684,29796.8
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,27.2799,307255
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.267851,29736
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.631439,33649.8
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.142719,29732.2
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.136005,29728.4
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.606287,29728.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.21748,34946.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.195122,29728.4
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.82772,116495
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_268.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_268.txt
deleted file mode 100644
index 15ea4be9c3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_268.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,157.979,1.44264e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.415761,26271.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.347633,26290
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.75702,58178
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.7396,34349.2
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,78.2725,786440
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.187634,26517
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.305835,27039.2
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.148415,26541
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,126.218,1.25053e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.342577,26794.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.371102,27384.4
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.206737,26820.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,52.577,492786
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.325796,26844
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.01519,29879.6
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,63.0262,627772
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.293855,27044.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.00186,30287.6
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.197906,27037.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.168773,27041.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,143.678,1.45529e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.331372,27506
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.450589,28211
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.298071,27475.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,129.288,1.30893e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.376555,27609.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.472227,28387.2
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.309636,27590.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,53.0654,524344
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.33491,27610
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,0.95881,31636.8
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.17667,27633
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.187128,27640.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,144.088,1.49667e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.33226,28039
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.45448,28916.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.294891,28058
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,129.467,1.33103e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.37537,28122.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.465347,29030
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.309656,28107.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,53.2242,538073
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.334244,28115
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,0.963821,32736.6
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.165689,28107.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.170674,28111.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,82.4395,871762
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.273049,28294.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.255794,28282.8
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.191263,28290.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,76.4929,809843
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.329592,28416
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.31068,28416
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.285259,28400.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,32.3583,345417
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.28775,28397
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.53585,30361.6
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,53.1277,572101
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.289964,28649.6
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.631568,30663.8
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.153215,28642
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.171558,28657.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,74.9948,806547
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.293426,28860
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.29427,28875.2
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.275314,28863.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,77.3328,831845
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.326328,28966.8
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.293848,28925
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.281982,28898.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,32.3874,350317
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.381092,28887
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.544599,31049.6
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.126668,28898.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.152953,28898.4
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,74.9127,814790
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.297016,29062.2
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.301873,29073.6
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.274002,29058.4
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,77.4105,841822
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.336043,29130.6
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.304018,29138.2
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.2865,29138.2
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,32.126,350003
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.290866,29123
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.543569,31384.8
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.125439,29138.2
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.152377,29138.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,74.9284,821387
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.291358,29309.8
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.294385,29309.6
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.273348,29302.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,77.8189,853164
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.339505,29381.8
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.299902,29355.2
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.281707,29343.8
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,32.2492,354893
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.29162,29286.8
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.542928,31606.4
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.123826,29283
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.158681,29260.2
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,40.9547,462549
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.227443,29351.8
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.155289,29336.6
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.138879,29317.4
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,53.0785,586709
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.314917,29378.4
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.297381,29348
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.256945,29351.8
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,23.0526,264413
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.230917,29302.4
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.486327,31652.8
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,52.7449,591079
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.303096,29469.6
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.630525,31850.4
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.212543,29458.2
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.16711,29465.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,45.0617,503990
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.268997,29557.6
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.258495,29538.4
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.161042,29523
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,53.5478,595695
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.313111,29526.6
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.296901,29515.2
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.261228,29500
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,22.9725,266458
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.222629,29465.2
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.476951,31895.2
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.115852,29469.2
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.136691,29469.2
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,45.0946,504612
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.263403,29626.4
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.238034,29611.2
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.15543,29596
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,53.411,596209
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.309496,29609.6
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.295479,29563.4
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.260568,29563.4
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,23.2305,268111
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.223615,29513
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.480977,31969.4
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.108256,29497.6
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.135097,29497.6
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,44.8984,502389
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.270699,28171.6
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.233798,28156.2
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.160274,28140.8
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,53.5999,591389
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.31779,29643
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.298859,29635.2
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.258545,29600.8
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,22.9657,267075
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.222539,29539.8
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.493456,32057.4
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.130853,29543.6
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.146994,29555
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,45.0524,506095
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.264236,29685.2
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.237542,29677.6
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.16341,29670
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,53.3686,594822
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.322207,29663
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.298399,29659
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.267973,29643.6
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,23.1358,267504
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.223647,29590.2
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.445841,32146.2
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.119321,29582.4
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.153158,29582.4
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,45.066,506608
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.268908,29750.8
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.238743,29735.6
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.15765,29705
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,53.3066,595357
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.30606,29724.4
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.268152,29709.2
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.261214,29697.8
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,23.0216,268090
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.216491,29667.4
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.446301,32219.2
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.110777,29644.2
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.134303,29644.2
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,27.0053,301122
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.241759,29728.2
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.199448,29728.2
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.136262,29678.4
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,42.8956,480949
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.295755,29728.2
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.27715,29728.2
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.213292,29712.8
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,27.1434,302019
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.260773,29674.8
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.645756,33527.8
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,41.1084,452568
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.288068,29812
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.663823,33680.2
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.185196,29793
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.143673,29793
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,36.7699,418828
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.271647,29865.2
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.213202,29865.2
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.143001,29834.8
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,43.2289,486662
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.297765,29884.2
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.263953,29884.2
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.221432,29880.4
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,27.1151,303465
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.266584,29823.4
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.620457,32402.2
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.134022,29808.2
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.140991,29823.4
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,37.1188,421913
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.273137,29884.2
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.223877,29876.6
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.14247,29846.2
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,42.8424,479392
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.293854,29884.2
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.26901,29872.8
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.230361,29857.6
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,27.109,305105
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.268044,29808.2
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.624092,32406
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.136325,29793
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.13687,29777.8
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.615158,29793
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.33968,36316
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.192172,29796.8
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.92475,116746
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_269.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_269.txt
deleted file mode 100644
index ffe9c7f43d..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_269.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,160.729,1.45536e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.403914,26163
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.293496,26109.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,5.9709,67251.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,2.73814,34096.6
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,77.2551,773915
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.196837,26441
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.301368,26962.8
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.146015,26452.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,127.234,1.24772e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.342405,26639.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.387269,27225.6
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.212562,26662
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,52.6512,509299
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.323845,26704
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,0.994126,29766.2
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,63.0586,629511
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.30188,26892.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,0.971502,30135.4
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.180805,26900.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.163724,26900.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,143.476,1.44485e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.327626,27376.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.446218,28089.4
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.290584,27365.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,129.567,1.3028e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.378763,27505.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.472797,28294.4
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.293528,27501.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,53.5011,525808
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.335556,27487.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.00013,31518.4
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.165355,27507
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.162123,27507
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,144.947,1.49617e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.339415,27893.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.450275,28770.6
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.282021,27885.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,129.562,1.32464e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.371107,27965.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.471083,28881.2
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.31509,27977.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,53.1016,536493
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.337483,27939.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,0.979847,32568.4
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.157055,27946.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.159999,27950.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,82.9835,867403
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.275825,28141.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.315225,28095.6
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.182643,28103.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,76.7925,812172
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.34746,28306
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.3102,28283.2
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.276101,28290.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,32.5864,342571
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.292772,28244.8
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.505623,30228.4
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,53.0444,569632
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.290751,28496
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.608208,30517.8
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.155372,28488.2
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.164038,28488.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,74.8307,802514
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.28636,28749
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.297414,28741.4
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.268517,28725.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,77.597,837031
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.33971,28795.2
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.293253,28783.8
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.281912,28791.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,32.5679,350264
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.29166,28776
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.55274,30923.2
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.134687,28772
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.151436,28787.2
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,74.9734,811515
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.292447,29001.2
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.294846,28955.6
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.271256,28967
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,77.8673,843698
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.333534,29062.2
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.304983,29050.8
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.277649,29020.4
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,32.5221,352657
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.29802,28997.6
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.553066,31240.4
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.125944,28978.6
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.146981,28963.4
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,75.1733,818688
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.293906,29172.6
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.287685,29176.4
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.272254,29180.2
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,78.116,848689
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.335249,29207
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.309176,29214.6
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.281451,29214.6
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,32.5588,355049
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.296836,29142.4
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.54504,31477.2
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.129221,29157.6
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.161413,29161.6
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,41.3082,464069
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.24133,29237.8
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.157612,29237.8
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.134022,29207.2
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,53.4512,586710
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.31349,29207.4
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.28963,29207.4
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.252632,29211.2
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,23.1345,263623
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.230143,29203.6
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.488023,31546.4
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,52.8845,588491
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.28999,29291
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.624598,31683.2
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.207493,29291
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.172324,29260.6
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,44.9031,498448
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.265106,29477.4
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.26782,29466
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.166316,29458.4
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,53.4766,590058
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.303569,29477
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.296952,29477
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.261432,29480.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,23.1321,265588
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.226117,29416
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.479498,31849.6
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.115353,29400.8
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.13456,29385.6
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,44.9036,501895
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.264075,29522.8
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.261374,29526.6
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.154642,29526.6
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,53.4845,597344
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.311979,29502
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.291371,29506
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.261817,29502
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,23.1998,267699
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.228498,29466.8
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.483664,31946.4
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.106975,29466.8
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.13262,29466.8
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,44.8166,502763
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.257995,29532.2
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.238661,29544.2
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.157516,29532.2
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,53.5401,592689
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.307743,29539.2
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.292395,29516.2
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.256261,29508.4
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,23.04,266388
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.226175,29481.8
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.481668,31995.6
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.110341,29470.4
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.135212,29474.2
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,45.0617,501693
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.275665,29643
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.292344,29643
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.157817,29593.2
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,53.538,594523
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.307729,29586
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.291947,29570.6
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.25909,29543.6
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,23.189,266549
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.226994,29501.6
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.489975,32046.2
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.117747,29505.4
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.147903,29505.4
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,45.0959,503532
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.268753,29662.8
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.238936,29662.8
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.159443,29628.2
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,53.9496,597935
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.317989,29624.8
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.270002,29620.8
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.258123,29559.2
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,23.5419,269971
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.225202,29509.6
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.434609,32080.8
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.106355,29513.4
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.134822,29521
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,27.0386,305748
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.238398,29597.4
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.17998,29578.2
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.133651,29547.4
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,42.9802,477477
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.295012,29605.2
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.268101,29605.2
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.223218,29563
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,27.4824,306594
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.277599,29528.6
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.631325,33431
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,41.2329,465223
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.329508,29655.6
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.656246,33565.6
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.16467,29655.6
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.149323,29655.6
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,37.076,418406
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.26597,29732
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.224844,29716.8
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.139506,29716.8
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,43.43,486118
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.295089,29747.4
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.279307,29751.2
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.23589,29735.8
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,27.2939,306835
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.270392,29705.4
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.636041,33607.8
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.139308,29701.6
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.139647,29701.6
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,36.897,417413
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.276343,29747.4
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.213509,29732.2
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.138284,29717
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,43.1599,481978
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.291519,29755
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.276235,29762.6
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.222891,29755
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,27.6381,310049
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.274859,29709.4
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.630793,33611.8
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.136069,29686.6
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.13566,29682.8
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.623741,29682.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.25289,36186.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.135909,29682.8
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,9.72458,114871
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp32_perf_fp32_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp32_perf_fp32_120.txt
deleted file mode 100644
index 8d9d8fee56..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp32_perf_fp32_120.txt
+++ /dev/null
@@ -1,687 +0,0 @@
-Conv1,352.727,2.71583e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.49045,21979.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.298591,21994.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Pool1,14.1205,117030
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-BatchNorm1,3.20576,31212
-BatchNorm1_f2h,0,0
-BatchNorm1_h2f,0,0
-Conv2,96.997,770268
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.399084,22871.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-BatchNorm2,0.640881,23419.8
-BatchNorm2_f2h,0,0
-BatchNorm2_h2f,0,0
-Relu2,0.363673,22882.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Conv3,269.812,2.26471e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.433215,23528.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-BatchNorm3,0.618777,24230
-BatchNorm3_f2h,0,0
-BatchNorm3_h2f,0,0
-Relu3,0.424281,23547.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,144.543,1.18734e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.403065,24240
-Add4_f2h,0,0
-Add4_h2f,0,0
-BatchNorm4,1.53223,30464
-BatchNorm4_f2h,0,0
-BatchNorm4_h2f,0,0
-Conv5,166.193,1.48125e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.356293,23675.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-BatchNorm5,1.51066,30515
-BatchNorm5_f2h,0,0
-BatchNorm5_h2f,0,0
-Add6,0.26439,23664.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu4,0.309632,23672.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Conv6,217.966,2.03416e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add7,0.406316,25935.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-BatchNorm6,0.614763,26893.2
-BatchNorm6_f2h,0,0
-BatchNorm6_h2f,0,0
-Relu5,0.349894,25912.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv7,268.095,2.47292e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add8,0.43847,26289.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-BatchNorm7,0.62083,27255.2
-BatchNorm7_f2h,0,0
-BatchNorm7_h2f,0,0
-Relu6,0.353343,26301.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv8,173.448,1.58116e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add9,0.390066,26442.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-BatchNorm8,1.54885,34110.6
-BatchNorm8_f2h,0,0
-BatchNorm8_h2f,0,0
-Add10,0.280256,26442.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu7,0.320128,26442.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Conv9,219.91,2.1566e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add11,0.410745,27254
-Add11_f2h,0,0
-Add11_h2f,0,0
-BatchNorm9,0.608754,29295.2
-BatchNorm9_f2h,0,0
-BatchNorm9_h2f,0,0
-Relu8,0.34229,27250
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv10,405.412,3.80103e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add12,0.423705,25999.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-BatchNorm10,0.627186,27026.2
-BatchNorm10_f2h,0,0
-BatchNorm10_h2f,0,0
-Relu9,0.341772,25999.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv11,274.518,2.45326e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add13,0.354445,25078.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-BatchNorm11,1.53738,33075
-BatchNorm11_f2h,0,0
-BatchNorm11_h2f,0,0
-Add14,0.283603,23765.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu10,0.315775,23769.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Conv12,128.551,1.20479e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add15,0.345157,25545.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-BatchNorm12,0.428863,25545.2
-BatchNorm12_f2h,0,0
-BatchNorm12_h2f,0,0
-Relu11,0.290988,25544.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv13,157.593,1.47273e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add16,0.404562,25638.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-BatchNorm13,0.441138,25627.2
-BatchNorm13_f2h,0,0
-BatchNorm13_h2f,0,0
-Relu12,0.307039,25631.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv14,147.479,1.33218e+06
-Conv14_f2h,0,0
-Conv14_h2f,0,0
-Add17,0.321491,25208.2
-Add17_f2h,0,0
-Add17_h2f,0,0
-BatchNorm14,0.897682,29482.6
-BatchNorm14_f2h,0,0
-BatchNorm14_h2f,0,0
-Conv15,189.674,1.72808e+06
-Conv15_f2h,0,0
-Conv15_h2f,0,0
-Add18,0.35511,25306.6
-Add18_f2h,0,0
-Add18_h2f,0,0
-BatchNorm15,0.923762,29649
-BatchNorm15_f2h,0,0
-BatchNorm15_h2f,0,0
-Add19,0.288403,25333.6
-Add19_f2h,0,0
-Add19_h2f,0,0
-Relu13,0.314989,25326
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Conv16,151.332,1.38495e+06
-Conv16_f2h,0,0
-Conv16_h2f,0,0
-Add20,0.364755,25440
-Add20_f2h,0,0
-Add20_h2f,0,0
-BatchNorm16,0.45182,25413.6
-BatchNorm16_f2h,0,0
-BatchNorm16_h2f,0,0
-Relu14,0.30583,25436.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Conv17,242.979,2.15965e+06
-Conv17_f2h,0,0
-Conv17_h2f,0,0
-Add21,0.408454,24674.6
-Add21_f2h,0,0
-Add21_h2f,0,0
-BatchNorm17,0.454418,24693.6
-BatchNorm17_f2h,0,0
-BatchNorm17_h2f,0,0
-Relu15,0.323397,24697.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Conv18,135.456,1.18545e+06
-Conv18_f2h,0,0
-Conv18_h2f,0,0
-Add22,0.325971,24537.4
-Add22_f2h,0,0
-Add22_h2f,0,0
-BatchNorm18,0.928587,28905.8
-BatchNorm18_f2h,0,0
-BatchNorm18_h2f,0,0
-Add23,0.266848,24537.4
-Add23_f2h,0,0
-Add23_h2f,0,0
-Relu16,0.221516,24552.6
-Relu16_f2h,0,0
-Relu16_h2f,0,0
-Conv19,109.495,1.00602e+06
-Conv19_f2h,0,0
-Conv19_h2f,0,0
-Add24,0.352371,25164.8
-Add24_f2h,0,0
-Add24_h2f,0,0
-BatchNorm19,0.430892,25138.2
-BatchNorm19_f2h,0,0
-BatchNorm19_h2f,0,0
-Relu17,0.309568,25153.4
-Relu17_f2h,0,0
-Relu17_h2f,0,0
-Conv20,240.505,2.12346e+06
-Conv20_f2h,0,0
-Conv20_h2f,0,0
-Add25,0.405664,24427.2
-Add25_f2h,0,0
-Add25_h2f,0,0
-BatchNorm20,0.451731,24434.8
-BatchNorm20_f2h,0,0
-BatchNorm20_h2f,0,0
-Relu18,0.318406,24442
-Relu18_f2h,0,0
-Relu18_h2f,0,0
-Conv21,134.238,1.15705e+06
-Conv21_f2h,0,0
-Conv21_h2f,0,0
-Add26,0.328722,24259
-Add26_f2h,0,0
-Add26_h2f,0,0
-BatchNorm21,0.910079,28655
-BatchNorm21_f2h,0,0
-BatchNorm21_h2f,0,0
-Add27,0.244902,24259.2
-Add27_f2h,0,0
-Add27_h2f,0,0
-Relu19,0.208806,24278.2
-Relu19_f2h,0,0
-Relu19_h2f,0,0
-Conv22,109.363,995075
-Conv22_f2h,0,0
-Conv22_h2f,0,0
-Add28,0.359077,24915.2
-Add28_f2h,0,0
-Add28_h2f,0,0
-BatchNorm22,0.445318,24922.8
-BatchNorm22_f2h,0,0
-BatchNorm22_h2f,0,0
-Relu20,0.313215,24930.4
-Relu20_f2h,0,0
-Relu20_h2f,0,0
-Conv23,238.499,2.08729e+06
-Conv23_f2h,0,0
-Conv23_h2f,0,0
-Add29,0.415065,24247.8
-Add29_f2h,0,0
-Add29_h2f,0,0
-BatchNorm23,0.465695,24263.4
-BatchNorm23_f2h,0,0
-BatchNorm23_h2f,0,0
-Relu21,0.322854,24275.2
-Relu21_f2h,0,0
-Relu21_h2f,0,0
-Conv24,131.667,1.13298e+06
-Conv24_f2h,0,0
-Conv24_h2f,0,0
-Add30,0.325343,24183.2
-Add30_f2h,0,0
-Add30_h2f,0,0
-BatchNorm24,0.904709,28612.4
-BatchNorm24_f2h,0,0
-BatchNorm24_h2f,0,0
-Add31,0.222137,24210.4
-Add31_f2h,0,0
-Add31_h2f,0,0
-Relu22,0.210669,24226.4
-Relu22_f2h,0,0
-Relu22_h2f,0,0
-Conv25,63.9229,584253
-Conv25_f2h,0,0
-Conv25_h2f,0,0
-Add32,0.287955,24537
-Add32_f2h,0,0
-Add32_h2f,0,0
-BatchNorm25,0.339839,24537
-BatchNorm25_f2h,0,0
-BatchNorm25_h2f,0,0
-Relu23,0.22014,24498.8
-Relu23_f2h,0,0
-Relu23_h2f,0,0
-Conv26,100.177,906648
-Conv26_f2h,0,0
-Conv26_h2f,0,0
-Add33,0.372223,24919.4
-Add33_f2h,0,0
-Add33_h2f,0,0
-BatchNorm26,0.365586,24907.8
-BatchNorm26_f2h,0,0
-BatchNorm26_h2f,0,0
-Relu24,0.279231,24911.6
-Relu24_f2h,0,0
-Relu24_h2f,0,0
-Conv27,76.8815,689259
-Conv27_f2h,0,0
-Conv27_h2f,0,0
-Add34,0.305549,24946.2
-Add34_f2h,0,0
-Add34_h2f,0,0
-BatchNorm27,0.564005,27250.8
-BatchNorm27_f2h,0,0
-BatchNorm27_h2f,0,0
-Conv28,116.268,1.06217e+06
-Conv28_f2h,0,0
-Conv28_h2f,0,0
-Add35,0.325055,25244.2
-Add35_f2h,0,0
-Add35_h2f,0,0
-BatchNorm28,0.683973,27606
-BatchNorm28_f2h,0,0
-BatchNorm28_h2f,0,0
-Add36,0.209746,25267
-Add36_f2h,0,0
-Add36_h2f,0,0
-Relu25,0.177894,25289.8
-Relu25_f2h,0,0
-Relu25_h2f,0,0
-Conv29,82.3367,767483
-Conv29_f2h,0,0
-Conv29_h2f,0,0
-Add37,0.316652,25480
-Add37_f2h,0,0
-Add37_h2f,0,0
-BatchNorm29,0.346803,25483.8
-BatchNorm29_f2h,0,0
-BatchNorm29_h2f,0,0
-Relu26,0.263852,25491.4
-Relu26_f2h,0,0
-Relu26_h2f,0,0
-Conv30,129.534,1.19781e+06
-Conv30_f2h,0,0
-Conv30_h2f,0,0
-Add38,0.371858,25514.4
-Add38_f2h,0,0
-Add38_h2f,0,0
-BatchNorm30,0.372434,25506.8
-BatchNorm30_f2h,0,0
-BatchNorm30_h2f,0,0
-Relu27,0.265414,25525.8
-Relu27_f2h,0,0
-Relu27_h2f,0,0
-Conv31,76.4658,705852
-Conv31_f2h,0,0
-Conv31_h2f,0,0
-Add39,0.30581,25430.6
-Add39_f2h,0,0
-Add39_h2f,0,0
-BatchNorm31,0.585222,27823.2
-BatchNorm31_f2h,0,0
-BatchNorm31_h2f,0,0
-Add40,0.127334,25442
-Add40_f2h,0,0
-Add40_h2f,0,0
-Relu28,0.160691,25445.8
-Relu28_f2h,0,0
-Relu28_h2f,0,0
-Conv32,67.8801,645673
-Conv32_f2h,0,0
-Conv32_h2f,0,0
-Add41,0.317267,25816
-Add41_f2h,0,0
-Add41_h2f,0,0
-BatchNorm32,0.341324,25823.6
-BatchNorm32_f2h,0,0
-BatchNorm32_h2f,0,0
-Relu29,0.273017,25819.6
-Relu29_f2h,0,0
-Relu29_h2f,0,0
-Conv33,128.942,1.20726e+06
-Conv33_f2h,0,0
-Conv33_h2f,0,0
-Add42,0.375462,25739.4
-Add42_f2h,0,0
-Add42_h2f,0,0
-BatchNorm33,0.365759,25762.2
-BatchNorm33_f2h,0,0
-BatchNorm33_h2f,0,0
-Relu30,0.268441,25766
-Relu30_f2h,0,0
-Relu30_h2f,0,0
-Conv34,75.9416,688671
-Conv34_f2h,0,0
-Conv34_h2f,0,0
-Add43,0.312371,25720.4
-Add43_f2h,0,0
-Add43_h2f,0,0
-BatchNorm34,0.55925,28139.6
-BatchNorm34_f2h,0,0
-BatchNorm34_h2f,0,0
-Add44,0.135948,25724.2
-Add44_f2h,0,0
-Add44_h2f,0,0
-Relu31,0.157631,25731.8
-Relu31_f2h,0,0
-Relu31_h2f,0,0
-Conv35,68.6611,660372
-Conv35_f2h,0,0
-Conv35_h2f,0,0
-Add45,0.313395,26095.8
-Add45_f2h,0,0
-Add45_h2f,0,0
-BatchNorm35,0.363064,26103.4
-BatchNorm35_f2h,0,0
-BatchNorm35_h2f,0,0
-Relu32,0.273715,26091.8
-Relu32_f2h,0,0
-Relu32_h2f,0,0
-Conv36,128.796,1.21921e+06
-Conv36_f2h,0,0
-Conv36_h2f,0,0
-Add46,0.375922,26038.2
-Add46_f2h,0,0
-Add46_h2f,0,0
-BatchNorm36,0.369222,26045.8
-BatchNorm36_f2h,0,0
-BatchNorm36_h2f,0,0
-Relu33,0.278444,26045.8
-Relu33_f2h,0,0
-Relu33_h2f,0,0
-Conv37,75.7666,717855
-Conv37_f2h,0,0
-Conv37_h2f,0,0
-Add47,0.297465,25915.8
-Add47_f2h,0,0
-Add47_h2f,0,0
-BatchNorm37,0.538463,28353.6
-BatchNorm37_f2h,0,0
-BatchNorm37_h2f,0,0
-Add48,0.13081,25927.2
-Add48_f2h,0,0
-Add48_h2f,0,0
-Relu34,0.159641,25927.2
-Relu34_f2h,0,0
-Relu34_h2f,0,0
-Conv38,68.59,660503
-Conv38_f2h,0,0
-Conv38_h2f,0,0
-Add49,0.315603,26301.2
-Add49_f2h,0,0
-Add49_h2f,0,0
-BatchNorm38,0.3467,26278
-BatchNorm38_f2h,0,0
-BatchNorm38_h2f,0,0
-Relu35,0.27127,26289.4
-Relu35_f2h,0,0
-Relu35_h2f,0,0
-Conv39,129.905,1.23748e+06
-Conv39_f2h,0,0
-Conv39_h2f,0,0
-Add50,0.376045,26129.4
-Add50_f2h,0,0
-Add50_h2f,0,0
-BatchNorm39,0.373753,26133.2
-BatchNorm39_f2h,0,0
-BatchNorm39_h2f,0,0
-Relu36,0.284358,26140.8
-Relu36_f2h,0,0
-Relu36_h2f,0,0
-Conv40,75.3331,714353
-Conv40_f2h,0,0
-Conv40_h2f,0,0
-Add51,0.312057,26099
-Add51_f2h,0,0
-Add51_h2f,0,0
-BatchNorm40,0.612383,28529.2
-BatchNorm40_f2h,0,0
-BatchNorm40_h2f,0,0
-Add52,0.130508,26083.8
-Add52_f2h,0,0
-Add52_h2f,0,0
-Relu37,0.166067,26099
-Relu37_f2h,0,0
-Relu37_h2f,0,0
-Conv41,69.9504,680018
-Conv41_f2h,0,0
-Conv41_h2f,0,0
-Add53,0.322118,26458.6
-Add53_f2h,0,0
-Add53_h2f,0,0
-BatchNorm41,0.357381,26462.4
-BatchNorm41_f2h,0,0
-BatchNorm41_h2f,0,0
-Relu38,0.276204,26470
-Relu38_f2h,0,0
-Relu38_h2f,0,0
-Conv42,129.6,1.23621e+06
-Conv42_f2h,0,0
-Conv42_h2f,0,0
-Add54,0.378489,26309
-Add54_f2h,0,0
-Add54_h2f,0,0
-BatchNorm42,0.366976,26312.8
-BatchNorm42_f2h,0,0
-BatchNorm42_h2f,0,0
-Relu39,0.274348,26328.2
-Relu39_f2h,0,0
-Relu39_h2f,0,0
-Conv43,75.2967,719905
-Conv43_f2h,0,0
-Conv43_h2f,0,0
-Add55,0.310246,26263.4
-Add55_f2h,0,0
-Add55_h2f,0,0
-BatchNorm43,0.589254,28762.6
-BatchNorm43_f2h,0,0
-BatchNorm43_h2f,0,0
-Add56,0.129183,26278.8
-Add56_f2h,0,0
-Add56_h2f,0,0
-Relu40,0.157555,26286.4
-Relu40_f2h,0,0
-Relu40_h2f,0,0
-Conv44,36.9811,371583
-Conv44_f2h,0,0
-Conv44_h2f,0,0
-Add57,0.281792,26466
-Add57_f2h,0,0
-Add57_h2f,0,0
-BatchNorm44,0.229702,26450.6
-BatchNorm44_f2h,0,0
-BatchNorm44_h2f,0,0
-Relu41,0.154208,26439
-Relu41_f2h,0,0
-Relu41_h2f,0,0
-Conv45,86.571,848544
-Conv45_f2h,0,0
-Conv45_h2f,0,0
-Add58,0.324044,26657.4
-Add58_f2h,0,0
-Add58_h2f,0,0
-BatchNorm45,0.326054,26657.4
-BatchNorm45_f2h,0,0
-BatchNorm45_h2f,0,0
-Relu42,0.263474,26660.8
-Relu42_f2h,0,0
-Relu42_h2f,0,0
-Conv46,42.9393,428148
-Conv46_f2h,0,0
-Conv46_h2f,0,0
-Add59,0.265369,26725.2
-Add59_f2h,0,0
-Add59_h2f,0,0
-BatchNorm46,0.575724,30517.4
-BatchNorm46_f2h,0,0
-BatchNorm46_h2f,0,0
-Conv47,87.2954,864913
-Conv47_f2h,0,0
-Conv47_h2f,0,0
-Add60,0.29422,27020
-Add60_f2h,0,0
-Add60_h2f,0,0
-BatchNorm47,0.596255,29568.6
-BatchNorm47_f2h,0,0
-BatchNorm47_h2f,0,0
-Add61,0.122092,27020.2
-Add61_f2h,0,0
-Add61_h2f,0,0
-Relu43,0.160172,27020.2
-Relu43_f2h,0,0
-Relu43_h2f,0,0
-Conv48,52.3917,528531
-Conv48_f2h,0,0
-Conv48_h2f,0,0
-Add62,0.281618,27346.6
-Add62_f2h,0,0
-Add62_h2f,0,0
-BatchNorm48,0.32663,27331.4
-BatchNorm48_f2h,0,0
-BatchNorm48_h2f,0,0
-Relu44,0.224127,27331.4
-Relu44_f2h,0,0
-Relu44_h2f,0,0
-Conv49,91.6831,928447
-Conv49_f2h,0,0
-Conv49_h2f,0,0
-Add63,0.330559,27327.4
-Add63_f2h,0,0
-Add63_h2f,0,0
-BatchNorm49,0.338815,27327.4
-BatchNorm49_f2h,0,0
-BatchNorm49_h2f,0,0
-Relu45,0.259699,27327.4
-Relu45_f2h,0,0
-Relu45_h2f,0,0
-Conv50,42.0593,428805
-Conv50_f2h,0,0
-Conv50_h2f,0,0
-Add64,0.264121,27467.8
-Add64_f2h,0,0
-Add64_h2f,0,0
-BatchNorm50,0.521874,30047
-BatchNorm50_f2h,0,0
-BatchNorm50_h2f,0,0
-Add65,0.111501,27467.8
-Add65_f2h,0,0
-Add65_h2f,0,0
-Relu46,0.144992,27467.8
-Relu46_f2h,0,0
-Relu46_h2f,0,0
-Conv51,55.3189,566082
-Conv51_f2h,0,0
-Conv51_h2f,0,0
-Add66,0.286432,27642.8
-Add66_f2h,0,0
-Add66_h2f,0,0
-BatchNorm51,0.326393,27642.8
-BatchNorm51_f2h,0,0
-BatchNorm51_h2f,0,0
-Relu47,0.215097,27627.6
-Relu47_f2h,0,0
-Relu47_h2f,0,0
-Conv52,91.6897,930911
-Conv52_f2h,0,0
-Conv52_h2f,0,0
-Add67,0.331532,27715.4
-Add67_f2h,0,0
-Add67_h2f,0,0
-BatchNorm52,0.346547,27669.6
-BatchNorm52_f2h,0,0
-BatchNorm52_h2f,0,0
-Relu48,0.262815,27677.2
-Relu48_f2h,0,0
-Relu48_h2f,0,0
-Conv53,42.7118,442905
-Conv53_f2h,0,0
-Conv53_h2f,0,0
-Add68,0.264089,27745.6
-Add68_f2h,0,0
-Add68_h2f,0,0
-BatchNorm53,0.51813,30339.8
-BatchNorm53_f2h,0,0
-BatchNorm53_h2f,0,0
-Add69,0.109933,27745.6
-Add69_f2h,0,0
-Add69_h2f,0,0
-Relu49,0.148806,27745.6
-Relu49_f2h,0,0
-Relu49_h2f,0,0
-Pool2,0.74028,27745.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Mul1,1.56576,38195.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add70,0.138682,27757
-Add70_f2h,0,0
-Add70_h2f,0,0
-Softmax1,11.7675,135069
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_120.txt
deleted file mode 100644
index 35b105d97a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_120.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,111.207,1.12166e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.212219,14227.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.174597,14218.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,576.179,5.78929e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.227266,14260.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.175426,14254.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.16613,56288.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,192.271,1.94297e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.208485,14757.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.125772,14751.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,316.57,3.39462e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.216479,15116.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.172252,15107.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.36676,43823.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,109.625,1.16678e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.212504,15340.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.15622,15340.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,190.588,2.11006e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.206818,15480.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.157071,15465.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,190.242,2.12378e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.215202,15600.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.159916,15602.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.16016,31188.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,72.6437,816386
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.202409,15848.7
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.147737,15841.1
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,132.01,1.54934e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.206082,16210.3
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.147573,16212.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,132.995,1.58373e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.207378,16535.1
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.149116,16527.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.33644,18180.1
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,67.9618,814281
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.171717,16353.1
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.139711,16343.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,66.7829,804755
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.159362,16150.7
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.133334,16143.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.6739,792408
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.161289,15975.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.143932,15960.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.352235,15960.4
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.433457,17406.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0949944,15952.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0750587,15952.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.207237,15889.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0552669,15878.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.30158,18260
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_151.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_151.txt
deleted file mode 100644
index 8280ad73a8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_151.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,109.883,1.04254e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.215451,13361.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.165695,13375.1
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,285.834,2.77915e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.222789,13616.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.172303,13622.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.30943,55247.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,110.169,1.06864e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.210159,14139.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.168447,14147.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,163.861,1.67941e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.217589,14343.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.170895,14335.9
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.42049,42315.5
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,64.8149,666999
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.203035,14561.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.152188,14567.1
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,98.8989,1.05592e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.208495,14710.7
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.15037,14703.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,99.2197,1.06738e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.208639,14860.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.147836,14862.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.03008,29737.9
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,48.8777,528973
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.189513,14951
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.146172,14951
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,109.707,1.19456e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.18853,14881.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.151094,14866.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,110.228,1.19521e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.184844,14846.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.143065,14838.5
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.29591,14838.5
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,44.6806,488767
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.160754,14734.7
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.131202,14734.7
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,41.6508,457221
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.143548,14620.7
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.130312,14620.7
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,41.437,452738
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.143321,14506.3
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.125548,14506.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.260923,14506.3
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.344148,15706.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0640094,14502.5
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0617083,14502.5
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.16853,14494.9
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0530205,14483.5
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.21648,15206.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_152.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_152.txt
deleted file mode 100644
index 2bb5c473d4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_152.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,108.529,1.02679e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.209208,13383.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.167855,13359.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,282.701,2.74934e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.221096,13657.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.172502,13660.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.22041,55982
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,109.437,1.06802e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.215816,14157.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.172437,14164.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,162.933,1.66082e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.21342,14348.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.173186,14356.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.29587,43105.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,64.4613,665724
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.204095,14601.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.156335,14593.7
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,98.4478,1.05503e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.204312,14733.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.154844,14728.5
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,99.1785,1.06485e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.206322,14865.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.149935,14849.9
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.0148,29706.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,48.9101,529182
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.1767,14974
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.14277,14974
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,109.931,1.20504e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.184658,14880.1
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.14318,14865
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,110.398,1.18763e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.186411,14836.3
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.151529,14821.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.28786,14821.2
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,43.2478,471643
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.159807,14755.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.135913,14736.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.6118,444300
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.148268,14631.9
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.129462,14628.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,40.4124,440481
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.138815,14516
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.126019,14514.1
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.25027,14514.1
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.343537,15719.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0633243,14510.3
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0601566,14508.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.169481,14500.7
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0528224,14491.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.28432,18006.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_153.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_153.txt
deleted file mode 100644
index 2fc0ebfbe5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_153.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,98.1298,916184
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.214213,13206.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.169397,13212.3
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,274.817,2.6343e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.233183,13445.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.175609,13449.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.15712,54522.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,108.185,1.02872e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.224367,13944.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.172946,13939
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,162.697,1.63629e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.21757,14147.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.169247,14139.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.32408,42447.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,64.3237,648420
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.206869,14355.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.152015,14351.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,98.5349,1.03792e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.20861,14496.7
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.15237,14491
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,99.0745,1.04968e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.21197,14616.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.154018,14616.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,1.93814,29234.9
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,49.0738,521483
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.184223,14739.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.150082,14739.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,110.77,1.19663e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.190623,14672
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.145014,14666.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,111.379,1.19415e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.19268,14609.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.144002,14609.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.2826,14609.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,43.1364,466132
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.167301,14554.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.129804,14546.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.7482,440811
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.147442,14454.5
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.135721,14450.7
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,40.3888,436274
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.145516,14360.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.12935,14360.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.260203,14360.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.330046,15549.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0703355,14356.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0632124,14356.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.179161,14347.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0531199,14341.5
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.23427,18637
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_154.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_154.txt
deleted file mode 100644
index d1e7a3a32f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_154.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,98.9913,918542
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.218239,13141.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.165186,13143.5
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,275.986,2.62748e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.2159,13389.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.174021,13383.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.20344,54275.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,107.914,1.03712e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.213314,13875.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.167669,13881.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,162.166,1.62723e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.212911,14090
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.170021,14074.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.38287,40823.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,64.6072,652231
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.205071,14320.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.150616,14320.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,98.8683,1.04122e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.202149,14464.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.15302,14462.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,99.1482,1.0451e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.205432,14599.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.151234,14580.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.09322,29160.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,49.348,523166
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.181372,14688.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.142569,14681.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,110.688,1.18679e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.179871,14635.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.143346,14628.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,111.263,1.18976e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.190357,14592.3
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.150431,14592.3
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.26548,14594.2
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,44.5816,477628
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.154038,14527.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.134607,14502.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,42.0021,452542
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.143509,14418.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.128159,14402.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,41.8429,448163
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.143912,14330.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.12701,14328.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.264712,14328.3
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.340673,15511.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0637851,14320.7
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0602874,14320.7
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.169948,14301.5
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0527903,14290.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.25248,16424.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_155.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_155.txt
deleted file mode 100644
index 6e417397af..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_155.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,119.079,1.13372e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.210005,13453.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.188552,13430.7
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,355.973,3.45901e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.220786,13700.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.176869,13700.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.35758,54877.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,139.06,1.36761e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.21308,14299.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.176134,14282.9
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,218.057,2.25294e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.214319,14592.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.171458,14604.1
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.30404,43091.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,77.7244,808220
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.206946,14810.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.156415,14810.3
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,127.822,1.38616e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.207512,14921.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.156837,14923.5
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,129.446,1.38598e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.20501,14961.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.157324,14967.3
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.08451,29946.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,48.0753,522084
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.187858,15131.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.145384,15124.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,107.055,1.18645e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.181317,15033.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.145423,14263.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,107.51,1.17693e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.185138,14993
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.147769,14985.3
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.30975,14985.3
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,44.3652,492728
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.150988,14892.1
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.132361,14880.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,41.8004,465073
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.138956,14770.9
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.131059,14763.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,41.7483,459855
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.147753,14649.3
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.130773,14647.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.263538,14645.5
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.329131,15904.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0637437,14643.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0625727,14643.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.1695,14649.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0540413,14639.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.22707,16820
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_156.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_156.txt
deleted file mode 100644
index a7a3c079d3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_156.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,117.965,1.12843e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.212223,13545.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.177131,13547
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,352.332,3.44013e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.222315,13788.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.178828,13792.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.33666,55900.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,144.329,1.42745e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.217807,14370.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.167439,14361.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,227.662,2.37447e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.214821,14753.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.17828,14743.5
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.35686,44256.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,76.6474,799817
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.205151,14973.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.154582,14958.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,125.967,1.37685e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.208716,15044.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.154658,15046.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,127.389,1.3932e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.208715,15119.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.156092,15113.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.20968,30228.7
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,67.34,738285
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.188687,15261.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.149209,15254
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,151.955,1.6812e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.198655,15126.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.148448,15126.9
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,152.453,1.67297e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.198367,15084.3
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.151446,15084.3
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.3114,15835.8
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,43.612,480765
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.162652,14993
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.135993,14977.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.4848,451786
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.140524,14873.7
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.130396,14864.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,40.5455,449157
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.145813,14771.5
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.13126,14771.5
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.262312,14767.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.317774,16064.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0628124,14767.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0632797,14767.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.169733,14763.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0537853,14727.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.25415,15435.2
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_157.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_157.txt
deleted file mode 100644
index df8e948520..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_157.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,118.591,1.14231e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.211477,13518.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.175544,13526
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,365.717,3.56044e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.223941,13767.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.172991,13771.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.27455,55127.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,144.824,1.42982e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.210773,14349.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.170319,14359.5
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,228.204,2.36499e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.217378,14718.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.177535,14732.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.40404,44209.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,84.5471,887202
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.209512,14953.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.152454,14954.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,136.417,1.49071e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.207992,15110.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.154435,15102.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,136.919,1.49826e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.213855,15262.9
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.154204,15228.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.13481,30471.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,66.3913,730502
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.194143,15355.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.146505,15347.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,150.66,1.67668e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.194239,15221.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.146274,15223.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,151.261,1.66942e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.198802,15159.5
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.141375,15161.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.31095,15161.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,67.2387,742398
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.166943,15040.9
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.143394,15040.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,66.694,737648
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.163673,14898.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.136943,14146.7
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.5494,726618
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.159468,14763.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.140969,14756.1
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.335684,14756.1
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.437198,16068.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0864414,14754.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.068441,14754.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.203656,14763.7
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0602685,14727.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.37626,16915.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_158.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_158.txt
deleted file mode 100644
index 48144ad26d..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_158.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,106.983,1.00411e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.210629,13238.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.172882,13242
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,344.771,3.29724e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.220501,13510.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.169097,13498.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.22127,53964.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,135.473,1.31689e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.221522,14040.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.165369,14029
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,213.54,2.19071e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.216853,14399.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.172425,14388.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.35143,43187.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,77.0486,796273
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.208639,14597.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.150649,14599.1
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,127.247,1.35788e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.212402,14722.7
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.156031,14718.9
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,129.02,1.37963e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.207794,14768.1
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.153868,14762.3
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.10291,29534
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,48.8761,521961
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.183522,14914.3
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.151078,14899.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,109.464,1.19365e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.182351,14795.1
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.144255,14795.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,109.675,1.18545e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.200335,14763.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.153033,14755.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.29705,14757.7
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,43.2376,470940
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.160943,14658.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.133474,14651.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.5273,441563
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.147081,14588.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.131423,14588.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,40.3788,438782
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.14229,14470.9
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.127119,14469
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.266085,14469
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.322078,15695.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0704924,14467.1
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0651867,14467.1
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.183628,14461.5
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0578333,14451.9
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.25522,16626.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_159.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_159.txt
deleted file mode 100644
index a7a8b5d9a3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_159.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,107.189,1.00849e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.207365,13280.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.17292,13284
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,343.892,3.28643e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.217768,13544.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.169167,13527.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.25727,52111.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,142.501,1.38264e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.210776,14105
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.181512,13389.7
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,226.638,2.31586e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.215717,14469.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.164821,14454.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.40595,43400.5
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,77.7878,803696
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.198363,14667.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.147814,14661.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,128.387,1.37552e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.203436,14735.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.158428,14739.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,129.231,1.38054e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.204719,14786.9
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.155094,14771.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.13345,28803.5
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,66.8006,717605
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.191442,14937.5
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.144985,14929.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,150.439,1.62404e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.192338,14866.7
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.155097,14851.5
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,150.623,1.62379e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.196022,14853
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.153343,14853
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.33276,14853
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,45.0171,486081
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.155516,14746.5
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.136383,14738.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,41.9719,460951
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.146412,14617.1
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.132364,14617.1
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,42.0338,458946
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.140834,14546.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.128191,14529.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.254411,14527.7
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.338062,15790.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.064758,14523.7
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0617914,14531.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.1691,14511.9
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0533373,14488.7
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.28845,17263.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_160.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_160.txt
deleted file mode 100644
index 00bd498faf..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_160.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,107.49,999074
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.213132,13320.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.174021,13322
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,354.074,3.36895e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.228753,13538.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.167845,13544.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.29692,54894.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,140.792,1.36179e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.219112,14078.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.172956,14059.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,224.316,2.25972e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.21981,14418.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.169202,14426.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.43881,43309
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,83.9515,863295
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.205323,14653.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.146229,14648
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,136.011,1.44758e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.20973,14823.7
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.152873,14802.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,137.215,1.4667e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.212818,14941
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.152124,14942.9
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.17019,29893.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,65.5767,708698
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.189967,15086.3
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.14302,15080.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,148.706,1.62724e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.200789,14988.5
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.145301,14988.5
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,148.989,1.62081e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.198015,14995
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.148131,14987.3
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.29803,14979.7
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,67.4482,733743
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.171928,14878.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.139219,14871.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,66.67,728621
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.162904,14746.5
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.136966,14731.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.448,719795
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.160841,14595.9
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.150066,14595.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.342087,14595.9
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.438522,15877.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0986266,14595.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0763738,14595.9
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.22044,14593.9
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0572574,14555.3
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.33809,18173.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_161.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_161.txt
deleted file mode 100644
index feb7f4fb58..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_161.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,122.356,1.18071e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.217269,13550.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.177503,13548.9
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,382.136,3.71194e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.231176,13771.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.170626,13773.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.51154,59313
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,153.62,1.50244e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.216216,14278.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.166966,14273.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,240.599,2.48672e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.218568,14606.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.168764,14608
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.50083,43849.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,85.4118,885302
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.205653,14847.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.157608,14850
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,137.528,1.47878e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.215291,14998.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.16191,14996.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,138.46,1.50439e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.213874,15129.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.158985,15131.3
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.18009,30274.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,65.5398,716742
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.192725,15286.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.148821,15278.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,147.328,1.63476e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.196587,15200.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.152191,15200.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,147.655,1.62946e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.203564,15188.9
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.146652,15181.3
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.33757,15181.3
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,45.1015,498280
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.165846,15063.3
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.137698,15063.3
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,41.8254,467945
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.14821,14955.3
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.129154,14955.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,41.5928,464354
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.139353,14848.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.128457,14831.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.263067,14827.7
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.310286,16138.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0623098,14820
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0601242,14820
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.167861,14814.3
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0523582,14806.7
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.29183,15558.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_162.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_162.txt
deleted file mode 100644
index 9b5c953995..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_162.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,121.846,1.16738e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.212412,13516.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.178789,13514.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,384.141,3.7133e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.231051,13746.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.177932,13742.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.45027,57081.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,153.459,1.50471e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.214725,14267
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.171525,14267
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,240.892,2.49179e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.218776,14586.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.17662,14581
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.45359,43766.5
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,85.7184,890846
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.209708,14837.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.15669,14821.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,137.968,1.48479e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.210277,15003.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.156783,14996.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,138.221,1.48677e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.212373,15138.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.159698,14358.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.20662,28739
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,67.4394,734068
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.194104,15259.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.152735,15254
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,152.01,1.6804e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.196335,15152.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.144492,15145.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,152.442,1.67338e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.198825,15090.1
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.149068,15090.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.32739,15854.8
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,43.8544,486266
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.158716,14975.9
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.136873,14976.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.6245,452680
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.141148,14890.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.132601,14873.1
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,40.7973,451642
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.144862,14769.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.130998,14769.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.263745,14765.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.308542,16078
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0646363,14759.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0626012,14759.9
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.17296,14737.1
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.054352,14733.3
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.29299,16198.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_163.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_163.txt
deleted file mode 100644
index d440002ea7..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_163.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,123.074,1.16249e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.209051,13398.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.191187,13396.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,387.127,3.70547e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.228162,13622.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.177708,13624.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.64827,60156.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,153.745,1.47919e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.213951,14199.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.161765,14184.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,241.52,2.48585e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.216591,14533.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.173689,14537.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.50728,44349.7
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,85.5884,886377
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.203624,14767.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.154396,14769.3
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,137.718,1.48781e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.20637,14946.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.153317,14931.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,138.826,1.50322e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.215826,15089.1
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.154261,15081.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.24472,30148
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,67.0749,727506
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.203157,15219.5
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.151103,15219.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,152.012,1.66178e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.195445,15084.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.144671,15084.9
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,152.243,1.66198e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.20133,15028
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.145701,15012.3
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.36108,15768
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,67.7113,740849
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.174435,14920.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.137577,14913.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,66.9715,735257
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.161385,14763.7
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.133417,14752.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.9213,729651
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.161471,14637.9
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.133977,14637.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.353121,14637.9
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.437642,15954.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0868186,14630.3
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0680221,14630.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.198892,14607.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0548159,14590.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.32261,17525.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_164.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_164.txt
deleted file mode 100644
index 97bdfb6ee4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_164.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,122.304,1.16425e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.22244,13396.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.167529,13398.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,385.788,3.69328e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.233774,13607.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.177596,13592.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.62557,59661.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,154.367,1.49639e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.218431,14122
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.16101,14120.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,242.416,2.46897e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.212968,14442.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.170434,14446.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.59689,43382.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,86.0534,891989
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.216181,14680.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.151864,14672.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,138.333,1.48603e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.21461,14840.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.152239,14844.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,139.12,1.49611e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.210792,14988.5
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.154162,14980.5
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.24126,29963
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,63.7203,690521
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.194194,15122.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.145977,15122.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,145.538,1.60083e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.199887,15051.7
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.149733,15045.9
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,146.134,1.59581e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.196728,15021.7
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.147234,15014
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.35989,15765.3
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,67.6506,741016
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.172299,14933.9
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.137385,14909.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,67.0314,732208
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.167451,14757.7
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.14326,14742.5
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.6426,726541
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.158204,14609.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.131308,14609.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.346862,14607.3
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.433082,15904.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0896701,14603.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0712543,14603.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.204255,14599.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.054438,14586.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.31779,16777.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_165.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_165.txt
deleted file mode 100644
index 9fa8b59e7c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_165.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,112.873,1.0638e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.213458,13339.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.168226,13337.5
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,377.315,3.59071e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.230875,13527.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.177864,13520.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.50336,56198.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,150.829,1.45033e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.221214,14030.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.163929,14032.5
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,237.978,2.41404e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.213643,14351.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.164613,14355.5
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.60345,43812.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,84.4675,871314
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.202239,14601
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.141782,14589.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,136.621,1.45704e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.209455,14786
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.160031,14765.3
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,137.759,1.47485e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.212015,14933.9
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.15494,14936.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.1892,29872.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,66.57,713100
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.192287,15038.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.14533,15030.7
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,149.752,1.63643e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.199592,14973.5
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.144655,14958.5
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,150.309,1.63364e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.202968,14940.3
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.151817,14940.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.33758,16438.3
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,43.663,477395
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.159298,14856.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.135266,14848.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.7325,451800
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.14446,14759.9
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.130627,14757.9
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,40.5139,440631
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.142671,14641.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.128082,14632.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.267195,14632.2
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.318507,15929.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0672828,14628.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0653947,14628.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.180437,14634.1
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0546011,14613
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.29277,17523.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_166.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_166.txt
deleted file mode 100644
index c02af1e840..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_166.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,113.923,1.0714e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.224281,13198.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.172939,13207.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,377.565,3.55303e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.232702,13412.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.183432,13402.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.79093,59497.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,152.642,1.46039e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.218856,13923.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.170549,13925.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,242.518,2.43482e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.217657,14219.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.175279,14208
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.61946,42671.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,86.8015,876756
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.210207,14456.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.153731,14458.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,139.051,1.45895e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.214408,14599.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.158703,14603.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,140.091,1.48103e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.213349,14755.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.157353,14753.5
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.24246,29509
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,67.5266,720106
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.189298,14898
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.149983,14892.3
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,151.07,1.63582e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.198152,14828.1
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.143526,14830.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,151.253,1.62572e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.195547,14815
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.146559,14813.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.35976,17048.6
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,44.7287,483954
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.158703,14708.5
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.128092,14708.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,42.4148,463991
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.145136,14572.9
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.130265,14580.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,42.0307,456143
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.147618,14513.1
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.12759,14495.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.263502,14494
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.354842,15770.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0722334,14484.5
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.060207,14492.1
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.169365,14484.5
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0524412,14459.7
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.30475,14432.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_167.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_167.txt
deleted file mode 100644
index ac8c32b7b4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_167.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,112.969,1.0537e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.224619,13171.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.171842,13175.7
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,375.282,3.5406e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.231525,13425.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.185541,13414
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.53881,57140.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,148.81,1.4295e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.226965,13961.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.172766,13948.1
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,237.478,2.39285e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.222108,14295.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.164779,14307.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.47277,42946.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,85.8913,877468
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.209499,14525.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.144451,14519.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,137.789,1.46302e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.214536,14685
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.159196,14685
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,138.859,1.47172e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.215512,14798.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.153186,14802.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.18734,29612.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,66.1955,706900
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.199449,14958.5
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.143065,14958.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,149.895,1.62668e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.198046,14870.7
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.143423,14855.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,150.421,1.61886e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.204549,14841.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.14949,14833.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.35309,16317.6
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,67.6529,729839
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.170831,14737
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.14078,14737
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,66.8327,724828
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.162245,14603.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.138508,14603.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.8383,719891
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.161839,14465.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.13192,14454
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.33868,14454
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.441764,15755
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.100489,14455.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0759517,14463.5
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.220239,14450.1
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0604093,14446.3
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.31403,18772.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_168.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_168.txt
deleted file mode 100644
index 33036bcf77..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_168.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,111.324,1.03407e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.215211,13171.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.165667,13175
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,374.938,3.54103e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.231886,13425.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.178543,13421.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.77089,59207.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,149.667,1.43127e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.216287,13948
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.159698,13957.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,237.146,2.39731e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.214361,14282.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.171519,14288.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.53431,42885.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,84.1279,858904
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.215365,14519.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.152021,14498.3
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,136.524,1.4466e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.206584,14678.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.155285,14680.9
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,137.591,1.46436e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.211666,14814.9
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.149423,14809.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.19634,29624.1
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,64.4376,688748
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.192901,14929.7
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.14534,14172.7
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,147.632,1.60435e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.19077,14868.1
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.13885,14852.9
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,148.083,1.59761e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.197036,14846.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.142261,14831
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.35841,16301.6
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,67.7453,732092
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.174658,14725.5
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.133679,14716
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,67.1865,715178
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.161689,14611.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.133145,14603.9
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.9726,713842
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.159023,14458.5
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.131705,14446.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.338762,14446.9
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.415287,15740.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0879354,14450.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0723069,14458.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.202501,14442.9
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0578684,14438.9
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.34202,18015.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_261.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_261.txt
deleted file mode 100644
index 93c47b97c8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_261.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,43.1987,439477
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.185701,14185.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.167212,14173.7
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,365.047,3.69466e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.213368,14294.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.166652,14290.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.62853,62022.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,109.777,1.10721e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.203845,14693.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.177138,14689.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,199.71,2.12246e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.207055,14822.3
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.169183,14826.1
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.47566,43762.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,80.3271,858992
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.197515,15103.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.148857,15088.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,132.428,1.46241e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.198482,15177.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.149909,15139.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,133.266,1.4725e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.203231,15258.1
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.152713,15246.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.05306,30494.5
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,56.6371,629761
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.183782,15307.1
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.143042,15307.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,96.762,1.08074e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.192254,15468.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.144018,15463.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,97.3713,1.09327e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.1935,15628.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.143263,15621
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.37687,16416.1
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,29.2038,337845
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.159132,15690.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.131049,15690.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,27.1636,320391
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.147283,15740.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.127967,15717.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,27.0862,320402
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.14783,15795.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.128261,15795.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.273925,15795.7
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.336792,17109.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0655929,15788
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0629178,15788
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.172825,15786.1
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0526299,15763.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.3047,15743.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_262.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_262.txt
deleted file mode 100644
index 3068a69795..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_262.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,43.6399,444116
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.182037,14215.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.168831,14215.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,363.235,3.70204e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.217743,14335
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.186277,14331.3
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.69022,63146.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,109.704,1.1038e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.207161,14740.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.177455,14744.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,199.947,2.10971e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.209439,14830
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.180335,14826.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.53269,44510.5
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,80.5834,859763
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.197746,15133.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.156879,15135.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,132.891,1.47015e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.20717,15185.3
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.158905,15181.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,133.429,1.45962e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.205048,15255.5
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.159791,15253.9
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.08726,30513.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,56.6426,630938
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.184066,15349.9
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.150946,15326.7
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,96.6681,1.09416e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.186546,15497.7
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.149449,15492.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,97.1915,1.10773e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.187295,15642
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.150443,15644
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.28879,15646
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,29.1626,334532
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.160447,15723.5
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.139119,15708.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,27.0591,320761
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.150464,15761.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.132024,15761.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,27.0213,321893
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.141298,15840.3
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.130947,15817.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.274712,15809.7
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.32196,17131.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0674812,15800.1
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0629949,15792.5
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.175653,15792.5
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.057062,15777.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.27115,19731.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_263.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_263.txt
deleted file mode 100644
index 2b5bf1c1e3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_263.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,52.6999,546384
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.198037,14537.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.167414,14533.7
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,526.24,5.46434e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.212491,14757.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.167157,14757.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.60362,63177.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,155.518,1.62047e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.209423,15167.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.173631,15165.7
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,290.087,3.16438e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.212972,15275.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.175055,15281.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.51486,45826.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,112.17,1.24284e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.197132,15611.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.155762,15611
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,183.325,2.06789e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.206267,15624.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.157141,14829.7
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,183.851,2.06724e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.20941,15653.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.161016,15644.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.16421,30495
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,67.3042,765185
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.187221,15779.5
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.147635,15771.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,118.119,1.37467e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.200239,15964.3
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.145609,15968.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,118.979,1.3983e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.197608,16191.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.157228,16195
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.33445,17012.9
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,39.8238,470470
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.177077,16258.1
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.142783,16250.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,38.458,463975
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.157321,16277.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.135417,16277.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,38.7033,466826
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.158914,16259.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.137087,16236.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.328145,16236.2
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.391175,17681.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0729914,16228.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0665116,16228.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.187836,16196
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.054979,16178.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.31858,16989
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_264.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_264.txt
deleted file mode 100644
index 6c04634691..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_264.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,52.9083,548904
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.188988,14577
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.167944,14565.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,527.515,5.47719e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.223637,14765
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.184175,14763.3
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.67222,65025.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,155.107,1.62576e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.211368,15163.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.182633,15165.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,289.121,3.155e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.210738,15292.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.177333,15298.5
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.62499,45132
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,111.879,1.23559e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.198725,15603.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.155925,15597.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,182.352,2.07077e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.205867,15653.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.162169,15649.7
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,183.248,2.07693e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.209864,15688.1
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.16077,15674.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.18998,30561.9
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,67.1793,763391
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.187416,15790.9
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.151935,15787.1
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,118.002,1.37394e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.197512,16029
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.149721,16000.5
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,119.32,1.40387e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.197349,16223.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.155624,16208.5
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.36869,16208.5
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,39.8666,472541
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.166613,16269.5
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.145027,16254.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,38.6649,469343
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.155164,16289.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.140386,16251
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,38.7998,473514
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.15557,16303.3
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.140963,16280.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.321259,16280.3
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.363988,17724
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0740189,16270.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0704318,16255.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.184969,16247.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0610524,16228.5
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.31395,17837.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_265.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_265.txt
deleted file mode 100644
index 2befd2504c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_265.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,52.861,551786
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.19061,14577.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.168297,14573.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,527.678,5.49846e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.219905,14759.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.17852,14765.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.68147,63992.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,155.786,1.62992e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.204639,15171.5
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.176396,15171.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,289.384,3.16146e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.209138,15283.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.170386,15281.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.57568,45853.1
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,112.025,1.24576e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.198178,15611.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.149458,15607.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,182.531,2.07462e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.208482,15640.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.155228,15630.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,183.398,2.07862e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.205967,15678.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.155631,15680.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.12333,31353.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,67.3123,767247
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.188332,15814.1
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.14655,15783.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,118.415,1.38105e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.196828,16023.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.154956,16008
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,118.755,1.40144e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.194117,16244.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.149375,16246.7
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.32173,17056.9
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,39.546,471887
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.168028,16286.7
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.143801,16277.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,38.6609,468370
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.157347,16316
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.133135,16298.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,38.9068,474441
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.1615,16324.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.13646,16316.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.310766,16309.1
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.390116,17764.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.073539,16307.1
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.066435,16297.5
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.18516,16288
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.056915,16238.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.36499,18581.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_266.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_266.txt
deleted file mode 100644
index 9aa16c0226..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_266.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,48.9633,498139
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.188476,14343.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.164991,14324.3
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,462.396,4.69388e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.215061,14407.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.17005,14405.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.69412,64206.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,141.884,1.45206e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.208936,14813
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.176863,14807.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,261.73,2.76762e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.197772,14986.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.152431,14982.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.57217,43455.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,109.702,1.18372e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.196226,15375.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.154367,15364.3
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,171.412,1.91128e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.20068,15427.3
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.152178,15421.5
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,172.213,1.92151e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.204831,15488.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.156274,15475.3
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.13947,30956.3
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,68.3793,769846
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.189381,15641.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.139174,15641.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,119.871,1.38514e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.188025,15875.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.146639,15070.9
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,120.711,1.39485e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.200629,16125.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.145631,16118.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.32165,16120
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,46.4299,545064
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.168111,16114.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.139743,16091.3
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.1753,538166
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.156773,16061.1
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.132601,16045.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,45.9512,540783
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.156547,16019.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.133813,16017.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.322718,15224.4
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.417709,17440.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.081014,16021.5
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0698172,15224.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.197542,15982.9
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0593022,15952.1
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.34739,17536.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_267.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_267.txt
deleted file mode 100644
index d819d236d8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_267.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,50.5904,517863
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.188715,14278.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.169212,14268.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,462.506,4.68297e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.216411,14363.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.174258,14369.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.78763,64820
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,143.038,1.45618e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.213804,14752.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.176261,14748.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,263.493,2.77383e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.209698,14883.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.165583,14887.5
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.64322,43186.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,110.266,1.18977e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.202728,15279
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.156834,15265.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,171.977,1.91069e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.206021,15347.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.155583,15343.3
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,173.056,1.92271e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.202786,15420
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.156931,15406.5
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.09972,30828.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,68.7968,772608
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.188124,15559.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.147458,15551.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,120.378,1.38145e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.201964,15816.3
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.146732,15804.9
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,120.941,1.40528e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.196776,16032.5
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.148233,16024.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.32654,16816.2
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,46.7253,548974
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.171974,16053.3
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.143167,16045.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.3979,538337
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.157426,15989.9
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.136745,15982.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,45.2571,537055
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.155308,15955.9
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.136764,15955.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.320939,15955.9
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.401793,17374.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0745373,15954
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0677216,15946.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.185097,15938.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.065555,15111.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.30353,18278.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_268.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_268.txt
deleted file mode 100644
index f83aa5a470..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_268.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,49.905,511744
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.187695,14326.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.165718,14327.1
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,462.989,4.69437e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.217333,14423
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.174975,14424.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.61147,61400
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,141.977,1.44141e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.221525,14816.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.183052,14824.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,262.246,2.76769e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.207554,15004.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.159272,14997.1
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.60644,45018.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,109.646,1.19077e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.198764,15389.5
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.158476,15381.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,171.827,1.9213e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.200399,15435.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.15741,15412.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,173.02,1.93411e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.199449,15517.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.152559,15502.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.19238,30218.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,68.6716,767159
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.185068,15624.1
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.15005,15626
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,120.362,1.39028e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.191695,15864
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.151887,15856.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,121.256,1.41644e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.196053,16110.7
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.146831,16095.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.3246,16095.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,46.4277,543855
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.169686,16091.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.142594,16074.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.518,542008
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.158536,16065
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.136418,16065
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,45.2,538987
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.153132,16017.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.134201,16002.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.319528,16002.3
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.428074,17417.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0757755,15994.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0674875,16002.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.187858,16000.3
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0532381,15969.5
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.32115,18330.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_269.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_269.txt
deleted file mode 100644
index c6c446c551..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_269.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,49.0628,504810
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.189618,14396.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.170162,14370
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,462.758,4.70641e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.217557,14449.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.172396,14446.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.61854,63635.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,142.229,1.44652e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.209029,14822.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.166812,14811.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,261.43,2.77124e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.20309,15000.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.141414,14987.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.54698,44965.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,109.407,1.19114e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.190463,15387.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.143935,15368.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,171.196,1.91364e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.197474,15456
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.159666,15450.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,171.828,1.91883e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.201657,15532.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.158601,15509.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.06754,31025.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,68.4646,768095
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.186047,15679.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.143929,15671.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,119.875,1.37399e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.193714,15896.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.149308,15890.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,120.764,1.41222e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.193797,16146.9
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.143548,16146.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.3176,16955.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,46.5517,549285
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.16548,16142.9
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.140348,16142.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.1182,540106
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.153814,16093.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.1375,16093.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,45.1643,540580
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.153097,16069.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.134207,16069.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.306737,16060
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.406349,17482.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0792187,16036.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0665277,16044.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.185083,16035
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0555928,16017.7
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.3666,18389.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp32_perf_fp32_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp32_perf_fp32_120.txt
deleted file mode 100644
index abb7299681..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp32_perf_fp32_120.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,154.154,1.54199e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.220064,13668.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.13983,13662.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1317.67,1.13951e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.23238,11673.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.214768,11673.1
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,8.51273,72467
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,329.448,2.71108e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.217445,12526.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.124732,12532.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,436.005,4.05173e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.226307,13727.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.135171,13737.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,5.7522,61258.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,153.512,1.50259e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.219411,14562.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.169039,14539
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,288.878,3.01868e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.207609,15309.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.11494,15307.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,267.814,2.9785e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.212941,16040.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.163091,16021.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.73827,34424.5
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,112.878,1.27665e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.208707,16607
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.158825,16599.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,217.151,2.64636e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.21327,17216.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.154083,17212.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,197.886,2.48057e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.218889,18076
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.156406,18069.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,2.04499,35244
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,95.0991,1.20515e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.187971,17730.3
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.143526,17715.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,87.9554,1.13323e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.185939,17476
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.141891,17451.9
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,91.9726,1.16314e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.178909,17143.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.139305,17133.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.401583,17132
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,1.19045,19597.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.133037,17132
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.129558,17124.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.382002,17109.1
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.115561,17099.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.06808,17067.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_120.txt
deleted file mode 100644
index ecd5f1f273..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_120.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,105.152,1.06919e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.21909,14374.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.177375,14376.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,561.762,5.73341e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.232674,14455.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.193036,14449.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,4.83179,57022.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,187.585,1.9176e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.210258,14954.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.12759,14958.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,308.952,3.36508e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.213938,15426.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.185135,15418.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.13165,39204.1
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,106.808,1.1383e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.208075,15658.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.161487,15658.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,185.985,2.10699e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.209723,15859.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.161192,15846.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,186.116,2.12537e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.212821,16028.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.172169,16022.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.01809,32051.1
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,71.5588,819917
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.199356,16227.3
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.149874,16206.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,129.81,1.55325e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.232808,16487
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.155906,16464
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,130.779,1.57442e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.201823,16709.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.150386,16701.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.26897,16701.6
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,67.4552,811202
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.172434,16504
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.141912,16496.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,66.4187,805044
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.168658,16281.7
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.140319,16272.1
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.0525,792083
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.166613,16093.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.14109,16078.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.362695,16078.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.441015,17518.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0906938,16071
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.075007,16078.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.206635,16027
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0573946,16004.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.45245,29565.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_151.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_151.txt
deleted file mode 100644
index c526f42320..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_151.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,107.388,1.01178e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.217218,13386
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.176124,13391.7
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,278.896,2.7073e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.22901,13702.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.176969,13689.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,4.89328,54801.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,108.141,1.05324e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.217774,14204.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.174108,14210.5
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,161.17,1.65817e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.224632,14471
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.180677,14463.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.26071,41930.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,64.1092,645366
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.223637,14708.5
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.155749,14687.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,97.7646,1.05705e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.211176,14884.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.155685,14872.9
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,98.1629,1.06561e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.218981,15025.5
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.154259,15017.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.09586,31586.3
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,48.7943,519587
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.187666,15119.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.148316,15111.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,109.364,1.20515e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.19005,14980.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.153285,14965
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,109.637,1.19884e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.203541,14885
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.153398,14885
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.29977,14886.9
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,44.3599,486354
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.159106,14765.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.133414,14758
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,41.5714,458799
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.147109,14647.5
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.132844,14639.9
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,41.5871,454928
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.144597,14508.1
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.126316,14506.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.257944,14502.4
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.343559,15689
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0686942,14510
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.061484,14510
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.17542,14502.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.054467,14483.3
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.33096,24739.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_152.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_152.txt
deleted file mode 100644
index 22b1e98bab..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_152.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,107.908,1.02402e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.212172,13402.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.186924,13406.9
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,277.563,2.6994e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.228968,13714.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.173701,13702.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,4.86616,54839.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,107.244,1.04932e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.22834,14249.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.175698,14247.1
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,159.887,1.65401e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.22323,14512.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.173961,14520.9
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.20118,42108.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,63.708,657531
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.213141,14765.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.156402,14765.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,96.9116,1.05306e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.20861,14935.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.154582,14937.7
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,97.7397,1.06611e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.210031,15090.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.155897,15086.5
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,1.95418,30190.7
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,48.6037,530072
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.185836,15168.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.172722,15168.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,109.428,1.21203e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.193663,15044.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.145718,15037
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,109.937,1.20456e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.196133,14906.5
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.146338,14900.7
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.29171,14885.3
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,43.267,474744
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.155106,14799.7
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.132636,14799.7
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.6441,447564
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.150284,14691.3
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.128828,14689.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,40.4371,444369
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.142124,14594.3
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.128037,14594.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.25515,14592.4
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.319064,15796.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0651101,14588.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0606424,14586.7
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.178978,14571.3
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0551675,14558
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.26003,23364.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_153.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_153.txt
deleted file mode 100644
index b3b55cf48c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_153.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,96.0034,900818
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.211548,13286.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.179759,13276.9
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,267.491,2.58029e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.231195,13563.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.1787,13570.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,4.94008,53273.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,105.705,1.02375e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.218053,14068.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.181458,14066.7
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,158.017,1.59746e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.21758,14317.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.188793,14323.1
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.11806,37617
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,63.4123,639537
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.200844,14561.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.1607,14553.3
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,96.4935,1.03518e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.203944,14741.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.153634,14743.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,97.3106,1.04844e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.209634,14918.5
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.154399,14914.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,1.92911,29829.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,48.6112,527726
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.179531,15000.9
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.146549,14985.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,109.788,1.20457e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.192236,14862
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.148607,14846.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,110.181,1.19503e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.186904,14719.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.149225,14719.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.27559,14721.5
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,43.1905,471396
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.161407,14649.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.136101,14641.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.5769,443246
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.14295,14546.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.132989,14546.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,40.4194,437376
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.145609,14460.1
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.128675,14452.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.265003,14450.5
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.338891,15639
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0663772,14446.7
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0625948,14437.1
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.175772,14443
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0563163,14410.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.46747,27308.1
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_154.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_154.txt
deleted file mode 100644
index 92ec828264..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_154.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,97.3448,902774
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.210172,13227.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.170262,13227.5
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,268.012,2.57212e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.224734,13479.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.178885,13478.1
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.01295,54678.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,105.258,1.01806e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.218056,14017.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.174386,14024.9
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,157.903,1.60454e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.221938,14271.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.178588,14269.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.2309,38459.7
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,63.5376,644852
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.205026,14492.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.154486,14489.1
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,96.3055,1.01845e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.210549,14673.3
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.15438,14679
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,97.519,1.04651e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.213899,14821.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.148971,14821.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.04105,29658.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,48.7265,529032
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.183804,14962.9
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.147218,14949.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,110.53,1.20448e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.19061,14805.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.147106,14774.9
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,110.335,1.19139e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.191861,14719.5
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.146937,14719.5
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.29885,14719.5
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,44.3436,482803
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.158427,14615
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.132246,14615
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,42.107,458106
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.156572,14494.9
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.129564,14493
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,41.6751,449618
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.14228,14373.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.131922,14371.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.258302,14366.2
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.318139,15539.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0647836,14364.3
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0625307,14364.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.171909,14358.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0550554,14345.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.40385,25968.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_155.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_155.txt
deleted file mode 100644
index 189a292990..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_155.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,114.652,1.10512e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.208811,13621.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.170303,13606
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,342.146,3.37742e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.21853,13933.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.170575,13933.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,4.84349,55767.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,134.622,1.34003e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.212524,14525
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.173682,14523.1
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,210.24,2.22411e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.208926,14928.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.164172,14917.5
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.1884,41744.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,75.4834,810192
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.199346,15154.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.14182,15150.9
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,124.381,1.38114e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.206242,15277.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.154275,15275.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,125.539,1.37689e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.20549,15403.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.155395,15397.9
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,1.97444,30801.5
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,46.9719,515309
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.189151,15460.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.146434,15462.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,105.6,1.19399e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.18374,15303.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.146457,15303.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,106.068,1.18485e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.181388,15172.3
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.142831,15174.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.23698,15174.2
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,44.3997,498499
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.153365,15051.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.128223,15051.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,41.414,467223
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.142194,14939.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.124713,14931.7
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,41.3199,462608
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.141583,14833.9
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.125522,14816.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.259893,14812.8
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.308267,16073.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0629309,14808.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0599613,14808.9
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.16757,14820.3
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.053683,14793.9
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.48006,28157.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_156.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_156.txt
deleted file mode 100644
index 04a89e57ad..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_156.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,114.563,1.10576e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.212958,13613.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.181583,13602.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,340.917,3.36151e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.225963,13930
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.176498,13937.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,4.9187,56460.7
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,140.943,1.38834e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.216638,14542.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.177244,14528.9
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,223.057,2.34305e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.210335,14963.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.179829,14957.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.11012,41106.1
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,75.249,799106
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.207253,15192.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.152751,15173.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,123.141,1.36856e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.206875,15321.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.153938,15321.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,124.296,1.38348e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.209314,15443.7
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.154937,15422.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.07298,30851.3
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,66.0923,740815
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.1879,15521.5
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.148905,15513.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,150.117,1.6834e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.198159,15337.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.15086,15309.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,150.573,1.66636e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.19173,15167
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.145801,15157.5
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.2818,15916.7
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,43.7546,484715
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.158777,15050.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.132258,15033.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.5319,456047
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.150447,14926.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.128162,14926.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,40.3782,449441
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.143823,14824.9
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.135957,14815.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.269835,14811.3
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.330065,16106.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0638014,14809.3
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0603612,14809.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.169862,14801.7
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0552574,14795.9
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.42526,27375.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_157.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_157.txt
deleted file mode 100644
index 3355365dc0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_157.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,115.17,1.10581e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.214194,13640.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.192386,13615.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,352.209,3.47654e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.230485,13933.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.170828,13941.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,4.84019,55787.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,140.527,1.39151e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.213058,14547.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.169804,14540.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,222.489,2.35001e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.213909,14961.5
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.166185,14965.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.19234,41113.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,82.8968,887210
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.199749,15221.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.149455,15213.9
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,133.074,1.4816e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.20723,15405.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.15884,15407.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,134.342,1.49634e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.20782,15601.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.155199,15570.9
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,1.96713,31145.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,65.5628,733081
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.194386,15675.7
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.151017,15668.1
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,149.146,1.69338e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.194975,15453
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.148482,15437.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,149.707,1.67093e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.192037,15269.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.15029,15261.5
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.28831,16018.5
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,66.8287,746356
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.175512,15126.7
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.142684,15126.7
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,66.4599,737366
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.173397,14956.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.142738,14956.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.0923,727416
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.160898,14820.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.134687,14803.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.342158,14803.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.443492,16116
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0908923,14801.7
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0679643,14801.7
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.206552,14815
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0579709,14790.3
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.50242,28888.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_158.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_158.txt
deleted file mode 100644
index d519f756fb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_158.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,104.005,982620
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.21652,13347.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.169279,13339.7
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,334.014,3.21074e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.221484,13673.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.170626,13677.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,4.94483,54335.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,132.86,1.30108e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.214875,14223.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.17502,14211.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,210.266,2.16598e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.214658,14610.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.156284,14608.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.1546,40850.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,75.9882,804125
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.196178,14830.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.140562,14826.9
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,125.195,1.36587e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.205058,14977.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.15524,14962.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,126.555,1.3718e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.209429,15067.7
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.153865,15062
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,1.96234,30133.5
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,48.135,524368
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.190892,15149.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.147606,15141.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,108.527,1.20022e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.186191,14991.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.146015,14983.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,108.975,1.18932e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.194197,14878.9
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.146274,14871.3
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.25158,14873.2
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,43.5028,476529
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.165455,14791.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.137708,14789.7
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.7184,449443
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.147759,14685.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.133196,14065.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,40.3824,440233
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.14014,14590.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.13109,14582.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.259518,14578.7
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.320593,15807.1
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0654427,14578.7
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0630333,14574.7
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.171532,14561.3
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0548188,14557.5
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.35723,25528.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_159.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_159.txt
deleted file mode 100644
index 84702c96ff..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_159.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,104.253,985547
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.207788,13312.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.173077,13310.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,334.204,3.22428e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.219388,13655.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.175544,13649.9
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,4.91091,54652.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,138.704,1.35367e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.211772,14223.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.176668,14226.9
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,220.793,2.28541e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.21269,14654.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.175724,14639.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.17622,40976.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,76.6529,798480
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.20397,14842.5
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.154671,14844.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,126.51,1.37455e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.20499,14949.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.158063,14949.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,127.097,1.38155e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.204562,15060.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.17084,15046.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.04432,30093
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,65.8159,717678
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.185471,15145.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.162956,15131.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,148.493,1.63189e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.191887,14987.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.149036,14987.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,149.129,1.60367e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.198638,14896.5
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.154646,14888.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.29081,14888.9
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,45.1633,493666
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.161356,14797.9
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.134595,14780.7
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,41.6363,459435
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.138162,14657.3
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.133961,14657.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,41.8212,456622
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.147074,14577.5
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.132837,14569.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.271144,14567.8
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.330119,15834.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0699677,14559.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.065619,14565.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.179292,14559.7
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0594974,14536.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.41956,26942.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_160.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_160.txt
deleted file mode 100644
index 7cca3d0807..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_160.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,104.742,992497
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.21037,13347.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.16525,13342
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,344.837,3.32325e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.220447,13617.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.178066,13608.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,4.932,52311.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,137.356,1.34175e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.218773,14181.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.170047,14173.7
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,218.658,2.25012e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.215397,14600.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.171292,14604.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.1774,41599.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,82.2516,853073
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.207183,14859.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.153026,14851.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,133.087,1.44949e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.210549,15069.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.155439,15048.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,133.645,1.46219e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.214248,15231.1
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.158274,15234.9
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,1.98313,30477.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,64.8349,714762
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.189154,15372.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.154758,15349.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,147.444,1.64201e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.197356,15130.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.147513,15130.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,148.167,1.63163e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.197666,15023.5
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.14807,15017.7
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.31627,15017.7
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,67.0279,734773
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.174793,14903.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.13868,14903.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,66.4437,728622
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.161017,14773.5
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.139244,14765.9
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.4334,719978
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.16756,14627
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.135622,14627
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.348855,14625
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.445341,15906.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0952541,14625
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0703513,14625
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.214754,14617.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0575292,14595.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.50441,28527.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_161.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_161.txt
deleted file mode 100644
index 74007f27b8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_161.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,118.468,1.14623e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.214315,13651.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.175205,13644
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,374.619,3.66935e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.225039,13876.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.174059,13869.1
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,4.88226,55459.3
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,148.591,1.47676e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.211617,14452.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.172885,14447.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,233.337,2.43933e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.21253,14816.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.164812,14829.5
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.09059,39985
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,82.7071,877005
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.199592,15105.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.143519,15099.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,133.588,1.47736e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.2094,15344.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.153999,15336.5
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,134.532,1.49512e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.20918,15536.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.189119,15536.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.03014,31081.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,64.3475,722642
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.190492,15614.7
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.14638,15599.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,145.423,1.64795e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.193577,15447.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.151512,15449.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,146.18,1.63677e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.198859,15326.1
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.145824,15318.5
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.2778,16085.1
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,44.7669,499811
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.162642,15205.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.134566,15175.3
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,41.8066,474123
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.149816,15052
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.128774,15036.7
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,41.5007,466513
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.145311,14948.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.128626,14948.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.266741,14940.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.323131,16235.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0658558,14933
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0638906,14931
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.180242,14921.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0556859,14894.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.45246,27563.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_162.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_162.txt
deleted file mode 100644
index f7efa42363..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_162.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,118.741,1.14215e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.211128,13546.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.179547,13562
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,376.67,3.66557e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.221951,13796.7
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.17246,13806.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.12953,56655
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,148.969,1.46177e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.213608,14386.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.179263,14395.7
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,233.791,2.4429e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.213327,14786.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.169298,14794.1
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.28015,42155.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,83.645,880490
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.205535,15033
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.153039,15038.7
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,134.725,1.48053e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.206322,15275.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.159394,15269.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,135.171,1.49384e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.208434,15464.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.150521,15454.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.05906,30917.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,66.57,743440
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.189336,15540.5
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.151551,15525.3
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,150.398,1.69329e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.193999,15353.3
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.145695,15330.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,151.252,1.6769e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.194469,15178.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.148597,15153.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.2828,15153.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,43.8369,485190
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.158233,15065.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.136079,14287.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.8194,454889
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.145096,14950.3
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.130073,14933.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,40.3038,449279
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.142847,14845.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.128367,14824.5
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.266353,14824.5
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.316289,16117.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0665598,14820.7
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0616826,14820.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.177961,14820.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.058387,14809.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.50631,28880.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_163.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_163.txt
deleted file mode 100644
index 18b0d55b59..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_163.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,118.318,1.12989e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.210421,13493.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.179951,13493.9
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,375.458,3.64614e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.220639,13794.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.169705,13792.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.11998,55876.9
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,148.871,1.46612e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.21539,14357.9
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.172556,14321.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,234.573,2.43904e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.209803,14795.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.165148,14791.9
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.24031,40255.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,83.4016,881618
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.203858,15051.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.139942,15051.9
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,134.274,1.47313e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.2066,15250
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.155855,15252
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,134.932,1.48941e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.210235,15437.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.156099,15422.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.07714,30856.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,65.9037,731946
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.197762,15546.3
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.148146,15548.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,149.404,1.67873e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.200399,15316.7
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.1454,15309
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,150.301,1.66482e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.200472,15174.5
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.143446,15176.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.28163,15178.3
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,66.7982,737711
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.171948,15025.1
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.138243,15023.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,66.5307,732768
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.166908,14868.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.137001,14851.1
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.4203,725713
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.159138,14708.5
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.136812,14693.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.339854,14693.3
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.437914,15990.5
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0918619,14693.3
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0708121,14699
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.211887,14689.5
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0560191,14670.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.50579,28671.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_164.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_164.txt
deleted file mode 100644
index 9f999df508..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_164.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,119.105,1.13978e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.211033,13430.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.167749,13438.5
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,375.932,3.64148e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.229423,13737.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.176364,13735.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.13915,54962.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,149.104,1.46648e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.21909,14292.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.170073,14294.5
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,234.99,2.43435e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.212175,14702.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.162175,14683.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.36468,44082
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,83.4489,880213
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.219653,14952.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.147356,14942.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,135.007,1.47847e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.209525,14399.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.156454,14405.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,135.414,1.47701e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.211666,15344.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.162901,15346.1
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.11945,30699.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,62.6361,693395
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.193426,15437.7
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.145029,15430.1
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,143.801,1.61282e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.201596,15272.7
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.143922,15265.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,144.863,1.60107e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.201522,15151.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.143903,15136.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.29956,15136.1
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,66.9644,740178
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.167829,15037.1
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.137228,15027.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,66.7179,738882
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.162258,14873.9
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.136319,14864.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.4668,728791
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.161548,14741.3
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.134018,14716.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.355752,14716.3
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.426999,16011.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0889654,14710.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0696798,14712.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.196501,14698.9
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0560892,14674
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.51014,28632.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_165.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_165.txt
deleted file mode 100644
index 4adb029c93..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_165.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,110.403,1.0538e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.206396,13407.9
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.172418,13381.3
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,369.42,3.51865e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.232094,13605.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.184005,13594.3
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.25174,55789.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,146.687,1.42799e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.215631,14137.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.166502,14139.1
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,232.684,2.38229e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.211449,14547.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.166805,14543.9
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.31008,42195.3
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,82.879,866005
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.20573,14781.5
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.149058,14781.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,133.915,1.44632e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.208213,15000
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.163017,14996.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,134.348,1.46492e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.208552,15225.3
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.156473,15217.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.11087,30446.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,65.5125,719686
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.199349,15307.1
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.148121,15307.1
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,148.209,1.6465e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.197042,15149.1
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.149401,15126.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,148.888,1.63166e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.201605,15025.7
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.150009,15018
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.29136,15760
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,43.3923,474948
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.161455,14925.5
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.137384,14910.3
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,40.3975,451244
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.144238,14784.3
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.133177,14784.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,40.5106,447381
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.146268,14703.3
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.127609,14693.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.258312,14689.8
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.337,15969.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0632411,14685.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0661947,14683.9
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.178165,14674.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0551931,14668.5
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.50292,28539.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_166.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_166.txt
deleted file mode 100644
index 743386a526..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_166.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,110.662,1.03491e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.217679,13276.5
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.17133,13261.1
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,367.799,3.51004e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.22844,13533.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.181439,13528.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.27467,55958
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,147.144,1.42659e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.226341,14089.7
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.168815,14093.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,233.945,2.39505e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.216162,14470.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.164748,14468.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.30686,40888.7
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,84.0718,870818
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.207064,14727.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.147397,14717.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,134.92,1.457e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.206853,14954.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.151842,14941.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,136.018,1.47743e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.217714,15149.5
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.157173,15143.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.12146,29513.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,66.5401,727501
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.198636,15232.9
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.14662,15234.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,149.575,1.65533e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.196645,15065.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.146306,15060
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,150.11,1.63542e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.197458,14941.9
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.142492,14939.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.32995,15681.8
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,44.9229,489101
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.163202,14830.1
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.135376,14820.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,42.0917,462904
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.14365,14704.5
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.129526,14689.3
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,41.892,456969
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.138956,14603.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.124233,14595.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.26276,14593.9
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.325396,15871.8
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0643294,14582.3
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.061238,14580.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.179935,14565
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0559071,14553.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.4314,26993.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_167.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_167.txt
deleted file mode 100644
index 0c0ce45f0e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_167.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,109.543,1.03105e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.212667,13238
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.172652,13238
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,365.749,3.47448e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.229461,13535.3
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.176459,13524
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.30098,55153.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,146.898,1.42171e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.220315,14084
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.17229,14078.3
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,232.114,2.36594e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.230623,14460.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.18078,14464.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.52148,43402.7
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,84.4898,873339
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.208968,14735.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.152662,14714.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,135.934,1.46153e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.20965,14923.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.156329,14910.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,136.37,1.46507e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.21453,15097.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.159624,15089.9
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.14539,30952
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,65.7065,720241
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.193589,15194.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.14613,15179.3
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,148.704,1.63405e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.200008,14987.3
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.152262,14989.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,149.107,1.62284e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.197483,14903.9
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.149141,14903.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.32388,15556.3
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,67.2184,727451
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.172287,14775.1
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.150789,14759.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,66.6351,725542
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.161106,14622.3
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.134348,14607
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.3955,716563
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.158357,14503.9
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.133651,14503.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.338183,14502
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.433975,15787.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0982811,14502
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0712315,14500.1
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.206466,14496.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0560892,14477
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.49951,28276.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_168.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_168.txt
deleted file mode 100644
index b99d2b30cb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_168.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,109.964,1.02782e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.209144,13222.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.169343,13224.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,368.175,3.47789e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.227381,13499.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.176818,13511.3
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.26897,54774.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,146.055,1.40689e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.221614,14043.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.173208,14049.5
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,232.12,2.35205e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.213768,13728.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.173074,13724.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.28884,39738.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,82.6531,845603
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.208508,14704.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.15445,14698.9
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,134.491,1.42239e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.230827,14908
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.164619,14904.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,134.587,1.43033e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.211865,15074.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.158169,15074.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.08325,30154.9
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,63.6095,673646
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.191759,15173.7
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.144873,15173.7
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,146.355,1.61068e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.1958,15011.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.148201,14254.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,147.031,1.59311e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.198533,14895.5
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.146693,14890
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.30927,15534.8
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,67.098,716002
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.170178,14771.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.138127,14771.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,66.6769,723222
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.16181,14638.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.138575,14636.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,66.7153,718120
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.161426,14494.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.137394,14494.7
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.350289,14494.7
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.444647,15757.3
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0872635,14490.7
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0668315,14488.7
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.202018,14482.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0601662,14465.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.49285,28251.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_261.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_261.txt
deleted file mode 100644
index 2c67cf74bc..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_261.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,41.6586,423473
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.187628,14233.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.176332,14223.5
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,358.486,3.66111e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.214203,14409.9
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.186911,14396.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.2347,59034.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,108.454,1.1053e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.207125,14796.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.17886,14798
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,196.633,2.11168e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.209125,14958.1
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.179363,14958.1
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.34564,44910.7
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,79.9596,867722
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.194216,15261.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.157282,15238.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,130.911,1.45884e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.202597,15359.5
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.162287,15336.7
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,131.525,1.46554e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.205582,15443.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.159151,15430.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,1.97105,30845.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,55.995,627529
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.182146,15494.7
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.151023,15494.7
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,95.7772,1.09279e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.186005,15612.5
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.149868,15584.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,96.2079,1.10512e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.187922,15707.1
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.150709,15693.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.2862,15695.9
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,29.1307,337100
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.159001,15768.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.142741,15768.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,26.9254,321724
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.147507,15836.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.132421,15824.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,26.7507,320416
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.145679,15874.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.133932,15866.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.265982,15865
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.341304,17188.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0655133,15863.1
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0640187,15868.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.17404,15859.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0565371,15828.5
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.47606,29997.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_262.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_262.txt
deleted file mode 100644
index f14d74e574..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_262.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,43.1483,438855
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.186162,14160.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.171186,14162.1
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,356.871,3.63702e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.217039,14342.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.172565,14346.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.31939,58857.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,108.839,1.10217e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.200377,14765.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.176156,14769.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,196.699,2.10974e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.209192,14919.7
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.172882,14922.1
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.41312,42508
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,79.8326,866160
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.19389,15208.1
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.151925,15208.1
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,130.48,1.4529e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.202415,15350
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.152498,15344.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,131.45,1.4663e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.199487,15397
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.155215,15393.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.06297,30770.3
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,56.2601,633310
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.177609,15487.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.149679,15468.1
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,95.6505,1.09524e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.187724,15589.3
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.143983,15578
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,96.5862,1.10719e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.185352,15675.9
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.146383,15655.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.26552,15655.1
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,29.1971,336277
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.157839,15719.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.143161,15709.5
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,27.1253,321619
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.150597,15778.5
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.135462,15765.1
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,26.9022,318713
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.150757,15811.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.132566,15811.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.263371,15811.2
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.34292,17127
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0661213,15803.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0634942,15805.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.17012,15780.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0542846,15763.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.50226,30795.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_263.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_263.txt
deleted file mode 100644
index 247eb7b675..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_263.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,50.9234,535554
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.195403,14636.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.175452,14620.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,518.426,5.43526e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.216277,14809.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.17759,14817.1
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.42519,61606.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,153.413,1.61083e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.220498,15255.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.213352,15243.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,285.114,3.14387e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.210978,15424.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.178655,15426.7
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.39062,46304.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,110.742,1.24232e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.200325,15777.7
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.159061,15764.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,179.691,2.06888e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.208322,15842.7
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.159765,15831.3
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,180.557,2.0787e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.212366,15884.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.160652,15869.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.03828,31748.3
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,66.7223,767171
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.193103,15989.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.149388,15974
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,116.697,1.36662e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.198892,16141.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.147193,16133.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,117.454,1.39289e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.20044,16298.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.148207,16279.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.2809,16271.7
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,39.1437,468696
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.171871,16351.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.145081,16344.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,37.9035,462892
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.154028,16350.1
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.140879,16350.1
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,38.3762,468250
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.160856,16378
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.139458,16370.3
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.325403,16370.3
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.390788,17815.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0718173,16353
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0646715,16345.3
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.179743,16337.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0558013,16312.7
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.54232,32547.9
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_264.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_264.txt
deleted file mode 100644
index 878f8d8716..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_264.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,50.0958,528452
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.191519,14650.7
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.168265,14641.5
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,520.446,5.44998e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.219042,14838.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.18717,14832.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.47338,60801.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,153.435,1.61459e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.207189,15238.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.17886,15240
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,286.302,3.13869e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.210693,15386.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.178072,15392.3
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.40118,46257
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,110.89,1.23083e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.206536,15762.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.152713,15754.7
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,180.001,2.05725e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.209068,15808.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.156223,15812.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,180.808,2.06412e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.209374,15878.9
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.156092,15882.7
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.06083,31759.7
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,66.3801,770597
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.187692,15987.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.146994,15979.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,116.85,1.37985e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.195074,16154.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.151641,16149
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,117.299,1.39156e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.199852,16304.1
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.149961,16290.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.29325,16290.8
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,39.5039,472278
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.175125,16359.3
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.143804,16336.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,38.041,463769
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.153064,16356.5
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.132758,16348.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,38.424,469832
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.153106,16374.3
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.13542,16359
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.308513,16359
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.359895,17810.2
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0677469,16355.1
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0626397,16339.7
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.181554,16337.7
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0631007,16318.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.49855,31858.5
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_265.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_265.txt
deleted file mode 100644
index 36826ba6b9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_265.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,50.5423,515981
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.198585,14612.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.175669,14613
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,519.793,5.43925e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.217084,14836.1
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.179282,14822.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.45173,61576.4
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,153.575,1.61708e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.210207,15253.3
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.175103,15245.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,286.184,3.15231e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.211147,15413.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.175359,15407.5
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.47327,46233.9
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,110.892,1.22944e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.199208,15754.9
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.157708,15754.9
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,179.987,2.06213e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.206335,15835.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.15973,15831.3
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,180.9,2.07261e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.208188,15880.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.167141,15873.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.05679,31752.1
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,66.9345,771288
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.189138,15930
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.154322,15931.9
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,116.787,1.36718e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.197935,16112.7
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.149583,16107
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,117.508,1.38856e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.198248,16269.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.153516,16269.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.3271,16267.8
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,39.253,454324
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.173289,16321.1
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.147554,16305.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,38.1583,449501
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.159545,16317.7
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.140188,16317.7
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,38.5135,467491
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.158671,16341.7
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.139055,16326.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.331435,16326.4
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.347447,17051
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0876053,16318.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0679453,16320.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.185029,16308.7
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0602749,16308.7
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.50535,31811.6
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_266.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_266.txt
deleted file mode 100644
index 45da23840f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_266.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,48.0955,491356
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.191413,14327.3
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.172201,14321.7
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,454.861,4.63547e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.220444,14471.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.179081,14471.7
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.67009,62297.1
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,140.383,1.42932e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.213221,14891.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.183014,14887.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,258.397,2.75077e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.216508,15089.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.176981,15093
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.4418,45302
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,108.651,1.18832e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.206623,15510.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.171596,15516.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,169.658,1.91587e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.210312,15607.1
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.159036,15609
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,170.257,1.91032e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.207119,15681.1
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.159353,15679.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.04852,31350.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,67.8856,758920
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.195292,15799.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.152092,15801.1
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,119.303,1.38995e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.205631,15950.3
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.146229,15954.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,120.097,1.39083e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.20164,16125.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.148373,16120.1
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.32921,16122
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,46.3915,547948
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.178716,16108.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.139413,16108.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.2391,541629
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.159026,16059.1
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.140956,15241.1
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,45.4586,542408
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.163135,16044.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.14157,16036.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.342292,16029
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.412724,17430.4
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0718044,16025.1
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0707743,16032.7
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.190146,16007.7
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0562845,15969.7
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.53685,31872.4
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_267.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_267.txt
deleted file mode 100644
index 1244da47a9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_267.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,48.3476,494655
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.199381,14333.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.17701,14322.1
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,455.144,4.63328e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.217317,14442.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.182137,14450.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.75186,63638.8
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,140.298,1.42313e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.211183,14877.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.175836,14880.7
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,258.572,2.75308e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.213746,15082.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.179867,15087.9
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.54198,45280.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,108.426,1.18182e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.220027,15493.3
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.160268,15487.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,169.827,1.90984e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.212972,15587.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.163737,15584.1
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,169.68,1.91164e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.214572,15687
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.162475,15671.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.12413,31351.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,67.8137,771936
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.197829,15793.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.154044,15785.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,119.404,1.38237e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.200329,15955.9
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.151564,15940.7
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,120.341,1.4096e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.204632,16135.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.158905,16127.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.33683,16945.6
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,46.5145,548025
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.177717,16100.9
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.147538,16100.9
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.2055,540629
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.162277,16047.7
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.142898,16040
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,45.2164,540444
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.159416,16009.9
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.140572,16007.9
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.351511,16007.9
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.403469,17430.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0745818,16000.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0660413,16000.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.183557,15998.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0557373,15965.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.50264,31135.3
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_268.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_268.txt
deleted file mode 100644
index 83fb5487c9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_268.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,46.9584,482258
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.188003,14421.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.172437,14408
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,454.234,4.64545e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.218751,14535
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.174255,14537
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.64122,63324.5
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,139.332,1.43474e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.208296,14950.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.178604,14925.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,256.972,2.74482e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.203167,15159
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.165541,15149.5
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.38018,44708.7
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,108.133,1.17891e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.197692,15569.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.155263,15556
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,168.781,1.91535e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.209381,15658.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.161279,15654.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,169.877,1.91614e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.200197,15749.7
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.162008,14958.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.01499,29922.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,67.6277,768950
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.186021,15854.5
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.147129,15854.5
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,118.852,1.38048e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.199301,16026.7
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.146917,16019.1
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,119.584,1.40499e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.200978,16202
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.148972,16203.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.2899,16198.2
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,46.1641,545492
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.173177,16188.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.138792,16171.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,44.9602,537586
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.15638,16133.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.136962,16133.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,44.9592,538474
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.156457,16098.3
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.13454,16090.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.319717,16082.9
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.428202,17501.7
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0808825,16082.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0676285,16082.9
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.195874,16073.3
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.0597115,16055.9
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.50665,31335
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_269.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_269.txt
deleted file mode 100644
index b1d756884b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_269.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,48.5249,496066
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.192949,14310.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.17317,14307.1
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,455.058,4.63206e+06
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.217346,14457.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.182591,14451.5
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,5.5584,60076.6
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,140.254,1.42519e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.212344,14875.1
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.176223,14878.9
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,258.159,2.73933e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.211557,15083.9
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.167519,15081.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,3.49324,44516.7
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,108.704,1.18796e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.202456,15498
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.164021,15486.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,169.675,1.90747e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.206693,15580.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.166293,15572.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,171.573,1.92222e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.205378,15648.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.158453,15648.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.0528,31303.5
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,68.2187,770960
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.189029,15786
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.145011,15780.3
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,119.598,1.39175e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.195483,15934.7
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.147113,15936.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,120.545,1.41227e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.198587,16099.3
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.149212,16085.9
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.34542,18520.9
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,46.2039,541168
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.170264,16085.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.142588,16060.7
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,45.0235,533320
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.16268,16011
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.136076,16011
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,44.8935,533198
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.159618,15988.5
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.138479,15988.5
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.328334,15980.8
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,0.415604,17399.6
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.0753406,15982.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.0624704,15982.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.185115,15950.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.054502,15934.9
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.49926,31108.7
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp32_perf_fp32_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp32_perf_fp32_120.txt
deleted file mode 100644
index ef02bae604..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp32_perf_fp32_120.txt
+++ /dev/null
@@ -1,150 +0,0 @@
-Conv1,154.121,1.5352e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.214927,13584.1
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.136566,13578.3
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1325.28,1.13532e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.236611,11594.5
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.220182,11607.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,8.68528,74729.2
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,331.497,2.71722e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.216182,12478.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.124275,12482.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,432.035,4.00485e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.186777,13886
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.115949,13886.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,5.62394,57859.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,150.64,1.48113e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.226653,14698.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.171148,14679.5
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,288.325,3.0245e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.205001,15516.9
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.121619,15505.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,264.666,2.98708e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.214553,16309
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.164982,16309
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,2.61044,35073.9
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,112.96,1.29306e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.212323,16783.5
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.155334,16775.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,212.916,2.61474e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.217145,17265.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.15167,17260.3
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,198.796,2.50427e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.21831,17883.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.158639,17875.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,1.99705,35751.2
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,93.2211,1.15799e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.20279,17562.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.144169,17537.1
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,87.9998,1.12374e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.184723,17274.3
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.155062,17272.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,90.4476,1.13632e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.180918,17000.1
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.142857,17000.1
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,0.403823,16998.2
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,1.2391,19436.9
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,0.131792,16984.9
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.129337,16976.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,0.301634,16992.5
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.113517,16961.9
-Add15_f2h,0,0
-Add15_h2f,0,0
-Softmax1,1.2847,19459.8
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_120.txt
deleted file mode 100644
index 589d3a4ac0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_120.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,455.534,4.66689e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.467397,28685.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.353119,28666.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,3585.02,3.23302e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.29772,26928.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.564293,25742
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,18.7601,185486
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,826.856,8.46898e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.449964,29892.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.359794,29870
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,1786.12,1.82074e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.468817,29916.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.352812,29890.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,9.09037,107589
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,478.69,5.23312e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.41733,32107.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.244485,32102.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,831.879,9.90367e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.435262,33751.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.356005,33721.4
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,840.699,1.02418e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.452997,34772.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.353574,34718.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,6.35883,93697.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,323.333,4.00937e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.41452,35776.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.295627,35705
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,578.318,7.63711e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.437893,36712.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.327973,36701.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,584.029,7.80435e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.435525,37234.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.338002,37234.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,4.18449,72631.2
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,157.441,2.07545e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.419966,37340.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.301151,37340.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,153.651,2.09778e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.388005,37251.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.296562,37240.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,153.779,2.0978e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.392287,37089.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.295282,37094.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.64765,38909
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.2883,224476
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.52821,62175
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.284005,36871.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.55282,59052
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.598219,40063.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.268076,36932.8
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.37863,41627.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.353387,38523.4
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,35.025,504571
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_151.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_151.txt
deleted file mode 100644
index 2971ceeea6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_151.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,422.717,4.83781e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.456913,31601.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.359327,31608.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1351.01,1.46175e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.470782,30563.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.368472,30559.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,11.228,130152
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,626.09,6.72717e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.451845,31359.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.295218,31306.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,648.454,7.58183e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.436287,33020.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.363244,33016.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.28285,102264
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,311.013,3.59628e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.434258,33894.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.335039,33833.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,428.34,5.33657e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.439071,34778.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.370583,34793.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,423.839,5.3622e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.43703,35548.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.364132,35537.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.63221,78037.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,212.129,2.72828e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.42188,36041
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.347705,36018.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,332.695,4.41954e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.431877,36682.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.34686,36667.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,332.14,4.46479e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.430725,37207.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.346392,37157.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.78426,74326.6
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,97.827,1.3133e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.388683,37420.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.306886,37344.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,92.2453,1.28561e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.354264,37397.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.297324,37378.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,92.9713,1.25568e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.34988,37401.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.289862,37405.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.47196,37393.8
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,16.902,217429
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.44447,61995.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.202687,37222.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.34326,58863.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.417669,40296.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.129376,37226.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,0.987714,41852.8
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.258476,38778.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.5777,499794
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_152.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_152.txt
deleted file mode 100644
index 06209b3983..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_152.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,419.903,4.80075e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.450078,31672
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.351653,31680.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1354.26,1.46644e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.473668,30639.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.374443,30632
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,10.7878,128710
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,629.827,6.75962e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.454795,31321.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.272812,31314.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,644.683,7.58624e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.443385,32989.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.358603,33001
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,8.00249,107248
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,308.4,3.61338e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.43619,33863.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.367026,33841.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,423.632,5.3215e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.435115,34743.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.367647,34759
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,421.208,5.34379e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.434041,35567
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.373271,35582.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.04297,72888.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,212.04,2.72156e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.417727,36076.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.337746,36061
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,330.815,4.41945e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.430757,36785.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.336632,36759
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,330.747,4.48595e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.42286,37325.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.342738,37287.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.48344,72664.8
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,97.3527,1.30997e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.392498,37432.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.301055,37432.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,93.0202,1.29313e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.354111,37462.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.294315,37436.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,92.5392,1.28842e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.349381,37493.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.306501,37485.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.38795,37474.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.395,220404
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.46958,61965
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.184447,37284.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.32265,58890.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.401323,40369.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.122387,37307.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.00145,43413.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.123065,37311
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,35.3043,513809
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_153.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_153.txt
deleted file mode 100644
index 1643c576ca..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_153.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,397.099,4.49506e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.464709,31356.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.343282,31356.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1336.38,1.42703e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.470538,30228.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.378373,30205.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,10.9964,128980
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,626.936,6.64333e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.448285,30804.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.284934,30789
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,637.781,7.37639e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.444574,32500.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.362661,32508
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,8.10626,105809
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,302.937,3.50525e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.428292,33395.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.346002,33407.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,424.344,5.24194e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.434564,34333
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.35514,34283.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,417.409,5.23588e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.451583,35186
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.369868,35194.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.50473,80917.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,209.276,2.65827e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.414558,35686.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.341151,35683
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,329.01,4.37324e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.423173,36457.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.338411,36431
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,329.79,4.42685e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.436888,36984.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.340741,36969.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.87922,73954.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,97.0445,1.29492e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.391,37108.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.304063,37081.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,92.307,1.27898e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.350232,37150.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.299244,37104.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,92.5061,1.27908e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.339915,37192.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.301516,37135.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.49239,37135.4
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.0233,216437
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.44092,61606
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.191008,36964.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.32535,58478.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.399518,40008.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.130444,36984
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,0.999721,41606.8
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.237637,37084.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.9984,502847
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_154.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_154.txt
deleted file mode 100644
index 8cbd143a8a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_154.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,395.41,4.48572e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.459621,31490.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.364497,31468.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1328.47,1.42515e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.468785,30320.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.380856,30327.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,10.8877,127334
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,610.254,6.4922e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.458974,31088
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.279231,31080.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,634.81,7.41724e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.448696,32754.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.35749,32750.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,8.46483,113000
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,302.409,3.51463e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.433527,33631.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.345092,33616.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,421.615,5.24439e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.431966,34539.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.360196,34555.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,415.836,5.24529e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.437291,35324.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.373394,35309.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.32959,75931.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,209.433,2.67102e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.42028,35870
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.337439,35847.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,328.216,4.36791e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.427327,36583.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.348831,36576.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,328.848,4.39311e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.437989,37115.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.350975,37100
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.72423,74230.8
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,96.8875,1.30091e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.394866,37311.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.309835,37285.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,91.7807,1.28302e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.347807,37292.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.293189,37281.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,92.4581,1.28597e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.346776,37350
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.293771,37323.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.45509,37304.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.3207,217244
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.50188,61790.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.221772,37118.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.40265,58712.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.506123,40192.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.134297,37152.6
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,0.977603,41760
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.241554,38732
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,35.1237,504354
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_155.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_155.txt
deleted file mode 100644
index 918a2cfce4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_155.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,462.746,5.246e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.45827,31208
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.363135,31219.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1795.36,1.85564e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.500798,29407.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.361573,29400.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,12.464,135697
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,657.109,6.98681e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.442558,31339.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.285369,31297.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,814.728,9.52013e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.446494,33129.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.360696,33103.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.6773,104305
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,357.389,4.22886e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.430322,34345.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.25004,34334.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,525.941,6.72597e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.444279,35538.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.373759,35527
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,535.496,6.91661e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.457022,36269
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.364248,36254
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.64234,79736.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,226.052,2.92868e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.425669,36883.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.351659,36868.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,357.102,4.87983e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.429515,37632.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.347782,37610.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,357.738,4.92851e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.435691,38087.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.342629,38060.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.96874,74207.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,102.78,1.33726e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.406462,38231
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.317036,38219.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,99.1507,1.39934e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.367941,38262
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.301356,38246.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,99.6024,1.40954e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.364204,38247.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.297618,38232.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.44887,38244
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,16.5915,219518
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.50539,63628.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.263711,38024
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.47619,60419.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.46332,41238.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.137439,38072.8
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.01981,42813.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.266642,39639.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,35.0942,522733
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_156.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_156.txt
deleted file mode 100644
index fba1d66b52..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_156.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,462.456,5.24759e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.470366,31236.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.386437,31233.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1786.08,1.85547e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.49011,29461.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.370277,29415.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,12.5777,141552
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,663.094,7.01318e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.437688,31327.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.275621,31346.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,823.458,9.69014e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.449541,33274.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.357054,33236.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.39743,99651.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,356.259,4.23242e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.423742,34467.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.252715,34452
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,521.528,6.72026e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.432165,35629.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.353637,35636.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,534.428,6.93684e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.434514,36372
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.349861,36364.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.33523,76385
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,264.163,3.45833e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.422949,37001.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.344703,37016.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,454.981,6.23324e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.432491,37652.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.341759,37618.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,449.836,6.22413e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.429278,38118.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.338193,38042
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.72389,74164
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,103.435,1.40574e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.407595,38286
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.311096,38241.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,98.8123,1.3937e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.364689,38296.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.299948,38281.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,98.9485,1.39474e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.362276,38295.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.305662,38280.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.43221,38258
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,16.6626,224054
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.4331,63837.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.27479,38030
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.47038,60583.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.449291,41256.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.136051,38082.8
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.031,42885
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.248115,39714.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.754,518805
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_157.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_157.txt
deleted file mode 100644
index b282e13d31..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_157.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,468.293,5.32815e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.459716,31350.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.361305,31347
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1806.4,1.8695e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.484337,29411.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.36887,29407.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,12.5641,141800
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,659.204,7.01771e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.439812,31373.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.290232,31335.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,840.833,9.84169e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.452702,33122.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.358027,33099.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,9.36531,122385
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,357.921,4.22258e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.43866,34342.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.256524,34323.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,534.371,6.81454e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.443589,35461.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.37228,35465
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,542.12,6.99337e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.440734,36303.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.366692,36280.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.83277,81524.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,262.187,3.38728e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.423952,36929.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.352555,36906
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,451.982,6.17926e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.450955,37599.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.359506,37557
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,450.047,6.21311e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.429361,38157.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.346597,38058
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.9487,74195.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,157.702,2.17e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.404139,38346.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.313132,38289.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,152.988,2.17418e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.376485,38459.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.30309,38429.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,153.741,2.19347e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.374002,38562.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.297509,38559
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.52252,42397.8
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,20.9558,246174
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.50233,63426.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.266917,38277.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.56151,60267
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.508606,41435
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.152332,38311.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.08275,44550.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.201663,38348.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.6795,520432
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_158.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_158.txt
deleted file mode 100644
index 1d3d729848..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_158.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,440.391,4.93952e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.457003,30781
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.393682,30762
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1768.36,1.80247e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.473995,28941
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.368388,28918.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,12.2497,134357
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,637.755,6.69794e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.437727,30922.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.274207,30906.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,807.055,9.28151e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.452458,32744
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.353163,32717.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.66591,104685
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,348.993,4.06804e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.432177,33951.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.264075,33940.4
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,516.937,6.57625e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.438206,35229.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.354039,35203.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,527.592,6.77733e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.445438,36025.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.363307,36010.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.53962,82711.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,223.042,2.87593e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.434948,36631
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.340484,36638.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,354.42,4.80514e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.433405,37407.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.343896,37408
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,355.301,4.87592e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.431262,37947.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.359659,37924.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.53255,68235.8
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,103.145,1.4059e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.396575,38067
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.320997,38049.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,97.3921,1.38125e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.364581,38127.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.291557,38102
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,98.0876,1.38361e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.37388,38184.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.300728,38139.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.45106,40011.8
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.5523,230987
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.50426,63542
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.26862,37919.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.47382,60366.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.445022,41136.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.128671,37939.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.03307,42710.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.304607,39548.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.6686,514672
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_159.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_159.txt
deleted file mode 100644
index 6008d5f448..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_159.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,442.356,4.97196e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.462584,30844.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.370206,30817.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1768.67,1.8098e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.475421,28921.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.360697,28918.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,14.2141,156890
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,645.945,6.74373e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.444785,30872.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.260088,30830.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,813.457,9.41369e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.445118,32775.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.369771,32787.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,8.4375,109650
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,348.398,4.06891e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.430449,33993.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.264927,33977.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,517.785,6.58243e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.440625,35167.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.365381,35167.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,524.29,6.73461e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.442187,36030
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.361554,36018.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.41749,75571
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,259.883,3.36736e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.422353,36693
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.349579,36677.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,448.786,6.09741e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.426001,37333.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.34275,37318.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,443.92,6.11344e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.438495,37857.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.35484,37851.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.82873,75683.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,101.987,1.38326e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.404247,38031.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.308447,38020.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,98.3819,1.38671e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.363224,38076.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.296389,38061.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,98.0606,1.38627e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.360787,38114.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.309348,38054.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.47454,38027.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,16.686,221396
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.53037,63461
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.271225,37853
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.46261,60279
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.453169,41083.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.129932,37932.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.00204,42720
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.265541,39542.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,35.0496,515967
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_160.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_160.txt
deleted file mode 100644
index cf13d7359d..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_160.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,439.53,4.95415e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.470738,30974.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.357822,30978.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1775.42,1.8209e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.461854,29032
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.355947,29020.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,12.3663,136579
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,649.944,6.83858e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.52767,30841.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.340959,30845.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,816.027,9.38261e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.45667,32764.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.379954,32756.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.26986,94885.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,351.701,4.11514e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.413343,34040.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.251686,34032.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,521.937,6.63458e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.442014,35203
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.36243,35210.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,530.806,6.8045e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.436114,36034
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.377022,35999.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.34536,82602.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,258.532,3.34657e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.429067,36681.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.346635,36655
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,443.115,5.97557e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.438622,37390.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.346532,37402.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,444.201,6.10648e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.432062,37943.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.343135,37870.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.93087,77591
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,156.695,2.14927e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.406808,38181.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.315108,38185.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,152.801,2.15966e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.369144,38314.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.295499,38310.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,152.641,2.16467e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.372076,38386
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.292274,38325.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.4523,38329
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.5014,230710
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.51477,63950.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.305273,38109
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.46625,59128.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.578813,41342.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.152736,38150
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.034,42990.2
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.241906,38183
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.7851,517607
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_161.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_161.txt
deleted file mode 100644
index 233279f082..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_161.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,478.02,5.46943e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.465918,31498.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,1.0835,37990
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,2012.03,2.06938e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.467992,28891.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.371358,28918.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,10.9606,124367
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,658.983,6.96979e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.453086,31441.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.310206,31441.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,897.926,1.04895e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.446583,33380.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.361222,33392
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.61052,105274
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,381.524,4.56242e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.421905,34695.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.272114,34667.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,582.714,7.53659e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.436203,35850.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.391563,35839.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,589.591,7.67717e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.430814,36612.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.354577,36589.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.17412,76754.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,269.766,3.53152e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.431499,37341
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.341713,37329.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,459.852,6.34445e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.43468,38053
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.344503,38045.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,457.161,6.36878e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.428471,38513
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.342924,38482.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,4.06549,76943.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,158.162,2.19765e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.401719,38764.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.308338,38737.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,153.356,2.19658e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.374117,38775
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.308625,38775
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,153.861,2.19994e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.375191,38816.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.307033,38805.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.61365,40707.4
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,16.921,229242
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.52435,64934
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.284293,38565.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.55004,61680.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.469035,41867.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.152223,38580.8
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.1082,43465.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.25902,40228.4
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,35.3303,523513
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_162.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_162.txt
deleted file mode 100644
index 48c3d43c77..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_162.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,475.325,5.4428e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.471466,31519.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.3742,31462.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,2020.22,2.0755e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.508951,28887.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.358884,28857.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,11.2409,124215
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,656.323,7.02688e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.44049,31453.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.299179,31411.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,892.972,1.04325e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.451563,33487.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.363717,33460.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,6.78994,95413.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,381.923,4.57262e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.423358,34762.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.239269,34735.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,577.293,7.4957e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.445502,35934.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.365464,35915.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,593.824,7.74848e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.451172,36608.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.355052,36627.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.24424,78714.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,268.715,3.53072e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.436618,37359.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.336767,37344.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,455.751,6.29107e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.434417,38081.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.349431,38073.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,466.098,6.50048e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.443306,38491.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.344352,38423.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,4.05167,78744.6
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,157.831,2.18693e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.412754,38725.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.305323,38664.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,154.577,2.20612e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.381516,38732.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.30065,38705.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,153.16,2.19708e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.375505,38816.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.30179,38805
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.56629,44533
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.5104,234215
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.4864,64875.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.288939,38557.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.54192,61492.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.520995,41833.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.17054,38584.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.08611,43484.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.243653,38599.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.679,521041
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_163.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_163.txt
deleted file mode 100644
index 28c204df0f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_163.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,479.238,5.47122e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.472881,31532
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.370654,31509.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,2009.93,2.071e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.475536,28956.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.372708,28906.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,11.867,128952
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,655.007,7.00299e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.455499,29912.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.306564,29904.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,891.666,1.04793e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.450814,33510.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.359422,33479.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.20779,98681.2
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,383.212,4.59584e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.426141,34769.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.252728,34731.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,581.782,7.53428e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.445591,35919.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.357694,35938.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,594.196,7.74715e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.440305,36627.2
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.367193,36619.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.3188,78652.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,267.855,3.51516e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.435704,37363.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.352267,37302.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,463.867,6.40406e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.437777,37971.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.349809,37964.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,464.395,6.47535e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.432613,38369
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.346564,38376.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.87553,74836.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,161.566,2.2398e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.412286,38661.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.310443,38623.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,156.559,2.23724e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.383142,38736.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.310756,38721.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,156.895,2.24982e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.389003,38851.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.309067,38789.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.54096,40718.4
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.4789,233751
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.53203,64756
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.274219,38557.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.58088,61451.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.516374,41810.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.200409,38580.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.14305,43454.2
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.266034,40205.4
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.769,521154
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_164.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_164.txt
deleted file mode 100644
index ba32130b28..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_164.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,477.833,5.45557e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.469412,31456.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.36227,31464.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,2015.78,2.06643e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.47699,28845.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.372528,28838.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,11.1428,121236
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,652.718,7.00219e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.456913,31419.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.311314,31381.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,900.165,1.05334e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.445285,33365.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.354392,33323.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,8.12858,110012
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,380.239,4.53905e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.428068,34643.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.24837,34635.8
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,587.874,7.54955e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.441552,35736.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.36629,35732.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,587.373,7.61664e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.442706,36554
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.35916,36535
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.84492,85826.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,267.762,3.51378e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.433412,37242
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.351883,37253.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,459.17,6.29912e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.438225,37946.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.353125,37939.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,459.939,6.39416e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.432228,38460.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.360254,38442.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,4.37304,76782.6
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,159.158,2.20692e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.417258,38618.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.322084,38588.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,158.785,2.26098e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.38058,38774.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.303506,38733
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,156.775,2.24646e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.38451,38782.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.314642,38752.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.54075,38764
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.7869,243173
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.41893,64810.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.266853,38532
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.55642,61472
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.51491,41799.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.189989,38531.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.10684,43410.2
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.262994,40183
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,35.1529,527155
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_165.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_165.txt
deleted file mode 100644
index c59e479d32..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_165.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,455.03,5.1799e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.474071,31222
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.357803,31222
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1993.53,2.02599e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.490026,28544.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.389419,28517.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,12.0146,129931
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,635.387,6.69338e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.438711,31119.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.293996,31131.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,878.929,1.02748e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.453002,33173.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.365246,33177.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.53228,101166
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,372.759,4.41161e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.422795,34486
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.256888,34444.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,575.548,7.3956e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.439851,35651.8
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.364133,35613.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,581.808,7.49618e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.443313,36417.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.357905,36413.8
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.62854,85542.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,264.059,3.44035e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.440401,37144.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.348472,37117.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,450.917,6.16922e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.433373,37960.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.355365,37911.2
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,457.997,6.36621e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.432715,38355.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.349317,38348.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.7753,74725.8
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,157.693,2.18022e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.41155,38563.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.322738,38552.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,152.831,2.1846e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.374059,38631.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.302943,38619.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,153.721,2.20276e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.372421,38657.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.299263,38642.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.55718,42488.2
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.8352,238008
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.50222,64212.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.273887,38394.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.5249,61001.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.458295,41641
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.154111,38444.6
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.06721,43254.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.23811,38467.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.8366,521695
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_166.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_166.txt
deleted file mode 100644
index c64f4f3e52..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_166.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,455.276,5.15856e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.470359,31149.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.386129,31145.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1996.3,2.01844e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.480478,28456.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.364843,28445
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,13.1877,143849
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,638.453,6.70373e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.431281,31082
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.261586,31043.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,880.677,1.02525e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.45194,33154
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.348607,33101
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.77753,104127
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,373.494,4.4186e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.434212,34390.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.239826,34384.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,572.542,7.35019e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.445362,35601.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.361758,35590
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,587.187,7.55765e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.432727,36379.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.345611,36349.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.38383,78143.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,265.481,3.44615e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.43203,37082
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.340134,37047.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,457.768,6.28078e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.433913,37802.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.337374,37742.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,461.356,6.38625e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.448081,38215.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.334123,38165.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,4.02037,76324.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,157.571,2.17779e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.412523,38442.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.308018,38415.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,151.67,2.1545e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.382015,38556.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.297944,38510.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,152.985,2.18104e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.364958,38608.6
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.292722,38593.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.54659,42345.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.0312,230236
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,3.12421,72351.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.266975,38342.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.50593,61281.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.462366,41629.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.160268,38415.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.0812,43265.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.265125,40039.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.389,510390
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_167.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_167.txt
deleted file mode 100644
index 54673f6021..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_167.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,451.842,5.16033e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.452958,31297.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.357886,31285.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1989.47,2.03189e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.480286,28587
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.363979,28587
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,12.3533,134286
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,637.258,6.74518e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.456464,31197.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.29685,31158.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,868.231,1.01561e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.513251,33295.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.366276,33295.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,6.86558,94687.4
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,373.103,4.43476e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.435043,34535
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.254092,34509.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,568.595,7.33462e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.442815,35774.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.367025,35762.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,581.724,7.54411e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.44195,36516.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.376818,36459.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.18811,76563.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,263.492,3.43995e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.437937,37228.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.350263,37225.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,456.802,6.28936e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.445771,37918.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.348363,37907.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,455.376,6.28292e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.437924,38408.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.347474,38374.6
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.87436,78635.8
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,160.082,2.21344e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.412855,38662.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.320402,38616.6
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,153.655,2.19322e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.37941,38745.6
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.300702,38715.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,157.898,2.25175e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.397201,38870.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.304267,38817
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.56307,42609.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.2303,233792
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.48179,65167.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.267499,38554.4
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.50431,60194.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.645949,43541.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.222649,38638.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.18469,43603.2
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.268466,40305.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.5977,523272
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_168.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_168.txt
deleted file mode 100644
index 89db826f20..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_168.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,455.361,5.15884e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.462295,31145.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.348824,31137.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1995.56,2.02344e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.491447,28479.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.384536,28460.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,12.3928,132727
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,636.833,6.72095e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.445655,31089.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.272409,31051.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,879.67,1.02163e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.445156,33127.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.359966,33135.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.38661,102607
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,372.157,4.40128e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.429316,34394.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.24878,34356.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,572.823,7.33369e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.44659,35587
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.366212,35563.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,587.311,7.54241e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.439447,36326.6
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.355396,36288.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.36879,76206.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,263.153,3.41952e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.426865,37082.6
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.345381,37032.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,455.644,6.22858e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.428452,37785.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.345034,37743.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,461.078,6.37041e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.441771,38219
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.34899,38181.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,4.02953,76389.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,159.459,2.18848e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.419249,38479.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.302764,38465
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,153.928,2.19033e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.390789,38585.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.295538,38528
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,155.492,2.2155e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.381342,38626
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.306968,38599.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.53701,40509.2
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,16.9248,227446
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.4997,64466
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.274335,38386.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.54606,61207.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.516683,41628.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.227859,38394.6
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.15581,43260.8
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.258853,40068.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.6115,519004
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_261.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_261.txt
deleted file mode 100644
index e4548f7e02..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_261.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,158.533,1.70071e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.608726,29506.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.700086,29513.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1612.88,1.72645e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.440343,30937.8
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.360055,30937.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,48.8482,521160
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,617.178,6.62728e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.392593,31410
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.261708,31364
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,838.687,9.64846e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.405412,33103.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.32588,33092.4
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.15102,94244
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,312.613,3.66479e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.400798,33924.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.366008,33901.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,561.898,6.91021e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.40202,34422.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.352472,34388.2
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,566.636,7.01164e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.41356,34754.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.360101,34720
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.63076,78106
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,222.176,2.74714e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.404613,35239.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.338552,35227.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,370.64,4.79864e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.399108,35681.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.356363,35654.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,375.592,4.86329e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.403716,35914.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.348452,35899.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.77606,68143
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,107.212,1.38908e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.375371,35974.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.30003,35928.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,104.675,1.38702e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.347679,35823.2
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.290034,35804
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,103.943,1.3766e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.344331,35724.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.308018,35697.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.36247,35674.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,16.6724,208084
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.51761,59723.4
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.251244,35503
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.44899,56794.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.496343,38581
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.122719,35568.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,0.966106,40148.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.239468,35621.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.9406,481371
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_262.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_262.txt
deleted file mode 100644
index 10e7503edf..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_262.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,161.775,1.71091e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.646858,29303.2
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.70038,29269
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,1616.13,1.72554e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.444843,30853
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.378718,30849.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,43.8828,470734
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,622.413,6.66894e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.4006,31322.2
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.240914,31291.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,840.715,9.67157e+06
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.409157,32995.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.324543,32965.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.28431,95528.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,313.057,3.66414e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.407569,32159.2
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.347211,32132.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,565.199,6.98615e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.412817,34278
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.376728,34278
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,571.402,7.08326e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.415205,34648
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.359423,34609.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.42833,77877.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,223.315,2.77603e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.402314,35079.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.338814,35048.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,374.653,4.82776e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.403198,35528.6
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.34396,35471.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,376.729,4.8737e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.410859,35773.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.336914,35770.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.95248,73240.2
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,108.964,1.40477e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.385227,35845.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.302072,35815.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,104.682,1.38706e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.354814,35643.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.29758,35620.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,105.504,1.38483e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.347557,35540.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.292427,35513.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.37679,37298.4
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.9789,220505
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.42912,59088.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.255692,35347
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.37765,56103
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.405227,38306.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.124992,35373.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.00088,39794.4
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.245631,36880.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.8187,480667
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_263.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_263.txt
deleted file mode 100644
index b7b8bf2f63..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_263.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,191.244,2.05055e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.427005,29448.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.405195,29437
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,2487.73,2.6007e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.453194,31370.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.363793,31294.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,18.1172,201366
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,828.895,9.10771e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.399441,32356.4
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.239199,32322.6
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,1237,1.45583e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.426282,34091.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.359966,34068.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.02861,98706.8
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,417.602,5.03829e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.384861,35003.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.244683,35000
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,771.959,9.86912e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.41635,35389.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.364319,35370
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,793.706,1.00655e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.41605,35386.8
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.35642,35379
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.2233,76126
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,288.912,3.68753e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.425668,35922.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.336254,35900.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,509.037,6.70579e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.416388,36378.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.337548,36369
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,511.52,6.73398e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.42227,36633.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.346916,36595
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.88723,73228
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,142.071,1.87156e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.390155,36636.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.299205,36609.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,141.007,1.89372e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.373725,36482.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.978996,43770.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,140.041,1.87996e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.363351,36285.4
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.313297,36258.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.47037,41611.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.1496,218005
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.76381,64189.2
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.273976,36041.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.61064,57895.6
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.561251,39172
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.161829,36113
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.09352,40776.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.272631,37718.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.759,490105
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_264.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_264.txt
deleted file mode 100644
index eba4402010..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_264.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,185.939,2.01791e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.449643,30210.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.343512,30229.4
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,2501.14,2.62453e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.451575,31454.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.368913,31385.8
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,17.6366,203769
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,840.18,9.23985e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.392555,32314
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.230943,32298.8
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,1241.48,1.46389e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.422672,34077.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.344183,34043
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,6.79683,93570.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,419.176,5.00392e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.386577,35025.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.247319,34995
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,774.162,9.89205e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.427769,35412
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.368358,35370
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,797.068,1.00928e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.419845,35440
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.354776,35386.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,4.93236,70696.2
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,288.804,3.67853e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.409738,35955.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.348677,35956.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,508.633,6.69683e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.408344,36369.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.33811,36342.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,512.264,6.77793e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.409164,36564.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.348613,36515
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.76312,73057.8
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,142.653,1.88521e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.388529,36604.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.305624,36597
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,141.645,1.91591e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.364497,36482.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.297535,36470.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,140.293,1.88476e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.363978,36259.2
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.309911,36259.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.44807,39896.4
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.5247,220418
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.52067,60871
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.276351,36098.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.46145,56290
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.552175,39172.2
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.147897,36112.2
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.08342,40749.2
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.267576,37671.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.5875,488365
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_265.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_265.txt
deleted file mode 100644
index db65e012b6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_265.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,181.593,1.97488e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.435338,30282.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.362987,30290
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,2487.56,2.62436e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.451518,31514.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.359396,31522.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,16.0458,183438
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,839.503,9.23851e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.406814,32390.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.239935,32379.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,1231.36,1.45609e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.41331,34183.2
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.342853,34141
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,6.27933,90448.6
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,418.496,5.08746e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.379108,35090.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.259602,35079.2
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,771.965,9.87368e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.412676,35435.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.359735,35438.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,792.082,1.00742e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.418122,35520
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.345508,35493.4
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,4.98809,76145
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,288.537,3.68433e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.412907,36023
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.334289,35993.8
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,507.543,6.69858e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.412267,34598.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.356229,34590.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,511.552,6.75649e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.404273,36646.4
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.328203,36619.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.74345,73209.2
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,141.885,1.87443e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.382647,36661
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.30273,36634.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,141.71,1.88134e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.370231,36525.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.291647,36506.8
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,139.32,1.87431e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.369893,36344.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.297732,36344.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.35279,36314.4
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,18.1132,230015
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.70198,62321.6
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.268184,36131.6
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.56545,57748.8
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.531895,39215.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.159673,36175.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.09838,40793.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.305682,37742.8
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.577,493091
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_266.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_266.txt
deleted file mode 100644
index 2e8956e67f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_266.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,173.636,1.88423e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.382109,29883.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.23427,29876
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,2203.56,2.29131e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.443966,30586.2
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.357886,30563.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,11.8112,129441
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,808.594,8.60936e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.406653,31346.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.229196,31293.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,1103.71,1.26443e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.39701,33358.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.263282,33339.2
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.49989,104971
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,385.211,4.5293e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.371255,34566.4
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.239333,34562.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,720.44,9.09088e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.433539,35193.2
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.353361,35208.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,733.435,9.28106e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.409885,35524.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.363396,35448
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.01246,74445.8
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,279.884,3.55576e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.407653,36195.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.350302,36172.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,492.124,6.51622e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.410564,36792
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.338763,36730.8
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,489.535,6.56928e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.406942,37234.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.340312,37223.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.98477,72556.2
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,140.565,1.86739e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.393834,37238.6
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.301565,37207.8
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,134.25,1.83479e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.373726,37082
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.298366,37074.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,137.186,1.85803e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.368363,36891.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.304748,36876.6
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.31005,36876.6
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.7354,228173
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.44129,61585.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.279455,36686
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.60228,58413
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.599172,39775.4
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.226284,36739.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.20536,41365.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.290533,38291.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.3464,493073
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_267.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_267.txt
deleted file mode 100644
index 5f30e77188..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_267.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,173.43,1.88699e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.432465,29975.4
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.265816,29906.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,2189.87,2.27982e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.447863,30708.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.377387,30700.6
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,11.7131,130270
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,811.672,8.66601e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.409406,31388.6
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.24188,31392.2
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,1087.52,1.25855e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.404529,33567.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.265292,33563.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.50409,105686
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,385.257,4.55787e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.402116,34794.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.244031,34768
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,716.365,9.13693e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.411127,35388
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.363588,35380.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,728.465,9.28599e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.424139,35713
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.358654,35705
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.58631,81955.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,279.577,3.55842e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.406104,34465.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.345726,34461.6
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,490.52,6.53554e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.411645,36880.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.342597,36876.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,489.563,6.56559e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.407865,37260.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.339583,37272.2
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.94554,70759.4
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,139.657,1.85316e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.406366,37318.4
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.305451,37299.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,134.329,1.84175e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.377727,37139.8
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.305733,37128.2
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,136.158,1.85446e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.384023,37040.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.294015,37029.4
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.35178,37033.2
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,16.9017,220361
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.51832,62185.8
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.266738,36835.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.56782,59012.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.579926,39988.8
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.24677,36861.6
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.23178,41586.6
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.293713,38444.2
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.6911,500087
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_268.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_268.txt
deleted file mode 100644
index a1c035492a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_268.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,172.568,1.85223e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.489962,29910.6
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.351377,29906.8
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,2199.18,2.29044e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.448868,30589.6
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.352606,30593.2
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,11.7867,131114
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,815.728,8.70656e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.397655,31312.8
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.221695,31286
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,1090.71,1.25589e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.40275,33476.8
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.249778,33457.8
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.08774,100305
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,385.327,4.53715e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.375781,32933.8
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.228966,32937.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,720.12,9.15787e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.41354,35222.4
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.365994,35242
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,727.57,9.21097e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.430333,35650.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.345554,35628.2
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.28686,78318.6
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,281.544,3.586e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.403525,36241.2
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.328901,36238.4
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,489.049,6.51004e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.41068,36818.8
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.341932,36826.4
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,491.617,6.59014e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.413425,37181.6
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.345412,37155
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,3.67613,74222.6
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,138.692,1.8372e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.39196,37215.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.299634,37189.2
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,134.834,1.84099e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.367915,37044.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.289253,37002
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,135.723,1.85095e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.384183,36883.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.292472,36856.8
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.37167,36868.2
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,16.7821,218357
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.44907,61929
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.268958,36693.2
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.58351,58774.4
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.565629,39812.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.213682,36708.4
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.18764,42986
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.152147,36769.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.6579,496474
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_269.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_269.txt
deleted file mode 100644
index ed7cda546e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_269.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,177.146,1.92423e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.389764,29834
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.235153,29830.2
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,2208.16,2.28783e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,0.449534,30547.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.395057,30539.4
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,13.5003,147919
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,807.798,8.61431e+06
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.400164,31351
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.23141,31332
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,1093.81,1.25446e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.415601,33495.4
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.306174,33491.6
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,7.76251,107105
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,384.777,4.55696e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.38947,34718.6
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.231743,34680.6
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,717.842,9.11881e+06
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.41612,35364
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.359,35337.6
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,732.336,9.28302e+06
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.417719,35656
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.352651,35602.6
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,5.13727,74716.4
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,279.723,3.56724e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.404785,34507.8
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.327915,34500.2
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,493.197,6.53166e+06
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.404516,36769.4
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.347365,36761.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,491.166,6.59488e+06
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.419332,37215.2
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.34819,37192.4
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,4.09067,72479
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,140.255,1.87033e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.391978,37234.2
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.307012,37211.4
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,135.678,1.82971e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.373272,37059
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.294117,37028.6
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,137.144,1.86828e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.36892,36914.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.294406,36903.2
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,2.33905,36899.4
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,17.0487,221581
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,2.49858,62079
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.26908,36728
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,2.60495,58928.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.591433,39885.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.24277,36751
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.24619,41498.8
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.295,38375.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,34.9505,503463
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp32_perf_fp32_120.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp32_perf_fp32_120.txt
deleted file mode 100644
index e5708b3dbd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp32_perf_fp32_120.txt
+++ /dev/null
@@ -1,159 +0,0 @@
-Conv1,669.347,5.79186e+06
-Conv1_f2h,0,0
-Conv1_h2f,0,0
-Add1,0.44224,25607.8
-Add1_f2h,0,0
-Add1_h2f,0,0
-Relu1,0.340281,25592.6
-Relu1_f2h,0,0
-Relu1_h2f,0,0
-Conv2,7139.74,5.46598e+07
-Conv2_f2h,0,0
-Conv2_h2f,0,0
-Add2,1.47909,22940.4
-Add2_f2h,0,0
-Add2_h2f,0,0
-Relu2,0.578853,20923
-Relu2_f2h,0,0
-Relu2_h2f,0,0
-Pool1,60.6459,467598
-Pool1_f2h,0,0
-Pool1_h2f,0,0
-Conv3,1541.99,1.25939e+07
-Conv3_f2h,0,0
-Conv3_h2f,0,0
-Add3,0.449702,25722
-Add3_f2h,0,0
-Add3_h2f,0,0
-Relu3,0.310771,25726.4
-Relu3_f2h,0,0
-Relu3_h2f,0,0
-Conv4,2598.25,2.36776e+07
-Conv4_f2h,0,0
-Conv4_h2f,0,0
-Add4,0.799391,29436.6
-Add4_f2h,0,0
-Add4_h2f,0,0
-Relu4,0.599462,29567
-Relu4_f2h,0,0
-Relu4_h2f,0,0
-Pool2,21.6318,230524
-Pool2_f2h,0,0
-Pool2_h2f,0,0
-Conv5,744.163,7.87985e+06
-Conv5_f2h,0,0
-Conv5_h2f,0,0
-Add5,0.431494,32244
-Add5_f2h,0,0
-Add5_h2f,0,0
-Relu5,0.286745,32244
-Relu5_f2h,0,0
-Relu5_h2f,0,0
-Conv6,1576.37,1.74915e+07
-Conv6_f2h,0,0
-Conv6_h2f,0,0
-Add6,0.467628,33194.6
-Add6_f2h,0,0
-Add6_h2f,0,0
-Relu6,0.387506,33190.8
-Relu6_f2h,0,0
-Relu6_h2f,0,0
-Conv7,1744.67,1.92649e+07
-Conv7_f2h,0,0
-Conv7_h2f,0,0
-Add7,0.43932,32155.4
-Add7_f2h,0,0
-Add7_h2f,0,0
-Relu7,0.295117,32174
-Relu7_f2h,0,0
-Relu7_h2f,0,0
-Pool3,11.0778,134872
-Pool3_f2h,0,0
-Pool3_h2f,0,0
-Conv8,493.441,5.963e+06
-Conv8_f2h,0,0
-Conv8_h2f,0,0
-Add8,0.433138,35087.4
-Add8_f2h,0,0
-Add8_h2f,0,0
-Relu8,0.380959,35076
-Relu8_f2h,0,0
-Relu8_h2f,0,0
-Conv9,910.678,1.20378e+07
-Conv9_f2h,0,0
-Conv9_h2f,0,0
-Add9,0.450066,38405.2
-Add9_f2h,0,0
-Add9_h2f,0,0
-Relu9,0.407878,38378.6
-Relu9_f2h,0,0
-Relu9_h2f,0,0
-Conv10,897.248,1.2633e+07
-Conv10_f2h,0,0
-Conv10_h2f,0,0
-Add10,0.449388,40364.8
-Add10_f2h,0,0
-Add10_h2f,0,0
-Relu10,0.386828,40322.8
-Relu10_f2h,0,0
-Relu10_h2f,0,0
-Pool4,6.03651,96459
-Pool4_f2h,0,0
-Pool4_h2f,0,0
-Conv11,246.238,3.40541e+06
-Conv11_f2h,0,0
-Conv11_h2f,0,0
-Add11,0.428479,40601.8
-Add11_f2h,0,0
-Add11_h2f,0,0
-Relu11,0.351365,40546
-Relu11_f2h,0,0
-Relu11_h2f,0,0
-Conv12,238.715,3.44638e+06
-Conv12_f2h,0,0
-Conv12_h2f,0,0
-Add12,0.414546,40587.4
-Add12_f2h,0,0
-Add12_h2f,0,0
-Relu12,0.273535,40563.4
-Relu12_f2h,0,0
-Relu12_h2f,0,0
-Conv13,236.791,3.49246e+06
-Conv13_f2h,0,0
-Conv13_h2f,0,0
-Add13,0.413625,40604.8
-Add13_f2h,0,0
-Add13_h2f,0,0
-Relu13,0.31061,40577
-Relu13_f2h,0,0
-Relu13_h2f,0,0
-Pool5,3.85133,81098.8
-Pool5_f2h,0,0
-Pool5_h2f,0,0
-Mul1,30.3164,365908
-Mul1_f2h,0,0
-Mul1_h2f,0,0
-Add14,3.40558,74419
-Add14_f2h,0,0
-Add14_h2f,0,0
-Relu14,0.301913,40155.8
-Relu14_f2h,0,0
-Relu14_h2f,0,0
-Mul2,4.51905,81648.2
-Mul2_f2h,0,0
-Mul2_h2f,0,0
-Add15,0.766091,44549.6
-Add15_f2h,0,0
-Add15_h2f,0,0
-Relu15,0.310483,40165.6
-Relu15_f2h,0,0
-Relu15_h2f,0,0
-Mul3,1.77246,50359.2
-Mul3_f2h,0,0
-Mul3_h2f,0,0
-Add16,0.339148,40173.6
-Add16_f2h,0,0
-Add16_h2f,0,0
-Softmax1,65.736,976903
-Softmax1_f2h,0,0
-Softmax1_h2f,0,0
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet/alexnet_valid_soc.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet/alexnet_valid_soc.txt
deleted file mode 100644
index 1b7aeb981c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet/alexnet_valid_soc.txt
+++ /dev/null
@@ -1,231 +0,0 @@
-2725.121326
-+++++
-conf1 1 1 78.78 0.0
-1 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-2 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 tanh fp32 1
-4 gpu conv fp32 11 add fp32 1 tanh fp32 1
-5 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-6 gpu mul fp32 11 add fp32 1
-7 gpu softmax fp32 1
------
-+++++
-conf2 2.1233638648528457 1.6150951710244676 78.3544 0.42560000000000286
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 tanh fp16 12
-4 gpu conv fp16 12 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf3 2.051295134864554 1.6122580072322763 78.3278 0.4522000000000048
-1 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12
-4 gpu conv fp16 12 add fp16 12 tanh fp16 12
-5 gpu conv samp_fp16 269 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf4 2.188609573694276 1.688911612634961 78.30120000000001 0.47879999999999256
-1 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 268 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 tanh fp16 12
-4 gpu conv fp16 12 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf5 2.0570505767108007 1.6000014977491621 78.2214 0.5585999999999984
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 265 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12
-4 gpu conv fp16 12 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf6 2.009166522889861 1.5755494376470724 78.1948 0.5852000000000004
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 269 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12
-4 gpu conv fp16 12 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf7 2.0188668300066377 1.5976556515195433 78.06179999999999 0.7182000000000102
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 268 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 266 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf8 2.1797184471932716 1.6767378001241562 78.06179999999999 0.7182000000000102
-1 gpu conv samp_fp16 263 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 263 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 tanh fp16 12
-4 gpu conv fp16 12 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf9 2.064914192886025 1.6203964986881603 78.06179999999999 0.7182000000000102
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 263 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 269 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf10 2.2070171560926672 1.7194657877315815 78.0352 0.7447999999999979
-1 gpu conv samp_fp16 263 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 265 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12
-4 gpu conv fp16 12 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf11 2.0161469236407057 1.5964768988685245 78.0086 0.7713999999999999
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 269 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 269 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf12 2.157846755426679 1.6765250202752133 78.0086 0.7713999999999999
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 269 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf13 2.0319664118931096 1.6183541826275754 77.98200000000001 0.7979999999999876
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 269 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12
-4 gpu conv fp16 12 add fp16 12 tanh fp16 12
-5 gpu conv samp_fp16 269 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf14 2.354997704376988 1.7779732164691666 77.98200000000001 0.7979999999999876
-1 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 tanh fp16 12
-4 gpu conv fp16 12 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf15 2.3463673263694 1.8510470086526165 77.98200000000001 0.7979999999999876
-1 gpu conv samp_fp16 264 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 263 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf16 2.284714727579521 1.7855758235498087 77.7692 1.0108000000000033
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-5 gpu conv samp_fp16 269 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf17 2.3463673263694 1.8510470086526165 77.68939999999999 1.0906000000000091
-1 gpu conv samp_fp16 264 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 263 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf18 2.427840309027486 1.9007943438562696 77.68939999999999 1.0906000000000091
-1 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 263 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf19 2.4671009475732766 1.9246545843862224 77.47659999999999 1.3034000000000106
-1 gpu conv samp_fp16 264 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf20 2.5567127702266332 1.9773019485322874 77.2638 1.5161999999999978
-1 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf21 2.557898283218207 1.9895818051250724 77.2372 1.5427999999999997
-1 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12
-5 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf22 2.557898283218207 1.9895818051250724 77.21060000000001 1.5693999999999875
-1 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12
-5 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
-+++++
-conf23 2.6457265307759883 2.029290916760937 77.1574 1.6226000000000056
-1 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-6 gpu mul fp16 12 add fp16 12
-7 gpu softmax fp16 12
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet2/alexnet2_valid_soc.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet2/alexnet2_valid_soc.txt
deleted file mode 100644
index a888b5ee5a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet2/alexnet2_valid_soc.txt
+++ /dev/null
@@ -1,188 +0,0 @@
-1129.3450630000002
-+++++
-conf1 1 1 84.76 0.0
-1 gpu conv fp32 11 add fp32 1 tanh fp32 1
-2 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 tanh fp32 1
-4 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-5 gpu conv fp32 11 add fp32 1 tanh fp32 1
-6 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-7 gpu mul fp32 11 add fp32 1
-8 gpu softmax fp32 1
------
-+++++
-conf2 2.2258170210610477 1.3875307929727092 84.74 0.020000000000010232
-1 gpu conv fp16 11 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 151 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf3 2.3673182996864846 1.4566777038051897 84.49999999999999 0.2600000000000193
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 153 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf4 2.24614762418964 1.41800542976017 84.25999999999999 0.5000000000000142
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 158 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 268 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf5 2.304084258604824 1.4284953488024343 84.228 0.5320000000000107
-1 gpu conv fp16 11 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 151 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 267 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf6 2.3377766277342653 1.4440340860007412 84.228 0.5320000000000107
-1 gpu conv fp16 11 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 153 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-6 gpu conv fp16 12 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf7 2.24614762418964 1.41800542976017 84.17479999999999 0.5852000000000146
-1 gpu conv fp16 11 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 158 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 268 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf8 2.3673182996864846 1.4566777038051897 84.095 0.6650000000000063
-1 gpu conv fp16 11 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 153 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf9 2.2463714607055545 1.417884448648111 83.8024 0.9575999999999993
-1 gpu conv fp16 11 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 158 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 266 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf10 2.389025803395913 1.4732901147183992 83.77579999999999 0.9842000000000155
-1 gpu conv fp16 11 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 153 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 268 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf11 2.288831273542033 1.435952475412438 83.61619999999999 1.143800000000013
-1 gpu conv fp16 11 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 158 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf12 2.288831273542033 1.435952475412438 83.58959999999999 1.170400000000015
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 158 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf13 2.389025803395913 1.4732901147183992 83.58959999999999 1.170400000000015
-1 gpu conv fp16 11 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 153 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 268 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf14 2.3892790238475423 1.4731595166090572 83.4566 1.3034000000000106
-1 gpu conv fp16 11 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 153 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 266 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf15 2.390450803781405 1.4707319718833016 83.3768 1.3832000000000022
-1 gpu conv fp16 11 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 153 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 266 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 157 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf16 2.4373708430335537 1.49267343110314 83.3768 1.3832000000000022
-1 gpu conv fp16 11 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 153 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
-+++++
-conf17 2.4373708430335537 1.49267343110314 83.2704 1.48960000000001
-1 gpu conv fp16 12 add fp16 12 tanh fp16 12
-2 gpu conv perf_fp16 153 add fp16 12 tanh fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 tanh fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 261 add fp16 12 tanh fp16 12
-6 gpu conv perf_fp16 160 add fp16 12 tanh fp16 12 pool_max fp16 12
-7 gpu mul fp16 12 add fp16 12
-8 gpu softmax fp16 12
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/resnet18/resnet18_valid_soc.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/resnet18/resnet18_valid_soc.txt
deleted file mode 100644
index 942789c1c4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/resnet18/resnet18_valid_soc.txt
+++ /dev/null
@@ -1,1576 +0,0 @@
-2593.3013975999997
-+++++
-conf1 1 1 89.42 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1
-3 gpu conv fp32 11 add fp32 1
-4 gpu add fp32 11
-5 gpu relu fp32 11
-6 gpu conv fp32 11 add fp32 1 relu fp32 1
-7 gpu conv fp32 11 add fp32 1
-8 gpu add fp32 11
-9 gpu relu fp32 11
-10 gpu conv fp32 11 add fp32 1 relu fp32 1
-11 gpu conv fp32 11 add fp32 1
-12 gpu add fp32 11
-13 gpu relu fp32 11
-14 gpu conv fp32 11 add fp32 1 relu fp32 1
-15 gpu conv fp32 11 add fp32 1
-16 gpu conv fp32 11 add fp32 1
-17 gpu add fp32 11
-18 gpu relu fp32 11
-19 gpu conv fp32 11 add fp32 1 relu fp32 1
-20 gpu conv fp32 11 add fp32 1
-21 gpu add fp32 11
-22 gpu relu fp32 11
-23 gpu conv fp32 11 add fp32 1 relu fp32 1
-24 gpu conv fp32 11 add fp32 1
-25 gpu add fp32 11
-26 gpu relu fp32 11
-27 gpu conv fp32 11 add fp32 1 relu fp32 1
-28 gpu conv fp32 11 add fp32 1
-29 gpu conv fp32 11 add fp32 1
-30 gpu add fp32 11
-31 gpu relu fp32 11
-32 gpu conv fp32 11 add fp32 1 relu fp32 1
-33 gpu conv fp32 11 add fp32 1
-34 gpu add fp32 11
-35 gpu relu fp32 11
-36 gpu conv fp32 11 add fp32 1 relu fp32 1
-37 gpu conv fp32 11 add fp32 1
-38 gpu add fp32 11
-39 gpu relu fp32 11
-40 gpu pool_mean fp32 11
-41 gpu mul fp32 11 add fp32 1
-42 gpu softmax fp32 1
------
-+++++
-conf2 1.8227860146926984 1.3592380545823108 88.28 1.1400000000000006
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 162 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 159 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 166 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv fp16 12 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf3 1.772745264351603 1.3340968704252147 88.2 1.2199999999999989
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 159 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 166 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf4 1.831301934833889 1.3636544094268177 88.2 1.2199999999999989
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf5 1.7541385118416233 1.323200331238725 88.12 1.2999999999999972
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 166 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf6 1.750881760437994 1.3214899710791683 88.12 1.2999999999999972
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 159 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 166 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 268 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf7 1.9207420870636576 1.4105446231099241 88.1 1.3200000000000074
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 159 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 159 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 159 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 160 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 151 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 165 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 268 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv fp16 11 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-37 gpu conv fp16 12 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf8 1.897654446584276 1.3943617562849198 88.1 1.3200000000000074
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 263 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 153 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 151 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 165 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv perf_fp16 154 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-37 gpu conv samp_fp16 262 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf9 1.9276001243246026 1.4155139358802007 88.08 1.3400000000000034
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 168 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 159 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 159 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 160 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 151 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 165 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 268 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv fp16 11 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 155 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf10 1.8877611861107602 1.3945090937373315 88.03999999999999 1.3800000000000097
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 154 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 159 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 166 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf11 1.884015904997108 1.386748889441216 87.96000000000001 1.4599999999999937
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 263 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 153 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 151 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 165 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 268 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv perf_fp16 154 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-37 gpu conv samp_fp16 262 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf12 1.815742308450095 1.3541765419789824 87.83999999999999 1.5800000000000125
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 262 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv fp16 11 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf13 1.928011277898605 1.414528053850526 87.83999999999999 1.5800000000000125
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 159 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 159 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 160 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 151 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 165 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 268 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv fp16 11 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 155 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf14 1.8702574116471649 1.3838796270391824 87.8 1.6200000000000045
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 269 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf15 1.9390257777318618 1.4193909923193697 87.8 1.6200000000000045
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 159 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 159 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 151 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 165 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 268 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv perf_fp16 154 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 155 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf16 1.8505712546542585 1.372601565984325 87.76 1.6599999999999966
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf17 1.931335957581042 1.4149043748735137 87.74 1.6800000000000068
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 157 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf18 1.8390656100510818 1.3668229301466752 87.68 1.7399999999999949
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf19 1.9360126662655235 1.416245073512222 87.64 1.7800000000000011
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 155 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 264 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf20 1.826739398491775 1.3609522133620269 87.62 1.7999999999999972
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 153 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv samp_fp16 262 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 165 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf21 1.8243322012642802 1.3542277148411042 87.62 1.7999999999999972
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 263 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv samp_fp16 266 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf22 1.8245510435946863 1.3601414031759373 87.58 1.8400000000000034
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv samp_fp16 269 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 153 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf23 1.9832010015590205 1.4407797001367388 87.56 1.8599999999999994
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 159 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 159 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 151 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 165 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 261 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv fp16 11 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 155 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf24 1.831958859203629 1.3643626254848584 87.5 1.9200000000000017
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 165 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 151 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf25 1.827209961997738 1.3576190436536635 87.5 1.9200000000000017
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 263 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 159 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 151 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 165 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 268 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv perf_fp16 154 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-37 gpu conv samp_fp16 262 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf26 1.9532893879837718 1.4253186875342474 87.5 1.9200000000000017
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 153 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 168 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 262 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv fp16 11 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf27 1.8598315807624513 1.376813374656673 87.48 1.9399999999999977
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf28 1.8545931630272876 1.3744725755811524 87.48 1.9399999999999977
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 267 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 152 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf29 1.9088935397779812 1.4033062374488858 87.44 1.980000000000004
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 163 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 267 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf30 1.8306014158563824 1.3613821654101905 87.44 1.980000000000004
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 164 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 265 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 168 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 262 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf31 1.9755297077095708 1.4378811225069261 87.44 1.980000000000004
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 159 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 159 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 151 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 268 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv perf_fp16 154 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 155 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf32 1.827200177575606 1.356175543415313 87.38 2.0400000000000063
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 156 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 264 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv perf_fp16 167 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf33 1.8517276001191023 1.3729319418960464 87.38 2.0400000000000063
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-7 gpu conv fp16 12 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv samp_fp16 269 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 157 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 12 relu fp16 12
-24 gpu conv perf_fp16 160 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv samp_fp16 268 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 12 relu fp16 12
-37 gpu conv samp_fp16 269 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf34 1.8938192956663813 1.3919348631813433 87.38 2.0400000000000063
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 263 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 153 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 151 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 165 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 268 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv fp16 11 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-37 gpu conv samp_fp16 262 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
-+++++
-conf35 1.8989539669005067 1.3938360809175603 87.36 2.0600000000000023
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv fp16 12 add fp16 12 relu fp16 12
-3 gpu conv fp16 12 add fp16 12
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 263 add fp16 12
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 12 relu fp16 12
-11 gpu conv perf_fp16 154 add fp16 12
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 12 relu fp16 12
-15 gpu conv fp16 12 add fp16 12
-16 gpu conv fp16 11 add fp16 12
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 153 add fp16 12 relu fp16 12
-20 gpu conv perf_fp16 151 add fp16 12
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 157 add fp16 12 relu fp16 12
-24 gpu conv samp_fp16 268 add fp16 12
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-28 gpu conv fp16 12 add fp16 12
-29 gpu conv perf_fp16 154 add fp16 12
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 12 relu fp16 12
-33 gpu conv fp16 12 add fp16 12
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-37 gpu conv samp_fp16 262 add fp16 12
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 12
-42 gpu softmax fp16 12
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar10/vgg16_cifar10_valid_soc.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar10/vgg16_cifar10_valid_soc.txt
deleted file mode 100644
index 789f4e21cf..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar10/vgg16_cifar10_valid_soc.txt
+++ /dev/null
@@ -1,1027 +0,0 @@
-3994.0731450000017
-+++++
-conf1 1 1 89.22 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 relu fp32 1
-4 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-5 gpu conv fp32 11 add fp32 1 relu fp32 1
-6 gpu conv fp32 11 add fp32 1 relu fp32 1
-7 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-8 gpu conv fp32 11 add fp32 1 relu fp32 1
-9 gpu conv fp32 11 add fp32 1 relu fp32 1
-10 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-11 gpu conv fp32 11 add fp32 1 relu fp32 1
-12 gpu conv fp32 11 add fp32 1 relu fp32 1
-13 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-14 gpu mul fp32 11 add fp32 1 relu fp32 1
-15 gpu mul fp32 11 add fp32 1
-16 gpu softmax fp32 1
------
-+++++
-conf2 2.3049904288987464 1.6887800235455193 89.14 0.0799999999999983
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 155 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 269 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 153 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-9 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-13 gpu conv fp16 11 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf3 2.357615734902983 1.7226289827534114 89.14 0.0799999999999983
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 155 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 269 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 153 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-9 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf4 2.3831343547359976 1.7374446557158316 88.84 0.37999999999999545
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 162 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 269 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 153 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-9 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf5 2.3696393667573616 1.7284732038695636 88.8 0.4200000000000017
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 162 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 269 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 153 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-9 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 265 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf6 2.4444787116056292 1.7833916898567774 88.58 0.6400000000000006
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf7 2.40209759505425 1.7661661942711917 88.58 0.6400000000000006
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf8 2.528892013058046 1.8332619869789675 88.08 1.1400000000000006
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-10 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf9 2.5283008295291105 1.8324605771289624 88.06 1.1599999999999966
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf10 2.5562616043247313 1.847605117430125 88.03999999999999 1.1800000000000068
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf11 2.5337351216813757 1.836759334487813 88.03999999999999 1.1800000000000068
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf12 2.556171297969468 1.8482604143790797 88.03999999999999 1.1800000000000068
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf13 2.5562385363337343 1.8481145682015834 88.03999999999999 1.1800000000000068
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf14 2.556612910921585 1.8486422226408725 88.03999999999999 1.1800000000000068
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf15 2.5419253262471346 1.8395765136023223 88.02 1.2000000000000028
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 263 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf16 2.4937721600323406 1.8116328904640306 88.0 1.2199999999999989
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv perf_fp16 162 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf17 2.5545877208248187 1.8465313171321942 88.0 1.2199999999999989
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv samp_fp16 266 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf18 2.528537397828869 1.8330988121074523 88.0 1.2199999999999989
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf19 2.531670576114998 1.8357132731685366 88.0 1.2199999999999989
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf20 2.5294693760803577 1.8335105878862015 87.98 1.2399999999999949
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 268 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf21 2.5582293136941723 1.8476583031165972 87.98 1.2399999999999949
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 156 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf22 2.556327374925176 1.8481587827658859 87.98 1.2399999999999949
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf23 2.557806470696261 1.8492020211230846 87.98 1.2399999999999949
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf24 2.5545697480449 1.8464092920718178 87.96000000000001 1.259999999999991
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv samp_fp16 267 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf25 2.528206406642683 1.832658178797549 87.96000000000001 1.259999999999991
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf26 2.556533707152568 1.8484262997816934 87.96000000000001 1.259999999999991
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf27 2.5393059900815325 1.837123626585959 87.94 1.2800000000000011
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 265 add fp16 12 relu fp16 12
-12 gpu conv samp_fp16 269 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf28 2.5486219361262235 1.845481069177171 87.94 1.2800000000000011
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf29 2.5485321687357825 1.8461348600374907 87.94 1.2800000000000011
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf30 2.5657339222733015 1.8517901869245543 87.92 1.2999999999999972
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv samp_fp16 263 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf31 2.581139532058275 1.860666047394923 87.92 1.2999999999999972
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf32 2.5098654459068945 1.8297655130336108 87.92 1.2999999999999972
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf33 2.528587182046725 1.8312521826965082 87.9 1.3199999999999932
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 156 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv samp_fp16 266 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf34 2.517311952294846 1.8204468250382393 87.9 1.3199999999999932
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv fp16 11 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf35 2.517311952294846 1.8204468250382393 87.9 1.3199999999999932
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv fp16 11 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf36 2.517311952294846 1.8204468250382393 87.9 1.3199999999999932
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv fp16 11 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf37 2.517311952294846 1.8204468250382393 87.9 1.3199999999999932
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv fp16 11 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf38 2.5346932948358267 1.8376287813464989 87.9 1.3199999999999932
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 265 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf39 2.4914548049246 1.8095620501702707 87.86 1.3599999999999994
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv perf_fp16 162 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv samp_fp16 268 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf40 2.5809312104420865 1.8607657818447936 87.86 1.3599999999999994
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf41 2.5120056276901925 1.824277681148882 87.83999999999999 1.3800000000000097
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 268 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv samp_fp16 266 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf42 2.556168516896762 1.849243225747987 87.83999999999999 1.3800000000000097
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 153 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf43 2.512713457130698 1.8053797549107755 87.82 1.4000000000000057
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 269 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 153 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-9 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf44 2.509447559327321 1.8294109824358684 87.82 1.4000000000000057
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf45 2.532043246184595 1.8347717424454622 87.74 1.480000000000004
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv samp_fp16 265 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf46 2.4911011329750212 1.795311376068545 87.68 1.539999999999992
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 155 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 153 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-9 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv samp_fp16 269 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf47 2.549746515565958 1.8283676275816687 87.66000000000001 1.559999999999988
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv fp16 12 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-9 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf48 2.51145215830771 1.8254971754777813 87.64 1.5799999999999983
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 266 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf49 2.513356522647888 1.826263067419964 87.58 1.6400000000000006
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 269 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf50 2.513356522647888 1.826263067419964 87.53999999999999 1.6800000000000068
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 269 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf51 2.4881677905203494 1.8127135485543127 87.4 1.8199999999999932
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 269 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 269 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf52 2.51145215830771 1.8254971754777813 87.36 1.8599999999999994
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 266 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf53 2.4757784613808234 1.7991027289904775 87.26 1.9599999999999937
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 269 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv fp16 11 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf54 2.5913526715019284 1.8695479088125426 87.24 1.980000000000004
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv perf_fp16 163 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv perf_fp16 151 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar100/vgg16_cifar100_valid_soc.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar100/vgg16_cifar100_valid_soc.txt
deleted file mode 100644
index ef6509b99b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar100/vgg16_cifar100_valid_soc.txt
+++ /dev/null
@@ -1,210 +0,0 @@
-3845.438677999999
-+++++
-conf1 1 1 68.42 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 relu fp32 1
-4 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-5 gpu conv fp32 11 add fp32 1 relu fp32 1
-6 gpu conv fp32 11 add fp32 1 relu fp32 1
-7 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-8 gpu conv fp32 11 add fp32 1 relu fp32 1
-9 gpu conv fp32 11 add fp32 1 relu fp32 1
-10 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-11 gpu conv fp32 11 add fp32 1 relu fp32 1
-12 gpu conv fp32 11 add fp32 1 relu fp32 1
-13 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-14 gpu mul fp32 11 add fp32 1 relu fp32 1
-15 gpu mul fp32 11 add fp32 1
-16 gpu softmax fp32 1
------
-+++++
-conf2 2.4361074671227554 1.7555866253938424 67.22 1.2000000000000028
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 163 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 269 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv fp16 11 add fp16 12 relu fp16 12
-12 gpu conv fp16 11 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 264 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf3 2.602684148359414 1.8286503060252126 67.10000000000001 1.3199999999999932
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 156 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv fp16 11 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf4 2.661880095451371 1.886369953641946 67.06 1.3599999999999994
-1 gpu conv fp16 12 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 156 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf5 2.5990656605003855 1.8588553950032938 66.84 1.5799999999999983
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 163 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf6 2.5884968081531485 1.8594972115815722 66.8 1.6200000000000045
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 165 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf7 2.4323231936537972 1.8028228076034056 66.8 1.6200000000000045
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 269 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf8 2.575472326184571 1.8375078883357683 66.72 1.7000000000000028
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv perf_fp16 161 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-12 gpu conv fp16 11 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf9 2.4912510106198957 1.848807665058795 66.58 1.8400000000000034
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 266 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf10 2.4323231936537972 1.8028228076034056 66.53999999999999 1.8800000000000097
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 269 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv perf_fp16 152 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
-+++++
-conf11 2.4027045398540046 1.7853827712848849 66.47999999999999 1.940000000000012
-1 gpu conv fp16 11 add fp16 12 relu fp16 12
-2 gpu conv samp_fp16 269 add fp16 12 relu fp16 12 pool_max fp16 12
-3 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-4 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-5 gpu conv fp16 12 add fp16 12 relu fp16 12
-6 gpu conv samp_fp16 261 add fp16 12 relu fp16 12
-7 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-8 gpu conv perf_fp16 155 add fp16 12 relu fp16 12
-9 gpu conv samp_fp16 262 add fp16 12 relu fp16 12
-10 gpu conv samp_fp16 262 add fp16 12 relu fp16 12 pool_max fp16 12
-11 gpu conv perf_fp16 160 add fp16 12 relu fp16 12
-12 gpu conv perf_fp16 151 add fp16 12 relu fp16 12
-13 gpu conv samp_fp16 261 add fp16 12 relu fp16 12 pool_max fp16 12
-14 gpu mul fp16 12 add fp16 12 relu fp16 12
-15 gpu mul fp16 12 add fp16 12
-16 gpu softmax fp16 12
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/README.md b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/README.md
deleted file mode 100644
index e69de29bb2..0000000000
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/alexnet2_cifar10/alexnet2_cifar10.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/alexnet2_cifar10/alexnet2_cifar10.txt
deleted file mode 100644
index 208f154e02..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/alexnet2_cifar10/alexnet2_cifar10.txt
+++ /dev/null
@@ -1,23 +0,0 @@
-1114.3009809999999
-+++++
-conf1 1 1 84.76 0.0
-1 gpu conv fp32 11 add fp32 1 tanh fp32 1
-2 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 tanh fp32 1
-4 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-5 gpu conv fp32 11 add fp32 1 tanh fp32 1
-6 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-7 gpu mul fp32 11 add fp32 1
-8 gpu softmax fp32 1
------
-+++++
-conf2 1.678391931801309 1.4393008204786808 84.76 0.0
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/alexnet_cifar10/alexnet_cifar10.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/alexnet_cifar10/alexnet_cifar10.txt
deleted file mode 100644
index eba22e3f01..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/alexnet_cifar10/alexnet_cifar10.txt
+++ /dev/null
@@ -1,421 +0,0 @@
-2592.187221
-+++++
-conf1 1 1 78.78 0.0
-1 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-2 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 tanh fp32 1
-4 gpu conv fp32 11 add fp32 1 tanh fp32 1
-5 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-6 gpu mul fp32 11 add fp32 1
-7 gpu softmax fp32 1
------
-+++++
-conf2 1.7593976485873195 1.6193399031642917 78.78 0.0
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf3 2.081712090729918 1.9102226906341664 78.53999999999999 0.2400000000000091
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf4 2.081712090729918 1.9102226906341664 78.53999999999999 0.2400000000000091
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf5 2.2627828537139263 2.065683616898884 78.34 0.4399999999999977
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf6 2.3527290658539215 2.145832257234814 78.10000000000001 0.6799999999999926
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf7 2.3527290658539215 2.145832257234814 78.10000000000001 0.6799999999999926
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf8 2.3527290658539215 2.145832257234814 78.10000000000001 0.6799999999999926
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf9 2.2247938983110425 2.060416584958474 77.98 0.7999999999999972
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf10 2.2247938983110425 2.060416584958474 77.98 0.7999999999999972
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf11 2.4370818494175888 2.250857540113024 77.98 0.7999999999999972
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf12 2.432854949808342 2.2424500615508003 77.9 0.8799999999999955
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf13 2.432854949808342 2.2424500615508003 77.9 0.8799999999999955
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf14 2.432854949808342 2.2424500615508003 77.9 0.8799999999999955
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf15 2.228328207535687 2.0675123320068267 77.82 0.960000000000008
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf16 2.228328207535687 2.0675123320068267 77.82 0.960000000000008
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf17 2.3417491169395532 2.1355030360671465 77.78 1.0
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf18 2.3417491169395532 2.1355030360671465 77.78 1.0
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf19 2.3417491169395532 2.1355030360671465 77.78 1.0
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf20 2.5243776633638846 2.324968713897418 77.78 1.0
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf21 2.5243776633638846 2.324968713897418 77.78 1.0
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf22 2.5243776633638846 2.324968713897418 77.78 1.0
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf23 2.5371416718362823 2.3372173527293847 77.56 1.2199999999999989
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf24 2.5371416718362823 2.3372173527293847 77.56 1.2199999999999989
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf25 2.472472828611022 2.286262888143739 77.48 1.2999999999999972
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf26 2.574475112841438 2.3637004022727544 77.4 1.3799999999999955
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf27 2.1200397577541747 1.951741010849448 77.3 1.480000000000004
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf28 2.1200397577541747 1.951741010849448 77.3 1.480000000000004
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf29 2.5289288699015304 2.334007588396142 77.2 1.5799999999999983
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf30 2.5289288699015304 2.334007588396142 77.2 1.5799999999999983
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf31 2.5289288699015304 2.334007588396142 77.2 1.5799999999999983
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf32 2.541739061163583 2.3463519042470864 77.18 1.5999999999999943
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf33 2.541739061163583 2.3463519042470864 77.18 1.5999999999999943
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf34 2.580258965052788 2.3848508703934153 76.96 1.8200000000000074
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf35 2.580258965052788 2.3848508703934153 76.96 1.8200000000000074
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf36 2.4768386387310675 2.295002745725082 76.94 1.8400000000000034
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf37 2.5713008246729716 2.3684101116633007 76.94 1.8400000000000034
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 269 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf38 2.5713008246729716 2.3684101116633007 76.94 1.8400000000000034
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 269 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf39 2.5670585645212847 2.3720992406158463 76.92 1.8599999999999994
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf40 2.5670585645212847 2.3720992406158463 76.92 1.8599999999999994
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf41 2.5760229577267673 2.3777906009584133 76.9 1.8799999999999955
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 269 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf42 2.5760229577267673 2.3777906009584133 76.9 1.8799999999999955
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 269 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/alexnet_imagenet/alexnet_imagenet.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/alexnet_imagenet/alexnet_imagenet.txt
deleted file mode 100644
index 8ae986b90c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/alexnet_imagenet/alexnet_imagenet.txt
+++ /dev/null
@@ -1,289 +0,0 @@
-2739.950736
-+++++
-conf1 1 1 56.3 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 relu fp32 1
-4 gpu conv fp32 11 add fp32 1 relu fp32 1
-5 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-6 gpu mul fp32 11 add fp32 1 relu fp32 1
-7 gpu mul fp32 11 add fp32 1 relu fp32 1
-8 gpu mul fp32 11 add fp32 1
-9 gpu softmax fp32 1
------
-+++++
-conf2 1.802133644103582 1.8186433204507424 55.76 0.5399999999999991
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf3 2.0227701930718065 2.043112495268932 55.42 0.8799999999999955
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 166 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf4 1.8063132288735129 1.8239088223620996 54.96 1.3399999999999963
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf5 1.8063132288735129 1.8239088223620996 54.96 1.3399999999999963
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf6 1.8063132288735129 1.8239088223620996 54.96 1.3399999999999963
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf7 2.085011755614172 2.122606306624671 54.92 1.3799999999999955
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 159 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf8 2.085011755614172 2.122606306624671 54.92 1.3799999999999955
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 159 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf9 1.8052659214923805 1.8217111622759978 54.82 1.4799999999999969
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf10 2.0146435217865446 2.0367475358800102 54.58 1.7199999999999989
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf11 1.9101312060368951 1.9552389688678584 54.24 2.059999999999995
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 157 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf12 1.9101312060368951 1.9552389688678584 54.24 2.059999999999995
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 157 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf13 1.9101312060368951 1.9552389688678584 54.24 2.059999999999995
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 157 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf14 2.019868378233057 2.0433540129730265 54.17999999999999 2.1200000000000045
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf15 2.019868378233057 2.0433540129730265 54.17999999999999 2.1200000000000045
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf16 2.028037341700216 2.049760395549724 53.98 2.3200000000000003
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 166 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf17 2.028037341700216 2.049760395549724 53.98 2.3200000000000003
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 166 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf18 2.028037341700216 2.049760395549724 53.98 2.3200000000000003
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 166 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf19 1.8052659214923805 1.8217111622759978 53.879999999999995 2.4200000000000017
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 11 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf20 1.8052659214923805 1.8217111622759978 53.879999999999995 2.4200000000000017
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 11 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf21 2.0267172350289036 2.046985186681549 53.86 2.4399999999999977
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 166 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf22 2.0267172350289036 2.046985186681549 53.86 2.4399999999999977
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 166 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf23 2.0267172350289036 2.046985186681549 53.86 2.4399999999999977
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 166 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf24 2.0185588815268836 2.0405961127674277 53.559999999999995 2.740000000000002
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/lenet_keras/lenet_keras.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/lenet_keras/lenet_keras.txt
deleted file mode 100644
index da88f7cd26..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/lenet_keras/lenet_keras.txt
+++ /dev/null
@@ -1,409 +0,0 @@
-282.5141369999999
-+++++
-conf1 1 1 98.7 0.0
-1 gpu conv fp32 11 add fp32 1 pool_max fp32 1 tanh fp32 1
-2 gpu conv fp32 11 add fp32 1 pool_max fp32 1 tanh fp32 1
-3 gpu mul fp32 11 add fp32 1 tanh fp32 1
-4 gpu mul fp32 11 add fp32 1 tanh fp32 1
-5 gpu softmax fp32 1
------
-+++++
-conf2 1.9343699741206566 2.1183040240042 98.68 0.01999999999999602
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 265 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf3 1.9343699741206566 2.1183040240042 98.68 0.01999999999999602
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 265 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf4 1.8936889628815377 2.139779619692146 98.68 0.01999999999999602
-1 gpu conv perf_fp16 152 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf5 1.8936889628815377 2.139779619692146 98.68 0.01999999999999602
-1 gpu conv perf_fp16 152 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf6 1.6415764141643088 1.8012120076077847 98.66 0.04000000000000625
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 265 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf7 1.9358279784215788 2.1233340385374495 98.66 0.04000000000000625
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf8 1.9358279784215788 2.1233340385374495 98.66 0.04000000000000625
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf9 1.6319327047042609 1.8046853367113418 98.64 0.060000000000002274
-1 gpu conv fp16 12 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf10 1.6319327047042609 1.8046853367113418 98.64 0.060000000000002274
-1 gpu conv fp16 12 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf11 1.6319327047042609 1.8046853367113418 98.64 0.060000000000002274
-1 gpu conv fp16 12 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf12 1.6319327047042609 1.8046853367113418 98.64 0.060000000000002274
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf13 1.6319327047042609 1.8046853367113418 98.64 0.060000000000002274
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf14 1.5602284338468988 1.7102497386784767 98.61999999999999 0.0800000000000125
-1 gpu conv fp16 12 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf15 1.5602284338468988 1.7102497386784767 98.61999999999999 0.0800000000000125
-1 gpu conv fp16 12 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf16 1.5602284338468988 1.7102497386784767 98.61999999999999 0.0800000000000125
-1 gpu conv fp16 12 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf17 1.8224050632690918 1.9936046569348063 98.61999999999999 0.0800000000000125
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf18 1.8224050632690918 1.9936046569348063 98.61999999999999 0.0800000000000125
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf19 1.8224050632690918 1.9936046569348063 98.61999999999999 0.0800000000000125
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf20 2.2168527051833635 2.453341076720038 98.61999999999999 0.0800000000000125
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf21 2.2168527051833635 2.453341076720038 98.61999999999999 0.0800000000000125
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf22 1.9040998718547615 2.1501783570812565 98.61999999999999 0.0800000000000125
-1 gpu conv perf_fp16 151 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf23 1.9040998718547615 2.1501783570812565 98.61999999999999 0.0800000000000125
-1 gpu conv perf_fp16 151 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf24 1.5630416487818 1.7451546885860074 98.6 0.10000000000000853
-1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf25 1.5630416487818 1.7451546885860074 98.6 0.10000000000000853
-1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf26 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
-1 gpu conv fp16 12 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf27 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
-1 gpu conv fp16 12 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf28 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf29 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf30 2.1941568976363475 2.4445764373737644 98.6 0.10000000000000853
-1 gpu conv samp_fp16 269 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf31 2.1941568976363475 2.4445764373737644 98.6 0.10000000000000853
-1 gpu conv samp_fp16 269 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf32 1.5602284338468988 1.7102497386784767 98.58 0.12000000000000455
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf33 1.5602284338468988 1.7102497386784767 98.58 0.12000000000000455
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf34 1.5602284338468988 1.7102497386784767 98.58 0.12000000000000455
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf35 1.9209933607603906 2.123109543083542 98.58 0.12000000000000455
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf36 1.9209933607603906 2.123109543083542 98.58 0.12000000000000455
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf37 1.9209933607603906 2.123109543083542 98.58 0.12000000000000455
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf38 1.8406161850501603 2.037849502542524 98.58 0.12000000000000455
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf39 1.8406161850501603 2.037849502542524 98.58 0.12000000000000455
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf40 1.8445326456180258 2.087601822059355 98.58 0.12000000000000455
-1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf41 1.8445326456180258 2.087601822059355 98.58 0.12000000000000455
-1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf42 1.8649226857257986 2.1076025277601325 98.56 0.14000000000000057
-1 gpu conv perf_fp16 168 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf43 1.8649226857257986 2.1076025277601325 98.56 0.14000000000000057
-1 gpu conv perf_fp16 168 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf44 1.8463058650555446 2.067271423078985 98.56 0.14000000000000057
-1 gpu conv perf_fp16 157 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf45 1.8463058650555446 2.067271423078985 98.56 0.14000000000000057
-1 gpu conv perf_fp16 157 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf46 1.9234076467497994 2.1864740913112275 98.56 0.14000000000000057
-1 gpu conv perf_fp16 153 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf47 1.9234076467497994 2.1864740913112275 98.56 0.14000000000000057
-1 gpu conv perf_fp16 153 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf48 1.6319327047042609 1.8046853367113418 98.54 0.1599999999999966
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf49 1.6350106933897723 1.8435952834193967 98.52 0.18000000000000682
-1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf50 1.6350106933897723 1.8435952834193967 98.52 0.18000000000000682
-1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf51 1.6510114896409525 1.8591762752048948 98.48 0.21999999999999886
-1 gpu conv perf_fp16 168 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/mobilenet_cifar10/mobilenet_cifar10.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/mobilenet_cifar10/mobilenet_cifar10.txt
deleted file mode 100644
index 93ca37c00a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/mobilenet_cifar10/mobilenet_cifar10.txt
+++ /dev/null
@@ -1,871 +0,0 @@
-4077.307063200001
-+++++
-conf1 1 1 84.42 0.0
-1 gpu conv fp32 11
-2 gpu batchnorm fp32 11
-3 gpu relu fp32 11
-4 gpu group_conv fp32 11
-5 gpu batchnorm fp32 11
-6 gpu relu fp32 11
-7 gpu conv fp32 11
-8 gpu batchnorm fp32 11
-9 gpu relu fp32 11
-10 gpu group_conv fp32 11
-11 gpu batchnorm fp32 11
-12 gpu relu fp32 11
-13 gpu conv fp32 11
-14 gpu batchnorm fp32 11
-15 gpu relu fp32 11
-16 gpu group_conv fp32 11
-17 gpu batchnorm fp32 11
-18 gpu relu fp32 11
-19 gpu conv fp32 11
-20 gpu batchnorm fp32 11
-21 gpu relu fp32 11
-22 gpu group_conv fp32 11
-23 gpu batchnorm fp32 11
-24 gpu relu fp32 11
-25 gpu conv fp32 11
-26 gpu batchnorm fp32 11
-27 gpu relu fp32 11
-28 gpu group_conv fp32 11
-29 gpu batchnorm fp32 11
-30 gpu relu fp32 11
-31 gpu conv fp32 11
-32 gpu batchnorm fp32 11
-33 gpu relu fp32 11
-34 gpu group_conv fp32 11
-35 gpu batchnorm fp32 11
-36 gpu relu fp32 11
-37 gpu conv fp32 11
-38 gpu batchnorm fp32 11
-39 gpu relu fp32 11
-40 gpu group_conv fp32 11
-41 gpu batchnorm fp32 11
-42 gpu relu fp32 11
-43 gpu conv fp32 11
-44 gpu batchnorm fp32 11
-45 gpu relu fp32 11
-46 gpu group_conv fp32 11
-47 gpu batchnorm fp32 11
-48 gpu relu fp32 11
-49 gpu conv fp32 11
-50 gpu batchnorm fp32 11
-51 gpu relu fp32 11
-52 gpu group_conv fp32 11
-53 gpu batchnorm fp32 11
-54 gpu relu fp32 11
-55 gpu conv fp32 11
-56 gpu batchnorm fp32 11
-57 gpu relu fp32 11
-58 gpu group_conv fp32 11
-59 gpu batchnorm fp32 11
-60 gpu relu fp32 11
-61 gpu conv fp32 11
-62 gpu batchnorm fp32 11
-63 gpu relu fp32 11
-64 gpu group_conv fp32 11
-65 gpu batchnorm fp32 11
-66 gpu relu fp32 11
-67 gpu conv fp32 11
-68 gpu batchnorm fp32 11
-69 gpu relu fp32 11
-70 gpu group_conv fp32 11
-71 gpu batchnorm fp32 11
-72 gpu relu fp32 11
-73 gpu conv fp32 11
-74 gpu batchnorm fp32 11
-75 gpu relu fp32 11
-76 gpu group_conv fp32 11
-77 gpu batchnorm fp32 11
-78 gpu relu fp32 11
-79 gpu conv fp32 11
-80 gpu batchnorm fp32 11
-81 gpu relu fp32 11
-82 gpu pool_mean fp32 11
-83 gpu mul fp32 11 add fp32 1
-84 gpu softmax fp32 1
------
-+++++
-conf2 1.504059255565631 1.4598468219902432 81.86 2.5600000000000023
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 152
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 152
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf3 1.5040783418076804 1.459845395800413 81.86 2.5600000000000023
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 152
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 152
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf4 1.5042737817275433 1.4598464522370567 81.74 2.680000000000007
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 155
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 152
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf5 1.5042737817275433 1.4598464522370567 81.74 2.680000000000007
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 155
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 152
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf6 1.5070383438802568 1.463241585164149 81.69999999999999 2.720000000000013
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 168
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 152
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 152
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf7 1.5070575058058588 1.463240152333617 81.58 2.8400000000000034
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 168
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 152
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 152
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf8 1.5039678813445672 1.4598454486222088 81.56 2.8599999999999994
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 152
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 152
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 153
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf9 1.5038655354281372 1.4599130636549171 81.46 2.960000000000008
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 161
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 152
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 152
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 153
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf10 1.4785375660713596 1.4280520288797043 84.42 0.0
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv fp16 12
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv fp16 12
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv fp16 12
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv fp16 12
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv fp16 12
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/resnet18_cifar10/resnet18_cifar10.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/resnet18_cifar10/resnet18_cifar10.txt
deleted file mode 100644
index d1d75a011e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/resnet18_cifar10/resnet18_cifar10.txt
+++ /dev/null
@@ -1,91 +0,0 @@
-2484.981244
-+++++
-conf1 1 1 89.42 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1
-3 gpu conv fp32 11 add fp32 1
-4 gpu add fp32 11
-5 gpu relu fp32 11
-6 gpu conv fp32 11 add fp32 1 relu fp32 1
-7 gpu conv fp32 11 add fp32 1
-8 gpu add fp32 11
-9 gpu relu fp32 11
-10 gpu conv fp32 11 add fp32 1 relu fp32 1
-11 gpu conv fp32 11 add fp32 1
-12 gpu add fp32 11
-13 gpu relu fp32 11
-14 gpu conv fp32 11 add fp32 1 relu fp32 1
-15 gpu conv fp32 11 add fp32 1
-16 gpu conv fp32 11 add fp32 1
-17 gpu add fp32 11
-18 gpu relu fp32 11
-19 gpu conv fp32 11 add fp32 1 relu fp32 1
-20 gpu conv fp32 11 add fp32 1
-21 gpu add fp32 11
-22 gpu relu fp32 11
-23 gpu conv fp32 11 add fp32 1 relu fp32 1
-24 gpu conv fp32 11 add fp32 1
-25 gpu add fp32 11
-26 gpu relu fp32 11
-27 gpu conv fp32 11 add fp32 1 relu fp32 1
-28 gpu conv fp32 11 add fp32 1
-29 gpu conv fp32 11 add fp32 1
-30 gpu add fp32 11
-31 gpu relu fp32 11
-32 gpu conv fp32 11 add fp32 1 relu fp32 1
-33 gpu conv fp32 11 add fp32 1
-34 gpu add fp32 11
-35 gpu relu fp32 11
-36 gpu conv fp32 11 add fp32 1 relu fp32 1
-37 gpu conv fp32 11 add fp32 1
-38 gpu add fp32 11
-39 gpu relu fp32 11
-40 gpu pool_mean fp32 11
-41 gpu mul fp32 11 add fp32 1
-42 gpu softmax fp32 1
------
-+++++
-conf2 1.3617910209460897 1.3866827244386561 89.42 0.0
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv fp16 12 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv fp16 12 add fp16 1 relu fp16 1
-11 gpu conv fp16 12 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 12 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12 add fp16 1 relu fp16 1
-20 gpu conv fp16 12 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv fp16 12 add fp16 1 relu fp16 1
-24 gpu conv fp16 12 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 12 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv fp16 12 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/resnet50_imagenet/resnet50_imagenet.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/resnet50_imagenet/resnet50_imagenet.txt
deleted file mode 100644
index a045011580..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/resnet50_imagenet/resnet50_imagenet.txt
+++ /dev/null
@@ -1,1233 +0,0 @@
-7161.053769000008
-+++++
-conf1 1 1 75.7 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-2 gpu batchnorm fp32 11
-3 gpu conv fp32 11 add fp32 1
-4 gpu batchnorm fp32 11
-5 gpu relu fp32 11
-6 gpu conv fp32 11 add fp32 1
-7 gpu batchnorm fp32 11
-8 gpu relu fp32 11
-9 gpu conv fp32 11 add fp32 1
-10 gpu batchnorm fp32 11
-11 gpu conv fp32 11 add fp32 1
-12 gpu batchnorm fp32 11
-13 gpu add fp32 11
-14 gpu relu fp32 11
-15 gpu conv fp32 11 add fp32 1
-16 gpu batchnorm fp32 11
-17 gpu relu fp32 11
-18 gpu conv fp32 11 add fp32 1
-19 gpu batchnorm fp32 11
-20 gpu relu fp32 11
-21 gpu conv fp32 11 add fp32 1
-22 gpu batchnorm fp32 11
-23 gpu add fp32 11
-24 gpu relu fp32 11
-25 gpu conv fp32 11 add fp32 1
-26 gpu batchnorm fp32 11
-27 gpu relu fp32 11
-28 gpu conv fp32 11 add fp32 1
-29 gpu batchnorm fp32 11
-30 gpu relu fp32 11
-31 gpu conv fp32 11 add fp32 1
-32 gpu batchnorm fp32 11
-33 gpu add fp32 11
-34 gpu relu fp32 11
-35 gpu conv fp32 11 add fp32 1
-36 gpu batchnorm fp32 11
-37 gpu relu fp32 11
-38 gpu conv fp32 11 add fp32 1
-39 gpu batchnorm fp32 11
-40 gpu relu fp32 11
-41 gpu conv fp32 11 add fp32 1
-42 gpu batchnorm fp32 11
-43 gpu conv fp32 11 add fp32 1
-44 gpu batchnorm fp32 11
-45 gpu add fp32 11
-46 gpu relu fp32 11
-47 gpu conv fp32 11 add fp32 1
-48 gpu batchnorm fp32 11
-49 gpu relu fp32 11
-50 gpu conv fp32 11 add fp32 1
-51 gpu batchnorm fp32 11
-52 gpu relu fp32 11
-53 gpu conv fp32 11 add fp32 1
-54 gpu batchnorm fp32 11
-55 gpu add fp32 11
-56 gpu relu fp32 11
-57 gpu conv fp32 11 add fp32 1
-58 gpu batchnorm fp32 11
-59 gpu relu fp32 11
-60 gpu conv fp32 11 add fp32 1
-61 gpu batchnorm fp32 11
-62 gpu relu fp32 11
-63 gpu conv fp32 11 add fp32 1
-64 gpu batchnorm fp32 11
-65 gpu add fp32 11
-66 gpu relu fp32 11
-67 gpu conv fp32 11 add fp32 1
-68 gpu batchnorm fp32 11
-69 gpu relu fp32 11
-70 gpu conv fp32 11 add fp32 1
-71 gpu batchnorm fp32 11
-72 gpu relu fp32 11
-73 gpu conv fp32 11 add fp32 1
-74 gpu batchnorm fp32 11
-75 gpu add fp32 11
-76 gpu relu fp32 11
-77 gpu conv fp32 11 add fp32 1
-78 gpu batchnorm fp32 11
-79 gpu relu fp32 11
-80 gpu conv fp32 11 add fp32 1
-81 gpu batchnorm fp32 11
-82 gpu relu fp32 11
-83 gpu conv fp32 11 add fp32 1
-84 gpu batchnorm fp32 11
-85 gpu conv fp32 11 add fp32 1
-86 gpu batchnorm fp32 11
-87 gpu add fp32 11
-88 gpu relu fp32 11
-89 gpu conv fp32 11 add fp32 1
-90 gpu batchnorm fp32 11
-91 gpu relu fp32 11
-92 gpu conv fp32 11 add fp32 1
-93 gpu batchnorm fp32 11
-94 gpu relu fp32 11
-95 gpu conv fp32 11 add fp32 1
-96 gpu batchnorm fp32 11
-97 gpu add fp32 11
-98 gpu relu fp32 11
-99 gpu conv fp32 11 add fp32 1
-100 gpu batchnorm fp32 11
-101 gpu relu fp32 11
-102 gpu conv fp32 11 add fp32 1
-103 gpu batchnorm fp32 11
-104 gpu relu fp32 11
-105 gpu conv fp32 11 add fp32 1
-106 gpu batchnorm fp32 11
-107 gpu add fp32 11
-108 gpu relu fp32 11
-109 gpu conv fp32 11 add fp32 1
-110 gpu batchnorm fp32 11
-111 gpu relu fp32 11
-112 gpu conv fp32 11 add fp32 1
-113 gpu batchnorm fp32 11
-114 gpu relu fp32 11
-115 gpu conv fp32 11 add fp32 1
-116 gpu batchnorm fp32 11
-117 gpu add fp32 11
-118 gpu relu fp32 11
-119 gpu conv fp32 11 add fp32 1
-120 gpu batchnorm fp32 11
-121 gpu relu fp32 11
-122 gpu conv fp32 11 add fp32 1
-123 gpu batchnorm fp32 11
-124 gpu relu fp32 11
-125 gpu conv fp32 11 add fp32 1
-126 gpu batchnorm fp32 11
-127 gpu add fp32 11
-128 gpu relu fp32 11
-129 gpu conv fp32 11 add fp32 1
-130 gpu batchnorm fp32 11
-131 gpu relu fp32 11
-132 gpu conv fp32 11 add fp32 1
-133 gpu batchnorm fp32 11
-134 gpu relu fp32 11
-135 gpu conv fp32 11 add fp32 1
-136 gpu batchnorm fp32 11
-137 gpu add fp32 11
-138 gpu relu fp32 11
-139 gpu conv fp32 11 add fp32 1
-140 gpu batchnorm fp32 11
-141 gpu relu fp32 11
-142 gpu conv fp32 11 add fp32 1
-143 gpu batchnorm fp32 11
-144 gpu relu fp32 11
-145 gpu conv fp32 11 add fp32 1
-146 gpu batchnorm fp32 11
-147 gpu conv fp32 11 add fp32 1
-148 gpu batchnorm fp32 11
-149 gpu add fp32 11
-150 gpu relu fp32 11
-151 gpu conv fp32 11 add fp32 1
-152 gpu batchnorm fp32 11
-153 gpu relu fp32 11
-154 gpu conv fp32 11 add fp32 1
-155 gpu batchnorm fp32 11
-156 gpu relu fp32 11
-157 gpu conv fp32 11 add fp32 1
-158 gpu batchnorm fp32 11
-159 gpu add fp32 11
-160 gpu relu fp32 11
-161 gpu conv fp32 11 add fp32 1
-162 gpu batchnorm fp32 11
-163 gpu relu fp32 11
-164 gpu conv fp32 11 add fp32 1
-165 gpu batchnorm fp32 11
-166 gpu relu fp32 11
-167 gpu conv fp32 11 add fp32 1
-168 gpu batchnorm fp32 11
-169 gpu add fp32 11
-170 gpu relu fp32 11
-171 gpu pool_max fp32 11
-172 gpu mul fp32 11 add fp32 1
-173 gpu softmax fp32 1
------
-+++++
-conf2 1.8254789092281507 1.4527803526239977 75.7 0.0
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu batchnorm fp16 12
-3 gpu conv fp16 12 add fp16 1
-4 gpu batchnorm fp16 12
-5 gpu relu fp16 12
-6 gpu conv fp16 12 add fp16 1
-7 gpu batchnorm fp16 12
-8 gpu relu fp16 12
-9 gpu conv fp16 12 add fp16 1
-10 gpu batchnorm fp16 12
-11 gpu conv fp16 12 add fp16 1
-12 gpu batchnorm fp16 12
-13 gpu add fp16 12
-14 gpu relu fp16 12
-15 gpu conv fp16 12 add fp16 1
-16 gpu batchnorm fp16 12
-17 gpu relu fp16 12
-18 gpu conv fp16 12 add fp16 1
-19 gpu batchnorm fp16 12
-20 gpu relu fp16 12
-21 gpu conv fp16 12 add fp16 1
-22 gpu batchnorm fp16 12
-23 gpu add fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12 add fp16 1
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu conv fp16 12 add fp16 1
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12 add fp16 1
-32 gpu batchnorm fp16 12
-33 gpu add fp16 12
-34 gpu relu fp16 12
-35 gpu conv fp16 12 add fp16 1
-36 gpu batchnorm fp16 12
-37 gpu relu fp16 12
-38 gpu conv fp16 12 add fp16 1
-39 gpu batchnorm fp16 12
-40 gpu relu fp16 12
-41 gpu conv fp16 12 add fp16 1
-42 gpu batchnorm fp16 12
-43 gpu conv fp16 12 add fp16 1
-44 gpu batchnorm fp16 12
-45 gpu add fp16 12
-46 gpu relu fp16 12
-47 gpu conv fp16 12 add fp16 1
-48 gpu batchnorm fp16 12
-49 gpu relu fp16 12
-50 gpu conv fp16 12 add fp16 1
-51 gpu batchnorm fp16 12
-52 gpu relu fp16 12
-53 gpu conv fp16 12 add fp16 1
-54 gpu batchnorm fp16 12
-55 gpu add fp16 12
-56 gpu relu fp16 12
-57 gpu conv fp16 12 add fp16 1
-58 gpu batchnorm fp16 12
-59 gpu relu fp16 12
-60 gpu conv fp16 12 add fp16 1
-61 gpu batchnorm fp16 12
-62 gpu relu fp16 12
-63 gpu conv fp16 12 add fp16 1
-64 gpu batchnorm fp16 12
-65 gpu add fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12 add fp16 1
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu conv fp16 12 add fp16 1
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv fp16 12 add fp16 1
-74 gpu batchnorm fp16 12
-75 gpu add fp16 12
-76 gpu relu fp16 12
-77 gpu conv fp16 12 add fp16 1
-78 gpu batchnorm fp16 12
-79 gpu relu fp16 12
-80 gpu conv fp16 12 add fp16 1
-81 gpu batchnorm fp16 12
-82 gpu relu fp16 12
-83 gpu conv fp16 12 add fp16 1
-84 gpu batchnorm fp16 12
-85 gpu conv fp16 12 add fp16 1
-86 gpu batchnorm fp16 12
-87 gpu add fp16 12
-88 gpu relu fp16 12
-89 gpu conv fp16 12 add fp16 1
-90 gpu batchnorm fp16 12
-91 gpu relu fp16 12
-92 gpu conv fp16 12 add fp16 1
-93 gpu batchnorm fp16 12
-94 gpu relu fp16 12
-95 gpu conv fp16 12 add fp16 1
-96 gpu batchnorm fp16 12
-97 gpu add fp16 12
-98 gpu relu fp16 12
-99 gpu conv fp16 12 add fp16 1
-100 gpu batchnorm fp16 12
-101 gpu relu fp16 12
-102 gpu conv fp16 12 add fp16 1
-103 gpu batchnorm fp16 12
-104 gpu relu fp16 12
-105 gpu conv fp16 12 add fp16 1
-106 gpu batchnorm fp16 12
-107 gpu add fp16 12
-108 gpu relu fp16 12
-109 gpu conv fp16 12 add fp16 1
-110 gpu batchnorm fp16 12
-111 gpu relu fp16 12
-112 gpu conv fp16 12 add fp16 1
-113 gpu batchnorm fp16 12
-114 gpu relu fp16 12
-115 gpu conv fp16 12 add fp16 1
-116 gpu batchnorm fp16 12
-117 gpu add fp16 12
-118 gpu relu fp16 12
-119 gpu conv fp16 12 add fp16 1
-120 gpu batchnorm fp16 12
-121 gpu relu fp16 12
-122 gpu conv fp16 12 add fp16 1
-123 gpu batchnorm fp16 12
-124 gpu relu fp16 12
-125 gpu conv fp16 12 add fp16 1
-126 gpu batchnorm fp16 12
-127 gpu add fp16 12
-128 gpu relu fp16 12
-129 gpu conv fp16 12 add fp16 1
-130 gpu batchnorm fp16 12
-131 gpu relu fp16 12
-132 gpu conv fp16 12 add fp16 1
-133 gpu batchnorm fp16 12
-134 gpu relu fp16 12
-135 gpu conv fp16 12 add fp16 1
-136 gpu batchnorm fp16 12
-137 gpu add fp16 12
-138 gpu relu fp16 12
-139 gpu conv fp16 12 add fp16 1
-140 gpu batchnorm fp16 12
-141 gpu relu fp16 12
-142 gpu conv fp16 12 add fp16 1
-143 gpu batchnorm fp16 12
-144 gpu relu fp16 12
-145 gpu conv fp16 12 add fp16 1
-146 gpu batchnorm fp16 12
-147 gpu conv fp16 12 add fp16 1
-148 gpu batchnorm fp16 12
-149 gpu add fp16 12
-150 gpu relu fp16 12
-151 gpu conv fp16 12 add fp16 1
-152 gpu batchnorm fp16 12
-153 gpu relu fp16 12
-154 gpu conv fp16 12 add fp16 1
-155 gpu batchnorm fp16 12
-156 gpu relu fp16 12
-157 gpu conv fp16 12 add fp16 1
-158 gpu batchnorm fp16 12
-159 gpu add fp16 12
-160 gpu relu fp16 12
-161 gpu conv fp16 12 add fp16 1
-162 gpu batchnorm fp16 12
-163 gpu relu fp16 12
-164 gpu conv fp16 12 add fp16 1
-165 gpu batchnorm fp16 12
-166 gpu relu fp16 12
-167 gpu conv fp16 12 add fp16 1
-168 gpu batchnorm fp16 12
-169 gpu add fp16 12
-170 gpu relu fp16 12
-171 gpu pool_max fp16 12
-172 gpu mul fp16 12 add fp16 1
-173 gpu softmax fp32 1
------
-+++++
-conf3 1.8521749055745271 1.472492365706726 75.02 0.6800000000000068
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu batchnorm fp16 12
-3 gpu conv fp16 12 add fp16 1
-4 gpu batchnorm fp16 12
-5 gpu relu fp16 12
-6 gpu conv fp16 12 add fp16 1
-7 gpu batchnorm fp16 12
-8 gpu relu fp16 12
-9 gpu conv fp16 12 add fp16 1
-10 gpu batchnorm fp16 12
-11 gpu conv fp16 12 add fp16 1
-12 gpu batchnorm fp16 12
-13 gpu add fp16 12
-14 gpu relu fp16 12
-15 gpu conv fp16 12 add fp16 1
-16 gpu batchnorm fp16 12
-17 gpu relu fp16 12
-18 gpu conv fp16 12 add fp16 1
-19 gpu batchnorm fp16 12
-20 gpu relu fp16 12
-21 gpu conv fp16 12 add fp16 1
-22 gpu batchnorm fp16 12
-23 gpu add fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12 add fp16 1
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu conv perf_fp16 160 add fp16 1
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 11 add fp16 1
-32 gpu batchnorm fp16 12
-33 gpu add fp16 12
-34 gpu relu fp16 12
-35 gpu conv fp16 12 add fp16 1
-36 gpu batchnorm fp16 12
-37 gpu relu fp16 12
-38 gpu conv fp16 12 add fp16 1
-39 gpu batchnorm fp16 12
-40 gpu relu fp16 12
-41 gpu conv fp16 12 add fp16 1
-42 gpu batchnorm fp16 12
-43 gpu conv fp16 12 add fp16 1
-44 gpu batchnorm fp16 12
-45 gpu add fp16 12
-46 gpu relu fp16 12
-47 gpu conv fp16 12 add fp16 1
-48 gpu batchnorm fp16 12
-49 gpu relu fp16 12
-50 gpu conv fp16 12 add fp16 1
-51 gpu batchnorm fp16 12
-52 gpu relu fp16 12
-53 gpu conv fp16 12 add fp16 1
-54 gpu batchnorm fp16 12
-55 gpu add fp16 12
-56 gpu relu fp16 12
-57 gpu conv fp16 12 add fp16 1
-58 gpu batchnorm fp16 12
-59 gpu relu fp16 12
-60 gpu conv fp16 12 add fp16 1
-61 gpu batchnorm fp16 12
-62 gpu relu fp16 12
-63 gpu conv fp16 12 add fp16 1
-64 gpu batchnorm fp16 12
-65 gpu add fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12 add fp16 1
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu conv fp16 12 add fp16 1
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv fp16 12 add fp16 1
-74 gpu batchnorm fp16 12
-75 gpu add fp16 12
-76 gpu relu fp16 12
-77 gpu conv fp16 12 add fp16 1
-78 gpu batchnorm fp16 12
-79 gpu relu fp16 12
-80 gpu conv fp16 12 add fp16 1
-81 gpu batchnorm fp16 12
-82 gpu relu fp16 12
-83 gpu conv fp16 12 add fp16 1
-84 gpu batchnorm fp16 12
-85 gpu conv fp16 12 add fp16 1
-86 gpu batchnorm fp16 12
-87 gpu add fp16 12
-88 gpu relu fp16 12
-89 gpu conv fp16 12 add fp16 1
-90 gpu batchnorm fp16 12
-91 gpu relu fp16 12
-92 gpu conv fp16 12 add fp16 1
-93 gpu batchnorm fp16 12
-94 gpu relu fp16 12
-95 gpu conv fp16 11 add fp16 1
-96 gpu batchnorm fp16 12
-97 gpu add fp16 12
-98 gpu relu fp16 12
-99 gpu conv fp16 12 add fp16 1
-100 gpu batchnorm fp16 12
-101 gpu relu fp16 12
-102 gpu conv perf_fp16 164 add fp16 1
-103 gpu batchnorm fp16 12
-104 gpu relu fp16 12
-105 gpu conv fp16 12 add fp16 1
-106 gpu batchnorm fp16 12
-107 gpu add fp16 12
-108 gpu relu fp16 12
-109 gpu conv fp16 12 add fp16 1
-110 gpu batchnorm fp16 12
-111 gpu relu fp16 12
-112 gpu conv fp16 12 add fp16 1
-113 gpu batchnorm fp16 12
-114 gpu relu fp16 12
-115 gpu conv fp16 12 add fp16 1
-116 gpu batchnorm fp16 12
-117 gpu add fp16 12
-118 gpu relu fp16 12
-119 gpu conv fp16 12 add fp16 1
-120 gpu batchnorm fp16 12
-121 gpu relu fp16 12
-122 gpu conv fp16 12 add fp16 1
-123 gpu batchnorm fp16 12
-124 gpu relu fp16 12
-125 gpu conv fp16 12 add fp16 1
-126 gpu batchnorm fp16 12
-127 gpu add fp16 12
-128 gpu relu fp16 12
-129 gpu conv fp16 12 add fp16 1
-130 gpu batchnorm fp16 12
-131 gpu relu fp16 12
-132 gpu conv fp16 12 add fp16 1
-133 gpu batchnorm fp16 12
-134 gpu relu fp16 12
-135 gpu conv fp16 12 add fp16 1
-136 gpu batchnorm fp16 12
-137 gpu add fp16 12
-138 gpu relu fp16 12
-139 gpu conv fp16 12 add fp16 1
-140 gpu batchnorm fp16 12
-141 gpu relu fp16 12
-142 gpu conv fp16 12 add fp16 1
-143 gpu batchnorm fp16 12
-144 gpu relu fp16 12
-145 gpu conv fp16 12 add fp16 1
-146 gpu batchnorm fp16 12
-147 gpu conv fp16 12 add fp16 1
-148 gpu batchnorm fp16 12
-149 gpu add fp16 12
-150 gpu relu fp16 12
-151 gpu conv fp16 12 add fp16 1
-152 gpu batchnorm fp16 12
-153 gpu relu fp16 12
-154 gpu conv fp16 12 add fp16 1
-155 gpu batchnorm fp16 12
-156 gpu relu fp16 12
-157 gpu conv fp16 12 add fp16 1
-158 gpu batchnorm fp16 12
-159 gpu add fp16 12
-160 gpu relu fp16 12
-161 gpu conv fp16 12 add fp16 1
-162 gpu batchnorm fp16 12
-163 gpu relu fp16 12
-164 gpu conv fp16 12 add fp16 1
-165 gpu batchnorm fp16 12
-166 gpu relu fp16 12
-167 gpu conv fp16 12 add fp16 1
-168 gpu batchnorm fp16 12
-169 gpu add fp16 12
-170 gpu relu fp16 12
-171 gpu pool_max fp16 12
-172 gpu mul fp16 12 add fp16 1
-173 gpu softmax fp32 1
------
-+++++
-conf4 1.8509087142956673 1.4713858340895483 74.68 1.019999999999996
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu batchnorm fp16 12
-3 gpu conv fp16 12 add fp16 1
-4 gpu batchnorm fp16 12
-5 gpu relu fp16 12
-6 gpu conv fp16 12 add fp16 1
-7 gpu batchnorm fp16 12
-8 gpu relu fp16 12
-9 gpu conv fp16 12 add fp16 1
-10 gpu batchnorm fp16 12
-11 gpu conv fp16 12 add fp16 1
-12 gpu batchnorm fp16 12
-13 gpu add fp16 12
-14 gpu relu fp16 12
-15 gpu conv fp16 12 add fp16 1
-16 gpu batchnorm fp16 12
-17 gpu relu fp16 12
-18 gpu conv fp16 12 add fp16 1
-19 gpu batchnorm fp16 12
-20 gpu relu fp16 12
-21 gpu conv fp16 12 add fp16 1
-22 gpu batchnorm fp16 12
-23 gpu add fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12 add fp16 1
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu conv perf_fp16 160 add fp16 1
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12 add fp16 1
-32 gpu batchnorm fp16 12
-33 gpu add fp16 12
-34 gpu relu fp16 12
-35 gpu conv fp16 12 add fp16 1
-36 gpu batchnorm fp16 12
-37 gpu relu fp16 12
-38 gpu conv fp16 12 add fp16 1
-39 gpu batchnorm fp16 12
-40 gpu relu fp16 12
-41 gpu conv fp16 12 add fp16 1
-42 gpu batchnorm fp16 12
-43 gpu conv fp16 12 add fp16 1
-44 gpu batchnorm fp16 12
-45 gpu add fp16 12
-46 gpu relu fp16 12
-47 gpu conv fp16 12 add fp16 1
-48 gpu batchnorm fp16 12
-49 gpu relu fp16 12
-50 gpu conv fp16 12 add fp16 1
-51 gpu batchnorm fp16 12
-52 gpu relu fp16 12
-53 gpu conv fp16 12 add fp16 1
-54 gpu batchnorm fp16 12
-55 gpu add fp16 12
-56 gpu relu fp16 12
-57 gpu conv fp16 12 add fp16 1
-58 gpu batchnorm fp16 12
-59 gpu relu fp16 12
-60 gpu conv fp16 12 add fp16 1
-61 gpu batchnorm fp16 12
-62 gpu relu fp16 12
-63 gpu conv fp16 12 add fp16 1
-64 gpu batchnorm fp16 12
-65 gpu add fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12 add fp16 1
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu conv fp16 12 add fp16 1
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv fp16 12 add fp16 1
-74 gpu batchnorm fp16 12
-75 gpu add fp16 12
-76 gpu relu fp16 12
-77 gpu conv fp16 12 add fp16 1
-78 gpu batchnorm fp16 12
-79 gpu relu fp16 12
-80 gpu conv fp16 12 add fp16 1
-81 gpu batchnorm fp16 12
-82 gpu relu fp16 12
-83 gpu conv fp16 12 add fp16 1
-84 gpu batchnorm fp16 12
-85 gpu conv fp16 12 add fp16 1
-86 gpu batchnorm fp16 12
-87 gpu add fp16 12
-88 gpu relu fp16 12
-89 gpu conv fp16 12 add fp16 1
-90 gpu batchnorm fp16 12
-91 gpu relu fp16 12
-92 gpu conv fp16 12 add fp16 1
-93 gpu batchnorm fp16 12
-94 gpu relu fp16 12
-95 gpu conv fp16 12 add fp16 1
-96 gpu batchnorm fp16 12
-97 gpu add fp16 12
-98 gpu relu fp16 12
-99 gpu conv fp16 12 add fp16 1
-100 gpu batchnorm fp16 12
-101 gpu relu fp16 12
-102 gpu conv fp16 12 add fp16 1
-103 gpu batchnorm fp16 12
-104 gpu relu fp16 12
-105 gpu conv fp16 12 add fp16 1
-106 gpu batchnorm fp16 12
-107 gpu add fp16 12
-108 gpu relu fp16 12
-109 gpu conv fp16 12 add fp16 1
-110 gpu batchnorm fp16 12
-111 gpu relu fp16 12
-112 gpu conv fp16 12 add fp16 1
-113 gpu batchnorm fp16 12
-114 gpu relu fp16 12
-115 gpu conv fp16 12 add fp16 1
-116 gpu batchnorm fp16 12
-117 gpu add fp16 12
-118 gpu relu fp16 12
-119 gpu conv fp16 12 add fp16 1
-120 gpu batchnorm fp16 12
-121 gpu relu fp16 12
-122 gpu conv fp16 12 add fp16 1
-123 gpu batchnorm fp16 12
-124 gpu relu fp16 12
-125 gpu conv fp16 12 add fp16 1
-126 gpu batchnorm fp16 12
-127 gpu add fp16 12
-128 gpu relu fp16 12
-129 gpu conv fp16 12 add fp16 1
-130 gpu batchnorm fp16 12
-131 gpu relu fp16 12
-132 gpu conv fp16 12 add fp16 1
-133 gpu batchnorm fp16 12
-134 gpu relu fp16 12
-135 gpu conv fp16 12 add fp16 1
-136 gpu batchnorm fp16 12
-137 gpu add fp16 12
-138 gpu relu fp16 12
-139 gpu conv fp16 12 add fp16 1
-140 gpu batchnorm fp16 12
-141 gpu relu fp16 12
-142 gpu conv fp16 12 add fp16 1
-143 gpu batchnorm fp16 12
-144 gpu relu fp16 12
-145 gpu conv fp16 12 add fp16 1
-146 gpu batchnorm fp16 12
-147 gpu conv fp16 12 add fp16 1
-148 gpu batchnorm fp16 12
-149 gpu add fp16 12
-150 gpu relu fp16 12
-151 gpu conv fp16 12 add fp16 1
-152 gpu batchnorm fp16 12
-153 gpu relu fp16 12
-154 gpu conv fp16 12 add fp16 1
-155 gpu batchnorm fp16 12
-156 gpu relu fp16 12
-157 gpu conv fp16 12 add fp16 1
-158 gpu batchnorm fp16 12
-159 gpu add fp16 12
-160 gpu relu fp16 12
-161 gpu conv fp16 12 add fp16 1
-162 gpu batchnorm fp16 12
-163 gpu relu fp16 12
-164 gpu conv fp16 12 add fp16 1
-165 gpu batchnorm fp16 12
-166 gpu relu fp16 12
-167 gpu conv fp16 12 add fp16 1
-168 gpu batchnorm fp16 12
-169 gpu add fp16 12
-170 gpu relu fp16 12
-171 gpu pool_max fp16 12
-172 gpu mul fp16 12 add fp16 1
-173 gpu softmax fp32 1
------
-+++++
-conf5 1.8538077719438253 1.4749308494814874 73.82 1.8800000000000097
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu batchnorm fp16 12
-3 gpu conv fp16 12 add fp16 1
-4 gpu batchnorm fp16 12
-5 gpu relu fp16 12
-6 gpu conv fp16 12 add fp16 1
-7 gpu batchnorm fp16 12
-8 gpu relu fp16 12
-9 gpu conv fp16 12 add fp16 1
-10 gpu batchnorm fp16 12
-11 gpu conv fp16 12 add fp16 1
-12 gpu batchnorm fp16 12
-13 gpu add fp16 12
-14 gpu relu fp16 12
-15 gpu conv fp16 12 add fp16 1
-16 gpu batchnorm fp16 12
-17 gpu relu fp16 12
-18 gpu conv fp16 12 add fp16 1
-19 gpu batchnorm fp16 12
-20 gpu relu fp16 12
-21 gpu conv fp16 12 add fp16 1
-22 gpu batchnorm fp16 12
-23 gpu add fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12 add fp16 1
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu conv perf_fp16 160 add fp16 1
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 11 add fp16 1
-32 gpu batchnorm fp16 12
-33 gpu add fp16 12
-34 gpu relu fp16 12
-35 gpu conv fp16 12 add fp16 1
-36 gpu batchnorm fp16 12
-37 gpu relu fp16 12
-38 gpu conv fp16 12 add fp16 1
-39 gpu batchnorm fp16 12
-40 gpu relu fp16 12
-41 gpu conv fp16 12 add fp16 1
-42 gpu batchnorm fp16 12
-43 gpu conv fp16 12 add fp16 1
-44 gpu batchnorm fp16 12
-45 gpu add fp16 12
-46 gpu relu fp16 12
-47 gpu conv fp16 12 add fp16 1
-48 gpu batchnorm fp16 12
-49 gpu relu fp16 12
-50 gpu conv fp16 12 add fp16 1
-51 gpu batchnorm fp16 12
-52 gpu relu fp16 12
-53 gpu conv perf_fp16 153 add fp16 1
-54 gpu batchnorm fp16 12
-55 gpu add fp16 12
-56 gpu relu fp16 12
-57 gpu conv fp16 12 add fp16 1
-58 gpu batchnorm fp16 12
-59 gpu relu fp16 12
-60 gpu conv fp16 12 add fp16 1
-61 gpu batchnorm fp16 12
-62 gpu relu fp16 12
-63 gpu conv fp16 12 add fp16 1
-64 gpu batchnorm fp16 12
-65 gpu add fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12 add fp16 1
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu conv fp16 12 add fp16 1
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv fp16 12 add fp16 1
-74 gpu batchnorm fp16 12
-75 gpu add fp16 12
-76 gpu relu fp16 12
-77 gpu conv fp16 12 add fp16 1
-78 gpu batchnorm fp16 12
-79 gpu relu fp16 12
-80 gpu conv fp16 12 add fp16 1
-81 gpu batchnorm fp16 12
-82 gpu relu fp16 12
-83 gpu conv fp16 12 add fp16 1
-84 gpu batchnorm fp16 12
-85 gpu conv fp16 12 add fp16 1
-86 gpu batchnorm fp16 12
-87 gpu add fp16 12
-88 gpu relu fp16 12
-89 gpu conv fp16 12 add fp16 1
-90 gpu batchnorm fp16 12
-91 gpu relu fp16 12
-92 gpu conv fp16 12 add fp16 1
-93 gpu batchnorm fp16 12
-94 gpu relu fp16 12
-95 gpu conv fp16 11 add fp16 1
-96 gpu batchnorm fp16 12
-97 gpu add fp16 12
-98 gpu relu fp16 12
-99 gpu conv fp16 12 add fp16 1
-100 gpu batchnorm fp16 12
-101 gpu relu fp16 12
-102 gpu conv perf_fp16 164 add fp16 1
-103 gpu batchnorm fp16 12
-104 gpu relu fp16 12
-105 gpu conv fp16 12 add fp16 1
-106 gpu batchnorm fp16 12
-107 gpu add fp16 12
-108 gpu relu fp16 12
-109 gpu conv fp16 12 add fp16 1
-110 gpu batchnorm fp16 12
-111 gpu relu fp16 12
-112 gpu conv fp16 12 add fp16 1
-113 gpu batchnorm fp16 12
-114 gpu relu fp16 12
-115 gpu conv fp16 12 add fp16 1
-116 gpu batchnorm fp16 12
-117 gpu add fp16 12
-118 gpu relu fp16 12
-119 gpu conv samp_fp16 268 add fp16 1
-120 gpu batchnorm fp16 12
-121 gpu relu fp16 12
-122 gpu conv fp16 12 add fp16 1
-123 gpu batchnorm fp16 12
-124 gpu relu fp16 12
-125 gpu conv fp16 12 add fp16 1
-126 gpu batchnorm fp16 12
-127 gpu add fp16 12
-128 gpu relu fp16 12
-129 gpu conv fp16 12 add fp16 1
-130 gpu batchnorm fp16 12
-131 gpu relu fp16 12
-132 gpu conv fp16 12 add fp16 1
-133 gpu batchnorm fp16 12
-134 gpu relu fp16 12
-135 gpu conv fp16 11 add fp16 1
-136 gpu batchnorm fp16 12
-137 gpu add fp16 12
-138 gpu relu fp16 12
-139 gpu conv fp16 12 add fp16 1
-140 gpu batchnorm fp16 12
-141 gpu relu fp16 12
-142 gpu conv fp16 12 add fp16 1
-143 gpu batchnorm fp16 12
-144 gpu relu fp16 12
-145 gpu conv fp16 12 add fp16 1
-146 gpu batchnorm fp16 12
-147 gpu conv fp16 12 add fp16 1
-148 gpu batchnorm fp16 12
-149 gpu add fp16 12
-150 gpu relu fp16 12
-151 gpu conv fp16 12 add fp16 1
-152 gpu batchnorm fp16 12
-153 gpu relu fp16 12
-154 gpu conv fp16 12 add fp16 1
-155 gpu batchnorm fp16 12
-156 gpu relu fp16 12
-157 gpu conv fp16 11 add fp16 1
-158 gpu batchnorm fp16 12
-159 gpu add fp16 12
-160 gpu relu fp16 12
-161 gpu conv fp16 12 add fp16 1
-162 gpu batchnorm fp16 12
-163 gpu relu fp16 12
-164 gpu conv fp16 12 add fp16 1
-165 gpu batchnorm fp16 12
-166 gpu relu fp16 12
-167 gpu conv fp16 12 add fp16 1
-168 gpu batchnorm fp16 12
-169 gpu add fp16 12
-170 gpu relu fp16 12
-171 gpu pool_max fp16 12
-172 gpu mul fp16 12 add fp16 1
-173 gpu softmax fp32 1
------
-+++++
-conf6 1.8538077719438253 1.4749308494814874 73.7 2.0
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu batchnorm fp16 12
-3 gpu conv fp16 12 add fp16 1
-4 gpu batchnorm fp16 12
-5 gpu relu fp16 12
-6 gpu conv fp16 12 add fp16 1
-7 gpu batchnorm fp16 12
-8 gpu relu fp16 12
-9 gpu conv fp16 12 add fp16 1
-10 gpu batchnorm fp16 12
-11 gpu conv fp16 12 add fp16 1
-12 gpu batchnorm fp16 12
-13 gpu add fp16 12
-14 gpu relu fp16 12
-15 gpu conv fp16 12 add fp16 1
-16 gpu batchnorm fp16 12
-17 gpu relu fp16 12
-18 gpu conv fp16 12 add fp16 1
-19 gpu batchnorm fp16 12
-20 gpu relu fp16 12
-21 gpu conv fp16 12 add fp16 1
-22 gpu batchnorm fp16 12
-23 gpu add fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12 add fp16 1
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu conv perf_fp16 160 add fp16 1
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 11 add fp16 1
-32 gpu batchnorm fp16 12
-33 gpu add fp16 12
-34 gpu relu fp16 12
-35 gpu conv fp16 12 add fp16 1
-36 gpu batchnorm fp16 12
-37 gpu relu fp16 12
-38 gpu conv fp16 12 add fp16 1
-39 gpu batchnorm fp16 12
-40 gpu relu fp16 12
-41 gpu conv fp16 12 add fp16 1
-42 gpu batchnorm fp16 12
-43 gpu conv fp16 12 add fp16 1
-44 gpu batchnorm fp16 12
-45 gpu add fp16 12
-46 gpu relu fp16 12
-47 gpu conv fp16 12 add fp16 1
-48 gpu batchnorm fp16 12
-49 gpu relu fp16 12
-50 gpu conv fp16 12 add fp16 1
-51 gpu batchnorm fp16 12
-52 gpu relu fp16 12
-53 gpu conv perf_fp16 153 add fp16 1
-54 gpu batchnorm fp16 12
-55 gpu add fp16 12
-56 gpu relu fp16 12
-57 gpu conv fp16 12 add fp16 1
-58 gpu batchnorm fp16 12
-59 gpu relu fp16 12
-60 gpu conv fp16 12 add fp16 1
-61 gpu batchnorm fp16 12
-62 gpu relu fp16 12
-63 gpu conv fp16 12 add fp16 1
-64 gpu batchnorm fp16 12
-65 gpu add fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12 add fp16 1
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu conv fp16 12 add fp16 1
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv fp16 12 add fp16 1
-74 gpu batchnorm fp16 12
-75 gpu add fp16 12
-76 gpu relu fp16 12
-77 gpu conv fp16 12 add fp16 1
-78 gpu batchnorm fp16 12
-79 gpu relu fp16 12
-80 gpu conv fp16 12 add fp16 1
-81 gpu batchnorm fp16 12
-82 gpu relu fp16 12
-83 gpu conv fp16 12 add fp16 1
-84 gpu batchnorm fp16 12
-85 gpu conv fp16 12 add fp16 1
-86 gpu batchnorm fp16 12
-87 gpu add fp16 12
-88 gpu relu fp16 12
-89 gpu conv fp16 12 add fp16 1
-90 gpu batchnorm fp16 12
-91 gpu relu fp16 12
-92 gpu conv fp16 12 add fp16 1
-93 gpu batchnorm fp16 12
-94 gpu relu fp16 12
-95 gpu conv fp16 11 add fp16 1
-96 gpu batchnorm fp16 12
-97 gpu add fp16 12
-98 gpu relu fp16 12
-99 gpu conv fp16 12 add fp16 1
-100 gpu batchnorm fp16 12
-101 gpu relu fp16 12
-102 gpu conv perf_fp16 164 add fp16 1
-103 gpu batchnorm fp16 12
-104 gpu relu fp16 12
-105 gpu conv fp16 12 add fp16 1
-106 gpu batchnorm fp16 12
-107 gpu add fp16 12
-108 gpu relu fp16 12
-109 gpu conv fp16 12 add fp16 1
-110 gpu batchnorm fp16 12
-111 gpu relu fp16 12
-112 gpu conv fp16 12 add fp16 1
-113 gpu batchnorm fp16 12
-114 gpu relu fp16 12
-115 gpu conv fp16 12 add fp16 1
-116 gpu batchnorm fp16 12
-117 gpu add fp16 12
-118 gpu relu fp16 12
-119 gpu conv samp_fp16 268 add fp16 1
-120 gpu batchnorm fp16 12
-121 gpu relu fp16 12
-122 gpu conv fp16 12 add fp16 1
-123 gpu batchnorm fp16 12
-124 gpu relu fp16 12
-125 gpu conv fp16 12 add fp16 1
-126 gpu batchnorm fp16 12
-127 gpu add fp16 12
-128 gpu relu fp16 12
-129 gpu conv fp16 12 add fp16 1
-130 gpu batchnorm fp16 12
-131 gpu relu fp16 12
-132 gpu conv fp16 12 add fp16 1
-133 gpu batchnorm fp16 12
-134 gpu relu fp16 12
-135 gpu conv fp16 12 add fp16 1
-136 gpu batchnorm fp16 12
-137 gpu add fp16 12
-138 gpu relu fp16 12
-139 gpu conv fp16 12 add fp16 1
-140 gpu batchnorm fp16 12
-141 gpu relu fp16 12
-142 gpu conv fp16 12 add fp16 1
-143 gpu batchnorm fp16 12
-144 gpu relu fp16 12
-145 gpu conv fp16 12 add fp16 1
-146 gpu batchnorm fp16 12
-147 gpu conv fp16 12 add fp16 1
-148 gpu batchnorm fp16 12
-149 gpu add fp16 12
-150 gpu relu fp16 12
-151 gpu conv fp16 12 add fp16 1
-152 gpu batchnorm fp16 12
-153 gpu relu fp16 12
-154 gpu conv fp16 12 add fp16 1
-155 gpu batchnorm fp16 12
-156 gpu relu fp16 12
-157 gpu conv fp16 12 add fp16 1
-158 gpu batchnorm fp16 12
-159 gpu add fp16 12
-160 gpu relu fp16 12
-161 gpu conv fp16 12 add fp16 1
-162 gpu batchnorm fp16 12
-163 gpu relu fp16 12
-164 gpu conv fp16 12 add fp16 1
-165 gpu batchnorm fp16 12
-166 gpu relu fp16 12
-167 gpu conv fp16 12 add fp16 1
-168 gpu batchnorm fp16 12
-169 gpu add fp16 12
-170 gpu relu fp16 12
-171 gpu pool_max fp16 12
-172 gpu mul fp16 12 add fp16 1
-173 gpu softmax fp32 1
------
-+++++
-conf7 1.8577902325643394 1.478552049679054 72.82 2.8800000000000097
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu batchnorm fp16 12
-3 gpu conv fp16 12 add fp16 1
-4 gpu batchnorm fp16 12
-5 gpu relu fp16 12
-6 gpu conv fp16 12 add fp16 1
-7 gpu batchnorm fp16 12
-8 gpu relu fp16 12
-9 gpu conv fp16 12 add fp16 1
-10 gpu batchnorm fp16 12
-11 gpu conv fp16 12 add fp16 1
-12 gpu batchnorm fp16 12
-13 gpu add fp16 12
-14 gpu relu fp16 12
-15 gpu conv fp16 12 add fp16 1
-16 gpu batchnorm fp16 12
-17 gpu relu fp16 12
-18 gpu conv fp16 12 add fp16 1
-19 gpu batchnorm fp16 12
-20 gpu relu fp16 12
-21 gpu conv fp16 12 add fp16 1
-22 gpu batchnorm fp16 12
-23 gpu add fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12 add fp16 1
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu conv perf_fp16 160 add fp16 1
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 11 add fp16 1
-32 gpu batchnorm fp16 12
-33 gpu add fp16 12
-34 gpu relu fp16 12
-35 gpu conv fp16 12 add fp16 1
-36 gpu batchnorm fp16 12
-37 gpu relu fp16 12
-38 gpu conv fp16 12 add fp16 1
-39 gpu batchnorm fp16 12
-40 gpu relu fp16 12
-41 gpu conv fp16 12 add fp16 1
-42 gpu batchnorm fp16 12
-43 gpu conv fp16 12 add fp16 1
-44 gpu batchnorm fp16 12
-45 gpu add fp16 12
-46 gpu relu fp16 12
-47 gpu conv fp16 12 add fp16 1
-48 gpu batchnorm fp16 12
-49 gpu relu fp16 12
-50 gpu conv fp16 12 add fp16 1
-51 gpu batchnorm fp16 12
-52 gpu relu fp16 12
-53 gpu conv fp16 11 add fp16 1
-54 gpu batchnorm fp16 12
-55 gpu add fp16 12
-56 gpu relu fp16 12
-57 gpu conv samp_fp16 268 add fp16 1
-58 gpu batchnorm fp16 12
-59 gpu relu fp16 12
-60 gpu conv fp16 12 add fp16 1
-61 gpu batchnorm fp16 12
-62 gpu relu fp16 12
-63 gpu conv fp16 12 add fp16 1
-64 gpu batchnorm fp16 12
-65 gpu add fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12 add fp16 1
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu conv fp16 12 add fp16 1
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv fp16 12 add fp16 1
-74 gpu batchnorm fp16 12
-75 gpu add fp16 12
-76 gpu relu fp16 12
-77 gpu conv fp16 12 add fp16 1
-78 gpu batchnorm fp16 12
-79 gpu relu fp16 12
-80 gpu conv fp16 12 add fp16 1
-81 gpu batchnorm fp16 12
-82 gpu relu fp16 12
-83 gpu conv fp16 12 add fp16 1
-84 gpu batchnorm fp16 12
-85 gpu conv fp16 12 add fp16 1
-86 gpu batchnorm fp16 12
-87 gpu add fp16 12
-88 gpu relu fp16 12
-89 gpu conv fp16 12 add fp16 1
-90 gpu batchnorm fp16 12
-91 gpu relu fp16 12
-92 gpu conv fp16 12 add fp16 1
-93 gpu batchnorm fp16 12
-94 gpu relu fp16 12
-95 gpu conv fp16 11 add fp16 1
-96 gpu batchnorm fp16 12
-97 gpu add fp16 12
-98 gpu relu fp16 12
-99 gpu conv fp16 12 add fp16 1
-100 gpu batchnorm fp16 12
-101 gpu relu fp16 12
-102 gpu conv perf_fp16 164 add fp16 1
-103 gpu batchnorm fp16 12
-104 gpu relu fp16 12
-105 gpu conv fp16 12 add fp16 1
-106 gpu batchnorm fp16 12
-107 gpu add fp16 12
-108 gpu relu fp16 12
-109 gpu conv fp16 12 add fp16 1
-110 gpu batchnorm fp16 12
-111 gpu relu fp16 12
-112 gpu conv fp16 12 add fp16 1
-113 gpu batchnorm fp16 12
-114 gpu relu fp16 12
-115 gpu conv fp16 12 add fp16 1
-116 gpu batchnorm fp16 12
-117 gpu add fp16 12
-118 gpu relu fp16 12
-119 gpu conv samp_fp16 268 add fp16 1
-120 gpu batchnorm fp16 12
-121 gpu relu fp16 12
-122 gpu conv fp16 12 add fp16 1
-123 gpu batchnorm fp16 12
-124 gpu relu fp16 12
-125 gpu conv fp16 12 add fp16 1
-126 gpu batchnorm fp16 12
-127 gpu add fp16 12
-128 gpu relu fp16 12
-129 gpu conv fp16 12 add fp16 1
-130 gpu batchnorm fp16 12
-131 gpu relu fp16 12
-132 gpu conv fp16 12 add fp16 1
-133 gpu batchnorm fp16 12
-134 gpu relu fp16 12
-135 gpu conv perf_fp16 158 add fp16 1
-136 gpu batchnorm fp16 12
-137 gpu add fp16 12
-138 gpu relu fp16 12
-139 gpu conv fp16 12 add fp16 1
-140 gpu batchnorm fp16 12
-141 gpu relu fp16 12
-142 gpu conv fp16 12 add fp16 1
-143 gpu batchnorm fp16 12
-144 gpu relu fp16 12
-145 gpu conv fp16 12 add fp16 1
-146 gpu batchnorm fp16 12
-147 gpu conv fp16 12 add fp16 1
-148 gpu batchnorm fp16 12
-149 gpu add fp16 12
-150 gpu relu fp16 12
-151 gpu conv fp16 12 add fp16 1
-152 gpu batchnorm fp16 12
-153 gpu relu fp16 12
-154 gpu conv fp16 12 add fp16 1
-155 gpu batchnorm fp16 12
-156 gpu relu fp16 12
-157 gpu conv fp16 11 add fp16 1
-158 gpu batchnorm fp16 12
-159 gpu add fp16 12
-160 gpu relu fp16 12
-161 gpu conv fp16 12 add fp16 1
-162 gpu batchnorm fp16 12
-163 gpu relu fp16 12
-164 gpu conv fp16 12 add fp16 1
-165 gpu batchnorm fp16 12
-166 gpu relu fp16 12
-167 gpu conv fp16 12 add fp16 1
-168 gpu batchnorm fp16 12
-169 gpu add fp16 12
-170 gpu relu fp16 12
-171 gpu pool_max fp16 12
-172 gpu mul fp16 12 add fp16 1
-173 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/vgg16_cifar10/vgg16_cifar10.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/vgg16_cifar10/vgg16_cifar10.txt
deleted file mode 100644
index f4e185f358..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/vgg16_cifar10/vgg16_cifar10.txt
+++ /dev/null
@@ -1,58 +0,0 @@
-3776.508929999999
-+++++
-conf1 1 1 89.96 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 relu fp32 1
-4 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-5 gpu conv fp32 11 add fp32 1 relu fp32 1
-6 gpu conv fp32 11 add fp32 1 relu fp32 1
-7 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-8 gpu conv fp32 11 add fp32 1 relu fp32 1
-9 gpu conv fp32 11 add fp32 1 relu fp32 1
-10 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-11 gpu conv fp32 11 add fp32 1 relu fp32 1
-12 gpu conv fp32 11 add fp32 1 relu fp32 1
-13 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-14 gpu mul fp32 11 add fp32 1 relu fp32 1
-15 gpu mul fp32 11 add fp32 1
-16 gpu softmax fp32 1
------
-+++++
-conf2 2.4192803184847484 2.2393153800931898 89.22 0.7399999999999949
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf3 2.1240075032467187 1.9749367321301132 88.64 1.3199999999999932
-1 gpu conv fp16 11 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 167 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv fp16 11 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-12 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 269 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/vgg16_cifar100/vgg16_cifar100.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/vgg16_cifar100/vgg16_cifar100.txt
deleted file mode 100644
index b55bb668b1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/vgg16_cifar100/vgg16_cifar100.txt
+++ /dev/null
@@ -1,77 +0,0 @@
-3768.819777999999
-+++++
-conf1 1 1 66.5 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 relu fp32 1
-4 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-5 gpu conv fp32 11 add fp32 1 relu fp32 1
-6 gpu conv fp32 11 add fp32 1 relu fp32 1
-7 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-8 gpu conv fp32 11 add fp32 1 relu fp32 1
-9 gpu conv fp32 11 add fp32 1 relu fp32 1
-10 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-11 gpu conv fp32 11 add fp32 1 relu fp32 1
-12 gpu conv fp32 11 add fp32 1 relu fp32 1
-13 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-14 gpu mul fp32 11 add fp32 1 relu fp32 1
-15 gpu mul fp32 11 add fp32 1
-16 gpu softmax fp32 1
------
-+++++
-conf2 2.2793321208062913 2.0502797911533945 66.42 0.0799999999999983
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 269 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 268 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 267 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 268 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-12 gpu conv fp16 11 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf3 2.2793321208062913 2.0502797911533945 66.42 0.0799999999999983
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 269 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 268 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 267 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 268 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-12 gpu conv fp16 11 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf4 2.664296720624579 2.427276363573644 64.7 1.7999999999999972
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/vgg16_imagenet/vgg16_imagenet.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/vgg16_imagenet/vgg16_imagenet.txt
deleted file mode 100644
index d0a23ffb10..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/empirical/vgg16_imagenet/vgg16_imagenet.txt
+++ /dev/null
@@ -1,41 +0,0 @@
-19194.623482
-+++++
-conf1 1 1 72.84 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 relu fp32 1
-4 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-5 gpu conv fp32 11 add fp32 1 relu fp32 1
-6 gpu conv fp32 11 add fp32 1 relu fp32 1
-7 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-8 gpu conv fp32 11 add fp32 1 relu fp32 1
-9 gpu conv fp32 11 add fp32 1 relu fp32 1
-10 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-11 gpu conv fp32 11 add fp32 1 relu fp32 1
-12 gpu conv fp32 11 add fp32 1 relu fp32 1
-13 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-14 gpu mul fp32 11 add fp32 1 relu fp32 1
-15 gpu mul fp32 11 add fp32 1 relu fp32 1
-16 gpu mul fp32 11 add fp32 1
-17 gpu softmax fp32 1
------
-+++++
-conf2 1.7719381411481732 1.5850925672384186 72.84 0.0
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv fp16 12 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/alexnet2_cifar10/alexnet2_cifar10.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/alexnet2_cifar10/alexnet2_cifar10.txt
deleted file mode 100644
index 6ec4a06d3d..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/alexnet2_cifar10/alexnet2_cifar10.txt
+++ /dev/null
@@ -1,419 +0,0 @@
-1114.3009809999999
-+++++
-conf1 1 1 84.98 0.0
-1 gpu conv fp32 11 add fp32 1 tanh fp32 1
-2 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 tanh fp32 1
-4 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-5 gpu conv fp32 11 add fp32 1 tanh fp32 1
-6 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-7 gpu mul fp32 11 add fp32 1
-8 gpu softmax fp32 1
------
-+++++
-conf2 2.4248748377353113 2.0815908534183163 84.5 0.480000000000004
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf3 2.4055188425519614 2.0586265720811823 84.48 0.5
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 269 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf4 2.4156140842962985 2.0617867479342706 84.28 0.7000000000000028
-1 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 163 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf5 2.396416918342732 2.0506214971794585 84.02 0.960000000000008
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 151 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf6 2.463002582910052 2.1171077568609458 83.84 1.1400000000000006
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 167 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf7 2.360283215266004 2.0255245321874304 83.78 1.2000000000000028
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf8 2.4140791541736157 2.0671513522247653 83.74000000000001 1.2399999999999949
-1 gpu conv fp16 11 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 160 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf9 2.457753689612079 2.1086250651240137 83.7 1.2800000000000011
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 163 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf10 2.459170454055443 2.1111925341396343 83.7 1.2800000000000011
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 164 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf11 2.4135986141645764 2.060453960420927 83.62 1.3599999999999994
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf12 2.4631278039012106 2.1092094797926637 83.58 1.4000000000000057
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf13 2.535761391794481 2.16998336112692 83.58 1.4000000000000057
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf14 2.289006193945062 1.961240158652051 83.54 1.4399999999999977
-1 gpu conv perf_fp16 167 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 155 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf15 2.4257674844112573 2.0808440756495563 83.5 1.480000000000004
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 161 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf16 2.458122368488622 2.109531159729078 83.48 1.5
-1 gpu conv fp16 11 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 162 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf17 2.281072202152105 1.9539314420536427 83.46000000000001 1.519999999999996
-1 gpu conv fp16 11 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 160 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 160 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf18 2.4572171342078444 2.1088933553775697 83.46000000000001 1.519999999999996
-1 gpu conv fp16 11 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 163 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf19 2.3017607719030058 1.9782265708150768 83.42 1.5600000000000023
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 162 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 162 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf20 2.379206814483014 2.047909200292713 83.39999999999999 1.5800000000000125
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 151 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf21 2.4636282705302537 2.1162281156388527 83.39999999999999 1.5800000000000125
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 160 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf22 2.461590101374146 2.1108493881199184 83.22 1.7600000000000051
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 161 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf23 2.537054645442804 2.167568834938183 83.22 1.7600000000000051
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf24 2.4631604723407885 2.1099694757102845 83.17999999999999 1.8000000000000114
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf25 2.4636282705302537 2.1162281156388527 83.14 1.8400000000000034
-1 gpu conv fp16 11 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 160 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf26 2.462588899729088 2.109477918791931 83.14 1.8400000000000034
-1 gpu conv fp16 11 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf27 2.4638085754689025 2.1071960926343603 83.1 1.8800000000000097
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf28 2.4640079766123635 2.110326453157297 83.08 1.9000000000000057
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf29 2.459337622764853 2.107249218450713 83.06 1.9200000000000017
-1 gpu conv fp16 11 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 162 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf30 2.538176340059405 2.173287257415721 83.02000000000001 1.9599999999999937
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 164 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf31 2.3905426931959846 2.044333576277581 83.02000000000001 1.9599999999999937
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 160 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf32 2.459337622764853 2.107249218450713 83.0 1.980000000000004
-1 gpu conv fp16 11 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 162 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf33 2.458968579288317 2.1063450826631396 82.89999999999999 2.0800000000000125
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 163 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf34 2.2912974651603877 1.9670210508860688 82.8 2.180000000000007
-1 gpu conv perf_fp16 168 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 155 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 160 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf35 2.4648489763056327 2.113931670664391 82.66 2.3200000000000074
-1 gpu conv fp16 11 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 160 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf36 2.4599076869402854 2.1077397371200193 82.6 2.3800000000000097
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 162 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf37 2.4636282705302537 2.1162281156388527 82.54 2.4399999999999977
-1 gpu conv fp16 11 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 160 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
-+++++
-conf38 2.591814267389778 2.222680944458784 82.26 2.719999999999999
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1
-2 gpu conv perf_fp16 154 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1
-6 gpu conv perf_fp16 157 add fp16 1 tanh fp16 1 pool_max fp16 1
-7 gpu mul fp16 12 add fp16 1
-8 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/alexnet_cifar10/alexnet_cifar10.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/alexnet_cifar10/alexnet_cifar10.txt
deleted file mode 100644
index a9ccba6eb6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/alexnet_cifar10/alexnet_cifar10.txt
+++ /dev/null
@@ -1,511 +0,0 @@
-2592.187221
-+++++
-conf1 1 1 79.28 0.0
-1 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-2 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 tanh fp32 1
-4 gpu conv fp32 11 add fp32 1 tanh fp32 1
-5 gpu conv fp32 11 add fp32 1 tanh fp32 1 pool_max fp32 1
-6 gpu mul fp32 11 add fp32 1
-7 gpu softmax fp32 1
------
-+++++
-conf2 1.7593976485873195 1.6193399031642917 79.23 0.04999999999999716
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf3 2.092625440752526 1.9139078015388271 78.96 0.3200000000000074
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf4 1.8870195448805414 1.7296919053025768 78.8 0.480000000000004
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf5 2.1184804041774554 1.9598989563949536 78.75999999999999 0.5200000000000102
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf6 2.1184804041774554 1.9598989563949536 78.75999999999999 0.5200000000000102
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf7 2.0933825381386364 1.9150743378318535 78.64 0.6400000000000006
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf8 2.081712090729918 1.9102226906341664 78.5 0.7800000000000011
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf9 2.081712090729918 1.9102226906341664 78.5 0.7800000000000011
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf10 2.2662606588487595 2.066560750795139 78.48 0.7999999999999972
-1 gpu conv samp_fp16 264 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf11 2.121684761285686 1.966318179285323 78.48 0.7999999999999972
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf12 2.3417491169395532 2.1355030360671465 78.38000000000001 0.8999999999999915
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf13 2.2247938983110425 2.060416584958474 78.38000000000001 0.8999999999999915
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf14 2.2247938983110425 2.060416584958474 78.38000000000001 0.8999999999999915
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf15 2.2247938983110425 2.060416584958474 78.38000000000001 0.8999999999999915
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf16 2.2627828537139263 2.065683616898884 78.32000000000001 0.9599999999999937
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf17 2.2627828537139263 2.065683616898884 78.32000000000001 0.9599999999999937
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf18 2.2627828537139263 2.065683616898884 78.32000000000001 0.9599999999999937
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf19 2.146571989407323 1.95711703610764 78.18 1.0999999999999943
-1 gpu conv samp_fp16 264 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf20 2.303316973793268 2.1036463961913276 78.10000000000001 1.1799999999999926
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf21 2.436875653706139 2.2434837737118056 78.08 1.2000000000000028
-1 gpu conv samp_fp16 264 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf22 2.436875653706139 2.2434837737118056 78.08 1.2000000000000028
-1 gpu conv samp_fp16 264 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf23 2.436875653706139 2.2434837737118056 78.08 1.2000000000000028
-1 gpu conv samp_fp16 264 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf24 2.1106508925330925 1.9419233584234938 78.06 1.2199999999999989
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf25 2.3203534290038634 2.116965679235447 78.06 1.2199999999999989
-1 gpu conv samp_fp16 264 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf26 2.3527290658539215 2.145832257234814 78.03999999999999 1.240000000000009
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf27 2.3527290658539215 2.145832257234814 78.03999999999999 1.240000000000009
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv fp16 12 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf28 2.432854949808342 2.2424500615508003 78.0 1.2800000000000011
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf29 2.432854949808342 2.2424500615508003 78.0 1.2800000000000011
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf30 2.432854949808342 2.2424500615508003 78.0 1.2800000000000011
-1 gpu conv samp_fp16 263 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf31 2.3137982135449207 2.1281257317083417 77.84 1.4399999999999977
-1 gpu conv samp_fp16 264 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 265 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf32 2.1198074418988333 1.9522214255218437 77.82 1.460000000000008
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf33 2.246924974355375 2.065289762405701 77.8 1.480000000000004
-1 gpu conv samp_fp16 264 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 269 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf34 2.263614734554485 2.090777846534249 77.74 1.5400000000000063
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf35 2.263614734554485 2.090777846534249 77.74 1.5400000000000063
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf36 2.263614734554485 2.090777846534249 77.74 1.5400000000000063
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf37 2.5289288699015304 2.334007588396142 77.72 1.5600000000000023
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf38 2.5289288699015304 2.334007588396142 77.72 1.5600000000000023
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf39 2.3117594882585775 2.1152397180868943 77.56 1.7199999999999989
-1 gpu conv samp_fp16 264 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf40 2.452732477854469 2.264573687601476 77.56 1.7199999999999989
-1 gpu conv perf_fp16 167 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf41 2.452732477854469 2.264573687601476 77.56 1.7199999999999989
-1 gpu conv perf_fp16 167 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf42 2.382518688546389 2.178614303992064 77.5 1.7800000000000011
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf43 2.382518688546389 2.178614303992064 77.5 1.7800000000000011
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf44 2.3900667100485924 2.188128526401265 77.48 1.7999999999999972
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf45 2.3900667100485924 2.188128526401265 77.48 1.7999999999999972
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf46 2.3900667100485924 2.188128526401265 77.48 1.7999999999999972
-1 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf47 2.4835281673276515 2.279527076032239 77.3 1.980000000000004
-1 gpu conv samp_fp16 264 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf48 2.4835281673276515 2.279527076032239 77.3 1.980000000000004
-1 gpu conv samp_fp16 264 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf49 2.1553694968551302 1.9959124044028933 77.18 2.0999999999999943
-1 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 265 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 268 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf50 2.5877520959724816 2.3763616521050364 77.03999999999999 2.240000000000009
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
-+++++
-conf51 2.5877520959724816 2.3763616521050364 77.03999999999999 2.240000000000009
-1 gpu conv samp_fp16 261 add fp16 1 tanh fp16 1 pool_max fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 267 add fp16 1 tanh fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 tanh fp16 1
-5 gpu conv fp16 12 add fp16 1 tanh fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1
-7 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/alexnet_imagenet/alexnet_imagenet.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/alexnet_imagenet/alexnet_imagenet.txt
deleted file mode 100644
index b0e42a5aaa..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/alexnet_imagenet/alexnet_imagenet.txt
+++ /dev/null
@@ -1,229 +0,0 @@
-2739.950736
-+++++
-conf1 1 1 56.3 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 relu fp32 1
-4 gpu conv fp32 11 add fp32 1 relu fp32 1
-5 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-6 gpu mul fp32 11 add fp32 1 relu fp32 1
-7 gpu mul fp32 11 add fp32 1 relu fp32 1
-8 gpu mul fp32 11 add fp32 1
-9 gpu softmax fp32 1
------
-+++++
-conf2 1.802133644103582 1.8186433204507424 55.76 0.5399999999999991
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf3 1.7574572103878898 1.7673706184460103 55.58 0.7199999999999989
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1
-5 gpu conv samp_fp16 268 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf4 2.0227701930718065 2.043112495268932 55.42 0.8799999999999955
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 166 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf5 1.9872634777043927 2.002789650227035 55.120000000000005 1.1799999999999926
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 164 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf6 1.8204253918445088 1.843736069756362 54.84 1.4599999999999937
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 154 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf7 1.9308336510645352 1.934889049414224 54.74 1.5599999999999952
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 168 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf8 2.0146435217865446 2.0367475358800102 54.58 1.7199999999999989
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf9 2.0101709494490696 2.0329911158023064 54.400000000000006 1.8999999999999915
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 164 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf10 2.0052132441967916 2.0284931705407003 54.300000000000004 1.999999999999993
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 168 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf11 2.010827434817262 2.036001862538864 54.2 2.0999999999999943
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 164 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 154 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf12 2.019868378233057 2.0433540129730265 54.17999999999999 2.1200000000000045
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf13 1.9923471030291253 2.009177323959059 54.120000000000005 2.1799999999999926
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 164 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf14 1.9923471030291253 2.009177323959059 54.120000000000005 2.1799999999999926
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 164 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf15 2.028037341700216 2.049760395549724 54.0 2.299999999999997
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 166 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf16 1.9910730364852436 2.006510848093771 53.54 2.759999999999998
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 164 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf17 2.1567475543719614 2.159142310265706 53.300000000000004 2.999999999999993
-1 gpu conv perf_fp16 164 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 166 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf18 2.1567475543719614 2.159142310265706 53.300000000000004 2.999999999999993
-1 gpu conv perf_fp16 164 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 166 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
-+++++
-conf19 2.0232690820426464 2.0527698121318476 53.300000000000004 2.999999999999993
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu conv perf_fp16 168 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 11 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-5 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-6 gpu mul fp16 12 add fp16 1 relu fp16 1
-7 gpu mul fp16 12 add fp16 1 relu fp16 1
-8 gpu mul fp16 12 add fp16 1
-9 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/lenet_keras/lenet_keras.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/lenet_keras/lenet_keras.txt
deleted file mode 100644
index b4e51dff42..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/lenet_keras/lenet_keras.txt
+++ /dev/null
@@ -1,409 +0,0 @@
-282.5141369999999
-+++++
-conf1 1 1 98.7 0.0
-1 gpu conv fp32 11 add fp32 1 pool_max fp32 1 tanh fp32 1
-2 gpu conv fp32 11 add fp32 1 pool_max fp32 1 tanh fp32 1
-3 gpu mul fp32 11 add fp32 1 tanh fp32 1
-4 gpu mul fp32 11 add fp32 1 tanh fp32 1
-5 gpu softmax fp32 1
------
-+++++
-conf2 1.828613181003043 2.071721708828981 98.65 0.04999999999999716
-1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf3 1.8936889628815377 2.139779619692146 98.65 0.04999999999999716
-1 gpu conv perf_fp16 152 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf4 1.8936889628815377 2.139779619692146 98.65 0.04999999999999716
-1 gpu conv perf_fp16 152 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf5 1.8936889628815377 2.139779619692146 98.65 0.04999999999999716
-1 gpu conv perf_fp16 152 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf6 1.8247639611533713 2.0227145446958756 98.64 0.060000000000002274
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf7 1.8247639611533713 2.0227145446958756 98.64 0.060000000000002274
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf8 1.8406161850501603 2.037849502542524 98.64 0.060000000000002274
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf9 1.8406161850501603 2.037849502542524 98.64 0.060000000000002274
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf10 1.8406161850501603 2.037849502542524 98.64 0.060000000000002274
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf11 1.8663357888260776 2.115790921611576 98.64 0.060000000000002274
-1 gpu conv perf_fp16 155 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf12 1.8663357888260776 2.115790921611576 98.64 0.060000000000002274
-1 gpu conv perf_fp16 155 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf13 1.8663357888260776 2.115790921611576 98.64 0.060000000000002274
-1 gpu conv perf_fp16 155 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf14 1.8645645142051612 2.1037012333044935 98.61999999999999 0.0800000000000125
-1 gpu conv perf_fp16 167 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf15 1.8645645142051612 2.1037012333044935 98.61999999999999 0.0800000000000125
-1 gpu conv perf_fp16 167 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf16 1.8645645142051612 2.1037012333044935 98.61999999999999 0.0800000000000125
-1 gpu conv perf_fp16 167 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf17 2.2168527051833635 2.453341076720038 98.61999999999999 0.0800000000000125
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf18 2.2168527051833635 2.453341076720038 98.61999999999999 0.0800000000000125
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf19 2.2168527051833635 2.453341076720038 98.61999999999999 0.0800000000000125
-1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf20 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
-1 gpu conv fp16 12 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf21 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
-1 gpu conv fp16 12 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf22 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
-1 gpu conv fp16 12 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf23 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf24 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf25 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf26 2.200653361151419 2.425091789360736 98.6 0.10000000000000853
-1 gpu conv samp_fp16 266 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf27 2.200653361151419 2.425091789360736 98.6 0.10000000000000853
-1 gpu conv samp_fp16 266 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf28 1.8406161850501603 2.037849502542524 98.58 0.12000000000000455
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf29 1.8406161850501603 2.037849502542524 98.58 0.12000000000000455
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf30 1.8406161850501603 2.037849502542524 98.58 0.12000000000000455
-1 gpu conv fp16 11 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf31 1.8445326456180258 2.087601822059355 98.58 0.12000000000000455
-1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf32 1.8445326456180258 2.087601822059355 98.58 0.12000000000000455
-1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf33 1.8445326456180258 2.087601822059355 98.58 0.12000000000000455
-1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf34 1.8916677984300285 2.155437579874673 98.58 0.12000000000000455
-1 gpu conv perf_fp16 158 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf35 1.8916677984300285 2.155437579874673 98.58 0.12000000000000455
-1 gpu conv perf_fp16 158 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf36 1.8916677984300285 2.155437579874673 98.58 0.12000000000000455
-1 gpu conv perf_fp16 158 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf37 1.8649226857257986 2.1076025277601325 98.56 0.14000000000000057
-1 gpu conv perf_fp16 168 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf38 1.8649226857257986 2.1076025277601325 98.56 0.14000000000000057
-1 gpu conv perf_fp16 168 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf39 1.8649226857257986 2.1076025277601325 98.56 0.14000000000000057
-1 gpu conv perf_fp16 168 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf40 1.8463058650555446 2.067271423078985 98.56 0.14000000000000057
-1 gpu conv perf_fp16 157 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf41 1.8463058650555446 2.067271423078985 98.56 0.14000000000000057
-1 gpu conv perf_fp16 157 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf42 1.8463058650555446 2.067271423078985 98.56 0.14000000000000057
-1 gpu conv perf_fp16 157 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf43 1.9234076467497994 2.1864740913112275 98.56 0.14000000000000057
-1 gpu conv perf_fp16 153 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf44 1.9234076467497994 2.1864740913112275 98.56 0.14000000000000057
-1 gpu conv perf_fp16 153 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf45 1.9234076467497994 2.1864740913112275 98.56 0.14000000000000057
-1 gpu conv perf_fp16 153 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf46 1.8698191484268973 2.13979218727595 98.54 0.1599999999999966
-1 gpu conv perf_fp16 159 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf47 1.8698191484268973 2.13979218727595 98.54 0.1599999999999966
-1 gpu conv perf_fp16 159 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf48 1.8575043605938137 2.092057786757256 98.52 0.18000000000000682
-1 gpu conv perf_fp16 165 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf49 1.8575043605938137 2.092057786757256 98.52 0.18000000000000682
-1 gpu conv perf_fp16 165 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf50 1.8575043605938137 2.092057786757256 98.52 0.18000000000000682
-1 gpu conv perf_fp16 165 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
-+++++
-conf51 1.8534621507951072 2.1231113105788597 98.44000000000001 0.2599999999999909
-1 gpu conv perf_fp16 159 add fp16 1 pool_max fp16 1 tanh fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 pool_max fp16 1 tanh fp16 1
-3 gpu mul fp16 12 add fp16 1 tanh fp16 1
-4 gpu mul fp16 12 add fp16 1 tanh fp16 1
-5 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/mobilenet_cifar10/mobilenet_cifar10.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/mobilenet_cifar10/mobilenet_cifar10.txt
deleted file mode 100644
index b4d8bd893c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/mobilenet_cifar10/mobilenet_cifar10.txt
+++ /dev/null
@@ -1,3220 +0,0 @@
-4077.307063200001
-+++++
-conf1 1 1 84.42 0.0
-1 gpu conv fp32 11
-2 gpu batchnorm fp32 11
-3 gpu relu fp32 11
-4 gpu group_conv fp32 11
-5 gpu batchnorm fp32 11
-6 gpu relu fp32 11
-7 gpu conv fp32 11
-8 gpu batchnorm fp32 11
-9 gpu relu fp32 11
-10 gpu group_conv fp32 11
-11 gpu batchnorm fp32 11
-12 gpu relu fp32 11
-13 gpu conv fp32 11
-14 gpu batchnorm fp32 11
-15 gpu relu fp32 11
-16 gpu group_conv fp32 11
-17 gpu batchnorm fp32 11
-18 gpu relu fp32 11
-19 gpu conv fp32 11
-20 gpu batchnorm fp32 11
-21 gpu relu fp32 11
-22 gpu group_conv fp32 11
-23 gpu batchnorm fp32 11
-24 gpu relu fp32 11
-25 gpu conv fp32 11
-26 gpu batchnorm fp32 11
-27 gpu relu fp32 11
-28 gpu group_conv fp32 11
-29 gpu batchnorm fp32 11
-30 gpu relu fp32 11
-31 gpu conv fp32 11
-32 gpu batchnorm fp32 11
-33 gpu relu fp32 11
-34 gpu group_conv fp32 11
-35 gpu batchnorm fp32 11
-36 gpu relu fp32 11
-37 gpu conv fp32 11
-38 gpu batchnorm fp32 11
-39 gpu relu fp32 11
-40 gpu group_conv fp32 11
-41 gpu batchnorm fp32 11
-42 gpu relu fp32 11
-43 gpu conv fp32 11
-44 gpu batchnorm fp32 11
-45 gpu relu fp32 11
-46 gpu group_conv fp32 11
-47 gpu batchnorm fp32 11
-48 gpu relu fp32 11
-49 gpu conv fp32 11
-50 gpu batchnorm fp32 11
-51 gpu relu fp32 11
-52 gpu group_conv fp32 11
-53 gpu batchnorm fp32 11
-54 gpu relu fp32 11
-55 gpu conv fp32 11
-56 gpu batchnorm fp32 11
-57 gpu relu fp32 11
-58 gpu group_conv fp32 11
-59 gpu batchnorm fp32 11
-60 gpu relu fp32 11
-61 gpu conv fp32 11
-62 gpu batchnorm fp32 11
-63 gpu relu fp32 11
-64 gpu group_conv fp32 11
-65 gpu batchnorm fp32 11
-66 gpu relu fp32 11
-67 gpu conv fp32 11
-68 gpu batchnorm fp32 11
-69 gpu relu fp32 11
-70 gpu group_conv fp32 11
-71 gpu batchnorm fp32 11
-72 gpu relu fp32 11
-73 gpu conv fp32 11
-74 gpu batchnorm fp32 11
-75 gpu relu fp32 11
-76 gpu group_conv fp32 11
-77 gpu batchnorm fp32 11
-78 gpu relu fp32 11
-79 gpu conv fp32 11
-80 gpu batchnorm fp32 11
-81 gpu relu fp32 11
-82 gpu pool_mean fp32 11
-83 gpu mul fp32 11 add fp32 1
-84 gpu softmax fp32 1
------
-+++++
-conf2 1.4930855091460031 1.447990050940341 83.72 0.7000000000000028
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv fp16 12
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf3 1.493397883226807 1.449591062426989 83.72 0.7000000000000028
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 163
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf4 1.4934429016801338 1.4500582352111675 83.72 0.7000000000000028
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 168
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf5 1.4938214813031556 1.450038222978811 83.72 0.7000000000000028
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 157
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf6 1.4933879828131855 1.449975636202813 83.72 0.7000000000000028
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 160
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf7 1.492663093331302 1.4487067754520524 83.7 0.7199999999999989
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 167
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf8 1.495724395088184 1.4507925552157772 83.56 0.8599999999999994
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 162
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf9 1.496506307637598 1.4521705950285135 83.36 1.0600000000000023
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 162
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf10 1.496532672928805 1.4521696542076958 83.36 1.0600000000000023
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 156
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf11 1.4988418058849937 1.4555327556053628 83.28 1.1400000000000006
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 168
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 164
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 158
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf12 1.4994289979945077 1.4562439330251535 83.28 1.1400000000000006
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 168
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 152
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 153
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf13 1.4952028793065038 1.450369851058777 83.14 1.2800000000000011
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 162
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 151
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 155
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 156
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf14 1.4933978285280285 1.448265686258097 83.12 1.2999999999999972
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 158
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf15 1.491958833559989 1.4459262032919467 83.08 1.3400000000000034
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 157
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 155
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf16 1.4937317297990984 1.4498121856525021 83.02000000000001 1.3999999999999915
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 156
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 158
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf17 1.4963413808686974 1.4522391736954623 82.86 1.5600000000000023
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 165
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf18 1.4942172827099065 1.4504631324933321 82.86 1.5600000000000023
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 157
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 158
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf19 1.4963964073376739 1.4525461321361477 82.86 1.5600000000000023
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 158
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf20 1.4932583049858652 1.4472547227714012 82.84 1.5799999999999983
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv samp_fp16 266
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf21 1.4964326545281064 1.4526263046333605 82.82000000000001 1.5999999999999943
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 152
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 158
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf22 1.4966042483929347 1.4527859961226985 82.82000000000001 1.5999999999999943
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 152
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 153
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf23 1.4966008974318024 1.4527415844509437 82.78 1.6400000000000006
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 155
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 158
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf24 1.4932738366973777 1.448820445466833 82.64 1.7800000000000011
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 164
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 151
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 155
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 156
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 157
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf25 1.4940402684133964 1.447332235394843 82.48 1.9399999999999977
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv samp_fp16 261
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf26 1.4981764588414919 1.4530714150549078 82.39999999999999 2.0200000000000102
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 151
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 161
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 156
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf27 1.5004334658773033 1.4549115105608688 82.3 2.1200000000000045
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 151
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 155
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 156
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 156
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf28 1.5006808163336343 1.4553824345285296 82.3 2.1200000000000045
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 151
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 155
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 156
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf29 1.4999870719460484 1.4571625511374704 82.28 2.1400000000000006
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 168
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 152
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 165
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf30 1.500042366879961 1.4574715946270216 82.28 2.1400000000000006
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 168
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 152
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 158
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf31 1.500214789632402 1.4576323532660131 82.28 2.1400000000000006
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 168
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 163
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 164
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 152
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 151
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 153
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf32 1.4927009086066445 1.4484049211953174 82.26 2.1599999999999966
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 164
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 151
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 161
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 156
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf33 1.5003438014588875 1.4538240352408085 82.22 2.200000000000003
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 151
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 155
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 152
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf34 1.5041587978616728 1.4610492456195174 82.02000000000001 2.3999999999999915
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 168
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 161
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 155
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 152
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 158
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf35 1.5000040131742656 1.4555601139156464 81.88 2.5400000000000063
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv fp16 12
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv perf_fp16 152
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 12
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv perf_fp16 161
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 151
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 155
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 151
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 167
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf36 1.4950571524902583 1.451478376045808 81.84 2.5799999999999983
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 164
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv perf_fp16 161
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 161
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 155
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 155
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 153
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
-+++++
-conf37 1.4975271575548847 1.4532126224638244 81.44 2.980000000000004
-1 gpu conv fp16 12
-2 gpu batchnorm fp16 12
-3 gpu relu fp16 12
-4 gpu group_conv fp16 12
-5 gpu batchnorm fp16 12
-6 gpu relu fp16 12
-7 gpu conv fp16 12
-8 gpu batchnorm fp16 12
-9 gpu relu fp16 12
-10 gpu group_conv fp16 12
-11 gpu batchnorm fp16 12
-12 gpu relu fp16 12
-13 gpu conv fp16 12
-14 gpu batchnorm fp16 12
-15 gpu relu fp16 12
-16 gpu group_conv fp16 12
-17 gpu batchnorm fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 164
-20 gpu batchnorm fp16 12
-21 gpu relu fp16 12
-22 gpu group_conv fp16 12
-23 gpu batchnorm fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu group_conv fp16 12
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12
-32 gpu batchnorm fp16 12
-33 gpu relu fp16 12
-34 gpu group_conv fp16 12
-35 gpu batchnorm fp16 12
-36 gpu relu fp16 12
-37 gpu conv fp16 11
-38 gpu batchnorm fp16 12
-39 gpu relu fp16 12
-40 gpu group_conv fp16 12
-41 gpu batchnorm fp16 12
-42 gpu relu fp16 12
-43 gpu conv fp16 12
-44 gpu batchnorm fp16 12
-45 gpu relu fp16 12
-46 gpu group_conv fp16 12
-47 gpu batchnorm fp16 12
-48 gpu relu fp16 12
-49 gpu conv perf_fp16 155
-50 gpu batchnorm fp16 12
-51 gpu relu fp16 12
-52 gpu group_conv fp16 12
-53 gpu batchnorm fp16 12
-54 gpu relu fp16 12
-55 gpu conv perf_fp16 155
-56 gpu batchnorm fp16 12
-57 gpu relu fp16 12
-58 gpu group_conv fp16 12
-59 gpu batchnorm fp16 12
-60 gpu relu fp16 12
-61 gpu conv perf_fp16 151
-62 gpu batchnorm fp16 12
-63 gpu relu fp16 12
-64 gpu group_conv fp16 12
-65 gpu batchnorm fp16 12
-66 gpu relu fp16 12
-67 gpu conv perf_fp16 155
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu group_conv fp16 12
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv perf_fp16 152
-74 gpu batchnorm fp16 12
-75 gpu relu fp16 12
-76 gpu group_conv fp16 12
-77 gpu batchnorm fp16 12
-78 gpu relu fp16 12
-79 gpu conv perf_fp16 153
-80 gpu batchnorm fp16 12
-81 gpu relu fp16 12
-82 gpu pool_mean fp16 12
-83 gpu mul fp16 12 add fp16 1
-84 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/resnet18_cifar10/resnet18_cifar10.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/resnet18_cifar10/resnet18_cifar10.txt
deleted file mode 100644
index 654cffbf63..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/resnet18_cifar10/resnet18_cifar10.txt
+++ /dev/null
@@ -1,2296 +0,0 @@
-2484.981244
-+++++
-conf1 1 1 89.56 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1
-3 gpu conv fp32 11 add fp32 1
-4 gpu add fp32 11
-5 gpu relu fp32 11
-6 gpu conv fp32 11 add fp32 1 relu fp32 1
-7 gpu conv fp32 11 add fp32 1
-8 gpu add fp32 11
-9 gpu relu fp32 11
-10 gpu conv fp32 11 add fp32 1 relu fp32 1
-11 gpu conv fp32 11 add fp32 1
-12 gpu add fp32 11
-13 gpu relu fp32 11
-14 gpu conv fp32 11 add fp32 1 relu fp32 1
-15 gpu conv fp32 11 add fp32 1
-16 gpu conv fp32 11 add fp32 1
-17 gpu add fp32 11
-18 gpu relu fp32 11
-19 gpu conv fp32 11 add fp32 1 relu fp32 1
-20 gpu conv fp32 11 add fp32 1
-21 gpu add fp32 11
-22 gpu relu fp32 11
-23 gpu conv fp32 11 add fp32 1 relu fp32 1
-24 gpu conv fp32 11 add fp32 1
-25 gpu add fp32 11
-26 gpu relu fp32 11
-27 gpu conv fp32 11 add fp32 1 relu fp32 1
-28 gpu conv fp32 11 add fp32 1
-29 gpu conv fp32 11 add fp32 1
-30 gpu add fp32 11
-31 gpu relu fp32 11
-32 gpu conv fp32 11 add fp32 1 relu fp32 1
-33 gpu conv fp32 11 add fp32 1
-34 gpu add fp32 11
-35 gpu relu fp32 11
-36 gpu conv fp32 11 add fp32 1 relu fp32 1
-37 gpu conv fp32 11 add fp32 1
-38 gpu add fp32 11
-39 gpu relu fp32 11
-40 gpu pool_mean fp32 11
-41 gpu mul fp32 11 add fp32 1
-42 gpu softmax fp32 1
------
-+++++
-conf2 1.767527790869615 1.7962938589450996 88.96 0.6000000000000085
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 162 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 167 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 167 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-20 gpu conv perf_fp16 155 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv fp16 12 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 160 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 155 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf3 1.7676486174436143 1.7967155014984917 88.78 0.7800000000000011
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 162 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 167 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 167 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv perf_fp16 160 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-20 gpu conv perf_fp16 155 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv fp16 12 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 160 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 155 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf4 1.7674352647250422 1.792910560846682 88.7 0.8599999999999994
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 162 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 167 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 167 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 168 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv fp16 12 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 160 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 155 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf5 1.8655703338511067 1.8930089896922888 88.53999999999999 1.0200000000000102
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 167 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 158 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv perf_fp16 159 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 165 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 157 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf6 1.9070428103729684 1.9172857853336078 88.38000000000001 1.1799999999999926
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 157 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 152 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv samp_fp16 266 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-20 gpu conv perf_fp16 152 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv samp_fp16 261 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 155 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf7 1.769778590701739 1.7956222622694236 88.24 1.3200000000000074
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv fp16 12 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv samp_fp16 268 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf8 1.841404652091802 1.8677947628418006 88.24 1.3200000000000074
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 168 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 162 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf9 1.8679349428783487 1.8995927920729931 88.22 1.3400000000000034
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 159 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 160 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 168 add fp16 1 relu fp16 1
-20 gpu conv perf_fp16 161 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf10 1.876937310100899 1.9041581451399825 88.1 1.460000000000008
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 158 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf11 1.842140004857965 1.8673692956620238 88.06 1.5
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 166 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 167 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf12 1.9070567138857761 1.9165525910492667 88.02 1.5400000000000063
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 157 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 152 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv samp_fp16 266 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 261 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 152 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 155 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf13 1.9185835698271805 1.9328202469403 87.98 1.5799999999999983
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 157 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 152 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv samp_fp16 266 add fp16 1
-16 gpu conv perf_fp16 160 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-20 gpu conv perf_fp16 152 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 152 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 155 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf14 1.781744853993609 1.8082995958456516 87.92 1.6400000000000006
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 166 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 168 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv perf_fp16 159 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 165 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv samp_fp16 265 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv samp_fp16 268 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf15 1.9185835698271805 1.9328202469403 87.92 1.6400000000000006
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 157 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 152 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv samp_fp16 266 add fp16 1
-16 gpu conv perf_fp16 160 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-20 gpu conv perf_fp16 152 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 152 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 12 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 155 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf16 1.875261840315855 1.8986912653657988 87.88 1.6800000000000068
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 159 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 156 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 12 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf17 1.9013559086026153 1.9230901214481015 87.86 1.7000000000000028
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf18 1.9185835698271805 1.9328202469403 87.83999999999999 1.720000000000013
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 157 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 152 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv samp_fp16 266 add fp16 1
-16 gpu conv perf_fp16 160 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-20 gpu conv perf_fp16 152 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 152 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 155 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf19 1.8770503055325798 1.9007923328014182 87.82 1.740000000000009
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 162 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 158 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 151 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf20 1.8774136276932418 1.90365663123621 87.82 1.740000000000009
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 158 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf21 1.943143041264842 1.9591958561422729 87.82 1.740000000000009
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf22 1.870789918969847 1.8863625217899933 87.8 1.7600000000000051
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 264 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf23 1.7445941809066292 1.7754934270309912 87.78 1.7800000000000011
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 162 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 167 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 167 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv perf_fp16 160 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-20 gpu conv perf_fp16 155 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv fp16 12 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 160 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv perf_fp16 166 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 155 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf24 1.9065930313550916 1.928938946228637 87.78 1.7800000000000011
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 167 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf25 1.9021824494907031 1.9237134505552098 87.78 1.7800000000000011
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 154 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf26 1.9017271009017505 1.9211078231701697 87.78 1.7800000000000011
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 162 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf27 1.8187224917656395 1.820406007609536 87.76 1.7999999999999972
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv samp_fp16 264 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf28 1.9070855899343322 1.9285210655709735 87.76 1.7999999999999972
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv samp_fp16 268 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf29 1.9013559086026153 1.9230901214481015 87.74 1.8200000000000074
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf30 1.8772990284718367 1.9022146647342513 87.72 1.8400000000000034
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 162 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 158 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf31 1.9013559086026153 1.9230901214481015 87.68 1.8799999999999955
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf32 1.9020502478364545 1.923319572598976 87.66000000000001 1.8999999999999915
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf33 1.7516394053514481 1.7809034526471939 87.62 1.9399999999999977
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 157 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 162 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 167 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 167 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv perf_fp16 160 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-20 gpu conv perf_fp16 155 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv fp16 12 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 160 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv perf_fp16 166 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 155 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf34 1.7814953252955337 1.8122658147993431 87.62 1.9399999999999977
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 162 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 167 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 167 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv perf_fp16 160 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-20 gpu conv perf_fp16 155 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv fp16 12 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 160 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv perf_fp16 166 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 155 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf35 1.887538247557846 1.9103369445911678 87.62 1.9399999999999977
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 158 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 159 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf36 1.9107566783735581 1.9273803227885578 87.6 1.960000000000008
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 157 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf37 1.9013559086026153 1.9230901214481015 87.58 1.980000000000004
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 12 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf38 1.8984089819969947 1.9195632881772446 87.58 1.980000000000004
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf39 1.9020502478364545 1.923319572598976 87.52 2.0400000000000063
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf40 1.9020502478364545 1.923319572598976 87.52 2.0400000000000063
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf41 1.9013559086026153 1.9230901214481015 87.5 2.0600000000000023
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf42 1.9013559086026153 1.9230901214481015 87.46000000000001 2.0999999999999943
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv fp16 11 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf43 1.9196179152539186 1.9443459719929068 87.44 2.1200000000000045
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 153 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf44 1.9020502478364545 1.923319572598976 87.4 2.1599999999999966
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf45 1.9152817031040366 1.9357432559063958 87.4 2.1599999999999966
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf46 1.915754791147898 1.9373322475753219 87.4 2.1599999999999966
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf47 1.9130551004051772 1.9409232417921056 87.38 2.180000000000007
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv perf_fp16 153 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf48 1.9421147660673033 1.9584555432766413 87.38 2.180000000000007
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf49 1.9052849920081363 1.9300100333661123 87.32 2.240000000000009
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 153 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf50 1.9154322863033566 1.934908329027621 87.3 2.260000000000005
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv perf_fp16 151 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
-+++++
-conf51 1.9079703554020564 1.9287218218306195 86.96000000000001 2.5999999999999943
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1
-3 gpu conv fp16 12 add fp16 1
-4 gpu add fp16 12
-5 gpu relu fp16 12
-6 gpu conv perf_fp16 153 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 161 add fp16 1
-8 gpu add fp16 12
-9 gpu relu fp16 12
-10 gpu conv perf_fp16 154 add fp16 1 relu fp16 1
-11 gpu conv perf_fp16 151 add fp16 1
-12 gpu add fp16 12
-13 gpu relu fp16 12
-14 gpu conv fp16 12 add fp16 1 relu fp16 1
-15 gpu conv fp16 12 add fp16 1
-16 gpu conv fp16 11 add fp16 1
-17 gpu add fp16 12
-18 gpu relu fp16 12
-19 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-20 gpu conv samp_fp16 262 add fp16 1
-21 gpu add fp16 12
-22 gpu relu fp16 12
-23 gpu conv perf_fp16 158 add fp16 1 relu fp16 1
-24 gpu conv perf_fp16 153 add fp16 1
-25 gpu add fp16 12
-26 gpu relu fp16 12
-27 gpu conv fp16 12 add fp16 1 relu fp16 1
-28 gpu conv fp16 12 add fp16 1
-29 gpu conv samp_fp16 261 add fp16 1
-30 gpu add fp16 12
-31 gpu relu fp16 12
-32 gpu conv fp16 12 add fp16 1 relu fp16 1
-33 gpu conv fp16 12 add fp16 1
-34 gpu add fp16 12
-35 gpu relu fp16 12
-36 gpu conv fp16 12 add fp16 1 relu fp16 1
-37 gpu conv perf_fp16 152 add fp16 1
-38 gpu add fp16 12
-39 gpu relu fp16 12
-40 gpu pool_mean fp16 12
-41 gpu mul fp16 12 add fp16 1
-42 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/resnet50_imagenet/resnet50_imagenet.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/resnet50_imagenet/resnet50_imagenet.txt
deleted file mode 100644
index 094eed413b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/resnet50_imagenet/resnet50_imagenet.txt
+++ /dev/null
@@ -1,1057 +0,0 @@
-7161.053769000008
-+++++
-conf1 1 1 75.7 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-2 gpu batchnorm fp32 11
-3 gpu conv fp32 11 add fp32 1
-4 gpu batchnorm fp32 11
-5 gpu relu fp32 11
-6 gpu conv fp32 11 add fp32 1
-7 gpu batchnorm fp32 11
-8 gpu relu fp32 11
-9 gpu conv fp32 11 add fp32 1
-10 gpu batchnorm fp32 11
-11 gpu conv fp32 11 add fp32 1
-12 gpu batchnorm fp32 11
-13 gpu add fp32 11
-14 gpu relu fp32 11
-15 gpu conv fp32 11 add fp32 1
-16 gpu batchnorm fp32 11
-17 gpu relu fp32 11
-18 gpu conv fp32 11 add fp32 1
-19 gpu batchnorm fp32 11
-20 gpu relu fp32 11
-21 gpu conv fp32 11 add fp32 1
-22 gpu batchnorm fp32 11
-23 gpu add fp32 11
-24 gpu relu fp32 11
-25 gpu conv fp32 11 add fp32 1
-26 gpu batchnorm fp32 11
-27 gpu relu fp32 11
-28 gpu conv fp32 11 add fp32 1
-29 gpu batchnorm fp32 11
-30 gpu relu fp32 11
-31 gpu conv fp32 11 add fp32 1
-32 gpu batchnorm fp32 11
-33 gpu add fp32 11
-34 gpu relu fp32 11
-35 gpu conv fp32 11 add fp32 1
-36 gpu batchnorm fp32 11
-37 gpu relu fp32 11
-38 gpu conv fp32 11 add fp32 1
-39 gpu batchnorm fp32 11
-40 gpu relu fp32 11
-41 gpu conv fp32 11 add fp32 1
-42 gpu batchnorm fp32 11
-43 gpu conv fp32 11 add fp32 1
-44 gpu batchnorm fp32 11
-45 gpu add fp32 11
-46 gpu relu fp32 11
-47 gpu conv fp32 11 add fp32 1
-48 gpu batchnorm fp32 11
-49 gpu relu fp32 11
-50 gpu conv fp32 11 add fp32 1
-51 gpu batchnorm fp32 11
-52 gpu relu fp32 11
-53 gpu conv fp32 11 add fp32 1
-54 gpu batchnorm fp32 11
-55 gpu add fp32 11
-56 gpu relu fp32 11
-57 gpu conv fp32 11 add fp32 1
-58 gpu batchnorm fp32 11
-59 gpu relu fp32 11
-60 gpu conv fp32 11 add fp32 1
-61 gpu batchnorm fp32 11
-62 gpu relu fp32 11
-63 gpu conv fp32 11 add fp32 1
-64 gpu batchnorm fp32 11
-65 gpu add fp32 11
-66 gpu relu fp32 11
-67 gpu conv fp32 11 add fp32 1
-68 gpu batchnorm fp32 11
-69 gpu relu fp32 11
-70 gpu conv fp32 11 add fp32 1
-71 gpu batchnorm fp32 11
-72 gpu relu fp32 11
-73 gpu conv fp32 11 add fp32 1
-74 gpu batchnorm fp32 11
-75 gpu add fp32 11
-76 gpu relu fp32 11
-77 gpu conv fp32 11 add fp32 1
-78 gpu batchnorm fp32 11
-79 gpu relu fp32 11
-80 gpu conv fp32 11 add fp32 1
-81 gpu batchnorm fp32 11
-82 gpu relu fp32 11
-83 gpu conv fp32 11 add fp32 1
-84 gpu batchnorm fp32 11
-85 gpu conv fp32 11 add fp32 1
-86 gpu batchnorm fp32 11
-87 gpu add fp32 11
-88 gpu relu fp32 11
-89 gpu conv fp32 11 add fp32 1
-90 gpu batchnorm fp32 11
-91 gpu relu fp32 11
-92 gpu conv fp32 11 add fp32 1
-93 gpu batchnorm fp32 11
-94 gpu relu fp32 11
-95 gpu conv fp32 11 add fp32 1
-96 gpu batchnorm fp32 11
-97 gpu add fp32 11
-98 gpu relu fp32 11
-99 gpu conv fp32 11 add fp32 1
-100 gpu batchnorm fp32 11
-101 gpu relu fp32 11
-102 gpu conv fp32 11 add fp32 1
-103 gpu batchnorm fp32 11
-104 gpu relu fp32 11
-105 gpu conv fp32 11 add fp32 1
-106 gpu batchnorm fp32 11
-107 gpu add fp32 11
-108 gpu relu fp32 11
-109 gpu conv fp32 11 add fp32 1
-110 gpu batchnorm fp32 11
-111 gpu relu fp32 11
-112 gpu conv fp32 11 add fp32 1
-113 gpu batchnorm fp32 11
-114 gpu relu fp32 11
-115 gpu conv fp32 11 add fp32 1
-116 gpu batchnorm fp32 11
-117 gpu add fp32 11
-118 gpu relu fp32 11
-119 gpu conv fp32 11 add fp32 1
-120 gpu batchnorm fp32 11
-121 gpu relu fp32 11
-122 gpu conv fp32 11 add fp32 1
-123 gpu batchnorm fp32 11
-124 gpu relu fp32 11
-125 gpu conv fp32 11 add fp32 1
-126 gpu batchnorm fp32 11
-127 gpu add fp32 11
-128 gpu relu fp32 11
-129 gpu conv fp32 11 add fp32 1
-130 gpu batchnorm fp32 11
-131 gpu relu fp32 11
-132 gpu conv fp32 11 add fp32 1
-133 gpu batchnorm fp32 11
-134 gpu relu fp32 11
-135 gpu conv fp32 11 add fp32 1
-136 gpu batchnorm fp32 11
-137 gpu add fp32 11
-138 gpu relu fp32 11
-139 gpu conv fp32 11 add fp32 1
-140 gpu batchnorm fp32 11
-141 gpu relu fp32 11
-142 gpu conv fp32 11 add fp32 1
-143 gpu batchnorm fp32 11
-144 gpu relu fp32 11
-145 gpu conv fp32 11 add fp32 1
-146 gpu batchnorm fp32 11
-147 gpu conv fp32 11 add fp32 1
-148 gpu batchnorm fp32 11
-149 gpu add fp32 11
-150 gpu relu fp32 11
-151 gpu conv fp32 11 add fp32 1
-152 gpu batchnorm fp32 11
-153 gpu relu fp32 11
-154 gpu conv fp32 11 add fp32 1
-155 gpu batchnorm fp32 11
-156 gpu relu fp32 11
-157 gpu conv fp32 11 add fp32 1
-158 gpu batchnorm fp32 11
-159 gpu add fp32 11
-160 gpu relu fp32 11
-161 gpu conv fp32 11 add fp32 1
-162 gpu batchnorm fp32 11
-163 gpu relu fp32 11
-164 gpu conv fp32 11 add fp32 1
-165 gpu batchnorm fp32 11
-166 gpu relu fp32 11
-167 gpu conv fp32 11 add fp32 1
-168 gpu batchnorm fp32 11
-169 gpu add fp32 11
-170 gpu relu fp32 11
-171 gpu pool_max fp32 11
-172 gpu mul fp32 11 add fp32 1
-173 gpu softmax fp32 1
------
-+++++
-conf2 1.8254789092281507 1.4527803526239977 75.7 0.0
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu batchnorm fp16 12
-3 gpu conv fp16 12 add fp16 1
-4 gpu batchnorm fp16 12
-5 gpu relu fp16 12
-6 gpu conv fp16 12 add fp16 1
-7 gpu batchnorm fp16 12
-8 gpu relu fp16 12
-9 gpu conv fp16 12 add fp16 1
-10 gpu batchnorm fp16 12
-11 gpu conv fp16 12 add fp16 1
-12 gpu batchnorm fp16 12
-13 gpu add fp16 12
-14 gpu relu fp16 12
-15 gpu conv fp16 12 add fp16 1
-16 gpu batchnorm fp16 12
-17 gpu relu fp16 12
-18 gpu conv fp16 12 add fp16 1
-19 gpu batchnorm fp16 12
-20 gpu relu fp16 12
-21 gpu conv fp16 12 add fp16 1
-22 gpu batchnorm fp16 12
-23 gpu add fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12 add fp16 1
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu conv fp16 12 add fp16 1
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12 add fp16 1
-32 gpu batchnorm fp16 12
-33 gpu add fp16 12
-34 gpu relu fp16 12
-35 gpu conv fp16 12 add fp16 1
-36 gpu batchnorm fp16 12
-37 gpu relu fp16 12
-38 gpu conv fp16 12 add fp16 1
-39 gpu batchnorm fp16 12
-40 gpu relu fp16 12
-41 gpu conv fp16 12 add fp16 1
-42 gpu batchnorm fp16 12
-43 gpu conv fp16 12 add fp16 1
-44 gpu batchnorm fp16 12
-45 gpu add fp16 12
-46 gpu relu fp16 12
-47 gpu conv fp16 12 add fp16 1
-48 gpu batchnorm fp16 12
-49 gpu relu fp16 12
-50 gpu conv fp16 12 add fp16 1
-51 gpu batchnorm fp16 12
-52 gpu relu fp16 12
-53 gpu conv fp16 12 add fp16 1
-54 gpu batchnorm fp16 12
-55 gpu add fp16 12
-56 gpu relu fp16 12
-57 gpu conv fp16 12 add fp16 1
-58 gpu batchnorm fp16 12
-59 gpu relu fp16 12
-60 gpu conv fp16 12 add fp16 1
-61 gpu batchnorm fp16 12
-62 gpu relu fp16 12
-63 gpu conv fp16 12 add fp16 1
-64 gpu batchnorm fp16 12
-65 gpu add fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12 add fp16 1
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu conv fp16 12 add fp16 1
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv fp16 12 add fp16 1
-74 gpu batchnorm fp16 12
-75 gpu add fp16 12
-76 gpu relu fp16 12
-77 gpu conv fp16 12 add fp16 1
-78 gpu batchnorm fp16 12
-79 gpu relu fp16 12
-80 gpu conv fp16 12 add fp16 1
-81 gpu batchnorm fp16 12
-82 gpu relu fp16 12
-83 gpu conv fp16 12 add fp16 1
-84 gpu batchnorm fp16 12
-85 gpu conv fp16 12 add fp16 1
-86 gpu batchnorm fp16 12
-87 gpu add fp16 12
-88 gpu relu fp16 12
-89 gpu conv fp16 12 add fp16 1
-90 gpu batchnorm fp16 12
-91 gpu relu fp16 12
-92 gpu conv fp16 12 add fp16 1
-93 gpu batchnorm fp16 12
-94 gpu relu fp16 12
-95 gpu conv fp16 12 add fp16 1
-96 gpu batchnorm fp16 12
-97 gpu add fp16 12
-98 gpu relu fp16 12
-99 gpu conv fp16 12 add fp16 1
-100 gpu batchnorm fp16 12
-101 gpu relu fp16 12
-102 gpu conv fp16 12 add fp16 1
-103 gpu batchnorm fp16 12
-104 gpu relu fp16 12
-105 gpu conv fp16 12 add fp16 1
-106 gpu batchnorm fp16 12
-107 gpu add fp16 12
-108 gpu relu fp16 12
-109 gpu conv fp16 12 add fp16 1
-110 gpu batchnorm fp16 12
-111 gpu relu fp16 12
-112 gpu conv fp16 12 add fp16 1
-113 gpu batchnorm fp16 12
-114 gpu relu fp16 12
-115 gpu conv fp16 12 add fp16 1
-116 gpu batchnorm fp16 12
-117 gpu add fp16 12
-118 gpu relu fp16 12
-119 gpu conv fp16 12 add fp16 1
-120 gpu batchnorm fp16 12
-121 gpu relu fp16 12
-122 gpu conv fp16 12 add fp16 1
-123 gpu batchnorm fp16 12
-124 gpu relu fp16 12
-125 gpu conv fp16 12 add fp16 1
-126 gpu batchnorm fp16 12
-127 gpu add fp16 12
-128 gpu relu fp16 12
-129 gpu conv fp16 12 add fp16 1
-130 gpu batchnorm fp16 12
-131 gpu relu fp16 12
-132 gpu conv fp16 12 add fp16 1
-133 gpu batchnorm fp16 12
-134 gpu relu fp16 12
-135 gpu conv fp16 12 add fp16 1
-136 gpu batchnorm fp16 12
-137 gpu add fp16 12
-138 gpu relu fp16 12
-139 gpu conv fp16 12 add fp16 1
-140 gpu batchnorm fp16 12
-141 gpu relu fp16 12
-142 gpu conv fp16 12 add fp16 1
-143 gpu batchnorm fp16 12
-144 gpu relu fp16 12
-145 gpu conv fp16 12 add fp16 1
-146 gpu batchnorm fp16 12
-147 gpu conv fp16 12 add fp16 1
-148 gpu batchnorm fp16 12
-149 gpu add fp16 12
-150 gpu relu fp16 12
-151 gpu conv fp16 12 add fp16 1
-152 gpu batchnorm fp16 12
-153 gpu relu fp16 12
-154 gpu conv fp16 12 add fp16 1
-155 gpu batchnorm fp16 12
-156 gpu relu fp16 12
-157 gpu conv fp16 12 add fp16 1
-158 gpu batchnorm fp16 12
-159 gpu add fp16 12
-160 gpu relu fp16 12
-161 gpu conv fp16 12 add fp16 1
-162 gpu batchnorm fp16 12
-163 gpu relu fp16 12
-164 gpu conv fp16 12 add fp16 1
-165 gpu batchnorm fp16 12
-166 gpu relu fp16 12
-167 gpu conv fp16 12 add fp16 1
-168 gpu batchnorm fp16 12
-169 gpu add fp16 12
-170 gpu relu fp16 12
-171 gpu pool_max fp16 12
-172 gpu mul fp16 12 add fp16 1
-173 gpu softmax fp32 1
------
-+++++
-conf3 1.8254789092281507 1.4527803526239977 75.7 0.0
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu batchnorm fp16 12
-3 gpu conv fp16 12 add fp16 1
-4 gpu batchnorm fp16 12
-5 gpu relu fp16 12
-6 gpu conv fp16 12 add fp16 1
-7 gpu batchnorm fp16 12
-8 gpu relu fp16 12
-9 gpu conv fp16 12 add fp16 1
-10 gpu batchnorm fp16 12
-11 gpu conv fp16 12 add fp16 1
-12 gpu batchnorm fp16 12
-13 gpu add fp16 12
-14 gpu relu fp16 12
-15 gpu conv fp16 12 add fp16 1
-16 gpu batchnorm fp16 12
-17 gpu relu fp16 12
-18 gpu conv fp16 12 add fp16 1
-19 gpu batchnorm fp16 12
-20 gpu relu fp16 12
-21 gpu conv fp16 12 add fp16 1
-22 gpu batchnorm fp16 12
-23 gpu add fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12 add fp16 1
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu conv fp16 12 add fp16 1
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12 add fp16 1
-32 gpu batchnorm fp16 12
-33 gpu add fp16 12
-34 gpu relu fp16 12
-35 gpu conv fp16 12 add fp16 1
-36 gpu batchnorm fp16 12
-37 gpu relu fp16 12
-38 gpu conv fp16 12 add fp16 1
-39 gpu batchnorm fp16 12
-40 gpu relu fp16 12
-41 gpu conv fp16 12 add fp16 1
-42 gpu batchnorm fp16 12
-43 gpu conv fp16 12 add fp16 1
-44 gpu batchnorm fp16 12
-45 gpu add fp16 12
-46 gpu relu fp16 12
-47 gpu conv fp16 12 add fp16 1
-48 gpu batchnorm fp16 12
-49 gpu relu fp16 12
-50 gpu conv fp16 12 add fp16 1
-51 gpu batchnorm fp16 12
-52 gpu relu fp16 12
-53 gpu conv fp16 12 add fp16 1
-54 gpu batchnorm fp16 12
-55 gpu add fp16 12
-56 gpu relu fp16 12
-57 gpu conv fp16 12 add fp16 1
-58 gpu batchnorm fp16 12
-59 gpu relu fp16 12
-60 gpu conv fp16 12 add fp16 1
-61 gpu batchnorm fp16 12
-62 gpu relu fp16 12
-63 gpu conv fp16 12 add fp16 1
-64 gpu batchnorm fp16 12
-65 gpu add fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12 add fp16 1
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu conv fp16 12 add fp16 1
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv fp16 12 add fp16 1
-74 gpu batchnorm fp16 12
-75 gpu add fp16 12
-76 gpu relu fp16 12
-77 gpu conv fp16 12 add fp16 1
-78 gpu batchnorm fp16 12
-79 gpu relu fp16 12
-80 gpu conv fp16 12 add fp16 1
-81 gpu batchnorm fp16 12
-82 gpu relu fp16 12
-83 gpu conv fp16 12 add fp16 1
-84 gpu batchnorm fp16 12
-85 gpu conv fp16 12 add fp16 1
-86 gpu batchnorm fp16 12
-87 gpu add fp16 12
-88 gpu relu fp16 12
-89 gpu conv fp16 12 add fp16 1
-90 gpu batchnorm fp16 12
-91 gpu relu fp16 12
-92 gpu conv fp16 12 add fp16 1
-93 gpu batchnorm fp16 12
-94 gpu relu fp16 12
-95 gpu conv fp16 12 add fp16 1
-96 gpu batchnorm fp16 12
-97 gpu add fp16 12
-98 gpu relu fp16 12
-99 gpu conv fp16 12 add fp16 1
-100 gpu batchnorm fp16 12
-101 gpu relu fp16 12
-102 gpu conv fp16 12 add fp16 1
-103 gpu batchnorm fp16 12
-104 gpu relu fp16 12
-105 gpu conv fp16 12 add fp16 1
-106 gpu batchnorm fp16 12
-107 gpu add fp16 12
-108 gpu relu fp16 12
-109 gpu conv fp16 12 add fp16 1
-110 gpu batchnorm fp16 12
-111 gpu relu fp16 12
-112 gpu conv fp16 12 add fp16 1
-113 gpu batchnorm fp16 12
-114 gpu relu fp16 12
-115 gpu conv fp16 12 add fp16 1
-116 gpu batchnorm fp16 12
-117 gpu add fp16 12
-118 gpu relu fp16 12
-119 gpu conv fp16 12 add fp16 1
-120 gpu batchnorm fp16 12
-121 gpu relu fp16 12
-122 gpu conv fp16 12 add fp16 1
-123 gpu batchnorm fp16 12
-124 gpu relu fp16 12
-125 gpu conv fp16 12 add fp16 1
-126 gpu batchnorm fp16 12
-127 gpu add fp16 12
-128 gpu relu fp16 12
-129 gpu conv fp16 12 add fp16 1
-130 gpu batchnorm fp16 12
-131 gpu relu fp16 12
-132 gpu conv fp16 12 add fp16 1
-133 gpu batchnorm fp16 12
-134 gpu relu fp16 12
-135 gpu conv fp16 12 add fp16 1
-136 gpu batchnorm fp16 12
-137 gpu add fp16 12
-138 gpu relu fp16 12
-139 gpu conv fp16 12 add fp16 1
-140 gpu batchnorm fp16 12
-141 gpu relu fp16 12
-142 gpu conv fp16 12 add fp16 1
-143 gpu batchnorm fp16 12
-144 gpu relu fp16 12
-145 gpu conv fp16 12 add fp16 1
-146 gpu batchnorm fp16 12
-147 gpu conv fp16 12 add fp16 1
-148 gpu batchnorm fp16 12
-149 gpu add fp16 12
-150 gpu relu fp16 12
-151 gpu conv fp16 12 add fp16 1
-152 gpu batchnorm fp16 12
-153 gpu relu fp16 12
-154 gpu conv fp16 12 add fp16 1
-155 gpu batchnorm fp16 12
-156 gpu relu fp16 12
-157 gpu conv fp16 12 add fp16 1
-158 gpu batchnorm fp16 12
-159 gpu add fp16 12
-160 gpu relu fp16 12
-161 gpu conv fp16 12 add fp16 1
-162 gpu batchnorm fp16 12
-163 gpu relu fp16 12
-164 gpu conv fp16 12 add fp16 1
-165 gpu batchnorm fp16 12
-166 gpu relu fp16 12
-167 gpu conv fp16 12 add fp16 1
-168 gpu batchnorm fp16 12
-169 gpu add fp16 12
-170 gpu relu fp16 12
-171 gpu pool_max fp16 12
-172 gpu mul fp16 12 add fp16 1
-173 gpu softmax fp32 1
------
-+++++
-conf4 1.8254789092281507 1.4527803526239977 75.7 0.0
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu batchnorm fp16 12
-3 gpu conv fp16 12 add fp16 1
-4 gpu batchnorm fp16 12
-5 gpu relu fp16 12
-6 gpu conv fp16 12 add fp16 1
-7 gpu batchnorm fp16 12
-8 gpu relu fp16 12
-9 gpu conv fp16 12 add fp16 1
-10 gpu batchnorm fp16 12
-11 gpu conv fp16 12 add fp16 1
-12 gpu batchnorm fp16 12
-13 gpu add fp16 12
-14 gpu relu fp16 12
-15 gpu conv fp16 12 add fp16 1
-16 gpu batchnorm fp16 12
-17 gpu relu fp16 12
-18 gpu conv fp16 12 add fp16 1
-19 gpu batchnorm fp16 12
-20 gpu relu fp16 12
-21 gpu conv fp16 12 add fp16 1
-22 gpu batchnorm fp16 12
-23 gpu add fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12 add fp16 1
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu conv fp16 12 add fp16 1
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12 add fp16 1
-32 gpu batchnorm fp16 12
-33 gpu add fp16 12
-34 gpu relu fp16 12
-35 gpu conv fp16 12 add fp16 1
-36 gpu batchnorm fp16 12
-37 gpu relu fp16 12
-38 gpu conv fp16 12 add fp16 1
-39 gpu batchnorm fp16 12
-40 gpu relu fp16 12
-41 gpu conv fp16 12 add fp16 1
-42 gpu batchnorm fp16 12
-43 gpu conv fp16 12 add fp16 1
-44 gpu batchnorm fp16 12
-45 gpu add fp16 12
-46 gpu relu fp16 12
-47 gpu conv fp16 12 add fp16 1
-48 gpu batchnorm fp16 12
-49 gpu relu fp16 12
-50 gpu conv fp16 12 add fp16 1
-51 gpu batchnorm fp16 12
-52 gpu relu fp16 12
-53 gpu conv fp16 12 add fp16 1
-54 gpu batchnorm fp16 12
-55 gpu add fp16 12
-56 gpu relu fp16 12
-57 gpu conv fp16 12 add fp16 1
-58 gpu batchnorm fp16 12
-59 gpu relu fp16 12
-60 gpu conv fp16 12 add fp16 1
-61 gpu batchnorm fp16 12
-62 gpu relu fp16 12
-63 gpu conv fp16 12 add fp16 1
-64 gpu batchnorm fp16 12
-65 gpu add fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12 add fp16 1
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu conv fp16 12 add fp16 1
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv fp16 12 add fp16 1
-74 gpu batchnorm fp16 12
-75 gpu add fp16 12
-76 gpu relu fp16 12
-77 gpu conv fp16 12 add fp16 1
-78 gpu batchnorm fp16 12
-79 gpu relu fp16 12
-80 gpu conv fp16 12 add fp16 1
-81 gpu batchnorm fp16 12
-82 gpu relu fp16 12
-83 gpu conv fp16 12 add fp16 1
-84 gpu batchnorm fp16 12
-85 gpu conv fp16 12 add fp16 1
-86 gpu batchnorm fp16 12
-87 gpu add fp16 12
-88 gpu relu fp16 12
-89 gpu conv fp16 12 add fp16 1
-90 gpu batchnorm fp16 12
-91 gpu relu fp16 12
-92 gpu conv fp16 12 add fp16 1
-93 gpu batchnorm fp16 12
-94 gpu relu fp16 12
-95 gpu conv fp16 12 add fp16 1
-96 gpu batchnorm fp16 12
-97 gpu add fp16 12
-98 gpu relu fp16 12
-99 gpu conv fp16 12 add fp16 1
-100 gpu batchnorm fp16 12
-101 gpu relu fp16 12
-102 gpu conv fp16 12 add fp16 1
-103 gpu batchnorm fp16 12
-104 gpu relu fp16 12
-105 gpu conv fp16 12 add fp16 1
-106 gpu batchnorm fp16 12
-107 gpu add fp16 12
-108 gpu relu fp16 12
-109 gpu conv fp16 12 add fp16 1
-110 gpu batchnorm fp16 12
-111 gpu relu fp16 12
-112 gpu conv fp16 12 add fp16 1
-113 gpu batchnorm fp16 12
-114 gpu relu fp16 12
-115 gpu conv fp16 12 add fp16 1
-116 gpu batchnorm fp16 12
-117 gpu add fp16 12
-118 gpu relu fp16 12
-119 gpu conv fp16 12 add fp16 1
-120 gpu batchnorm fp16 12
-121 gpu relu fp16 12
-122 gpu conv fp16 12 add fp16 1
-123 gpu batchnorm fp16 12
-124 gpu relu fp16 12
-125 gpu conv fp16 12 add fp16 1
-126 gpu batchnorm fp16 12
-127 gpu add fp16 12
-128 gpu relu fp16 12
-129 gpu conv fp16 12 add fp16 1
-130 gpu batchnorm fp16 12
-131 gpu relu fp16 12
-132 gpu conv fp16 12 add fp16 1
-133 gpu batchnorm fp16 12
-134 gpu relu fp16 12
-135 gpu conv fp16 12 add fp16 1
-136 gpu batchnorm fp16 12
-137 gpu add fp16 12
-138 gpu relu fp16 12
-139 gpu conv fp16 12 add fp16 1
-140 gpu batchnorm fp16 12
-141 gpu relu fp16 12
-142 gpu conv fp16 12 add fp16 1
-143 gpu batchnorm fp16 12
-144 gpu relu fp16 12
-145 gpu conv fp16 12 add fp16 1
-146 gpu batchnorm fp16 12
-147 gpu conv fp16 12 add fp16 1
-148 gpu batchnorm fp16 12
-149 gpu add fp16 12
-150 gpu relu fp16 12
-151 gpu conv fp16 12 add fp16 1
-152 gpu batchnorm fp16 12
-153 gpu relu fp16 12
-154 gpu conv fp16 12 add fp16 1
-155 gpu batchnorm fp16 12
-156 gpu relu fp16 12
-157 gpu conv fp16 12 add fp16 1
-158 gpu batchnorm fp16 12
-159 gpu add fp16 12
-160 gpu relu fp16 12
-161 gpu conv fp16 12 add fp16 1
-162 gpu batchnorm fp16 12
-163 gpu relu fp16 12
-164 gpu conv fp16 12 add fp16 1
-165 gpu batchnorm fp16 12
-166 gpu relu fp16 12
-167 gpu conv fp16 12 add fp16 1
-168 gpu batchnorm fp16 12
-169 gpu add fp16 12
-170 gpu relu fp16 12
-171 gpu pool_max fp16 12
-172 gpu mul fp16 12 add fp16 1
-173 gpu softmax fp32 1
------
-+++++
-conf5 1.8323072136026506 1.457112696128105 74.76 0.9399999999999977
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu batchnorm fp16 12
-3 gpu conv fp16 12 add fp16 1
-4 gpu batchnorm fp16 12
-5 gpu relu fp16 12
-6 gpu conv fp16 12 add fp16 1
-7 gpu batchnorm fp16 12
-8 gpu relu fp16 12
-9 gpu conv fp16 12 add fp16 1
-10 gpu batchnorm fp16 12
-11 gpu conv fp16 12 add fp16 1
-12 gpu batchnorm fp16 12
-13 gpu add fp16 12
-14 gpu relu fp16 12
-15 gpu conv fp16 12 add fp16 1
-16 gpu batchnorm fp16 12
-17 gpu relu fp16 12
-18 gpu conv fp16 12 add fp16 1
-19 gpu batchnorm fp16 12
-20 gpu relu fp16 12
-21 gpu conv fp16 12 add fp16 1
-22 gpu batchnorm fp16 12
-23 gpu add fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12 add fp16 1
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu conv fp16 12 add fp16 1
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12 add fp16 1
-32 gpu batchnorm fp16 12
-33 gpu add fp16 12
-34 gpu relu fp16 12
-35 gpu conv fp16 12 add fp16 1
-36 gpu batchnorm fp16 12
-37 gpu relu fp16 12
-38 gpu conv fp16 12 add fp16 1
-39 gpu batchnorm fp16 12
-40 gpu relu fp16 12
-41 gpu conv fp16 12 add fp16 1
-42 gpu batchnorm fp16 12
-43 gpu conv fp16 12 add fp16 1
-44 gpu batchnorm fp16 12
-45 gpu add fp16 12
-46 gpu relu fp16 12
-47 gpu conv fp16 12 add fp16 1
-48 gpu batchnorm fp16 12
-49 gpu relu fp16 12
-50 gpu conv fp16 12 add fp16 1
-51 gpu batchnorm fp16 12
-52 gpu relu fp16 12
-53 gpu conv fp16 12 add fp16 1
-54 gpu batchnorm fp16 12
-55 gpu add fp16 12
-56 gpu relu fp16 12
-57 gpu conv fp16 12 add fp16 1
-58 gpu batchnorm fp16 12
-59 gpu relu fp16 12
-60 gpu conv fp16 12 add fp16 1
-61 gpu batchnorm fp16 12
-62 gpu relu fp16 12
-63 gpu conv fp16 12 add fp16 1
-64 gpu batchnorm fp16 12
-65 gpu add fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12 add fp16 1
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu conv fp16 12 add fp16 1
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv fp16 12 add fp16 1
-74 gpu batchnorm fp16 12
-75 gpu add fp16 12
-76 gpu relu fp16 12
-77 gpu conv fp16 12 add fp16 1
-78 gpu batchnorm fp16 12
-79 gpu relu fp16 12
-80 gpu conv fp16 12 add fp16 1
-81 gpu batchnorm fp16 12
-82 gpu relu fp16 12
-83 gpu conv fp16 12 add fp16 1
-84 gpu batchnorm fp16 12
-85 gpu conv fp16 12 add fp16 1
-86 gpu batchnorm fp16 12
-87 gpu add fp16 12
-88 gpu relu fp16 12
-89 gpu conv fp16 12 add fp16 1
-90 gpu batchnorm fp16 12
-91 gpu relu fp16 12
-92 gpu conv fp16 12 add fp16 1
-93 gpu batchnorm fp16 12
-94 gpu relu fp16 12
-95 gpu conv fp16 12 add fp16 1
-96 gpu batchnorm fp16 12
-97 gpu add fp16 12
-98 gpu relu fp16 12
-99 gpu conv perf_fp16 157 add fp16 1
-100 gpu batchnorm fp16 12
-101 gpu relu fp16 12
-102 gpu conv fp16 12 add fp16 1
-103 gpu batchnorm fp16 12
-104 gpu relu fp16 12
-105 gpu conv fp16 12 add fp16 1
-106 gpu batchnorm fp16 12
-107 gpu add fp16 12
-108 gpu relu fp16 12
-109 gpu conv fp16 12 add fp16 1
-110 gpu batchnorm fp16 12
-111 gpu relu fp16 12
-112 gpu conv fp16 12 add fp16 1
-113 gpu batchnorm fp16 12
-114 gpu relu fp16 12
-115 gpu conv fp16 12 add fp16 1
-116 gpu batchnorm fp16 12
-117 gpu add fp16 12
-118 gpu relu fp16 12
-119 gpu conv fp16 12 add fp16 1
-120 gpu batchnorm fp16 12
-121 gpu relu fp16 12
-122 gpu conv fp16 12 add fp16 1
-123 gpu batchnorm fp16 12
-124 gpu relu fp16 12
-125 gpu conv fp16 12 add fp16 1
-126 gpu batchnorm fp16 12
-127 gpu add fp16 12
-128 gpu relu fp16 12
-129 gpu conv fp16 12 add fp16 1
-130 gpu batchnorm fp16 12
-131 gpu relu fp16 12
-132 gpu conv fp16 12 add fp16 1
-133 gpu batchnorm fp16 12
-134 gpu relu fp16 12
-135 gpu conv fp16 12 add fp16 1
-136 gpu batchnorm fp16 12
-137 gpu add fp16 12
-138 gpu relu fp16 12
-139 gpu conv fp16 12 add fp16 1
-140 gpu batchnorm fp16 12
-141 gpu relu fp16 12
-142 gpu conv fp16 12 add fp16 1
-143 gpu batchnorm fp16 12
-144 gpu relu fp16 12
-145 gpu conv fp16 12 add fp16 1
-146 gpu batchnorm fp16 12
-147 gpu conv fp16 12 add fp16 1
-148 gpu batchnorm fp16 12
-149 gpu add fp16 12
-150 gpu relu fp16 12
-151 gpu conv fp16 12 add fp16 1
-152 gpu batchnorm fp16 12
-153 gpu relu fp16 12
-154 gpu conv fp16 12 add fp16 1
-155 gpu batchnorm fp16 12
-156 gpu relu fp16 12
-157 gpu conv fp16 12 add fp16 1
-158 gpu batchnorm fp16 12
-159 gpu add fp16 12
-160 gpu relu fp16 12
-161 gpu conv fp16 12 add fp16 1
-162 gpu batchnorm fp16 12
-163 gpu relu fp16 12
-164 gpu conv perf_fp16 152 add fp16 1
-165 gpu batchnorm fp16 12
-166 gpu relu fp16 12
-167 gpu conv fp16 12 add fp16 1
-168 gpu batchnorm fp16 12
-169 gpu add fp16 12
-170 gpu relu fp16 12
-171 gpu pool_max fp16 12
-172 gpu mul fp16 12 add fp16 1
-173 gpu softmax fp32 1
------
-+++++
-conf6 1.8333922701839533 1.4589203187717397 74.53999999999999 1.1600000000000108
-1 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-2 gpu batchnorm fp16 12
-3 gpu conv fp16 12 add fp16 1
-4 gpu batchnorm fp16 12
-5 gpu relu fp16 12
-6 gpu conv fp16 12 add fp16 1
-7 gpu batchnorm fp16 12
-8 gpu relu fp16 12
-9 gpu conv fp16 12 add fp16 1
-10 gpu batchnorm fp16 12
-11 gpu conv fp16 12 add fp16 1
-12 gpu batchnorm fp16 12
-13 gpu add fp16 12
-14 gpu relu fp16 12
-15 gpu conv fp16 12 add fp16 1
-16 gpu batchnorm fp16 12
-17 gpu relu fp16 12
-18 gpu conv fp16 12 add fp16 1
-19 gpu batchnorm fp16 12
-20 gpu relu fp16 12
-21 gpu conv fp16 12 add fp16 1
-22 gpu batchnorm fp16 12
-23 gpu add fp16 12
-24 gpu relu fp16 12
-25 gpu conv fp16 12 add fp16 1
-26 gpu batchnorm fp16 12
-27 gpu relu fp16 12
-28 gpu conv fp16 12 add fp16 1
-29 gpu batchnorm fp16 12
-30 gpu relu fp16 12
-31 gpu conv fp16 12 add fp16 1
-32 gpu batchnorm fp16 12
-33 gpu add fp16 12
-34 gpu relu fp16 12
-35 gpu conv fp16 12 add fp16 1
-36 gpu batchnorm fp16 12
-37 gpu relu fp16 12
-38 gpu conv fp16 12 add fp16 1
-39 gpu batchnorm fp16 12
-40 gpu relu fp16 12
-41 gpu conv fp16 12 add fp16 1
-42 gpu batchnorm fp16 12
-43 gpu conv fp16 12 add fp16 1
-44 gpu batchnorm fp16 12
-45 gpu add fp16 12
-46 gpu relu fp16 12
-47 gpu conv fp16 12 add fp16 1
-48 gpu batchnorm fp16 12
-49 gpu relu fp16 12
-50 gpu conv fp16 12 add fp16 1
-51 gpu batchnorm fp16 12
-52 gpu relu fp16 12
-53 gpu conv fp16 12 add fp16 1
-54 gpu batchnorm fp16 12
-55 gpu add fp16 12
-56 gpu relu fp16 12
-57 gpu conv fp16 12 add fp16 1
-58 gpu batchnorm fp16 12
-59 gpu relu fp16 12
-60 gpu conv fp16 12 add fp16 1
-61 gpu batchnorm fp16 12
-62 gpu relu fp16 12
-63 gpu conv fp16 12 add fp16 1
-64 gpu batchnorm fp16 12
-65 gpu add fp16 12
-66 gpu relu fp16 12
-67 gpu conv fp16 12 add fp16 1
-68 gpu batchnorm fp16 12
-69 gpu relu fp16 12
-70 gpu conv fp16 12 add fp16 1
-71 gpu batchnorm fp16 12
-72 gpu relu fp16 12
-73 gpu conv fp16 12 add fp16 1
-74 gpu batchnorm fp16 12
-75 gpu add fp16 12
-76 gpu relu fp16 12
-77 gpu conv fp16 12 add fp16 1
-78 gpu batchnorm fp16 12
-79 gpu relu fp16 12
-80 gpu conv fp16 12 add fp16 1
-81 gpu batchnorm fp16 12
-82 gpu relu fp16 12
-83 gpu conv fp16 12 add fp16 1
-84 gpu batchnorm fp16 12
-85 gpu conv fp16 12 add fp16 1
-86 gpu batchnorm fp16 12
-87 gpu add fp16 12
-88 gpu relu fp16 12
-89 gpu conv fp16 12 add fp16 1
-90 gpu batchnorm fp16 12
-91 gpu relu fp16 12
-92 gpu conv fp16 12 add fp16 1
-93 gpu batchnorm fp16 12
-94 gpu relu fp16 12
-95 gpu conv fp16 12 add fp16 1
-96 gpu batchnorm fp16 12
-97 gpu add fp16 12
-98 gpu relu fp16 12
-99 gpu conv perf_fp16 157 add fp16 1
-100 gpu batchnorm fp16 12
-101 gpu relu fp16 12
-102 gpu conv samp_fp16 267 add fp16 1
-103 gpu batchnorm fp16 12
-104 gpu relu fp16 12
-105 gpu conv fp16 12 add fp16 1
-106 gpu batchnorm fp16 12
-107 gpu add fp16 12
-108 gpu relu fp16 12
-109 gpu conv fp16 12 add fp16 1
-110 gpu batchnorm fp16 12
-111 gpu relu fp16 12
-112 gpu conv fp16 12 add fp16 1
-113 gpu batchnorm fp16 12
-114 gpu relu fp16 12
-115 gpu conv fp16 12 add fp16 1
-116 gpu batchnorm fp16 12
-117 gpu add fp16 12
-118 gpu relu fp16 12
-119 gpu conv fp16 12 add fp16 1
-120 gpu batchnorm fp16 12
-121 gpu relu fp16 12
-122 gpu conv fp16 12 add fp16 1
-123 gpu batchnorm fp16 12
-124 gpu relu fp16 12
-125 gpu conv fp16 12 add fp16 1
-126 gpu batchnorm fp16 12
-127 gpu add fp16 12
-128 gpu relu fp16 12
-129 gpu conv fp16 12 add fp16 1
-130 gpu batchnorm fp16 12
-131 gpu relu fp16 12
-132 gpu conv fp16 12 add fp16 1
-133 gpu batchnorm fp16 12
-134 gpu relu fp16 12
-135 gpu conv fp16 12 add fp16 1
-136 gpu batchnorm fp16 12
-137 gpu add fp16 12
-138 gpu relu fp16 12
-139 gpu conv fp16 12 add fp16 1
-140 gpu batchnorm fp16 12
-141 gpu relu fp16 12
-142 gpu conv fp16 12 add fp16 1
-143 gpu batchnorm fp16 12
-144 gpu relu fp16 12
-145 gpu conv fp16 12 add fp16 1
-146 gpu batchnorm fp16 12
-147 gpu conv fp16 12 add fp16 1
-148 gpu batchnorm fp16 12
-149 gpu add fp16 12
-150 gpu relu fp16 12
-151 gpu conv fp16 12 add fp16 1
-152 gpu batchnorm fp16 12
-153 gpu relu fp16 12
-154 gpu conv fp16 12 add fp16 1
-155 gpu batchnorm fp16 12
-156 gpu relu fp16 12
-157 gpu conv fp16 12 add fp16 1
-158 gpu batchnorm fp16 12
-159 gpu add fp16 12
-160 gpu relu fp16 12
-161 gpu conv fp16 12 add fp16 1
-162 gpu batchnorm fp16 12
-163 gpu relu fp16 12
-164 gpu conv perf_fp16 152 add fp16 1
-165 gpu batchnorm fp16 12
-166 gpu relu fp16 12
-167 gpu conv fp16 12 add fp16 1
-168 gpu batchnorm fp16 12
-169 gpu add fp16 12
-170 gpu relu fp16 12
-171 gpu pool_max fp16 12
-172 gpu mul fp16 12 add fp16 1
-173 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/vgg16_cifar10/vgg16_cifar10.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/vgg16_cifar10/vgg16_cifar10.txt
deleted file mode 100644
index 2b325a9fe2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/vgg16_cifar10/vgg16_cifar10.txt
+++ /dev/null
@@ -1,913 +0,0 @@
-3776.508929999999
-+++++
-conf1 1 1 89.96 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 relu fp32 1
-4 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-5 gpu conv fp32 11 add fp32 1 relu fp32 1
-6 gpu conv fp32 11 add fp32 1 relu fp32 1
-7 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-8 gpu conv fp32 11 add fp32 1 relu fp32 1
-9 gpu conv fp32 11 add fp32 1 relu fp32 1
-10 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-11 gpu conv fp32 11 add fp32 1 relu fp32 1
-12 gpu conv fp32 11 add fp32 1 relu fp32 1
-13 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-14 gpu mul fp32 11 add fp32 1 relu fp32 1
-15 gpu mul fp32 11 add fp32 1
-16 gpu softmax fp32 1
------
-+++++
-conf2 2.1225958306417145 1.9771056444390926 89.91 0.04999999999999716
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 167 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 155 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 267 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 161 add fp16 1 relu fp16 1
-12 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf3 2.090180991844805 1.9532689756636086 89.82 0.14000000000000057
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 167 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 269 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 155 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 161 add fp16 1 relu fp16 1
-12 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf4 2.169931036393396 2.0048851858669283 89.53999999999999 0.4200000000000017
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 155 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv perf_fp16 162 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 264 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 269 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf5 2.1012179398201756 1.9325098819632314 89.42 0.539999999999992
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 155 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 264 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv fp16 11 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 269 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf6 2.2313002482945326 2.069581185407626 89.38000000000001 0.5799999999999841
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 158 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 269 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv fp16 12 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 269 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 264 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf7 2.143061101834193 1.9675759235961738 89.3 0.6599999999999966
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 155 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 265 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 264 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 269 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf8 2.199379444387758 2.0314348091429677 89.2 0.7599999999999909
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 155 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 264 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf9 2.3236298452294624 2.156907976575644 89.03999999999999 0.9200000000000017
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 159 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv fp16 11 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf10 2.3224369486241603 2.1560351277882046 89.03999999999999 0.9200000000000017
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 159 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv fp16 11 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf11 2.358467412507993 2.1904290636262784 89.02 0.9399999999999977
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 159 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 266 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 264 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf12 2.3633503986583126 2.1980949050120437 88.88000000000001 1.079999999999984
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 159 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 264 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf13 2.4903388172036043 2.3063593441573564 88.82 1.1400000000000006
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf14 2.508156996742662 2.3204109539869595 88.78 1.1799999999999926
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 156 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf15 2.4818531813049622 2.2910866330696744 88.75999999999999 1.2000000000000028
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 263 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf16 2.4591564896606 2.272664410995804 88.74 1.2199999999999989
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 263 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf17 2.5370582721089496 2.3464665753522405 88.72 1.2399999999999949
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf18 2.438100014978735 2.257620696759345 88.7 1.259999999999991
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 263 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf19 2.4776935382337006 2.2949598026093168 88.7 1.259999999999991
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf20 2.4380041604279596 2.254330054479329 88.68 1.279999999999987
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 157 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf21 2.4745444350223327 2.2883888475386525 88.64 1.3199999999999932
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf22 2.4136652022060625 2.2360545757445407 88.52 1.4399999999999977
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf23 2.510093966915115 2.316437144001897 88.52 1.4399999999999977
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 266 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf24 2.475990790728594 2.28127562431577 88.5 1.4599999999999937
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv fp16 11 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf25 2.4761929121466926 2.290365501363375 88.5 1.4599999999999937
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf26 2.4763575559033875 2.291312348847263 88.5 1.4599999999999937
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf27 2.600249602991055 2.4123747341424644 88.06 1.8999999999999915
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 165 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf28 2.596077615026303 2.4115375655840245 88.02 1.9399999999999977
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 166 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf29 2.580888020555937 2.3840829703999833 87.88 2.0799999999999983
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv samp_fp16 269 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf30 2.556352783745439 2.3641413704751537 87.8 2.1599999999999966
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv samp_fp16 269 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf31 2.5559756082494527 2.3677471703724575 87.78 2.1799999999999926
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 11 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf32 2.597413373332546 2.4091972878097585 87.76 2.1999999999999886
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf33 2.4797467027434656 2.2874608793842612 87.74 2.219999999999999
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf34 2.593675604602072 2.400513932866452 87.7 2.259999999999991
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 156 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 264 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf35 2.6300759173431336 2.432687374579977 87.62 2.339999999999989
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 266 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf36 2.5907083037103864 2.4042762580264356 87.6 2.3599999999999994
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 156 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf37 2.6143261650366187 2.423427684623993 87.6 2.3599999999999994
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 156 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf38 2.6144436259117203 2.4231961521843344 87.6 2.3599999999999994
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 156 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf39 2.662088796913144 2.4660859696742032 87.6 2.3599999999999994
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf40 2.6210428708834517 2.423389791646294 87.58 2.3799999999999955
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 156 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 265 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf41 2.6399924349243533 2.4443864221157914 87.58 2.3799999999999955
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 156 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf42 2.616443708384916 2.4217582570150697 87.58 2.3799999999999955
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf43 2.6883473596205225 2.5036952786284137 87.5 2.4599999999999937
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv perf_fp16 166 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 156 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf44 2.6117356623585875 2.420771216556161 87.48 2.4799999999999898
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 156 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf45 2.6359174040106708 2.444231592562593 87.48 2.4799999999999898
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf46 2.56504192294198 2.371871906722655 87.44 2.519999999999996
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv fp16 11 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf47 2.5652588453899727 2.3816996471861174 87.44 2.519999999999996
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf48 2.68806951500876 2.5007647690311425 87.14 2.819999999999993
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv perf_fp16 166 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv perf_fp16 156 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/vgg16_cifar100/vgg16_cifar100.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/vgg16_cifar100/vgg16_cifar100.txt
deleted file mode 100644
index 2c29bedd09..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/vgg16_cifar100/vgg16_cifar100.txt
+++ /dev/null
@@ -1,970 +0,0 @@
-3768.819777999999
-+++++
-conf1 1 1 66.5 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 relu fp32 1
-4 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-5 gpu conv fp32 11 add fp32 1 relu fp32 1
-6 gpu conv fp32 11 add fp32 1 relu fp32 1
-7 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-8 gpu conv fp32 11 add fp32 1 relu fp32 1
-9 gpu conv fp32 11 add fp32 1 relu fp32 1
-10 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-11 gpu conv fp32 11 add fp32 1 relu fp32 1
-12 gpu conv fp32 11 add fp32 1 relu fp32 1
-13 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-14 gpu mul fp32 11 add fp32 1 relu fp32 1
-15 gpu mul fp32 11 add fp32 1
-16 gpu softmax fp32 1
------
-+++++
-conf2 2.2877724452131787 2.08025704453875 66.45 0.04999999999999716
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 153 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 269 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 268 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 162 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 266 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf3 2.5314658805383816 2.30737681453141 66.45 0.04999999999999716
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf4 2.044123178914057 1.8616966918258782 66.32000000000001 0.1799999999999926
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 168 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 268 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 11 add fp16 1 relu fp16 1
-13 gpu conv fp16 11 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf5 2.231179358259141 2.0317825813373864 66.18 0.3199999999999932
-1 gpu conv fp16 11 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 161 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv fp16 12 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 269 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 265 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf6 2.2474834421641057 2.0338639876373272 65.88000000000001 0.6199999999999903
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 268 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 267 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 265 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 267 add fp16 1 relu fp16 1
-13 gpu conv fp16 11 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf7 2.22281439516094 2.0205460706906377 65.88000000000001 0.6199999999999903
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 155 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 268 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 161 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 167 add fp16 1 relu fp16 1
-12 gpu conv perf_fp16 161 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf8 2.1625085012968484 1.94560449637282 65.88000000000001 0.6199999999999903
-1 gpu conv fp16 11 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 266 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv fp16 11 add fp16 1 relu fp16 1
-10 gpu conv fp16 11 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 264 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 263 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf9 2.639337323402163 2.3960416499256825 65.8 0.7000000000000028
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 269 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf10 2.672718090670276 2.4276905528801507 65.68 0.8199999999999932
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf11 2.699089631751789 2.446114054498494 65.68 0.8199999999999932
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf12 2.6003752638648767 2.3553067802112344 65.64 0.8599999999999994
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv fp16 11 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf13 2.638763904718665 2.395072565223988 65.64 0.8599999999999994
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 268 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf14 2.6003752638648767 2.3553067802112344 65.64 0.8599999999999994
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv fp16 11 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf15 2.6003752638648767 2.3553067802112344 65.64 0.8599999999999994
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv fp16 11 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf16 2.6732183804279006 2.4287517162140326 65.62 0.8799999999999955
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf17 2.6728394017929027 2.428768169588016 65.60000000000001 0.8999999999999915
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf18 2.4549989178389238 2.2406620346549433 65.56 0.9399999999999977
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 156 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf19 2.673556689244081 2.429092581627209 65.52 0.980000000000004
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf20 2.6525635304451756 2.406830663552284 65.5 1.0
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 263 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf21 2.6692288605087553 2.423462800937785 65.5 1.0
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf22 2.583650505571873 2.3471533059252194 65.48 1.019999999999996
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 263 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf23 2.6474572655420125 2.400471260394867 65.48 1.019999999999996
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 265 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf24 2.4710116424304736 2.2555966923178996 65.46 1.0400000000000063
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 161 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 266 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf25 2.557911102074785 2.3292661683311526 65.46 1.0400000000000063
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 156 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf26 2.6032957018479532 2.367574146141511 65.44 1.0600000000000023
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 163 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf27 2.6029968728098916 2.3672068592437223 65.44 1.0600000000000023
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf28 2.602540311129756 2.3691028781436954 65.44 1.0600000000000023
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 167 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf29 2.602756708588441 2.3708111025211718 65.44 1.0600000000000023
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 168 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf30 2.603240857443844 2.3662875785790183 65.44 1.0600000000000023
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 157 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf31 2.602882717372841 2.368011704225619 65.44 1.0600000000000023
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf32 2.67999343314603 2.4305182001043826 65.4 1.0999999999999943
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf33 2.670314990364046 2.4275308713267485 65.38000000000001 1.1199999999999903
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf34 2.650982630033638 2.405821467700663 65.36 1.1400000000000006
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 263 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf35 2.6507266317871756 2.405938171802741 65.36 1.1400000000000006
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 265 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf36 2.6523068534836174 2.406695716686769 65.34 1.1599999999999966
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 264 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf37 2.6533198495191073 2.4077689394073865 65.34 1.1599999999999966
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 264 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf38 2.64630900155657 2.4073892305914986 65.32 1.1800000000000068
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 152 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf39 2.6725522534379413 2.42903505877629 65.32 1.1800000000000068
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf40 2.6435249267602225 2.403536258709464 65.3 1.2000000000000028
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 161 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf41 2.6442059720503557 2.4037376163252024 65.3 1.2000000000000028
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf42 2.6536933126724027 2.4077527693156053 65.3 1.2000000000000028
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 264 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf43 2.6442798101298948 2.4056031584129225 65.3 1.2000000000000028
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf44 2.603921271336049 2.3665955131107683 65.28 1.2199999999999989
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 157 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf45 2.4967248028856828 2.2748997625822716 65.25999999999999 1.240000000000009
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 157 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf46 2.4963953691980665 2.2764932409573166 65.25999999999999 1.240000000000009
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 266 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf47 2.678944927989822 2.4251978482969956 65.24 1.2600000000000051
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 264 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf48 2.6727135417173904 2.428897140422096 65.22 1.2800000000000011
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf49 2.600256135586627 2.355428067042657 65.16 1.3400000000000034
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 151 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-12 gpu conv fp16 11 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf50 2.264460006128871 2.058037581586567 64.9 1.5999999999999943
-1 gpu conv fp16 11 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 155 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv perf_fp16 165 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv fp16 12 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 269 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 164 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 263 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 265 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
-+++++
-conf51 2.2817447204106736 2.0758846029697513 64.84 1.6599999999999966
-1 gpu conv fp16 11 add fp16 1 relu fp16 1
-2 gpu conv perf_fp16 155 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv perf_fp16 165 add fp16 1 relu fp16 1
-4 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv fp16 12 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 267 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-9 gpu conv perf_fp16 155 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv perf_fp16 160 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 265 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1
-16 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/vgg16_imagenet/vgg16_imagenet.txt b/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/vgg16_imagenet/vgg16_imagenet.txt
deleted file mode 100644
index 108a101c81..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/predictive/vgg16_imagenet/vgg16_imagenet.txt
+++ /dev/null
@@ -1,561 +0,0 @@
-19194.623482
-+++++
-conf1 1 1 72.84 0.0
-1 gpu conv fp32 11 add fp32 1 relu fp32 1
-2 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-3 gpu conv fp32 11 add fp32 1 relu fp32 1
-4 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-5 gpu conv fp32 11 add fp32 1 relu fp32 1
-6 gpu conv fp32 11 add fp32 1 relu fp32 1
-7 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-8 gpu conv fp32 11 add fp32 1 relu fp32 1
-9 gpu conv fp32 11 add fp32 1 relu fp32 1
-10 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-11 gpu conv fp32 11 add fp32 1 relu fp32 1
-12 gpu conv fp32 11 add fp32 1 relu fp32 1
-13 gpu conv fp32 11 add fp32 1 relu fp32 1 pool_max fp32 1
-14 gpu mul fp32 11 add fp32 1 relu fp32 1
-15 gpu mul fp32 11 add fp32 1 relu fp32 1
-16 gpu mul fp32 11 add fp32 1
-17 gpu softmax fp32 1
------
-+++++
-conf2 2.0787477568568082 1.7725701909562666 72.76 0.0799999999999983
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv fp16 12 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf3 2.2877881266029436 1.9268677640464096 72.04 0.7999999999999972
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf4 2.493698381711785 2.0336802939709626 72.02 0.8200000000000074
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf5 2.164723960411776 1.8442442134020163 71.94 0.9000000000000057
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf6 2.53794461743687 2.069640641367895 71.67999999999999 1.1600000000000108
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 156 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf7 1.7943268128686711 1.6103705347377417 71.58 1.2600000000000051
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv fp16 12 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 266 add fp16 1 relu fp16 1
-10 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv fp16 11 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf8 1.8143284638396158 1.6288620764171362 71.5 1.3400000000000034
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv fp16 12 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 266 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv fp16 11 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf9 2.5462742331906263 2.076061630349781 71.48 1.3599999999999994
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 167 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 156 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf10 2.526515422129153 2.063839193109964 71.39999999999999 1.440000000000012
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv fp16 11 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf11 2.1596661517243856 1.8351710968407349 71.34 1.5
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 267 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 268 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 156 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf12 2.3444383477958337 1.981259839350623 71.22 1.6200000000000045
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf13 1.8402020049200172 1.652343405000522 71.2 1.6400000000000006
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 266 add fp16 1 relu fp16 1
-10 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-13 gpu conv fp16 11 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf14 2.6420417968257306 2.167425635999969 71.12 1.7199999999999989
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 167 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 155 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf15 2.543198098440602 2.0805826545876145 71.1 1.740000000000009
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf16 2.6224991911009328 2.1476958232678807 70.89999999999999 1.940000000000012
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 167 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf17 2.5978010917593752 2.131515210392801 70.8 2.0400000000000063
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 157 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf18 2.623210258119482 2.156636511928761 70.76 2.0799999999999983
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf19 2.598187894495609 2.1322228990374104 70.76 2.0799999999999983
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf20 2.640464221374653 2.1682626030871295 70.76 2.0799999999999983
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 167 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf21 2.659563405662692 2.1881035849678936 70.54 2.299999999999997
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf22 2.636584103560761 2.1652496021557557 70.39999999999999 2.440000000000012
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 165 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf23 2.6315080449303547 2.161259580137757 70.38 2.460000000000008
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 162 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf24 2.7367939789033153 2.263326406058847 70.34 2.5
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 160 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf25 2.712182817327382 2.2404693918737233 70.24000000000001 2.5999999999999943
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 168 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf26 2.660510795888948 2.187299344706456 70.22 2.6200000000000045
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 159 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-9 gpu conv fp16 12 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf27 2.457573203839654 2.0936930776435383 70.1 2.740000000000009
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv fp16 12 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-10 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv samp_fp16 262 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv samp_fp16 261 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
-+++++
-conf28 2.7452293174567757 2.2593302388139347 69.92 2.9200000000000017
-1 gpu conv fp16 12 add fp16 1 relu fp16 1
-2 gpu conv samp_fp16 262 add fp16 1 relu fp16 1 pool_max fp16 1
-3 gpu conv fp16 12 add fp16 1 relu fp16 1
-4 gpu conv perf_fp16 159 add fp16 1 relu fp16 1 pool_max fp16 1
-5 gpu conv fp16 12 add fp16 1 relu fp16 1
-6 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-7 gpu conv samp_fp16 266 add fp16 1 relu fp16 1 pool_max fp16 1
-8 gpu conv fp16 12 add fp16 1 relu fp16 1
-9 gpu conv samp_fp16 261 add fp16 1 relu fp16 1
-10 gpu conv perf_fp16 152 add fp16 1 relu fp16 1 pool_max fp16 1
-11 gpu conv fp16 12 add fp16 1 relu fp16 1
-12 gpu conv fp16 12 add fp16 1 relu fp16 1
-13 gpu conv perf_fp16 151 add fp16 1 relu fp16 1 pool_max fp16 1
-14 gpu mul fp16 12 add fp16 1 relu fp16 1
-15 gpu mul fp16 12 add fp16 1 relu fp16 1
-16 gpu mul fp16 12 add fp16 1
-17 gpu softmax fp32 1
------
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/autotuner/img_pareto_curve.py b/hpvm/projects/hpvm-tensor-rt/autotuner/autotuner/img_pareto_curve.py
deleted file mode 100644
index 5192eb7e58..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/autotuner/img_pareto_curve.py
+++ /dev/null
@@ -1,169 +0,0 @@
-import os
-import shutil
-
-
-AL_THRESHOLD = 0.1
-
-
-class Config:
- def __init__(self):
- self.avg_accuracy = 0
- self.avg_loss = 0
- self.speedup = 1
- self.fname = ""
- self.flags = []
-
-
-def loadConfigsFromDir(result_dir):
- def parseTopLine(x):
-
- toks = x.split()
-
- speedup = 1.0
- for tok in toks:
- if "avg_accuracy" in tok:
- avg_accuracy = float(tok.split("=")[1])
- if "speedup" in tok:
- speedup = float(tok.split("=")[1])
- return avg_accuracy, speedup
-
- def skipFile(fname):
-
- skip_files = {}
- skip_files["confidence_summary.txt"] = 1
- skip_files["promise_confs.txt"] = 1
-
- if "accuracy" in fname: # *_accuracy files should be skipped
- return True
-
- if "norms" in fname: # *_accuracy files should be skipped
- return True
-
- if ".#" in fname: # *_accuracy files should be skipped
- return True
-
- # if "_promise" in fname: # *_accuracy files should be skipped
- # return True
-
- if not fname[-1].isdigit():
- return True
-
- if fname in skip_files:
- return True
- else:
- return False
-
- config_arr = []
- file_names = os.listdir(result_dir)
-
- for fname in file_names:
- if not skipFile(fname):
-
- fpath = result_dir + fname
- config = Config()
- f = open(fpath, "r")
-
- it = 0
- for x in f:
- if x.strip == "":
- continue
- if it == 0:
- avg_accuracy, speedup = parseTopLine(x)
- config.avg_accuracy = avg_accuracy
- config.avg_loss = -avg_accuracy
- config.speedup = speedup
- config.fname = fname
- else:
- flag = int(x.strip())
- config.flags.append(flag)
- it += 1
-
- config_arr.append(config)
-
- return config_arr
-
-
-class Configuration:
- def __init__(self, name, speedup, energy, accuracy, accuracy_loss):
- self.name = name
- self.speedup = speedup
- self.energy = energy
- self.accuracy = accuracy
- self.accuracy_loss = accuracy_loss
-
- def __repr__(self):
- return repr((self.name, self.speedup, self.energy, self.accuracy, self.accuracy_loss))
-
- @classmethod
- def from_config(cls, config):
- return cls(config.fname, config.speedup, 0, config.avg_accuracy, config.avg_loss)
-
- @staticmethod
- def energy_points(configurations):
- return [
- (conf.energy, conf.accuracy)
- for conf in configurations
- ]
-
- @staticmethod
- def speedup_points(configurations):
- return [
- (conf.speedup, conf.accuracy)
- for conf in configurations
- ]
-
-
-def is_pareto_efficient(configs, values, value_margins):
- import numpy as np
- from pprint import pprint
-
- np_values = np.array(values)
- np_margins = np.array(value_margins)
- is_efficient = np.ones(np_values.shape[0], dtype=bool)
- for i, c in enumerate(np_values):
- if is_efficient[i]:
- # Keep any point with a higher value
- is_efficient[is_efficient] = np.any(
- np_values[is_efficient] + np_margins >= c, axis=1
- )
- is_efficient[i] = True # And keep self
- return (np.array(configs)[is_efficient]).tolist()
-
-
-def findParetoConfigs(base_dir, psnr_band_size):
- result_dir = base_dir + "/pareto/"
- try:
- os.mkdir(result_dir)
- except:
- print "could not create dir"
-
- input_dir = base_dir + "/high_confidence/"
- config_arr = loadConfigsFromDir(input_dir)
- configurations = [Configuration.from_config(
- config) for config in config_arr]
- # energy_points = Configuration.energy_points(configurations)
- speedup_points = Configuration.speedup_points(configurations)
-
- # No Pareto Selection if list is < 50 configurations
- if len(configurations) < 50:
- speedup_pareto = configurations # Include all in Pareto Frontier
- else:
- # energy_pareto = is_pareto_efficient(configurations, energy_points, ...)
- speedup_pareto = is_pareto_efficient(
- configurations, speedup_points, [-1e-2, psnr_band_size]
- )
- print("len(configurations) = ", len(configurations))
- print("len(speedup_pareto) = ", len(speedup_pareto))
-
- for conf in speedup_pareto:
- #dst_path = conf.name.replace("full_results", "pareto")
- src_path = input_dir + conf.name
- dst_path = result_dir + conf.name
- shutil.copy(src_path, dst_path)
-
-
-if __name__ == "__main__":
- from sys import argv
- psnr_band_size = float(argv[2]) if len(argv) > 2 else 3.0
- print("*psnr_band_size = ", psnr_band_size)
- findParetoConfigs(argv[1], psnr_band_size)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/alexnet_layers.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/alexnet_layers.txt
deleted file mode 100644
index 5741a41ba3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/alexnet_layers.txt
+++ /dev/null
@@ -1,6 +0,0 @@
-conv add tanh pool
-conv add tanh pool
-conv add tanh
-conv add tanh
-conv add tanh pool
-dense add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/alexnet_tensors.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/alexnet_tensors.txt
deleted file mode 100644
index 2a925c986f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/alexnet_tensors.txt
+++ /dev/null
@@ -1,26 +0,0 @@
-#Conv1,4
-Conv
-Add
-Relu
-Pool
-#Conv2,4
-Conv
-Add
-Relu
-Pool
-#Conv3,3
-Conv
-Add
-Relu
-#Conv4,3
-Conv
-Add
-Relu
-#Conv5,4
-Conv
-Add
-Relu
-Pool
-#FC1,2
-Mul
-Add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/dev_knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/dev_knobs.txt
deleted file mode 100644
index bba4dc88d5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/dev_knobs.txt
+++ /dev/null
@@ -1,6 +0,0 @@
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/knobs.txt
deleted file mode 100644
index 050fc61180..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/knobs.txt
+++ /dev/null
@@ -1,6 +0,0 @@
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/knobs1.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/knobs1.txt
deleted file mode 100644
index d2fc2c9493..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/knobs1.txt
+++ /dev/null
@@ -1,6 +0,0 @@
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/op_cost.txt
deleted file mode 100644
index 04336fca27..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet/op_cost.txt
+++ /dev/null
@@ -1,6 +0,0 @@
-11894784.000000
-39321600.000000
-21233664.000000
-28311552.000000
-18874368.000000
-20480.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/alexnet2_layers.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/alexnet2_layers.txt
deleted file mode 100644
index 00059a38ce..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/alexnet2_layers.txt
+++ /dev/null
@@ -1,7 +0,0 @@
-conv add tanh
-conv add tanh pool
-conv add tanh
-conv add tanh pool
-conv add tanh
-conv add tanh pool
-dense add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/alexnet2_tensors.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/alexnet2_tensors.txt
deleted file mode 100644
index 747c7221ba..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/alexnet2_tensors.txt
+++ /dev/null
@@ -1,30 +0,0 @@
-#Conv1,3
-Conv
-Add
-Relu
-#Conv2,4
-Conv
-Add
-Relu
-Pool
-#Conv3,3
-Conv
-Add
-Relu
-#Conv4,4
-Conv
-Add
-Relu
-Pool
-#Conv5,3
-Conv
-Add
-Relu
-#Conv6,4
-Conv
-Add
-Relu
-Pool
-#FC1,2
-Mul
-Add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/dev_knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/dev_knobs.txt
deleted file mode 100644
index 0324aecdca..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/dev_knobs.txt
+++ /dev/null
@@ -1,7 +0,0 @@
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/knobs.txt
deleted file mode 100644
index c873eeddcd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/knobs.txt
+++ /dev/null
@@ -1,7 +0,0 @@
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/knobs1.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/knobs1.txt
deleted file mode 100644
index 063ba473d6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/knobs1.txt
+++ /dev/null
@@ -1,7 +0,0 @@
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/op_cost.txt
deleted file mode 100644
index 5a5722f202..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2/op_cost.txt
+++ /dev/null
@@ -1,7 +0,0 @@
-88473.601562
-943718.375000
-471859.187500
-943718.375000
-471859.187500
-943718.375000
-2048.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2_canny_hpvm/layers.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2_canny_hpvm/layers.txt
deleted file mode 100644
index 01f40077d4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2_canny_hpvm/layers.txt
+++ /dev/null
@@ -1,13 +0,0 @@
-conv add tanh
-conv add tanh pool
-conv add tanh
-conv add tanh pool
-conv add tanh
-conv add tanh pool
-dense add
-reduce
-conv
-conv
-conv
-reduce
-reduce
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2_canny_hpvm/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2_canny_hpvm/op_cost.txt
deleted file mode 100644
index 80ff2706a4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet2_canny_hpvm/op_cost.txt
+++ /dev/null
@@ -1,13 +0,0 @@
-468.076
-947.434
-255.422
-348.769
-256.658
-1.05427
-1.05427
-107.5062
-666.888
-432.622
-252.458
-11.51922
-2.01168
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet_imagenet/layer_composition.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet_imagenet/layer_composition.txt
deleted file mode 100644
index b2bf962cd6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet_imagenet/layer_composition.txt
+++ /dev/null
@@ -1,8 +0,0 @@
-conv add activation pool
-conv add activation pool
-conv add activation
-conv add activation
-conv add activation pool
-dense add activation
-dense add activation
-dense add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet_imagenet/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet_imagenet/op_cost.txt
deleted file mode 100644
index ec3b8b5f37..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet_imagenet/op_cost.txt
+++ /dev/null
@@ -1,8 +0,0 @@
-1457111.000000
-4478976.000000
-2242805.750000
-2990407.750000
-1993605.125000
-754974.750000
-335544.312500
-81920.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet_imagenet/quant_ranges2.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet_imagenet/quant_ranges2.txt
deleted file mode 100644
index 36c9c390b5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/alexnet_imagenet/quant_ranges2.txt
+++ /dev/null
@@ -1,8 +0,0 @@
-0.0 255.0 -0.5503702693581581 0.5811487324237921 -2.802485 1.648145 0.0 1572.3096923828125
-0.0 1572.3096923828125 -0.2867645202279091 0.26272463005783797 -0.47985682 0.501206 0.0 3183.7813264160477
-0.0 3183.7813264160477 -0.16606662392616273 0.15785247704386754 -0.42038992 0.5545839 0.0 1765.4451872558668
-0.0 1765.4451872558668 -0.10464580833911895 0.11035470351576919 -1.4275751 0.9042998 0.0 1345.5418548586083
-0.0 1345.5418548586083 -0.09240880391001702 0.10250756608694818 -0.45662758 2.4040315 0.0 1227.3563232421875
-0.0 1227.3563232421875 -0.030517672039568428 0.02963459612801672 -0.07124679 0.09377053 0.0 1034.5966391601676
-0.0 1034.5966391601676 -0.038392101023346184 0.039147199764847845 -0.050027702 0.1841282 0.0 839.0697069702154
-0.0 839.0697069702154 -0.05494491942599416 0.08549865524470925 -0.16314922 0.15416704 -608.3993963623047 1082.8444653320819
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/blend/knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/blend/knobs.txt
deleted file mode 100644
index 3bcc8d25cc..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/blend/knobs.txt
+++ /dev/null
@@ -1,17 +0,0 @@
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,41,43,45
-11,41,43,45
-11
-11,41,43,45
-11,41,43,45
-11
-11
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,41,43,45
-11,41,43,45
-11
-11,41,43,45
-11,41,43,45
-11
-11,12
-11
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/blend/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/blend/op_cost.txt
deleted file mode 100644
index 9575b4b7c6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/blend/op_cost.txt
+++ /dev/null
@@ -1,17 +0,0 @@
-4425.6
-112.809
-10.9522
-675.831
-113.067
-11.4471
-674.881
-686.27
-4334.16
-112.544
-11.1993
-674.836
-112.973
-11.7161
-675.575
-685.598
-686.733
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/blend/tuner_conf_template.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/blend/tuner_conf_template.txt
deleted file mode 100644
index 4cd90c098b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/blend/tuner_conf_template.txt
+++ /dev/null
@@ -1,20 +0,0 @@
-+++++
-conf1 1.5 90 1.0 2.0
-1 gpu conv fp32 1
-2 gpu reduce fp32 1
-3 gpu reduce fp32 1
-4 gpu map2 fp32 1
-5 gpu reduce fp32 1
-6 gpu reduce fp32 1
-7 gpu map2 fp32 1
-8 gpu map2 fp32 1
-9 gpu conv fp32 1
-10 gpu reduce fp32 1
-11 gpu reduce fp32 1
-12 gpu map2 fp32 1
-13 gpu reduce fp32 1
-14 gpu reduce fp32 1
-15 gpu map2 fp32 1
-16 gpu map2 fp32 1
-17 gpu map2 fp32 1
------
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/canny/knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/canny/knobs.txt
deleted file mode 100644
index bd4820f412..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/canny/knobs.txt
+++ /dev/null
@@ -1,9 +0,0 @@
-11,41,43,45
-11
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11
-11,41,43,45
-11,41,43,45
-11
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/canny/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/canny/op_cost.txt
deleted file mode 100644
index 64e2c8bec1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/canny/op_cost.txt
+++ /dev/null
@@ -1,9 +0,0 @@
-87.1338
-10.9506
-703.070
-520.145
-335.735
-19.1877
-9.27351
-2.46408
-17.7423
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/canny/tuner_conf_template.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/canny/tuner_conf_template.txt
deleted file mode 100644
index 17b30a46e0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/canny/tuner_conf_template.txt
+++ /dev/null
@@ -1,12 +0,0 @@
-+++++
-conf1 1.5 90 1.0 2.0
-1 gpu reduce fp16 1
-2 gpu map1 fp16 1
-3 gpu conv fp16 1
-4 gpu conv fp16 1
-5 gpu conv fp16 1
-6 gpu map2 fp16 1
-7 gpu reduce fp16 1
-8 gpu reduce fp16 1
-9 gpu map2 fp16 1
------
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/fft/knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/fft/knobs.txt
deleted file mode 100644
index 15a06d67cb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/fft/knobs.txt
+++ /dev/null
@@ -1,4 +0,0 @@
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,42,44,46
-12,42,44,46
-12
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/fft/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/fft/op_cost.txt
deleted file mode 100644
index 910bcf3579..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/fft/op_cost.txt
+++ /dev/null
@@ -1,4 +0,0 @@
-1672.868
-198.9254
-9876.189
-28158.14
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/fft/tuner_conf_template.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/fft/tuner_conf_template.txt
deleted file mode 100644
index 49cee56932..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/fft/tuner_conf_template.txt
+++ /dev/null
@@ -1,7 +0,0 @@
-+++++
-conf1 1.5 90 1.0 2.0
-1 gpu conv fp32 1
-2 gpu reduce fp32 1
-3 gpu reduce fp32 1
-4 gpu map2 fp32 1
------
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/global_knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/global_knobs.txt
deleted file mode 100644
index ee2cd80cb6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/global_knobs.txt
+++ /dev/null
@@ -1,69 +0,0 @@
-fp32,11 -1 1.0 tensorConvolution tensorConvApprox dev conv_fc_red
-fp16,12 -1 1.5 tensorConvolution tensorConvApproxHalf2 install conv_fc_red
-perf,121 1,2,0 2.0 tensorConvolution tensorConvApprox dev conv
-perf,122 1,2,1 2.0 tensorConvolution tensorConvApprox dev conv
-perf,123 2,1,0 2.0 tensorConvolution tensorConvApprox dev conv
-perf,124 2,1,1 2.0 tensorConvolution tensorConvApprox dev conv
-perf,125 1,3,0 1.5 tensorConvolution tensorConvApprox dev conv
-perf,126 1,3,1 1.5 tensorConvolution tensorConvApprox dev conv
-perf,127 1,3,2 1.5 tensorConvolution tensorConvApprox dev conv
-perf,128 3,1,0 1.5 tensorConvolution tensorConvApprox dev conv
-perf,129 3,1,1 1.5 tensorConvolution tensorConvApprox dev conv
-perf,130 3,1,2 1.5 tensorConvolution tensorConvApprox dev conv
-perf,131 1,4,0 1.33 tensorConvolution tensorConvApprox dev conv
-perf,132 1,4,1 1.33 tensorConvolution tensorConvApprox dev conv
-perf,133 1,4,2 1.33 tensorConvolution tensorConvApprox dev conv
-perf,134 1,4,3 1.33 tensorConvolution tensorConvApprox dev conv
-perf,135 4,1,0 1.33 tensorConvolution tensorConvApprox dev conv
-perf,136 4,1,1 1.33 tensorConvolution tensorConvApprox dev conv
-perf,137 4,1,2 1.33 tensorConvolution tensorConvApprox dev conv
-perf,138 4,1,3 1.33 tensorConvolution tensorConvApprox dev conv
-perf_fp16,151 1,2,0 3.0 tensorConvolution tensorConvApprox install conv
-perf_fp16,152 1,2,1 3.0 tensorConvolution tensorConvApprox install conv
-perf_fp16,153 2,1,0 3.0 tensorConvolution tensorConvApprox install conv
-perf_fp16,154 2,1,1 3.0 tensorConvolution tensorConvApprox install conv
-perf_fp16,155 1,3,0 2.25 tensorConvolution tensorConvApprox install conv
-perf_fp16,156 1,3,1 2.25 tensorConvolution tensorConvApprox install conv
-perf_fp16,157 1,3,2 2.25 tensorConvolution tensorConvApprox install conv
-perf_fp16,158 3,1,0 2.25 tensorConvolution tensorConvApprox install conv
-perf_fp16,159 3,1,1 2.25 tensorConvolution tensorConvApprox install conv
-perf_fp16,160 3,1,2 2.25 tensorConvolution tensorConvApprox install conv
-perf_fp16,161 1,4,0 2.0 tensorConvolution tensorConvApprox install conv
-perf_fp16,162 1,4,1 2.0 tensorConvolution tensorConvApprox install conv
-perf_fp16,163 1,4,2 2.0 tensorConvolution tensorConvApprox install conv
-perf_fp16,164 1,4,3 2.0 tensorConvolution tensorConvApprox install conv
-perf_fp16,165 4,1,0 2.0 tensorConvolution tensorConvApprox install conv
-perf_fp16,166 4,1,1 2.0 tensorConvolution tensorConvApprox install conv
-perf_fp16,167 4,1,2 2.0 tensorConvolution tensorConvApprox install conv
-perf_fp16,168 4,1,3 2.0 tensorConvolution tensorConvApprox install conv
-samp,231 2,0,1 2.0 tensorConvolution tensorConvApprox dev conv
-samp,232 2,1,1 2.0 tensorConvolution tensorConvApprox dev conv
-samp,233 3,0,1 1.5 tensorConvolution tensorConvApprox dev conv
-samp,234 3,1,1 1.5 tensorConvolution tensorConvApprox dev conv
-samp,235 3,2,1 1.5 tensorConvolution tensorConvApprox dev conv
-samp,236 4,0,1 1.33 tensorConvolution tensorConvApprox dev conv
-samp,237 4,1,1 1.33 tensorConvolution tensorConvApprox dev conv
-samp,238 4,2,1 1.33 tensorConvolution tensorConvApprox dev conv
-samp,239 4,3,1 1.33 tensorConvolution tensorConvApprox dev conv
-samp_fp16,261 2,0,1 3.0 tensorConvolution tensorConvApprox install conv
-samp_fp16,262 2,1,1 3.0 tensorConvolution tensorConvApprox install conv
-samp_fp16,263 3,0,1 2.25 tensorConvolution tensorConvApprox install conv
-samp_fp16,264 3,1,1 2.25 tensorConvolution tensorConvApprox install conv
-samp_fp16,265 3,2,1 2.25 tensorConvolution tensorConvApprox install conv
-samp_fp16,266 4,0,1 2.0 tensorConvolution tensorConvApprox install conv
-samp_fp16,267 4,1,1 2.0 tensorConvolution tensorConvApprox install conv
-samp_fp16,268 4,2,1 2.0 tensorConvolution tensorConvApprox install conv
-samp_fp16,269 4,3,1 2.0 tensorConvolution tensorConvApprox install conv
-red_samp,41 1 1.5 tensorReduction tensorReduction dev red
-red_samp,42 1 2.25 tensorReduction tensorReduction dev red
-red_samp,43 1 1.4 tensorReduction tensorReduction dev red
-red_samp,44 1 2 tensorReduction tensorReduction dev red
-red_samp,45 1 1.25 tensorReduction tensorReduction dev red
-red_samp,46 1 1.8 tensorReduction tensorReduction dev red
-swing_level,1 1 12 tensorConvolution tensorConvApprox install conv_fc
-swing_level,2 1 10 tensorConvolution tensorConvApprox install conv_fc
-swing_level,3 1 9 tensorConvolution tensorConvApprox install conv_fc
-swing_level,4 1 8 tensorConvolution tensorConvApprox install conv_fc
-swing_level,5 1 6 tensorConvolution tensorConvApprox install conv_fc
-swing_level,6 1 5 tensorConvolution tensorConvApprox install conv_fc
-swing_level,7 1 4 tensorConvolution tensorConvApprox install conv_fc
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/global_knobs_dnn.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/global_knobs_dnn.txt
deleted file mode 100644
index 2180997527..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/global_knobs_dnn.txt
+++ /dev/null
@@ -1,25 +0,0 @@
-fp32,11 -1 1.0 tensorConvolution tensorConvolution
-fp16,12 -1 1.5 tensorConvolution tensorHalfConvolution
-perf,21 1,2,0 2.25 tensorConvolution tensorConvPerfCuda
-perf,22 1,2,1 2.25 tensorConvolution tensorConvPerfCuda
-perf,23 1,3,0 1.88 tensorConvolution tensorConvPerfCuda
-perf,24 1,3,1 1.88 tensorConvolution tensorConvPerfCuda
-perf,25 1,3,2 1.88 tensorConvolution tensorConvPerfCuda
-perf,26 2,1,0 2.25 tensorConvolution tensorConvPerfCuda
-perf,27 2,1,1 2.25 tensorConvolution tensorConvPerfCuda
-perf,28 3,1,0 1.88 tensorConvolution tensorConvPerfCuda
-perf,29 3,1,1 1.88 tensorConvolution tensorConvPerfCuda
-perf,30 3,1,2 1.88 tensorConvolution tensorConvPerfCuda
-samp,31 2,0 2.25 tensorConvolution tensorConvSampSim
-samp,32 2,1 2.25 tensorConvolution tensorConvSampSim
-samp,33 4,0 1.8 tensorConvolution tensorConvSampSim
-samp,34 4,1 1.8 tensorConvolution tensorConvSampSim
-samp,35 4,2 1.8 tensorConvolution tensorConvSampSim
-samp,36 4,3 1.8 tensorConvolution tensorConvSampSim
-swing_level,1 1 12
-swing_level,2 1 10
-swing_level,3 1 9
-swing_level,4 1 8
-swing_level,5 1 6
-swing_level,6 1 5
-swing_level,7 1 4
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/global_knobs_old.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/global_knobs_old.txt
deleted file mode 100644
index c632abbd47..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/global_knobs_old.txt
+++ /dev/null
@@ -1,31 +0,0 @@
-fp32,11 -1 1.0 tensorConvolution tensorConvolution
-fp16,12 -1 1.5 tensorConvolution tensorHalfConvolution
-perf,21 1,2,0 2.0 tensorConvolution tensorConvPerfCuda
-perf,22 1,2,1 2.0 tensorConvolution tensorConvPerfCuda
-perf,23 1,3,0 1.5 tensorConvolution tensorConvPerfCuda
-perf,24 1,3,1 1.5 tensorConvolution tensorConvPerfCuda
-perf,25 1,3,2 1.5 tensorConvolution tensorConvPerfCuda
-perf,26 2,1,0 2.0 tensorConvolution tensorConvPerfCuda
-perf,27 2,1,1 2.0 tensorConvolution tensorConvPerfCuda
-perf,28 3,1,0 1.5 tensorConvolution tensorConvPerfCuda
-perf,29 3,1,1 1.5 tensorConvolution tensorConvPerfCuda
-perf,30 3,1,2 1.5 tensorConvolution tensorConvPerfCuda
-samp,31 2,0 2.0 tensorConvolution tensorConvSampSim
-samp,32 2,1 2.0 tensorConvolution tensorConvSampSim
-samp,33 4,0 1.5 tensorConvolution tensorConvSampSim
-samp,34 4,1 1.5 tensorConvolution tensorConvSampSim
-samp,35 4,2 1.5 tensorConvolution tensorConvSampSim
-samp,36 4,3 1.5 tensorConvolution tensorConvSampSim
-reduction_samp,41 1 1.5
-reduction_samp,42 1 2.25
-reduction_samp,43 1 1.4
-reduction_samp,44 1 2
-reduction_samp,45 1 1.25
-reduction_samp,46 1 1.8
-swing_level,1 1 12
-swing_level,2 1 10
-swing_level,3 1 9
-swing_level,4 1 8
-swing_level,5 1 6
-swing_level,6 1 5
-swing_level,7 1 4
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/dev_knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/dev_knobs.txt
deleted file mode 100644
index 94b9e6ebd3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/dev_knobs.txt
+++ /dev/null
@@ -1,4 +0,0 @@
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11
-11
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/knobs.txt
deleted file mode 100644
index 8973c89f7a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/knobs.txt
+++ /dev/null
@@ -1,4 +0,0 @@
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12
-12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/knobs1.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/knobs1.txt
deleted file mode 100644
index be1ce58c95..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/knobs1.txt
+++ /dev/null
@@ -1,4 +0,0 @@
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12
-11,12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/lenet_layers.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/lenet_layers.txt
deleted file mode 100644
index 5c28aa6dca..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/lenet_layers.txt
+++ /dev/null
@@ -1,4 +0,0 @@
-conv add pool tanh
-conv add pool tanh
-dense add tanh
-dense add tanh
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/lenet_tensors.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/lenet_tensors.txt
deleted file mode 100644
index f26403376f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/lenet_tensors.txt
+++ /dev/null
@@ -1,18 +0,0 @@
-#Conv1,4
-Conv
-Add
-Relu
-Pool
-#Conv2,4
-Conv
-Add
-Relu
-Pool
-#FC1,3
-Mul
-Add
-Relu
-#FC1,3
-Mul
-Add
-Relu
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/op_cost.txt
deleted file mode 100644
index 74b1b668e2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/lenet/op_cost.txt
+++ /dev/null
@@ -1,4 +0,0 @@
-62720.000000
-1003520.000000
-321126.406250
-1024.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/dev_knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/dev_knobs.txt
deleted file mode 100644
index 7e8de16a08..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/dev_knobs.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/knobs.txt
deleted file mode 100644
index 900ad3944d..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/knobs.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/knobs1.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/knobs1.txt
deleted file mode 100644
index 6719acb97a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/knobs1.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/mobilenet_layer_comp.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/mobilenet_layer_comp.txt
deleted file mode 100644
index adcfbfed53..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/mobilenet_layer_comp.txt
+++ /dev/null
@@ -1,83 +0,0 @@
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-pool_mean
-dense add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/mobilenet_ops.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/mobilenet_ops.txt
deleted file mode 100644
index 6481664b86..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/mobilenet_ops.txt
+++ /dev/null
@@ -1,167 +0,0 @@
-#Conv1,1
-Conv1
-#NML1,1
-BatchNorm1
-#NML2,1
-Relu1
-#NML3,1
-Conv2
-#NML4,1
-BatchNorm2
-#NML5,1
-Relu2
-#Conv3,1
-Conv3
-#NML6,1
-BatchNorm3
-#NML7,1
-Relu3
-#NML8,1
-Conv4
-#NML9,1
-BatchNorm4
-#NML10,1
-Relu4
-#Conv5,1
-Conv5
-#NML11,1
-BatchNorm5
-#NML12,1
-Relu5
-#NML13,1
-Conv6
-#NML14,1
-BatchNorm6
-#NML15,1
-Relu6
-#Conv7,1
-Conv7
-#NML16,1
-BatchNorm7
-#NML17,1
-Relu7
-#NML18,1
-Conv8
-#NML19,1
-BatchNorm8
-#NML20,1
-Relu8
-#Conv9,1
-Conv9
-#NML21,1
-BatchNorm9
-#NML22,1
-Relu9
-#NML23,1
-Conv10
-#NML24,1
-BatchNorm10
-#NML25,1
-Relu10
-#Conv11,1
-Conv11
-#NML26,1
-BatchNorm11
-#NML27,1
-Relu11
-#NML28,1
-Conv12
-#NML29,1
-BatchNorm12
-#NML30,1
-Relu12
-#Conv13,1
-Conv13
-#NML31,1
-BatchNorm13
-#NML32,1
-Relu13
-#NML33,1
-Conv14
-#NML34,1
-BatchNorm14
-#NML35,1
-Relu14
-#Conv15,1
-Conv15
-#NML36,1
-BatchNorm15
-#NML37,1
-Relu15
-#NML38,1
-Conv16
-#NML39,1
-BatchNorm16
-#NML40,1
-Relu16
-#Conv17,1
-Conv17
-#NML41,1
-BatchNorm17
-#NML42,1
-Relu17
-#NML43,1
-Conv18
-#NML44,1
-BatchNorm18
-#NML45,1
-Relu18
-#Conv19,1
-Conv19
-#NML46,1
-BatchNorm19
-#NML47,1
-Relu19
-#NML48,1
-Conv20
-#NML49,1
-BatchNorm20
-#NML50,1
-Relu20
-#Conv21,1
-Conv21
-#NML51,1
-BatchNorm21
-#NML52,1
-Relu21
-#NML53,1
-Conv22
-#NML54,1
-BatchNorm22
-#NML55,1
-Relu22
-#Conv23,1
-Conv23
-#NML56,1
-BatchNorm23
-#NML57,1
-Relu23
-#NML58,1
-Conv24
-#NML59,1
-BatchNorm24
-#NML60,1
-Relu24
-#Conv25,1
-Conv25
-#NML61,1
-BatchNorm25
-#NML62,1
-Relu25
-#NML63,1
-Conv26
-#NML64,1
-BatchNorm26
-#NML65,1
-Relu26
-#Conv27,1
-Conv27
-#NML66,1
-BatchNorm27
-#NML67,1
-Relu27
-#NML68,1
-Pool1
-#FC1,2
-Mul1
-Add1
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/op_cost.txt
deleted file mode 100644
index 673e704b7e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet/op_cost.txt
+++ /dev/null
@@ -1,28 +0,0 @@
-88473.601562
-29491.199219
-209715.203125
-14745.599609
-209715.203125
-29491.199219
-419430.406250
-7372.799805
-209715.203125
-14745.599609
-419430.406250
-3686.399902
-209715.203125
-7372.799805
-419430.406250
-7372.799805
-419430.406250
-7372.799805
-419430.406250
-7372.799805
-419430.406250
-7372.799805
-419430.406250
-1843.199951
-209715.203125
-3686.399902
-419430.406250
-1024.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/dev_knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/dev_knobs.txt
deleted file mode 100644
index 9b93811d3b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/dev_knobs.txt
+++ /dev/null
@@ -1,8 +0,0 @@
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/knobs.txt
deleted file mode 100644
index c7273f3fc6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/knobs.txt
+++ /dev/null
@@ -1,8 +0,0 @@
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/knobs1.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/knobs1.txt
deleted file mode 100644
index 719d96e481..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/knobs1.txt
+++ /dev/null
@@ -1,8 +0,0 @@
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/mobilenet_shallow_layer_comp.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/mobilenet_shallow_layer_comp.txt
deleted file mode 100644
index 8ba22c7e6f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/mobilenet_shallow_layer_comp.txt
+++ /dev/null
@@ -1,41 +0,0 @@
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-pool_mean
-dense add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/mobilenet_shallow_ops.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/mobilenet_shallow_ops.txt
deleted file mode 100644
index ee6dcf89d9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/mobilenet_shallow_ops.txt
+++ /dev/null
@@ -1,47 +0,0 @@
-#Conv1,1
-Conv1
-#NML1,1
-BatchNorm1
-#NML2,1
-Relu1
-#NML3,1
-Conv2
-#NML4,1
-BatchNorm2
-#NML5,1
-Relu2
-#Conv3,1
-Conv3
-#NML6,1
-BatchNorm3
-#NML7,1
-Relu3
-#NML8,1
-Conv4
-#NML9,1
-BatchNorm4
-#NML10,1
-Relu4
-#Conv5,1
-Conv5
-#NML11,1
-BatchNorm5
-#NML12,1
-Relu5
-#NML13,1
-Conv6
-#NML14,1
-BatchNorm6
-#NML15,1
-Relu6
-#Conv7,1
-Conv7
-#NML16,1
-BatchNorm7
-#NML17,1
-Relu7
-#NML18,1
-Pool1
-#FC1,2
-Mul1
-Add1
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/op_cost.txt
deleted file mode 100644
index 7266441905..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_shallow/op_cost.txt
+++ /dev/null
@@ -1,8 +0,0 @@
-265420.812500
-629145.625000
-629145.625000
-1258291.250000
-629145.625000
-1258291.250000
-629145.625000
-6144.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_torch/layers.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_torch/layers.txt
deleted file mode 100644
index a93fac1dae..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_torch/layers.txt
+++ /dev/null
@@ -1,81 +0,0 @@
-conv
-batchnorm
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-depthwise_conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-dense add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_torch/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_torch/op_cost.txt
deleted file mode 100644
index 44d50dbe00..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/mobilenet_torch/op_cost.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-44236.80078
-104857.6019
-104857.6019
-209715.2037
-104857.6019
-209715.2037
-104857.6019
-209715.2037
-209715.2037
-209715.2037
-209715.2037
-209715.2037
-104857.6019
-209715.2037
-256.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/dev_knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/dev_knobs.txt
deleted file mode 100644
index 2a07f89372..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/dev_knobs.txt
+++ /dev/null
@@ -1,22 +0,0 @@
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/knobs.txt
deleted file mode 100644
index eadcb5ebff..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/knobs.txt
+++ /dev/null
@@ -1,22 +0,0 @@
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/knobs1.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/knobs1.txt
deleted file mode 100644
index b7ff033cec..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/knobs1.txt
+++ /dev/null
@@ -1,22 +0,0 @@
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/knobs2.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/knobs2.txt
deleted file mode 100644
index ec3e26a51f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/knobs2.txt
+++ /dev/null
@@ -1,22 +0,0 @@
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/op_cost.txt
deleted file mode 100644
index fdba070cfc..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/op_cost.txt
+++ /dev/null
@@ -1,22 +0,0 @@
-44236.800781
-235929.593750
-235929.593750
-235929.593750
-235929.593750
-235929.593750
-235929.593750
-117964.796875
-235929.593750
-13107.200195
-235929.593750
-235929.593750
-235929.593750
-235929.593750
-117964.796875
-235929.593750
-13107.200195
-235929.593750
-235929.593750
-235929.593750
-235929.593750
-64.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/resnet_layers.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/resnet_layers.txt
deleted file mode 100644
index 43f0024925..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/resnet_layers.txt
+++ /dev/null
@@ -1,41 +0,0 @@
-conv add activation
-conv add activation
-conv add
-add
-activation
-conv add activation
-conv add
-add
-activation
-conv add activation
-conv add
-add
-activation
-conv add activation
-conv add
-conv add
-add
-activation
-conv add activation
-conv add
-add
-activation
-conv add activation
-conv add
-add
-activation
-conv add activation
-conv add
-conv add
-add
-activation
-conv add activation
-conv add
-add
-activation
-conv add activation
-conv add
-add
-activation
-pool_mean
-dense add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/resnet_tensors.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/resnet_tensors.txt
deleted file mode 100644
index ee0e8456da..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet/resnet_tensors.txt
+++ /dev/null
@@ -1,114 +0,0 @@
-#Conv1,3
-Conv
-Add
-Relu
-#Conv2,3
-Conv
-Add
-Relu
-#Conv3,2
-Conv
-Add
-#NML1,1
-Add
-#NML2,1
-Relu
-#Conv4,3
-Conv
-Add
-Relu
-#Conv5,2
-Conv
-Add
-#NML3,1
-Add
-#NML4,1
-Relu
-#Conv6,3
-Conv
-Add
-Relu
-#Conv7,2
-Conv
-Add
-#NML5,1
-Add
-#NML6,1
-Relu
-#Conv8,3
-Conv
-Add
-Relu
-#Conv9,2
-Conv
-Add
-#Conv10,2
-Conv
-Add
-#NML7,1
-Add
-#NML8,1
-Relu
-#Conv11,3
-Conv
-Add
-Relu
-#Conv12,2
-Conv
-Add
-#NML9,1
-Add
-#NML10,1
-Relu
-#Conv13,3
-Conv
-Add
-Relu
-#Conv14,2
-Conv
-Add
-#NML11,1
-Add
-#NML12,1
-Relu
-#Conv15,3
-Conv
-Add
-Relu
-#Conv16,2
-Conv
-Add
-#Conv17,2
-Conv
-Add
-#NML13,1
-Add
-#NML14,1
-Relu
-#Conv18,3
-Conv
-Add
-Relu
-#Conv19,2
-Conv
-Add
-#NML15,1
-Add
-#NML16,1
-Relu
-#Conv20,3
-Conv
-Add
-Relu
-#Conv21,2
-Conv
-Add
-#NML17,1
-Add
-#NML18,1
-Relu
-#NML19,1
-Pool
-#FC1,2
-Mul
-Add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet18_torch/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet18_torch/op_cost.txt
deleted file mode 100644
index 6fb1aef66a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet18_torch/op_cost.txt
+++ /dev/null
@@ -1,21 +0,0 @@
-88473.60156
-1887436.833
-1887436.833
-1887436.833
-1887436.833
-3774873.667
-1887436.833
-26214.40046
-1887436.833
-1887436.833
-3774873.667
-1887436.833
-13107.20023
-1887436.833
-1887436.833
-3774873.667
-1887436.833
-6553.600116
-1887436.833
-1887436.833
-64.0000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet18_torch/resnet_layers.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet18_torch/resnet_layers.txt
deleted file mode 100644
index 2e51c67842..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet18_torch/resnet_layers.txt
+++ /dev/null
@@ -1,59 +0,0 @@
-conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-conv
-batchnorm
-conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-conv
-batchnorm
-conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-conv
-batchnorm
-conv
-batchnorm
-conv
-batchnorm
-activation
-conv
-batchnorm
-activation
-activation
-pool_mean
-dense add
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/dev_knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/dev_knobs.txt
deleted file mode 100644
index 44fabd399e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/dev_knobs.txt
+++ /dev/null
@@ -1,54 +0,0 @@
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/op_cost.txt
deleted file mode 100644
index 51a116031c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/op_cost.txt
+++ /dev/null
@@ -1,54 +0,0 @@
-1180139.500000
-123904.000000
-1115136.000000
-495616.000000
-495616.000000
-495616.000000
-1115136.000000
-495616.000000
-495616.000000
-1115136.000000
-495616.000000
-238878.718750
-1156055.000000
-513802.250000
-955514.875000
-513802.250000
-1156055.000000
-513802.250000
-513802.250000
-1156055.000000
-513802.250000
-513802.250000
-1156055.000000
-513802.250000
-256901.125000
-1156055.000000
-513802.250000
-1027604.500000
-513802.250000
-1156055.000000
-513802.250000
-513802.250000
-1156055.000000
-513802.250000
-513802.250000
-1156055.000000
-513802.250000
-513802.250000
-1156055.000000
-513802.250000
-513802.250000
-1156055.000000
-513802.250000
-256901.125000
-1156055.000000
-513802.250000
-1027604.500000
-513802.250000
-1156055.000000
-513802.250000
-513802.250000
-1156055.000000
-513802.250000
-20480.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/quant_ranges2.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/quant_ranges2.txt
deleted file mode 100644
index efd6050d42..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/quant_ranges2.txt
+++ /dev/null
@@ -1,54 +0,0 @@
--123.68 151.061 -0.5747970953583718 0.56459772187472 -0.0058212606 0.015389577 0.0 934.6561442260769
--5.810299396514893 12.891858446121272 -0.544614662528038 0.24515408918262138 -0.00013975719 0.00010588932 -10.319062399864197 8.678670206069992
-0.0 6.145776463985475 -0.18031905634701254 0.21207299317419936 -0.00017183031 0.000101588186 -7.6816201620101925 9.201453342437766
-0.0 10.43091783905031 -0.22560040648281576 0.22525191849470233 -0.00024638147 0.00014042071 -3.83010697555542 3.860411907196095
--5.810299396514893 12.891858446121272 -0.43263624683022495 0.2949749290347107 -0.000119998615 9.139888e-05 -8.061470986366272 8.141587300300685
-0.0 6.200384625434893 -0.1698226538300514 0.14235900431871545 -8.860863e-05 8.355517e-05 -5.481491597175599 4.2078081378936965
-0.0 4.47170594310761 -0.15525014728307723 0.15632940307259882 -8.171968e-05 4.0018913e-05 -3.6843703370094296 3.4650170893669143
-0.0 4.6161603989601145 -0.15933965119719506 0.17397216582299158 -0.00012988421 0.00010135813 -1.5977122792005538 1.4737758718729186
-0.0 6.8487525005340615 -0.14451123128831386 0.14304637806118328 -4.2361393e-05 6.6440014e-05 -4.931722650527954 4.449415233612065
-0.0 4.033760806083722 -0.12324695030599832 0.12555985279381346 -5.0523813e-05 4.329417e-05 -3.2670734379291533 3.6830093951225296
-0.0 6.667089067935957 -0.1683492512255907 0.16272910618782485 -6.668582e-05 3.8562874e-05 -2.2874248178005216 1.6114515174627524
-0.0 7.2551298189163305 -0.14964802296459673 0.1396080917269055 -1.9781537e-05 2.6307946e-05 -5.558738519668579 4.226346692085322
-0.0 4.103045602321853 -0.09806137733161449 0.11390491172671352 -4.056617e-05 2.3085524e-05 -5.205046501636505 5.010882857322848
-0.0 3.8779746284484986 -0.1485816574841738 0.1498906888067731 -4.8492308e-05 6.3164814e-05 -1.5636358127593994 1.4471412793398377
-0.0 7.2551298189163305 -0.16395657777786254 0.16875704464316676 -2.4847817e-05 2.8032688e-05 -4.694492742061615 4.153118049144783
-0.0 7.047195756912359 -0.09732106301933527 0.10529463365674105 -1.8920293e-05 1.3819959e-05 -5.122300206184387 3.822466685295119
-0.0 3.601598421573673 -0.08627650760114193 0.08969089165330116 -2.4561228e-05 3.960287e-05 -2.9372031297683714 2.453564183473647
-0.0 3.8704470412732075 -0.12190281387418508 0.1344742950797146 -6.2474406e-05 4.5489418e-05 -1.1263611575365067 0.9419634058475754
-0.0 7.360276422501315 -0.11249553721398114 0.1438516600430019 -1.775357e-05 1.5396989e-05 -6.671636656284333 7.135764349460775
-0.0 3.7979030499458766 -0.10526792272925377 0.1349908015877019 -3.4459194e-05 3.4367415e-05 -4.684974233627319 5.264288017273344
-0.0 4.181413197040673 -0.1404084537923336 0.12109696540981926 -6.0512917e-05 5.7992842e-05 -1.707221627831459 0.9086800929904393
-0.0 7.209602268219356 -0.16052422896027566 0.19026783116161974 -2.5455045e-05 2.27223e-05 -4.75133854341507 4.969562854290029
-0.0 5.027068079471608 -0.11333685562014578 0.10740291688591388 -3.8223963e-05 3.5434005e-05 -4.555457803249359 3.7164909157753137
-0.0 5.087393596649491 -0.11392991304397583 0.14321510225534495 -3.1888136e-05 3.15042e-05 -1.5135304692983627 1.3553025794029487
-0.0 8.094761092185998 -0.10484503862261772 0.1209987572357194 -1.4100494e-05 1.0806584e-05 -4.084207750797272 3.8462553339005128
-0.0 4.2752423663139805 -0.07607016738504171 0.07963053999841248 -1.6701442e-05 1.5241814e-05 -6.879458940029145 4.920539072513741
-0.0 3.51579514932658 -0.11533396770060061 0.10761716394126962 -3.1804513e-05 2.6629998e-05 -1.3364267847537994 1.0267766494750994
-0.0 8.094761092185998 -0.09670568963885307 0.10261806760728373 -1.7871134e-05 1.2980788e-05 -4.165599856376648 3.9349664554596693
-0.0 7.114320134163032 -0.06487256648391485 0.08072723559290183 -1.0383376e-05 7.2508515e-06 -7.300614297866821 4.876313820362272
-0.0 3.3421912226677954 -0.06480042573064566 0.07086511892825366 -1.8462242e-05 2.0371868e-05 -3.3948297231197357 2.547599082708394
-0.0 3.0945872743134055 -0.0918739195242524 0.0884740859493657 -3.932561e-05 3.442098e-05 -1.020233019709587 0.5999923423528848
-0.0 6.7569902172089265 -0.07119932436943054 0.09078719960898206 -1.09133825e-05 1.1075285e-05 -4.42014742231369 3.863946924209598
-0.0 3.1060702869893984 -0.05620019323378801 0.0647352593019614 -1.7693064e-05 2.5484389e-05 -2.1564681169986724 1.7775597308874955
-0.0 3.5050717415809824 -0.07137715590745211 0.08277502278238583 -3.7179158e-05 3.6062916e-05 -1.2057263028621674 0.7878368514776246
-0.0 6.819144511223078 -0.06814771068841219 0.09314906919747618 -1.3129389e-05 1.2350107e-05 -4.264177570819855 3.6667083778381695
-0.0 3.010977530002802 -0.05523281182348728 0.06323543420434069 -2.8231338e-05 2.693129e-05 -2.403562089204788 2.0931886761188707
-0.0 3.2924690744877205 -0.07265310526639222 0.0840946944877511 -3.2401527e-05 3.4846238e-05 -1.3499533622264863 0.5710797674655921
-0.0 6.86581426668181 -0.06556392236053944 0.10085416895151492 -1.34982765e-05 1.5712916e-05 -4.0751124567985535 3.8311849322319937
-0.0 3.305581816196461 -0.05189245833083987 0.06286057442426854 -2.7664553e-05 2.7124273e-05 -2.2815526587963104 1.6054227759838495
-0.0 4.0827409029009045 -0.0667651962414384 0.0779341282621039 -2.5976125e-05 3.067452e-05 -1.531399680972099 0.9400330729484727
-0.0 7.124539953709001 -0.07115359698235989 0.10833600817620766 -1.1960556e-05 1.597267e-05 -3.499965760946274 6.504482283592731
-0.0 5.178686481952853 -0.05611966584250331 0.06787234821170791 -2.637223e-05 2.5699503e-05 -4.342721074581146 3.5257833659650792
-0.0 7.024678532601377 -0.06805450225621462 0.08350853440165543 -2.051169e-05 2.4474735e-05 -3.132657457351684 1.5121956378223056
-0.0 7.723012191295651 -0.06772804392129184 0.0790314114466323 -1.0622113e-05 1.1603386e-05 -4.100774613380432 3.367799842357637
-0.0 3.30561476135256 -0.04215342074632644 0.04599864583462508 -2.3764493e-05 2.0471329e-05 -3.442601467609405 1.980771198987965
-0.0 2.7743033297061963 -0.05116425715386867 0.07435814850032654 -3.0402798e-05 3.17601e-05 -0.6644622380137444 0.6883783437013742
-0.0 7.723012191295651 -0.04721667753905058 0.06949943191558243 -1.7287093e-05 1.273127e-05 -2.0042565877437593 2.5162933135032794
-0.0 13.089089687347467 -0.04412535773590207 0.06208482719958153 -5.398635e-06 4.8724514e-06 -10.966248041152955 6.323037844181082
-0.0 2.665933228015901 -0.036525007653981444 0.04293334979563958 -2.6018542e-05 3.12493e-05 -2.3321639671325682 1.520020196676264
-0.0 2.3228017909526826 -0.0457546092569828 0.07228027954698127 -2.8726334e-05 4.8297712e-05 -0.5039844071269035 0.5125965433716879
-0.0 14.212855710030006 -0.05184821262955666 0.07706690635532187 -7.475739e-06 6.432327e-06 -8.724605677604675 7.675269114017488
-0.0 2.9809948239326776 -0.032554453313350676 0.03890361279249199 -2.651515e-05 3.1177675e-05 -1.8098392320871355 0.9923119935989424
-0.0 3.0153360161781455 -0.04308305019512773 0.07738747298717986 -2.5071722e-05 4.40581e-05 -0.5571711810231208 0.6459895361661969
-0.0 6.169458543777802 -0.0922410073429346 0.2492780588418327 -0.024867358 0.028531343 -8.56706305217743 11.648686022758701
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/resnet50_layers.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/resnet50_layers.txt
deleted file mode 100644
index f0b6ebedc9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/resnet50_layers.txt
+++ /dev/null
@@ -1,172 +0,0 @@
-conv add activation pool
-batchnorm
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-add
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-add
-activation
-pool
-dense add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/resnet50_tensors.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/resnet50_tensors.txt
deleted file mode 100644
index 70ec0ff111..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/resnet50_imagenet/resnet50_tensors.txt
+++ /dev/null
@@ -1,172 +0,0 @@
-Conv1,100,3,224,224,64,3,7,7
-#tensorBatchNorm1
-Conv2,100,64,55,55,64,64,1,1
-#tensorBatchNorm2
-#tensorRelu1
-Conv3,100,64,55,55,64,64,3,3
-#tensorBatchNorm3
-#tensorRelu2
-Conv4,100,64,55,55,256,64,1,1
-#tensorBatchNorm4
-Conv5,100,64,55,55,256,64,1,1
-#tensorBatchNorm5
-#tensorAdd1
-#tensorRelu3
-Conv6,100,256,55,55,64,256,1,1
-#tensorBatchNorm6
-#tensorRelu4
-Conv7,100,64,55,55,64,64,3,3
-#tensorBatchNorm7
-#tensorRelu5
-Conv8,100,64,55,55,256,64,1,1
-#tensorBatchNorm8
-#tensorAdd2
-#tensorRelu6
-Conv9,100,256,55,55,64,256,1,1
-#tensorBatchNorm9
-#tensorRelu7
-Conv10,100,64,55,55,64,64,3,3
-#tensorBatchNorm10
-#tensorRelu8
-Conv11,100,64,55,55,256,64,1,1
-#tensorBatchNorm11
-#tensorAdd3
-#tensorRelu9
-Conv12,100,256,55,55,128,256,1,1
-#tensorBatchNorm12
-#tensorRelu10
-Conv13,100,128,28,28,128,128,3,3
-#tensorBatchNorm13
-#tensorRelu11
-Conv14,100,128,28,28,512,128,1,1
-#tensorBatchNorm14
-Conv15,100,256,55,55,512,256,1,1
-#tensorBatchNorm15
-#tensorAdd4
-#tensorRelu12
-Conv16,100,512,28,28,128,512,1,1
-#tensorBatchNorm16
-#tensorRelu13
-Conv17,100,128,28,28,128,128,3,3
-#tensorBatchNorm17
-#tensorRelu14
-Conv18,100,128,28,28,512,128,1,1
-#tensorBatchNorm18
-#tensorAdd5
-#tensorRelu15
-Conv19,100,512,28,28,128,512,1,1
-#tensorBatchNorm19
-#tensorRelu16
-Conv20,100,128,28,28,128,128,3,3
-#tensorBatchNorm20
-#tensorRelu17
-Conv21,100,128,28,28,512,128,1,1
-#tensorBatchNorm21
-#tensorAdd6
-#tensorRelu18
-Conv22,100,512,28,28,128,512,1,1
-#tensorBatchNorm22
-#tensorRelu19
-Conv23,100,128,28,28,128,128,3,3
-#tensorBatchNorm23
-#tensorRelu20
-Conv24,100,128,28,28,512,128,1,1
-#tensorBatchNorm24
-#tensorAdd7
-#tensorRelu21
-Conv25,100,512,28,28,256,512,1,1
-#tensorBatchNorm25
-#tensorRelu22
-Conv26,100,256,14,14,256,256,3,3
-#tensorBatchNorm26
-#tensorRelu23
-Conv27,100,256,14,14,1024,256,1,1
-#tensorBatchNorm27
-Conv28,100,512,28,28,1024,512,1,1
-#tensorBatchNorm28
-#tensorAdd8
-#tensorRelu24
-Conv29,100,1024,14,14,256,1024,1,1
-#tensorBatchNorm29
-#tensorRelu25
-Conv30,100,256,14,14,256,256,3,3
-#tensorBatchNorm30
-#tensorRelu26
-Conv31,100,256,14,14,1024,256,1,1
-#tensorBatchNorm31
-#tensorAdd9
-#tensorRelu27
-Conv32,100,1024,14,14,256,1024,1,1
-#tensorBatchNorm32
-#tensorRelu28
-Conv33,100,256,14,14,256,256,3,3
-#tensorBatchNorm33
-#tensorRelu29
-Conv34,100,256,14,14,1024,256,1,1
-#tensorBatchNorm34
-#tensorAdd10
-#tensorRelu30
-Conv35,100,1024,14,14,256,1024,1,1
-#tensorBatchNorm35
-#tensorRelu31
-Conv36,100,256,14,14,256,256,3,3
-#tensorBatchNorm36
-#tensorRelu32
-Conv37,100,256,14,14,1024,256,1,1
-#tensorBatchNorm37
-#tensorAdd11
-#tensorRelu33
-Conv38,100,1024,14,14,256,1024,1,1
-#tensorBatchNorm38
-#tensorRelu34
-Conv39,100,256,14,14,256,256,3,3
-#tensorBatchNorm39
-#tensorRelu35
-Conv40,100,256,14,14,1024,256,1,1
-#tensorBatchNorm40
-#tensorAdd12
-#tensorRelu36
-Conv41,100,1024,14,14,256,1024,1,1
-#tensorBatchNorm41
-#tensorRelu37
-Conv42,100,256,14,14,256,256,3,3
-#tensorBatchNorm42
-#tensorRelu38
-Conv43,100,256,14,14,1024,256,1,1
-#tensorBatchNorm43
-#tensorAdd13
-#tensorRelu39
-Conv44,100,1024,14,14,512,1024,1,1
-#tensorBatchNorm44
-#tensorRelu40
-Conv45,100,512,7,7,512,512,3,3
-#tensorBatchNorm45
-#tensorRelu41
-Conv46,100,512,7,7,2048,512,1,1
-#tensorBatchNorm46
-Conv47,100,1024,14,14,2048,1024,1,1
-#tensorBatchNorm47
-#tensorAdd14
-#tensorRelu42
-Conv48,100,2048,7,7,512,2048,1,1
-#tensorBatchNorm48
-#tensorRelu43
-Conv49,100,512,7,7,512,512,3,3
-#tensorBatchNorm49
-#tensorRelu44
-Conv50,100,512,7,7,2048,512,1,1
-#tensorBatchNorm50
-#tensorAdd15
-#tensorRelu45
-Conv51,100,2048,7,7,512,2048,1,1
-#tensorBatchNorm51
-#tensorRelu46
-Conv52,100,512,7,7,512,512,3,3
-#tensorBatchNorm52
-#tensorRelu47
-Conv53,100,512,7,7,2048,512,1,1
-#tensorBatchNorm53
-#tensorAdd16
-#tensorRelu48
-#tensorPooling1
-FC1,100,2048,2048,1000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/dev_knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/dev_knobs.txt
deleted file mode 100644
index 8970b19bfb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/dev_knobs.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11
-11
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/knobs.txt
deleted file mode 100644
index d238fa1036..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/knobs.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12
-12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/knobs1.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/knobs1.txt
deleted file mode 100644
index fb54e7f077..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/knobs1.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12
-11,12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/op_cost.txt
deleted file mode 100644
index 5f58ebcc04..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/op_cost.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-88473.601562
-1887436.750000
-943718.375000
-1887436.750000
-943718.375000
-1887436.750000
-1887436.750000
-943718.375000
-1887436.750000
-1887436.750000
-471859.187500
-471859.187500
-471859.187500
-13107.200195
-256.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/vgg16_layers.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/vgg16_layers.txt
deleted file mode 100644
index 79818d6f01..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/vgg16_layers.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-conv add activation
-conv add activation pool
-conv add activation
-conv add activation pool
-conv add activation
-conv add activation
-conv add activation pool
-conv add activation
-conv add activation
-conv add activation pool
-conv add activation
-conv add activation
-conv add activation pool
-dense add activation
-dense add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/vgg16_tensors.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/vgg16_tensors.txt
deleted file mode 100644
index a524e1e74e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10/vgg16_tensors.txt
+++ /dev/null
@@ -1,64 +0,0 @@
-#Conv1,3
-Conv
-Add
-Relu
-#Conv2,4
-Conv
-Add
-Relu
-Pool
-#Conv3,3
-Conv
-Add
-Relu
-#Conv4,4
-Conv
-Add
-Relu
-Pool
-#Conv5,3
-Conv
-Add
-Relu
-#Conv6,3
-Conv
-Add
-Relu
-#Conv7,4
-Conv
-Add
-Relu
-Pool
-#Conv8,3
-Conv
-Add
-Relu
-#Conv9,3
-Conv
-Add
-Relu
-#Conv10,4
-Conv
-Add
-Relu
-Pool
-#Conv11,3
-Conv
-Add
-Relu
-#Conv12,3
-Conv
-Add
-Relu
-#Conv13,4
-Conv
-Add
-Relu
-Pool
-#FC1,3
-Mul
-Add
-Relu
-#FC2,2
-Mul
-Add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/dev_knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/dev_knobs.txt
deleted file mode 100644
index 8970b19bfb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/dev_knobs.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11
-11
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/knobs.txt
deleted file mode 100644
index d238fa1036..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/knobs.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12
-12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/knobs1.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/knobs1.txt
deleted file mode 100644
index fb54e7f077..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/knobs1.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12
-11,12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/op_cost.txt
deleted file mode 100644
index 8c6daad2e2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/op_cost.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-884736.000000
-18874368.000000
-9437184.000000
-18874368.000000
-9437184.000000
-18874368.000000
-18874368.000000
-9437184.000000
-18874368.000000
-18874368.000000
-4718592.000000
-4718592.000000
-4718592.000000
-131072.000000
-25600.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/vgg16_layers.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/vgg16_layers.txt
deleted file mode 100644
index 79818d6f01..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/vgg16_layers.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-conv add activation
-conv add activation pool
-conv add activation
-conv add activation pool
-conv add activation
-conv add activation
-conv add activation pool
-conv add activation
-conv add activation
-conv add activation pool
-conv add activation
-conv add activation
-conv add activation pool
-dense add activation
-dense add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/vgg16_tensors.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/vgg16_tensors.txt
deleted file mode 100644
index a524e1e74e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar100/vgg16_tensors.txt
+++ /dev/null
@@ -1,64 +0,0 @@
-#Conv1,3
-Conv
-Add
-Relu
-#Conv2,4
-Conv
-Add
-Relu
-Pool
-#Conv3,3
-Conv
-Add
-Relu
-#Conv4,4
-Conv
-Add
-Relu
-Pool
-#Conv5,3
-Conv
-Add
-Relu
-#Conv6,3
-Conv
-Add
-Relu
-#Conv7,4
-Conv
-Add
-Relu
-Pool
-#Conv8,3
-Conv
-Add
-Relu
-#Conv9,3
-Conv
-Add
-Relu
-#Conv10,4
-Conv
-Add
-Relu
-Pool
-#Conv11,3
-Conv
-Add
-Relu
-#Conv12,3
-Conv
-Add
-Relu
-#Conv13,4
-Conv
-Add
-Relu
-Pool
-#FC1,3
-Mul
-Add
-Relu
-#FC2,2
-Mul
-Add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10_torch/layers.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10_torch/layers.txt
deleted file mode 100644
index ef3d0ebcf7..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10_torch/layers.txt
+++ /dev/null
@@ -1,46 +0,0 @@
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-pool
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-pool
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-pool
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-pool
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-conv add
-batchnorm
-activation
-pool
-pool_mean
-dense add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10_torch/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10_torch/op_cost.txt
deleted file mode 100644
index 10dc83f865..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_cifar10_torch/op_cost.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-88473.60156
-1887436.833
-943718.4167
-1887436.833
-943718.4167
-1887436.833
-1887436.833
-943718.4167
-1887436.833
-1887436.833
-471859.2083
-471859.2083
-471859.2083
-13107.200195
-256.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/dev_knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/dev_knobs.txt
deleted file mode 100644
index 793b41f5b8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/dev_knobs.txt
+++ /dev/null
@@ -1,16 +0,0 @@
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-11
-11
-11
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/knobs.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/knobs.txt
deleted file mode 100644
index d238fa1036..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/knobs.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
-12
-12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/knobs1.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/knobs1.txt
deleted file mode 100644
index fb54e7f077..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/knobs1.txt
+++ /dev/null
@@ -1,15 +0,0 @@
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12,21,22,23,24,25,26,27,28,31,32,33,34
-11,12
-11,12
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/op_cost.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/op_cost.txt
deleted file mode 100644
index 77754f5f8b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/op_cost.txt
+++ /dev/null
@@ -1,16 +0,0 @@
-88473.601562
-1887436.750000
-943718.375000
-1887436.750000
-943718.375000
-1887436.750000
-1887436.750000
-943718.375000
-1887436.750000
-1887436.750000
-471859.187500
-471859.187500
-471859.187500
-13107.200195
-13107.200195
-256.000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/quant_ranges2.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/quant_ranges2.txt
deleted file mode 100644
index 92b6cbdabe..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/quant_ranges2.txt
+++ /dev/null
@@ -1,16 +0,0 @@
--123.68 151.061 -0.5682651399970055 0.5677501424551024 -0.015828926 2.064037 0.0 455.8094849853551
-0.0 455.8094849853551 -0.13156980648636818 0.2164201746285022 -1.0271513 0.9052184 0.0 2768.0925808105603
-0.0 2768.0925808105603 -0.18644111251831055 0.202149114727974 -0.17922063 0.36547425 0.0 2843.0992189941426
-0.0 2843.0992189941426 -0.10804861642420292 0.12427636455744764 -0.59533477 0.63375777 0.0 6242.117490722692
-0.0 6242.117490722692 -0.08040237371623515 0.09835810117424044 -0.20097896 0.34949613 0.0 5991.169433105955
-0.0 5991.169433105955 -0.05306418750435114 0.06628044287860436 -0.18124875 0.274845 0.0 5665.713403320435
-0.0 5665.713403320435 -0.05084674355760217 0.07320860563218634 -0.14288792 0.59477174 0.0 7507.031136718913
-0.0 7507.031136718913 -0.04523278899490833 0.053042236261070186 -0.14548235 0.3148451 0.0 4485.656244140948
-0.0 4485.656244140948 -0.02917514201253653 0.03586270406842279 -0.08428453 0.18237582 0.0 2199.244487060563
-0.0 2199.244487060563 -0.029496615380048753 0.04047201693058028 -0.19835947 0.33766547 0.0 1430.6242337648137
-0.0 1430.6242337648137 -0.031951379626989365 0.04218719156458998 -0.3508028 0.6397485 0.0 740.3813181152382
-0.0 740.3813181152382 -0.028522676015272738 0.03794213477522136 -0.9171057 0.7597668 0.0 379.07848284912643
-0.0 379.07848284912643 -0.02821479567326605 0.03854479804635069 -0.50036746 9.431553 0.0 200.11104328918475
-0.0 200.11104328918475 -0.007091613108757884 0.008147951829247227 -0.78005254 0.8555075 0.0 58.32595815658672
-0.0 58.32595815658672 -0.012781758182682096 0.01437051862943929 -0.012339931 1.2154555 0.0 17.291313619614062
-0.0 17.291313619614062 -0.02119149128906429 0.02715564412623694 -0.773357 0.6615543 -11.322624863624572 24.837094623566433
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/vgg16_layers.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/vgg16_layers.txt
deleted file mode 100644
index bfa2a2700a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/vgg16_layers.txt
+++ /dev/null
@@ -1,16 +0,0 @@
-conv add activation
-conv add activation pool
-conv add activation
-conv add activation pool
-conv add activation
-conv add activation
-conv add activation pool
-conv add activation
-conv add activation
-conv add activation pool
-conv add activation
-conv add activation
-conv add activation pool
-dense add activation
-dense add activation
-dense add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/vgg16_tensors.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/vgg16_tensors.txt
deleted file mode 100644
index b9afe8dd34..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/data/vgg16_imagenet/vgg16_tensors.txt
+++ /dev/null
@@ -1,68 +0,0 @@
-#Conv1,3
-Conv
-Add
-Relu
-#Conv2,4
-Conv
-Add
-Relu
-Pool
-#Conv3,3
-Conv
-Add
-Relu
-#Conv4,4
-Conv
-Add
-Relu
-Pool
-#Conv5,3
-Conv
-Add
-Relu
-#Conv6,3
-Conv
-Add
-Relu
-#Conv7,4
-Conv
-Add
-Relu
-Pool
-#Conv8,3
-Conv
-Add
-Relu
-#Conv9,3
-Conv
-Add
-Relu
-#Conv10,4
-Conv
-Add
-Relu
-Pool
-#Conv11,3
-Conv
-Add
-Relu
-#Conv12,3
-Conv
-Add
-Relu
-#Conv13,4
-Conv
-Add
-Relu
-Pool
-#FC1,3
-Mul
-Add
-Relu
-#FC2,3
-Mul
-Add
-Relu
-#FC3,2
-Mul
-Add
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/.gitignore b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/.gitignore
deleted file mode 100644
index 9eb809777b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/.gitignore
+++ /dev/null
@@ -1,51 +0,0 @@
-*.py[cod]
-
-# C extensions
-*.so
-
-# Packages
-*.egg
-*.egg-info
-dist
-build
-eggs
-parts
-bin
-var
-sdist
-develop-eggs
-.installed.cfg
-lib
-lib64
-
-# Installer logs
-pip-log.txt
-
-# Unit test / coverage reports
-.coverage
-.tox
-nosetests.xml
-
-# Translations
-*.mo
-
-# Mr Developer
-.mr.developer.cfg
-.project
-.pydevproject
-
-#vim
-*.swp
-
-#virtualenv
-venv
-.ropeproject
-opentuner.log
-.*.swo
-opentuner.db
-.idea
-
-# SMB ROM (for SMB demo)
-smb.nes
-
-MANIFEST
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/AUTHORS.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/AUTHORS.txt
deleted file mode 100644
index 620e549e23..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/AUTHORS.txt
+++ /dev/null
@@ -1,8 +0,0 @@
-Jason Ansel
-Sam Fingeret
-Shoaib Kamil
-Deepak Narayanan
-Jonathan Ragan-Kelley
-Kalyan Veeramachaneni
-Kevin Wu
-Minshu Zhan
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/CHANGES.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/CHANGES.txt
deleted file mode 100644
index a0af442222..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/CHANGES.txt
+++ /dev/null
@@ -1,3 +0,0 @@
-v0.5.0, 2015-02-10 -- Refactoring and bugfixes.
-v0.4.0, 2014-10-26 -- Add api and bugfixes.
-v0.3.0, 2014-08-11 -- Initial release.
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/LICENSE.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/LICENSE.txt
deleted file mode 100644
index 2b602e192b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/LICENSE.txt
+++ /dev/null
@@ -1,22 +0,0 @@
-The MIT License (MIT)
-
-Copyright (c) 2014 Jason Ansel
-
-Permission is hereby granted, free of charge, to any person obtaining a copy
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in
-all copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
-THE SOFTWARE.
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/MANIFEST.in b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/MANIFEST.in
deleted file mode 100644
index 376b77ae8f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/MANIFEST.in
+++ /dev/null
@@ -1 +0,0 @@
-include *.txt *.md
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/README.md b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/README.md
deleted file mode 100644
index 729f35553a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/README.md
+++ /dev/null
@@ -1,116 +0,0 @@
-OpenTuner
-=========
-
-Program autotuning has been demonstrated in many domains to achieve better
-or more portable performance. However, autotuners themselves are often not
-very portable between projects because using a domain informed search space
-representation is critical to achieving good results and because no single
-search technique performs best for all problems.
-
-OpenTuner is a new framework for building domain-specific multi-objective
-program autotuners. OpenTuner supports fully customizable configuration
-representations, an extensible technique representation to allow for
-domain-specific techniques, and an easy to use interface for communicating
-with the tuned program. A key capability inside OpenTuner is the use of
-ensembles of disparate search techniques simultaneously, techniques which
-perform well will receive larger testing budgets and techniques which perform
-poorly will be disabled.
-
-System dependencies
--------------------
-
-A list of system dependencies can be found in [debian-packages-deps][]
-which are primarily python 2.6+ (not 3.x) and sqlite3 (or your
-[supported][sqlalchemy-dialects] database backend of choice).
-
-On Ubuntu/Debian there can be installed with:
-
- sudo apt-get install `cat debian-packages-deps | tr '\n' ' '`
-
-[debian-packages-deps]: https://raw.github.com/jansel/opentuner/master/debian-packages-deps
-[sqlalchemy-dialects]: http://docs.sqlalchemy.org/en/rel_0_8/dialects/index.html
-
-
-Installation
--------------------
-OpenTuner (and dependencies) can be installed with
-
- sudo pip install opentuner
-
-or
-
- pip install --user opentuner
-
-This will not install any of the example programs.
-
-
-Development installation
--------------------
-For development (running OpenTuner out of a git checkout), a list of python
-dependencies can be found in [requirements.txt][] these can be installed
-system-wide with `pip`.
-
- sudo apt-get install python-pip
- sudo pip install -r requirements.txt
-
-Or you can use virtual env to create a isolated python environment by running:
-
- python ./venv-bootstrap.py
-
-which will create a ./venv/bin/python (./venv/Scripts/python.exe on windows)
-with all the required packages installed.
-
-[requirements.txt]: https://raw.github.com/jansel/opentuner/master/requirements.txt
-
-
-Checking Installation
----------------------
-
-Quickly checking that a successful installation has been made, may be performed
-by running an example program such as:
-
- ./examples/rosenbrock/rosenbrock.py
-
-
-Tutorials
----------
-
-- [Optimizing Block Matrix Multiplication][gettingstarted]
-- [Creating OpenTuner Techniques][technique-tutorial].
-
-[gettingstarted]: http://opentuner.org/tutorial/gettingstarted/
-[technique-tutorial]: http://opentuner.org/tutorial/techniques/
-
-
-Papers
----------
-
-- [OpenTuner: An Extensible Framework for Program Autotuning][paper1]. <br>
- Jason Ansel, Shoaib Kamil, Kalyan Veeramachaneni, Jonathan Ragan-Kelley,
- Jeffrey Bosboom, Una-May O'Reilly, Saman Amarasinghe. <br>
- International Conference on Parallel Architectures and Compilation
- Techniques. <br>
- Edmonton, Canada. August, 2014. [Slides][slides1]. [Bibtex][bibtex1].
-
-[paper1]: http://groups.csail.mit.edu/commit/papers/2014/ansel-pact14-opentuner.pdf
-[bibtex1]: http://groups.csail.mit.edu/commit/bibtex.cgi?key=ansel:pact:2014
-[slides1]: http://groups.csail.mit.edu/commit/papers/2014/ansel-pact14-opentuner-slides.pdf
-
-
-Contributing Code
------------------
-
-The preferred way to contribute code to OpenTuner is to fork the project
-on github and [submit a pull request][pull-req].
-
-[pull-req]: https://www.openshift.com/wiki/github-workflow-for-submitting-pull-requests
-
-
-Support
--------
-OpenTuner is supported in part by the United States Department of Energy
-[X-Stack][xstack] program as part of [D-TEC][dtec].
-
-[xstack]: http://science.energy.gov/ascr/research/computer-science/ascr-x-stack-portfolio/
-[dtec]: http://www.dtec-xstack.org/
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/__init__.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/__init__.py
deleted file mode 100644
index e69de29bb2..0000000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/accuracy_tuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/accuracy_tuner.py
deleted file mode 100644
index 5977fe7ee5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/accuracy_tuner.py
+++ /dev/null
@@ -1,198 +0,0 @@
-#!/usr/bin/env python
-#
-
-import adddeps # fix sys.path
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-
-
-output_dir = ""
-flag_ranges = []
-tuning_flags = []
-binary_name = ""
-accuracy_threshold = 10.0
-opt_confs_index = 9
-evaluated_configs = {}
-
-
-def extractTotalOverhead(file_name):
-
- total_comps = 0.0
- file = open(file_name, "r")
- for x in file:
- words = x.split()
- total_comps += float(words[opt_confs_index])
-
- print total_comps
- return total_comps
-
-
-def getAccuracy(file_name):
-
- file = open(file_name, "r")
- acc_str = file.read()
- accuracy = float(acc_str)
- print accuracy
- return accuracy
-
-
-def createFlagsFile(file_name, cfg):
-
- f = open(file_name, "w+")
- cmd_config = ""
- for flag in tuning_flags:
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-class ClangFlagsTuner(MeasurementInterface):
-
- def __init__(self, args):
- objective = ThresholdAccuracyMinimizeTime(accuracy_threshold)
- input_manager = FixedInputManager(size=num_flags)
- self.configs_list = []
-
- super(ClangFlagsTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- for flag in tuning_flags:
- manipulator.add_parameter(
- EnumParameter(flag, flag_ranges
- # [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
- )) #default is needed, optimizations don't work without it(tried and tested)
- return manipulator
-
-
- def run(self, desired_result, input, limit):
- """
- Compile and run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
-
- # NOTE: creates the file with flags read by the runtime
- createFlagsFile("opentuner_flags", cfg)
-
- run_cmd = binary_name
- print run_cmd
- run_result_call_program = self.call_program(run_cmd)
- #print run_result_call_program
-
- total_comps = extractTotalOverhead("accuracy_summary")
- accuracy = getAccuracy("final_accuracy")
-
- #Result = opentuner.resultsdb.models.Result(time=total_comps)
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- Result.accuracy = accuracy
-
- if accuracy > accuracy_threshold:
- if accuracy not in evaluated_configs:
- config_tuple = (total_comps, accuracy, cfg)
- self.configs_list.append(config_tuple)
- evaluated_configs[accuracy] = 1
- shutil.copy('accuracy_summary', output_dir + '/' + binary_name + '_' + str(accuracy))
-
-
- return Result
-
-
- def save_final_config(self, configuration):
- """
- called at the end of autotuning with the best resultsdb.models.Configuration
- """
- print "Final configuration", configuration.data
-
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- createFlagsFile("opentuner_flags", configuration.data)
- run_cmd = binary_name
- run_result_call_program = self.call_program(run_cmd)
-
- accuracy = getAccuracy("final_accuracy")
- shutil.copy('accuracy_summary', result_dir + '/' + binary_name + '_final_' + str(accuracy) )
-
- sorted_list = sorted(self.configs_list, key = lambda tup: tup[0])
- print sorted_list[0:10]
-
- top_elems = 20
- if len(sorted_list) < top_elems:
- top_elems = len(sorted_list)
-
-
- for i in range(top_elems):
- createFlagsFile("opentuner_flags", sorted_list[i][2])
- run_cmd = binary_name
- run_result_call_program = self.call_program(run_cmd)
- accuracy = getAccuracy("final_accuracy")
- shutil.copy('accuracy_summary', result_dir + '/' + binary_name + '_' + str(accuracy) + "_rank_" + str(i) )
-
-
- #os.mkdir(result_dir + "full_results")
-
-
-
-
-if __name__ == '__main__':
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='name of binary to run')
- argparser.add_argument('--num-flags', type=int, help='num of flags to tune for')
- argparser.add_argument('--error-range', type=int, help='num of flags to tune for')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='accuracy threshold')
-
-
- args = argparser.parse_args()
- binary_name = str(args.binary)
- print("binary_name = ", binary_name)
- num_flags = int(args.num_flags)
- error_range = int(args.error_range)
- accuracy_threshold = float(args.accuracy)
- print("accuracy = ", accuracy_threshold)
- result_dir = args.result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
-
- output_dir = result_dir + "/full_results"
- print output_dir
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- if not os.path.exists(output_dir):
- print("Creating output directory = ", output_dir)
- os.mkdir(output_dir)
-
- for j in range(error_range):
- flag_ranges.append(j)
-
- print("flag_ranges = ", flag_ranges)
-
- for i in range(num_flags):
- tuning_flags.append("flag" + str(i))
-
- ClangFlagsTuner.main(argparser.parse_args())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/accuracy_tuner_piped.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/accuracy_tuner_piped.py
deleted file mode 100644
index 6d46c5762e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/accuracy_tuner_piped.py
+++ /dev/null
@@ -1,269 +0,0 @@
-#!/usr/bin/env python
-#
-# Optimize blocksize of apps/mmm_block.cpp
-#
-# This is an extremely simplified version meant only for tutorials
-#
-import adddeps # fix sys.path
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-import subprocess
-import threading
-import psutil
-
-from measure_confidence import dump_high_confidence_files
-from select_top_results import select_top_results
-from time import sleep
-
-
-output_dir = ""
-flag_ranges = []
-tuning_flags = []
-binary_name = ""
-accuracy_threshold = 10.0
-opt_confs_index = 9
-evaluated_configs = {}
-orig_result_dir = ""
-
-
-def extractTotalOverhead(file_name):
-
- total_comps = 0.0
- file = open(file_name, "r")
- for x in file:
- words = x.split()
- total_comps += float(words[opt_confs_index])
-
- print total_comps
- return total_comps
-
-
-def getAccuracy(file_name):
-
- file = open(file_name, "r")
- acc_str = file.read()
- file.close()
-
- accuracy = float(acc_str)
-
- try:
- accuracy = float(acc_str)
- except:
- return 20
-
- print accuracy
- return accuracy
-
-
-
-def kill(proc_pid):
- process = psutil.Process(proc_pid)
- for proc in process.children(recursive=True):
- proc.kill()
- process.kill()
-
-
-
-def createFlagsFile(file_name, cfg):
-
- f = open(file_name, "w+")
- cmd_config = ""
- for flag in tuning_flags:
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-class ClangFlagsTuner(MeasurementInterface):
-
- def __init__(self, args):
- objective = ThresholdAccuracyMinimizeTime(accuracy_threshold)
- input_manager = FixedInputManager(size=num_flags)
- self.configs_list = []
-
- super(ClangFlagsTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
- FNULL = open(os.devnull, 'wb')
- #run_result_call_program = self.call_program(run_cmd)
- self.start_process = subprocess.Popen([binary_name, "opentuner_run"]) #, stdout=FNULL);
-
- try:
- os.mkfifo("/tmp/myfifo")
- except OSError, e:
- print("FIFO exists")
-
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- for flag in tuning_flags:
- manipulator.add_parameter(
- EnumParameter(flag, flag_ranges
- # [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
- )) #default is needed, optimizations don't work without it(tried and tested)
- return manipulator
-
-
- def run(self, desired_result, input, limit):
-
- """
- Run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
-
- # NOTE: creates the file with flags read by the runtime
- createFlagsFile("opentuner_flags", cfg)
-
- run_cmd = binary_name
- print run_cmd
- #run_result_call_program = self.call_program(run_cmd)
-
- # Using Named Pipes to signal execution to the DNN outer thread
- fifo = open("/tmp/myfifo", "w")
- fifo.write("start_run")
- fifo.close()
-
- print "Waiting for process to signal back - when done processing one run"
-
- fifo2 = open("/tmp/myfifo", "r")
- fifo2.read()
- fifo2.close()
-
- print "Process Signalled back"
-
- total_comps = extractTotalOverhead("accuracy_summary")
- accuracy = getAccuracy("final_accuracy")
-
-
- #Result = opentuner.resultsdb.models.Result(time=total_comps)
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- Result.accuracy = accuracy
-
- if accuracy > accuracy_threshold:
- if accuracy not in evaluated_configs:
- config_tuple = (total_comps, accuracy, cfg)
- self.configs_list.append(config_tuple)
- evaluated_configs[accuracy] = 1
- shutil.copy('accuracy_summary', output_dir + '/' + binary_name + '_' + str(accuracy))
-
-
- print "done with one run"
-
- return Result
-
-
- def save_final_config(self, configuration):
-
- print "Dumping High Confidence results"
- sleep(5)
-
- # Only dumping files with 95% confidence
- dump_high_confidence_files(binary_name, orig_result_dir, accuracy_threshold, 95)
- select_top_results(orig_result_dir + "/high_confidence")
-
-
- #self.start_process.kill()
- kill(self.start_process.pid)
-
- """
- called at the end of autotuning with the best resultsdb.models.Configuration
- """
- print "Final configuration", configuration.data
-
- return
-
-
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- createFlagsFile("opentuner_flags", configuration.data)
- run_cmd = binary_name
- run_result_call_program = self.call_program(run_cmd)
-
- accuracy = getAccuracy("final_accuracy")
- shutil.copy('accuracy_summary', result_dir + '/' + binary_name + '_final_' + str(accuracy) )
-
- sorted_list = sorted(self.configs_list, key = lambda tup: tup[0])
- print sorted_list[0:10]
-
- top_elems = 20
- if len(sorted_list) < top_elems:
- top_elems = len(sorted_list)
-
-
- for i in range(top_elems):
- createFlagsFile("opentuner_flags", sorted_list[i][2])
- run_cmd = binary_name
- run_result_call_program = self.call_program(run_cmd)
- accuracy = getAccuracy("final_accuracy")
- shutil.copy('accuracy_summary', result_dir + '/' + binary_name + '_' + str(accuracy) + "_rank_" + str(i) )
-
-
- #os.mkdir(result_dir + "full_results")
-
-
-
-
-if __name__ == '__main__':
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='name of binary to run')
- argparser.add_argument('--num-flags', type=int, help='num of flags to tune for')
- argparser.add_argument('--error-range', type=int, help='num of flags to tune for')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='accuracy threshold')
-
-
- args = argparser.parse_args()
- binary_name = str(args.binary)
- print("binary_name = ", binary_name)
- num_flags = int(args.num_flags)
- error_range = int(args.error_range)
- accuracy_threshold = float(args.accuracy)
- print("accuracy = ", accuracy_threshold)
- result_dir = args.result_dir
- orig_result_dir = result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
-
- output_dir = result_dir + "/full_results"
- print output_dir
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- if not os.path.exists(output_dir):
- print("Creating output directory = ", output_dir)
- os.mkdir(output_dir)
-
- for j in range(error_range):
- flag_ranges.append(j)
-
- print("flag_ranges = ", flag_ranges)
-
- for i in range(num_flags):
- tuning_flags.append("flag" + str(i))
-
- ClangFlagsTuner.main(argparser.parse_args())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/adddeps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/adddeps.py
deleted file mode 100644
index 61fd4757d6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/adddeps.py
+++ /dev/null
@@ -1,5 +0,0 @@
-# we would prefer a symbolic link, but it does not work on windows
-import os
-target = os.path.join(os.path.dirname(__file__),
- '../opentuner/utils/adddeps.py')
-execfile(target, dict(__file__=target))
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/algo_tuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/algo_tuner.py
deleted file mode 100644
index b8145e1798..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/algo_tuner.py
+++ /dev/null
@@ -1,318 +0,0 @@
-#!/usr/bin/env python
-#
-# Algorithmic Approximation Tuning
-# Purpose: Tunes for Perforation, Sampling, Numerical Precision (FP16)
-
-
-import adddeps
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-import subprocess
-import threading
-import psutil
-
-from measure_confidence2 import dump_promise_confidence_files3
-from measure_confidence2 import getConfidence, getMinAccuracy
-from select_top_results import select_top_results
-from time import sleep
-from pareto_curve import findParetoConfigs
-
-
-
-
-class TunerData:
- def __init__(self):
- self.binary_path = ""
- self.output_dir = ""
- self.num_layers = 0
- self.knobs_list = []
- self.knobs_speedup = {}
- self.accuracy_threshold = 0
- self.test_id = 0
- self.layer_costs = []
- self.tuning_flags = []
- self.autotuner_runs = 0
-
-
-
-tunerData = TunerData()
-
-
-def readCostFile(file_path):
-
- layer_costs = []
- f = open(file_path)
- for x in f:
- cost = float(x.strip())
- layer_costs.append(cost)
-
- print ("len(layer_costs) = ", layer_costs)
- f.close()
-
- return layer_costs
-
-
-
-def getAccuracy(file_name):
-
- file = open(file_name, "r")
- acc_str = file.read()
- file.close()
- accuracy = float(acc_str)
-
- try:
- accuracy = float(acc_str)
- except:
- return 20
-
- print (accuracy)
- return accuracy
-
-
-
-def createFlagsFile(file_name, cfg):
-
- f = open(file_name, "w+")
- cmd_config = ""
- for i in range(tunerData.num_layers): # flag in tunerData.tuning_flags:
- flag = tunerData.tuning_flags[i]
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-
-def readLayerKnobs(file_path):
-
- f = open(file_path, "r")
- knobs_list = []
- for x in f:
- knobs = []
- vals = x.split(",")
- for val in vals:
- knobs.append(int(val))
-
- knobs_list.append(knobs)
-
- print ("knobs_list = ", knobs_list)
-
- return knobs_list
-
-
-
-def readKnobConfig(file_path):
-
- knobs_speedup = {}
- f = open(file_path, "r")
- for x in f:
- toks = x.split("\t")
- ID = int(toks[0].split(",")[1])
-
- speedup = float(toks[2])
- knobs_speedup[ID] = speedup
-
- print ("knobs_speedup = ", knobs_speedup)
-
- return knobs_speedup
-
-
-
-
-def getConfigCost(cfg):
-
- orig_cost = 0.0
- total_cost = 0.0
- for it in range(tunerData.num_layers):
- flag = tunerData.tuning_flags[it]
- flag_value = cfg[flag]
- op_cost = tunerData.layer_costs[it]
- speedup = tunerData.knobs_speedup[flag_value]
-
- total_cost += (op_cost * 1.0 / speedup * 1.0)
- orig_cost += op_cost
-
- it += 1
-
- speedup = (orig_cost * 1.0) / (total_cost * 1.0)
-
- return total_cost, speedup
-
-
-
-def appendTopLine(f_path, accuracy, total_runs, total_comps, speedup):
-
- f_str = open(f_path, "r").read()
-
- f_out = open(f_path, "w+")
-
- f_out.write("total_runs=" + str(total_runs) + "\tconfidence=100.0" + "\tavg_accuracy=" + str(accuracy) + "\tconfig_cost=" + str(total_comps) + "\tspeedup=" + str(speedup) + "\n" )
- f_out.write(f_str)
-
- f_out.close()
-
-
-
-
-
-class ClangFlagsTuner(MeasurementInterface):
-
- def __init__(self, args):
- objective = ThresholdAccuracyMinimizeTime(tunerData.accuracy_threshold)
- input_manager = FixedInputManager(size=tunerData.num_layers)
- self.configs_list = []
-
- super(ClangFlagsTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
-
- for i in range(tunerData.num_layers):
- tunerData.tuning_flags.append("flag" + str(i))
-
-
- #for flag in tunerData.tuning_flags:
- for ind in range(tunerData.num_layers):
- flag = tunerData.tuning_flags[ind]
- manipulator.add_parameter(
- EnumParameter(flag, tunerData.knobs_list[ind]))
-
- print ("ind = ", ind, " len = ", len(tunerData.knobs_list))
- print (tunerData.knobs_list[ind])
- ind += 1
-
- return manipulator
-
-
-
- def run(self, desired_result, input, limit):
-
- """
- Run a given configuration then
- return performance
- """
- global test_id
-
- cfg = desired_result.configuration.data
-
- # NOTE: creates the file with flags read by the runtime
- createFlagsFile("promise_flags", cfg)
-
- run_cmd = tunerData.binary_path
- print "\nbinary_path = ", run_cmd
-
-
- total_runs = 1 # NOTE: Single run sufficient in Algorithmic Approx Tuner
- FNULL = open(os.devnull, 'wb')
- p = subprocess.Popen([run_cmd, str(total_runs)], stdout = FNULL)
- p.wait()
-
-
- accuracy = getAccuracy("final_accuracy")
-
- # getConfigCost returns the cost associated with the selected configuration
- total_comps, speedup = getConfigCost(cfg)
-
-
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- #Result.accuracy = accuracy
- min_accuracy = getMinAccuracy("run_accuracies.txt")
- print ("min_accuracy = ", min_accuracy)
- Result.accuracy = min_accuracy
-
- if min_accuracy > tunerData.accuracy_threshold:
- config_tuple = (total_comps, accuracy, cfg)
- self.configs_list.append(config_tuple)
- f_path = tunerData.output_dir + '/' + tunerData.binary_path + '_' + str(tunerData.test_id)
- shutil.copy('promise_flags', f_path)
-
- appendTopLine(f_path, accuracy, total_runs, total_comps, speedup)
-
- f_acc = open(tunerData.output_dir + '/' + tunerData.binary_path + '_' + str(tunerData.test_id) + "_accuracy", "w")
- f_acc.write(str(accuracy))
- f_acc.close()
-
-
- tunerData.test_id += 1
-
- return Result
-
-
- def save_final_config(self, configuration):
-
- print "Done with Autotuning Run \n"
- sleep(2)
-
- print "Final configuration", configuration.data
-
- return
-
-
-
-
-if __name__ == '__main__':
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='path to target binary')
- argparser.add_argument('--num-layers', type=int, help='num of flags to tune')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='result directory')
- argparser.add_argument('--cost-file', help='layer description')
- argparser.add_argument('--knobs-config', help='knob settings and ID mapping')
- argparser.add_argument('--layer-knobs', help='per-layer Knobs')
-
-
- args = argparser.parse_args()
-
- tunerData.binary_path = str(args.binary)
- tunerData.num_layers = int(args.num_layers)
- tunerData.accuracy_threshold = float(args.accuracy)
-
-
- # NOTE: Reading the cost file (with No of ops) to better guide the Autotuner
- cost_file_path = args.cost_file
- tunerData.layer_costs = readCostFile(cost_file_path)
-
-
- tunerData.knobs_list = readLayerKnobs(args.layer_knobs)
- tunerData.knobs_speedup = readKnobConfig(args.knobs_config)
-
- result_dir = args.result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
- tunerData.output_dir = result_dir + "/high_confidence/"
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- if not os.path.exists(tunerData.output_dir):
- print("Creating output directory = ", tunerData.output_dir)
- os.mkdir(tunerData.output_dir)
-
-
-
- ClangFlagsTuner.main(argparser.parse_args())
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/algo_tuner2.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/algo_tuner2.py
deleted file mode 100644
index 4ca0062f93..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/algo_tuner2.py
+++ /dev/null
@@ -1,339 +0,0 @@
-#!/usr/bin/env python
-#
-# Algorithmic Approximation Tuning
-# Purpose: Tunes for Perforation, Sampling, Numerical Precision (FP16)
-
-
-import adddeps
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-import subprocess
-import threading
-import psutil
-
-from measure_confidence2 import dump_promise_confidence_files4
-from measure_confidence2 import getConfidence, getMinAccuracy
-from select_top_results import select_top_results
-from time import sleep
-from pareto_curve import findParetoConfigs
-
-
-
-
-class TunerData:
- def __init__(self):
- self.binary_path = ""
- self.output_dir = ""
- self.num_layers = 0
- self.knobs_list = []
- self.knobs_speedup = {}
- self.accuracy_threshold = 0
- self.test_id = 0
- self.layer_costs = []
- self.tuning_flags = []
- self.autotuner_runs = 0
-
-
-
-tunerData = TunerData()
-
-
-orig_result_dir = ""
-
-
-def readCostFile(file_path):
-
- layer_costs = []
- f = open(file_path)
- for x in f:
- cost = float(x.strip())
- layer_costs.append(cost)
-
- print ("len(layer_costs) = ", layer_costs)
- f.close()
-
- return layer_costs
-
-
-
-def getAccuracy(file_name):
-
- file = open(file_name, "r")
- acc_str = file.read()
- file.close()
- accuracy = float(acc_str)
-
- try:
- accuracy = float(acc_str)
- except:
- return 20
-
- print (accuracy)
- return accuracy
-
-
-
-def createFlagsFile(file_name, cfg):
-
- f = open(file_name, "w+")
- cmd_config = ""
- for i in range(tunerData.num_layers): # flag in tunerData.tuning_flags:
- flag = tunerData.tuning_flags[i]
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-
-def readLayerKnobs(file_path):
-
- f = open(file_path, "r")
- knobs_list = []
- for x in f:
- knobs = []
- vals = x.split(",")
- for val in vals:
- knobs.append(int(val))
-
- knobs_list.append(knobs)
-
- print ("knobs_list = ", knobs_list)
-
- return knobs_list
-
-
-
-def readKnobConfig(file_path):
-
- knobs_speedup = {}
- f = open(file_path, "r")
- for x in f:
- toks = x.split("\t")
- ID = int(toks[0].split(",")[1])
-
- speedup = float(toks[2])
- knobs_speedup[ID] = speedup
-
- print ("knobs_speedup = ", knobs_speedup)
-
- return knobs_speedup
-
-
-
-
-def getConfigCost(cfg):
-
- orig_cost = 0.0
- total_cost = 0.0
- for it in range(tunerData.num_layers):
- flag = tunerData.tuning_flags[it]
- flag_value = cfg[flag]
- op_cost = tunerData.layer_costs[it]
- speedup = tunerData.knobs_speedup[flag_value]
-
- total_cost += (op_cost * 1.0 / speedup * 1.0)
- orig_cost += op_cost
-
- it += 1
-
- speedup = (orig_cost * 1.0) / (total_cost * 1.0)
-
- return total_cost, speedup
-
-
-
-def appendTopLine(f_path, accuracy, total_runs, total_comps, speedup):
-
- f_str = open(f_path, "r").read()
-
- f_out = open(f_path, "w+")
-
- f_out.write("total_runs=" + str(total_runs) + "\tconfidence=100.0" + "\tavg_accuracy=" + str(accuracy) + "\tconfig_cost=" + str(total_comps) + "\tspeedup=" + str(speedup) + "\n" )
- f_out.write(f_str)
-
- f_out.close()
-
-
-
-
-
-class ClangFlagsTuner(MeasurementInterface):
-
- def __init__(self, args):
- objective = ThresholdAccuracyMinimizeTime(tunerData.accuracy_threshold)
- input_manager = FixedInputManager(size=tunerData.num_layers)
- self.configs_list = []
-
- super(ClangFlagsTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
-
- for i in range(tunerData.num_layers):
- tunerData.tuning_flags.append("flag" + str(i))
-
-
- #for flag in tunerData.tuning_flags:
- for ind in range(tunerData.num_layers):
- flag = tunerData.tuning_flags[ind]
- manipulator.add_parameter(
- EnumParameter(flag, tunerData.knobs_list[ind]))
-
- print ("ind = ", ind, " len = ", len(tunerData.knobs_list))
- print (tunerData.knobs_list[ind])
- ind += 1
-
- return manipulator
-
-
-
- def run(self, desired_result, input, limit):
-
- """
- Run a given configuration then
- return performance
- """
- global test_id
-
- cfg = desired_result.configuration.data
-
- # NOTE: creates the file with flags read by the runtime
- createFlagsFile("promise_flags", cfg)
-
- run_cmd = tunerData.binary_path
- print "\nbinary_path = ", run_cmd
-
-
- input_size = 5000
- offset = 5000
-
- total_runs = 2 # NOTE: Single run sufficient in Algorithmic Approx Tuner
- FNULL = open(os.devnull, 'wb')
- p = subprocess.Popen([run_cmd, str(total_runs), str(tunerData.accuracy_threshold), str(1), str(input_size), str(offset) ], stdout = FNULL)
- p.wait()
-
- #total_runs = 2 # NOTE: Atleast two runs for promise tuner
- #FNULL = open(os.devnull, 'wb')
- #p = subprocess.Popen([run_cmd, str(total_runs)], stdout = FNULL)
- #p.wait()
-
-
- accuracy = getAccuracy("final_accuracy")
-
- # getConfigCost returns the cost associated with the selected configuration
- total_comps, speedup = getConfigCost(cfg)
-
-
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- #Result.accuracy = accuracy
- min_accuracy = getMinAccuracy("run_accuracies.txt")
- print ("min_accuracy = ", min_accuracy)
- Result.accuracy = min_accuracy
-
- if min_accuracy > tunerData.accuracy_threshold:
- config_tuple = (total_comps, accuracy, cfg)
- self.configs_list.append(config_tuple)
- f_path = tunerData.output_dir + '/' + tunerData.binary_path + '_' + str(tunerData.test_id)
- shutil.copy('promise_flags', f_path)
-
- appendTopLine(f_path, accuracy, total_runs, total_comps, speedup)
-
-
- tunerData.test_id += 1
-
- return Result
-
-
- def save_final_config(self, configuration):
-
- print "Done with Autotuning Run \n"
- sleep(2)
-
-
- #findParetoConfigs(orig_result_dir, layer_costs, accuracy_threshold)
-
- input_dir = orig_result_dir + "/full_results/"
- output_dir = orig_result_dir + "/high_confidence/"
-
- # Only dumping files with 95% confidence
- dump_promise_confidence_files4(tunerData.binary_path, input_dir, output_dir, tunerData.layer_file, tunerData.num_layers, tunerData.accuracy_threshold, tunerData.layer_costs, 95, tunerData.knobs_speedup)
-
-
- print "Final configuration", configuration.data
-
- return
-
-
-
-if __name__ == '__main__':
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='path to target binary')
- argparser.add_argument('--num-layers', type=int, help='num of flags to tune')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='result directory')
-
- argparser.add_argument('--layer-file', help='layer file')
-
- argparser.add_argument('--cost-file', help='layer description')
- argparser.add_argument('--knobs-config', help='knob settings and ID mapping')
- argparser.add_argument('--layer-knobs', help='per-layer Knobs')
-
-
- args = argparser.parse_args()
-
- tunerData.binary_path = str(args.binary)
- tunerData.num_layers = int(args.num_layers)
- tunerData.accuracy_threshold = float(args.accuracy)
-
- tunerData.layer_file = args.layer_file
-
- # NOTE: Reading the cost file (with No of ops) to better guide the Autotuner
- cost_file_path = args.cost_file
- tunerData.layer_costs = readCostFile(cost_file_path)
-
-
- tunerData.knobs_list = readLayerKnobs(args.layer_knobs)
- tunerData.knobs_speedup = readKnobConfig(args.knobs_config)
-
- result_dir = args.result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
- orig_result_dir = result_dir
- tunerData.output_dir = result_dir + "/full_results/"
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- if not os.path.exists(tunerData.output_dir):
- print("Creating output directory = ", tunerData.output_dir)
- os.mkdir(tunerData.output_dir)
-
-
-
- ClangFlagsTuner.main(argparser.parse_args())
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/approxhpvm_tuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/approxhpvm_tuner.py
deleted file mode 100644
index 9ae2266bf4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/approxhpvm_tuner.py
+++ /dev/null
@@ -1,262 +0,0 @@
-#!/usr/bin/env python
-#
-# Optimize blocksize of apps/mmm_block.cpp
-#
-# This is an extremely simplified version meant only for tutorials
-#
-import adddeps # fix sys.path
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-import subprocess
-import threading
-import psutil
-
-from measure_confidence2 import dump_high_confidence_files
-from select_top_results import select_top_results
-from time import sleep
-
-
-output_dir = ""
-flag_ranges = []
-tuning_flags = []
-binary_name = ""
-accuracy_threshold = 10.0
-opt_confs_index = 9
-evaluated_configs = {}
-orig_result_dir = ""
-
-
-def copyTunerRuntime():
- tensor_rt_path = os.environ["LLVM_SRC_ROOT"]
- if tensor_rt_path == "":
- print "LLVM_SRC_ROOT NOT SET"
- sys.exit(0)
-
- print "tensor_rt_path = ", tensor_rt_path
-
-
-
-
-def extractTotalOverhead(file_name):
-
- total_comps = 0.0
- file = open(file_name, "r")
- for x in file:
- words = x.split()
- total_comps += float(words[opt_confs_index])
-
- print total_comps
- return total_comps
-
-
-def getAccuracy(file_name):
-
- file = open(file_name, "r")
- acc_str = file.read()
- file.close()
-
- accuracy = float(acc_str)
-
- try:
- accuracy = float(acc_str)
- except:
- return 20
-
- print accuracy
- return accuracy
-
-
-
-def kill(proc_pid):
- process = psutil.Process(proc_pid)
- for proc in process.children(recursive=True):
- proc.kill()
- process.kill()
-
-
-
-def createFlagsFile(file_name, cfg):
-
- f = open(file_name, "w+")
- cmd_config = ""
- for flag in tuning_flags:
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-class ClangFlagsTuner(MeasurementInterface):
-
- def __init__(self, args):
-
- objective = ThresholdAccuracyMinimizeTime(accuracy_threshold)
- input_manager = FixedInputManager(size=num_flags)
- self.configs_list = []
-
- super(ClangFlagsTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- for flag in tuning_flags:
- manipulator.add_parameter(
- EnumParameter(flag, flag_ranges
- # [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
- )) #default is needed, optimizations don't work without it(tried and tested)
- return manipulator
-
-
- def run(self, desired_result, input, limit):
-
- """
- Run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
-
- # NOTE: creates the file with flags read by the runtime
- createFlagsFile("opentuner_flags", cfg)
-
- run_cmd = binary_name
- print "binary_name = ", run_cmd
- #run_result_call_program = self.call_program(run_cmd)
- #print "returned \n\n"
-
- FNULL = open(os.devnull, 'wb')
- p = subprocess.Popen(run_cmd, stdout = FNULL)
- p.wait()
-
- total_comps = extractTotalOverhead("accuracy_summary")
- accuracy = getAccuracy("final_accuracy")
-
- #Result = opentuner.resultsdb.models.Result(time=total_comps)
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- Result.accuracy = accuracy
-
- if accuracy > accuracy_threshold:
- if accuracy not in evaluated_configs:
- config_tuple = (total_comps, accuracy, cfg)
- self.configs_list.append(config_tuple)
- evaluated_configs[accuracy] = 1
- shutil.copy('accuracy_summary', output_dir + '/' + binary_name + '_' + str(accuracy))
-
-
- print "done with one run"
-
- return Result
-
-
- def save_final_config(self, configuration):
-
- print "Dumping High Confidence results"
- sleep(5)
-
- # Only dumping files with 95% confidence
- dump_high_confidence_files(binary_name, orig_result_dir, accuracy_threshold, 95)
- select_top_results(orig_result_dir + "/high_confidence")
-
-
- #self.start_process.kill()
- kill(self.start_process.pid)
-
- """
- called at the end of autotuning with the best resultsdb.models.Configuration
- """
- print "Final configuration", configuration.data
-
- return
-
-
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- createFlagsFile("opentuner_flags", configuration.data)
- run_cmd = binary_name
- run_result_call_program = self.call_program(run_cmd)
-
- accuracy = getAccuracy("final_accuracy")
- shutil.copy('accuracy_summary', result_dir + '/' + binary_name + '_final_' + str(accuracy) )
-
- sorted_list = sorted(self.configs_list, key = lambda tup: tup[0])
- print sorted_list[0:10]
-
- top_elems = 20
- if len(sorted_list) < top_elems:
- top_elems = len(sorted_list)
-
-
- for i in range(top_elems):
- createFlagsFile("opentuner_flags", sorted_list[i][2])
- run_cmd = binary_name
- run_result_call_program = self.call_program(run_cmd)
- accuracy = getAccuracy("final_accuracy")
- shutil.copy('accuracy_summary', result_dir + '/' + binary_name + '_' + str(accuracy) + "_rank_" + str(i) )
-
-
- #os.mkdir(result_dir + "full_results")
-
-
-
-
-if __name__ == '__main__':
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='name of binary to run')
- argparser.add_argument('--num-flags', type=int, help='num of flags to tune for')
- argparser.add_argument('--error-range', type=int, help='num of flags to tune for')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='accuracy threshold')
-
-
- args = argparser.parse_args()
- binary_name = str(args.binary)
- print("binary_name = ", binary_name)
- num_flags = int(args.num_flags)
- error_range = int(args.error_range)
- accuracy_threshold = float(args.accuracy)
- print("accuracy = ", accuracy_threshold)
- result_dir = args.result_dir
- orig_result_dir = result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
-
- output_dir = result_dir + "/full_results"
- print output_dir
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- if not os.path.exists(output_dir):
- print("Creating output directory = ", output_dir)
- os.mkdir(output_dir)
-
- for j in range(error_range):
- flag_ranges.append(j)
-
- print("flag_ranges = ", flag_ranges)
-
- for i in range(num_flags):
- tuning_flags.append("flag" + str(i))
-
- ClangFlagsTuner.main(argparser.parse_args())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/devtuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/devtuner.py
deleted file mode 100644
index 4d5da6afb6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/devtuner.py
+++ /dev/null
@@ -1,331 +0,0 @@
-#!/usr/bin/env python
-#
-# Development-time Tuner with Algorithmic Approximations:
-# Approximations: Perforation, Sampling with varying knobs for rate, skip offset
-
-
-import adddeps
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-import subprocess
-import threading
-import psutil
-import thread
-
-from select_top_results import select_top_results
-from time import sleep
-from pareto_curve import findParetoConfigs
-import utils
-
-
-
-class TunerData:
- def __init__(self):
- self.binary_path = ""
- self.output_dir = ""
- self.num_layers = 0
- self.knobs_list = []
- self.knobs_speedup = {}
- self.accuracy_threshold = 0
- self.accuracy_slack = 0
- self.test_id = 0
- self.layer_costs = []
- self.tuning_flags = []
- self.autotuner_runs = 0
- self.best_speedup = 1
- self.log_file = ""
-
- self.use_seed = True
-
-
-
-class DevTuner(MeasurementInterface):
-
-
- def initTunerData(self, args):
-
- self.tunerData.binary_path = str(args.binary)
- self.tunerData.num_layers = int(args.num_layers)
- self.tunerData.accuracy_threshold = float(args.accuracy)
- self.tunerData.accuracy_slack = float(args.accuracy_slack)
-
- # NOTE: Reading the cost file (with No of ops) to better guide the Autotuner
- cost_file_path = args.cost_file
- self.tunerData.layer_costs = utils.readCostFile(cost_file_path)
-
- self.tunerData.knobs_list = utils.readLayerKnobs(args.layer_knobs)
- self.tunerData.knobs_speedup = utils.readGlobalKnobConfig(args.knobs_config)
- self.tunerData.test_id = args.start_id
-
- result_dir = args.result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- self.tunerData.output_dir = result_dir + "/high_confidence/"
- if not os.path.exists(self.tunerData.output_dir):
- print("Creating output directory = ", self.tunerData.output_dir)
- os.mkdir(self.tunerData.output_dir)
-
-
- def createFIFO(self):
- try:
- os.mkfifo("/tmp/opentuner_fifo")
- except OSError, e:
- print("FIFO exists")
-
-
- def invokeBinary(self, runs):
-
- print ("\n\n\n\n SPAWNING BINARY *****\n\n***")
- run_cmd = self.tunerData.binary_path
- print "\nbinary_path = ", run_cmd
- FNULL = open(os.devnull, 'wb')
- self.process = subprocess.Popen([run_cmd, str(runs)], stdout = FNULL)
-
-
-
- def signalPipe(self):
-
- fifo = open("/tmp/opentuner_fifo", "w")
- fifo.write("start_run")
- fifo.close()
- #print "Waiting for process to signal back - when done processing one run"
-
-
-
- def pollOnProcess(self, self2):
-
- print (" self.piped_execution = ", self.piped_execution, "*** \n")
- sleep(5)
-
- while (not self.escape_poll_thread):
- poll = self.process.poll()
- #print ("POLLING")
-
- if poll is not None: # If process aborted, invoke another instance
- sleep(6)
-
- poll = self.process.poll()
- if not utils.check_pid(self.process.pid) and poll is not None: # Second check for process existence
- self.corrupted_run = True
- utils.process_kill(self.process.pid) # Kill existing process if exists
- self.invokeBinary(100000)
- self.signalPipe()
-
-
-
- def waitOnPipe(self):
-
- fifo2 = open("/tmp/opentuner_fifo", "r")
- fifo2.read()
- fifo2.close()
-
-
-
- def stopProcess(self):
-
- fifo = open("/tmp/opentuner_fifo", "w")
- fifo.write("stop_run")
- fifo.close()
- print "***** Sending Stop Signal ***** "
-
-
- def __init__(self, args):
-
- #print ("\n\n\n\n\n\******* ARGS[0] = ", args)
-
- self.tunerData = TunerData()
- self.initTunerData(args)
-
- # Adding knob to use piped execution instead
- self.piped_execution = True
- self.corrupted_run = False
- self.escape_poll_thread = False
-
- objective = ThresholdAccuracyMinimizeTime(self.tunerData.accuracy_threshold)
- input_manager = FixedInputManager(size=self.tunerData.num_layers)
- self.configs_list = []
- # initializing tuner related data
-
- super(DevTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
- log_path = self.tunerData.output_dir + "/log_file.txt"
- self.log_file = open(log_path, "a+")
-
-
- if self.piped_execution:
- self.createFIFO()
- self.invokeBinary(100000)
- print ("Invoking thread to launch a Polling THREAD ")
- sleep(10)
- thread.start_new_thread(self.pollOnProcess, (self, ))
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
-
- for i in range(self.tunerData.num_layers):
- self.tunerData.tuning_flags.append("flag" + str(i))
-
-
- for ind in range(self.tunerData.num_layers):
- flag = self.tunerData.tuning_flags[ind]
- manipulator.add_parameter(
- EnumParameter(flag, self.tunerData.knobs_list[ind]))
-
- print ("ind = ", ind, " len = ", len(self.tunerData.knobs_list))
- print (self.tunerData.knobs_list[ind])
- ind += 1
-
- return manipulator
-
-
-
- def seed_configurations(self):
- """Provide baseline config as seed if model uses seed."""
- baseline_config = {layer: 11 for layer in self.tunerData.tuning_flags}
- return [baseline_config] if self.tunerData.use_seed else []
-
-
-
- def run(self, desired_result, input, limit):
-
- """
- Run a given configuration then
- return performance
- """
- global test_id
-
- cfg = desired_result.configuration.data
-
-
- print ("cfg = ", cfg)
-
- # NOTE: creates flags file used by hpvm-tensor-rt
- utils.genLayerFlagsFile("promise_flags", cfg, self.tunerData)
-
-
- total_runs = 1 # NOTE: Single run sufficient in Algorithmic Approx Tuner
- if not self.piped_execution:
- run_cmd = self.tunerData.binary_path
- print "\nbinary_path = ", run_cmd
- FNULL = open(os.devnull, 'wb')
- p = subprocess.Popen([run_cmd, str(total_runs)], stdout = FNULL)
- p.wait()
-
-
- waitSignal = 0
- if self.piped_execution:
- self.signalPipe()
- waitSignal = self.waitOnPipe()
-
- accuracy = utils.readAccuracy("final_accuracy")
-
- if self.corrupted_run == True:
- accuracy = self.tunerData.accuracy_slack - 5 # Invalid Run
- print ("\n\n\n **** Corrupted Run **** Accuracy = ", accuracy, " --- \n\n\n")
- self.corrupted_run = False
-
- # getConfigCost returns the cost associated with the selected configuration
- total_comps, speedup = utils.computeConfigCost(cfg, self.tunerData)
-
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- Result.accuracy = accuracy
-
-
- if accuracy > self.tunerData.accuracy_slack:
- config_tuple = (total_comps, accuracy, cfg)
- self.configs_list.append(config_tuple)
- f_path = self.tunerData.output_dir + '/' + self.tunerData.binary_path + '_' + str(self.tunerData.test_id)
- shutil.copy('promise_flags', f_path)
-
- utils.addInfoToOutFile(f_path, accuracy, total_runs, total_comps, speedup)
-
- print ("------ Config Chosen with Accuracy = ", accuracy, " And Predicted Speedup = ", speedup, "\n")
-
- if speedup > self.tunerData.best_speedup:
- self.tunerData.best_speedup = speedup
-
-
-
- if self.tunerData.test_id % 100 == 0:
- self.log_file.write("** iteration = " + str(self.tunerData.test_id) + \
- " speedup = " + str(self.tunerData.best_speedup) + " \n")
-
-
-
-
- self.tunerData.test_id += 1
-
- return Result
-
-
-
- def save_final_config(self, configuration):
-
- # Indication to terminate polling thread
- self.escape_poll_thread = True
-
- if self.piped_execution:
- #self.stopProcess()
- utils.process_kill(self.process.pid)
- print ("Killed hanging process")
-
- print "Final configuration", configuration.data
-
- # Close log file
- self.log_file.close()
-
- print "Done with Autotuning Run \n"
- sleep(2)
-
- return
-
-
-
-
-if __name__ == '__main__':
-
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='path to target binary')
- argparser.add_argument('--num-layers', type=int, help='num of flags to tune')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--accuracy-slack', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='result directory')
- argparser.add_argument('--cost-file', help='layer description')
- argparser.add_argument('--knobs-config', help='knob settings and ID mapping')
- argparser.add_argument('--layer-knobs', help='per-layer Knobs')
- # NOTE: needed to have unique file-names across runs
- argparser.add_argument('--start-id', type=int, help='start id for naming output files')
-
-
- args = argparser.parse_args()
- #devTuner = DevTuner(args)
- print ("\n\n\n\n\ -- NOTE --- \n\n")
- DevTuner.main(argparser.parse_args())
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/gettingstarted.md b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/gettingstarted.md
deleted file mode 100644
index 8a442c5f44..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/gettingstarted.md
+++ /dev/null
@@ -1,215 +0,0 @@
----
-layout: default
-title: OpenTuner - Using OpenTuner
-permalink: /tutorial/gettingstarted/index.html
----
-
-Tutorial: Optimizing Block Matrix Multiplication
-================================================
-
-This tutorial assumes that you have checked out a copy of opentuner. For
-guidelines on how to get opentuner set up, refer [here][setup].
-
-[setup]: http://opentuner.org/tutorial/setup/
-
-Identifying a Program to Autotune
----------------------------------
-
-In order to do autotuning, you first need something to autotune. This will
-normally be your own program that you want to make either fast or better in
-some way. For this tutorial we will use a blocked version of matrix multiply
-as an example. We will use opentuner to find the optimal value of the block
-size parameter.
-
-We will autotune the sample code below(based off of modification of code
-found [here][matrix-multiply-code]), making sure to take the block size as
-a compile time constant to the program.
-
-[matrix-multiply-code]: http://csapp.cs.cmu.edu/public/waside/waside-blocking.pdf
-
-Save the sample code below to examples/tutorials/mmm_block.cpp
-
- #include <stdio.h>
- #include <cstdlib>
-
- #define N 100
-
- int main(int argc, const char** argv)
- {
-
- int n = BLOCK_SIZE * (N/BLOCK_SIZE);
- int a[N][N];
- int b[N][N];
- int c[N][N];
- int sum=0;
- for(int k1=0;k1<n;k1+=BLOCK_SIZE)
- {
- for(int j1=0;j1<n;j1+=BLOCK_SIZE)
- {
- for(int k1=0;k1<n;k1+=BLOCK_SIZE)
- {
- for(int i=0;i<n;i++)
- {
- for(int j=j1;j<j1+BLOCK_SIZE;j++)
- {
- sum = c[i][j];
- for(int k=k1;k<k1+BLOCK_SIZE;k++)
- {
- sum += a[i][k] * b[k][j];
- }
- c[i][j] = sum;
- }
- }
- }
- }
- }
- return 0;
- }
-
-Creating a New Autotuner with Opentuner
-------------------------------------
-Now we need to create a program that uses OpenTuner to optimize the program we just saved.
-
-Save the following code to examples/tutorials/mmm_tuner.py
-
- #!/usr/bin/env python
- #
- # Optimize blocksize of apps/mmm_block.cpp
- #
- # This is an extremely simplified version meant only for tutorials
- #
- import adddeps # fix sys.path
-
- import opentuner
- from opentuner import ConfigurationManipulator
- from opentuner import IntegerParameter
- from opentuner import MeasurementInterface
- from opentuner import Result
-
-
- class GccFlagsTuner(MeasurementInterface):
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- manipulator.add_parameter(
- IntegerParameter('blockSize', 1, 10))
- return manipulator
-
- def run(self, desired_result, input, limit):
- """
- Compile and run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
-
- gcc_cmd = 'g++ mmm_block.cpp '
- gcc_cmd += '-DBLOCK_SIZE='+ cfg['blockSize']
- gcc_cmd += ' -o ./tmp.bin'
-
- compile_result = self.call_program(gcc_cmd)
- assert compile_result['returncode'] == 0
-
- run_cmd = './tmp.bin'
-
- run_result = self.call_program(run_cmd)
- assert run_result['returncode'] == 0
-
- return Result(time=run_result['time'])
-
- def save_final_config(self, configuration):
- """called at the end of tuning"""
- print "Optimal block size written to mmm_final_config.json:", configuration.data
- self.manipulator().save_to_file(configuration.data,
- 'mmm_final_config.json')
-
-
- if __name__ == '__main__':
- argparser = opentuner.default_argparser()
- GccFlagsTuner.main(argparser.parse_args())
-
-
-This file consists of several components, each of which will be discussed in further detail below.
-
-Tuning Programs have a general structure as follows:
-
- from opentuner import MeasurementInterface
- from opentuner import Result
-
-Create an instance of class GccFlagsTuner, which tunes specified parameters using opentuner.
- class GccFlagsTuner(MeasurementInterface):
-
-The manipulator method defines the variable search space by specifying parameters that should be tuned by this instance of GccFlagsTuner
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- manipulator.add_parameter(
- IntegerParameter('blockSize', 1, 10))
- return manipulator
-
-The run method actually runs opentuner under the given configuration and returns the calculated performance under this configuration. In this example, the blockSize parameter to be tuned is input as a compile-time constant that takes on a value within the specified range each time it is run. However, opentuner also supports other methods of specifying these parameters that may be preferred in different use cases.
-
- def run(self, desired_result, input, limit):
- """
- Compile and run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
-
- gcc_cmd = 'g++ mmm_block.cpp '
- gcc_cmd += '-DBLOCK_SIZE='+ cfg['blockSize']
- gcc_cmd += ' -o ./tmp.bin'
-
- compile_result = self.call_program(gcc_cmd)
- assert compile_result['returncode'] == 0
-
- run_cmd = './tmp.bin'
-
- run_result = self.call_program(run_cmd)
- assert run_result['returncode'] == 0
-
- return Result(time=run_result['time'])
-
-We can actually display the result of running opentuner(the optimal block size for our multiplication problem) by creating a method, save_final_config() in our class. This saves a json dictionary of the optimal blockSize parameter found to the file mmm_final_config.json
-
- def save_final_config(self, configuration):
- """called at the end of tuning"""
- print "Optimal block size written to mmm_final_config.json:", configuration.data
- self.manipulator().save_to_file(configuration.data,
- 'mmm_final_config.json')
-
- if __name__ == '__main__':
- argparser = opentuner.default_argparser()
- GccFlagsTuner.main(argparser.parse_args())
-
-Generating and Viewing Results
-------------------------------
-
-Run the following command to autotune our program(The --no-dups flag hides warnings about duplicate results and the --stop-after parameter specifies that we are running opentuner for a maximum of 30 seconds):
-
- python mmm_tuner.py --no-dups --stop-after=30
-
-The results of each run configuration will be displayed as follows(output lines are truncated for readability here):
-
- [ 10s] INFO opentuner.search.plugin.DisplayPlugin: tests=10, best {'BLOCK_SIZE': 4}, cost time=0.0081, found by DifferentialEvolutionAlt[...]
- [ 19s] INFO opentuner.search.metatechniques: AUCBanditMetaTechniqueA: [('DifferentialEvolutionAlt', 477), ('UniformGreedyMutation', 18), ('NormalGreedyMutation', 5), ('RandomNelderMead', 1)]
- [ 20s] INFO opentuner.search.plugin.DisplayPlugin: tests=10, best {'BLOCK_SIZE': 4}, cost time=0.0081, found by DifferentialEvolutionAlt[...]
- [ 30s] INFO opentuner.search.plugin.DisplayPlugin: tests=10, best {'BLOCK_SIZE': 4}, cost time=0.0081, found by DifferentialEvolutionAlt[...]
- [ 30s] INFO opentuner.search.plugin.DisplayPlugin: tests=10, best {'BLOCK_SIZE': 4}, cost time=0.0081, found by DifferentialEvolutionAlt[...]
- Optimal block size written to mmm_final_config.json: {'BLOCK_SIZE': 4}
-
-
-Look up the optimal BlockSize value by inspecting the following created file:
-
- mmm_final_config.json
-
-In this example, the output file content was as follows:
-
- {'BLOCK_SIZE': 4}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/measure_confidence.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/measure_confidence.py
deleted file mode 100644
index dd7a050ac8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/measure_confidence.py
+++ /dev/null
@@ -1,290 +0,0 @@
-
-import argparse
-import os
-import sys
-from time import sleep
-
-
-def getAccuracy(file_name):
-
- if not os.path.exists(file_name):
- print("final_accuracy file not found ")
- sys.exit(0)
-
- file = open(file_name, "r")
- acc_str = file.read()
- accuracy = float(acc_str)
- print accuracy
- return accuracy
-
-
-total_runs = 40
-skip_lines = 0
-
-
-def test_func():
- print "test_func"
- sys.exit(0)
-
-
-def do_multiple_runs(binary_name, accuracy_threshold, confidence_threshold):
-
- #total_runs = 100.0
- successful_runs = 0.0
- total_acc = 0
-
- for i in range(int(total_runs)):
-
- fifo = open("/tmp/myfifo", "w")
- fifo.write("start_run")
- fifo.close()
-
- print "Waiting for process to signal back - when done processing one run"
-
- fifo2 = open("/tmp/myfifo", "r")
- fifo2.read()
- fifo2.close()
-
- print "Process Signalled back"
-
- accuracy = getAccuracy("final_accuracy")
- total_acc += accuracy
-
- if accuracy > accuracy_threshold:
- successful_runs += 1
-
- confidence = (successful_runs / (total_runs*1.0) ) * 100.0
- print("confidence = ", confidence)
- avg_acc = total_acc / total_runs
- print("average accuracy = ", avg_acc)
-
- return confidence, avg_acc
-
-
-def compute_confidence(binary_name, accuracy, confidence, result_dir, output_dir):
-
- confidence_list = []
-
- if not os.path.exists(result_dir):
- print("Path does not exist")
- sys.exit(0)
-
- file_names = os.listdir(result_dir)
- print file_names
-
- for file_name in file_names:
- # Skip sub-directories
- if os.path.isdir(result_dir + "/" + file_name):
- continue
-
- f = open(result_dir + "/" + file_name)
- tuner_file = open("opentuner_flags", "w+")
-
- index = 0
- results_str = ""
- for x in f:
- if index >= skip_lines:
- error_knob = int(float(x.split()[1]))
- print error_knob
- tuner_file.write(str(error_knob) + "\n")
-
- results_str += x
- index += 1
-
- tuner_file.close()
-
- run_confidence, avg_accuracy = do_multiple_runs(binary_name, accuracy, confidence)
-
- if run_confidence > 90:
- f2 = open(output_dir + "/" + file_name, "w+")
- f2.write("total_runs=" + str(total_runs) + "\t confidence=" + str(run_confidence) + "\t avg_accuracy=" + str(avg_accuracy) + "\n")
- f2.write(results_str)
- f2.close()
-
- conf_result = (run_confidence, avg_accuracy, file_name)
- confidence_list.append(conf_result)
-
- return confidence_list
-
-
-
-
-def compute_promise_confidence(binary_name, accuracy, confidence, result_dir, output_dir):
-
- confidence_list = []
-
- if not os.path.exists(result_dir):
- print("Path does not exist")
- sys.exit(0)
-
- file_names = os.listdir(result_dir)
- print file_names
-
- for file_name in file_names:
- # Skip sub-directories
- if os.path.isdir(result_dir + "/" + file_name):
- continue
-
- f = open(result_dir + "/" + file_name)
- tuner_file = open("opentuner_flags", "w+")
-
- config_str = f.read()
- tuner_file.write(config_str)
- tuner_file.close()
-
- run_confidence, avg_accuracy = do_multiple_runs(binary_name, accuracy, confidence)
-
- if run_confidence > 90:
- f2 = open(output_dir + "/" + file_name, "w+")
- f2.write("total_runs=" + str(total_runs) + "\t confidence=" + str(run_confidence) + "\t avg_accuracy=" + str(avg_accuracy) + "\n")
- f2.write(config_str)
- f2.close()
-
- flags_str = config_str.replace('\n', ',')
- conf_result = (run_confidence, avg_accuracy, file_name, flags_str)
- confidence_list.append(conf_result)
-
- return confidence_list
-
-
-
-
-def dump_high_confidence_files(binary, result_dir, accuracy, confidence):
-
- #result_dir = args.result_dir
- output_dir = result_dir + "/high_confidence"
- result_dir = result_dir + "/full_results"
-
- if not os.path.exists(output_dir):
- os.mkdir(output_dir)
-
-
- confidence_list = compute_confidence(binary, accuracy, confidence, result_dir, output_dir)
- print confidence_list
-
- # descending sort on confidence
- sorted_list = sorted(confidence_list, key = lambda tup: tup[0], reverse=True)
-
- output_file = open(output_dir + "/confidence_summary.txt", "w+")
- for x in sorted_list:
- output_file.write(str(x[0]) + "\t" + str(x[1]) + "\t" + str(x[2]) + "\n")
-
- output_file.close()
- print "Dumped Confidence Summary"
-
-
-def processLayerFile(layer_file_path):
-
- layer_sizes = []
- layer_file = open(layer_file_path, "r")
-
- for layer_size in layer_file:
- try:
- size = int(layer_size)
- layer_sizes.append(size)
- except:
- return layer_sizes
-
- return layer_sizes
-
-
-
-def getLayerConfigStr(config_str, layer_sizes, num_flags):
-
- new_config_str = ""
- config_vals = config_str.split(',')
- it_count = 0
- for val in config_vals:
- if val == "":
- continue
-
- config_val = int(val)
- # For FP32 and FP32 values, each tensor op needs to be annotated
- if config_val == 8 or config_val == 9:
- for i in range(layer_sizes[it_count] - 1):
- new_config_str += val + " "
- new_config_str += val
- if it_count < num_flags - 1:
- new_config_str += ","
- else:
- new_config_str += val
- if it_count < num_flags - 1:
- new_config_str += ","
-
- it_count += 1
-
- return new_config_str
-
-
-def dump_promise_confidence_files(binary, result_dir, layer_file_path, num_flags, accuracy, confidence):
-
- #result_dir = args.result_dir
- output_dir = result_dir + "/high_confidence"
- input_dir = result_dir + "/full_results"
-
- if not os.path.exists(output_dir):
- os.mkdir(output_dir)
-
-
- layer_sizes = processLayerFile(layer_file_path);
- print layer_sizes
- sleep(3)
-
- confidence_list = compute_promise_confidence(binary, accuracy, confidence, input_dir, output_dir)
- print confidence_list
-
- # Ascending sort on accuracy
- sorted_list = sorted(confidence_list, key = lambda tup: tup[1])
-
- promise_file = open(output_dir + "/promise_confs.txt", "w+")
- confidence_file = open(output_dir + "/confidence_summary.txt", "w+")
-
- max_configs = 50
- it_count = 0
- for x in sorted_list:
- if x[1] > accuracy and x[0] > confidence:
- config_str = getLayerConfigStr(x[3], layer_sizes, num_flags)
- promise_file.write(config_str + "\n")
- it_count += 1
- if it_count > max_configs:
- break
-
- confidence_file.write(str(x[0]) + "\t" + str(x[1]) + "\t" + str(x[3]) + "\n")
-
- promise_file.close()
- confidence_file.close()
-
- print "Dumped Confidence Summary"
-
-
-
-
-
-if __name__ == "__main__":
-
- argparser = argparse.ArgumentParser(description='runs best configs to get high confidence on accuracy')
- argparser.add_argument('--result-dir', help='Directory containing OpenTuner configurations')
- argparser.add_argument('--output-dir', help='Directory for storing output directory')
- argparser.add_argument('--binary', help='Binary name to run')
- argparser.add_argument('--accuracy', type=float, help='Accuracy constraint')
- argparser.add_argument('--confidence', type=float, help='Confidence threshold')
-
-
- args = argparser.parse_args()
- result_dir = args.result_dir
- output_dir = args.output_dir
- binary = args.binary
- accuracy = args.accuracy
- confidence = args.confidence
-
- confidence_list = compute_confidence(binary, accuracy, confidence, result_dir, output_dir)
- #print confidence_list
-
- sorted_list = sorted(confidence_list, key = lambda tup: tup[0], reverse=True)
-
- output_file = open(output_dir + "/confidence_summary.txt", "w+")
- for x in sorted_list:
- output_file.write(str(x[0]) + "\t" + str(x[1]) + "\t" + str(x[2]) + "\n")
-
- output_file.close()
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/measure_confidence2.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/measure_confidence2.py
deleted file mode 100644
index f5998ff3c8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/measure_confidence2.py
+++ /dev/null
@@ -1,664 +0,0 @@
-
-import argparse
-import os
-import sys
-import subprocess
-from time import sleep
-
-
-def getAccuracy(file_name):
-
- if not os.path.exists(file_name):
- print("final_accuracy file not found ")
- sys.exit(0)
-
- file = open(file_name, "r")
- acc_str = file.read()
- accuracy = float(acc_str)
- print accuracy
- return accuracy
-
-
-total_runs = 40.0
-fails_allowed = 3
-skip_lines = 0
-
-
-def test_func():
- print "test_func"
- sys.exit(0)
-
-
-
-def do_multiple_runs(binary_name, accuracy_threshold, confidence_threshold):
-
- successful_runs = 0.0
- unsuccessful_runs = 0.0
- total_acc = 0
-
- for i in range(int(total_runs)):
- FNULL = open(os.devnull, 'wb')
- p = subprocess.Popen(binary_name, stdout = FNULL)
- p.wait()
-
- accuracy = getAccuracy("final_accuracy")
- total_acc += accuracy
-
- if accuracy > accuracy_threshold:
- successful_runs += 1
- else:
- unsuccessful_runs += 1
-
- if unsuccessful_runs > 6:
- break
-
-
- confidence = (successful_runs / total_runs) * 100.0
- print("confidence = ", confidence)
- avg_acc = total_acc / total_runs
- print("average accuracy = ", avg_acc)
-
- return confidence, avg_acc
-
-
-
-def getConfidence(accuracy_outfile, acc_threshold):
-
- f = open(accuracy_outfile, "r")
-
- total_acc = 0.0
- failed = 0
- it = 0
-
- for x in f:
- acc = float(x.strip())
- if acc < acc_threshold:
- failed += 1
-
- total_acc += acc
- it += 1
-
- conf = (it * 1.0 - failed) / it * 100
- avg_acc = total_acc / it
-
- return conf, avg_acc
-
-
-
-def getMinAccuracy(accuracy_outfile):
-
- f = open(accuracy_outfile, "r")
-
- total_acc = 0.0
- failed = 0
- it = 0
-
- acc_list = []
- for x in f:
- acc = float(x.strip())
- acc_list.append(acc)
-
- return min(acc_list)
-
-
-# NOTE: invokes the binary with the number of runs
-def do_multiple_runs2(binary_name, accuracy_threshold, confidence_threshold):
-
- successful_runs = 0.0
- unsuccessful_runs = 0.0
- total_acc = 0
-
- FNULL = open(os.devnull, 'wb')
- p = subprocess.Popen([binary_name, str(int(total_runs)), str(accuracy_threshold), str(fails_allowed)], stdout = FNULL)
- p.wait()
-
- confidence, avg_acc = getConfidence("run_accuracies.txt", accuracy_threshold)
-
- print("confidence = ", confidence)
- print("average accuracy = ", avg_acc)
-
- return confidence, avg_acc
-
-
-
-
-
-def compute_confidence(binary_name, accuracy, confidence, result_dir, output_dir):
-
- confidence_list = []
-
- if not os.path.exists(result_dir):
- print("Path does not exist")
- sys.exit(0)
-
- file_names = os.listdir(result_dir)
- print file_names
-
- for file_name in file_names:
- # Skip sub-directories
- if os.path.isdir(result_dir + "/" + file_name):
- continue
-
- f = open(result_dir + "/" + file_name)
- tuner_file = open("opentuner_flags", "w+")
-
- index = 0
- results_str = ""
- for x in f:
- if index >= skip_lines:
- error_knob = int(float(x.split()[1]))
- print error_knob
- tuner_file.write(str(error_knob) + "\n")
-
- results_str += x
- index += 1
-
- tuner_file.close()
-
- run_confidence, avg_accuracy = do_multiple_runs2(binary_name, accuracy, confidence)
-
- if run_confidence >= 95:
- f2 = open(output_dir + "/" + file_name, "w+")
- f2.write("total_runs=" + str(total_runs) + "\t confidence=" + str(run_confidence) + "\t avg_accuracy=" + str(avg_accuracy) + "\n")
- f2.write(results_str)
- f2.close()
-
- conf_result = (run_confidence, avg_accuracy, file_name)
- confidence_list.append(conf_result)
-
- return confidence_list
-
-
-
-def dump_high_confidence_files(binary, result_dir, accuracy, confidence):
-
- #result_dir = args.result_dir
- output_dir = result_dir + "/high_confidence"
- result_dir = result_dir + "/full_results"
-
- if not os.path.exists(output_dir):
- os.mkdir(output_dir)
-
- confidence_list = compute_confidence(binary, accuracy, confidence, result_dir, output_dir)
- print confidence_list
-
- sorted_list = sorted(confidence_list, key = lambda tup: tup[0], reverse=True)
-
- output_file = open(output_dir + "/confidence_summary.txt", "w+")
- for x in sorted_list:
- output_file.write(str(x[0]) + "\t" + str(x[1]) + "\t" + str(x[2]) + "\n")
-
- output_file.close()
- print "Dumped Confidence Summary"
-
-
-
-
-def processLayerFile(layer_file_path):
-
- layer_sizes = []
- layer_file = open(layer_file_path, "r")
-
- for layer_desc in layer_file:
- try:
- toks = layer_desc.split(",")
- if len(toks) < 2: # Not layer size description
- continue
-
- size = int(toks[1])
- if "NML" in layer_desc:
- size = -1
- layer_sizes.append(size)
- except:
- return layer_sizes
-
- return layer_sizes
-
-
-
-def getLayerConfigStr(config_str, layer_sizes, num_flags):
-
- new_config_str = ""
- config_vals = config_str.split(',')
- it_count = 0
- layer_count = 0
-
- #for layer_size in val in config_vals:
- for layer_depth_size in layer_sizes:
-
- if layer_depth_size == -1:
- new_config_str += "8"
- layer_count += 1
- if layer_count < len(layer_sizes):
- new_config_str += ","
- continue
-
- val = config_vals[it_count]
- if val == "":
- continue
-
- config_val = int(val)
- # For FP32 and FP32 values, each tensor op needs to be annotated
- if config_val == 8 or config_val == 9:
- for i in range(layer_depth_size - 1):
- new_config_str += val + " "
- new_config_str += val
- if layer_count < len(layer_sizes) - 1:
- new_config_str += ","
- else:
- new_config_str += val
- if layer_count < len(layer_sizes) - 1:
- new_config_str += ","
-
- it_count += 1
- layer_count += 1
-
-
- return new_config_str
-
-
-
-def compute_promise_confidence(binary_name, accuracy, confidence, result_dir, output_dir):
-
- confidence_list = []
-
- if not os.path.exists(result_dir):
- print("Path does not exist")
- sys.exit(0)
-
- file_names = os.listdir(result_dir)
- print file_names
-
- for file_name in file_names:
- # Skip sub-directories
- if os.path.isdir(result_dir + "/" + file_name):
- continue
-
- f = open(result_dir + "/" + file_name)
- tuner_file = open("promise_flags", "w+")
-
- config_str = f.read()
- tuner_file.write(config_str)
- tuner_file.close()
-
- run_confidence, avg_accuracy = do_multiple_runs(binary_name, accuracy, confidence)
-
- if run_confidence >= 95:
- f2 = open(output_dir + "/" + file_name, "w+")
- f2.write("total_runs=" + str(total_runs) + "\t confidence=" + str(run_confidence) + "\t avg_accuracy=" + str(avg_accuracy) + "\n")
- f2.write(config_str)
- f2.close()
-
- flags_str = config_str.replace('\n', ',')
- conf_result = (run_confidence, avg_accuracy, file_name, flags_str)
- confidence_list.append(conf_result)
-
- return confidence_list
-
-
-
-def getConfigCost(layer_costs, config_str):
-
- tuning_flags = config_str.split("\n")
-
- it = 0
- orig_cost = 0.0
- total_cost = 0.0
- for flag in tuning_flags:
-
- flag_value = -1
- try:
- flag_value = int(flag)
- except:
- continue
-
- orig_cost += layer_costs[it]
-
- #print ("orig_cost = ", orig_cost, " flag_value = ", flag_value)
-
- if flag_value == 11:
- total_cost += layer_costs[it]
- elif flag_value == 10:
- total_cost += (layer_costs[it] / 1.3)
- elif flag_value == 8 or flag_value == 9:
- total_cost += (layer_costs[it] / 1.6)
- elif flag_value < 8:
- divisor = 5 + (7 - flag_value)
- total_cost += (layer_costs[it] / divisor)
-
- it += 1
-
- speedup = orig_cost * 1.0 / total_cost * 1.0
-
- return total_cost, speedup
-
-
-
-
-
-def getConfigCost2(layer_costs, knobs_speedup, config_flags):
-
- orig_cost = 0.0
- total_cost = 0.0
- for it in range(len(config_flags)):
- flag_value = config_flags[it]
- op_cost = layer_costs[it]
- speedup = knobs_speedup[flag_value]
-
- total_cost += (op_cost * 1.0 / speedup * 1.0)
- orig_cost += op_cost
-
- it += 1
-
- speedup = (orig_cost * 1.0) / (total_cost * 1.0)
-
- return total_cost, speedup
-
-
-
-
-
-def compute_promise_confidence2(binary_name, accuracy, confidence, layer_costs,
- result_dir, output_dir):
-
- confidence_list = []
-
- if not os.path.exists(result_dir):
- print("Path does not exist")
- sys.exit(0)
-
- file_names = os.listdir(result_dir)
- print file_names
-
- for file_name in file_names:
- # Skip sub-directories
- if os.path.isdir(result_dir + "/" + file_name):
- continue
-
- f = open(result_dir + "/" + file_name)
- tuner_file = open("promise_flags", "w+")
-
- config_str = f.read()
- tuner_file.write(config_str)
- tuner_file.close()
-
- #run_confidence, avg_accuracy = do_multiple_runs(binary_name, accuracy, confidence)
- run_confidence, avg_accuracy = do_multiple_runs2(binary_name, accuracy, confidence)
-
- if run_confidence >= 95:
- f2 = open(output_dir + "/" + file_name, "w+")
-
- config_cost, speedup = getConfigCost(layer_costs, config_str)
-
- f2.write("total_runs=" + str(total_runs) + "\t confidence=" + str(run_confidence) + "\t avg_accuracy=" + str(avg_accuracy) + "\t config_cost=" + str(config_cost) + "\t speedup=" + str(speedup) + "\n")
- f2.write(config_str)
- f2.close()
-
- flags_str = config_str.replace('\n', ',')
- conf_result = (run_confidence, avg_accuracy, file_name, flags_str)
- confidence_list.append(conf_result)
-
- return confidence_list
-
-
-
-
-
-def compute_promise_confidence3(binary_name, accuracy, confidence, layer_costs,
- result_dir, output_dir, knobs_speedup):
-
- confidence_list = []
-
- if not os.path.exists(result_dir):
- print("Path does not exist")
- sys.exit(0)
-
- file_names = os.listdir(result_dir)
- print file_names
-
- for file_name in file_names:
- # Skip sub-directories
- if os.path.isdir(result_dir + "/" + file_name):
- continue
-
- f = open(result_dir + "/" + file_name)
- tuner_file = open("promise_flags", "w+")
-
- config_flags = []
- config_str = ""
- it = 0
- for x in f:
-
- if it > 0:
- config_str += x
- config_flags.append(int(x.strip()))
- tuner_file.write(x)
- it += 1
-
- tuner_file.close()
-
-
- #run_confidence, avg_accuracy = do_multiple_runs(binary_name, accuracy, confidence)
- run_confidence, avg_accuracy = do_multiple_runs2(binary_name, accuracy, confidence)
-
- if run_confidence >= 95:
- f2 = open(output_dir + "/" + file_name, "w+")
-
- config_cost, speedup = getConfigCost2(layer_costs, knobs_speedup, config_flags)
-
- f2.write("total_runs=" + str(total_runs) + "\t confidence=" + str(run_confidence) + "\t avg_accuracy=" + str(avg_accuracy) + "\t config_cost=" + str(config_cost) + "\t speedup=" + str(speedup) + "\n")
- f2.write(config_str)
- f2.close()
-
- flags_str = config_str.replace('\n', ',')
- conf_result = (run_confidence, avg_accuracy, file_name, flags_str)
- confidence_list.append(conf_result)
-
- return confidence_list
-
-
-
-
-def dump_promise_confidence_files(binary, result_dir, layer_file_path,
- num_flags, accuracy, confidence):
-
- #result_dir = args.result_dir
- output_dir = result_dir + "/high_confidence"
- input_dir = result_dir + "/full_results"
-
- if not os.path.exists(output_dir):
- os.mkdir(output_dir)
-
- layer_sizes = processLayerFile(layer_file_path);
- print layer_sizes
- sleep(2)
-
- confidence_list = compute_promise_confidence(binary, accuracy, confidence, input_dir, output_dir)
- print confidence_list
-
- # Ascending sort on accuracy
- sorted_list = sorted(confidence_list, key = lambda tup: tup[1])
-
- promise_file = open(output_dir + "/promise_confs.txt", "w+")
- confidence_file = open(output_dir + "/confidence_summary.txt", "w+")
-
- max_configs = 50
- it_count = 0
- for x in sorted_list:
- if x[1] > accuracy and x[0] > confidence:
- config_str = getLayerConfigStr(x[3], layer_sizes, num_flags)
- promise_file.write(config_str + "\n")
- it_count += 1
- if it_count > max_configs:
- break
-
- confidence_file.write(str(x[0]) + "\t" + str(x[1]) + "\t" + str(x[3]) + "\n")
-
- promise_file.close()
- confidence_file.close()
-
- print "Dumped Confidence Summary"
-
-
-
-
-
-def dump_promise_confidence_files2(binary, result_dir, layer_file_path,
- num_flags, accuracy, layer_costs, confidence):
-
-
- #result_dir = args.result_dir
- output_dir = result_dir + "/high_confidence"
- input_dir = result_dir + "/full_results"
-
- if not os.path.exists(output_dir):
- os.mkdir(output_dir)
-
- layer_sizes = processLayerFile(layer_file_path);
- print layer_sizes
- sleep(2)
-
- confidence_list = compute_promise_confidence2(binary, accuracy, confidence, layer_costs, input_dir, output_dir)
- print confidence_list
-
- # Ascending sort on accuracy
- sorted_list = sorted(confidence_list, key = lambda tup: tup[1])
-
- promise_file = open(output_dir + "/promise_confs.txt", "w+")
- confidence_file = open(output_dir + "/confidence_summary.txt", "w+")
-
- max_configs = 50
- it_count = 0
- for x in sorted_list:
- if x[1] > accuracy and x[0] > confidence:
- config_str = getLayerConfigStr(x[3], layer_sizes, num_flags)
- promise_file.write(config_str + "\n")
- it_count += 1
- if it_count > max_configs:
- break
-
- confidence_file.write(str(x[0]) + "\t" + str(x[1]) + "\t" + str(x[3]) + "\n")
-
- promise_file.close()
- confidence_file.close()
-
- print "Dumped Confidence Summary"
-
-
-
-
-def dump_promise_confidence_files3(binary, input_dir, output_dir, layer_file_path,
- num_flags, accuracy, layer_costs, confidence):
-
-
- #result_dir = args.result_dir
- #output_dir = result_dir + "/high_confidence"
- #input_dir = result_dir + "/full_results"
-
- if not os.path.exists(output_dir):
- os.mkdir(output_dir)
-
- layer_sizes = processLayerFile(layer_file_path);
- print layer_sizes
- sleep(2)
-
- confidence_list = compute_promise_confidence2(binary, accuracy, confidence, layer_costs, input_dir, output_dir)
- print confidence_list
-
- # Ascending sort on accuracy
- sorted_list = sorted(confidence_list, key = lambda tup: tup[1])
-
- promise_file = open(output_dir + "/promise_confs.txt", "w+")
- confidence_file = open(output_dir + "/confidence_summary.txt", "w+")
-
- max_configs = 50
- it_count = 0
- for x in sorted_list:
- if x[1] > accuracy and x[0] > confidence:
- config_str = getLayerConfigStr(x[3], layer_sizes, num_flags)
- promise_file.write(config_str + "\n")
- it_count += 1
- if it_count > max_configs:
- break
-
- confidence_file.write(str(x[0]) + "\t" + str(x[1]) + "\t" + str(x[3]) + "\n")
-
- promise_file.close()
- confidence_file.close()
-
- print "Dumped Confidence Summary"
-
-
-
-
-
-def dump_promise_confidence_files4(binary, input_dir, output_dir, layer_file_path,
- num_flags, accuracy, layer_costs, confidence, knobs_speedup):
-
-
- if not os.path.exists(output_dir):
- os.mkdir(output_dir)
-
- layer_sizes = processLayerFile(layer_file_path);
- print layer_sizes
- sleep(2)
-
- confidence_list = compute_promise_confidence3(binary, accuracy, confidence, layer_costs, input_dir, output_dir, knobs_speedup)
- print confidence_list
-
- # Ascending sort on accuracy
- sorted_list = sorted(confidence_list, key = lambda tup: tup[1])
-
- promise_file = open(output_dir + "/promise_confs.txt", "w+")
- confidence_file = open(output_dir + "/confidence_summary.txt", "w+")
-
- max_configs = 50
- it_count = 0
- for x in sorted_list:
- if x[1] > accuracy and x[0] > confidence:
- config_str = getLayerConfigStr(x[3], layer_sizes, num_flags)
- promise_file.write(config_str + "\n")
- it_count += 1
- if it_count > max_configs:
- break
-
- confidence_file.write(str(x[0]) + "\t" + str(x[1]) + "\t" + str(x[3]) + "\n")
-
- promise_file.close()
- confidence_file.close()
-
- print "Dumped Confidence Summary"
-
-
-
-
-
-
-if __name__ == "__main__":
-
- argparser = argparse.ArgumentParser(description='runs best configs to get high confidence on accuracy')
- argparser.add_argument('--result-dir', help='Directory containing OpenTuner configurations')
- argparser.add_argument('--output-dir', help='Directory for storing output directory')
- argparser.add_argument('--binary', help='Binary name to run')
- argparser.add_argument('--accuracy', type=float, help='Accuracy constraint')
- argparser.add_argument('--confidence', type=float, help='Confidence threshold')
-
-
- args = argparser.parse_args()
- result_dir = args.result_dir
- output_dir = args.output_dir
- binary = args.binary
- accuracy = args.accuracy
- confidence = args.confidence
-
- confidence_list = compute_confidence(binary, accuracy, confidence, result_dir, output_dir)
- #print confidence_list
-
- sorted_list = sorted(confidence_list, key = lambda tup: tup[0], reverse=True)
-
- output_file = open(output_dir + "/confidence_summary.txt", "w+")
- for x in sorted_list:
- output_file.write(str(x[0]) + "\t" + str(x[1]) + "\t" + str(x[2]) + "\n")
-
- output_file.close()
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/pareto_curve.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/pareto_curve.py
deleted file mode 100644
index db8233994b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/pareto_curve.py
+++ /dev/null
@@ -1,264 +0,0 @@
-
-
-import os
-import shutil
-from measure_confidence2 import getConfigCost
-
-
-AL_THRESHOLD = 0.1
-
-
-class Config:
- def __init__(self):
- self.avg_accuracy = 0
- self.avg_loss = 0
- self.speedup = 1
- self.fname = ""
- self.flags = []
-
-
-
-
-def skipFile(fname):
-
- skip_files = {}
- skip_files["confidence_summary.txt"] = 1
- skip_files["promise_confs.txt"] = 1
-
- if "accuracy" in fname:
- return True
-
- if fname in skip_files:
- return True
- else:
- return False
-
-
-
-
-def loadConfigData(result_dir, layer_costs, baseline_accuracy):
-
- config_arr = []
-
- #result_dir += "/promise_tuner/high_confidence/"
- file_names = os.listdir(result_dir)
-
-
- for fname in file_names:
- if not skipFile(fname):
-
- fpath = result_dir + fname
- config = Config()
- f = open(fpath, "r")
-
- config_str = f.read()
- cost, speedup = getConfigCost(layer_costs, config_str)
-
- config.speedup = speedup
- config.fname = fname
-
- fpath2 = fpath + "_accuracy"
- f2 = open(fpath2, "r")
- acc_str = f2.read().strip()
- accuracy = float(acc_str)
-
- config.avg_accuracy = accuracy
- config.avg_loss = baseline_accuracy - accuracy
-
- config_arr.append(config)
-
-
- return config_arr
-
-
-
-
-class Configuration:
- def __init__(self, name, speedup, energy, accuracy, accuracy_loss):
- self.name = name
- self.speedup = speedup
- self.energy = energy
- self.accuracy = accuracy
- self.accuracy_loss = accuracy_loss
- def __repr__(self):
- return repr((self.name, self.speedup, self.energy, self.accuracy, self.accuracy_loss))
-
-configuration_objects = [
- Configuration('conf1', 1.05, 15, 85, 1.2),
- Configuration('conf2', 2.51, 12, 83, 1.4),
- Configuration('conf3', 2.05, 10, 84, 0.8),
-]
-
-def compute_pareto_points(configurations):
- speedupconfigurations = []
- energyconfigurations = []
- #sort configurations based on speedup
- sorted_configurations = sorted(configurations, key=lambda conf: conf.accuracy_loss)
-
- start_idx = 0
- while start_idx < len(sorted_configurations):
- end_idx = start_idx + 1;
- # find end_idx
- while end_idx < len(sorted_configurations) and (sorted_configurations[end_idx].accuracy_loss - sorted_configurations[start_idx].accuracy_loss < AL_THRESHOLD) :
- end_idx += 1
- # find best speedup end energy in this accuracy loss level
- sp = -1.0
- sp_idx = 0
- en = -1.0
- en_idx = 0
- for i in range(start_idx, end_idx):
- if sorted_configurations[i].speedup > sp:
- sp = sorted_configurations[i].speedup
- sp_idx = i
- if sorted_configurations[i].energy > en:
- en = sorted_configurations[i].energy
- en_idx = i
- sp_not_dominated = True
- # if not empty list of configurations
- if speedupconfigurations:
- if speedupconfigurations[-1].speedup >= sp:
- sp_not_dominated = False
- en_not_dominated = True
- # if not empty list of configurations
- if energyconfigurations:
- if energyconfigurations[-1].energy >= en:
- en_not_dominated = False
- if sp_not_dominated:
- speedupconfigurations.append(sorted_configurations[sp_idx])
- if en_not_dominated:
- energyconfigurations.append(sorted_configurations[en_idx])
- # outer while loop variable increment
- start_idx = end_idx
- return [speedupconfigurations, energyconfigurations]
-
-
-def compute_pareto_points_with_margin(configurations, speedup_band_width, energy_band_width):
- speedupconfigurations = []
- energyconfigurations = []
- #sort configurations based on speedup
- sorted_configurations = sorted(configurations, key=lambda conf: conf.accuracy_loss)
-
- idx_to_sp_conf_dict = {}
- idx_to_en_conf_dict = {}
-
- start_idx = 0
- while start_idx < len(sorted_configurations):
- end_idx = start_idx + 1;
- # find end_idx
- while end_idx < len(sorted_configurations) and (sorted_configurations[end_idx].accuracy_loss - sorted_configurations[start_idx].accuracy_loss < AL_THRESHOLD) :
- end_idx += 1
- # find best speedup end energy in this accuracy loss level
- sp = -1.0
- sp_idx = 0
- en = -1.0
- en_idx = 0
- for i in range(start_idx, end_idx):
- if sorted_configurations[i].speedup > sp:
- sp = sorted_configurations[i].speedup
- sp_idx = i
- if sorted_configurations[i].energy < en:
- en = sorted_configurations[i].energy
- en_idx = i
- sp_not_dominated = True
- # if not empty list of configurations
- if speedupconfigurations:
- if speedupconfigurations[-1].speedup >= sp:
- sp_not_dominated = False
- en_not_dominated = True
- # if not empty list of configurations
- if energyconfigurations:
- if energyconfigurations[-1].energy >= en:
- en_not_dominated = False
- if sp_not_dominated:
- speedupconfigurations.append(sorted_configurations[sp_idx])
- idx_to_sp_conf_dict[start_idx] = len(speedupconfigurations)-1
- if en_not_dominated:
- energyconfigurations.append(sorted_configurations[en_idx])
- idx_to_en_conf_dict[start_idx] = len(energyconfigurations)-1
- # outer while loop variable increment
- start_idx = end_idx
-
- # We want to add configurations in a band of a certain width around the curves
- # not possible to do during contruction, because the quality of the curve would
- # deteriorate quickly
-
- AdjustedSpeedupCurve = []
- AdjustedEnergyCurve = []
-
- start_idx = 0
- while start_idx < len(sorted_configurations):
- end_idx = start_idx + 1;
- # find end_idx
- while end_idx < len(sorted_configurations) and (sorted_configurations[end_idx].accuracy_loss - sorted_configurations[start_idx].accuracy_loss < AL_THRESHOLD) :
- end_idx += 1
- for i in range(start_idx, end_idx):
- if sorted_configurations[i].speedup + speedup_band_width >= speedupconfigurations[idx_to_sp_conf_dict[start_idx]].speedup:
- AdjustedSpeedupCurve.append(sorted_configurations[i])
- if sorted_configurations[i].energy + energy_band_width >= energyconfigurations[idx_to_en_conf_dict[start_idx]].energy:
- AdjustedEnergyCurve.append(sorted_configurations[i])
- # outer while loop variable increment
- start_idx = end_idx
-
- return [AdjustedSpeedupCurve, AdjustedEnergyCurve]
-
-
-
-def findParetoConfigs(base_dir, layer_costs, accuracy):
-
- result_dir = base_dir + "/pareto/"
- try:
- os.mkdir(result_dir)
- except:
- print "could not create dir"
-
- input_dir = base_dir + "/full_results/"
- #result_dir = "../build_tuner/tuner_results/alexnet_cifar10/loss_3/batch15"
- config_arr = loadConfigData(input_dir, layer_costs, accuracy)
-
- config_list = []
-
- it = 0
- for config in config_arr:
- config = Configuration(config.fname , config.speedup, 100, config.avg_accuracy, config.avg_loss)
- config_list.append(config)
-
-
- SPEEDUP_BAND_SIZE = 1.0
- ENERGY_BAND_SIZE = 10
-
- # No Pareto Selection if list is < 50 configurations
- if len(config_list) < 50:
- SPEEDUP_BAND_SIZE = 100 # Include all in Pareto Frontier
-
-
- print ("*SPEEDUP_BAND_SIZE = ", SPEEDUP_BAND_SIZE)
-
- ASC, AEC = compute_pareto_points_with_margin(config_list, SPEEDUP_BAND_SIZE, ENERGY_BAND_SIZE)
-
-
- print ("len(config_list) = ", len(config_list))
- print ("len(ASC) = ", len(ASC))
-
- #print (ASC)
- #print (config_list)
-
- for conf in ASC:
- #dst_path = conf.name.replace("full_results", "pareto")
- src_path = base_dir + "/full_results/" + conf.name
- dst_path = base_dir + "/pareto/" + conf.name
- shutil.copy(src_path, dst_path)
-
-
-
-if __name__ == "__main__":
-
- get_pareto_configs("")
-
- #SC, EC = compute_pareto_points(configuration_objects)
- #ASC, AEC = compute_pareto_points_with_margin(configuration_objects, SPEEDUP_BAND_SIZE, ENERGY_BAND_SIZE)
-
- #print(SC)
- #print(EC)
-
- #print(ASC)
- #print(AEC)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/promise_tuner2.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/promise_tuner2.py
deleted file mode 100644
index ca96ff16c2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/promise_tuner2.py
+++ /dev/null
@@ -1,220 +0,0 @@
-#!/usr/bin/env python
-#
-# Optimize blocksize of apps/mmm_block.cpp
-#
-# This is an extremely simplified version meant only for tutorials
-#
-import adddeps # fix sys.path
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-import subprocess
-import threading
-import psutil
-
-from measure_confidence2 import dump_promise_confidence_files
-from select_top_results import select_top_results
-from time import sleep
-
-
-layer_file = ""
-output_dir = ""
-flag_ranges = []
-tuning_flags = []
-binary_name = ""
-accuracy_threshold = 10.0
-evaluated_configs = {}
-orig_result_dir = ""
-gpu_layers = 0
-
-
-def getAccuracy(file_name):
-
- file = open(file_name, "r")
- acc_str = file.read()
- file.close()
- accuracy = float(acc_str)
-
- try:
- accuracy = float(acc_str)
- except:
- return 20
-
- print accuracy
- return accuracy
-
-
-
-def createFlagsFile(file_name, cfg):
-
- f = open(file_name, "w+")
- cmd_config = ""
- for flag in tuning_flags:
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-class ClangFlagsTuner(MeasurementInterface):
-
- def __init__(self, args):
- objective = ThresholdAccuracyMinimizeTime(accuracy_threshold)
- input_manager = FixedInputManager(size=num_flags)
- self.configs_list = []
-
- super(ClangFlagsTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
-
- # NOTE: Skipping first 'gpu_layers' to run on GPU
- for flag in tuning_flags[:gpu_layers]:
- manipulator.add_parameter(
- EnumParameter(flag, [8, 9]))
-
- for flag in tuning_flags[gpu_layers:]:
- manipulator.add_parameter(
- EnumParameter(flag, flag_ranges
- )) #default is needed, optimizations don't work without it(tried and tested)
- return manipulator
-
-
- def run(self, desired_result, input, limit):
-
- """
- Run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
-
- # NOTE: creates the file with flags read by the runtime
- createFlagsFile("promise_flags", cfg)
-
- run_cmd = binary_name
- print "binary_name = ", run_cmd
- #run_result_call_program = self.call_program(run_cmd)
- #print "returned \n\n"
-
- FNULL = open(os.devnull, 'wb')
- p = subprocess.Popen(run_cmd, stdout = FNULL)
- p.wait()
-
-
- accuracy = getAccuracy("final_accuracy")
- total_comps = abs(accuracy_threshold - accuracy)
-
-
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- Result.accuracy = accuracy
-
- if accuracy > accuracy_threshold:
- #if accuracy not in evaluated_configs:
- config_tuple = (total_comps, accuracy, cfg)
- self.configs_list.append(config_tuple)
- evaluated_configs[accuracy] = 1
- shutil.copy('promise_flags', output_dir + '/' + binary_name + '_' + str(accuracy))
-
-
- print "done with one run"
-
- return Result
-
-
- def save_final_config(self, configuration):
-
- print "Dumping High Confidence results \n"
- sleep(20)
-
- # Only dumping files with 95% confidence
- dump_promise_confidence_files(binary_name, orig_result_dir, layer_file, num_flags, accuracy_threshold, 95)
- #select_top_results(orig_result_dir + "/high_confidence")
-
-
-
- """
- called at the end of autotuning with the best resultsdb.models.Configuration
- """
- print "Final configuration", configuration.data
-
- return
-
-
-
-
-
-if __name__ == '__main__':
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='path to target binary')
- argparser.add_argument('--num-flags', type=int, help='num of flags to tune')
- argparser.add_argument('--start-range', type=int, help='start range in tuning')
- argparser.add_argument('--error-range', type=int, help='range of error values used in tuning')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='result directory')
- argparser.add_argument('--layer-file', help='layer description')
- argparser.add_argument('--gpu-layers', type=int, help='first N layers to run on GPU')
-
-
- args = argparser.parse_args()
- binary_name = str(args.binary)
- print("binary_name = ", binary_name)
- num_flags = int(args.num_flags)
- start_range = int(args.start_range)
- error_range = int(args.error_range)
- accuracy_threshold = float(args.accuracy)
- print("accuracy = ", accuracy_threshold)
- result_dir = args.result_dir
- orig_result_dir = result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
- gpu_layers = args.gpu_layers
-
-
- output_dir = result_dir + "/full_results"
- print output_dir
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- if not os.path.exists(output_dir):
- print("Creating output directory = ", output_dir)
- os.mkdir(output_dir)
-
- #for j in range(error_range):
- # flag_ranges.append(j)
-
- for j in range(start_range, error_range):
- flag_ranges.append(j)
-
-
- print("flag_ranges = ", flag_ranges)
-
- # File with layer description
- layer_file = args.layer_file
-
- for i in range(num_flags):
- tuning_flags.append("flag" + str(i))
-
- ClangFlagsTuner.main(argparser.parse_args())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/promise_tuner3.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/promise_tuner3.py
deleted file mode 100644
index 04ce0d6158..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/promise_tuner3.py
+++ /dev/null
@@ -1,314 +0,0 @@
-#!/usr/bin/env python
-#
-# Optimize blocksize of apps/mmm_block.cpp
-#
-# This is an extremely simplified version meant only for tutorials
-#
-import adddeps # fix sys.path
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-import subprocess
-import threading
-import psutil
-
-from measure_confidence2 import dump_promise_confidence_files3
-from measure_confidence2 import getConfidence, getMinAccuracy
-from select_top_results import select_top_results
-from time import sleep
-from pareto_curve import findParetoConfigs
-
-
-layer_file = ""
-output_dir = ""
-flag_ranges = []
-tuning_flags = []
-binary_name = ""
-accuracy_threshold = 10.0
-evaluated_configs = {}
-orig_result_dir = ""
-gpu_layers = 0
-
-test_id = 0
-
-layer_costs = []
-
-
-def readCostFile(file_path):
-
- f = open(file_path)
- for x in f:
- cost = float(x.strip())
- layer_costs.append(cost)
-
- print ("len(layer_costs) = ", layer_costs)
- f.close()
-
-
-
-def getAccuracy(file_name):
-
- file = open(file_name, "r")
- acc_str = file.read()
- file.close()
- accuracy = float(acc_str)
-
- try:
- accuracy = float(acc_str)
- except:
- return 20
-
- print accuracy
- return accuracy
-
-
-
-def createFlagsFile(file_name, cfg):
-
- f = open(file_name, "w+")
- cmd_config = ""
- for flag in tuning_flags:
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-
-def getConfigCost(cfg):
-
- it = 0
- total_cost = 0.0
- for flag in tuning_flags:
- flag_value = cfg[flag]
- if flag_value == 11:
- total_cost += layer_costs[it]
- elif flag_value == 10:
- total_cost += (layer_costs[it] / 1.3)
- elif flag_value == 8 or flag_value == 9:
- total_cost += (layer_costs[it] / 1.6)
- elif flag_value < 8:
- divisor = 5 + (7 - flag_value)
- total_cost += (layer_costs[it] / divisor)
-
- it += 1
-
- return total_cost
-
-
-
-
-class ClangFlagsTuner(MeasurementInterface):
-
- def __init__(self, args):
- objective = ThresholdAccuracyMinimizeTime(accuracy_threshold)
- input_manager = FixedInputManager(size=num_flags)
- self.configs_list = []
-
- super(ClangFlagsTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
-
- flags_arr = []
- for i in range (8, error_range):
- flags_arr.append(i)
-
- # NOTE: Skipping first 'gpu_layers' to run on GPU
- for flag in tuning_flags[:gpu_layers]:
- manipulator.add_parameter(
- EnumParameter(flag, flags_arr))
-
- ind = gpu_layers
- for flag in tuning_flags[gpu_layers:]:
- if ind in skip_layers:
- manipulator.add_parameter(
- EnumParameter(flag, flags_arr))
- print ("8 ..... 11")
- else:
- manipulator.add_parameter(
- EnumParameter(flag, flag_ranges
- )) #default is needed, optimizations don't work without it(tried and tested)
- print ("1 .... 11")
- ind += 1
-
-
- return manipulator
-
-
- def run(self, desired_result, input, limit):
-
- """
- Run a given configuration then
- return performance
- """
- global test_id
-
- cfg = desired_result.configuration.data
-
- # NOTE: creates the file with flags read by the runtime
- createFlagsFile("promise_flags", cfg)
-
- run_cmd = binary_name
- print "\nbinary_name = ", run_cmd
- #run_result_call_program = self.call_program(run_cmd)
-
-
- total_runs = 2
- FNULL = open(os.devnull, 'wb')
- #p = subprocess.Popen(run_cmd, stdout = FNULL)
- p = subprocess.Popen([run_cmd, str(total_runs)], stdout = FNULL)
- p.wait()
-
-
- accuracy = getAccuracy("final_accuracy")
-
- # Get Confidence for multiple runs
- conf, avg_acc = getConfidence("run_accuracies.txt", accuracy_threshold)
-
- # getConfigCost returns the cost associated with the selected configuration
- total_comps = getConfigCost(cfg)
-
-
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- #Result.accuracy = accuracy
- min_accuracy = getMinAccuracy("run_accuracies.txt")
- print ("min_accuracy = ", min_accuracy)
- Result.accuracy = min_accuracy
-
- # Only pass conf if conf == 100
- if min_accuracy > accuracy_threshold and conf == 100:
- print ("conf = ", conf, " avg_acc = ", avg_acc)
- #if accuracy not in evaluated_configs:
- config_tuple = (total_comps, accuracy, cfg)
- self.configs_list.append(config_tuple)
- evaluated_configs[accuracy] = 1
- shutil.copy('promise_flags', output_dir + '/' + binary_name + '_' + str(test_id))
-
- f_acc = open(output_dir + '/' + binary_name + '_' + str(test_id) + "_accuracy", "w")
- f_acc.write(str(accuracy))
- f_acc.close()
-
-
- test_id += 1
-
- return Result
-
-
- def save_final_config(self, configuration):
-
- print "Dumping High Confidence results \n"
- sleep(2)
-
-
- findParetoConfigs(orig_result_dir, layer_costs, accuracy_threshold)
-
- input_dir = orig_result_dir + "/pareto/"
- output_dir = orig_result_dir + "/high_confidence/"
-
- # Only dumping files with 95% confidence
- dump_promise_confidence_files3(binary_name, input_dir, output_dir, layer_file, num_flags, accuracy_threshold, layer_costs, 95)
- #select_top_results(orig_result_dir + "/high_confidence")
-
-
- """
- called at the end of autotuning with the best resultsdb.models.Configuration
- """
- print "Final configuration", configuration.data
-
- return
-
-
-
-error_range = 11
-
-
-if __name__ == '__main__':
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='path to target binary')
- argparser.add_argument('--num-flags', type=int, help='num of flags to tune')
- argparser.add_argument('--start-range', type=int, help='start range in tuning')
- argparser.add_argument('--error-range', type=int, help='range of error values used in tuning')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='result directory')
- argparser.add_argument('--layer-file', help='layer description')
- argparser.add_argument('--cost-file', help='layer description')
- argparser.add_argument('--gpu-layers', type=int, help='first N layers to run on GPU')
- argparser.add_argument('--skip-layers', help='layer IDs to run on GPU')
-
-
- args = argparser.parse_args()
- binary_name = str(args.binary)
- print("binary_name = ", binary_name)
- num_flags = int(args.num_flags)
- start_range = int(args.start_range)
- error_range = int(args.error_range)
- accuracy_threshold = float(args.accuracy)
- print("accuracy = ", accuracy_threshold)
- result_dir = args.result_dir
- orig_result_dir = result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
- gpu_layers = args.gpu_layers
- skip_layers_str = args.skip_layers
-
- skip_layers = []
- layer_ids = skip_layers_str.split("_")
- for layer_id in layer_ids:
- skip_layers.append(int(layer_id))
-
- print ("skip_layers = ", skip_layers)
-
- # NOTE: Reading the cost file (with No of ops) to better guide the Autotuner
- cost_file_path = args.cost_file
- readCostFile(cost_file_path)
-
-
- output_dir = result_dir + "/full_results"
- print output_dir
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- if not os.path.exists(output_dir):
- print("Creating output directory = ", output_dir)
- os.mkdir(output_dir)
-
- #for j in range(error_range):
- # flag_ranges.append(j)
-
- for j in range(start_range, error_range):
- flag_ranges.append(j)
-
-
- print("flag_ranges = ", flag_ranges)
-
- # File with layer description
- layer_file = args.layer_file
-
- for i in range(num_flags):
- tuning_flags.append("flag" + str(i))
-
- ClangFlagsTuner.main(argparser.parse_args())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/promise_tuner_piped.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/promise_tuner_piped.py
deleted file mode 100644
index cf84c503b0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/promise_tuner_piped.py
+++ /dev/null
@@ -1,231 +0,0 @@
-#!/usr/bin/env python
-#
-# Optimize blocksize of apps/mmm_block.cpp
-#
-# This is an extremely simplified version meant only for tutorials
-#
-import adddeps # fix sys.path
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-import subprocess
-import threading
-import psutil
-
-from measure_confidence import dump_promise_confidence_files
-from select_top_results import select_top_results
-from time import sleep
-
-
-layer_file = ""
-output_dir = ""
-flag_ranges = []
-tuning_flags = []
-binary_name = ""
-accuracy_threshold = 10.0
-evaluated_configs = {}
-orig_result_dir = ""
-
-
-def getAccuracy(file_name):
-
- file = open(file_name, "r")
- acc_str = file.read()
- file.close()
- accuracy = float(acc_str)
-
- try:
- accuracy = float(acc_str)
- except:
- return 20
-
- print accuracy
- return accuracy
-
-
-
-def kill(proc_pid):
- process = psutil.Process(proc_pid)
- for proc in process.children(recursive=True):
- proc.kill()
- process.kill()
-
-
-
-def createFlagsFile(file_name, cfg):
-
- f = open(file_name, "w+")
- cmd_config = ""
- for flag in tuning_flags:
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-class ClangFlagsTuner(MeasurementInterface):
-
- def __init__(self, args):
- objective = ThresholdAccuracyMinimizeTime(accuracy_threshold)
- input_manager = FixedInputManager(size=num_flags)
- self.configs_list = []
-
- super(ClangFlagsTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
- FNULL = open(os.devnull, 'wb')
- #run_result_call_program = self.call_program(run_cmd)
- self.start_process = subprocess.Popen([binary_name, "opentuner_run"] , stdout=FNULL);
-
- try:
- os.mkfifo("/tmp/myfifo")
- except OSError, e:
- print("FIFO exists")
-
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- for flag in tuning_flags:
- manipulator.add_parameter(
- EnumParameter(flag, flag_ranges
- # [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
- )) #default is needed, optimizations don't work without it(tried and tested)
- return manipulator
-
-
- def run(self, desired_result, input, limit):
-
- """
- Run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
-
- # NOTE: creates the file with flags read by the runtime
- createFlagsFile("opentuner_flags", cfg)
-
- run_cmd = binary_name
- print run_cmd
- #run_result_call_program = self.call_program(run_cmd)
-
- # Using Named Pipes to signal execution to the DNN outer thread
- fifo = open("/tmp/myfifo", "w")
- fifo.write("start_run")
- fifo.close()
-
- print "Waiting for process to signal back - when done processing one run"
-
- fifo2 = open("/tmp/myfifo", "r")
- fifo2.read()
- fifo2.close()
-
- print "Process Signalled back"
-
- accuracy = getAccuracy("final_accuracy")
- total_comps = abs(accuracy_threshold - accuracy)
-
-
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- Result.accuracy = accuracy
-
- if accuracy > accuracy_threshold:
- #if accuracy not in evaluated_configs:
- config_tuple = (total_comps, accuracy, cfg)
- self.configs_list.append(config_tuple)
- evaluated_configs[accuracy] = 1
- shutil.copy('opentuner_flags', output_dir + '/' + binary_name + '_' + str(accuracy))
-
-
- print "done with one run"
-
- return Result
-
-
- def save_final_config(self, configuration):
-
- print "Dumping High Confidence results \n"
- sleep(20)
-
- # Only dumping files with 95% confidence
- dump_promise_confidence_files(binary_name, orig_result_dir, layer_file, num_flags, accuracy_threshold, 95)
- #select_top_results(orig_result_dir + "/high_confidence")
-
-
- #self.start_process.kill()
- kill(self.start_process.pid)
-
- """
- called at the end of autotuning with the best resultsdb.models.Configuration
- """
- print "Final configuration", configuration.data
-
- return
-
-
-
-
-
-if __name__ == '__main__':
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='path to target binary')
- argparser.add_argument('--num-flags', type=int, help='num of flags to tune')
- argparser.add_argument('--error-range', type=int, help='range of error values used in tuning')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='result directory')
- argparser.add_argument('--layer-file', help='layer description')
-
-
- args = argparser.parse_args()
- binary_name = str(args.binary)
- print("binary_name = ", binary_name)
- num_flags = int(args.num_flags)
- error_range = int(args.error_range)
- accuracy_threshold = float(args.accuracy)
- print("accuracy = ", accuracy_threshold)
- result_dir = args.result_dir
- orig_result_dir = result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
-
- output_dir = result_dir + "/full_results"
- print output_dir
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- if not os.path.exists(output_dir):
- print("Creating output directory = ", output_dir)
- os.mkdir(output_dir)
-
- for j in range(error_range):
- flag_ranges.append(j)
-
- print("flag_ranges = ", flag_ranges)
-
- # File with layer description
- layer_file = args.layer_file
-
- for i in range(num_flags):
- tuning_flags.append("flag" + str(i))
-
- ClangFlagsTuner.main(argparser.parse_args())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/psnr_tuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/psnr_tuner.py
deleted file mode 100644
index eb126de3aa..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/psnr_tuner.py
+++ /dev/null
@@ -1,318 +0,0 @@
-#!/usr/bin/env python
-#
-# Algorithmic Approximation Tuning
-# Purpose: Tunes for Perforation, Sampling, Numerical Precision (FP16)
-
-
-import adddeps
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-import subprocess
-import threading
-import psutil
-
-from measure_confidence2 import dump_promise_confidence_files3
-from measure_confidence2 import getConfidence, getMinAccuracy
-from select_top_results import select_top_results
-from time import sleep
-from pareto_curve import findParetoConfigs
-
-
-
-
-class TunerData:
- def __init__(self):
- self.binary_path = ""
- self.output_dir = ""
- self.num_layers = 0
- self.knobs_list = []
- self.knobs_speedup = {}
- self.accuracy_threshold = 0
- self.test_id = 0
- self.layer_costs = []
- self.tuning_flags = []
- self.autotuner_runs = 0
-
-
-
-
-tunerData = TunerData()
-
-
-
-
-def readCostFile(file_path):
-
- layer_costs = []
- f = open(file_path)
- for x in f:
- cost = float(x.strip())
- layer_costs.append(cost)
-
- print ("len(layer_costs) = ", layer_costs)
- f.close()
-
- return layer_costs
-
-
-
-def getPSNR(file_name):
- with open(file_name) as f:
- try:
- raw_str = f.read()
- violation, avg_psnr = [float(s) for s in raw_str.split(",")]
- except:
- return None, None
- print (100 - violation, avg_psnr)
- return 100 - violation, avg_psnr
-
-
-
-def createFlagsFile(file_name, cfg):
-
- f = open(file_name, "w+")
- cmd_config = ""
- for i in range(tunerData.num_layers): # flag in tunerData.tuning_flags:
- flag = tunerData.tuning_flags[i]
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-
-def readLayerKnobs(file_path):
-
- f = open(file_path, "r")
- knobs_list = []
- for x in f:
- knobs = []
- vals = x.split(",")
- for val in vals:
- knobs.append(int(val))
-
- knobs_list.append(knobs)
-
- print ("knobs_list = ", knobs_list)
-
- return knobs_list
-
-
-
-def readKnobConfig(file_path):
-
- knobs_speedup = {}
- f = open(file_path, "r")
- for x in f:
- toks = x.split("\t")
- ID = int(toks[0].split(",")[1])
-
- speedup = float(toks[2])
- knobs_speedup[ID] = speedup
-
- print ("knobs_speedup = ", knobs_speedup)
-
- return knobs_speedup
-
-
-
-
-def getConfigCost(cfg):
-
- orig_cost = 0.0
- total_cost = 0.0
- for it in range(tunerData.num_layers):
- flag = tunerData.tuning_flags[it]
- flag_value = cfg[flag]
- op_cost = tunerData.layer_costs[it]
- speedup = tunerData.knobs_speedup[flag_value]
-
- total_cost += (op_cost * 1.0 / speedup * 1.0)
- orig_cost += op_cost
-
- it += 1
-
- speedup = (orig_cost * 1.0) / (total_cost * 1.0)
-
- return total_cost, speedup
-
-
-
-def appendTopLine(f_path, accuracy, total_runs, total_comps, speedup):
-
- f_str = open(f_path, "r").read()
-
- f_out = open(f_path, "w+")
- f_out.write("avg_accuracy=" + str(accuracy) + "\tconfig_cost=" + str(total_comps) + "\tspeedup=" + str(speedup) + "\n" )
- f_out.write(f_str)
-
- f_out.close()
-
-
-
-def dumpAccuracyFile(accuracy):
-
- f_acc = open(tunerData.output_dir + '/' + tunerData.binary_path + '_' + str(tunerData.test_id) + "_accuracy", "w")
- f_acc.write(str(accuracy))
- f_acc.close()
-
-
-
-class ClangFlagsTuner(MeasurementInterface):
-
- def __init__(self, args):
- objective = ThresholdAccuracyMinimizeTime(tunerData.accuracy_threshold)
- input_manager = FixedInputManager(size=tunerData.num_layers)
- self.configs_list = []
-
- super(ClangFlagsTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
-
- for i in range(tunerData.num_layers):
- tunerData.tuning_flags.append("flag" + str(i))
-
-
- #for flag in tunerData.tuning_flags:
- for ind in range(tunerData.num_layers):
- flag = tunerData.tuning_flags[ind]
-
- manipulator.add_parameter(EnumParameter(flag, tunerData.knobs_list[ind]))
- print ("ind = ", ind, " len = ", len(tunerData.knobs_list))
- print (tunerData.knobs_list[ind])
-
- ind += 1
-
- return manipulator
-
-
-
- def run(self, desired_result, input, limit):
-
- """
- Run a given configuration then
- return performance
- """
- global test_id
-
- cfg = desired_result.configuration.data
-
- # NOTE: creates the file with flags read by the runtime
- createFlagsFile("promise_flags", cfg)
-
- run_cmd = tunerData.binary_path
- print "\nbinary_path = ", run_cmd
-
- input_size = 5000
- offset = 5000
-
- total_runs = 1 # NOTE: Single run sufficient in Algorithmic Approx Tuner
- FNULL = open(os.devnull, 'wb')
- p = subprocess.Popen([run_cmd], stdout = FNULL)
- p.wait()
- if p.returncode != 0:
- # Something went wrong
- sys.stderr.write("Child program returned non-zero; you may want to stop and check.")
-
- success_rate, avg_psnr = getPSNR("final_accuracy")
-
- # getConfigCost returns the cost associated with the selected configuration
- total_comps, speedup = getConfigCost(cfg)
-
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- Result.accuracy = success_rate
-
- if success_rate > tunerData.accuracy_threshold:
- config_tuple = (total_comps, success_rate, cfg)
- self.configs_list.append(config_tuple)
- f_path = tunerData.output_dir + '/' + tunerData.binary_path + '_' + str(tunerData.test_id)
- shutil.copy('promise_flags', f_path)
-
- appendTopLine(f_path, avg_psnr, total_runs, total_comps, speedup)
-
- # dumpAccuracyFile(accuracy)
-
-
- tunerData.test_id += 1
-
- return Result
-
-
- def save_final_config(self, configuration):
-
- print "Done with Autotuning Run \n"
- sleep(2)
-
- print "Final configuration", configuration.data
-
- return
-
-
-
-
-if __name__ == '__main__':
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='path to target binary')
- argparser.add_argument('--num-layers', type=int, help='num of flags to tune')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='result directory')
- argparser.add_argument('--cost-file', help='layer description')
- argparser.add_argument('--knobs-config', help='knob settings and ID mapping')
- argparser.add_argument('--layer-knobs', help='per-layer Knobs')
-
-
- args = argparser.parse_args()
-
- tunerData.binary_path = str(args.binary)
- tunerData.num_layers = int(args.num_layers)
- tunerData.accuracy_threshold = float(args.accuracy)
-
- # NOTE: Reading the cost file (with No of ops) to better guide the Autotuner
- cost_file_path = args.cost_file
- tunerData.layer_costs = readCostFile(cost_file_path)
-
- tunerData.knobs_list = readLayerKnobs(args.layer_knobs)
- tunerData.knobs_speedup = readKnobConfig(args.knobs_config)
-
-
- result_dir = args.result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
- tunerData.output_dir = result_dir + "/high_confidence/"
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- if not os.path.exists(tunerData.output_dir):
- print("Creating output directory = ", tunerData.output_dir)
- os.mkdir(tunerData.output_dir)
-
-
-
- ClangFlagsTuner.main(argparser.parse_args())
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/select_top_results.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/select_top_results.py
deleted file mode 100644
index 7ee878e5f8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/select_top_results.py
+++ /dev/null
@@ -1,101 +0,0 @@
-
-
-import argparse
-import sys
-import os
-
-
-log_index = 9
-linear_index = 10
-quad_index = 11
-
-top_k = 10
-skip_lines = 1
-
-
-def dump_results(sorted_list, k, result_dir, sub_dir):
-
- ref_dir = result_dir + "/" + sub_dir
- if not os.path.exists(ref_dir):
- os.mkdir(ref_dir)
-
- for i in range(min(k, len(sorted_list)) ):
- file_name = sorted_list[i][1]
- file_name = ref_dir + "/" + file_name + "_rank_" + str(i)
- f = open(file_name, "w+")
- f.write(str(sorted_list[i][2]) + "\t")
- f.write(str(sorted_list[i][3]) + "\t")
- f.write(str(sorted_list[i][4]) + "\n")
- f.write(sorted_list[i][0])
- f.close()
-
-
-
-
-def select_top_results(result_dir):
-
- if not os.path.exists(result_dir):
- print("Path does not exist")
- sys.exit(0)
-
- file_names = os.listdir(result_dir)
- print file_names
-
- results_arr = []
-
- for file_name in file_names:
-
- if file_name == "confidence_summary.txt":
- continue
-
- # Skip sub-directories
- if os.path.isdir(result_dir + "/" + file_name):
- continue
-
- log_result = 0.0
- linear_result = 0.0
- quad_result = 0.0
- file_str = ""
-
- index = 0
- f = open(result_dir + "/" + file_name)
- for x in f:
- if index >= skip_lines:
- words = x.split()
- log_result += float(words[log_index])
- linear_result += float(words[linear_index])
- quad_result += float(words[quad_index])
- file_str += x
-
- index += 1
-
-
- file_result = (file_str, file_name, log_result, linear_result, quad_result)
- results_arr.append(file_result)
-
-
- sorted_list = sorted(results_arr, key = lambda tup: tup[2])
- dump_results(sorted_list, top_k, result_dir, "log")
-
- sorted_list = sorted(results_arr, key = lambda tup: tup[3])
- dump_results(sorted_list, top_k, result_dir, "linear")
-
- sorted_list = sorted(results_arr, key = lambda tup: tup[4])
- dump_results(sorted_list, top_k, result_dir, "quad")
-
-
-#def select_top_configuration(result_dir):
-
-
-if __name__ == "__main__":
-
- argparser = argparse.ArgumentParser(description='runs best configs to get high confidence on accuracy')
- argparser.add_argument('--result-dir', help='Directory containing OpenTuner configurations')
-
- args = argparser.parse_args()
- result_dir = args.result_dir
-
- select_top_results(result_dir)
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/utils.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/utils.py
deleted file mode 100644
index 47429d9599..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/utils.py
+++ /dev/null
@@ -1,157 +0,0 @@
-
-
-import psutil
-from time import sleep
-import os
-
-
-
-def readCostFile(cost_file_path):
-
- layer_costs = []
- f = open(cost_file_path)
- for x in f:
- cost = float(x.strip())
- layer_costs.append(cost)
-
- print ("-Layer count = ", layer_costs)
- f.close()
-
- return layer_costs
-
-
-
-def readAccuracy(accuray_res_file):
-
- file = open(accuray_res_file, "r")
- accuracy_str = file.read()
- file.close()
- accuracy = 0
-
- try:
- accuracy = float(accuracy_str)
- except:
- accuracy = 0
-
- print ("*Configuration Accuracy = ", accuracy, "\n\n")
- return accuracy
-
-
-def genLayerFlagsFile(flags_file_path, cfg, tunerData):
-
- f = open(flags_file_path, "w+")
- cmd_config = ""
- for i in range(tunerData.num_layers): # flag in tunerData.tuning_flags:
- flag = tunerData.tuning_flags[i]
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-
-def readLayerKnobs(layer_knobs_path):
-
- f = open(layer_knobs_path, "r")
- knobs_list = []
- for x in f:
- knobs = []
- vals = x.split(",")
- for val in vals:
- knobs.append(int(val))
-
- knobs_list.append(knobs)
-
- print ("\n **** Global Approximation Knobs List = \n", knobs_list)
-
- return knobs_list
-
-
-
-def readGlobalKnobConfig(global_knobs_path):
-
- knobs_speedup = {}
- f = open(global_knobs_path, "r")
- for x in f:
- toks = x.split("\t")
- ID = int(toks[0].split(",")[1])
-
- speedup = float(toks[2])
- knobs_speedup[ID] = speedup
-
- print ("knobs_speedup = ", knobs_speedup)
-
- return knobs_speedup
-
-
-
-
-def computeConfigCost(cfg, tunerData):
-
- orig_cost = 0.0
- total_cost = 0.0
- for it in range(tunerData.num_layers):
- flag = tunerData.tuning_flags[it]
- flag_value = cfg[flag]
- op_cost = tunerData.layer_costs[it]
- speedup = tunerData.knobs_speedup[flag_value]
-
- total_cost += (op_cost * 1.0 / speedup * 1.0)
- orig_cost += op_cost
-
-
-
- speedup = (orig_cost * 1.0) / (total_cost * 1.0)
-
- return total_cost, speedup
-
-
-
-def addInfoToOutFile(f_path, accuracy, total_runs, total_comps, speedup):
-
- f_str = open(f_path, "r").read()
-
- f_out = open(f_path, "w+")
-
- f_out.write("total_runs=" + str(total_runs) + "\tconfidence=100.0" + "\tavg_accuracy=" + \
- str(accuracy) + "\tconfig_cost=" + str(total_comps) + "\tspeedup=" + str(speedup) + "\n" )
- f_out.write(f_str)
-
- f_out.close()
-
-
-
-def check_pid(pid):
-
- """ Check For the existence of a unix pid. """
- try:
- os.kill(pid, 0)
- except OSError:
- return False
- else:
- return True
-
-
-
-def process_kill(proc_pid):
-
- if not check_pid(proc_pid):
- return # Return if process does not exist
-
- process = psutil.Process(proc_pid)
-
- try:
- for proc in process.children(recursive=True):
- proc.kill()
- process.kill()
-
- print ("\n\n\n\n\n\ %%%%% Killed Process \n\n\n\n")
-
- except:
- print ("\n\n\n\n PROCESS NOT KILLED ------- \n\n\n\n\n\n\n")
-
- #sleep(20)
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/debian-packages-deps b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/debian-packages-deps
deleted file mode 100644
index ea49289a87..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/debian-packages-deps
+++ /dev/null
@@ -1,9 +0,0 @@
-build-essential
-git
-gnuplot
-libfreetype6-dev
-libpng-dev
-libsqlite3-dev
-python-dev
-python-pip
-sqlite3
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/Makefile b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/Makefile
deleted file mode 100644
index 1c028b3a91..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/Makefile
+++ /dev/null
@@ -1,177 +0,0 @@
-# Makefile for Sphinx documentation
-#
-
-# You can set these variables from the command line.
-SPHINXOPTS =
-SPHINXBUILD = sphinx-build
-PAPER =
-BUILDDIR = build
-
-# User-friendly check for sphinx-build
-ifeq ($(shell which $(SPHINXBUILD) >/dev/null 2>&1; echo $$?), 1)
-$(error The '$(SPHINXBUILD)' command was not found. Make sure you have Sphinx installed, then set the SPHINXBUILD environment variable to point to the full path of the '$(SPHINXBUILD)' executable. Alternatively you can add the directory with the executable to your PATH. If you don't have Sphinx installed, grab it from http://sphinx-doc.org/)
-endif
-
-# Internal variables.
-PAPEROPT_a4 = -D latex_paper_size=a4
-PAPEROPT_letter = -D latex_paper_size=letter
-ALLSPHINXOPTS = -d $(BUILDDIR)/doctrees $(PAPEROPT_$(PAPER)) $(SPHINXOPTS) source
-# the i18n builder cannot share the environment and doctrees with the others
-I18NSPHINXOPTS = $(PAPEROPT_$(PAPER)) $(SPHINXOPTS) source
-
-.PHONY: help clean html dirhtml singlehtml pickle json htmlhelp qthelp devhelp epub latex latexpdf text man changes linkcheck doctest gettext
-
-help:
- @echo "Please use \`make <target>' where <target> is one of"
- @echo " html to make standalone HTML files"
- @echo " dirhtml to make HTML files named index.html in directories"
- @echo " singlehtml to make a single large HTML file"
- @echo " pickle to make pickle files"
- @echo " json to make JSON files"
- @echo " htmlhelp to make HTML files and a HTML help project"
- @echo " qthelp to make HTML files and a qthelp project"
- @echo " devhelp to make HTML files and a Devhelp project"
- @echo " epub to make an epub"
- @echo " latex to make LaTeX files, you can set PAPER=a4 or PAPER=letter"
- @echo " latexpdf to make LaTeX files and run them through pdflatex"
- @echo " latexpdfja to make LaTeX files and run them through platex/dvipdfmx"
- @echo " text to make text files"
- @echo " man to make manual pages"
- @echo " texinfo to make Texinfo files"
- @echo " info to make Texinfo files and run them through makeinfo"
- @echo " gettext to make PO message catalogs"
- @echo " changes to make an overview of all changed/added/deprecated items"
- @echo " xml to make Docutils-native XML files"
- @echo " pseudoxml to make pseudoxml-XML files for display purposes"
- @echo " linkcheck to check all external links for integrity"
- @echo " doctest to run all doctests embedded in the documentation (if enabled)"
-
-clean:
- rm -rf $(BUILDDIR)/*
-
-html:
- $(SPHINXBUILD) -b html $(ALLSPHINXOPTS) $(BUILDDIR)/html
- @echo
- @echo "Build finished. The HTML pages are in $(BUILDDIR)/html."
-
-dirhtml:
- $(SPHINXBUILD) -b dirhtml $(ALLSPHINXOPTS) $(BUILDDIR)/dirhtml
- @echo
- @echo "Build finished. The HTML pages are in $(BUILDDIR)/dirhtml."
-
-singlehtml:
- $(SPHINXBUILD) -b singlehtml $(ALLSPHINXOPTS) $(BUILDDIR)/singlehtml
- @echo
- @echo "Build finished. The HTML page is in $(BUILDDIR)/singlehtml."
-
-pickle:
- $(SPHINXBUILD) -b pickle $(ALLSPHINXOPTS) $(BUILDDIR)/pickle
- @echo
- @echo "Build finished; now you can process the pickle files."
-
-json:
- $(SPHINXBUILD) -b json $(ALLSPHINXOPTS) $(BUILDDIR)/json
- @echo
- @echo "Build finished; now you can process the JSON files."
-
-htmlhelp:
- $(SPHINXBUILD) -b htmlhelp $(ALLSPHINXOPTS) $(BUILDDIR)/htmlhelp
- @echo
- @echo "Build finished; now you can run HTML Help Workshop with the" \
- ".hhp project file in $(BUILDDIR)/htmlhelp."
-
-qthelp:
- $(SPHINXBUILD) -b qthelp $(ALLSPHINXOPTS) $(BUILDDIR)/qthelp
- @echo
- @echo "Build finished; now you can run "qcollectiongenerator" with the" \
- ".qhcp project file in $(BUILDDIR)/qthelp, like this:"
- @echo "# qcollectiongenerator $(BUILDDIR)/qthelp/OpenTuner.qhcp"
- @echo "To view the help file:"
- @echo "# assistant -collectionFile $(BUILDDIR)/qthelp/OpenTuner.qhc"
-
-devhelp:
- $(SPHINXBUILD) -b devhelp $(ALLSPHINXOPTS) $(BUILDDIR)/devhelp
- @echo
- @echo "Build finished."
- @echo "To view the help file:"
- @echo "# mkdir -p $$HOME/.local/share/devhelp/OpenTuner"
- @echo "# ln -s $(BUILDDIR)/devhelp $$HOME/.local/share/devhelp/OpenTuner"
- @echo "# devhelp"
-
-epub:
- $(SPHINXBUILD) -b epub $(ALLSPHINXOPTS) $(BUILDDIR)/epub
- @echo
- @echo "Build finished. The epub file is in $(BUILDDIR)/epub."
-
-latex:
- $(SPHINXBUILD) -b latex $(ALLSPHINXOPTS) $(BUILDDIR)/latex
- @echo
- @echo "Build finished; the LaTeX files are in $(BUILDDIR)/latex."
- @echo "Run \`make' in that directory to run these through (pdf)latex" \
- "(use \`make latexpdf' here to do that automatically)."
-
-latexpdf:
- $(SPHINXBUILD) -b latex $(ALLSPHINXOPTS) $(BUILDDIR)/latex
- @echo "Running LaTeX files through pdflatex..."
- $(MAKE) -C $(BUILDDIR)/latex all-pdf
- @echo "pdflatex finished; the PDF files are in $(BUILDDIR)/latex."
-
-latexpdfja:
- $(SPHINXBUILD) -b latex $(ALLSPHINXOPTS) $(BUILDDIR)/latex
- @echo "Running LaTeX files through platex and dvipdfmx..."
- $(MAKE) -C $(BUILDDIR)/latex all-pdf-ja
- @echo "pdflatex finished; the PDF files are in $(BUILDDIR)/latex."
-
-text:
- $(SPHINXBUILD) -b text $(ALLSPHINXOPTS) $(BUILDDIR)/text
- @echo
- @echo "Build finished. The text files are in $(BUILDDIR)/text."
-
-man:
- $(SPHINXBUILD) -b man $(ALLSPHINXOPTS) $(BUILDDIR)/man
- @echo
- @echo "Build finished. The manual pages are in $(BUILDDIR)/man."
-
-texinfo:
- $(SPHINXBUILD) -b texinfo $(ALLSPHINXOPTS) $(BUILDDIR)/texinfo
- @echo
- @echo "Build finished. The Texinfo files are in $(BUILDDIR)/texinfo."
- @echo "Run \`make' in that directory to run these through makeinfo" \
- "(use \`make info' here to do that automatically)."
-
-info:
- $(SPHINXBUILD) -b texinfo $(ALLSPHINXOPTS) $(BUILDDIR)/texinfo
- @echo "Running Texinfo files through makeinfo..."
- make -C $(BUILDDIR)/texinfo info
- @echo "makeinfo finished; the Info files are in $(BUILDDIR)/texinfo."
-
-gettext:
- $(SPHINXBUILD) -b gettext $(I18NSPHINXOPTS) $(BUILDDIR)/locale
- @echo
- @echo "Build finished. The message catalogs are in $(BUILDDIR)/locale."
-
-changes:
- $(SPHINXBUILD) -b changes $(ALLSPHINXOPTS) $(BUILDDIR)/changes
- @echo
- @echo "The overview file is in $(BUILDDIR)/changes."
-
-linkcheck:
- $(SPHINXBUILD) -b linkcheck $(ALLSPHINXOPTS) $(BUILDDIR)/linkcheck
- @echo
- @echo "Link check complete; look for any errors in the above output " \
- "or in $(BUILDDIR)/linkcheck/output.txt."
-
-doctest:
- $(SPHINXBUILD) -b doctest $(ALLSPHINXOPTS) $(BUILDDIR)/doctest
- @echo "Testing of doctests in the sources finished, look at the " \
- "results in $(BUILDDIR)/doctest/output.txt."
-
-xml:
- $(SPHINXBUILD) -b xml $(ALLSPHINXOPTS) $(BUILDDIR)/xml
- @echo
- @echo "Build finished. The XML files are in $(BUILDDIR)/xml."
-
-pseudoxml:
- $(SPHINXBUILD) -b pseudoxml $(ALLSPHINXOPTS) $(BUILDDIR)/pseudoxml
- @echo
- @echo "Build finished. The pseudo-XML files are in $(BUILDDIR)/pseudoxml."
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/rtd-requirements.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/rtd-requirements.txt
deleted file mode 100644
index e30d149ed5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/rtd-requirements.txt
+++ /dev/null
@@ -1,5 +0,0 @@
-argparse>=1.2.1
-django==1.6.1
-fn>=0.2.12
-SQLAlchemy>=0.8.2
-virtualenv==1.9.1
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/conf.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/conf.py
deleted file mode 100644
index a27fabf403..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/conf.py
+++ /dev/null
@@ -1,261 +0,0 @@
-# -*- coding: utf-8 -*-
-#
-# OpenTuner documentation build configuration file, created by
-# sphinx-quickstart on Sat Jan 3 04:13:12 2015.
-#
-# This file is execfile()d with the current directory set to its
-# containing dir.
-#
-# Note that not all possible configuration values are present in this
-# autogenerated file.
-#
-# All configuration values have a default; values that are commented out
-# serve to show the default.
-
-import sys
-import os
-
-# If extensions (or modules to document with autodoc) are in another directory,
-# add these directories to sys.path here. If the directory is relative to the
-# documentation root, use os.path.abspath to make it absolute, like shown here.
-sys.path.insert(0, os.path.abspath('../..'))
-
-# -- General configuration ------------------------------------------------
-
-# If your documentation needs a minimal Sphinx version, state it here.
-#needs_sphinx = '1.0'
-
-# Add any Sphinx extension module names here, as strings. They can be
-# extensions coming with Sphinx (named 'sphinx.ext.*') or your custom
-# ones.
-extensions = [
- 'sphinx.ext.autodoc',
- 'sphinx.ext.pngmath',
-]
-
-# Add any paths that contain templates here, relative to this directory.
-templates_path = ['_templates']
-
-# The suffix of source filenames.
-source_suffix = '.rst'
-
-# The encoding of source files.
-#source_encoding = 'utf-8-sig'
-
-# The master toctree document.
-master_doc = 'index'
-
-# General information about the project.
-project = u'OpenTuner'
-copyright = u'2015, Jason Ansel'
-
-# The version info for the project you're documenting, acts as replacement for
-# |version| and |release|, also used in various other places throughout the
-# built documents.
-#
-# The short X.Y version.
-version = '0.0'
-# The full version, including alpha/beta/rc tags.
-release = '0.0'
-
-# The language for content autogenerated by Sphinx. Refer to documentation
-# for a list of supported languages.
-#language = None
-
-# There are two options for replacing |today|: either, you set today to some
-# non-false value, then it is used:
-#today = ''
-# Else, today_fmt is used as the format for a strftime call.
-#today_fmt = '%B %d, %Y'
-
-# List of patterns, relative to source directory, that match files and
-# directories to ignore when looking for source files.
-exclude_patterns = []
-
-# The reST default role (used for this markup: `text`) to use for all
-# documents.
-#default_role = None
-
-# If true, '()' will be appended to :func: etc. cross-reference text.
-#add_function_parentheses = True
-
-# If true, the current module name will be prepended to all description
-# unit titles (such as .. function::).
-#add_module_names = True
-
-# If true, sectionauthor and moduleauthor directives will be shown in the
-# output. They are ignored by default.
-#show_authors = False
-
-# The name of the Pygments (syntax highlighting) style to use.
-pygments_style = 'sphinx'
-
-# A list of ignored prefixes for module index sorting.
-#modindex_common_prefix = []
-
-# If true, keep warnings as "system message" paragraphs in the built documents.
-#keep_warnings = False
-
-
-# -- Options for HTML output ----------------------------------------------
-
-# The theme to use for HTML and HTML Help pages. See the documentation for
-# a list of builtin themes.
-html_theme = 'default'
-
-# Theme options are theme-specific and customize the look and feel of a theme
-# further. For a list of options available for each theme, see the
-# documentation.
-#html_theme_options = {}
-
-# Add any paths that contain custom themes here, relative to this directory.
-#html_theme_path = []
-
-# The name for this set of Sphinx documents. If None, it defaults to
-# "<project> v<release> documentation".
-#html_title = None
-
-# A shorter title for the navigation bar. Default is the same as html_title.
-#html_short_title = None
-
-# The name of an image file (relative to this directory) to place at the top
-# of the sidebar.
-#html_logo = None
-
-# The name of an image file (within the static path) to use as favicon of the
-# docs. This file should be a Windows icon file (.ico) being 16x16 or 32x32
-# pixels large.
-#html_favicon = None
-
-# Add any paths that contain custom static files (such as style sheets) here,
-# relative to this directory. They are copied after the builtin static files,
-# so a file named "default.css" will overwrite the builtin "default.css".
-html_static_path = ['_static']
-
-# Add any extra paths that contain custom files (such as robots.txt or
-# .htaccess) here, relative to this directory. These files are copied
-# directly to the root of the documentation.
-#html_extra_path = []
-
-# If not '', a 'Last updated on:' timestamp is inserted at every page bottom,
-# using the given strftime format.
-#html_last_updated_fmt = '%b %d, %Y'
-
-# If true, SmartyPants will be used to convert quotes and dashes to
-# typographically correct entities.
-#html_use_smartypants = True
-
-# Custom sidebar templates, maps document names to template names.
-#html_sidebars = {}
-
-# Additional templates that should be rendered to pages, maps page names to
-# template names.
-#html_additional_pages = {}
-
-# If false, no module index is generated.
-#html_domain_indices = True
-
-# If false, no index is generated.
-#html_use_index = True
-
-# If true, the index is split into individual pages for each letter.
-#html_split_index = False
-
-# If true, links to the reST sources are added to the pages.
-#html_show_sourcelink = True
-
-# If true, "Created using Sphinx" is shown in the HTML footer. Default is True.
-#html_show_sphinx = True
-
-# If true, "(C) Copyright ..." is shown in the HTML footer. Default is True.
-#html_show_copyright = True
-
-# If true, an OpenSearch description file will be output, and all pages will
-# contain a <link> tag referring to it. The value of this option must be the
-# base URL from which the finished HTML is served.
-#html_use_opensearch = ''
-
-# This is the file name suffix for HTML files (e.g. ".xhtml").
-#html_file_suffix = None
-
-# Output file base name for HTML help builder.
-htmlhelp_basename = 'OpenTunerdoc'
-
-
-# -- Options for LaTeX output ---------------------------------------------
-
-latex_elements = {
-# The paper size ('letterpaper' or 'a4paper').
-#'papersize': 'letterpaper',
-
-# The font size ('10pt', '11pt' or '12pt').
-#'pointsize': '10pt',
-
-# Additional stuff for the LaTeX preamble.
-#'preamble': '',
-}
-
-# Grouping the document tree into LaTeX files. List of tuples
-# (source start file, target name, title,
-# author, documentclass [howto, manual, or own class]).
-latex_documents = [
- ('index', 'OpenTuner.tex', u'OpenTuner Documentation',
- u'Jason Ansel', 'manual'),
-]
-
-# The name of an image file (relative to this directory) to place at the top of
-# the title page.
-#latex_logo = None
-
-# For "manual" documents, if this is true, then toplevel headings are parts,
-# not chapters.
-#latex_use_parts = False
-
-# If true, show page references after internal links.
-#latex_show_pagerefs = False
-
-# If true, show URL addresses after external links.
-#latex_show_urls = False
-
-# Documents to append as an appendix to all manuals.
-#latex_appendices = []
-
-# If false, no module index is generated.
-#latex_domain_indices = True
-
-
-# -- Options for manual page output ---------------------------------------
-
-# One entry per manual page. List of tuples
-# (source start file, name, description, authors, manual section).
-man_pages = [
- ('index', 'opentuner', u'OpenTuner Documentation',
- [u'Jason Ansel'], 1)
-]
-
-# If true, show URL addresses after external links.
-#man_show_urls = False
-
-
-# -- Options for Texinfo output -------------------------------------------
-
-# Grouping the document tree into Texinfo files. List of tuples
-# (source start file, target name, title, author,
-# dir menu entry, description, category)
-texinfo_documents = [
- ('index', 'OpenTuner', u'OpenTuner Documentation',
- u'Jason Ansel', 'OpenTuner', 'One line description of project.',
- 'Miscellaneous'),
-]
-
-# Documents to append as an appendix to all manuals.
-#texinfo_appendices = []
-
-# If false, no module index is generated.
-#texinfo_domain_indices = True
-
-# How to display URL addresses: 'footnote', 'no', or 'inline'.
-#texinfo_show_urls = 'footnote'
-
-# If true, do not generate a @detailmenu in the "Top" node's menu.
-#texinfo_no_detailmenu = False
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/index.rst b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/index.rst
deleted file mode 100644
index 48f7468982..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/index.rst
+++ /dev/null
@@ -1,27 +0,0 @@
-.. OpenTuner documentation master file, created by
- sphinx-quickstart on Sat Jan 3 04:13:12 2015.
- You can adapt this file completely to your liking, but it should at least
- contain the root `toctree` directive.
-
-Welcome to OpenTuner's documentation!
-=====================================
-This is still under construction
-
-
-Contents:
-
-.. toctree::
- :maxdepth: 2
-
- params
- techniques
-
-
-
-Indices and tables
-==================
-
-* :ref:`genindex`
-* :ref:`modindex`
-* :ref:`search`
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/params.rst b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/params.rst
deleted file mode 100644
index b8d08cd300..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/params.rst
+++ /dev/null
@@ -1,339 +0,0 @@
-.. currentmodule:: opentuner.search.manipulator
-
-****************
-Parameters
-****************
-
-This will be an overview of parameters in OpenTuner.
-
-Each Parameter instance is created with a name. Most methods in parameters operate on configurations, dict-like objects spawned by the ConfigurationManipulator. Configurations contain values corresponding to a collection of instances of named parameters.
-
-A Parameter’s methods may mutate the value in a configuration corresponding to the name of the particular parameter instance. These methods are called operators.
-
-==============================
-Parameter Types and Operators
-==============================
-
-Each parameter has a set of operators. These operators take in a set of parent configurations and mutate the corresponding parameter value in the first configuration according to the parent values. Operators form the set of available transformations for search techniques to generate new configurations to test.
-
-Operator methods can be identified by the prefix 'op#_', where # is the number of required input configurations. The prefix 'opn\_' specifies an arbitrary number of input configurations, as a list. The first argument into an operator is always the configuration that will be mutated. This is followed by the required parent configurations, then any required arguments, and finally optional arguments.
-
-Any operators defined for a Parameter are inherited by its subclasses.
-
------------------
-Parameter
------------------
-This is an abstract base interface for parameters.
-
-.. autoclass:: Parameter
-
- .. automethod:: op1_randomize
-
- .. automethod:: op3_swarm
-
- .. automethod:: op4_set_linear
-
- .. automethod:: opn_stochastic_mix
-
-
--------------------------
-Primitive Parameter
--------------------------
-.. autoclass:: PrimitiveParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`Parameter.op1_randomize`,
- :meth:`Parameter.op3_swarm`,
- :meth:`Parameter.opn_stochastic_mix`
-
- .. automethod:: op1_normal_mutation
-
- **This paragraph can have examples for the above operator**
-
- .. automethod:: op4_set_linear
-
-
-------------------------
-Numeric Parameter
-------------------------
-.. autoclass:: NumericParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`PrimitiveParameter.op1_normal_mutation`,
- :meth:`Parameter.op3_swarm`,
- :meth:`PrimitiveParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`
-
- .. automethod:: op1_randomize
-
- .. automethod:: op1_scale
-
- .. automethod:: op3_difference
-
- .. automethod:: opn_sum
-
-
-------------------------
-Integer Parameter
-------------------------
-.. autoclass:: IntegerParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`PrimitiveParameter.op1_normal_mutation`,
- :meth:`NumericParameter.op1_randomize`,
- :meth:`NumericParameter.op1_scale`,
- :meth:`NumericParameter.op3_difference`,
- :meth:`PrimitiveParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`,
- :meth:`NumericParameter.opn_sum`
-
- .. automethod:: op3_swarm
-
-
-------------------------
-Float Parameter
-------------------------
-.. autoclass:: FloatParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`PrimitiveParameter.op1_normal_mutation`,
- :meth:`NumericParameter.op1_randomize`,
- :meth:`NumericParameter.op1_scale`,
- :meth:`NumericParameter.op3_difference`,
- :meth:`PrimitiveParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`,
- :meth:`NumericParameter.opn_sum`
-
- .. automethod:: op3_swarm
-
-
-------------------------
-ScaledNumericParameter
-------------------------
-.. autoclass:: ScaledNumericParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`PrimitiveParameter.op1_normal_mutation`,
- :meth:`NumericParameter.op1_randomize`,
- :meth:`NumericParameter.op1_scale`,
- :meth:`NumericParameter.op3_difference`,
- :meth:`Parameter.op3_swarm`,
- :meth:`PrimitiveParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`,
- :meth:`NumericParameter.opn_sum`
-
-
-------------------------
-LogIntegerParameter
-------------------------
-.. autoclass:: LogIntegerParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`PrimitiveParameter.op1_normal_mutation`,
- :meth:`NumericParameter.op1_randomize`,
- :meth:`NumericParameter.op1_scale`,
- :meth:`NumericParameter.op3_difference`,
- :meth:`FloatParameter.op3_swarm`,
- :meth:`PrimitiveParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`,
- :meth:`NumericParameter.opn_sum`
-
-
-------------------------
-LogFloatParameter
-------------------------
-.. autoclass:: LogFloatParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`PrimitiveParameter.op1_normal_mutation`,
- :meth:`NumericParameter.op1_randomize`,
- :meth:`NumericParameter.op1_scale`,
- :meth:`NumericParameter.op3_difference`,
- :meth:`FloatParameter.op3_swarm`,
- :meth:`PrimitiveParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`,
- :meth:`NumericParameter.opn_sum`
-
-
-------------------------
-PowerOfTwoParameter
-------------------------
-.. autoclass:: LogFloatParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`PrimitiveParameter.op1_normal_mutation`,
- :meth:`NumericParameter.op1_randomize`,
- :meth:`NumericParameter.op1_scale`,
- :meth:`NumericParameter.op3_difference`,
- :meth:`IntegerParameter.op3_swarm`,
- :meth:`PrimitiveParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`,
- :meth:`NumericParameter.opn_sum`
-
-
-------------------------
-Complex Parameter
-------------------------
-.. autoclass:: ComplexParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`Parameter.op3_swarm`,
- :meth:`Parameter.opn_stochastic_mix`
-
- .. automethod:: op1_randomize
-
- .. automethod:: op4_set_linear
-
-
-------------------------
-Boolean Parameter
-------------------------
-.. autoclass:: BooleanParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`Parameter.op3_swarm`,
- :meth:`ComplexParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`
-
- .. automethod:: op1_flip
-
- .. automethod:: op1_randomize
-
- .. automethod:: op3_swarm
-
---------------------------
-Switch Parameter
---------------------------
-.. autoclass:: SwitchParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`Parameter.op3_swarm`,
- :meth:`ComplexParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`
-
- .. automethod:: op1_randomize
-
---------------------------
-Enum Parameter
---------------------------
-.. autoclass:: EnumParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`Parameter.op3_swarm`,
- :meth:`ComplexParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`
-
- .. automethod:: op1_randomize
-
-
---------------------------
-Permutation Parameter
---------------------------
-.. autoclass:: PermutationParameter
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`ComplexParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`
-
- .. automethod:: op1_randomize
-
- .. automethod:: op1_small_random_change
-
- .. automethod:: op2_random_swap
-
- .. automethod:: op2_random_invert
-
- .. automethod:: op3_cross
-
- .. automethod:: op3_cross_PX
-
- .. automethod:: op3_cross_PMX
-
- .. automethod:: op3_cross_CX
-
- .. automethod:: op3_cross_OX1
-
- .. automethod:: op3_cross_OX3
-
- .. automethod:: op3_swarm
-
---------------------------
-Array
---------------------------
-.. autoclass:: Array
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`ComplexParameter.op1_randomize`,
- :meth:`ComplexParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`
-
- .. automethod:: op3_cross
-
- .. automethod:: op3_swarm
-
-
---------------------------
-BooleanArray
---------------------------
-.. autoclass:: BooleanArray
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`Array.op3_cross`,
- :meth:`Array.op3_swarm`,
- :meth:`ComplexParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`
-
- .. automethod:: op1_randomize
-
- .. automethod:: op3_swarm_parallel
-
-
---------------------------
-FloatArray
---------------------------
-.. autoclass:: FloatArray
- :show-inheritance:
-
- *Inherited Operators:*
-
- :meth:`Array.op3_cross`,
- :meth:`Array.op3_swarm`,
- :meth:`ComplexParameter.op4_set_linear`,
- :meth:`Parameter.opn_stochastic_mix`
-
- .. automethod:: op1_randomize
-
- .. automethod:: op3_swarm_parallel
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/techniques.rst b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/techniques.rst
deleted file mode 100644
index 3bbebddedb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/docs/source/techniques.rst
+++ /dev/null
@@ -1,51 +0,0 @@
-.. currentmodule:: opentuner.search.composableevolutionarytechniques
-
-********************
-Current Techniques
-********************
-
-OpenTuner has a library of existing search techniques.
-
-=================================
-Composable Search Techniques
-=================================
-
-A ComposableEvolutionaryTechnique allows for composition between the search technique and any operators. Creating a ComposableEvolutionaryTechnique requires implementing 3 methods:
-
- * :meth:`minimum_number_of_parents <ComposableEvolutionaryTechnique.minimum_number_of_parents>`
- * :meth:`get_parents <ComposableEvolutionaryTechnique.get_parents>`
- * :meth:`update_population <ComposableEvolutionaryTechnique.update_population>`
-
-Additionally, the following methods may be overridden for further customization
-
- * :meth:`make_population_member <ComposableEvolutionaryTechnique.make_population_member>`
- * :meth:`select_parameters <ComposableEvolutionaryTechnique.select_parameters>`
- * :meth:`get_default_operator <ComposableEvolutionaryTechnique.get_default_operator>`
-
-The following methods are useful when choosing parents or updating the population:
-
- * :meth:`lt <ComposableEvolutionaryTechnique.lt>`
- * :meth:`lte <ComposableEvolutionaryTechnique.lte>`
- * :meth:`get_global_best_configuration <ComposableEvolutionaryTechnique.get_global_best_configuration>`
-
-A ComposableEvolutionaryTechnique will yields configurations generated by successive iterations of applying operators on the configurations returned by :meth:`get_parents <ComposableEvolutionaryTechnique.get_parents>` and updating the population with the new configuration through :meth:`update_population <ComposableEvolutionaryTechnique.update_population>`
-
-.. autoclass:: ComposableEvolutionaryTechnique
-
- .. automethod:: minimum_number_of_parents
-
- .. automethod:: get_parents
-
- .. automethod:: update_population
-
- .. automethod:: make_population_member
-
- .. automethod:: select_parameters
-
- .. automethod:: get_default_operator
-
- .. automethod:: lt
-
- .. automethod:: lte
-
- .. automethod:: get_global_best_configuration
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/.gitignore b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/.gitignore
deleted file mode 100644
index f525a6259b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/.gitignore
+++ /dev/null
@@ -1,4 +0,0 @@
-*-journal
-stats
-opentuner.log
-opentuner.db
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/.gitignore b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/.gitignore
deleted file mode 100644
index fab2c2b13c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/.gitignore
+++ /dev/null
@@ -1,5 +0,0 @@
-tmp.bin
-cc_flags.json
-gccflags_final_config.cmd
-gccflags_final_config.json
-cc_params.json
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/adddeps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/adddeps.py
deleted file mode 100644
index ede22a8fcd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/adddeps.py
+++ /dev/null
@@ -1,6 +0,0 @@
-# we would prefer a symbolic link, but it does not work on windows
-import os
-target = os.path.join(os.path.dirname(__file__),
- '../../opentuner/utils/adddeps.py')
-execfile(target, dict(__file__=target))
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/.gitignore b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/.gitignore
deleted file mode 100644
index f06d3e01a2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/.gitignore
+++ /dev/null
@@ -1 +0,0 @@
-fft.c
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/matrixmultiply.cpp b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/matrixmultiply.cpp
deleted file mode 100644
index 9989ffbf4a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/matrixmultiply.cpp
+++ /dev/null
@@ -1,80 +0,0 @@
-// based on: http://blogs.msdn.com/b/xiangfan/archive/2009/04/28/optimize-your-code-matrix-multiplication.aspx
-// by Xiang Fan
-
-#include <algorithm>
-#include <iostream>
-
-#define N 512
-
-
-template<class T>
-T** make_test_matrix() {
- T** data = new T*[N];
- for (int i = 0; i < N; i++) {
- data[i] = new T[N];
- }
- for(int i = 0; i < N; i++) {
- for(int j = 0; j < N; j++) {
- data[i][j] = (int) i * j;
- }
- }
- return data;
-}
-
-
-
-template<typename T>
-void Transpose(int size, T** __restrict__ m)
-{
- for (int i = 0; i < size; i++) {
- for (int j = i + 1; j < size; j++) {
- std::swap(m[i][j], m[j][i]);
- }
- }
-}
-template<typename T>
-void SeqMatrixMult3(int size, T** __restrict__ m1, T** __restrict__ m2,
- T** __restrict__ result) {
- Transpose(size, m2);
- for (int i = 0; i < size; i++) {
- for (int j = 0; j < size; j++) {
- T c = 0;
- for (int k = 0; k < size; k++) {
- c += m1[i][k] * m2[j][k];
- }
- result[i][j] = c;
- }
- }
- Transpose(size, m2);
-}
-
-
-template<typename T>
-void test() {
- T** a = make_test_matrix<T>();
- T** b = make_test_matrix<T>();
- T** c = make_test_matrix<T>();
- SeqMatrixMult3(N, a, b, c);
-
-
- T avg = 0;
- for(int i = 0; i < N; i++) {
- for(int j = 0; j < N; j++) {
- avg += c[i][j] / (T)(N*N);
- }
- }
- // print out average so caller can check answer
- std::cout << avg << std::endl;
-}
-
-
-int main(int argc, const char** argv) {
- test<float>();
- return 0;
-}
-
-
-
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/raytracer.cpp b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/raytracer.cpp
deleted file mode 100644
index 3cb1192c6a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/raytracer.cpp
+++ /dev/null
@@ -1,277 +0,0 @@
-/*
- A very basic raytracer example.
- Copyright (C) 2012 www.scratchapixel.com
-
- This program is free software: you can redistribute it and/or modify
- it under the terms of the GNU General Public License as published by
- the Free Software Foundation, either version 3 of the License, or
- (at your option) any later version.
-
- This program is distributed in the hope that it will be useful,
- but WITHOUT ANY WARRANTY; without even the implied warranty of
- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- GNU General Public License for more details.
-
- You should have received a copy of the GNU General Public License
- along with this program. If not, see <http://www.gnu.org/licenses/>.
-
- - changes 02/04/13: fixed flag in ofstream causing a bug under Windows,
- added default values for M_PI and INFINITY
- - changes 24/05/13: small change to way we compute the refraction direction
- vector (eta=ior if we are inside and 1/ior if we are outside the sphere)
-
- Compile with the following command: c++ -o raytracer -O3 -Wall raytracer.cpp
-
-*/
-
-#include <cstdlib>
-#include <cstdio>
-#include <cmath>
-#include <fstream>
-#include <vector>
-#include <iostream>
-#include <cassert>
-
-#if defined(__linux__) || defined(__APPLE__)
- // "Compiled for Linux
-#else
- // Windows doesn't define these values by default, Linux does
- #define M_PI 3.141592653589793
- #define INFINITY 1e8
-#endif
-
-template<typename T>
-class Vec3
-{
-public:
- T x, y, z;
- Vec3() : x(T(0)), y(T(0)), z(T(0)) {}
- Vec3(T xx) : x(xx), y(xx), z(xx) {}
- Vec3(T xx, T yy, T zz) : x(xx), y(yy), z(zz) {}
- Vec3& normalize()
- {
- T nor2 = length2();
- if (nor2 > 0) {
- T invNor = 1 / sqrt(nor2);
- x *= invNor, y *= invNor, z *= invNor;
- }
- return *this;
- }
- Vec3<T> operator * (const T &f) const { return Vec3<T>(x * f, y * f, z * f); }
- Vec3<T> operator * (const Vec3<T> &v) const { return Vec3<T>(x * v.x, y * v.y, z * v.z); }
- T dot(const Vec3<T> &v) const { return x * v.x + y * v.y + z * v.z; }
- Vec3<T> operator - (const Vec3<T> &v) const { return Vec3<T>(x - v.x, y - v.y, z - v.z); }
- Vec3<T> operator + (const Vec3<T> &v) const { return Vec3<T>(x + v.x, y + v.y, z + v.z); }
- Vec3<T>& operator += (const Vec3<T> &v) { x += v.x, y += v.y, z += v.z; return *this; }
- Vec3<T>& operator *= (const Vec3<T> &v) { x *= v.x, y *= v.y, z *= v.z; return *this; }
- Vec3<T> operator - () const { return Vec3<T>(-x, -y, -z); }
- T length2() const { return x * x + y * y + z * z; }
- T length() const { return sqrt(length2()); }
- friend std::ostream & operator << (std::ostream &os, const Vec3<T> &v)
- {
- os << "[" << v.x << " " << v.y << " " << v.z << "]";
- return os;
- }
-};
-
-template<typename T>
-class Sphere
-{
-public:
- Vec3<T> center; /// position of the sphere
- T radius, radius2; /// sphere radius and radius^2
- Vec3<T> surfaceColor, emissionColor; /// surface color and emission (light)
- T transparency, reflection; /// surface transparency and reflectivity
- Sphere(const Vec3<T> &c, const T &r, const Vec3<T> &sc,
- const T &refl = 0, const T &transp = 0, const Vec3<T> &ec = 0) :
- center(c), radius(r), radius2(r * r), surfaceColor(sc), emissionColor(ec),
- transparency(transp), reflection(refl)
- {}
- // compute a ray-sphere intersection using the geometric solution
- bool intersect(const Vec3<T> &rayorig, const Vec3<T> &raydir, T *t0 = NULL, T *t1 = NULL) const
- {
- Vec3<T> l = center - rayorig;
- T tca = l.dot(raydir);
- if (tca < 0) return false;
- T d2 = l.dot(l) - tca * tca;
- if (d2 > radius2) return false;
- T thc = sqrt(radius2 - d2);
- if (t0 != NULL && t1 != NULL) {
- *t0 = tca - thc;
- *t1 = tca + thc;
- }
-
- return true;
- }
-};
-
-#define MAX_RAY_DEPTH 5
-
-template<typename T>
-T mix(const T &a, const T &b, const T &mix)
-{
- return b * mix + a * (T(1) - mix);
-}
-
-// This is the main trace function. It takes a ray as argument (defined by its origin
-// and direction). We test if this ray intersects any of the geometry in the scene.
-// If the ray intersects an object, we compute the intersection point, the normal
-// at the intersection point, and shade this point using this information.
-// Shading depends on the surface property (is it transparent, reflective, diffuse).
-// The function returns a color for the ray. If the ray intersects an object that
-// is the color of the object at the intersection point, otherwise it returns
-// the background color.
-template<typename T>
-Vec3<T> trace(const Vec3<T> &rayorig, const Vec3<T> &raydir,
- const std::vector<Sphere<T> *> &spheres, const int &depth)
-{
- //if (raydir.length() != 1) std::cerr << "Error " << raydir << std::endl;
- T tnear = INFINITY;
- const Sphere<T> *sphere = NULL;
- // find intersection of this ray with the sphere in the scene
- for (unsigned i = 0; i < spheres.size(); ++i) {
- T t0 = INFINITY, t1 = INFINITY;
- if (spheres[i]->intersect(rayorig, raydir, &t0, &t1)) {
- if (t0 < 0) t0 = t1;
- if (t0 < tnear) {
- tnear = t0;
- sphere = spheres[i];
- }
- }
- }
- // if there's no intersection return black or background color
- if (!sphere) return Vec3<T>(2);
- Vec3<T> surfaceColor = 0; // color of the ray/surfaceof the object intersected by the ray
- Vec3<T> phit = rayorig + raydir * tnear; // point of intersection
- Vec3<T> nhit = phit - sphere->center; // normal at the intersection point
- nhit.normalize(); // normalize normal direction
- // If the normal and the view direction are not opposite to each other
- // reverse the normal direction. That also means we are inside the sphere so set
- // the inside bool to true. Finally reverse the sign of IdotN which we want
- // positive.
- T bias = 1e-4; // add some bias to the point from which we will be tracing
- bool inside = false;
- if (raydir.dot(nhit) > 0) nhit = -nhit, inside = true;
- if ((sphere->transparency > 0 || sphere->reflection > 0) && depth < MAX_RAY_DEPTH) {
- T facingratio = -raydir.dot(nhit);
- // change the mix value to tweak the effect
- T fresneleffect = mix<T>(pow(1 - facingratio, 3), 1, 0.1);
- // compute reflection direction (not need to normalize because all vectors
- // are already normalized)
- Vec3<T> refldir = raydir - nhit * 2 * raydir.dot(nhit);
- refldir.normalize();
- Vec3<T> reflection = trace(phit + nhit * bias, refldir, spheres, depth + 1);
- Vec3<T> refraction = 0;
- // if the sphere is also transparent compute refraction ray (transmission)
- if (sphere->transparency) {
- T ior = 1.1, eta = (inside) ? ior : 1 / ior; // are we inside or outside the surface?
- T cosi = -nhit.dot(raydir);
- T k = 1 - eta * eta * (1 - cosi * cosi);
- Vec3<T> refrdir = raydir * eta + nhit * (eta * cosi - sqrt(k));
- refrdir.normalize();
- refraction = trace(phit - nhit * bias, refrdir, spheres, depth + 1);
- }
- // the result is a mix of reflection and refraction (if the sphere is transparent)
- surfaceColor = (reflection * fresneleffect +
- refraction * (1 - fresneleffect) * sphere->transparency) * sphere->surfaceColor;
- }
- else {
- // it's a diffuse object, no need to raytrace any further
- for (unsigned i = 0; i < spheres.size(); ++i) {
- if (spheres[i]->emissionColor.x > 0) {
- // this is a light
- Vec3<T> transmission = 1;
- Vec3<T> lightDirection = spheres[i]->center - phit;
- lightDirection.normalize();
- for (unsigned j = 0; j < spheres.size(); ++j) {
- if (i != j) {
- T t0, t1;
- if (spheres[j]->intersect(phit + nhit * bias, lightDirection, &t0, &t1)) {
- transmission = 0;
- break;
- }
- }
- }
- surfaceColor += sphere->surfaceColor * transmission *
- std::max(T(0), nhit.dot(lightDirection)) * spheres[i]->emissionColor;
- }
- }
- }
-
- return surfaceColor + sphere->emissionColor;
-}
-
-// Main rendering function. We compute a camera ray for each pixel of the image
-// trace it and return a color. If the ray hits a sphere, we return the color of the
-// sphere at the intersection point, else we return the background color.
-template<typename T>
-unsigned int render(const std::vector<Sphere<T> *> &spheres)
-{
- unsigned width = 640, height = 480;
- Vec3<T> *image = new Vec3<T>[width * height], *pixel = image;
- T invWidth = 1 / T(width), invHeight = 1 / T(height);
- T fov = 30, aspectratio = width / T(height);
- T angle = tan(M_PI * 0.5 * fov / T(180));
- // Trace rays
- for (unsigned y = 0; y < height; ++y) {
- for (unsigned x = 0; x < width; ++x, ++pixel) {
- T xx = (2 * ((x + 0.5) * invWidth) - 1) * angle * aspectratio;
- T yy = (1 - 2 * ((y + 0.5) * invHeight)) * angle;
- Vec3<T> raydir(xx, yy, -1);
- raydir.normalize();
- *pixel = trace(Vec3<T>(0), raydir, spheres, 0);
- }
- }
-#if 0
- // Save result to a PPM image (keep these flags if you compile under Windows)
- std::ofstream ofs("./untitled.ppm", std::ios::out | std::ios::binary);
- ofs << "P6\n" << width << " " << height << "\n255\n";
- for (unsigned i = 0; i < width * height; ++i) {
- ofs << (unsigned char)(std::min(T(1), image[i].x) * 255) <<
- (unsigned char)(std::min(T(1), image[i].y) * 255) <<
- (unsigned char)(std::min(T(1), image[i].z) * 255);
- }
- ofs.close();
-#endif
-
- unsigned int bad_hash = 0;
- for (unsigned i = 0; i < width * height; ++i) {
- bad_hash = bad_hash*31 + (unsigned int)(std::min(T(1), image[i].x) * 255);
- bad_hash = bad_hash*31 + (unsigned int)(std::min(T(1), image[i].y) * 255);
- bad_hash = bad_hash*31 + (unsigned int)(std::min(T(1), image[i].z) * 255);
- }
- delete [] image;
-
- return bad_hash;
-}
-
-volatile unsigned int dont_optimize_me;
-
-int main(int argc, char **argv) {
- srand48(13);
- std::vector<Sphere<float> *> spheres;
- // position, radius, surface color, reflectivity, transparency, emission color
- spheres.push_back(new Sphere<float>(Vec3<float>(0, -10004, -20), 10000, Vec3<float>(0.2), 0, 0.0));
- spheres.push_back(new Sphere<float>(Vec3<float>(0, 0, -20), 4, Vec3<float>(1.00, 0.32, 0.36), 1, 0.5));
- spheres.push_back(new Sphere<float>(Vec3<float>(5, -1, -15), 2, Vec3<float>(0.90, 0.76, 0.46), 1, 0.0));
- spheres.push_back(new Sphere<float>(Vec3<float>(5, 0, -25), 3, Vec3<float>(0.65, 0.77, 0.97), 1, 0.0));
- spheres.push_back(new Sphere<float>(Vec3<float>(-5.5, 0, -15), 3, Vec3<float>(0.90, 0.90, 0.90), 1, 0.0));
- // light
- spheres.push_back(new Sphere<float>(Vec3<float>(0, 20, -30), 3, Vec3<float>(0), 0, 0, Vec3<float>(3)));
-
- dont_optimize_me = render<float>(spheres);
- __asm__ __volatile__ ("" ::: "memory"); // memory barrier
- if(dont_optimize_me == 0x4bd7c0e0) {
- //printf("CORRECT\n");
- } else {
- printf("ERROR: WRONG ANSWER\n");
- }
-
- while (!spheres.empty()) {
- Sphere<float> *sph = spheres.back();
- spheres.pop_back();
- delete sph;
- }
-
- return 0;
-}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/tsp_ga.cpp b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/tsp_ga.cpp
deleted file mode 100644
index 0e8f232cb0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/apps/tsp_ga.cpp
+++ /dev/null
@@ -1,548 +0,0 @@
-//
-// based on: https://bitbucket.org/knordkvist/tsp-ga/overview
-// by Kristoffer Nordkvist
-//
-#include <algorithm>
-#include <assert.h>
-#include <iostream>
-#include <limits>
-#include <math.h>
-#include <sstream>
-#include <stdio.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <stdlib.h>
-#include <string.h>
-#include <string>
-#include <time.h>
-
-class TSP
-{
- public:
- TSP(const double crossoverProbability, const double mutationProbability);
-
- /* The constants used in this project */
- static const unsigned int chromosones = 30, cities = 20, xMin = 0, xMax = 1000, yMin = 0, yMax = 500;
-
- /* Generate a random population of chromosones */
- void randomPopulation();
-
- /* Create a new population using crossover and mutation */
- void nextPopulation();
-
- /* Returns the fitness of the best chromosone */
- double getBestFitness() const;
-
- /* Returns a string representation of the best path */
- std::string getBestPathString() const;
-
- /* Returns the total distance of the best chromosone path */
- double getLowestTotalDistance() const;
-
- /* Returns the populations average length */
- double getAverageDistance() const;
- private:
- const double crossoverProbability, mutationProbability;
-
- /* Gets the total distance of the supplied path */
- double totalDistance(int const * const chromosone) const;
-
- /* The coordinates for each city, (x,y) for the first city is found in (citiesX[0], citiesY[0]) */
- double citiesX[cities], citiesY[cities];
-
- /* The chromosone containing the shortest path */
- int *bestChromosone;
-
- /* Contains the current population of chromosones */
- int (* solutions)[cities],
- /* The two chromosones with the best fitness functions */
- //bestChromosone1[cities], bestChromosone2[cities],
- /* Used to store the new chromosones when creating a new population */
- (* newPopulation)[cities];
-
- /* Returns a random double r, 0 <= r <= max */
- static double randomInclusive(const double max);
-
- /* Returns a random double r, 0 <= r < max */
- static double randomExclusive(const double max);
-
- /* True if the two chromosones represent the same path */
- static bool areChromosonesEqual(int const * const chromosoneA, int const * const chromosoneB);
-
- /* Evaluate the fitness the supplied chromosone */
- double evaluateFitness(const int * const chromosone) const;
-
- /* Selects a chromosone from the current population using Roulette Wheel Selection.
- * Using the algorithm described in http://www.obitko.com/tutorials/genetic-algorithms/selection.php.
- */
- int * rouletteSelection(double const * const fitness) const;
-
- /* Replace the element at offspringIndex with the first element found in other that does not exist in offspringToRepair */
- void repairOffspring(int * const offspringToRepair, int missingIndex, const int * const other);
-
- /* Might swap one gene with another, depending on the mutation probability */
- void mutate(int * const chromosone);
-
- /* Cross over the parents to form new offspring using Multi-Point Crossover, collisions are handled as shown in lecture 5.
- * The chromosones might be a copy of their parents, depending on the crossover probability.
- */
- void crossover(const int * const parentA, const int * const parentB, int * const offspringA, int * const offspringB);
-
- /* Checks if the supplied chromosone is in newPopulation */
- bool hasDuplicate(const int * const chromosone, size_t populationCount);
-
- /* Copies the supplied chromosone to the new population */
- void copyToNewPopulation(const int * const chromosone, size_t index);
-
- /* Make the chromosone represent a path, which is chosen by random */
- static void setRandomPath(int * const chromosone);
-};
-
-using namespace std;
-
-TSP::TSP(double crossoverProbability, double mutationProbability) : crossoverProbability(crossoverProbability),
- mutationProbability(mutationProbability), solutions(new int[chromosones][cities]), newPopulation(new int[chromosones][cities])
-{
- /* Seed the random number generator */
- //srand((unsigned int)time(NULL));
- srand(17);
- /* Use the same number to generate a specific sequence */
- //srand(0);
- /* Set random coordinates */
- for(size_t coordinateIndex = 0; coordinateIndex < cities; ++coordinateIndex)
- {
- /* 0 <= x <= xMax */
- citiesX[coordinateIndex] = randomInclusive(xMax);
- /* 0 <= y <= yMax */
- citiesY[coordinateIndex] = randomInclusive(yMax);
- }
-
- /* Generate random population */
- randomPopulation();
-}
-
-void TSP::randomPopulation()
-{
- /* Iterate throught each chromosone... */
- for(size_t chromosoneIndex = 0; chromosoneIndex < chromosones; ++chromosoneIndex)
- {
- /* ... and give it a random path */
- setRandomPath(solutions[chromosoneIndex]);
- }
-}
-
-double TSP::getBestFitness() const
-{
- return evaluateFitness(bestChromosone);
-}
-
-double TSP::getAverageDistance() const
-{
- double distance = 0;
- for(size_t chromosoneIndex = 0; chromosoneIndex < chromosones; ++chromosoneIndex)
- {
- distance += totalDistance(solutions[chromosoneIndex]);
- }
- return distance/chromosones;
-}
-
-string TSP::getBestPathString() const
-{
- stringstream path;
- for(size_t gene = 0; gene < cities; ++gene)
- {
- if(gene != 0)
- {
- path << ",";
- }
- path << bestChromosone[gene];
- }
- return path.str();
-}
-
-double TSP::getLowestTotalDistance() const
-{
- return totalDistance(bestChromosone);
-}
-
-void TSP::nextPopulation()
-{
- double fitness[chromosones];
- /* Fill an array with a fitness score for each chromosone,
- * the index of a score corresponds with the chromosone's index in solutions[index]
- */
- for(size_t chromosoneIndex = 0; chromosoneIndex < chromosones; ++chromosoneIndex)
- {
- fitness[chromosoneIndex] = evaluateFitness(solutions[chromosoneIndex]);
- }
-
- /* Use elitism, find and copy over the two best chromosones to the new population */
- int eliteIndex1 = 0, eliteIndex2 = 0;
- /* find the best solution */
- eliteIndex1 = max_element(fitness, fitness + chromosones) - fitness;
- this->bestChromosone = solutions[eliteIndex1];
-
- double highestFitness = 0;
- /* Find the second best solution */
- for(size_t chromosoneIndex = 0; chromosoneIndex < chromosones; ++chromosoneIndex)
- {
- if(chromosoneIndex != eliteIndex1 && fitness[chromosoneIndex] > highestFitness)
- {
- highestFitness = fitness[chromosoneIndex];
- eliteIndex2 = chromosoneIndex;
- }
- }
-
- /* Keep track of how many chromosones exists in the new population */
- size_t offspringCount = 0;
- /* Copy over the two best solutions to the new population */
- copyToNewPopulation(solutions[eliteIndex1], offspringCount);
- ++offspringCount;
- copyToNewPopulation(solutions[eliteIndex2], offspringCount);
- ++offspringCount;
-
- /* Create the rest of the new population, break this loop when the new population is complete */
- while(true)
- {
- int * parentA;
- int * parentB;
- parentA = rouletteSelection(fitness);
- parentB = rouletteSelection(fitness);
- while (parentB == parentA)
- {
- parentB = rouletteSelection(fitness);
- }
- int offspringA[cities];
- int offspringB[cities];
- crossover(parentA, parentB, offspringA, offspringB);
- mutate(offspringA);
- mutate(offspringB);
-
- /* Add to new population if an equal chromosone doesn't exist already */
- if(!hasDuplicate(offspringA, offspringCount))
- {
- copyToNewPopulation(offspringA, offspringCount);
- ++offspringCount;
- }
- /* We need to check if the new population is filled */
- if(offspringCount == chromosones)
- {
- break;
- }
- if(!hasDuplicate(offspringB, offspringCount))
- {
- copyToNewPopulation(offspringB, offspringCount);
- ++offspringCount;
- }
- /* Check again so that we don't accidentaly write all over the heap and have to spend an evening wondering why the heap is corrupt... :) */
- if(offspringCount == chromosones)
- {
- break;
- }
- }
-
- /*
- * We now have a new population,
- * now it needs to replace the current population
- * so that we don't go through the same population every time we run this function
- */
- for(size_t chromosoneIndex = 0; chromosoneIndex < chromosones; ++chromosoneIndex)
- {
- memcpy(solutions[chromosoneIndex], newPopulation[chromosoneIndex], sizeof(int) * cities);
- }
-}
-
-bool TSP::hasDuplicate(const int * const chromosone, size_t populationCount)
-{
- /* Iterate throught each chromosone in newPopulation and compare them gene by gene */
- for(size_t chromosoneIndex = 0; chromosoneIndex < populationCount; ++chromosoneIndex)
- {
- int genesCompared = 0;
- for(size_t gene = 0; gene < cities; ++gene)
- {
- if(chromosone[gene] != newPopulation[chromosoneIndex][gene])
- {
- /* These chromosones are not equal! */
- break;
- }
- ++genesCompared;
- }
-
- if(genesCompared == cities)
- {
- return true;
- }
- }
-
- return false;
-}
-
-void TSP::mutate(int * const chromosone)
-{
- /* 0.0 <= random <= 1 */
- {
- double random = randomInclusive(1);
- /* Nope, didn't happen */
- if(random > mutationProbability)
- {
- return;
- }
- }
-
- int tmp;
- int random1 = (int)randomExclusive(cities);
- int random2 = (int)randomExclusive(cities);
- while(random1 == random2)
- {
- random2 = (int)randomExclusive(cities);
- }
-
- tmp = chromosone[random1];
- chromosone[random1] = chromosone[random2];
- chromosone[random2] = tmp;
-
-}
-
-void TSP::crossover(int const * const parentA, const int * const parentB, int * offspringA, int * offspringB)
-{
- {
- /* There is a chance we don't perform a crossover,
- * in that case the offspring is a copy of the parents
- */
- /* 0.0 <= random <= 1 */
- double random = randomInclusive(1);
- /* The offspring is a copy of their parents */
- if(random > crossoverProbability)
- {
- memcpy(offspringA, parentA, sizeof(int) * cities);
- memcpy(offspringB, parentB, sizeof(int) * cities);
- return;
- }
- }
- /* Perform multi-point crossover to generate offspring */
-
- /* 0 <= cuttOffIndex <= cities */
- int cuttOffIndex1 = (int)randomInclusive(cities);
- int cuttOffIndex2 = (int)randomInclusive(cities);
- while(cuttOffIndex2 == cuttOffIndex1)
- {
- cuttOffIndex2 = (int)randomExclusive(cities);
- }
-
- unsigned int start;
- unsigned int end;
- if(cuttOffIndex1 < cuttOffIndex2)
- {
- start = cuttOffIndex1;
- end = cuttOffIndex2;
- }
- else
- {
- start = cuttOffIndex2;
- end = cuttOffIndex1;
- }
- /* Offspring A is initially copy of parent A */
- memcpy(offspringA, parentA, sizeof(int) * cities);
- /* Offspring B is initially copy of parent B */
- memcpy(offspringB, parentB, sizeof(int) * cities);
-
- /* Put a sequence of parent B in offspring A */
- memcpy(offspringA + start, parentB + start, sizeof(int) * (end - start));
- /* Put a sequence of parent A in offspring B */
- memcpy(offspringB + start, parentA + start, sizeof(int) * (end - start));
-
- /* Mark collisions in offspring with -1*/
- for(size_t cityIndex = 0; cityIndex < cities; ++cityIndex)
- {
- /* Index is part of the parent sequence */
- if((cityIndex >= start && cityIndex < end)) {
- /* Do nothing, we want to keep this sequence intact */
- }
- else
- {
- /* Check if the item at cityIndex also occurs somewhere in the copied substring */
- for(size_t substringIndex = start; substringIndex < end; ++substringIndex)
- {
- /* A duplicate, mark it */
- if(offspringA[cityIndex] == offspringA[substringIndex])
- {
- offspringA[cityIndex] = -1;
- }
- if(offspringB[cityIndex] == offspringB[substringIndex])
- {
- offspringB[cityIndex] = -1;
- }
- }
- }
-
- }
-
- /*
- * Go through the offspring,
- * if an element is marked we fill the hole with an element from the other offspring
- */
- for(size_t offspringIndex = 0; offspringIndex < cities; ++offspringIndex)
- {
- /* There is a hole here */
- if(offspringA[offspringIndex] == -1)
- {
- repairOffspring(offspringA, offspringIndex, offspringB);
- }
- if(offspringB[offspringIndex] == -1)
- {
- repairOffspring(offspringB, offspringIndex, offspringA);
- }
- }
-}
-
-void TSP::repairOffspring(int * const offspringToRepair, int missingIndex, const int * const other)
-{
- /* Iterate through the other offspring until we find an element which doesn't exist in the offspring we are repairing */
- for(size_t patchIndex = 0; patchIndex < cities; ++patchIndex)
- {
- /* Look for other[patchIndex] in offspringToRepair */
- int *missing = find(offspringToRepair, offspringToRepair + cities, other[patchIndex]);
-
- /* The element at other[patchIndex] is missing from offspringToRepair */
- if(missing == (offspringToRepair + cities))
- {
- //cout << "1:" << offspringToRepair[missingIndex] << endl;
- offspringToRepair[missingIndex] = other[patchIndex];
- //cout << "2:" << offspringToRepair[missingIndex] << endl;
- break;
- }
- }
-}
-
-void TSP::copyToNewPopulation(int const * const chromosone, size_t index)
-{
- assert(index < chromosones && "Index out of bounds");
- for(size_t i = 0; i < cities; ++i)
- {
- newPopulation[index][i] = chromosone[i];
- }
-
-}
-
-int * TSP::rouletteSelection(double const * const fitness) const
-{
- double sum = 0;
- /* Calculate sum of all chromosome fitnesses in population */
- for(size_t i = 0; i < chromosones; ++i)
- {
- sum += fitness[i];
- }
-
- /* 0.0 <= random <= sum */
- double random = randomInclusive(sum);
-
- sum = 0;
- /* Go through the population and sum fitnesses from 0 to sum s. When the sum s is greater or equal to r; stop and return the chromosome where you are */
- for(size_t i = 0; i < chromosones; ++i)
- {
- sum += fitness[i];
- if(sum >= random)
- {
- return solutions[i];
- }
- }
- assert(false && "A chromosone should have been picked by now");
- return(NULL);
-}
-
-void TSP::setRandomPath(int * chromosone)
-{
- for(size_t i = 0; i < cities; ++i)
- {
- chromosone[i] = i;
- }
-
- /*
- * Shuffle the chromosone using the Fisher–Yates shuffle.
- */
- for(size_t i = cities-1; i > 0; --i)
- {
- /* 0 <= random <= i */
- int random = (int)randomInclusive(i);
- int temp = chromosone[i];
- chromosone[i] = chromosone[random];
- chromosone[random] = temp;
- }
-}
-
-double TSP::evaluateFitness(int const * const chromosone) const
-{
- return 1/totalDistance(chromosone);
-}
-
-double TSP::totalDistance(int const * const chromosone) const
-{
- double distance = 0;
- /* Calculate the total distance between all cities */
- for(size_t i = 0; i < cities-1; ++i)
- {
- double dx = citiesX[chromosone[i]] - citiesX[chromosone[i+1]];
- double dy = citiesY[chromosone[i]] - citiesY[chromosone[i+1]];
-
- /* The distance between two points is the square root of (dx^2+dy^2) */
- distance += sqrt((pow(dx, 2.0) + pow(dy, 2.0)));
- }
- /* We complete the tour by adding the distance between the last and the first city */
- double dx = citiesX[chromosone[cities-1]] - citiesX[chromosone[0]];
- double dy = citiesY[chromosone[cities-1]] - citiesY[chromosone[0]];
- distance += sqrt((pow(dx, 2.0) + pow(dy, 2.0)));
-
- return distance;
-}
-
-double TSP::randomInclusive(double max)
-{
- /* Generate random number r, 0.0 <= r <= max */
- //return ((double)rand() / (double)RAND_MAX * max);
- return ((double)rand() * max) / (double)RAND_MAX;
-}
-
-double TSP::randomExclusive(double max)
-{
- /* Generate random number r, 0.0 <= r < max */
- //return ((double)rand() / ((double)RAND_MAX + 1) * max);
- return ((double)rand() * max) / ((double)RAND_MAX + 1);
-}
-
-int main(int argc, const char *argv[])
-{
- /* 90% mutation probability, 2% mutation probability */
- TSP *tsp = new TSP(0.9, 0.02);
- size_t generations = 0, generationsWithoutImprovement = 0;
- double bestFitness = -1;
- double initialAverage = tsp->getAverageDistance();
- /* We'll stop when we've gone 10k generations without improvement */
- while(generations < 10000)
- {
- tsp->nextPopulation();
- ++generations;
- double newFitness = tsp->getBestFitness();
- /* The new fitness is higher, the chromosone is better */
- if(newFitness > bestFitness)
- {
- bestFitness = newFitness;
- generationsWithoutImprovement = 0;
- //cout << "Best goal function: " << tsp->getBestFitness() << endl;
- }
- else
- {
- ++generationsWithoutImprovement;
- }
- }
- //cout << "DONE!" << endl;
- cout << "Number of generations: " << generations << endl;
- cout << "Best chromosone info: " << endl;
- cout << "\t-Path: " << tsp->getBestPathString() << endl;
- cout << "\t-Goal function: " << tsp->getBestFitness() << endl;
- cout << "\t-Distance: " << tsp->getLowestTotalDistance() << endl;
- cout << "Average distance: " << tsp->getAverageDistance() << endl;
- cout << "Initial average: " << initialAverage << endl;
- delete tsp;
- return 0;
-}
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/cc_param_defaults.json b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/cc_param_defaults.json
deleted file mode 100644
index 067a26573a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/cc_param_defaults.json
+++ /dev/null
@@ -1 +0,0 @@
-{"max-pipeline-region-insns": {"default": 200, "max": 0, "min": 0}, "ipa-cp-loop-hint-bonus": {"default": 64, "max": 0, "min": 0}, "lim-expensive": {"default": 20, "max": 0, "min": 0}, "uninit-control-dep-attempts": {"default": 1000, "max": 0, "min": 1}, "lto-partitions": {"default": 32, "max": 0, "min": 1}, "max-inline-recursive-depth-auto": {"default": 8, "max": 0, "min": 0}, "max-unroll-times": {"default": 8, "max": 0, "min": 0}, "max-tail-merge-comparisons": {"default": 10, "max": 0, "min": 0}, "early-inlining-insns": {"default": 11, "max": 0, "min": 0}, "prefetch-latency": {"default": 200, "max": 0, "min": 0}, "partial-inlining-entry-probability": {"default": 70, "max": 0, "min": 0}, "integer-share-limit": {"default": 251, "max": 2, "min": 2}, "tm-max-aggregate-size": {"default": 9, "max": 0, "min": 0}, "ira-max-conflict-table-size": {"default": 1000, "max": 0, "min": 0}, "asan-instrument-reads": {"default": 1, "max": 1, "min": 0}, "lto-min-partition": {"default": 1000, "max": 0, "min": 0}, "hot-bb-frequency-fraction": {"default": 1000, "max": 0, "min": 0}, "min-vect-loop-bound": {"default": 1, "max": 0, "min": 1}, "max-crossjump-edges": {"default": 100, "max": 0, "min": 0}, "sms-dfa-history": {"default": 0, "max": 0, "min": 0}, "tracer-max-code-growth": {"default": 100, "max": 0, "min": 0}, "max-pipeline-region-blocks": {"default": 15, "max": 0, "min": 0}, "gcse-after-reload-partial-fraction": {"default": 3, "max": 0, "min": 0}, "asan-stack": {"default": 1, "max": 1, "min": 0}, "asan-memintrin": {"default": 1, "max": 1, "min": 0}, "large-function-insns": {"default": 2700, "max": 0, "min": 0}, "scev-max-expr-size": {"default": 100, "max": 0, "min": 0}, "iv-consider-all-candidates-bound": {"default": 30, "max": 0, "min": 0}, "max-partial-antic-length": {"default": 100, "max": 0, "min": 0}, "prefetch-min-insn-to-mem-ratio": {"default": 3, "max": 0, "min": 0}, "min-crossjump-insns": {"default": 5, "max": 0, "min": 1}, "asan-use-after-return": {"default": 1, "max": 1, "min": 0}, "allow-load-data-races": {"default": 1, "max": 1, "min": 0}, "max-jump-thread-duplication-stmts": {"default": 15, "max": 0, "min": 0}, "tracer-min-branch-probability": {"default": 50, "max": 100, "min": 0}, "l2-cache-size": {"default": 512, "max": 0, "min": 0}, "max-cse-insns": {"default": 1000, "max": 0, "min": 0}, "sched-pressure-algorithm": {"default": 1, "max": 2, "min": 1}, "max-unrolled-insns": {"default": 200, "max": 0, "min": 0}, "ipa-cp-value-list-size": {"default": 8, "max": 0, "min": 0}, "graphite-max-nb-scop-params": {"default": 10, "max": 0, "min": 0}, "max-completely-peel-times": {"default": 16, "max": 0, "min": 0}, "min-inline-recursive-probability": {"default": 10, "max": 0, "min": 0}, "max-stores-to-sink": {"default": 2, "max": 0, "min": 0}, "sink-frequency-threshold": {"default": 75, "max": 100, "min": 0}, "builtin-expect-probability": {"default": 90, "max": 100, "min": 0}, "max-average-unrolled-insns": {"default": 80, "max": 0, "min": 0}, "tracer-min-branch-ratio": {"default": 10, "max": 100, "min": 0}, "inline-unit-growth": {"default": 30, "max": 0, "min": 0}, "max-early-inliner-iterations": {"default": 1, "max": 0, "min": 0}, "hot-bb-count-ws-permille": {"default": 999, "max": 1000, "min": 0}, "max-gcse-memory": {"default": 52428800, "max": 0, "min": 0}, "ggc-min-expand": {"default": 30, "max": 0, "min": 0}, "tree-reassoc-width": {"default": 0, "max": 0, "min": 0}, "max-once-peeled-insns": {"default": 400, "max": 0, "min": 0}, "max-inline-recursive-depth": {"default": 8, "max": 0, "min": 0}, "max-inline-insns-recursive": {"default": 450, "max": 0, "min": 0}, "ira-loop-reserved-regs": {"default": 2, "max": 0, "min": 0}, "align-loop-iterations": {"default": 4, "max": 0, "min": 0}, "gcse-cost-distance-ratio": {"default": 10, "max": 0, "min": 0}, "sched-mem-true-dep-cost": {"default": 1, "max": 0, "min": 0}, "gcse-unrestricted-cost": {"default": 3, "max": 0, "min": 0}, "max-inline-insns-recursive-auto": {"default": 450, "max": 0, "min": 0}, "max-cse-path-length": {"default": 10, "max": 0, "min": 1}, "switch-conversion-max-branch-ratio": {"default": 8, "max": 0, "min": 1}, "max-tracked-strlens": {"default": 1000, "max": 0, "min": 0}, "inline-min-speedup": {"default": 10, "max": 0, "min": 0}, "max-cselib-memory-locations": {"default": 500, "max": 0, "min": 0}, "max-tail-merge-iterations": {"default": 2, "max": 0, "min": 0}, "max-inline-insns-auto": {"default": 40, "max": 0, "min": 0}, "min-insn-to-prefetch-ratio": {"default": 9, "max": 0, "min": 0}, "max-slsr-cand-scan": {"default": 50, "max": 999999, "min": 1}, "min-nondebug-insn-uid": {"default": 0, "max": 0, "min": 1}, "max-sched-region-blocks": {"default": 10, "max": 0, "min": 0}, "vect-max-version-for-alignment-checks": {"default": 6, "max": 0, "min": 0}, "max-vartrack-size": {"default": 50000000, "max": 0, "min": 0}, "loop-max-datarefs-for-datadeps": {"default": 1000, "max": 0, "min": 0}, "asan-instrument-writes": {"default": 1, "max": 1, "min": 0}, "asan-globals": {"default": 1, "max": 1, "min": 0}, "large-function-growth": {"default": 100, "max": 0, "min": 0}, "max-last-value-rtl": {"default": 10000, "max": 0, "min": 0}, "selsched-max-sched-times": {"default": 2, "max": 0, "min": 0}, "sms-max-ii-factor": {"default": 100, "max": 0, "min": 0}, "max-hoist-depth": {"default": 30, "max": 0, "min": 0}, "comdat-sharing-probability": {"default": 20, "max": 0, "min": 0}, "allow-store-data-races": {"default": 1, "max": 1, "min": 0}, "omega-max-vars": {"default": 128, "max": 0, "min": 0}, "iv-max-considered-uses": {"default": 250, "max": 0, "min": 0}, "max-inline-insns-single": {"default": 400, "max": 0, "min": 0}, "simultaneous-prefetches": {"default": 3, "max": 0, "min": 0}, "ipa-max-agg-items": {"default": 16, "max": 0, "min": 0}, "max-peel-times": {"default": 16, "max": 0, "min": 0}, "min-size-for-stack-sharing": {"default": 32, "max": 0, "min": 0}, "ira-max-loops-num": {"default": 100, "max": 0, "min": 0}, "tracer-dynamic-coverage": {"default": 75, "max": 100, "min": 0}, "max-gcse-insertion-ratio": {"default": 20, "max": 0, "min": 0}, "tracer-min-branch-probability-feedback": {"default": 80, "max": 100, "min": 0}, "max-sched-insn-conflict-delay": {"default": 3, "max": 10, "min": 1}, "max-peeled-insns": {"default": 100, "max": 0, "min": 0}, "max-dse-active-local-stores": {"default": 5000, "max": 0, "min": 0}, "max-variable-expansions-in-unroller": {"default": 1, "max": 0, "min": 0}, "max-delay-slot-live-search": {"default": 333, "max": 0, "min": 0}, "min-spec-prob": {"default": 40, "max": 0, "min": 0}, "loop-invariant-max-bbs-in-loop": {"default": 10000, "max": 0, "min": 0}, "selsched-insns-to-rename": {"default": 2, "max": 0, "min": 0}, "max-completely-peel-loop-nest-depth": {"default": 8, "max": 0, "min": 0}, "allow-packed-store-data-races": {"default": 1, "max": 1, "min": 0}, "omega-eliminate-redundant-constraints": {"default": 0, "max": 1, "min": 0}, "omega-max-geqs": {"default": 256, "max": 0, "min": 0}, "l1-cache-line-size": {"default": 32, "max": 0, "min": 0}, "case-values-threshold": {"default": 0, "max": 0, "min": 0}, "max-pending-list-length": {"default": 32, "max": 0, "min": 0}, "sccvn-max-alias-queries-per-access": {"default": 1000, "max": 0, "min": 0}, "max-vartrack-expr-depth": {"default": 12, "max": 0, "min": 0}, "loop-block-tile-size": {"default": 51, "max": 0, "min": 0}, "sms-loop-average-count-threshold": {"default": 0, "max": 0, "min": 0}, "vect-max-peeling-for-alignment": {"default": -1, "max": 64, "min": -1}, "selsched-max-lookahead": {"default": 50, "max": 0, "min": 0}, "omega-max-keys": {"default": 500, "max": 0, "min": 0}, "sccvn-max-scc-size": {"default": 10000, "max": 0, "min": 10}, "predictable-branch-outcome": {"default": 2, "max": 50, "min": 0}, "ssp-buffer-size": {"default": 8, "max": 0, "min": 1}, "max-delay-slot-insn-search": {"default": 100, "max": 0, "min": 0}, "sms-min-sc": {"default": 2, "max": 1, "min": 1}, "lra-max-considered-reload-pseudos": {"default": 500, "max": 0, "min": 0}, "tracer-dynamic-coverage-feedback": {"default": 95, "max": 100, "min": 0}, "omega-max-eqs": {"default": 128, "max": 0, "min": 0}, "max-fields-for-field-sensitive": {"default": 0, "max": 0, "min": 0}, "max-sched-region-insns": {"default": 100, "max": 0, "min": 0}, "large-stack-frame-growth": {"default": 1000, "max": 0, "min": 0}, "omega-max-wild-cards": {"default": 18, "max": 0, "min": 0}, "max-sched-extend-regions-iters": {"default": 0, "max": 0, "min": 0}, "max-unswitch-insns": {"default": 50, "max": 0, "min": 0}, "ipcp-unit-growth": {"default": 10, "max": 0, "min": 0}, "max-unswitch-level": {"default": 3, "max": 0, "min": 0}, "l1-cache-size": {"default": 64, "max": 0, "min": 0}, "max-grow-copy-bb-insns": {"default": 8, "max": 0, "min": 0}, "max-iterations-computation-cost": {"default": 10, "max": 0, "min": 0}, "ipa-cp-array-index-hint-bonus": {"default": 48, "max": 0, "min": 0}, "ggc-min-heapsize": {"default": 4096, "max": 0, "min": 0}, "align-threshold": {"default": 100, "max": 0, "min": 1}, "graphite-max-bbs-per-function": {"default": 100, "max": 0, "min": 0}, "max-vartrack-reverse-op-size": {"default": 50, "max": 0, "min": 0}, "ipa-sra-ptr-growth-factor": {"default": 2, "max": 0, "min": 0}, "max-completely-peeled-insns": {"default": 100, "max": 0, "min": 0}, "ipa-cp-eval-threshold": {"default": 500, "max": 0, "min": 0}, "large-stack-frame": {"default": 256, "max": 0, "min": 0}, "max-modulo-backtrack-attempts": {"default": 40, "max": 0, "min": 0}, "omega-hash-table-size": {"default": 550, "max": 0, "min": 0}, "max-goto-duplication-insns": {"default": 8, "max": 0, "min": 0}, "max-sched-ready-insns": {"default": 100, "max": 0, "min": 0}, "max-iterations-to-track": {"default": 1000, "max": 0, "min": 0}, "scev-max-expr-complexity": {"default": 10, "max": 0, "min": 0}, "cxx-max-namespaces-for-diagnostic-help": {"default": 1000, "max": 0, "min": 0}, "max-reload-search-insns": {"default": 100, "max": 0, "min": 0}, "use-canonical-types": {"default": 1, "max": 1, "min": 0}, "gcse-after-reload-critical-fraction": {"default": 10, "max": 0, "min": 0}, "sched-state-edge-prob-cutoff": {"default": 10, "max": 100, "min": 0}, "sched-spec-prob-cutoff": {"default": 40, "max": 100, "min": 0}, "unlikely-bb-count-fraction": {"default": 20, "max": 10000, "min": 1}, "slp-max-insns-in-bb": {"default": 1000, "max": 0, "min": 0}, "max-peel-branches": {"default": 32, "max": 0, "min": 0}, "large-unit-insns": {"default": 10000, "max": 0, "min": 0}, "iv-always-prune-cand-set-bound": {"default": 10, "max": 0, "min": 0}, "vect-max-version-for-alias-checks": {"default": 10, "max": 0, "min": 0}, "max-predicted-iterations": {"default": 100, "max": 0, "min": 0}, "allow-packed-load-data-races": {"default": 1, "max": 1, "min": 0}}
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/gccflags.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/gccflags.py
deleted file mode 100755
index 1edc5bb8a3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/gccflags.py
+++ /dev/null
@@ -1,412 +0,0 @@
-#!/usr/bin/env python
-import adddeps # fix sys.path
-
-import math
-import argparse
-import ast
-import collections
-import json
-import logging
-import opentuner
-import os
-import random
-import re
-import shutil
-import subprocess
-import sys
-
-from opentuner.resultsdb.models import Result, TuningRun
-from opentuner.search import manipulator
-
-FLAGS_WORKING_CACHE_FILE = 'cc_flags.json'
-PARAMS_DEFAULTS_CACHE_FILE = 'cc_param_defaults.json'
-PARAMS_DEF_PATH = '~/gcc-4.9.0/gcc/params.def'
-PARAMS_WORKING_CACHE_FILE = 'cc_params.json'
-
-log = logging.getLogger('gccflags')
-
-argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
-argparser.add_argument('source', help='source file to compile')
-argparser.add_argument('--compile-template',
- default='{cc} {source} -o {output} -lpthread {flags}',
- help='command to compile {source} into {output} with'
- ' {flags}')
-argparser.add_argument('--compile-limit', type=float, default=30,
- help='kill gcc if it runs more than {default} sec')
-argparser.add_argument('--scaler', type=int, default=4,
- help='by what factor to try increasing parameters')
-argparser.add_argument('--cc', default='g++', help='g++ or gcc')
-argparser.add_argument('--output', default='./tmp.bin',
- help='temporary file for compiler to write to')
-argparser.add_argument('--debug', action='store_true',
- help='on gcc errors try to find minimal set '
- 'of args to reproduce error')
-argparser.add_argument('--force-killall', action='store_true',
- help='killall cc1plus before each collection')
-argparser.add_argument('--memory-limit', default=1024 ** 3, type=int,
- help='memory limit for child process')
-argparser.add_argument('--no-cached-flags', action='store_true',
- help='regenerate the lists of legal flags each time')
-argparser.add_argument('--flags-histogram', action='store_true',
- help='print out a histogram of flags')
-argparser.add_argument('--flag-importance',
- help='Test the importance of different flags from a '
- 'given json file.')
-
-
-class GccFlagsTuner(opentuner.measurement.MeasurementInterface):
- def __init__(self, *pargs, **kwargs):
- super(GccFlagsTuner, self).__init__(program_name=args.source, *pargs,
- **kwargs)
- self.gcc_version = self.extract_gcc_version()
- self.cc_flags = self.extract_working_flags()
- self.cc_param_defaults = self.extract_param_defaults()
- self.cc_params = self.extract_working_params()
-
- # these bugs are hardcoded for now
- # sets of options which causes gcc to barf
- if True:
- # These bugs were for gcc 4.7 on ubuntu
- self.cc_bugs = (['-fipa-matrix-reorg', '-fwhole-program'],
- ['-fno-tree-coalesce-inlined-vars'],
- ['-fno-inline-atomics'],
- ['-ftoplevel-reorder', '-fno-unit-at-a-time'])
- else:
- # Bugs for gcc 4.9 (work in progress, incomplete list)
- self.cc_bugs = (['-ftoplevel-reorder', '-fno-unit-at-a-time'], )
-
- self.result_list = {}
- self.parallel_compile = True
- try:
- os.stat('./tmp')
- except OSError:
- os.mkdir('./tmp')
- self.run_baselines()
-
- def run_baselines(self):
- log.info("baseline perfs -O0=%.4f -O1=%.4f -O2=%.4f -O3=%.4f",
- *[self.run_with_flags(['-O%d' % i], None).time
- for i in range(4)])
-
- def extract_gcc_version(self):
- m = re.search(r'([0-9]+)[.]([0-9]+)[.]([0-9]+)', subprocess.check_output([
- self.args.cc, '--version']))
- if m:
- gcc_version = tuple(map(int, m.group(1, 2, 3)))
- else:
- gcc_version = None
- log.debug('gcc version %s', gcc_version)
- return gcc_version
-
- def extract_working_flags(self):
- """
- Figure out which gcc flags work (don't cause gcc to barf) by running
- each one.
- """
- if os.path.isfile(FLAGS_WORKING_CACHE_FILE) and not args.no_cached_flags:
- # use cached version
- found_cc_flags = json.load(open(FLAGS_WORKING_CACHE_FILE))
- else:
- # extract flags from --help=optimizers
- optimizers, err = subprocess.Popen([self.args.cc, '--help=optimizers'],
- stdout=subprocess.PIPE).communicate()
- found_cc_flags = re.findall(r'^ (-f[a-z0-9-]+) ', optimizers,
- re.MULTILINE)
- log.info('Determining which of %s possible gcc flags work',
- len(found_cc_flags))
- found_cc_flags = filter(self.check_if_flag_works, found_cc_flags)
- json.dump(found_cc_flags, open(FLAGS_WORKING_CACHE_FILE, 'w'))
- return found_cc_flags
-
- def extract_param_defaults(self):
- """
- Get the default, minimum, and maximum for each gcc parameter.
- Requires source code for gcc to be in your home directory.
- This example ships with a cached version so it does not require source.
- """
- if os.path.isfile(PARAMS_DEFAULTS_CACHE_FILE) and not args.no_cached_flags:
- # use cached version
- param_defaults = json.load(open(PARAMS_DEFAULTS_CACHE_FILE))
- else:
- # default values of params need to be extracted from source code,
- # since they are not in --help
- param_defaults = dict()
- params_def = open(os.path.expanduser(PARAMS_DEF_PATH)).read()
- for m in re.finditer(r'DEFPARAM *\((([^")]|"[^"]*")*)\)', params_def):
- param_def_str = (m.group(1)
- # Hacks!!!
- .replace('GGC_MIN_EXPAND_DEFAULT', '30')
- .replace('GGC_MIN_HEAPSIZE_DEFAULT', '4096')
- .replace('50 * 1024 * 1024', '52428800'))
- try:
- name, desc, default, param_min, param_max = ast.literal_eval(
- '[' + param_def_str.split(',', 1)[1] + ']')
- param_defaults[name] = {'default': default,
- 'min': param_min,
- 'max': param_max}
- except:
- log.exception("error with %s", param_def_str)
- json.dump(param_defaults, open(PARAMS_DEFAULTS_CACHE_FILE, 'w'))
- return param_defaults
-
- def extract_working_params(self):
- """
- Figure out which gcc params work (don't cause gcc to barf) by running
- each one to test.
- """
- params, err = subprocess.Popen(
- [self.args.cc, '--help=params'], stdout=subprocess.PIPE).communicate()
- all_params = re.findall(r'^ ([a-z0-9-]+) ', params, re.MULTILINE)
- all_params = sorted(set(all_params) &
- set(self.cc_param_defaults.keys()))
- if os.path.isfile(PARAMS_WORKING_CACHE_FILE) and not args.no_cached_flags:
- # use cached version
- return json.load(open(PARAMS_WORKING_CACHE_FILE))
- else:
- log.info('Determining which of %s possible gcc params work',
- len(all_params))
- working_params = []
- for param in all_params:
- if self.check_if_flag_works('--param={}={}'.format(
- param, self.cc_param_defaults[param]['default'])):
- working_params.append(param)
- json.dump(working_params, open(PARAMS_WORKING_CACHE_FILE, 'w'))
- return working_params
-
- def check_if_flag_works(self, flag, try_inverted=True):
- cmd = args.compile_template.format(source=args.source, output=args.output,
- flags=flag, cc=args.cc)
- compile_result = self.call_program(cmd, limit=args.compile_limit)
- if compile_result['returncode'] != 0:
- log.warning("removing flag %s because it results in compile error", flag)
- return False
- if 'warning: this target' in compile_result['stderr']:
- log.warning("removing flag %s because not supported by target", flag)
- return False
- if 'has been renamed' in compile_result['stderr']:
- log.warning("removing flag %s because renamed", flag)
- return False
- if try_inverted and flag[:2] == '-f':
- if not self.check_if_flag_works(invert_gcc_flag(flag),
- try_inverted=False):
- log.warning("Odd... %s works but %s does not", flag,
- invert_gcc_flag(flag))
- return False
- return True
-
- def manipulator(self):
- m = manipulator.ConfigurationManipulator()
- m.add_parameter(manipulator.IntegerParameter('-O', 0, 3))
- for flag in self.cc_flags:
- m.add_parameter(manipulator.EnumParameter(flag, ['on', 'off', 'default']))
- for param in self.cc_params:
- defaults = self.cc_param_defaults[param]
- if defaults['max'] <= defaults['min']:
- defaults['max'] = float('inf')
- defaults['max'] = min(defaults['max'],
- max(1, defaults['default']) * args.scaler)
- defaults['min'] = max(defaults['min'],
- max(1, defaults['default']) / args.scaler)
-
- if param == 'l1-cache-line-size':
- # gcc requires this to be a power of two or it internal errors
- m.add_parameter(manipulator.PowerOfTwoParameter(param, 4, 256))
- elif defaults['max'] > 128:
- m.add_parameter(manipulator.LogIntegerParameter(
- param, defaults['min'], defaults['max']))
- else:
- m.add_parameter(manipulator.IntegerParameter(
- param, defaults['min'], defaults['max']))
-
- return m
-
- def cfg_to_flags(self, cfg):
- flags = ['-O%d' % cfg['-O']]
- for flag in self.cc_flags:
- if cfg[flag] == 'on':
- flags.append(flag)
- elif cfg[flag] == 'off':
- flags.append(invert_gcc_flag(flag))
-
- for param in self.cc_params:
- flags.append('--param=%s=%d' % (param, cfg[param]))
-
- # workaround sets of flags that trigger compiler crashes/hangs
- for bugset in self.cc_bugs:
- if len(set(bugset) & set(flags)) == len(bugset):
- flags.remove(bugset[-1])
- return flags
-
- def make_command(self, cfg):
- return args.compile_template.format(source=args.source, output=args.output,
- flags=' '.join(self.cfg_to_flags(cfg)),
- cc=args.cc)
-
- def get_tmpdir(self, result_id):
- return './tmp/%d' % result_id
-
- def cleanup(self, result_id):
- tmp_dir = self.get_tmpdir(result_id)
- shutil.rmtree(tmp_dir)
-
- def compile_and_run(self, desired_result, input, limit):
- cfg = desired_result.configuration.data
- compile_result = self.compile(cfg, 0)
- return self.run_precompiled(desired_result, input, limit, compile_result, 0)
-
- compile_results = {'ok': 0, 'timeout': 1, 'error': 2}
-
- def run_precompiled(self, desired_result, input, limit, compile_result,
- result_id):
- if self.args.force_killall:
- os.system('killall -9 cc1plus 2>/dev/null')
- # Make sure compile was successful
- if compile_result == self.compile_results['timeout']:
- return Result(state='TIMEOUT', time=float('inf'))
- elif compile_result == self.compile_results['error']:
- return Result(state='ERROR', time=float('inf'))
-
- tmp_dir = self.get_tmpdir(result_id)
- output_dir = '%s/%s' % (tmp_dir, args.output)
- try:
- run_result = self.call_program([output_dir], limit=limit,
- memory_limit=args.memory_limit)
- except OSError:
- return Result(state='ERROR', time=float('inf'))
-
- if run_result['returncode'] != 0:
- if run_result['timeout']:
- return Result(state='TIMEOUT', time=float('inf'))
- else:
- log.error('program error')
- return Result(state='ERROR', time=float('inf'))
-
- return Result(time=run_result['time'])
-
- def debug_gcc_error(self, flags):
- def fails(subflags):
- cmd = args.compile_template.format(source=args.source, output=args.output,
- flags=' '.join(subflags),
- cc=args.cc)
- compile_result = self.call_program(cmd, limit=args.compile_limit)
- return compile_result['returncode'] != 0
-
- if self.args.debug:
- while len(flags) > 8:
- log.error("compile error with %d flags, diagnosing...", len(flags))
- tmpflags = filter(lambda x: random.choice((True, False)), flags)
- if fails(tmpflags):
- flags = tmpflags
-
- # linear scan
- minimal_flags = []
- for i in xrange(len(flags)):
- tmpflags = minimal_flags + flags[i + 1:]
- if not fails(tmpflags):
- minimal_flags.append(flags[i])
- log.error("compiler crashes/hangs with flags: %s", minimal_flags)
-
- def compile(self, config_data, result_id):
- flags = self.cfg_to_flags(config_data)
- return self.compile_with_flags(flags, result_id)
-
- def compile_with_flags(self, flags, result_id):
- tmp_dir = self.get_tmpdir(result_id)
- try:
- os.stat(tmp_dir)
- except OSError:
- os.mkdir(tmp_dir)
- output_dir = '%s/%s' % (tmp_dir, args.output)
- cmd = args.compile_template.format(source=args.source, output=output_dir,
- flags=' '.join(flags),
- cc=args.cc)
-
- compile_result = self.call_program(cmd, limit=args.compile_limit,
- memory_limit=args.memory_limit)
- if compile_result['returncode'] != 0:
- if compile_result['timeout']:
- log.warning("gcc timeout")
- return self.compile_results['timeout']
- else:
- log.warning("gcc error %s", compile_result['stderr'])
- self.debug_gcc_error(flags)
- return self.compile_results['error']
- return self.compile_results['ok']
-
- def run_with_flags(self, flags, limit):
- return self.run_precompiled(None, None, limit,
- self.compile_with_flags(flags, 0), 0)
-
- def save_final_config(self, configuration):
- """called at the end of tuning"""
- print "Best flags written to gccflags_final_config.{json,cmd}"
- self.manipulator().save_to_file(configuration.data,
- 'gccflags_final_config.json')
- with open('gccflags_final_config.cmd', 'w') as fd:
- fd.write(self.make_command(configuration.data))
-
- def flags_histogram(self, session):
- counter = collections.Counter()
- q = session.query(TuningRun).filter_by(state='COMPLETE')
- total = q.count()
- for tr in q:
- print tr.program.name
- for flag in self.cfg_to_flags(tr.final_config.data):
- counter[flag] += 1.0 / total
- print counter.most_common(20)
-
- def flag_importance(self):
- """
- Test the importance of each flag by measuring the performance with that
- flag removed. Print out a table for paper
- """
- with open(self.args.flag_importance) as fd:
- best_cfg = json.load(fd)
- flags = self.cfg_to_flags(best_cfg)
- counter = collections.Counter()
- baseline_time = self.flags_mean_time(flags)
- for flag in flags[1:]:
- delta_flags = [f for f in flags if f != flag]
- flag_time = self.flags_mean_time(delta_flags)
- impact = max(0.0, flag_time - baseline_time)
- if math.isinf(impact):
- impact = 0.0
- counter[flag] = impact
- print flag, '{:.4f}'.format(impact)
- total_impact = sum(counter.values())
- remaining_impact = total_impact
- print r'\bf Flag & \bf Importance \\\hline'
- for flag, impact in counter.most_common(20):
- print r'{} & {:.1f}\% \\\hline'.format(flag, 100.0 * impact / total_impact)
- remaining_impact -= impact
- print r'{} other flags & {:.1f}% \\\hline'.format(
- len(flags) - 20, 100.0 * remaining_impact / total_impact)
-
- def flags_mean_time(self, flags, trials=10):
- precompiled = self.compile_with_flags(flags, 0)
- total = 0.0
- for _ in xrange(trials):
- total += self.run_precompiled(None, None, None, precompiled, 0).time
- return total / trials
-
- def prefix_hook(self, session):
- if self.args.flags_histogram:
- self.flags_histogram(session)
- sys.exit(0)
- if self.args.flag_importance:
- self.flag_importance()
- sys.exit(0)
-
-
-
-def invert_gcc_flag(flag):
- assert flag[:2] == '-f'
- if flag[2:5] != 'no-':
- return '-fno-' + flag[2:]
- return '-f' + flag[5:]
-
-
-if __name__ == '__main__':
- opentuner.init_logging()
- args = argparser.parse_args()
- GccFlagsTuner.main(args)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/gccflags_minimal.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/gccflags_minimal.py
deleted file mode 100755
index 0363b984a8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/gccflags/gccflags_minimal.py
+++ /dev/null
@@ -1,92 +0,0 @@
-#!/usr/bin/env python
-#
-# Autotune flags to g++ to optimize the performance of apps/raytracer.cpp
-#
-# This is an extremely simplified version meant only for tutorials
-#
-import adddeps # fix sys.path
-
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import EnumParameter
-from opentuner import IntegerParameter
-from opentuner import MeasurementInterface
-from opentuner import Result
-
-GCC_FLAGS = [
- 'align-functions', 'align-jumps', 'align-labels',
- 'align-loops', 'asynchronous-unwind-tables',
- 'branch-count-reg', 'branch-probabilities',
- # ... (176 total)
-]
-
-# (name, min, max)
-GCC_PARAMS = [
- ('early-inlining-insns', 0, 1000),
- ('gcse-cost-distance-ratio', 0, 100),
- ('iv-max-considered-uses', 0, 1000),
- # ... (145 total)
-]
-
-
-class GccFlagsTuner(MeasurementInterface):
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- manipulator.add_parameter(
- IntegerParameter('opt_level', 0, 3))
- for flag in GCC_FLAGS:
- manipulator.add_parameter(
- EnumParameter(flag,
- ['on', 'off', 'default']))
- for param, min, max in GCC_PARAMS:
- manipulator.add_parameter(
- IntegerParameter(param, min, max))
- return manipulator
-
- def compile(self, cfg, id):
- """
- Compile a given configuration in parallel
- """
- gcc_cmd = 'g++ apps/raytracer.cpp -o ./tmp{0}.bin'.format(id)
- gcc_cmd += ' -O{0}'.format(cfg['opt_level'])
- for flag in GCC_FLAGS:
- if cfg[flag] == 'on':
- gcc_cmd += ' -f{0}'.format(flag)
- elif cfg[flag] == 'off':
- gcc_cmd += ' -fno-{0}'.format(flag)
- for param, min, max in GCC_PARAMS:
- gcc_cmd += ' --param {0}={1}'.format(
- param, cfg[param])
- return self.call_program(gcc_cmd)
-
- def run_precompiled(self, desired_result, input, limit, compile_result, id):
- """
- Run a compile_result from compile() sequentially and return performance
- """
- assert compile_result['returncode'] == 0
-
- try:
- run_result = self.call_program('./tmp{0}.bin'.format(id))
- assert run_result['returncode'] == 0
- finally:
- self.call_program('rm ./tmp{0}.bin'.format(id))
-
- return Result(time=run_result['time'])
-
- def compile_and_run(self, desired_result, input, limit):
- """
- Compile and run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
- compile_result = self.compile(cfg, 0)
- return self.run_precompiled(desired_result, input, limit, compile_result, 0)
-
-if __name__ == '__main__':
- argparser = opentuner.default_argparser()
- GccFlagsTuner.main(argparser.parse_args())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/.gitignore b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/.gitignore
deleted file mode 100644
index ebdc2a395f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/.gitignore
+++ /dev/null
@@ -1,2 +0,0 @@
-dump-call-graph
-*.callgraph
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/adddeps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/adddeps.py
deleted file mode 100644
index ede22a8fcd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/adddeps.py
+++ /dev/null
@@ -1,6 +0,0 @@
-# we would prefer a symbolic link, but it does not work on windows
-import os
-target = os.path.join(os.path.dirname(__file__),
- '../../opentuner/utils/adddeps.py')
-execfile(target, dict(__file__=target))
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/bilateral_grid.cpp b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/bilateral_grid.cpp
deleted file mode 100644
index 6f1c97ffb8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/bilateral_grid.cpp
+++ /dev/null
@@ -1,93 +0,0 @@
-#include "Halide.h"
-#include <stdio.h>
-
-using namespace Halide;
-
-int main(int argc, char **argv) {
- // if (argc < 2) {
- // printf("Usage: bilateral_grid <s_sigma>\n");
- // // printf("Spatial sigma is a compile-time parameter, please provide it as an argument.\n"
- // // "(llvm's ptx backend doesn't handle integer mods by non-consts yet)\n");
- // return 0;
- // }
-
- ImageParam input(Float(32), 2);
- float r_sigma = 0.1;
- // int s_sigma = atoi(argv[1]);
- int s_sigma = 4;
- Var x("x"), y("y"), z("z"), c("c");
-
- // Add a boundary condition
- Func clamped("clamped");
- clamped(x, y) = input(clamp(x, 0, input.width()-1),
- clamp(y, 0, input.height()-1));
-
- // Construct the bilateral grid
- RDom r(0, s_sigma, 0, s_sigma);
- Expr val = clamped(x * s_sigma + r.x - s_sigma/2, y * s_sigma + r.y - s_sigma/2);
- val = clamp(val, 0.0f, 1.0f);
- Expr zi = cast<int>(val * (1.0f/r_sigma) + 0.5f);
- Func grid("grid"), histogram("histogram");
- histogram(x, y, zi, c) += select(c == 0, val, 1.0f);
-
- // Introduce a dummy function, so we can schedule the histogram within it
- grid(x, y, z, c) = histogram(x, y, z, c);
-
- // Blur the grid using a five-tap filter
- Func blurx("blurx"), blury("blury"), blurz("blurz");
- blurx(x, y, z, _) = grid(x-2, y, z, _) + grid(x-1, y, z, _)*4 + grid(x, y, z, _)*6 + grid(x+1, y, z, _)*4 + grid(x+2, y, z, _);
- blury(x, y, z, _) = blurx(x, y-2, z, _) + blurx(x, y-1, z, _)*4 + blurx(x, y, z, _)*6 + blurx(x, y+1, z, _)*4 + blurx(x, y+2, z, _);
- blurz(x, y, z, _) = blury(x, y, z-2, _) + blury(x, y, z-1, _)*4 + blury(x, y, z, _)*6 + blury(x, y, z+1, _)*4 + blury(x, y, z+2, _);
-
- // Take trilinear samples to compute the output
- val = clamp(clamped(x, y), 0.0f, 1.0f);
- Expr zv = val * (1.0f/r_sigma);
- zi = cast<int>(zv);
- Expr zf = zv - zi;
- Expr xf = cast<float>(x % s_sigma) / s_sigma;
- Expr yf = cast<float>(y % s_sigma) / s_sigma;
- Expr xi = x/s_sigma;
- Expr yi = y/s_sigma;
- Func interpolated("interpolated");
- interpolated(x, y, _) =
- lerp(lerp(lerp(blurz(xi, yi, zi, _), blurz(xi+1, yi, zi, _), xf),
- lerp(blurz(xi, yi+1, zi, _), blurz(xi+1, yi+1, zi, _), xf), yf),
- lerp(lerp(blurz(xi, yi, zi+1, _), blurz(xi+1, yi, zi+1, _), xf),
- lerp(blurz(xi, yi+1, zi+1, _), blurz(xi+1, yi+1, zi+1, _), xf), yf), zf);
-
- // Normalize
- Func bilateral_grid("bilateral_grid");
- bilateral_grid(x, y) = interpolated(x, y, 0)/interpolated(x, y, 1);
-
- AUTOTUNE_HOOK(bilateral_grid);
-
- char *target = getenv("HL_TARGET");
- if (target && std::string(target) == "ptx") {
-
- // GPU schedule
- grid.compute_root().reorder(z, c, x, y).cuda_tile(x, y, 8, 8);
-
- // Compute the histogram into shared memory before spilling it to global memory
- histogram.store_at(grid, Var("blockidx")).compute_at(grid, Var("threadidx"));
-
- blurx.compute_root().cuda_tile(x, y, z, 16, 16, 1);
- blury.compute_root().cuda_tile(x, y, z, 16, 16, 1);
- blurz.compute_root().cuda_tile(x, y, z, 8, 8, 4);
- bilateral_grid.compute_root().cuda_tile(x, y, s_sigma, s_sigma);
- } else {
-
- // CPU schedule
- grid.compute_root().reorder(c, z, x, y).parallel(y);
- histogram.compute_at(grid, x).unroll(c);
- blurx.compute_root().parallel(z).vectorize(x, 4);
- blury.compute_root().parallel(z).vectorize(x, 4);
- blurz.compute_root().parallel(z).vectorize(x, 4);
- bilateral_grid.compute_root().parallel(y).vectorize(x, 4);
- }
-
- BASELINE_HOOK(bilateral_grid);
-
- //bilateral_grid.compile_to_file("bilateral_grid", r_sigma, input);
-
- return 0;
-}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/bilateral_grid.settings b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/bilateral_grid.settings
deleted file mode 100644
index 7b829b779a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/bilateral_grid.settings
+++ /dev/null
@@ -1,10 +0,0 @@
-{"input_size": "2048, 2048",
- "functions": [
- {"name": "clamped", "vars": ["x", "y"], "calls": []},
- {"name": "histogram", "vars": ["x", "y", "c"], "calls": ["clamped"]},
- {"name": "grid", "vars": ["x", "y", "z", "c"], "calls": ["histogram"]},
- {"name": "blurx", "vars": ["x", "y", "z"], "calls": ["grid"]},
- {"name": "blury", "vars": ["x", "y", "z"], "calls": ["blurx"]},
- {"name": "blurz", "vars": ["x", "y", "z"], "calls": ["blury"]},
- {"name": "interpolated", "vars": ["x", "y"], "calls": ["blurz", "clamped"]},
- {"name": "bilateral_grid", "vars": ["x", "y"], "calls": ["interpolated"]}]}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/halide_blur.cpp b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/halide_blur.cpp
deleted file mode 100644
index 7a38dd45fd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/halide_blur.cpp
+++ /dev/null
@@ -1,32 +0,0 @@
-#include <Halide.h>
-using namespace Halide;
-
-#define AUTOTUNE_HOOK(x)
-#define BASELINE_HOOK(x)
-
-int main(int argc, char **argv) {
-
- ImageParam in_img(UInt(16), 2);
- Func blur_x("blur_x"), blur_y("blur_y");
- Var x("x"), y("y"), xi("xi"), yi("yi");
-
- Func input;
- input(x,y) = in_img(clamp(x, 1, in_img.width()-1),
- clamp(y, 1, in_img.height())-1);
-
- // The algorithm
- blur_x(x, y) = (input(x, y) + input(x+1, y) + input(x+2, y))/3;
- blur_y(x, y) = (blur_x(x, y) + blur_x(x, y+1) + blur_x(x, y+2))/3;
-
- AUTOTUNE_HOOK(blur_y);
-
- // How to schedule it
- blur_y.split(y, y, yi, 8).parallel(y).vectorize(x, 8);
- blur_x.store_at(blur_y, y).compute_at(blur_y, yi).vectorize(x, 8);
-
- BASELINE_HOOK(blur_y);
-
- blur_y.compile_to_file("halide_blur", in_img);
-
- return 0;
-}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/halide_blur.settings b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/halide_blur.settings
deleted file mode 100644
index af0deeac34..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/halide_blur.settings
+++ /dev/null
@@ -1,4 +0,0 @@
-{"input_size": "4096, 4096",
- "functions": [
- {"name": "blur_x", "vars": ["x", "y"], "calls": []},
- {"name": "blur_y", "vars": ["x", "y"], "calls": ["blur_x"]}]}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simple.cpp b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simple.cpp
deleted file mode 100644
index 74d141721d..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simple.cpp
+++ /dev/null
@@ -1,208 +0,0 @@
-#include "Halide.h"
-
-#define AUTOTUNE_HOOK(x)
-#define BASELINE_HOOK(x)
-
-using namespace Halide;
-
-#include <iostream>
-#include <limits>
-
-#include <sys/time.h>
-
-using std::vector;
-
-double now() {
- struct timeval tv;
- gettimeofday(&tv, NULL);
- static bool first_call = true;
- static time_t first_sec = 0;
- if (first_call) {
- first_call = false;
- first_sec = tv.tv_sec;
- }
- assert(tv.tv_sec >= first_sec);
- return (tv.tv_sec - first_sec) + (tv.tv_usec / 1000000.0);
-}
-
-int main(int argc, char **argv) {
- ImageParam input(Float(32), 3, "input");
-
- const unsigned int levels = 3;
-
- Func downsampled[levels];
- Func downx[levels];
- Func interpolated[levels];
- Func upsampled[levels];
- Func upsampledx[levels];
- Var x("x"), y("y"), c("c");
-
- downsampled[0] = Func("downsampled");
- downx[0] = Func("downx");
- interpolated[0] = Func("interpolated");
- upsampled[0] = Func("upsampled");
- upsampledx[0] = Func("upsampledx");
-
- Func clamped("clamped");
- clamped(x, y, c) = input(clamp(x, 0, input.width()-1), clamp(y, 0, input.height()-1), c);
-
- // This triggers a bug in llvm 3.3 (3.2 and trunk are fine), so we
- // rewrite it in a way that doesn't trigger the bug. The rewritten
- // form assumes the input alpha is zero or one.
- // downsampled[0](x, y, c) = select(c < 3, clamped(x, y, c) * clamped(x, y, 3), clamped(x, y, 3));
- downsampled[0](x, y, c) = clamped(x, y, c) * clamped(x, y, 3);
-
- for (unsigned int l = 1; l < levels; ++l) {
- downx[l] = Func("downx");
- downsampled[l] = Func("downsampled");
- downx[l](x, y, c) = (downsampled[l-1](x*2-1, y, c) +
- 2.0f * downsampled[l-1](x*2, y, c) +
- downsampled[l-1](x*2+1, y, c)) * 0.25f;
- downsampled[l](x, y, c) = (downx[l](x, y*2-1, c) +
- 2.0f * downx[l](x, y*2, c) +
- downx[l](x, y*2+1, c)) * 0.25f;
- }
- interpolated[levels-1] = Func("interpolated");
- interpolated[levels-1](x, y, c) = downsampled[levels-1](x, y, c);
- for (unsigned int l = levels-2; l < levels; --l) {
- upsampledx[l] = Func("upsampledx");
- upsampled[l] = Func("upsampled");
- interpolated[l] = Func("interpolated");
- upsampledx[l](x, y, c) = select((x % 2) == 0,
- interpolated[l+1](x/2, y, c),
- 0.5f * (interpolated[l+1](x/2, y, c) +
- interpolated[l+1](x/2+1, y, c)));
- upsampled[l](x, y, c) = select((y % 2) == 0,
- upsampledx[l](x, y/2, c),
- 0.5f * (upsampledx[l](x, y/2, c) +
- upsampledx[l](x, y/2+1, c)));
- interpolated[l](x, y, c) = downsampled[l](x, y, c) + (1.0f - downsampled[l](x, y, 3)) * upsampled[l](x, y, c);
- }
-
- Func normalize("normalize");
- normalize(x, y, c) = interpolated[0](x, y, c) / interpolated[0](x, y, 3);
-
- Func final("final");
- final(x, y, c) = normalize(x, y, c);
-
- AUTOTUNE_HOOK(final);
-
- int sched;
- char *target = getenv("HL_TARGET");
- if (target && std::string(target) == "ptx") {
- sched = 4;
- } else {
- sched = 2;
- }
-
- switch (sched) {
- case 0:
- {
- //std::cout << "Flat schedule." << std::endl;
- for (unsigned int l = 0; l < levels; ++l) {
- downsampled[l].compute_root();
- interpolated[l].compute_root();
- }
- final.compute_root();
- break;
- }
- case 1:
- {
- //std::cout << "Flat schedule with vectorization." << std::endl;
- for (unsigned int l = 0; l < levels; ++l) {
- downsampled[l].compute_root().vectorize(x,4);
- interpolated[l].compute_root().vectorize(x,4);
- }
- final.compute_root();
- break;
- }
- case 2:
- {
- Var xi, yi;
- //std::cout << "Flat schedule with parallelization + vectorization." << std::endl;
- clamped.compute_root().parallel(y).reorder(c, x, y).reorder_storage(c, x, y).vectorize(c, 4);
- for (unsigned int l = 1; l < levels-1; ++l) {
- if (l > 0) downsampled[l].compute_root().parallel(y).reorder(c, x, y).reorder_storage(c, x, y).vectorize(c, 4);
- interpolated[l].compute_root().parallel(y).reorder(c, x, y).reorder_storage(c, x, y).vectorize(c, 4);
- interpolated[l].unroll(x, 2).unroll(y, 2);
- }
- final.reorder(c, x, y).bound(c, 0, 3).parallel(y);
- final.tile(x, y, xi, yi, 2, 2).unroll(xi).unroll(yi);
- break;
- }
- case 3:
- {
- //std::cout << "Flat schedule with vectorization sometimes." << std::endl;
- for (unsigned int l = 0; l < levels; ++l) {
- if (l + 4 < levels) {
- Var yo,yi;
- downsampled[l].compute_root().vectorize(x,4);
- interpolated[l].compute_root().vectorize(x,4);
- } else {
- downsampled[l].compute_root();
- interpolated[l].compute_root();
- }
- }
- final.compute_root();
- break;
- }
- case 4:
- {
- //std::cout << "GPU schedule." << std::endl;
-
- // Some gpus don't have enough memory to process the entire
- // image, so we process the image in tiles.
- Var yo, yi, xo, xi;
- final.reorder(c, x, y).bound(c, 0, 3).vectorize(x, 4);
- final.tile(x, y, xo, yo, xi, yi, input.width()/4, input.height()/4);
- normalize.compute_at(final, xo).reorder(c, x, y).cuda_tile(x, y, 16, 16).unroll(c);
-
- // Start from level 1 to save memory - level zero will be computed on demand
- for (unsigned int l = 1; l < levels; ++l) {
- int tile_size = 32 >> l;
- if (tile_size < 1) tile_size = 1;
- if (tile_size > 16) tile_size = 16;
- downsampled[l].compute_root().cuda_tile(x, y, c, tile_size, tile_size, 4);
- interpolated[l].compute_at(final, xo).cuda_tile(x, y, c, tile_size, tile_size, 4);
- }
-
- break;
- }
- default:
- assert(0 && "No schedule with this number.");
- }
-
- BASELINE_HOOK(final);
-
-#if 0
- // JIT compile the pipeline eagerly, so we don't interfere with timing
- final.compile_jit();
-
- // Image<float> in_png = load<float>(argv[1]);
- Image<float> out(2048, 2048, 3);
- // assert(in_png.channels() == 4);
- // input.set(in_png);
- final.infer_input_bounds(out);
-
- std::cout << "Running... " << std::endl;
- double min = std::numeric_limits<double>::infinity();
- const unsigned int iters = 20;
-
- for (unsigned int x = 0; x < iters; ++x) {
- double before = now();
- final.realize(out);
- double after = now();
- double amt = after - before;
-
- std::cout << " " << amt * 1000 << std::endl;
- if (amt < min) min = amt;
-
- }
- std::cout << " took " << min * 1000 << " msec." << std::endl;
-
- // vector<Argument> args;
- // args.push_back(input);
- // final.compile_to_assembly("test.s", args);
- // save(out, argv[2]);
-#endif
-}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simple.settings b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simple.settings
deleted file mode 100644
index cffd184ed6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simple.settings
+++ /dev/null
@@ -1,185 +0,0 @@
-{
- "functions": [
- {
- "calls": [],
- "name": "clamped",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "clamped"
- ],
- "name": "downsampled",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downx$2"
- ],
- "name": "downsampled$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downx$3"
- ],
- "name": "downsampled$3",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled"
- ],
- "name": "downx$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$2"
- ],
- "name": "downx$3",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$3"
- ],
- "name": "interpolated$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$2",
- "upsampled$2"
- ],
- "name": "interpolated$3",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled",
- "upsampled$3"
- ],
- "name": "interpolated$4",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$4"
- ],
- "name": "normalize",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "upsampledx$2"
- ],
- "name": "upsampled$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "upsampledx$3"
- ],
- "name": "upsampled$3",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$2"
- ],
- "name": "upsampledx$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$3"
- ],
- "name": "upsampledx$3",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "normalize"
- ],
- "name": "final",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- }
- ],
- "input_size": "2048, 2048, 3"
-}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simplest.cpp b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simplest.cpp
deleted file mode 100644
index cf57055836..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simplest.cpp
+++ /dev/null
@@ -1,208 +0,0 @@
-#include "Halide.h"
-
-#define AUTOTUNE_HOOK(x)
-#define BASELINE_HOOK(x)
-
-using namespace Halide;
-
-#include <iostream>
-#include <limits>
-
-#include <sys/time.h>
-
-using std::vector;
-
-double now() {
- struct timeval tv;
- gettimeofday(&tv, NULL);
- static bool first_call = true;
- static time_t first_sec = 0;
- if (first_call) {
- first_call = false;
- first_sec = tv.tv_sec;
- }
- assert(tv.tv_sec >= first_sec);
- return (tv.tv_sec - first_sec) + (tv.tv_usec / 1000000.0);
-}
-
-int main(int argc, char **argv) {
- ImageParam input(Float(32), 3, "input");
-
- const unsigned int levels = 2;
-
- Func downsampled[levels];
- Func downx[levels];
- Func interpolated[levels];
- Func upsampled[levels];
- Func upsampledx[levels];
- Var x("x"), y("y"), c("c");
-
- downsampled[0] = Func("downsampled");
- downx[0] = Func("downx");
- interpolated[0] = Func("interpolated");
- upsampled[0] = Func("upsampled");
- upsampledx[0] = Func("upsampledx");
-
- Func clamped("clamped");
- clamped(x, y, c) = input(clamp(x, 0, input.width()-1), clamp(y, 0, input.height()-1), c);
-
- // This triggers a bug in llvm 3.3 (3.2 and trunk are fine), so we
- // rewrite it in a way that doesn't trigger the bug. The rewritten
- // form assumes the input alpha is zero or one.
- // downsampled[0](x, y, c) = select(c < 3, clamped(x, y, c) * clamped(x, y, 3), clamped(x, y, 3));
- downsampled[0](x, y, c) = clamped(x, y, c) * clamped(x, y, 3);
-
- for (unsigned int l = 1; l < levels; ++l) {
- downx[l] = Func("downx");
- downsampled[l] = Func("downsampled");
- downx[l](x, y, c) = (downsampled[l-1](x*2-1, y, c) +
- 2.0f * downsampled[l-1](x*2, y, c) +
- downsampled[l-1](x*2+1, y, c)) * 0.25f;
- downsampled[l](x, y, c) = (downx[l](x, y*2-1, c) +
- 2.0f * downx[l](x, y*2, c) +
- downx[l](x, y*2+1, c)) * 0.25f;
- }
- interpolated[levels-1] = Func("interpolated");
- interpolated[levels-1](x, y, c) = downsampled[levels-1](x, y, c);
- for (unsigned int l = levels-2; l < levels; --l) {
- upsampledx[l] = Func("upsampledx");
- upsampled[l] = Func("upsampled");
- interpolated[l] = Func("interpolated");
- upsampledx[l](x, y, c) = select((x % 2) == 0,
- interpolated[l+1](x/2, y, c),
- 0.5f * (interpolated[l+1](x/2, y, c) +
- interpolated[l+1](x/2+1, y, c)));
- upsampled[l](x, y, c) = select((y % 2) == 0,
- upsampledx[l](x, y/2, c),
- 0.5f * (upsampledx[l](x, y/2, c) +
- upsampledx[l](x, y/2+1, c)));
- interpolated[l](x, y, c) = downsampled[l](x, y, c) + (1.0f - downsampled[l](x, y, 3)) * upsampled[l](x, y, c);
- }
-
- Func normalize("normalize");
- normalize(x, y, c) = interpolated[0](x, y, c) / interpolated[0](x, y, 3);
-
- Func final("final");
- final(x, y, c) = normalize(x, y, c);
-
- AUTOTUNE_HOOK(final);
-
- int sched;
- char *target = getenv("HL_TARGET");
- if (target && std::string(target) == "ptx") {
- sched = 4;
- } else {
- sched = 2;
- }
-
- switch (sched) {
- case 0:
- {
- //std::cout << "Flat schedule." << std::endl;
- for (unsigned int l = 0; l < levels; ++l) {
- downsampled[l].compute_root();
- interpolated[l].compute_root();
- }
- final.compute_root();
- break;
- }
- case 1:
- {
- //std::cout << "Flat schedule with vectorization." << std::endl;
- for (unsigned int l = 0; l < levels; ++l) {
- downsampled[l].compute_root().vectorize(x,4);
- interpolated[l].compute_root().vectorize(x,4);
- }
- final.compute_root();
- break;
- }
- case 2:
- {
- Var xi, yi;
- //std::cout << "Flat schedule with parallelization + vectorization." << std::endl;
- clamped.compute_root().parallel(y).reorder(c, x, y).reorder_storage(c, x, y).vectorize(c, 4);
- for (unsigned int l = 1; l < levels-1; ++l) {
- if (l > 0) downsampled[l].compute_root().parallel(y).reorder(c, x, y).reorder_storage(c, x, y).vectorize(c, 4);
- interpolated[l].compute_root().parallel(y).reorder(c, x, y).reorder_storage(c, x, y).vectorize(c, 4);
- interpolated[l].unroll(x, 2).unroll(y, 2);
- }
- final.reorder(c, x, y).bound(c, 0, 3).parallel(y);
- final.tile(x, y, xi, yi, 2, 2).unroll(xi).unroll(yi);
- break;
- }
- case 3:
- {
- //std::cout << "Flat schedule with vectorization sometimes." << std::endl;
- for (unsigned int l = 0; l < levels; ++l) {
- if (l + 4 < levels) {
- Var yo,yi;
- downsampled[l].compute_root().vectorize(x,4);
- interpolated[l].compute_root().vectorize(x,4);
- } else {
- downsampled[l].compute_root();
- interpolated[l].compute_root();
- }
- }
- final.compute_root();
- break;
- }
- case 4:
- {
- //std::cout << "GPU schedule." << std::endl;
-
- // Some gpus don't have enough memory to process the entire
- // image, so we process the image in tiles.
- Var yo, yi, xo, xi;
- final.reorder(c, x, y).bound(c, 0, 3).vectorize(x, 4);
- final.tile(x, y, xo, yo, xi, yi, input.width()/4, input.height()/4);
- normalize.compute_at(final, xo).reorder(c, x, y).cuda_tile(x, y, 16, 16).unroll(c);
-
- // Start from level 1 to save memory - level zero will be computed on demand
- for (unsigned int l = 1; l < levels; ++l) {
- int tile_size = 32 >> l;
- if (tile_size < 1) tile_size = 1;
- if (tile_size > 16) tile_size = 16;
- downsampled[l].compute_root().cuda_tile(x, y, c, tile_size, tile_size, 4);
- interpolated[l].compute_at(final, xo).cuda_tile(x, y, c, tile_size, tile_size, 4);
- }
-
- break;
- }
- default:
- assert(0 && "No schedule with this number.");
- }
-
- BASELINE_HOOK(final);
-
-#if 0
- // JIT compile the pipeline eagerly, so we don't interfere with timing
- final.compile_jit();
-
- // Image<float> in_png = load<float>(argv[1]);
- Image<float> out(2048, 2048, 3);
- // assert(in_png.channels() == 4);
- // input.set(in_png);
- final.infer_input_bounds(out);
-
- std::cout << "Running... " << std::endl;
- double min = std::numeric_limits<double>::infinity();
- const unsigned int iters = 20;
-
- for (unsigned int x = 0; x < iters; ++x) {
- double before = now();
- final.realize(out);
- double after = now();
- double amt = after - before;
-
- std::cout << " " << amt * 1000 << std::endl;
- if (amt < min) min = amt;
-
- }
- std::cout << " took " << min * 1000 << " msec." << std::endl;
-
- // vector<Argument> args;
- // args.push_back(input);
- // final.compile_to_assembly("test.s", args);
- // save(out, argv[2]);
-#endif
-}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simplest.settings b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simplest.settings
deleted file mode 100644
index 5f22d20f5f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate-simplest.settings
+++ /dev/null
@@ -1,124 +0,0 @@
-{
- "functions": [
- {
- "calls": [],
- "name": "clamped",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "clamped"
- ],
- "name": "downsampled",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downx$2"
- ],
- "name": "downsampled$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled"
- ],
- "name": "downx$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$2"
- ],
- "name": "interpolated$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled",
- "upsampled$2"
- ],
- "name": "interpolated$3",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$3"
- ],
- "name": "normalize",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "upsampledx$2"
- ],
- "name": "upsampled$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$2"
- ],
- "name": "upsampledx$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "normalize"
- ],
- "name": "final",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- }
- ],
- "input_size": "2048, 2048, 3"
-}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate.cpp b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate.cpp
deleted file mode 100644
index 1ca4ae5bd3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate.cpp
+++ /dev/null
@@ -1,208 +0,0 @@
-#include "Halide.h"
-
-#define AUTOTUNE_HOOK(x)
-#define BASELINE_HOOK(x)
-
-using namespace Halide;
-
-#include <iostream>
-#include <limits>
-
-#include <sys/time.h>
-
-using std::vector;
-
-double now() {
- struct timeval tv;
- gettimeofday(&tv, NULL);
- static bool first_call = true;
- static time_t first_sec = 0;
- if (first_call) {
- first_call = false;
- first_sec = tv.tv_sec;
- }
- assert(tv.tv_sec >= first_sec);
- return (tv.tv_sec - first_sec) + (tv.tv_usec / 1000000.0);
-}
-
-int main(int argc, char **argv) {
- ImageParam input(Float(32), 3, "input");
-
- const unsigned int levels = 10;
-
- Func downsampled[levels];
- Func downx[levels];
- Func interpolated[levels];
- Func upsampled[levels];
- Func upsampledx[levels];
- Var x("x"), y("y"), c("c");
-
- downsampled[0] = Func("downsampled");
- downx[0] = Func("downx");
- interpolated[0] = Func("interpolated");
- upsampled[0] = Func("upsampled");
- upsampledx[0] = Func("upsampledx");
-
- Func clamped("clamped");
- clamped(x, y, c) = input(clamp(x, 0, input.width()-1), clamp(y, 0, input.height()-1), c);
-
- // This triggers a bug in llvm 3.3 (3.2 and trunk are fine), so we
- // rewrite it in a way that doesn't trigger the bug. The rewritten
- // form assumes the input alpha is zero or one.
- // downsampled[0](x, y, c) = select(c < 3, clamped(x, y, c) * clamped(x, y, 3), clamped(x, y, 3));
- downsampled[0](x, y, c) = clamped(x, y, c) * clamped(x, y, 3);
-
- for (unsigned int l = 1; l < levels; ++l) {
- downx[l] = Func("downx");
- downsampled[l] = Func("downsampled");
- downx[l](x, y, c) = (downsampled[l-1](x*2-1, y, c) +
- 2.0f * downsampled[l-1](x*2, y, c) +
- downsampled[l-1](x*2+1, y, c)) * 0.25f;
- downsampled[l](x, y, c) = (downx[l](x, y*2-1, c) +
- 2.0f * downx[l](x, y*2, c) +
- downx[l](x, y*2+1, c)) * 0.25f;
- }
- interpolated[levels-1] = Func("interpolated");
- interpolated[levels-1](x, y, c) = downsampled[levels-1](x, y, c);
- for (unsigned int l = levels-2; l < levels; --l) {
- upsampledx[l] = Func("upsampledx");
- upsampled[l] = Func("upsampled");
- interpolated[l] = Func("interpolated");
- upsampledx[l](x, y, c) = select((x % 2) == 0,
- interpolated[l+1](x/2, y, c),
- 0.5f * (interpolated[l+1](x/2, y, c) +
- interpolated[l+1](x/2+1, y, c)));
- upsampled[l](x, y, c) = select((y % 2) == 0,
- upsampledx[l](x, y/2, c),
- 0.5f * (upsampledx[l](x, y/2, c) +
- upsampledx[l](x, y/2+1, c)));
- interpolated[l](x, y, c) = downsampled[l](x, y, c) + (1.0f - downsampled[l](x, y, 3)) * upsampled[l](x, y, c);
- }
-
- Func normalize("normalize");
- normalize(x, y, c) = interpolated[0](x, y, c) / interpolated[0](x, y, 3);
-
- Func final("final");
- final(x, y, c) = normalize(x, y, c);
-
- AUTOTUNE_HOOK(final);
-
- int sched;
- char *target = getenv("HL_TARGET");
- if (target && std::string(target) == "ptx") {
- sched = 4;
- } else {
- sched = 2;
- }
-
- switch (sched) {
- case 0:
- {
- //std::cout << "Flat schedule." << std::endl;
- for (unsigned int l = 0; l < levels; ++l) {
- downsampled[l].compute_root();
- interpolated[l].compute_root();
- }
- final.compute_root();
- break;
- }
- case 1:
- {
- //std::cout << "Flat schedule with vectorization." << std::endl;
- for (unsigned int l = 0; l < levels; ++l) {
- downsampled[l].compute_root().vectorize(x,4);
- interpolated[l].compute_root().vectorize(x,4);
- }
- final.compute_root();
- break;
- }
- case 2:
- {
- Var xi, yi;
- //std::cout << "Flat schedule with parallelization + vectorization." << std::endl;
- clamped.compute_root().parallel(y).reorder(c, x, y).reorder_storage(c, x, y).vectorize(c, 4);
- for (unsigned int l = 1; l < levels-1; ++l) {
- if (l > 0) downsampled[l].compute_root().parallel(y).reorder(c, x, y).reorder_storage(c, x, y).vectorize(c, 4);
- interpolated[l].compute_root().parallel(y).reorder(c, x, y).reorder_storage(c, x, y).vectorize(c, 4);
- interpolated[l].unroll(x, 2).unroll(y, 2);
- }
- final.reorder(c, x, y).bound(c, 0, 3).parallel(y);
- final.tile(x, y, xi, yi, 2, 2).unroll(xi).unroll(yi);
- break;
- }
- case 3:
- {
- //std::cout << "Flat schedule with vectorization sometimes." << std::endl;
- for (unsigned int l = 0; l < levels; ++l) {
- if (l + 4 < levels) {
- Var yo,yi;
- downsampled[l].compute_root().vectorize(x,4);
- interpolated[l].compute_root().vectorize(x,4);
- } else {
- downsampled[l].compute_root();
- interpolated[l].compute_root();
- }
- }
- final.compute_root();
- break;
- }
- case 4:
- {
- //std::cout << "GPU schedule." << std::endl;
-
- // Some gpus don't have enough memory to process the entire
- // image, so we process the image in tiles.
- Var yo, yi, xo, xi;
- final.reorder(c, x, y).bound(c, 0, 3).vectorize(x, 4);
- final.tile(x, y, xo, yo, xi, yi, input.width()/4, input.height()/4);
- normalize.compute_at(final, xo).reorder(c, x, y).cuda_tile(x, y, 16, 16).unroll(c);
-
- // Start from level 1 to save memory - level zero will be computed on demand
- for (unsigned int l = 1; l < levels; ++l) {
- int tile_size = 32 >> l;
- if (tile_size < 1) tile_size = 1;
- if (tile_size > 16) tile_size = 16;
- downsampled[l].compute_root().cuda_tile(x, y, c, tile_size, tile_size, 4);
- interpolated[l].compute_at(final, xo).cuda_tile(x, y, c, tile_size, tile_size, 4);
- }
-
- break;
- }
- default:
- assert(0 && "No schedule with this number.");
- }
-
- BASELINE_HOOK(final);
-
-#if 0
- // JIT compile the pipeline eagerly, so we don't interfere with timing
- final.compile_jit();
-
- // Image<float> in_png = load<float>(argv[1]);
- Image<float> out(2048, 2048, 3);
- // assert(in_png.channels() == 4);
- // input.set(in_png);
- final.infer_input_bounds(out);
-
- std::cout << "Running... " << std::endl;
- double min = std::numeric_limits<double>::infinity();
- const unsigned int iters = 20;
-
- for (unsigned int x = 0; x < iters; ++x) {
- double before = now();
- final.realize(out);
- double after = now();
- double amt = after - before;
-
- std::cout << " " << amt * 1000 << std::endl;
- if (amt < min) min = amt;
-
- }
- std::cout << " took " << min * 1000 << " msec." << std::endl;
-
- // vector<Argument> args;
- // args.push_back(input);
- // final.compile_to_assembly("test.s", args);
- // save(out, argv[2]);
-#endif
-}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate.settings b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate.settings
deleted file mode 100644
index 3a51d80626..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/interpolate.settings
+++ /dev/null
@@ -1,612 +0,0 @@
-{
- "functions": [
- {
- "calls": [],
- "name": "clamped",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "clamped"
- ],
- "name": "downsampled",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downx$10"
- ],
- "name": "downsampled$10",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downx$2"
- ],
- "name": "downsampled$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downx$3"
- ],
- "name": "downsampled$3",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downx$4"
- ],
- "name": "downsampled$4",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downx$5"
- ],
- "name": "downsampled$5",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downx$6"
- ],
- "name": "downsampled$6",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downx$7"
- ],
- "name": "downsampled$7",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downx$8"
- ],
- "name": "downsampled$8",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downx$9"
- ],
- "name": "downsampled$9",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$9"
- ],
- "name": "downx$10",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled"
- ],
- "name": "downx$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$2"
- ],
- "name": "downx$3",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$3"
- ],
- "name": "downx$4",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$4"
- ],
- "name": "downx$5",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$5"
- ],
- "name": "downx$6",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$6"
- ],
- "name": "downx$7",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$7"
- ],
- "name": "downx$8",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$8"
- ],
- "name": "downx$9",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$2",
- "upsampled$9"
- ],
- "name": "interpolated$10",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled",
- "upsampled$10"
- ],
- "name": "interpolated$11",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$10"
- ],
- "name": "interpolated$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$9",
- "upsampled$2"
- ],
- "name": "interpolated$3",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$8",
- "upsampled$3"
- ],
- "name": "interpolated$4",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$7",
- "upsampled$4"
- ],
- "name": "interpolated$5",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$6",
- "upsampled$5"
- ],
- "name": "interpolated$6",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$5",
- "upsampled$6"
- ],
- "name": "interpolated$7",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$4",
- "upsampled$7"
- ],
- "name": "interpolated$8",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "downsampled$3",
- "upsampled$8"
- ],
- "name": "interpolated$9",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$11"
- ],
- "name": "normalize",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "upsampledx$10"
- ],
- "name": "upsampled$10",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "upsampledx$2"
- ],
- "name": "upsampled$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "upsampledx$3"
- ],
- "name": "upsampled$3",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "upsampledx$4"
- ],
- "name": "upsampled$4",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "upsampledx$5"
- ],
- "name": "upsampled$5",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "upsampledx$6"
- ],
- "name": "upsampled$6",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "upsampledx$7"
- ],
- "name": "upsampled$7",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "upsampledx$8"
- ],
- "name": "upsampled$8",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "upsampledx$9"
- ],
- "name": "upsampled$9",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$10"
- ],
- "name": "upsampledx$10",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$2"
- ],
- "name": "upsampledx$2",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$3"
- ],
- "name": "upsampledx$3",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$4"
- ],
- "name": "upsampledx$4",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$5"
- ],
- "name": "upsampledx$5",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$6"
- ],
- "name": "upsampledx$6",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$7"
- ],
- "name": "upsampledx$7",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$8"
- ],
- "name": "upsampledx$8",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "interpolated$9"
- ],
- "name": "upsampledx$9",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- },
- {
- "calls": [
- "normalize"
- ],
- "name": "final",
- "update_calls": [],
- "vars": [
- "x",
- "y",
- "c"
- ]
- }
- ],
- "input_size": "1024, 1024, 3"
-}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/wavelet.cpp b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/wavelet.cpp
deleted file mode 100644
index e2cb008790..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/wavelet.cpp
+++ /dev/null
@@ -1,93 +0,0 @@
-#include "Halide.h"
-
-#define AUTOTUNE_HOOK(x)
-#define BASELINE_HOOK(x)
-
-using namespace Halide;
-
-Var x("x"), y("y"), c("c");
-
-Func haar_x(Func in) {
- Func out;
- out(x, y, c) = select(c == 0,
- (in(2*x, y) + in(2*x+1, y)),
- (in(2*x, y) - in(2*x+1, y)))/2;
- out.unroll(c, 2);
- return out;
-}
-
-Func inverse_haar_x(Func in) {
- Func out;
- out(x, y) = select(x%2 == 0,
- in(x/2, y, 0) + in(x/2, y, 1),
- in(x/2, y, 0) - in(x/2, y, 1));
- out.unroll(x, 2);
- return out;
-}
-
-
-const float D0 = 0.4829629131445341f;
-const float D1 = 0.83651630373780772f;
-const float D2 = 0.22414386804201339f;
-const float D3 = -0.12940952255126034f;
-
-/*
-const float D0 = 0.34150635f;
-const float D1 = 0.59150635f;
-const float D2 = 0.15849365f;
-const float D3 = -0.1830127f;
-*/
-
-Func daubechies_x(Func in) {
- Func out;
- out(x, y, c) = select(c == 0,
- D0*in(2*x-1, y) + D1*in(2*x, y) + D2*in(2*x+1, y) + D3*in(2*x+2, y),
- D3*in(2*x-1, y) - D2*in(2*x, y) + D1*in(2*x+1, y) - D0*in(2*x+2, y));
- //out.unroll(c, 2);
- return out;
-}
-
-Func inverse_daubechies_x(Func in) {
- Func out("inv_daub_x");
- out(x, y) = select(x%2 == 0,
- D2*in(x/2, y, 0) + D1*in(x/2, y, 1) + D0*in(x/2+1, y, 0) + D3*in(x/2+1, y, 1),
- D3*in(x/2, y, 0) - D0*in(x/2, y, 1) + D1*in(x/2+1, y, 0) - D2*in(x/2+1, y, 1));
- //out.unroll(x, 2);
- return out;
-}
-
-int main(int argc, char **argv) {
-
- ImageParam image(Float(32), 2);
- ImageParam wavelet(Float(32), 3);
-
- // Add a boundary condition for daubechies
- Func clamped;
- clamped(x, y) = image(clamp(x, 0, image.width()-1),
- clamp(y, 0, image.height()-1));
- Func wavelet_clamped("wavelet_clamped");
- wavelet_clamped(x, y, c) = wavelet(clamp(x, 0, wavelet.width()-1),
- clamp(y, 0, wavelet.height()-1), c);
-
-
- // Func inv_haar_x = inverse_haar_x(wavelet_clamped);
- // inv_haar_x.compile_to_file("inverse_haar_x", wavelet);
-
- // Func for_haar_x = haar_x(clamped);
- // for_haar_x.compile_to_file("haar_x", image);
-
- Func inv_daub_x = inverse_daubechies_x(wavelet_clamped);
- //inv_daub_x.compile_to_file("inverse_daubechies_x", wavelet);
-
- AUTOTUNE_HOOK(inv_daub_x);
- inv_daub_x.unroll(x, 2).vectorize(x, 8).parallel(y);
- BASELINE_HOOK(inv_daub_x);
-
- // Func for_daub_x = daubechies_x(clamped);
- //for_daub_x.compile_to_file("daubechies_x", image);
-
- return 0;
-}
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/wavelet.settings b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/wavelet.settings
deleted file mode 100644
index 6fbb5c4006..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/apps/wavelet.settings
+++ /dev/null
@@ -1,4 +0,0 @@
-{"input_size": "2048, 2048",
- "functions": [
- {"name": "wavelet_clamped", "vars": ["x", "y", "c"], "calls": []},
- {"name": "inv_daub_x", "vars": ["x", "y"], "calls": ["wavelet_clamped"]}]}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/halidetuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/halidetuner.py
deleted file mode 100755
index 08e6732575..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/halidetuner.py
+++ /dev/null
@@ -1,682 +0,0 @@
-#!/usr/bin/env python
-# coding: utf-8
-#
-# Example of synthesizing Halide schedules using OpenTuner. This program
-# expects a compiled version of Halide to exist at ~/Halide or at the location
-# specified by --halide-dir.
-#
-# Halide programs must be modified by:
-# 1) Inserting AUTOTUNE_HOOK(Func) directly after the algorithm definition
-# in main()
-# 2) Creating a settings file that describes the functions and variables
-# (see apps/halide_blur.settings for an example)
-#
-# Halide can be found here: https://github.com/halide/Halide
-#
-
-import adddeps # fix sys.path
-
-import argparse
-import collections
-import hashlib
-import json
-import logging
-import math
-import os
-import re
-import subprocess
-import tempfile
-import textwrap
-from cStringIO import StringIO
-from fn import _
-from pprint import pprint
-
-import opentuner
-from opentuner.search.manipulator import ConfigurationManipulator
-from opentuner.search.manipulator import PowerOfTwoParameter
-from opentuner.search.manipulator import PermutationParameter
-from opentuner.search.manipulator import BooleanParameter
-from opentuner.search.manipulator import ScheduleParameter
-
-
-COMPILE_CMD = (
- '{args.cxx} "{cpp}" -o "{bin}" -I "{args.halide_dir}/include" '
- '"{args.halide_dir}/bin/$BUILD_PREFIX/libHalide.a" -ldl -lcurses -lpthread {args.cxxflags} '
- '-DAUTOTUNE_N="{args.input_size}" -DAUTOTUNE_TRIALS={args.trials} '
- '-DAUTOTUNE_LIMIT={limit} -fno-rtti')
-
-log = logging.getLogger('halide')
-
-parser = argparse.ArgumentParser(parents=opentuner.argparsers())
-parser.add_argument('source', help='Halide source file annotated with '
- 'AUTOTUNE_HOOK')
-parser.add_argument('--halide-dir', default=os.path.expanduser('~/Halide'),
- help='Installation directory for Halide')
-parser.add_argument('--input-size',
- help='Input size to test with')
-parser.add_argument('--trials', default=3, type=int,
- help='Number of times to test each schedule')
-parser.add_argument('--nesting', default=2, type=int,
- help='Maximum depth for generated loops')
-parser.add_argument('--max-split-factor', default=8, type=int,
- help='The largest value a single split() can add')
-parser.add_argument('--compile-command', default=COMPILE_CMD,
- help='How to compile generated C++ code')
-parser.add_argument('--cxx', default='c++',
- help='C++ compiler to use (e.g., g++ or clang++)')
-parser.add_argument('--cxxflags', default='',
- help='Extra flags to the C++ compiler')
-parser.add_argument('--tmp-dir',
- default=('/run/shm' if os.access('/run/shm', os.W_OK)
- else '/tmp'),
- help='Where to store generated tests')
-parser.add_argument('--settings-file',
- help='Override location of json encoded settings')
-parser.add_argument('--debug-error',
- help='Stop on errors matching a given string')
-parser.add_argument('--limit', type=float, default=30,
- help='Kill compile + runs taking too long (seconds)')
-parser.add_argument('--memory-limit', type=int, default=1024 ** 3,
- help='Set memory ulimit on unix based systems')
-parser.add_argument('--enable-unroll', action='store_true',
- help='Enable .unroll(...) generation')
-parser.add_argument('--enable-store-at', action='store_true',
- help='Never generate .store_at(...)')
-parser.add_argument('--gated-store-reorder', action='store_true',
- help='Only reorder storage if a special parameter is given')
-group = parser.add_mutually_exclusive_group()
-group.add_argument('--random-test', action='store_true',
- help='Generate a random configuration and run it')
-group.add_argument('--random-source', action='store_true',
- help='Generate a random configuration and print source ')
-group.add_argument('--make-settings-file', action='store_true',
- help='Create a skeleton settings file from call graph')
-
-
-# class HalideRandomConfig(opentuner.search.technique.SearchTechnique):
-# def desired_configuration(self):
-# '''
-# inject random configs with no compute_at() calls to kickstart the search process
-# '''
-# cfg = self.manipulator.random()
-# for k in cfg.keys():
-# if re.match('.*_compute_level', k):
-# cfg[k] = LoopLevel.INLINE
-# return cfg
-#
-# technique.register(bandittechniques.AUCBanditMetaTechnique([
-# HalideRandomConfig(),
-# differentialevolution.DifferentialEvolutionAlt(),
-# evolutionarytechniques.UniformGreedyMutation(),
-# evolutionarytechniques.NormalGreedyMutation(mutation_rate=0.3),
-# ], name = "HalideMetaTechnique"))
-
-
-class HalideTuner(opentuner.measurement.MeasurementInterface):
- def __init__(self, args):
- # args.technique = ['HalideMetaTechnique']
- super(HalideTuner, self).__init__(args, program_name=args.source)
- timing_prefix = open(os.path.join(os.path.dirname(__file__),
- 'timing_prefix.h')).read()
- self.template = timing_prefix + open(args.source).read()
- self.min_collection_cost = float('inf')
- if not args.settings_file:
- args.settings_file = os.path.splitext(args.source)[0] + '.settings'
- if not args.make_settings_file:
- with open(args.settings_file) as fd:
- self.settings = json.load(fd)
- self.post_dominators = post_dominators(self.settings)
- if not args.input_size:
- args.input_size = self.settings['input_size']
- else:
- self.settings = None
- self.post_dominators = None
- args.input_size = '1, 1'
- # set "program_version" based on hash of halidetuner.py, program source
- h = hashlib.md5()
- #with open(__file__) as src:
- # h.update(src.read())
- with open(args.source) as src:
- h.update(src.read())
- self._version = h.hexdigest()
-
- def compute_order_parameter(self, func):
- name = func['name']
- schedule_vars = []
- schedule_deps = dict()
- for var in func['vars']:
- schedule_vars.append((var, 0))
- for i in xrange(1, self.args.nesting):
- schedule_vars.append((var, i))
- schedule_deps[(var, i - 1)] = [(var, i)]
- return ScheduleParameter('{0}_compute_order'.format(name), schedule_vars,
- schedule_deps)
-
- def manipulator(self):
- """
- The definition of the manipulator is meant to mimic the Halide::Schedule
- data structure and defines the configuration space to search
- """
- manipulator = HalideConfigurationManipulator(self)
- manipulator.add_parameter(HalideComputeAtScheduleParameter(
- 'schedule', self.args, self.settings['functions'],
- self.post_dominators))
- for func in self.settings['functions']:
- name = func['name']
- manipulator.add_parameter(PermutationParameter(
- '{0}_store_order'.format(name), func['vars']))
- manipulator.add_parameter(
- BooleanParameter('{0}_store_order_enabled'.format(name)))
- manipulator.add_parameter(self.compute_order_parameter(func))
- for var in func['vars']:
- manipulator.add_parameter(PowerOfTwoParameter(
- '{0}_vectorize'.format(name), 1, self.args.max_split_factor))
- manipulator.add_parameter(PowerOfTwoParameter(
- '{0}_unroll'.format(name), 1, self.args.max_split_factor))
- manipulator.add_parameter(BooleanParameter(
- '{0}_parallel'.format(name)))
- for nesting in xrange(1, self.args.nesting):
- manipulator.add_parameter(PowerOfTwoParameter(
- '{0}_splitfactor_{1}_{2}'.format(name, nesting, var),
- 1, self.args.max_split_factor))
-
- return manipulator
-
- def cfg_to_schedule(self, cfg):
- """
- Produce a Halide schedule from a configuration dictionary
- """
- o = StringIO()
- cnt = 0
- temp_vars = list()
- schedule = ComputeAtStoreAtParser(cfg['schedule'], self.post_dominators)
- compute_at = schedule.compute_at
- store_at = schedule.store_at
-
- # build list of all used variable names
- var_names = dict()
- var_name_order = dict()
- for func in self.settings['functions']:
- name = func['name']
- compute_order = cfg['{0}_compute_order'.format(name)]
- for var in func['vars']:
- var_names[(name, var, 0)] = var
- for nesting in xrange(1, self.args.nesting):
- split_factor = cfg.get('{0}_splitfactor_{1}_{2}'.format(
- name, nesting, var), 0)
- if split_factor > 1 and (name, var, nesting - 1) in var_names:
- var_names[(name, var, nesting)] = '_{var}{cnt}'.format(
- func=name, var=var, nesting=nesting, cnt=cnt)
- temp_vars.append(var_names[(name, var, nesting)])
- cnt += 1
- var_name_order[name] = [var_names[(name, v, n)] for v, n in compute_order
- if (name, v, n) in var_names]
-
- # set a schedule for each function
- for func in self.settings['functions']:
- name = func['name']
- inner_var_name = var_name_order[name][-1] # innermost variable in the reordered list for this func
- vectorize = cfg['{0}_vectorize'.format(name)]
- if self.args.enable_unroll:
- unroll = cfg['{0}_unroll'.format(name)]
- else:
- unroll = 1
-
- print >> o, 'Halide::Func(funcs["%s"])' % name
-
- for var in func['vars']:
- # handle all splits
- for nesting in xrange(1, self.args.nesting):
- split_factor = cfg.get('{0}_splitfactor_{1}_{2}'.format(
- name, nesting, var), 0)
- if split_factor <= 1:
- break
-
- for nesting2 in xrange(nesting + 1, self.args.nesting):
- split_factor2 = cfg.get('{0}_splitfactor_{1}_{2}'.format(
- name, nesting2, var), 0)
- if split_factor2 <= 1:
- break
- split_factor *= split_factor2
- var_name = var_names[(name, var, nesting)]
- last_var_name = var_names[(name, var, nesting - 1)]
-
- # apply unroll, vectorize factors to all surrounding splits iff we're the innermost var
- if var_name == inner_var_name:
- split_factor *= unroll
- split_factor *= vectorize
-
- print >> o, '.split({0}, {0}, {1}, {2})'.format(
- last_var_name, var_name, split_factor)
-
- # drop unused variables and truncate (Halide supports only 10 reorders)
- if len(var_name_order[name]) > 1:
- print >> o, '.reorder({0})'.format(
- ', '.join(reversed(var_name_order[name][:10])))
-
- # reorder_storage
- store_order_enabled = cfg['{0}_store_order_enabled'.format(name)]
- if store_order_enabled or not self.args.gated_store_reorder:
- store_order = cfg['{0}_store_order'.format(name)]
- if len(store_order) > 1:
- print >> o, '.reorder_storage({0})'.format(', '.join(store_order))
-
- if unroll > 1:
- # apply unrolling to innermost var
- print >> o, '.unroll({0}, {1})'.format(
- var_name_order[name][-1], unroll * vectorize)
-
- if vectorize > 1:
- # apply vectorization to innermost var
- print >> o, '.vectorize({0}, {1})'.format(
- var_name_order[name][-1], vectorize)
-
- # compute_at(not root)
- if (compute_at[name] is not None and
- len(var_name_order[compute_at[name][0]]) >= compute_at[name][1]):
- at_func, at_idx = compute_at[name]
- try:
- at_var = var_name_order[at_func][-at_idx]
- print >> o, '.compute_at(Halide::Func(funcs["{0}"]), {1})'.format(at_func, at_var)
- if not self.args.enable_store_at:
- pass # disabled
- elif store_at[name] is None:
- print >> o, '.store_root()'
- elif store_at[name] != compute_at[name]:
- at_func, at_idx = store_at[name]
- at_var = var_name_order[at_func][-at_idx]
- print >> o, '.store_at(Halide::Func(funcs["{0}"]), {1})'.format(at_func, at_var)
- except IndexError:
- # this is expected when at_idx is too large
- # TODO: implement a cleaner fix
- pass
- # compute_root
- else:
- parallel = cfg['{0}_parallel'.format(name)]
- if parallel:
- # only apply parallelism to outermost var of root funcs
- print >> o, '.parallel({0})'.format(var_name_order[name][0])
- print >> o, '.compute_root()'
-
- print >> o, ';'
-
- if temp_vars:
- return 'Halide::Var {0};\n{1}'.format(
- ', '.join(temp_vars), o.getvalue())
- else:
- return o.getvalue()
-
- def schedule_to_source(self, schedule):
- """
- Generate a temporary Halide cpp file with schedule inserted
- """
-
- def repl_autotune_hook(match):
- tmpl = '''
- {
- std::map<std::string, Halide::Internal::Function> funcs = Halide::Internal::find_transitive_calls((%(func)s).function());
-
- %(sched)s
-
- _autotune_timing_stub(%(func)s);
- }'''
- return tmpl % {"sched": schedule.replace('\n', '\n '), "func": match.group(1)}
-
- source = re.sub(r'\n\s*AUTOTUNE_HOOK\(\s*([a-zA-Z0-9_]+)\s*\)',
- repl_autotune_hook, self.template)
- return source
-
- def run_schedule(self, schedule, limit):
- """
- Generate a temporary Halide cpp file with schedule inserted and run it
- with our timing harness found in timing_prefix.h.
- """
- return self.run_source(self.schedule_to_source(schedule), limit)
-
- def run_baseline(self):
- """
- Generate a temporary Halide cpp file with schedule inserted and run it
- with our timing harness found in timing_prefix.h.
- """
-
- def repl_autotune_hook(match):
- return '\n\n_autotune_timing_stub(%s);' % match.group(1)
-
- source = re.sub(r'\n\s*BASELINE_HOOK\(\s*([a-zA-Z0-9_]+)\s*\)',
- repl_autotune_hook, self.template)
- return self.run_source(source)
-
- def run_source(self, source, limit=0, extra_args=''):
- cmd = ''
- with tempfile.NamedTemporaryFile(suffix='.cpp', prefix='halide',
- dir=self.args.tmp_dir) as cppfile:
- cppfile.write(source)
- cppfile.flush()
- # binfile = os.path.splitext(cppfile.name)[0] + '.bin'
- # binfile = '/tmp/halide.bin'
- binfile = ''
- with tempfile.NamedTemporaryFile(suffix='.bin', prefix='halide',
- dir=self.args.tmp_dir, delete=False) as binfiletmp:
-
- binfile = binfiletmp.name # unique temp file to allow multiple concurrent tuner runs
- assert(binfile)
- cmd = self.args.compile_command.format(
- cpp=cppfile.name, bin=binfile, args=self.args,
- limit=math.ceil(limit) if limit < float('inf') else 0)
- cmd += ' ' + extra_args
- compile_result = self.call_program(cmd, limit=self.args.limit,
- memory_limit=self.args.memory_limit)
- if compile_result['returncode'] != 0:
- log.error('compile failed: %s', compile_result)
- return None
-
- try:
- result = self.call_program(binfile,
- limit=self.args.limit,
- memory_limit=self.args.memory_limit)
- stdout = result['stdout']
- stderr = result['stderr']
- returncode = result['returncode']
-
- if result['timeout']:
- log.info('compiler timeout %d (%.2f+%.0f cost)', self.args.limit,
- compile_result['time'], self.args.limit)
- return float('inf')
- elif returncode == 142 or returncode == -14:
- log.info('program timeout %d (%.2f+%.2f cost)', math.ceil(limit),
- compile_result['time'], result['time'])
- return None
- elif returncode != 0:
- log.error('invalid schedule (returncode=%d): %s', returncode,
- stderr.strip())
- with tempfile.NamedTemporaryFile(suffix='.cpp', prefix='halide-error',
- dir=self.args.tmp_dir, delete=False) as errfile:
- errfile.write(source)
- log.error('failed schedule logged to %s.\ncompile as `%s`.', errfile.name, cmd)
- if self.args.debug_error is not None and (
- self.args.debug_error in stderr
- or self.args.debug_error == ""):
- self.debug_schedule('/tmp/halideerror.cpp', source)
- return None
- else:
- try:
- time = json.loads(stdout)['time']
- except:
- log.exception('error parsing output: %s', result)
- return None
- log.info('success: %.4f (collection cost %.2f + %.2f)',
- time, compile_result['time'], result['time'])
- self.min_collection_cost = min(
- self.min_collection_cost, result['time'])
- return time
- finally:
- os.unlink(binfile)
-
- def run_cfg(self, cfg, limit=0):
- try:
- schedule = self.cfg_to_schedule(cfg)
- except:
- log.exception('error generating schedule')
- return None
- return self.run_schedule(schedule, limit)
-
- def run(self, desired_result, input, limit):
- time = self.run_cfg(desired_result.configuration.data, limit)
- if time is not None:
- return opentuner.resultsdb.models.Result(time=time)
- else:
- return opentuner.resultsdb.models.Result(state='ERROR',
- time=float('inf'))
-
- def save_final_config(self, configuration):
- """called at the end of tuning"""
- print 'Final Configuration:'
- print self.cfg_to_schedule(configuration.data)
-
- def debug_log_schedule(self, filename, source):
- open(filename, 'w').write(source)
- print 'offending schedule written to {0}'.format(filename)
-
- def debug_schedule(self, filename, source):
- self.debug_log_schedule(filename, source)
- raw_input('press ENTER to continue')
-
- def make_settings_file(self):
- dump_call_graph_dir = os.path.join(os.path.dirname(__file__),
- 'dump-call-graph')
- if not os.path.isdir(dump_call_graph_dir):
- subprocess.check_call(['git', 'clone',
- 'http://github.com/halide/dump-call-graph.git'])
- assert os.path.isdir(dump_call_graph_dir)
-
- dump_call_graph_cpp = os.path.join(dump_call_graph_dir, 'DumpCallGraph.cpp')
- callgraph_file = self.args.settings_file + '.callgraph'
-
- def repl_autotune_hook(match):
- return r'''dump_call_graph("%s", %s);
- printf("{\"time\": 0}\n");
- exit(0);''' % (callgraph_file, match.group(1))
-
- source = re.sub(r'\n\s*AUTOTUNE_HOOK\(\s*([a-zA-Z0-9_]+)\s*\)',
- repl_autotune_hook, self.template)
- # TODO: BUG! - this only works correctly if given an absolute path to the
- # program (or explicit settings file). Otherwise it generates the callgraph
- # in a tmp dir somewhere and fails to find it in a local path here.
- source = open(dump_call_graph_cpp).read() + source
- self.run_source(source, extra_args='-I{0}'.format(dump_call_graph_dir))
- callgraph = json.load(open(callgraph_file))
- settings = {'input_size': '1024, 1024', 'functions': callgraph}
- json.dump(settings, open(self.args.settings_file, 'w'), sort_keys=True,
- indent=2)
- print textwrap.dedent('''
-
- {0} has been generated based on call graph of program.
-
- This file likely needs some manual tweaks in order to work correctly.
- The input size should be changed to have the right number of dimensions.
- Any naming differences between variable names and function names must
- be applied manually. Some temporary variables not in the source code
- need to be manually removed.
-
- '''.format(self.args.settings_file))
-
-
-class ComputeAtStoreAtParser(object):
- """
- A recursive descent parser to force proper loop nesting, and enforce post
- dominator scheduling constraints
-
- For each function input will have tokens like:
- ('foo', 's') = store_at location for foo
- ('foo', '2'), ('foo', '1') = opening the loop nests for foo,
- the inner 2 variables
- ('foo', 'c') = the computation of foo, and closing all loop nests
-
- The order of these tokens define a loop nest tree which we reconstruct
- """
-
- def __init__(self, tokens, post_dominators):
- self.tokens = list(tokens) # input, processed back to front
- self.post_dominators = post_dominators
- self.compute_at = dict()
- self.store_at = dict()
- self.process_root()
-
- def process_root(self):
- old_len = len(self.tokens)
- out = []
- while self.tokens:
- if self.tokens[-1][1] == 's':
- # store at root
- self.store_at[self.tokens[-1][0]] = None
- out.append(self.tokens.pop())
- else:
- self.process_loopnest(out, [])
- self.tokens = list(reversed(out))
- assert old_len == len(self.tokens)
-
- def process_loopnest(self, out, stack):
- func, idx = self.tokens[-1]
- out.append(self.tokens.pop())
- if idx != 'c':
- raise Exception('Invalid schedule')
-
- self.compute_at[func] = None
- for targ_func, targ_idx in reversed(stack):
- if targ_func in self.post_dominators[func]:
- self.compute_at[func] = (targ_func, targ_idx)
- break
-
- close_tokens = [(f, i) for f, i in self.tokens if f == func and i != 's']
- while close_tokens:
- if self.tokens[-1] == close_tokens[-1]:
- # proper nesting
- close_tokens.pop()
- out.append(self.tokens.pop())
- elif self.tokens[-1][1] == 'c':
- self.process_loopnest(out, stack + close_tokens[-1:])
- elif self.tokens[-1][1] == 's':
- # self.tokens[-1] is computed at this level
- if func in self.post_dominators[self.tokens[-1][0]]:
- self.store_at[self.tokens[-1][0]] = close_tokens[-1]
- else:
- self.store_at[self.tokens[-1][0]] = None
- out.append(self.tokens.pop())
- else:
- # improper nesting, just close the loop and search/delete close_tokens
- out.extend(reversed(close_tokens))
- self.tokens = [x for x in self.tokens if x not in close_tokens]
- break
-
-
-class HalideConfigurationManipulator(ConfigurationManipulator):
- def __init__(self, halide_tuner):
- super(HalideConfigurationManipulator, self).__init__()
- self.halide_tuner = halide_tuner
-
- def hash_config(self, config):
- """
- Multiple configs can lead to the same schedule, so we provide a custom
- hash function that hashes the resulting schedule instead of the raw config.
- This will lead to fewer duplicate tests.
- """
- self.normalize(config)
- try:
- schedule = self.halide_tuner.cfg_to_schedule(config)
- return hashlib.sha256(schedule).hexdigest()
- except:
- log.warning('error hashing config', exc_info=True)
- return super(HalideConfigurationManipulator, self).hash_config(config)
-
-
-class HalideComputeAtScheduleParameter(ScheduleParameter):
- def __init__(self, name, args, functions, post_dominators):
- """
- Custom ScheduleParameter that normalizes using ComputeAtStoreAtParser
- """
- super(HalideComputeAtScheduleParameter, self).__init__(
- name, *self.gen_nodes_deps(args, functions))
- self.post_dominators = post_dominators
-
- def gen_nodes_deps(self, args, functions):
- """
- Compute the list of nodes and point-to-point deps to provide to base class
- """
- nodes = list()
- deps = collections.defaultdict(list)
- for func in functions:
- last = None
- for idx in reversed(['c'] + # 'c' = compute location (and close loops)
- range(1, len(func['vars']) * args.nesting + 1) +
- ['s']): # 's' = storage location
- name = (func['name'], idx)
- if last is not None:
- # variables must go in order
- deps[last].append(name)
- last = name
- nodes.append(name)
- if idx == 'c':
- # computes must follow call graph order
- for callee in func['calls']:
- deps[(callee, 'c')].append(name)
- return nodes, deps
-
- def normalize(self, cfg):
- """
- First enforce basic point-to-point deps (in base class), then call
- ComputeAtStoreAtParser to normalize schedule.
- """
- super(HalideComputeAtScheduleParameter, self).normalize(cfg)
- cfg[self.name] = ComputeAtStoreAtParser(cfg[self.name],
- self.post_dominators).tokens
-
-
-def post_dominators(settings):
- """
- Compute post dominator tree using textbook iterative algorithm for the
- call graph defined in settings
- """
- functions = [f['name'] for f in settings['functions']]
- calls = dict([(f['name'], set(f['calls'])) for f in settings['functions']])
- inverse_calls = collections.defaultdict(set)
- for k, callees in calls.items():
- for v in callees:
- inverse_calls[v].add(k)
- dom = {functions[-1]: set([functions[-1]])}
- for f in functions[:-1]:
- dom[f] = set(functions)
- change = True
- while change:
- change = False
- for f in functions[:-1]:
- old = dom[f]
- dom[f] = set([f]) | reduce(
- _ & _, [dom[c] for c in inverse_calls[f]], set(functions))
- if old != dom[f]:
- change = True
- return dom
-
-
-def random_test(args):
- """
- Generate and run a random schedule
- """
-
- opentuner.tuningrunmain.init_logging()
- m = HalideTuner(args)
- cfg = m.manipulator().random()
- pprint(cfg)
- print
- schedule = m.cfg_to_schedule(cfg)
- print schedule
- print
- print 'Schedule', m.run_schedule(schedule, 30)
- print 'Baseline', m.run_baseline()
-
-
-def random_source(args):
- """
- Dump the source code of a random schedule
- """
- opentuner.tuningrunmain.init_logging()
- m = HalideTuner(args)
- cfg = m.manipulator().random()
- schedule = m.cfg_to_schedule(cfg)
- source = m.schedule_to_source(schedule)
- print source
-
-
-def main(args):
- if args.random_test:
- random_test(args)
- elif args.random_source:
- random_source(args)
- elif args.make_settings_file:
- opentuner.tuningrunmain.init_logging()
- HalideTuner(args).make_settings_file()
- else:
- HalideTuner.main(args)
-
-
-if __name__ == '__main__':
- main(parser.parse_args())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/timing_prefix.h b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/timing_prefix.h
deleted file mode 100644
index d8bbc5f57b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/halide/timing_prefix.h
+++ /dev/null
@@ -1,100 +0,0 @@
-#include <Halide.h>
-#include <stdio.h>
-#include <sys/time.h>
-#include <unistd.h>
-
-#include <map>
-#include <string>
-
-// How many times to run (and take min)
-// #define AUTOTUNE_TRIALS 3
-
-// Limit in seconds to try running for (0 = no limit)
-// #define AUTOTUNE_LIMIT 0
-
-// Size to run with
-// #define AUTOTUNE_N 1024, 1024
-
-inline void _autotune_timing_stub(Halide::Func& func) {
- func.compile_jit();
-
- // TODO: this assumes scalar/non-Tuple outputs - should generalize to a Realization
- std::vector<Halide::Type> out_types = func.output_types();
- std::vector<buffer_t> out_raw_bufs;
- std::vector<Halide::Buffer> out_bufs;
-
- for (int i = 0; i < out_types.size(); i++) {
- // Use the Buffer constructor as a helper to set up the buffer_t,
- // but then throw away its allocation which we don't really want.
- Halide::Buffer bufinit(out_types[i], AUTOTUNE_N);
- out_raw_bufs.push_back(*bufinit.raw_buffer());
- out_raw_bufs[i].host = NULL;
- // TODO: free the host pointer?!
- out_bufs.push_back(Halide::Buffer(out_types[i], &out_raw_bufs[i]));
- assert(out_bufs[i].host_ptr() == NULL); // make sure we don't have an allocation
- }
- Halide::Realization output(out_bufs);
- func.infer_input_bounds(output);
- // assert(output[0].host_ptr()); // for now, the API doesn't seem to allocate outputs
-
- // TODO: this should go into Func::infer_input_bounds(Realization)
- for (int i = 0; i < output.size(); i++) {
- assert(!output[i].host_ptr()); // for now, the API doesn't seem to allocate outputs
- buffer_t buf = *output[i].raw_buffer();
-
- // Figure out how much memory to allocate for this buffer
- size_t min_idx = 0, max_idx = 0;
- for (int d = 0; d < 4; d++) {
- if (buf.stride[d] > 0) {
- min_idx += buf.min[d] * buf.stride[d];
- max_idx += (buf.min[d] + buf.extent[d] - 1) * buf.stride[d];
- } else {
- max_idx += buf.min[d] * buf.stride[d];
- min_idx += (buf.min[d] + buf.extent[d] - 1) * buf.stride[d];
- }
- }
- size_t total_size = (max_idx - min_idx);
- while (total_size & 0x1f) total_size++;
-
- // Allocate enough memory with the right dimensionality.
- Halide::Buffer buffer(output[i].type(), total_size,
- buf.extent[1] > 0 ? 1 : 0,
- buf.extent[2] > 0 ? 1 : 0,
- buf.extent[3] > 0 ? 1 : 0);
-
- // Rewrite the buffer fields to match the ones returned
- for (int d = 0; d < 4; d++) {
- buffer.raw_buffer()->min[d] = buf.min[d];
- buffer.raw_buffer()->stride[d] = buf.stride[d];
- buffer.raw_buffer()->extent[d] = buf.extent[d];
- }
-
- output[i] = buffer;
- }
-
- timeval t1, t2;
- double rv = 0;
- const unsigned int timeout = AUTOTUNE_LIMIT;
- alarm(timeout);
- for (int i = 0; i < AUTOTUNE_TRIALS; i++) {
- gettimeofday(&t1, NULL);
- func.realize(output);
- gettimeofday(&t2, NULL);
- alarm(0); // disable alarm
- double t = (t2.tv_sec - t1.tv_sec) + (t2.tv_usec - t1.tv_usec)/1000000.0;
- if(i == 0 || t < rv)
- rv = t;
- }
- printf("{\"time\": %.10f}\n", rv);
- exit(0);
-}
-
-
-#ifndef AUTOTUNE_HOOK
-#define AUTOTUNE_HOOK(x)
-#endif
-
-#ifndef BASELINE_HOOK
-#define BASELINE_HOOK(x)
-#endif
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/hpl/HPL.dat.mako b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/hpl/HPL.dat.mako
deleted file mode 100644
index 93354a2292..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/hpl/HPL.dat.mako
+++ /dev/null
@@ -1,31 +0,0 @@
-HPLinpack benchmark input file
-Innovative Computing Laboratory, University of Tennessee
-HPL.out output file name (if any)
-0 device out (6=stdout,7=stderr,file)
-1 # of problems sizes (N)
-${size} Ns
-1 # of NBs
-${blocksize} NBs
-${row_or_colmajor_pmapping} PMAP process mapping (0=Row-,1=Column-major)
-1 # of process grids (P x Q)
-2 Ps PxQ must equal nprocs
-2 Qs
-16.0 threshold
-1 # of panel fact
-${pfact} PFACTs (0=left, 1=Crout, 2=Right)
-1 # of recursive stopping criterium
-${nbmin} NBMINs (>= 1)
-1 # of panels in recursion
-${ndiv} NDIVs
-1 # of recursive panel fact.
-${rfact} RFACTs (0=left, 1=Crout, 2=Right)
-1 # of broadcast
-${bcast} BCASTs (0=1rg,1=1rM,2=2rg,3=2rM,4=Lng,5=LnM)
-1 # of lookahead depth
-${depth} DEPTHs (>=0)
-${swap} SWAP (0=bin-exch,1=long,2=mix)
-${swapping_threshold} swapping threshold (default had 64)
-${L1_transposed} L1 in (0=transposed,1=no-transposed) form
-${U_transposed} U in (0=transposed,1=no-transposed) form
-1 Equilibration (0=no,1=yes)
-${mem_alignment} memory alignment in double (> 0) (4,8,16)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/hpl/adddeps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/hpl/adddeps.py
deleted file mode 100644
index ede22a8fcd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/hpl/adddeps.py
+++ /dev/null
@@ -1,6 +0,0 @@
-# we would prefer a symbolic link, but it does not work on windows
-import os
-target = os.path.join(os.path.dirname(__file__),
- '../../opentuner/utils/adddeps.py')
-execfile(target, dict(__file__=target))
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/hpl/hpl.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/hpl/hpl.py
deleted file mode 100644
index 4cbbe79824..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/hpl/hpl.py
+++ /dev/null
@@ -1,98 +0,0 @@
-import adddeps #fix sys.path
-
-import argparse
-import logging
-
-import opentuner
-from opentuner.search.manipulator import (ConfigurationManipulator,
- IntegerParameter,
- FloatParameter)
-from opentuner.search.objective import MinimizeTime
-from opentuner.measurement import MeasurementInterface
-from opentuner.measurement.inputmanager import FixedInputManager
-from opentuner.tuningrunmain import TuningRunMain
-
-log = logging.getLogger(__name__)
-
-parser = argparse.ArgumentParser(parents=opentuner.argparsers())
-
-parser.add_argument('--size', type=int, default=800,
- help='dimensions for the HPL matrix')
-parser.add_argument('--nprocs', type=int, default=4,
- help='number of processors for each HPL run (minimum=4)')
-parser.add_argument('--xhpl', type=str, default="hpl-2.1/bin/OSX/xhpl",
- help='location of xhpl binary')
-
-class HPLinpack(MeasurementInterface):
- def run(self, desired_result, input, limit):
- self.output_hpl_datfile(desired_result.configuration.data)
- import subprocess, os
- binary = self.args.xhpl
- subprocess.call(["mpirun", "-np", str(self.args.nprocs), binary])
-
- val = self.get_time_from_hpl_output()
-
- return opentuner.resultsdb.models.Result(time=val)
-
- def manipulator(self):
- #FIXME: should some of these be expressed as booleans or switch parameters?
- #FIXME: how to express P and Q, given PxQ=nprocs, with nprocs being fixed?
- #FIXME: how to express logscaled parameter with a particular base?
- manipulator = ConfigurationManipulator()
- manipulator.add_parameter(IntegerParameter("blocksize", 1, 64))
- manipulator.add_parameter(IntegerParameter("row_or_colmajor_pmapping", 0, 1))
- manipulator.add_parameter(IntegerParameter("pfact", 0, 2))
- manipulator.add_parameter(IntegerParameter("nbmin", 1, 4))
- manipulator.add_parameter(IntegerParameter("ndiv", 2, 2))
- manipulator.add_parameter(IntegerParameter("rfact", 0, 4))
- manipulator.add_parameter(IntegerParameter("bcast", 0, 5))
- manipulator.add_parameter(IntegerParameter("depth", 0, 4))
- manipulator.add_parameter(IntegerParameter("swap", 0, 2))
- manipulator.add_parameter(IntegerParameter("swapping_threshold", 64, 128))
- manipulator.add_parameter(IntegerParameter("L1_transposed", 0, 1))
- manipulator.add_parameter(IntegerParameter("U_transposed", 0, 1))
- manipulator.add_parameter(IntegerParameter("mem_alignment", 4, 16))
-
- return manipulator
-
- def output_hpl_datfile(self, params):
- """HPL uses an input file to express the parameters, and this uses mako to render it."""
- params["size"] = self.args.size
- from mako.template import Template
- template = Template(filename="HPL.dat.mako")
- with open("HPL.dat", "w") as f:
- f.write(template.render(**params))
-
- def get_time_from_hpl_output(self, fname="HPL.out"):
- """Returns the elapsed time only, from the HPL output file"""
- #FIXME: clean up with REs
- elapsed = 0.0
- with open(fname) as f:
- line = f.readline()
- while (line[0:3] != "T/V"):
- line = f.readline()
- line = f.readline()
- while (line[0:3] != "T/V"):
- line = f.readline()
- f.readline() # line of dashes
- splitted = f.readline().split()
- elapsed = float(splitted[5])
-
- return elapsed
-
-
- def program_name(self):
- return "HPL"
-
- def program_version(self):
- return "size=%d,nprocs=%d" % (self.args.size, self.args.nprocs)
-
- def save_final_config(self, configuration):
- '''
- called at the end of autotuning with the best resultsdb.models.Configuration
- '''
- print "Final configuration", configuration.data
-
-if __name__ == '__main__':
- args = parser.parse_args()
- HPLinpack.main(args)
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/README.md b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/README.md
deleted file mode 100644
index f094e987f5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/README.md
+++ /dev/null
@@ -1,33 +0,0 @@
-This is an OpenTuner-based tuner that learns a series of button presses that complete the first level of Super Mario Bros. for the original Nintendo Entertainment System.
-
-## Dependencies
-
-- FCEUX, a NES emulator
-- `xvfb-run`, to run the emulator headless (optional, but speeds up tuning)
-- Super Mario Bros., assumed to be named `smb.nes`, which we can't help you get for legal reasons
-
-## Running
-
-Run the tuner with `./mario.py --technique=PSO_GA_Bandit`; it will launch FCEUX to run trials. You can experiment with other techniques or `--parallelism` (the number of trials to run in parallel) too.
-
-You can implement your own configuration representation by subclassing Representation and passing `--representation=YourRepresentation`. Your Representation class needs to provide a ConfigurationManipulator populated with parameters and a method to translate these parameters to button presses. There are already a few representations implemented to use as examples.
-
-You can implement your own fitness function by subclassing FitnessFunction and passing `--fitness-function=YourFunction`. Your function receives a win/loss boolean, the number of pixels moved to the right when the trial ended, and the number of frames that elapsed during the trial. Lower fitness scores are better. There are a few existing fitness functions; in particular, `ProgressTimesAverageSpeed` also tries to optimize speed.
-
-If you want to watch the trials (or don't have `xvfb-run` available), pass `--headful`.
-
-## Playing the results
-
-When a tuning run completes, the best configuration (as judged by the fitness function) is written to `<hostname>-<tuningrun>.fm2`. This file can be played back in FCEUX to watch the best configuration.
-
-You can also use the `--tuning-run=` option (passing the tuning run number in the best configuration `.fm2`) to generate a new-bests `.fm2`, which will contain each tuning trial that was the best configuration found so far during the tuning run, concatenated back-to-back. You also need to pass `--database` pointing to the database containing that tuning run, and if you passed `--representation` or `--fitness-function` during the tuning run, you need to pass the same values for those parameters. So your final command might look like `./mario.py --tuning-run=42 --database=opentuner.db/hostname.db --representation=NaiveRepresentation --fitness-function=ProgressTimesAverageSpeed > new-bests-42.fm2`.
-
-## TODO
-
-- use the [fm2 format](http://www.fceux.com/web/help/fceux.html?fm2.html)'s subtitle support in new-bests movies to show run number and fitness score
-
-## Links
-
-- [Videos showing OpenTuner playing Super Mario Bros](https://www.youtube.com/playlist?list=PLngnz1zPEA08FWy8wF9JbGqjlm-elHmlb)
-- [Slides describing representation and results](http://groups.csail.mit.edu/commit/papers/2014/ansel-pact14-opentuner-slides.pdf) (see slide 16)
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/adddeps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/adddeps.py
deleted file mode 100644
index ede22a8fcd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/adddeps.py
+++ /dev/null
@@ -1,6 +0,0 @@
-# we would prefer a symbolic link, but it does not work on windows
-import os
-target = os.path.join(os.path.dirname(__file__),
- '../../opentuner/utils/adddeps.py')
-execfile(target, dict(__file__=target))
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/fceux-hook.lua b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/fceux-hook.lua
deleted file mode 100644
index ce00288149..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/fceux-hook.lua
+++ /dev/null
@@ -1,25 +0,0 @@
-player_state_addr = 0x000E;
-player_state_dying = 6;
-player_float_addr = 0x001D;
-player_float_flagpole = 3;
-player_page_addr = 0x006D;
-player_horizpos_addr = 0x0086;
-minimum_frames = 197;
-
-emu.speedmode("maximum");
-while true do
- if (emu.framecount() > minimum_frames) then
- --dead?
- local dead = memory.readbyte(player_state_addr) == player_state_dying;
- --flagpole?
- local won = memory.readbyte(player_float_addr) == player_float_flagpole;
- if (dead or won) then
- local str = (dead and "died" or "won");
- local x_pos = math.floor(memory.readbyteunsigned(player_page_addr)*256 + memory.readbyteunsigned(player_horizpos_addr));
- local framecount = emu.framecount();
- io.write(str, " ", x_pos, " ", framecount, "\n");
- os.exit(0);
- end;
- end;
- emu.frameadvance();
-end
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/mario.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/mario.py
deleted file mode 100755
index d388321e1f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/mario/mario.py
+++ /dev/null
@@ -1,341 +0,0 @@
-#!/usr/bin/env python2
-
-"""OpenTuner plays Super Mario Bros. for NES
-
-We write a movie file and ask the emulator to play it back while running
-fceux-hook.lua, which checks for death/flagpole and prints the fitness to
-stdout where OpenTuner, as the parent process, can read it.
-"""
-
-import adddeps #fix sys.path
-import argparse
-import base64
-import pickle
-import tempfile
-import subprocess
-import re
-import zlib
-import abc
-import sys
-import os
-import traceback
-import collections
-import socket
-
-import opentuner
-from opentuner.search.manipulator import ConfigurationManipulator, IntegerParameter, EnumParameter, BooleanParameter
-from opentuner.measurement import MeasurementInterface
-from opentuner.measurement.inputmanager import FixedInputManager
-from opentuner.tuningrunmain import TuningRunMain
-from opentuner.search.objective import MinimizeTime
-
-def instantiate(class_name):
- return getattr(sys.modules[__name__], class_name)()
-
-argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
-argparser.add_argument('--tuning-run', type=int, help='concatenate new bests from given tuning run into single movie')
-argparser.add_argument('--headful', action='store_true', help='run headful (not headless) for debugging or live demo')
-argparser.add_argument('--xvfb-delay', type=int, default=0, help='delay between launching xvfb and fceux')
-argparser.add_argument('--representation', default='DurationRepresentation', type=instantiate, help='name of representation class')
-argparser.add_argument('--fitness-function', default='Progress', type=instantiate, help='name of fitness function class')
-
-def call_or_die(command, failmsg=None):
- try:
- p = subprocess.Popen(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
- stdout, stderr = p.communicate()
- return stdout, stderr, p.returncode
- except:
- print "Failed to execute", command
- traceback.print_exc()
- print "Child traceback:"
- print sys.exc_info()[1].child_traceback
- if failmsg:
- print failmsg
- sys.exit(1)
-
-# Functions for building FCEUX movie files (.fm2 files)
-
-def fm2_line(up, down, left, right, a, b, start, select, reset=False):
- """formats one frame of input with the given button presses"""
- return ''.join(('|1|' if reset else '|0|') +
- ('R' if right else '.') +
- ('L' if left else '.') +
- ('D' if down else '.') +
- ('U' if up else '.') +
- ('T' if start else '.') +
- ('D' if select else '.') +
- ('B' if b else '.') +
- ('A' if a else '.') +
- '|........||')
-
-def maxd(iterable, default):
- try:
- return max(iterable)
- except ValueError:
- return default
-
-def fm2_lines(up, down, left, right, a, b, start, select, reset=set(), minFrame=None, maxFrame=None):
- """formats many frames using the given button-press sets"""
- if minFrame is None:
- minFrame = 0
- if maxFrame is None:
- maxFrame = max(maxd(up, 0), maxd(down, 0), maxd(left, 0), maxd(right, 0), maxd(a, 0), maxd(b, 0), maxd(start, 0), maxd(select, 0), maxd(reset, 0)) + 1
- lines = list()
- for i in xrange(minFrame, maxFrame):
- lines.append(fm2_line(i in up, i in down, i in left, i in right, i in a, i in b, i in start, i in select, i in reset))
- return lines
-
-def fm2_smb_header():
- return ["version 3",
- "emuVersion 9828",
- "romFilename smb.nes",
- "romChecksum base64:jjYwGG411HcjG/j9UOVM3Q==",
- "guid 51473540-E9D7-11E3-ADFC-46CE3219C4E0",
- "fourscore 0",
- "port0 1",
- "port1 1",
- "port2 0"]
-
-def fm2_smb(left, right, down, b, a, header=True, padding=True, minFrame=None, maxFrame=None):
- reset = set()
- start = set()
- if padding:
- left = set([x+196 for x in left])
- right = set([x+196 for x in right])
- down = set([x+196 for x in down])
- b = set([x+196 for x in b])
- a = set([x+196 for x in a])
- reset.add(0)
- start.add(33)
- lines = fm2_lines(set(), down, left, right, a, b, start, set(), reset, minFrame, maxFrame)
- if header:
- return "\n".join(fm2_smb_header() + lines)
- else:
- return "\n".join(lines)
-
-display_numbers = collections.deque()
-
-def run_movie(fm2, args):
- with tempfile.NamedTemporaryFile(suffix=".fm2", delete=True) as f:
- f.write(fm2)
- f.flush()
- cmd = []
- if not args.headful:
- display = display_numbers.pop()
- cmd += ["xvfb-run", "-n", display, "-w", str(args.xvfb_delay), "-e", "/dev/stderr"]
- cmd += ["fceux", "--playmov", f.name, "--loadlua",
- "fceux-hook.lua", "--nogui", "--volume", "0", "--no-config", "1",
- "smb.nes"]
- stdout, stderr, returncode = call_or_die(cmd)
- if not args.headful:
- display_numbers.append(display)
- match = re.search(r"^(won|died) (\d+) (\d+)$", stdout, re.MULTILINE)
- if not match:
- print stderr
- print stdout
- raise ValueError
- wl = match.group(1)
- x_pos = int(match.group(2))
- framecount = int(match.group(3))
- return (wl, x_pos, framecount)
-
-class Representation(object):
- """Interface for pluggable tuning representations."""
- __metaclass__ = abc.ABCMeta
-
- @abc.abstractmethod
- def manipulator():
- """Return a ConfigurationManipulator for this representation."""
- pass
-
- @abc.abstractmethod
- def interpret(cfg):
- """Unpack this representation into button-press sets (L, R, D, B, A)."""
- pass
-
-class NaiveRepresentation(Representation):
- """Uses a parameter per (button, frame) pair."""
- def manipulator(self):
- m = ConfigurationManipulator()
- for i in xrange(0, 12000):
- m.add_parameter(BooleanParameter('L{}'.format(i)))
- m.add_parameter(BooleanParameter('R{}'.format(i)))
- m.add_parameter(BooleanParameter('D{}'.format(i)))
- m.add_parameter(BooleanParameter('B{}'.format(i)))
- m.add_parameter(BooleanParameter('A{}'.format(i)))
- return m
-
- def interpret(self, cfg):
- left = set()
- right = set()
- down = set()
- running = set()
- jumping = set()
- for i in xrange(0, 12000):
- if cfg['L{}'.format(i)]:
- left.add(i)
- if cfg['R{}'.format(i)]:
- right.add(i)
- if cfg['D{}'.format(i)]:
- down.add(i)
- if cfg['B{}'.format(i)]:
- running.add(i)
- if cfg['A{}'.format(i)]:
- jumping.add(i)
- return left, right, down, running, jumping
-
-class DurationRepresentation(Representation):
- def manipulator(self):
- m = ConfigurationManipulator()
- for i in xrange(0, 1000):
- #bias 3:1 in favor of moving right
- m.add_parameter(EnumParameter('move{}'.format(i), ["R", "L", "RB", "LB", "N", "LR", "LRB", "R2", "RB2", "R3", "RB3"]))
- m.add_parameter(IntegerParameter('move_duration{}'.format(i), 1, 60))
- #m.add_parameter(BooleanParameter("D"+str(i)))
- for i in xrange(0, 1000):
- m.add_parameter(IntegerParameter('jump_frame{}'.format(i), 0, 24000))
- m.add_parameter(IntegerParameter('jump_duration{}'.format(i), 1, 32))
- return m
-
- def interpret(self, cfg):
- left = set()
- right = set()
- down = set()
- running = set()
- start = 0
- for i in xrange(0, 1000):
- move = cfg['move{}'.format(i)]
- move_duration = cfg['move_duration{}'.format(i)]
- if "R" in move:
- right.update(xrange(start, start + move_duration))
- if "L" in move:
- left.update(xrange(start, start + move_duration))
- if "B" in move:
- running.update(xrange(start, start + move_duration))
- start += move_duration
- jumping = set()
- for i in xrange(0, 1000):
- jump_frame = cfg['jump_frame{}'.format(i)]
- jump_duration = cfg['jump_duration{}'.format(i)]
- jumping.update(xrange(jump_frame, jump_frame + jump_duration))
- return left, right, down, running, jumping
-
-class AlphabetRepresentation(Representation):
- def manipulator(self):
- m = ConfigurationManipulator()
- for i in xrange(0, 400*60):
- m.add_parameter(EnumParameter('{}'.format(i), xrange(0, 16)))
- return m
-
- def interpret(self, cfg):
- left = set()
- right = set()
- down = set()
- running = set()
- jumping = set()
- for i in xrange(0, 400*60):
- bits = cfg[str(i)]
- if bits & 1:
- left.add(i)
- if bits & 2:
- right.add(i)
- if bits & 4:
- running.add(i)
- if bits & 8:
- jumping.add(i)
- #if bits & 16:
- # down.add(i)
- return left, right, down, running, jumping
-
-class FitnessFunction(object):
- """Interface for pluggable fitness functions."""
- __metaclass__ = abc.ABCMeta
-
- @abc.abstractmethod
- def __call__(won, x_pos, elapsed_frames):
- """Return the fitness (float, lower is better)."""
- pass
-
-class Progress(FitnessFunction):
- def __call__(self, won, x_pos, elapsed_frames):
- return -float(x_pos)
-
-class ProgressPlusTimeRemaining(FitnessFunction):
- def __call__(self, won, x_pos, elapsed_frames):
- """x_pos plus 1 for each frame remaining on the timer on a win. This results in a large discontinuity at wins. This was the fitness function used for the OpenTuner paper, though the paper only discussed time-to-first-win."""
- return -float(x_pos + 400*60 - elapsed_frames) if won else -float(x_pos)
-
-class ProgressTimesAverageSpeed(FitnessFunction):
- def __call__(self, won, x_pos, elapsed_frames):
- return -x_pos * (float(x_pos)/elapsed_frames)
-
-class SMBMI(MeasurementInterface):
- def __init__(self, args):
- super(SMBMI, self).__init__(args)
- self.parallel_compile = True
- self.args = args
-
- def manipulator(self):
- return self.args.representation.manipulator()
-
- def compile(self, cfg, id):
- left, right, down, running, jumping = self.args.representation.interpret(cfg)
- fm2 = fm2_smb(left, right, down, running, jumping)
- try:
- wl, x_pos, framecount = run_movie(fm2, self.args)
- except ValueError:
- return opentuner.resultsdb.models.Result(state='ERROR', time=float('inf'))
- print wl, x_pos, framecount
- return opentuner.resultsdb.models.Result(state='OK', time=self.args.fitness_function("won" in wl, x_pos, framecount))
-
- def run_precompiled(self, desired_result, input, limit, compile_result, id):
- return compile_result
-
- def run(self, desired_result, input, limit):
- pass
-
- def save_final_config(self, cfg):
- left, right, down, running, jumping = args.representation.interpret(cfg.data)
- fm2 = fm2_smb(left, right, down, running, jumping)
- _, _, framecount = run_movie(fm2, self.args)
- filename = '{}-{}.fm2'.format(socket.gethostname(), self.driver.tuning_run.id)
- with open(filename, 'w') as f:
- f.write(fm2_smb(left, right, down, running, jumping, maxFrame=framecount))
-
-def new_bests_movie(args):
- stdout, stderr, returncode = call_or_die(["sqlite3", args.database, "select configuration_id from result where tuning_run_id = %d and was_new_best = 1 order by collection_date;" % args.tuning_run])
- if returncode:
- print "Error retrieving new-best configurations:", stderr
- sys.exit(1)
- cids = stdout.split()
- print '\n'.join(fm2_smb_header())
- for cid in cids:
- stdout, stderr, returncode = call_or_die(["sqlite3", args.database, "select quote(data) from configuration where id = %d;" % int(cid)])
- if returncode:
- print "Error retriving configuration data:", cid, stderr
- sys.exit(1)
- cfg = pickle.loads(zlib.decompress(base64.b16decode(stdout.strip()[2:-1])))
- left, right, down, running, jumping = args.representation.interpret(cfg)
- fm2 = fm2_smb(left, right, down, running, jumping)
- _, _, framecount = run_movie(fm2, args)
- print fm2_smb(left, right, down, running, jumping, header=False, maxFrame=framecount)
-
-if __name__ == '__main__':
- args = argparser.parse_args()
- call_or_die(["fceux", "--help"], failmsg="Is fceux on your PATH?")
- if not args.headful:
- call_or_die(["xvfb-run", "--help"], failmsg="Is xvfb-run on your PATH? (or, pass --headful)")
- for n in xrange(99, 99 + args.parallelism):
- display_numbers.append(str(n))
- if args.tuning_run:
- call_or_die(["sqlite3", "-version"], failmsg="Is sqlite3 on your PATH?")
- if args.database is not None:
- new_bests_movie(args)
- else:
- print "must specify --database"
- else:
- if os.path.isfile('smb.nes'):
- SMBMI.main(args)
- else:
- print "smb.nes not found"
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/.gitignore b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/.gitignore
deleted file mode 100644
index a6e67132d6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/.gitignore
+++ /dev/null
@@ -1 +0,0 @@
-linux_x86_64
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/README.md b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/README.md
deleted file mode 100644
index e4b4464686..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/README.md
+++ /dev/null
@@ -1,24 +0,0 @@
-Source for PetaBricks binaries can be found at:
- - https://github.com/petabricks/petabricks/
- - https://code.google.com/p/petabricks/
-
-
-Basic usage for running the raw programs is:
-```
-./Prog --config=CONFIG -n=N --time --accuracy --max-sec=TIMEOUT --trials=1
-
---config=<STRING>
- filename of the program configuration (see example in .cfg.default file)
---n=<INTEGER>
- generate a random input of the given size and run it
---time
- print timing results in xml format
---accuracy
- print out accuracy of answer
---max-sec=<NUMBER> (default: 1.79769e+308)
- terminate measurement if it exceeds the given number of seconds
-
-many more options are given by running ./Prog --help
-```
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/adddeps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/adddeps.py
deleted file mode 100644
index ede22a8fcd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/adddeps.py
+++ /dev/null
@@ -1,6 +0,0 @@
-# we would prefer a symbolic link, but it does not work on windows
-import os
-target = os.path.join(os.path.dirname(__file__),
- '../../opentuner/utils/adddeps.py')
-execfile(target, dict(__file__=target))
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/deps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/deps.py
deleted file mode 100644
index c03a106a85..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/deps.py
+++ /dev/null
@@ -1,19 +0,0 @@
-import os
-import sys
-
-project_root = os.path.normpath(os.path.join(
- os.path.dirname(os.path.abspath(__file__)), '../..'))
-sys.path.insert(0, project_root)
-
-
-try:
- from lxml import etree
-except ImportError:
- try:
- # Python 2.5
- import xml.etree.cElementTree as etree
- except ImportError:
- import xml.etree.ElementTree as etree
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/download_benchmarks.sh b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/download_benchmarks.sh
deleted file mode 100755
index aaf333b455..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/download_benchmarks.sh
+++ /dev/null
@@ -1,8 +0,0 @@
-#!/bin/sh
-if test -e linux_x86_64
-then
- echo "benchmarks already downloaded"
-else
- wget -O- http://people.csail.mit.edu/jansel/petabricks_benchmarks_linux_x86_64.tar.bz2 | tar jxv
-fi
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/import_old_result.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/import_old_result.py
deleted file mode 100755
index 9add4ed035..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/import_old_result.py
+++ /dev/null
@@ -1,116 +0,0 @@
-#!/usr/bin/env python
-
-import adddeps # fix sys.path
-
-import argparse
-import json
-import logging
-import os
-import re
-import sys
-import uuid
-import subprocess
-
-try:
- from lxml import etree
-except ImportError:
- import xml.etree.ElementTree as etree
-
-import opentuner
-from opentuner import resultsdb
-from datetime import datetime
-from datetime import timedelta
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-
-log = logging.getLogger(__name__)
-
-argparser = argparse.ArgumentParser()
-argparser.add_argument('--database', default='opentuner.db/import.db')
-argparser.add_argument('--limit', type=float, default=10)
-argparser.add_argument('program')
-argparser.add_argument('candidatelog')
-
-
-def run(args, cfg):
- limit = args.limit
- cmd = [args.program,
- '--time',
- '--accuracy',
- '--config=' + cfg,
- '--max-sec=%.10f' % args.limit,
- '-n=%d' % args.n]
- p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
- out, err = p.communicate()
-
- result = opentuner.resultsdb.models.Result()
- try:
- root = etree.XML(out)
- result.time = float(root.find('stats/timing').get('average'))
- result.accuracy = float(root.find('stats/accuracy').get('average'))
- if result.time < limit + 3600:
- result.state = 'OK'
- else:
- # time will be 2**31 if timeout
- result.state = 'TIMEOUT'
- except:
- log.exception('run error')
- log.warning('program crash, out = %s / err = %s', out, err)
- result.state = 'ERROR'
- result.time = float('inf')
- result.accuracy = float('-inf')
- return result
-
-
-def main(args):
- if '://' not in args.database:
- args.database = 'sqlite:///' + args.database
- engine, Session = opentuner.resultsdb.connect(args.database)
- session = Session()
-
- program_settings = json.load(open(args.program + '.settings'))
- args.n = program_settings['n']
- args.technique = ['Imported']
- objective = ThresholdAccuracyMinimizeTime(program_settings['accuracy'])
-
- tuningrun = resultsdb.models.TuningRun(
- uuid=uuid.uuid4().hex,
- name='import',
- args=args,
- start_date=datetime.now(),
- objective=objective,
- program_version=resultsdb.models.ProgramVersion.get(
- session, 'PetaBricksInterface', args.program, 'imported'),
- state='COMPLETE',
- )
- session.add(tuningrun)
-
- for gen, line in enumerate(open(args.candidatelog)):
- if line[0] != '#':
- line = re.split('\t', line)
- date = tuningrun.start_date + timedelta(seconds=float(line[0]))
- cfg = os.path.normpath(
- os.path.join(os.path.dirname(args.candidatelog), '..', line[5]))
- result = run(args, cfg)
- result.was_new_best = True
- result.tuning_run = tuningrun
- result.collection_date = date
- session.add(result)
- desired_result = resultsdb.models.DesiredResult(
- limit=args.limit,
- tuning_run=tuningrun,
- generation=gen,
- requestor='Imported',
- request_date=date,
- start_date=date,
- result=result,
- state='COMPLETE')
- session.add(desired_result)
- tuningrun.end_date = date
- print gen, date, result.time
-
- session.commit()
-
-
-if __name__ == '__main__':
- opentuner.tuningrunmain.init_logging()
- sys.exit(main(argparser.parse_args()))
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/pbtuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/pbtuner.py
deleted file mode 100755
index 7163294494..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/pbtuner.py
+++ /dev/null
@@ -1,188 +0,0 @@
-#!/usr/bin/env python
-
-import adddeps # fix sys.path
-
-import re
-import argparse
-import logging
-import subprocess
-import tempfile
-import json
-from pprint import pprint
-
-import opentuner
-from opentuner.search.manipulator import (ConfigurationManipulator,
- IntegerParameter,
- LogIntegerParameter,
- FloatParameter,
- LogFloatParameter,
- SelectorParameter,
- SwitchParameter,
- PermutationParameter, )
-
-try:
- from lxml import etree
-except ImportError:
- import xml.etree.ElementTree as etree
-
-from opentuner.measurement import MeasurementInterface
-from opentuner.measurement.inputmanager import FixedInputManager
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-
-log = logging.getLogger("pbtuner")
-
-parser = argparse.ArgumentParser(parents=opentuner.argparsers())
-parser.add_argument('program',
- help='PetaBricks binary program to autotune')
-parser.add_argument('--program-cfg-default',
- help="override default program config exemplar location")
-parser.add_argument('--program-cfg-output',
- help="location final autotuned configuration is written")
-parser.add_argument('--program-settings',
- help="override default program settings file location")
-parser.add_argument('--program-input',
- help="use only a given input for autotuning")
-parser.add_argument('--upper-limit', type=float, default=30,
- help="time limit to apply to initial test")
-parser.add_argument('--test-config', action='store_true')
-
-
-class PetaBricksInterface(MeasurementInterface):
- def __init__(self, args):
- self.program_settings = json.load(open(args.program_settings))
- input_manager = FixedInputManager(size=self.program_settings['n'])
- objective = ThresholdAccuracyMinimizeTime(self.program_settings['accuracy'])
-
- # pass many settings to parent constructor
- super(PetaBricksInterface, self).__init__(
- args, program_name=args.program,
- program_version=self.file_hash(args.program),
- input_manager=input_manager, objective=objective)
-
- def build_config(self, cfg):
- r = dict()
-
- # direct copy
- for k, v in cfg.iteritems():
- if k[0] != '.':
- r[k] = v
-
- for name, choices in self.choice_sites.items():
- param = self.manipulator.parameters_dict(cfg)['.' + name]
- lvl = 0
- for cutoff, choice in param.selector_iter(cfg):
- lvl += 1
- r['%s_lvl%d_rule' % (name, lvl)] = choice
- if lvl > 1:
- r['%s_lvl%d_cutoff' % (name, lvl)] = cutoff
-
- return r
-
- def run(self, desired_result, input, limit):
- limit = min(limit, self.args.upper_limit)
- with tempfile.NamedTemporaryFile(suffix='.petabricks.cfg') as cfgtmp:
- for k, v in self.build_config(desired_result.configuration.data).items():
- print >> cfgtmp, k, '=', v
- cfgtmp.flush()
- if args.program_input:
- input_opts = ['--iogen-run=' + args.program_input,
- '--iogen-n=%d' % input.input_class.size]
- else:
- input_opts = ['-n=%d' % input.input_class.size]
-
- cmd = [args.program,
- '--time',
- '--accuracy',
- '--max-sec=%.8f' % limit,
- '--config=' + cfgtmp.name] + input_opts
- log.debug("cmd: %s", ' '.join(cmd))
- p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
- out, err = p.communicate()
-
- result = opentuner.resultsdb.models.Result()
- try:
- root = etree.XML(out)
- result.time = float(root.find('stats/timing').get('average'))
- result.accuracy = float(root.find('stats/accuracy').get('average'))
- if result.time < limit + 3600:
- result.state = 'OK'
- else:
- #time will be 2**31 if timeout
- result.state = 'TIMEOUT'
- except:
- log.warning("program crash, out = %s / err = %s", out, err)
- result.state = 'ERROR'
- result.time = float('inf')
- result.accuracy = float('-inf')
- return result
-
- def save_final_config(self, configuration):
- """
- called at the end of autotuning with the best
- resultsdb.models.Configuration
- """
- with open(args.program_cfg_output, 'w') as fd:
- cfg = self.build_config(configuration.data)
- for k, v in sorted(cfg.items()):
- print >> fd, k, '=', v
- log.info("final configuration written to %s", args.program_cfg_output)
-
- def manipulator(self):
- """create the configuration manipulator, from example config"""
- upper_limit = self.program_settings['n'] + 1
- cfg = open(self.args.program_cfg_default).read()
- manipulator = ConfigurationManipulator()
-
- self.choice_sites = dict()
-
- for m in re.finditer(r" *([a-zA-Z0-9_-]+)[ =]+([0-9e.+-]+) *"
- r"[#] *([a-z]+).* ([0-9]+) to ([0-9]+)", cfg):
- k, v, valtype, minval, maxval = m.group(1, 2, 3, 4, 5)
- minval = float(minval)
- maxval = float(maxval)
- if upper_limit:
- maxval = min(maxval, upper_limit)
- assert valtype == 'int'
- #log.debug("param %s %f %f", k, minval, maxval)
-
- m1 = re.match(r'(.*)_lvl[0-9]+_rule', k)
- m2 = re.match(r'(.*)_lvl[0-9]+_cutoff', k)
- if m1:
- self.choice_sites[m1.group(1)] = int(maxval)
- elif m2:
- pass
- elif k == 'worker_threads':
- manipulator.add_parameter(IntegerParameter(k, 1, 16))
- elif k == 'distributedcutoff':
- pass
- elif minval == 0 and maxval < 64:
- manipulator.add_parameter(SwitchParameter(k, maxval))
- else:
- manipulator.add_parameter(LogIntegerParameter(k, minval, maxval))
-
- for name, choices in self.choice_sites.items():
- manipulator.add_parameter(
- SelectorParameter('.' + name, range(choices + 1),
- upper_limit / choices))
-
- self.manipulator = manipulator
- return manipulator
-
- def test_config(self):
- pprint(self.manipulator().random())
-
-
-if __name__ == '__main__':
- args = parser.parse_args()
- if not args.program_cfg_default:
- args.program_cfg_default = args.program + '.cfg.default'
- if not args.program_cfg_output:
- args.program_cfg_output = args.program + '.cfg'
- if not args.program_settings:
- args.program_settings = args.program + '.settings'
- if args.test_config:
- PetaBricksInterface(args).test_config()
- else:
- PetaBricksInterface.main(args)
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/testwrapper.sh b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/testwrapper.sh
deleted file mode 100755
index 2b6a94e57a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/petabricks/testwrapper.sh
+++ /dev/null
@@ -1,11 +0,0 @@
-#!/bin/bash
-COUNT=50
-for Z in `seq $COUNT`
-do
- for T in `./pbtuner.py --list-techniques $@`;
- do
- echo $Z/$COUNT $T
- ./pbtuner.py --technique=$T $@
- done
-done
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/py_api/adddeps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/py_api/adddeps.py
deleted file mode 100644
index ede22a8fcd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/py_api/adddeps.py
+++ /dev/null
@@ -1,6 +0,0 @@
-# we would prefer a symbolic link, but it does not work on windows
-import os
-target = os.path.join(os.path.dirname(__file__),
- '../../opentuner/utils/adddeps.py')
-execfile(target, dict(__file__=target))
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/py_api/api_example.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/py_api/api_example.py
deleted file mode 100755
index e87a8fffe1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/py_api/api_example.py
+++ /dev/null
@@ -1,58 +0,0 @@
-#!/usr/bin/python
-"""
-Examples usage of a Python API interface to opentuner.
-
-Unlike the other examples, this code lets the user control the main() of
-the program and calls into opentuner to get new configurations to test.
-"""
-
-import adddeps # add opentuner to path in dev mode
-
-import opentuner
-from opentuner.api import TuningRunManager
-from opentuner.measurement.interface import DefaultMeasurementInterface
-from opentuner.resultsdb.models import Result
-from opentuner.search.manipulator import ConfigurationManipulator
-from opentuner.search.manipulator import IntegerParameter
-import logging
-import argparse
-
-log = logging.getLogger(__name__)
-
-
-def test_func(cfg):
- x = cfg['x']
- y = (x - 10) * (x - 10)
- log.debug("f({}) -> {}".format(x, y))
- return y
-
-
-def main():
- parser = argparse.ArgumentParser(parents=opentuner.argparsers())
- args = parser.parse_args()
- manipulator = ConfigurationManipulator()
- manipulator.add_parameter(IntegerParameter('x', -200, 200))
- interface = DefaultMeasurementInterface(args=args,
- manipulator=manipulator,
- project_name='examples',
- program_name='api_test',
- program_version='0.1')
- api = TuningRunManager(interface, args)
- for x in xrange(500):
- desired_result = api.get_next_desired_result()
- if desired_result is None:
- # The search space for this example is very small, so sometimes
- # the techniques have trouble finding a config that hasn't already
- # been tested. Change this to a continue to make it try again.
- break
- cfg = desired_result.configuration.data
- result = Result(time=test_func(cfg))
- api.report_result(desired_result, result)
-
- best_cfg = api.get_best_configuration()
- api.finish()
- print 'best x found was', best_cfg['x']
-
-if __name__ == '__main__':
- main()
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/py_api/multiple_tuning_runs.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/py_api/multiple_tuning_runs.py
deleted file mode 100755
index 5e0918e3af..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/py_api/multiple_tuning_runs.py
+++ /dev/null
@@ -1,83 +0,0 @@
-#!/usr/bin/python
-"""
-Examples usage of a Python API interface to opentuner.
-
-Unlike the other examples, this code lets the user control the main() of
-the program and calls into opentuner to get new configurations to test.
-
-This version runs multiple tuning runs at once in a single process.
-"""
-
-import adddeps # add opentuner to path in dev mode
-
-import opentuner
-from opentuner.api import TuningRunManager
-from opentuner.measurement.interface import DefaultMeasurementInterface
-from opentuner.resultsdb.models import Result
-from opentuner.search.manipulator import ConfigurationManipulator
-from opentuner.search.manipulator import IntegerParameter
-import logging
-import argparse
-
-log = logging.getLogger(__name__)
-
-
-def test_func1(cfg):
- x = cfg['x']
- y = (x - 10) * (x - 10)
- log.debug("f({}) -> {}".format(x, y))
- return y
-
-
-def test_func2(cfg):
- x = cfg['x']
- y = (x + 10) * (x + 10)
- log.debug("f({}) -> {}".format(x, y))
- return y
-
-
-def test_func3(cfg):
- x = cfg['x']
- y = (x + 20) * (x + 20)
- log.debug("f({}) -> {}".format(x, y))
- return y
-
-
-def create_test_tuning_run(db):
- parser = argparse.ArgumentParser(parents=opentuner.argparsers())
- args = parser.parse_args()
- args.database = db
- manipulator = ConfigurationManipulator()
- manipulator.add_parameter(IntegerParameter('x', -200, 200))
- interface = DefaultMeasurementInterface(args=args,
- manipulator=manipulator,
- project_name='examples',
- program_name='api_test',
- program_version='0.1')
- api = TuningRunManager(interface, args)
- return api
-
-
-def main():
- apis = [create_test_tuning_run('sqlite:////tmp/a.db'),
- create_test_tuning_run('sqlite:////tmp/b.db'),
- create_test_tuning_run('sqlite:////tmp/c.db')]
- test_funcs = [test_func1, test_func2, test_func3]
- for x in xrange(100):
- for api, test_func in zip(apis, test_funcs):
- desired_result = api.get_next_desired_result()
- if desired_result is None:
- continue
- cfg = desired_result.configuration.data
- result = Result(time=test_func(cfg))
- api.report_result(desired_result, result)
-
- best_cfgs = [api.get_best_configuration() for api in apis]
- for api in apis:
- api.finish()
-
- print('best x configs: {}'.format(best_cfgs))
-
-if __name__ == '__main__':
- main()
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/.gitignore b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/.gitignore
deleted file mode 100644
index aa0571caf1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/.gitignore
+++ /dev/null
@@ -1,3 +0,0 @@
-rosenbrock.db
-*.db
-opentuner.log
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/adddeps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/adddeps.py
deleted file mode 100644
index ede22a8fcd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/adddeps.py
+++ /dev/null
@@ -1,6 +0,0 @@
-# we would prefer a symbolic link, but it does not work on windows
-import os
-target = os.path.join(os.path.dirname(__file__),
- '../../opentuner/utils/adddeps.py')
-execfile(target, dict(__file__=target))
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/rosenbrock.makefile b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/rosenbrock.makefile
deleted file mode 100755
index 7b9be87a9a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/rosenbrock.makefile
+++ /dev/null
@@ -1,24 +0,0 @@
-#!/usr/bin/make -f
-# use -j4 to run in parallel
-
-FN := rosenbrock
-DIMS := 4
-TECHNIQUES := $(shell ./rosenbrock.py --list-techniques)
-define test_loop
-DB="sqlite:///opentuner.db/$$RUN.db"; \
-for TEQ in $(TECHNIQUES); do \
- ./rosenbrock.py --function=$(FN) \
- --technique=$$TEQ \
- --dimensions=$(DIMS) \
- --database=$$DB; \
-done;
-endef
-
-default: run.1 run.2 run.3 run.4 run.5 run.6 run.7 run.8 run.9 run.10 run.11 \
-run.12 run.13 run.14 run.15 run.16 run.17 run.18 run.19 run.20 run.21 run.22 \
-run.23 run.24 run.25 run.26 run.27 run.28 run.29 run.30
-
-run.%:
- RUN=$* $(test_loop)
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/rosenbrock.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/rosenbrock.py
deleted file mode 100755
index da426f239b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/rosenbrock/rosenbrock.py
+++ /dev/null
@@ -1,85 +0,0 @@
-#!/usr/bin/env python
-#
-# This is a simple testcase purely for testing the autotuner
-#
-# http://en.wikipedia.org/wiki/Rosenbrock_function
-#
-# Also supports some other test functions taken from:
-# http://en.wikipedia.org/wiki/Test_functions_for_optimization
-#
-
-import adddeps # fix sys.path
-
-import argparse
-import logging
-
-import opentuner
-from opentuner.measurement import MeasurementInterface
-from opentuner.search.manipulator import ConfigurationManipulator
-from opentuner.search.manipulator import FloatParameter
-
-log = logging.getLogger(__name__)
-
-parser = argparse.ArgumentParser(parents=opentuner.argparsers())
-parser.add_argument('--dimensions', type=int, default=2,
- help='dimensions for the Rosenbrock function')
-parser.add_argument('--domain', type=float, default=1000,
- help='bound for variables in each dimension')
-parser.add_argument('--function', default='rosenbrock',
- choices=('rosenbrock', 'sphere', 'beale'),
- help='function to use')
-
-
-class Rosenbrock(MeasurementInterface):
- def run(self, desired_result, input, limit):
- cfg = desired_result.configuration.data
- val = 0.0
- if self.args.function == 'rosenbrock':
- # the actual rosenbrock function:
- for d in xrange(self.args.dimensions - 1):
- x0 = cfg[d]
- x1 = cfg[d + 1]
- val += 100.0 * (x1 - x0 ** 2) ** 2 + (x0 - 1) ** 2
- elif self.args.function == 'sphere':
- for d in xrange(self.args.dimensions):
- xi = cfg[d]
- val += xi ** 2
- elif self.args.function == 'beale':
- assert self.args.dimensions == 2
- assert self.args.domain == 4.5
- x = cfg[0]
- y = cfg[1]
- val = ((1.5 - x + x * y) ** 2 +
- (2.25 - x + x * y ** 2) ** 2 +
- (2.625 - x + x * y ** 3) ** 2)
- return opentuner.resultsdb.models.Result(time=val)
-
- def manipulator(self):
- manipulator = ConfigurationManipulator()
- for d in xrange(self.args.dimensions):
- manipulator.add_parameter(FloatParameter(d,
- -self.args.domain,
- self.args.domain))
- return manipulator
-
- def program_name(self):
- return self.args.function
-
- def program_version(self):
- return "%dx%d" % (self.args.dimensions, self.args.domain)
-
- def save_final_config(self, configuration):
- """
- called at the end of autotuning with the best resultsdb.models.Configuration
- """
- print "Final configuration", configuration.data
-
-
-if __name__ == '__main__':
- args = parser.parse_args()
- if args.function == 'beale':
- # fixed for this function
- args.domain = 4.5
- args.dimensions = 2
- Rosenbrock.main(args)
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/adddeps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/adddeps.py
deleted file mode 100644
index ede22a8fcd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/adddeps.py
+++ /dev/null
@@ -1,6 +0,0 @@
-# we would prefer a symbolic link, but it does not work on windows
-import os
-target = os.path.join(os.path.dirname(__file__),
- '../../opentuner/utils/adddeps.py')
-execfile(target, dict(__file__=target))
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/att48_d.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/att48_d.txt
deleted file mode 100644
index b93e36ccfa..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/att48_d.txt
+++ /dev/null
@@ -1,48 +0,0 @@
- 0 4727 1205 6363 3657 3130 2414 563 463 5654 1713 1604 2368 2201 1290 1004 3833 2258 3419 2267 2957 720 1700 5279 2578 6076 3465 2654 3625 3115 1574 3951 1748 2142 6755 2383 3306 1029 3530 825 2188 4820 3489 1947 6835 1542 2379 3744
- 4727 0 3588 2012 1842 6977 6501 5187 5028 2327 4148 4723 3635 3125 4907 3930 7463 6338 7243 5105 4043 4022 3677 2863 3106 1850 7173 6630 1204 6814 6001 3447 5253 2656 3123 6274 7183 5622 3085 4564 2756 1591 7027 6186 3472 5461 4390 2088
- 1205 3588 0 5163 2458 3678 3071 1742 1444 4462 1184 1520 1498 1103 1501 951 4298 2903 3967 2169 2209 652 828 4136 1518 4873 3954 3254 2446 3581 2441 2960 1966 950 5564 2916 3878 2035 2482 1027 1395 3617 3891 2686 5661 2023 1867 2560
- 6363 2012 5163 0 2799 8064 7727 6878 6581 1402 5366 5946 4679 4378 6225 5709 8417 7578 8296 6135 4802 5707 4982 2322 4178 320 8186 7800 2778 7859 7408 3763 6461 4223 1427 7451 8263 7131 3669 6011 4638 1681 7987 7502 1877 6758 5360 2844
- 3657 1842 2458 2799 0 5330 4946 4200 3824 2012 2573 3157 1924 1580 3427 3179 5749 4793 5577 3409 2223 3066 2185 1860 1401 2491 5486 5035 894 5141 4611 1669 3677 1590 3113 4682 5533 4352 1252 3227 2426 1169 5313 4706 3241 3962 2651 304
- 3130 6977 3678 8064 5330 0 743 3209 2670 6929 2831 2266 3407 3854 2178 4076 727 881 293 1930 3310 3672 3315 6199 3932 7745 365 482 5774 261 1659 4513 1746 4431 7910 769 207 2225 4435 2681 5053 6384 550 1224 7805 1670 2704 5230
- 2414 6501 3071 7727 4946 743 0 2468 1952 6673 2380 1795 3051 3405 1604 3382 1469 168 1020 1681 3110 2993 2827 6009 3552 7412 1104 267 5300 821 916 4348 1270 3890 7698 332 900 1484 4185 2049 4415 6051 1219 482 7635 1054 2432 4884
- 563 5187 1742 6878 4200 3209 2468 0 718 6203 2241 2051 2920 2762 1687 1304 3932 2331 3487 2669 3487 1175 2260 5840 3141 6596 3563 2728 4120 3240 1559 4507 2082 2658 7304 2512 3364 985 4091 1319 2544 5358 3632 1987 7391 1785 2879 4296
- 463 5028 1444 6581 3824 2670 1952 718 0 5789 1602 1343 2330 2291 970 1451 3376 1796 2959 1951 2835 1112 1725 5346 2628 6285 3007 2193 3889 2661 1122 3920 1372 2391 6883 1927 2845 611 3543 676 2590 4993 3039 1486 6934 1112 2196 3876
- 5654 2327 4462 1402 2012 6929 6673 6203 5789 0 4392 4947 3648 3501 5274 5183 7216 6535 7140 5022 3621 5077 4090 922 3207 1131 7014 6714 2437 6707 6477 2476 5432 3599 1102 6376 7121 6284 2497 5160 4318 937 6795 6507 1268 5773 4249 1914
- 1713 4148 1184 5366 2573 2831 2380 2241 1602 4392 0 586 766 1029 883 2040 3353 2224 3100 1049 1246 1625 503 3841 1196 5054 3042 2488 2945 2676 2087 2331 1114 1650 5459 2132 3037 1958 1997 931 2513 3701 2923 2137 5459 1394 711 2534
- 1604 4723 1520 5946 3157 2266 1795 2051 1343 4947 586 0 1299 1612 406 2208 2824 1639 2542 694 1586 1767 1050 4357 1770 5633 2498 1907 3520 2128 1558 2778 531 2171 6003 1552 2472 1538 2506 791 2912 4277 2403 1564 5983 827 892 3109
- 2368 3635 1498 4679 1924 3407 3051 2920 2330 3648 766 1299 0 646 1642 2446 3840 2905 3655 1488 730 2096 697 3076 533 4363 3567 3122 2453 3219 2842 1592 1791 1480 4706 2772 3610 2721 1232 1656 2550 3001 3403 2860 4697 2126 756 1836
- 2201 3125 1103 4378 1580 3854 3405 2762 2291 3501 1029 1612 646 0 1853 2026 4349 3247 4119 1997 1341 1753 606 3078 419 4070 4052 3517 1923 3690 3032 1866 2142 838 4593 3161 4060 2788 1380 1663 1932 2736 3915 3138 4647 2395 1351 1592
- 1290 4907 1501 6225 3427 2178 1604 1687 970 5274 883 406 1642 1853 0 2029 2803 1438 2466 986 1987 1593 1253 4716 2072 5915 2454 1764 3710 2082 1204 3164 497 2287 6342 1419 2379 1134 2867 554 2885 4569 2405 1289 6338 555 1297 3406
- 1004 3930 951 5709 3179 4076 3382 1304 1451 5183 2040 2208 2446 2026 2029 0 4759 3220 4368 2900 3151 442 1765 4960 2444 5443 4396 3610 2932 4034 2572 3891 2525 1590 6278 3313 4261 2033 3398 1476 1241 4287 4390 2928 6419 2428 2749 3337
- 3833 7463 4298 8417 5749 727 1469 3932 3376 7216 3353 2824 3840 4349 2803 4759 0 1601 477 2359 3617 4345 3851 6433 4372 8098 370 1206 6267 726 2384 4754 2335 4991 8148 1452 609 2949 4752 3331 5687 6746 437 1948 8005 2334 3098 5618
- 2258 6338 2903 7578 4793 881 168 2331 1796 6535 2224 1639 2905 3247 1438 3220 1601 0 1165 1563 2988 2829 2666 5882 3401 7263 1233 399 5138 923 794 4227 1117 3724 7565 286 1049 1348 4051 1881 4248 5903 1322 355 7508 887 2302 4736
- 3419 7243 3967 8296 5577 293 1020 3487 2959 7140 3100 2542 3655 4119 2466 4368 477 1165 0 2170 3520 3965 3588 6393 4183 7977 202 767 6041 438 1932 4706 2027 4711 8107 1061 132 2503 4652 2972 5344 6617 486 1501 7989 1962 2939 5469
- 2267 5105 2169 6135 3409 1930 1681 2669 1951 5022 1049 694 1488 1997 986 2900 2359 1563 2170 0 1430 2460 1547 4333 2019 5817 2079 1694 3910 1733 1813 2668 654 2694 6029 1366 2130 1991 2525 1474 3542 4455 1923 1641 5957 1071 777 3302
- 2957 4043 2209 4802 2223 3310 3110 3487 2835 3621 1246 1586 730 1341 1987 3151 3617 2988 3520 1430 0 2779 1387 2905 1062 4482 3398 3119 2922 3087 3115 1240 1953 2175 4607 2796 3501 3119 1136 2173 3268 3136 3189 3029 4527 2355 711 2042
- 720 4022 652 5707 3066 3672 2993 1175 1112 5077 1625 1767 2096 1753 1593 442 4345 2829 3965 2460 2779 0 1401 4781 2166 5427 3984 3212 2946 3620 2224 3603 2089 1496 6178 2906 3861 1719 3132 1040 1479 4211 3969 2553 6290 2012 2336 3189
- 1700 3677 828 4982 2185 3315 2827 2260 1725 4090 503 1050 697 606 1253 1765 3851 2666 3588 1547 1387 1401 0 3621 903 4675 3537 2954 2475 3169 2427 2254 1578 1148 5177 2598 3521 2194 1833 1074 2054 3340 3423 2541 5213 1801 1077 2190
- 5279 2863 4136 2322 1860 6199 6009 5840 5346 922 3841 4357 3076 3078 4716 4960 6433 5882 6393 4333 2905 4781 3621 0 2718 2042 6254 6024 2569 5966 5913 1687 4807 3384 1716 5699 6384 5787 1852 4687 4285 1272 6022 5892 1629 5178 3581 1639
- 2578 3106 1518 4178 1401 3932 3552 3141 2628 3207 1196 1770 533 419 2072 2444 4372 3401 4183 2019 1062 2166 903 2718 0 3864 4097 3635 1932 3748 3274 1448 2284 1164 4286 3283 4136 3086 967 1973 2285 2507 3935 3331 4312 2589 1284 1340
- 6076 1850 4873 320 2491 7745 7412 6596 6285 1131 5054 5633 4363 4070 5915 5443 8098 7263 7977 5817 4482 5427 4675 2042 3864 0 7866 7483 2515 7539 7101 3449 6146 3938 1375 7134 7944 6831 3349 5709 4397 1363 7667 7190 1798 6446 5041 2528
- 3465 7173 3954 8186 5486 365 1104 3563 3007 7014 3042 2498 3567 4052 2454 4396 370 1233 202 2079 3398 3984 3537 6254 4097 7866 0 839 5973 374 2019 4569 1996 4669 7970 1085 305 2581 4532 2976 5339 6509 287 1581 7844 1974 2838 5369
- 2654 6630 3254 7800 5035 482 267 2728 2193 6714 2488 1907 3122 3517 1764 3610 1206 399 767 1694 3119 3212 2954 6024 3635 7483 839 0 5427 558 1181 4349 1377 4044 7723 356 653 1744 4218 2241 4614 6121 955 743 7644 1231 2465 4957
- 3625 1204 2446 2778 894 5774 5300 4120 3889 2437 2945 3520 2453 1923 3710 2932 6267 5138 6041 3910 2922 2946 2475 2569 1932 2515 5973 5427 0 5612 4824 2550 4050 1498 3476 5071 5980 4470 2096 3388 1911 1501 5831 4994 3704 4264 3209 1196
- 3115 6814 3581 7859 5141 261 821 3240 2661 6707 2676 2128 3219 3690 2082 4034 726 923 438 1733 3087 3620 3169 5966 3748 7539 374 558 5612 0 1716 4280 1624 4298 7679 735 420 2263 4216 2606 4967 6179 400 1277 7567 1609 2501 5032
- 1574 6001 2441 7408 4611 1659 916 1559 1122 6477 2087 1558 2842 3032 1204 2572 2384 794 1932 1813 3115 2224 2427 5913 3274 7101 2019 1181 4824 1716 0 4330 1180 3346 7545 1023 1808 578 4062 1438 3693 5763 2115 440 7537 763 2404 4603
- 3951 3447 2960 3763 1669 4513 4348 4507 3920 2476 2331 2778 1592 1866 3164 3891 4754 4227 4706 2668 1240 3603 2254 1687 1448 3449 4569 4349 2550 4280 4330 0 3184 2510 3402 4031 4698 4281 533 3245 3612 2187 4339 4265 3296 3576 1941 1381
- 1748 5253 1966 6461 3677 1746 1270 2082 1372 5432 1114 531 1791 2142 497 2525 2335 1117 2027 654 1953 2089 1578 4807 2284 6146 1996 1377 4050 1624 1180 3184 0 2685 6475 1022 1952 1341 2963 1050 3358 4787 1926 1086 6436 422 1244 3619
- 2142 2656 950 4223 1590 4431 3890 2658 2391 3599 1650 2171 1480 838 2287 1590 4991 3724 4711 2694 2175 1496 1148 3384 1164 3938 4669 4044 1498 4298 3346 2510 2685 0 4697 3693 4636 2975 1981 1909 1124 2718 4565 3548 4830 2839 2140 1751
- 6755 3123 5564 1427 3113 7910 7698 7304 6883 1102 5459 6003 4706 4593 6342 6278 8148 7565 8107 6029 4607 6178 5177 1716 4286 1375 7970 7723 3476 7679 7545 3402 6475 4697 0 7393 8097 7370 3515 6249 5379 2001 7738 7556 461 6829 5267 3013
- 2383 6274 2916 7451 4682 769 332 2512 1927 6376 2132 1552 2772 3161 1419 3313 1452 286 1061 1366 2796 2906 2598 5699 3283 7134 1085 356 5071 735 1023 4031 1022 3693 7393 0 965 1542 3883 1913 4286 5772 1121 600 7322 902 2128 4608
- 3306 7183 3878 8263 5533 207 900 3364 2845 7121 3037 2472 3610 4060 2379 4261 609 1049 132 2130 3501 3861 3521 6384 4136 7944 305 653 5980 420 1808 4698 1952 4636 8097 965 0 2380 4629 2877 5250 6583 570 1380 7986 1866 2904 5432
- 1029 5622 2035 7131 4352 2225 1484 985 611 6284 1958 1538 2721 2788 1134 2033 2949 1348 2503 1991 3119 1719 2194 5787 3086 6831 2581 1744 4470 2263 578 4281 1341 2975 7370 1542 2380 0 3952 1127 3197 5518 2658 1002 7395 951 2429 4380
- 3530 3085 2482 3669 1252 4435 4185 4091 3543 2497 1997 2506 1232 1380 2867 3398 4752 4051 4652 2525 1136 3132 1833 1852 967 3349 4532 4218 2096 4216 4062 533 2963 1981 3515 3883 4629 3952 0 2873 3080 2012 4324 4046 3478 3328 1755 1000
- 825 4564 1027 6011 3227 2681 2049 1319 676 5160 931 791 1656 1663 554 1476 3331 1881 2972 1474 2173 1040 1074 4687 1973 5709 2976 2241 3388 2606 1438 3245 1050 1909 6249 1913 2877 1127 2873 0 2374 4392 2943 1659 6285 1012 1563 3254
- 2188 2756 1395 4638 2426 5053 4415 2544 2590 4318 2513 2912 2550 1932 2885 1241 5687 4248 5344 3542 3268 1479 2054 4285 2285 4397 5339 4614 1911 4967 3693 3612 3358 1124 5379 4286 5250 3197 3080 2374 0 3386 5284 3997 5585 3386 3125 2664
- 4820 1591 3617 1681 1169 6384 6051 5358 4993 937 3701 4277 3001 2736 4569 4287 6746 5903 6617 4455 3136 4211 3340 1272 2507 1363 6509 6121 1501 6179 5763 2187 4787 2718 2001 5772 6583 5518 2012 4392 3386 0 6314 5837 2205 5095 3680 1169
- 3489 7027 3891 7987 5313 550 1219 3632 3039 6795 2923 2403 3403 3915 2405 4390 437 1322 486 1923 3189 3969 3423 6022 3935 7667 287 955 5831 400 2115 4339 1926 4565 7738 1121 570 2658 4324 2943 5284 6314 0 1676 7603 1964 2662 5184
- 1947 6186 2686 7502 4706 1224 482 1987 1486 6507 2137 1564 2860 3138 1289 2928 1948 355 1501 1641 3029 2553 2541 5892 3331 7190 1581 743 4994 1277 440 4265 1086 3548 7556 600 1380 1002 4046 1659 3997 5837 1676 0 7521 744 2325 4670
- 6835 3472 5661 1877 3241 7805 7635 7391 6934 1268 5459 5983 4697 4647 6338 6419 8005 7508 7989 5957 4527 6290 5213 1629 4312 1798 7844 7644 3704 7567 7537 3296 6436 4830 461 7322 7986 7395 3478 6285 5585 2205 7603 7521 0 6805 5208 3102
- 1542 5461 2023 6758 3962 1670 1054 1785 1112 5773 1394 827 2126 2395 555 2428 2334 887 1962 1071 2355 2012 1801 5178 2589 6446 1974 1231 4264 1609 763 3576 422 2839 6829 902 1866 951 3328 1012 3386 5095 1964 744 6805 0 1644 3928
- 2379 4390 1867 5360 2651 2704 2432 2879 2196 4249 711 892 756 1351 1297 2749 3098 2302 2939 777 711 2336 1077 3581 1284 5041 2838 2465 3209 2501 2404 1941 1244 2140 5267 2128 2904 2429 1755 1563 3125 3680 2662 2325 5208 1644 0 2532
- 3744 2088 2560 2844 304 5230 4884 4296 3876 1914 2534 3109 1836 1592 3406 3337 5618 4736 5469 3302 2042 3189 2190 1639 1340 2528 5369 4957 1196 5032 4603 1381 3619 1751 3013 4608 5432 4380 1000 3254 2664 1169 5184 4670 3102 3928 2532 0
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/p01_d.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/p01_d.txt
deleted file mode 100644
index 0464ad3143..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/p01_d.txt
+++ /dev/null
@@ -1,15 +0,0 @@
- 0 29 82 46 68 52 72 42 51 55 29 74 23 72 46
- 29 0 55 46 42 43 43 23 23 31 41 51 11 52 21
- 82 55 0 68 46 55 23 43 41 29 79 21 64 31 51
- 46 46 68 0 82 15 72 31 62 42 21 51 51 43 64
- 68 42 46 82 0 74 23 52 21 46 82 58 46 65 23
- 52 43 55 15 74 0 61 23 55 31 33 37 51 29 59
- 72 43 23 72 23 61 0 42 23 31 77 37 51 46 33
- 42 23 43 31 52 23 42 0 33 15 37 33 33 31 37
- 51 23 41 62 21 55 23 33 0 29 62 46 29 51 11
- 55 31 29 42 46 31 31 15 29 0 51 21 41 23 37
- 29 41 79 21 82 33 77 37 62 51 0 65 42 59 61
- 74 51 21 51 58 37 37 33 46 21 65 0 61 11 55
- 23 11 64 51 46 51 51 33 29 41 42 61 0 62 23
- 72 52 31 43 65 29 46 31 51 23 59 11 62 0 59
- 46 21 51 64 23 59 33 37 11 37 61 55 23 59 0
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/p01_s.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/p01_s.txt
deleted file mode 100644
index 38afab553d..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/p01_s.txt
+++ /dev/null
@@ -1,16 +0,0 @@
- 1
-13
- 2
-15
- 9
- 5
- 7
- 3
-12
-14
-10
- 8
- 6
- 4
-11
- 1
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/tsp.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/tsp.py
deleted file mode 100755
index 0ddff51564..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tsp/tsp.py
+++ /dev/null
@@ -1,59 +0,0 @@
-#!/usr/bin/env python
-#
-# This is a simple testcase purely for testing the autotuner on permutations
-#
-# http://en.wikipedia.org/wiki/Travelling_salesman_problem
-#
-
-import adddeps #fix sys.path
-
-import argparse
-import logging
-
-import opentuner
-from opentuner.search.manipulator import (ConfigurationManipulator,
- PermutationParameter)
-from opentuner.search.objective import MinimizeTime
-from opentuner.measurement import MeasurementInterface
-from opentuner.measurement.inputmanager import FixedInputManager
-from opentuner.tuningrunmain import TuningRunMain
-
-
-parser = argparse.ArgumentParser(parents=opentuner.argparsers())
-parser.add_argument('data', help='distance matrix file')
-
-class TSP(MeasurementInterface):
- def __init__(self, args):
- super(TSP, self).__init__(args)
- data = args.data
- m = open(data).readlines()
- self.distance = [[int(i) for i in l.split()] for l in m]
-
- def run(self, desired_result, input, limit):
- cfg = desired_result.configuration.data
- p = cfg[0] # cheating: should use manipulator function
- t = self.eval_path(p)
- return opentuner.resultsdb.models.Result(time=t)
-
- def eval_path(self, p):
- """ Given permutation of cities as a list of indices,
- return total path length """
- out = sum(self.distance[p[i]][p[i+1]] for i in range(len(p)-1))
-## print out, p
- return out
-
- def manipulator(self):
- manipulator = ConfigurationManipulator()
- manipulator.add_parameter(PermutationParameter(0, range(len(self.distance))))
- return manipulator
-
- def solution(self):
- p = [1,13,2,15,9,5,7,3,12,14,10,8,6,4,11]
- return self.eval_path(p)
-
-
-
-if __name__ == '__main__':
- args = parser.parse_args()
- TSP.main(args)
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/#accuracy_tuner.py# b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/#accuracy_tuner.py#
deleted file mode 100644
index 2110d0d692..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/#accuracy_tuner.py#
+++ /dev/null
@@ -1,203 +0,0 @@
-#!/usr/bin/env python
-#
-# Optimize blocksize of apps/mmm_block.cpp
-#
-# This is an extremely simplified version meant only for tutorials
-#
-import adddeps # fix sys.path
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-
-
-output_dir = ""
-flag_ranges = []
-tuning_flags = []
-binary_name = ""
-accuracy_threshold = 10.0
-opt_confs_index = 9
-evaluated_configs = {}
-
-
-def extractTotalOverhead(file_name):
-
- total_comps = 0.0
- file = open(file_name, "r")
- for x in file:
- words = x.split()
- total_comps += float(words[opt_confs_index])
-
- print total_comps
- return total_comps
-
-
-def getAccuracy(file_name):
-
- file = open(file_name, "r")
- acc_str = file.read()
- accuracy = float(acc_str)
- print accuracy
- return accuracy
-
-
-def createFlagsFile(file_name, cfg):
-
- f = open(file_name, "w+")
- cmd_config = ""
- for flag in tuning_flags:
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-class ClangFlagsTuner(MeasurementInterface):
-
- def __init__(self, args):
- objective = ThresholdAccuracyMinimizeTime(accuracy_threshold)
- input_manager = FixedInputManager(size=num_flags)
- self.configs_list = []
-
- super(ClangFlagsTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- for flag in tuning_flags:
- manipulator.add_parameter(
- EnumParameter(flag, flag_ranges
- # [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
- )) #default is needed, optimizations don't work without it(tried and tested)
- return manipulator
-
-
- def run(self, desired_result, input, limit):
- """
- Compile and run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
-
- # NOTE: creates the file with flags read by the runtime
- createFlagsFile("opentuner_flags", cfg)
-
- run_cmd = binary_name
- print run_cmd
- run_result_call_program = self.call_program(run_cmd)
- #print run_result_call_program
-
- total_comps = extractTotalOverhead("accuracy_summary")
- accuracy = getAccuracy("final_accuracy")
-
- #Result = opentuner.resultsdb.models.Result(time=total_comps)
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- Result.accuracy = accuracy
-
- if accuracy > accuracy_threshold:
- if accuracy not in evaluated_configs:
- config_tuple = (total_comps, accuracy, cfg)
- self.configs_list.append(config_tuple)
- evaluated_configs[accuracy] = 1
- shutil.copy('accuracy_summary', output_dir + '/' + binary_name + '_' + str(accuracy))
-
-
- return Result
-
-
- def save_final_config(self, configuration):
- """
- called at the end of autotuning with the best resultsdb.models.Configuration
- """
- print "Final configuration", configuration.data
-
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- createFlagsFile("opentuner_flags", configuration.data)
- run_cmd = binary_name
- run_result_call_program = self.call_program(run_cmd)
-
- accuracy = getAccuracy("final_accuracy")
- shutil.copy('accuracy_summary', result_dir + '/' + binary_name + '_final_' + str(accuracy) )
-
- sorted_list = sorted(self.configs_list, key = lambda tup: tup[0])
- print sorted_list[0:10]
-
- top_elems = 20
- if len(sorted_list) < top_elems:
- top_elems = len(sorted_list)
-
-
- for i in range(top_elems):
- createFlagsFile("opentuner_flags", sorted_list[i][2])
- run_cmd = binary_name
- run_result_call_program = self.call_program(run_cmd)
- accuracy = getAccuracy("final_accuracy")
- shutil.copy('accuracy_summary', result_dir + '/' + binary_name + '_' + str(accuracy) + "_rank_" + str(i) )
-
-
- #os.mkdir(result_dir + "full_results")
-
-
-
-
-if __name__ == '__main__':
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='name of binary to run')
- argparser.add_argument('--num-flags', type=int, help='num of flags to tune for')
- argparser.add_argument('--error-range', type=int, help='num of flags to tune for')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='accuracy threshold')
-
-
- args = argparser.parse_args()
- binary_name = str(args.binary)
- print("binary_name = ", binary_name)
- num_flags = int(args.num_flags)
- error_range = int(args.error_range)
- accuracy_threshold = float(args.accuracy)
- print("accuracy = ", accuracy_threshold)
- result_dir = args.result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
-
- output_dir = result_dir + "/full_results"
- print output_dir
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- if not os.path.exists(output_dir):
- print("Creating output directory = ", output_dir)
- os.mkdir(output_dir)
-
- for j in range(error_range):
- flag_ranges.append(j)
-
- print("flag_ranges = ", flag_ranges)
-
- for i in range(num_flags):
- tuning_flags.append("flag" + str(i))
-
- ClangFlagsTuner.main(argparser.parse_args())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/__init__.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/__init__.py
deleted file mode 100644
index e69de29bb2..0000000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/accuracy_tuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/accuracy_tuner.py
deleted file mode 100644
index 5977fe7ee5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/accuracy_tuner.py
+++ /dev/null
@@ -1,198 +0,0 @@
-#!/usr/bin/env python
-#
-
-import adddeps # fix sys.path
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-
-
-output_dir = ""
-flag_ranges = []
-tuning_flags = []
-binary_name = ""
-accuracy_threshold = 10.0
-opt_confs_index = 9
-evaluated_configs = {}
-
-
-def extractTotalOverhead(file_name):
-
- total_comps = 0.0
- file = open(file_name, "r")
- for x in file:
- words = x.split()
- total_comps += float(words[opt_confs_index])
-
- print total_comps
- return total_comps
-
-
-def getAccuracy(file_name):
-
- file = open(file_name, "r")
- acc_str = file.read()
- accuracy = float(acc_str)
- print accuracy
- return accuracy
-
-
-def createFlagsFile(file_name, cfg):
-
- f = open(file_name, "w+")
- cmd_config = ""
- for flag in tuning_flags:
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-class ClangFlagsTuner(MeasurementInterface):
-
- def __init__(self, args):
- objective = ThresholdAccuracyMinimizeTime(accuracy_threshold)
- input_manager = FixedInputManager(size=num_flags)
- self.configs_list = []
-
- super(ClangFlagsTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- for flag in tuning_flags:
- manipulator.add_parameter(
- EnumParameter(flag, flag_ranges
- # [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
- )) #default is needed, optimizations don't work without it(tried and tested)
- return manipulator
-
-
- def run(self, desired_result, input, limit):
- """
- Compile and run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
-
- # NOTE: creates the file with flags read by the runtime
- createFlagsFile("opentuner_flags", cfg)
-
- run_cmd = binary_name
- print run_cmd
- run_result_call_program = self.call_program(run_cmd)
- #print run_result_call_program
-
- total_comps = extractTotalOverhead("accuracy_summary")
- accuracy = getAccuracy("final_accuracy")
-
- #Result = opentuner.resultsdb.models.Result(time=total_comps)
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- Result.accuracy = accuracy
-
- if accuracy > accuracy_threshold:
- if accuracy not in evaluated_configs:
- config_tuple = (total_comps, accuracy, cfg)
- self.configs_list.append(config_tuple)
- evaluated_configs[accuracy] = 1
- shutil.copy('accuracy_summary', output_dir + '/' + binary_name + '_' + str(accuracy))
-
-
- return Result
-
-
- def save_final_config(self, configuration):
- """
- called at the end of autotuning with the best resultsdb.models.Configuration
- """
- print "Final configuration", configuration.data
-
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- createFlagsFile("opentuner_flags", configuration.data)
- run_cmd = binary_name
- run_result_call_program = self.call_program(run_cmd)
-
- accuracy = getAccuracy("final_accuracy")
- shutil.copy('accuracy_summary', result_dir + '/' + binary_name + '_final_' + str(accuracy) )
-
- sorted_list = sorted(self.configs_list, key = lambda tup: tup[0])
- print sorted_list[0:10]
-
- top_elems = 20
- if len(sorted_list) < top_elems:
- top_elems = len(sorted_list)
-
-
- for i in range(top_elems):
- createFlagsFile("opentuner_flags", sorted_list[i][2])
- run_cmd = binary_name
- run_result_call_program = self.call_program(run_cmd)
- accuracy = getAccuracy("final_accuracy")
- shutil.copy('accuracy_summary', result_dir + '/' + binary_name + '_' + str(accuracy) + "_rank_" + str(i) )
-
-
- #os.mkdir(result_dir + "full_results")
-
-
-
-
-if __name__ == '__main__':
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='name of binary to run')
- argparser.add_argument('--num-flags', type=int, help='num of flags to tune for')
- argparser.add_argument('--error-range', type=int, help='num of flags to tune for')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='accuracy threshold')
-
-
- args = argparser.parse_args()
- binary_name = str(args.binary)
- print("binary_name = ", binary_name)
- num_flags = int(args.num_flags)
- error_range = int(args.error_range)
- accuracy_threshold = float(args.accuracy)
- print("accuracy = ", accuracy_threshold)
- result_dir = args.result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
-
- output_dir = result_dir + "/full_results"
- print output_dir
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- if not os.path.exists(output_dir):
- print("Creating output directory = ", output_dir)
- os.mkdir(output_dir)
-
- for j in range(error_range):
- flag_ranges.append(j)
-
- print("flag_ranges = ", flag_ranges)
-
- for i in range(num_flags):
- tuning_flags.append("flag" + str(i))
-
- ClangFlagsTuner.main(argparser.parse_args())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/accuracy_tuner_piped.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/accuracy_tuner_piped.py
deleted file mode 100644
index 6d46c5762e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/accuracy_tuner_piped.py
+++ /dev/null
@@ -1,269 +0,0 @@
-#!/usr/bin/env python
-#
-# Optimize blocksize of apps/mmm_block.cpp
-#
-# This is an extremely simplified version meant only for tutorials
-#
-import adddeps # fix sys.path
-
-import argparse
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-from opentuner.search.objective import ThresholdAccuracyMinimizeTime
-from opentuner.measurement.inputmanager import FixedInputManager
-import shutil
-import os
-import sys
-import subprocess
-import threading
-import psutil
-
-from measure_confidence import dump_high_confidence_files
-from select_top_results import select_top_results
-from time import sleep
-
-
-output_dir = ""
-flag_ranges = []
-tuning_flags = []
-binary_name = ""
-accuracy_threshold = 10.0
-opt_confs_index = 9
-evaluated_configs = {}
-orig_result_dir = ""
-
-
-def extractTotalOverhead(file_name):
-
- total_comps = 0.0
- file = open(file_name, "r")
- for x in file:
- words = x.split()
- total_comps += float(words[opt_confs_index])
-
- print total_comps
- return total_comps
-
-
-def getAccuracy(file_name):
-
- file = open(file_name, "r")
- acc_str = file.read()
- file.close()
-
- accuracy = float(acc_str)
-
- try:
- accuracy = float(acc_str)
- except:
- return 20
-
- print accuracy
- return accuracy
-
-
-
-def kill(proc_pid):
- process = psutil.Process(proc_pid)
- for proc in process.children(recursive=True):
- proc.kill()
- process.kill()
-
-
-
-def createFlagsFile(file_name, cfg):
-
- f = open(file_name, "w+")
- cmd_config = ""
- for flag in tuning_flags:
- flag_value = cfg[flag]
- cmd_config += str(flag_value) + "\n"
-
- f.write(cmd_config)
- f.close()
-
-
-class ClangFlagsTuner(MeasurementInterface):
-
- def __init__(self, args):
- objective = ThresholdAccuracyMinimizeTime(accuracy_threshold)
- input_manager = FixedInputManager(size=num_flags)
- self.configs_list = []
-
- super(ClangFlagsTuner, self).__init__(
- args, program_name=args.binary,
- program_version=self.file_hash(args.binary),
- input_manager=input_manager, objective=objective)
-
-
- FNULL = open(os.devnull, 'wb')
- #run_result_call_program = self.call_program(run_cmd)
- self.start_process = subprocess.Popen([binary_name, "opentuner_run"]) #, stdout=FNULL);
-
- try:
- os.mkfifo("/tmp/myfifo")
- except OSError, e:
- print("FIFO exists")
-
-
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- for flag in tuning_flags:
- manipulator.add_parameter(
- EnumParameter(flag, flag_ranges
- # [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
- )) #default is needed, optimizations don't work without it(tried and tested)
- return manipulator
-
-
- def run(self, desired_result, input, limit):
-
- """
- Run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
-
- # NOTE: creates the file with flags read by the runtime
- createFlagsFile("opentuner_flags", cfg)
-
- run_cmd = binary_name
- print run_cmd
- #run_result_call_program = self.call_program(run_cmd)
-
- # Using Named Pipes to signal execution to the DNN outer thread
- fifo = open("/tmp/myfifo", "w")
- fifo.write("start_run")
- fifo.close()
-
- print "Waiting for process to signal back - when done processing one run"
-
- fifo2 = open("/tmp/myfifo", "r")
- fifo2.read()
- fifo2.close()
-
- print "Process Signalled back"
-
- total_comps = extractTotalOverhead("accuracy_summary")
- accuracy = getAccuracy("final_accuracy")
-
-
- #Result = opentuner.resultsdb.models.Result(time=total_comps)
- Result = opentuner.resultsdb.models.Result()
- Result.time = total_comps
- Result.accuracy = accuracy
-
- if accuracy > accuracy_threshold:
- if accuracy not in evaluated_configs:
- config_tuple = (total_comps, accuracy, cfg)
- self.configs_list.append(config_tuple)
- evaluated_configs[accuracy] = 1
- shutil.copy('accuracy_summary', output_dir + '/' + binary_name + '_' + str(accuracy))
-
-
- print "done with one run"
-
- return Result
-
-
- def save_final_config(self, configuration):
-
- print "Dumping High Confidence results"
- sleep(5)
-
- # Only dumping files with 95% confidence
- dump_high_confidence_files(binary_name, orig_result_dir, accuracy_threshold, 95)
- select_top_results(orig_result_dir + "/high_confidence")
-
-
- #self.start_process.kill()
- kill(self.start_process.pid)
-
- """
- called at the end of autotuning with the best resultsdb.models.Configuration
- """
- print "Final configuration", configuration.data
-
- return
-
-
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- createFlagsFile("opentuner_flags", configuration.data)
- run_cmd = binary_name
- run_result_call_program = self.call_program(run_cmd)
-
- accuracy = getAccuracy("final_accuracy")
- shutil.copy('accuracy_summary', result_dir + '/' + binary_name + '_final_' + str(accuracy) )
-
- sorted_list = sorted(self.configs_list, key = lambda tup: tup[0])
- print sorted_list[0:10]
-
- top_elems = 20
- if len(sorted_list) < top_elems:
- top_elems = len(sorted_list)
-
-
- for i in range(top_elems):
- createFlagsFile("opentuner_flags", sorted_list[i][2])
- run_cmd = binary_name
- run_result_call_program = self.call_program(run_cmd)
- accuracy = getAccuracy("final_accuracy")
- shutil.copy('accuracy_summary', result_dir + '/' + binary_name + '_' + str(accuracy) + "_rank_" + str(i) )
-
-
- #os.mkdir(result_dir + "full_results")
-
-
-
-
-if __name__ == '__main__':
-
- argparser = argparse.ArgumentParser(parents=opentuner.argparsers())
- argparser.add_argument('--binary', help='name of binary to run')
- argparser.add_argument('--num-flags', type=int, help='num of flags to tune for')
- argparser.add_argument('--error-range', type=int, help='num of flags to tune for')
- argparser.add_argument('--accuracy', type=float, help='accuracy threshold')
- argparser.add_argument('--result-dir', help='accuracy threshold')
-
-
- args = argparser.parse_args()
- binary_name = str(args.binary)
- print("binary_name = ", binary_name)
- num_flags = int(args.num_flags)
- error_range = int(args.error_range)
- accuracy_threshold = float(args.accuracy)
- print("accuracy = ", accuracy_threshold)
- result_dir = args.result_dir
- orig_result_dir = result_dir
- if result_dir == "":
- print("Provide --result-dir ")
-
-
- output_dir = result_dir + "/full_results"
- print output_dir
- if not os.path.exists(result_dir):
- os.mkdir(result_dir)
-
- if not os.path.exists(output_dir):
- print("Creating output directory = ", output_dir)
- os.mkdir(output_dir)
-
- for j in range(error_range):
- flag_ranges.append(j)
-
- print("flag_ranges = ", flag_ranges)
-
- for i in range(num_flags):
- tuning_flags.append("flag" + str(i))
-
- ClangFlagsTuner.main(argparser.parse_args())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/adddeps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/adddeps.py
deleted file mode 100644
index 72de04cf55..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/adddeps.py
+++ /dev/null
@@ -1,5 +0,0 @@
-# we would prefer a symbolic link, but it does not work on windows
-import os
-target = os.path.join(os.path.dirname(__file__),
- '../../opentuner/utils/adddeps.py')
-execfile(target, dict(__file__=target))
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/gettingstarted.md b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/gettingstarted.md
deleted file mode 100644
index 8a442c5f44..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/gettingstarted.md
+++ /dev/null
@@ -1,215 +0,0 @@
----
-layout: default
-title: OpenTuner - Using OpenTuner
-permalink: /tutorial/gettingstarted/index.html
----
-
-Tutorial: Optimizing Block Matrix Multiplication
-================================================
-
-This tutorial assumes that you have checked out a copy of opentuner. For
-guidelines on how to get opentuner set up, refer [here][setup].
-
-[setup]: http://opentuner.org/tutorial/setup/
-
-Identifying a Program to Autotune
----------------------------------
-
-In order to do autotuning, you first need something to autotune. This will
-normally be your own program that you want to make either fast or better in
-some way. For this tutorial we will use a blocked version of matrix multiply
-as an example. We will use opentuner to find the optimal value of the block
-size parameter.
-
-We will autotune the sample code below(based off of modification of code
-found [here][matrix-multiply-code]), making sure to take the block size as
-a compile time constant to the program.
-
-[matrix-multiply-code]: http://csapp.cs.cmu.edu/public/waside/waside-blocking.pdf
-
-Save the sample code below to examples/tutorials/mmm_block.cpp
-
- #include <stdio.h>
- #include <cstdlib>
-
- #define N 100
-
- int main(int argc, const char** argv)
- {
-
- int n = BLOCK_SIZE * (N/BLOCK_SIZE);
- int a[N][N];
- int b[N][N];
- int c[N][N];
- int sum=0;
- for(int k1=0;k1<n;k1+=BLOCK_SIZE)
- {
- for(int j1=0;j1<n;j1+=BLOCK_SIZE)
- {
- for(int k1=0;k1<n;k1+=BLOCK_SIZE)
- {
- for(int i=0;i<n;i++)
- {
- for(int j=j1;j<j1+BLOCK_SIZE;j++)
- {
- sum = c[i][j];
- for(int k=k1;k<k1+BLOCK_SIZE;k++)
- {
- sum += a[i][k] * b[k][j];
- }
- c[i][j] = sum;
- }
- }
- }
- }
- }
- return 0;
- }
-
-Creating a New Autotuner with Opentuner
-------------------------------------
-Now we need to create a program that uses OpenTuner to optimize the program we just saved.
-
-Save the following code to examples/tutorials/mmm_tuner.py
-
- #!/usr/bin/env python
- #
- # Optimize blocksize of apps/mmm_block.cpp
- #
- # This is an extremely simplified version meant only for tutorials
- #
- import adddeps # fix sys.path
-
- import opentuner
- from opentuner import ConfigurationManipulator
- from opentuner import IntegerParameter
- from opentuner import MeasurementInterface
- from opentuner import Result
-
-
- class GccFlagsTuner(MeasurementInterface):
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- manipulator.add_parameter(
- IntegerParameter('blockSize', 1, 10))
- return manipulator
-
- def run(self, desired_result, input, limit):
- """
- Compile and run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
-
- gcc_cmd = 'g++ mmm_block.cpp '
- gcc_cmd += '-DBLOCK_SIZE='+ cfg['blockSize']
- gcc_cmd += ' -o ./tmp.bin'
-
- compile_result = self.call_program(gcc_cmd)
- assert compile_result['returncode'] == 0
-
- run_cmd = './tmp.bin'
-
- run_result = self.call_program(run_cmd)
- assert run_result['returncode'] == 0
-
- return Result(time=run_result['time'])
-
- def save_final_config(self, configuration):
- """called at the end of tuning"""
- print "Optimal block size written to mmm_final_config.json:", configuration.data
- self.manipulator().save_to_file(configuration.data,
- 'mmm_final_config.json')
-
-
- if __name__ == '__main__':
- argparser = opentuner.default_argparser()
- GccFlagsTuner.main(argparser.parse_args())
-
-
-This file consists of several components, each of which will be discussed in further detail below.
-
-Tuning Programs have a general structure as follows:
-
- from opentuner import MeasurementInterface
- from opentuner import Result
-
-Create an instance of class GccFlagsTuner, which tunes specified parameters using opentuner.
- class GccFlagsTuner(MeasurementInterface):
-
-The manipulator method defines the variable search space by specifying parameters that should be tuned by this instance of GccFlagsTuner
-
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- manipulator.add_parameter(
- IntegerParameter('blockSize', 1, 10))
- return manipulator
-
-The run method actually runs opentuner under the given configuration and returns the calculated performance under this configuration. In this example, the blockSize parameter to be tuned is input as a compile-time constant that takes on a value within the specified range each time it is run. However, opentuner also supports other methods of specifying these parameters that may be preferred in different use cases.
-
- def run(self, desired_result, input, limit):
- """
- Compile and run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
-
- gcc_cmd = 'g++ mmm_block.cpp '
- gcc_cmd += '-DBLOCK_SIZE='+ cfg['blockSize']
- gcc_cmd += ' -o ./tmp.bin'
-
- compile_result = self.call_program(gcc_cmd)
- assert compile_result['returncode'] == 0
-
- run_cmd = './tmp.bin'
-
- run_result = self.call_program(run_cmd)
- assert run_result['returncode'] == 0
-
- return Result(time=run_result['time'])
-
-We can actually display the result of running opentuner(the optimal block size for our multiplication problem) by creating a method, save_final_config() in our class. This saves a json dictionary of the optimal blockSize parameter found to the file mmm_final_config.json
-
- def save_final_config(self, configuration):
- """called at the end of tuning"""
- print "Optimal block size written to mmm_final_config.json:", configuration.data
- self.manipulator().save_to_file(configuration.data,
- 'mmm_final_config.json')
-
- if __name__ == '__main__':
- argparser = opentuner.default_argparser()
- GccFlagsTuner.main(argparser.parse_args())
-
-Generating and Viewing Results
-------------------------------
-
-Run the following command to autotune our program(The --no-dups flag hides warnings about duplicate results and the --stop-after parameter specifies that we are running opentuner for a maximum of 30 seconds):
-
- python mmm_tuner.py --no-dups --stop-after=30
-
-The results of each run configuration will be displayed as follows(output lines are truncated for readability here):
-
- [ 10s] INFO opentuner.search.plugin.DisplayPlugin: tests=10, best {'BLOCK_SIZE': 4}, cost time=0.0081, found by DifferentialEvolutionAlt[...]
- [ 19s] INFO opentuner.search.metatechniques: AUCBanditMetaTechniqueA: [('DifferentialEvolutionAlt', 477), ('UniformGreedyMutation', 18), ('NormalGreedyMutation', 5), ('RandomNelderMead', 1)]
- [ 20s] INFO opentuner.search.plugin.DisplayPlugin: tests=10, best {'BLOCK_SIZE': 4}, cost time=0.0081, found by DifferentialEvolutionAlt[...]
- [ 30s] INFO opentuner.search.plugin.DisplayPlugin: tests=10, best {'BLOCK_SIZE': 4}, cost time=0.0081, found by DifferentialEvolutionAlt[...]
- [ 30s] INFO opentuner.search.plugin.DisplayPlugin: tests=10, best {'BLOCK_SIZE': 4}, cost time=0.0081, found by DifferentialEvolutionAlt[...]
- Optimal block size written to mmm_final_config.json: {'BLOCK_SIZE': 4}
-
-
-Look up the optimal BlockSize value by inspecting the following created file:
-
- mmm_final_config.json
-
-In this example, the output file content was as follows:
-
- {'BLOCK_SIZE': 4}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/measure_confidence.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/measure_confidence.py
deleted file mode 100644
index 655bdb024f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/measure_confidence.py
+++ /dev/null
@@ -1,164 +0,0 @@
-
-import argparse
-import os
-import sys
-
-
-def getAccuracy(file_name):
-
- if not os.path.exists(file_name):
- print("final_accuracy file not found ")
- sys.exit(0)
-
- file = open(file_name, "r")
- acc_str = file.read()
- accuracy = float(acc_str)
- print accuracy
- return accuracy
-
-
-total_runs = 100.0
-skip_lines = 0
-
-
-def test_func():
- print "test_func"
- sys.exit(0)
-
-
-def do_multiple_runs(binary_name, accuracy_threshold, confidence_threshold):
-
- #total_runs = 100.0
- successful_runs = 0.0
- total_acc = 0
-
- for i in range(int(total_runs)):
-
- fifo = open("/tmp/myfifo", "w")
- fifo.write("start_run")
- fifo.close()
-
- print "Waiting for process to signal back - when done processing one run"
-
- fifo2 = open("/tmp/myfifo", "r")
- fifo2.read()
- fifo2.close()
-
- print "Process Signalled back"
-
- accuracy = getAccuracy("final_accuracy")
- total_acc += accuracy
-
- if accuracy > accuracy_threshold:
- successful_runs += 1
-
- confidence = (successful_runs / total_runs) * 100.0
- print("confidence = ", confidence)
- avg_acc = total_acc / total_runs
- print("average accuracy = ", avg_acc)
-
- return confidence, avg_acc
-
-
-def compute_confidence(binary_name, accuracy, confidence, result_dir, output_dir):
-
- confidence_list = []
-
- if not os.path.exists(result_dir):
- print("Path does not exist")
- sys.exit(0)
-
- file_names = os.listdir(result_dir)
- print file_names
-
- for file_name in file_names:
- # Skip sub-directories
- if os.path.isdir(result_dir + "/" + file_name):
- continue
-
- f = open(result_dir + "/" + file_name)
- tuner_file = open("opentuner_flags", "w+")
-
- index = 0
- results_str = ""
- for x in f:
- if index >= skip_lines:
- error_knob = int(float(x.split()[1]))
- print error_knob
- tuner_file.write(str(error_knob) + "\n")
-
- results_str += x
- index += 1
-
- tuner_file.close()
-
- run_confidence, avg_accuracy = do_multiple_runs(binary_name, accuracy, confidence)
-
- if run_confidence > 90:
- f2 = open(output_dir + "/" + file_name, "w+")
- f2.write("total_runs=" + str(total_runs) + "\t confidence=" + str(run_confidence) + "\t avg_accuracy=" + str(avg_accuracy) + "\n")
- f2.write(results_str)
- f2.close()
-
- conf_result = (run_confidence, avg_accuracy, file_name)
- confidence_list.append(conf_result)
-
- return confidence_list
-
-
-
-def dump_high_confidence_files(binary, result_dir, accuracy, confidence):
-
- #result_dir = args.result_dir
- output_dir = result_dir + "/high_confidence"
- result_dir = result_dir + "/full_results"
-
- if not os.path.exists(output_dir):
- os.mkdir(output_dir)
-
-
- confidence_list = compute_confidence(binary, accuracy, confidence, result_dir, output_dir)
- print confidence_list
-
- sorted_list = sorted(confidence_list, key = lambda tup: tup[0], reverse=True)
-
- output_file = open(output_dir + "/confidence_summary.txt", "w+")
- for x in sorted_list:
- output_file.write(str(x[0]) + "\t" + str(x[1]) + "\t" + str(x[2]) + "\n")
-
- output_file.close()
- print "Dumped Confidence Summary"
-
-
-
-
-
-
-if __name__ == "__main__":
-
- argparser = argparse.ArgumentParser(description='runs best configs to get high confidence on accuracy')
- argparser.add_argument('--result-dir', help='Directory containing OpenTuner configurations')
- argparser.add_argument('--output-dir', help='Directory for storing output directory')
- argparser.add_argument('--binary', help='Binary name to run')
- argparser.add_argument('--accuracy', type=float, help='Accuracy constraint')
- argparser.add_argument('--confidence', type=float, help='Confidence threshold')
-
-
- args = argparser.parse_args()
- result_dir = args.result_dir
- output_dir = args.output_dir
- binary = args.binary
- accuracy = args.accuracy
- confidence = args.confidence
-
- confidence_list = compute_confidence(binary, accuracy, confidence, result_dir, output_dir)
- #print confidence_list
-
- sorted_list = sorted(confidence_list, key = lambda tup: tup[0], reverse=True)
-
- output_file = open(output_dir + "/confidence_summary.txt", "w+")
- for x in sorted_list:
- output_file.write(str(x[0]) + "\t" + str(x[1]) + "\t" + str(x[2]) + "\n")
-
- output_file.close()
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/mmm_block.cpp b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/mmm_block.cpp
deleted file mode 100755
index 0bb76845f8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/mmm_block.cpp
+++ /dev/null
@@ -1,36 +0,0 @@
-#include <stdio.h>
-#include <cstdlib>
-
-#define N 100
-
-int main(int argc, const char** argv)
-{
-
- int n = BLOCK_SIZE * (N/BLOCK_SIZE);
- int a[N][N];
- int b[N][N];
- int c[N][N];
- int sum=0;
- for(int k1=0;k1<n;k1+=BLOCK_SIZE)
- {
- for(int j1=0;j1<n;j1+=BLOCK_SIZE)
- {
- for(int k1=0;k1<n;k1+=BLOCK_SIZE)
- {
- for(int i=0;i<n;i++)
- {
- for(int j=j1;j<j1+BLOCK_SIZE;j++)
- {
- sum = c[i][j];
- for(int k=k1;k<k1+BLOCK_SIZE;k++)
- {
- sum += a[i][k] * b[k][j];
- }
- c[i][j] = sum;
- }
- }
- }
- }
- }
- return 0;
-}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/mmm_tuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/mmm_tuner.py
deleted file mode 100644
index f92c4c3bfc..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/mmm_tuner.py
+++ /dev/null
@@ -1,116 +0,0 @@
-#!/usr/bin/env python
-#
-# Optimize blocksize of apps/mmm_block.cpp
-#
-# This is an extremely simplified version meant only for tutorials
-#
-import adddeps # fix sys.path
-
-import opentuner
-from opentuner import ConfigurationManipulator
-from opentuner import MeasurementInterface
-from opentuner import Result
-from opentuner import EnumParameter
-import os
-import sys
-
-
-tuning_flags = []
-
-binary_name = ""
-
-
-class ClangFlagsTuner(MeasurementInterface):
- def manipulator(self):
- """
- Define the search space by creating a
- ConfigurationManipulator
- """
- manipulator = ConfigurationManipulator()
- for flag in tuning_flags:
- manipulator.add_parameter(
- EnumParameter(flag,
- [0, 1, 2, 3, 4, 5, 6])) #default is needed, optimizations don't work without it(tried and tested)
- return manipulator
-
- def compile(self, cfg, id):
- """
- Compile a given configuration in parallel
- """
- cmd_config = ""
- for flag in tuning_flags:
- flag_value = cfg[flag]
- cmd_config += " " + flag_value
-
- run_cmd = binary_name + cmd_config
- return self.call_program(run_cmd)
-
- def run_precompiled(self, desired_result, input, limit, compile_result, id):
- """
- Run a compile_result from compile() sequentially and return performance
- """
- run_result_call_program = self.call_program(binary_filename.format(id))
- run_result_getFileSize = self.getFileSize(output_filename)
- self.store_size_list(run_result_getFileSize)
- return Result(size=run_result_getFileSize['binary_size'],time=run_result_call_program['time'])
-
- def run(self, desired_result, input, limit):
- """
- Compile and run a given configuration then
- return performance
- """
- cfg = desired_result.configuration.data
- self.store_config_list(cfg)
- compile_result = self.compile(cfg, 0)
- return self.run_precompiled(desired_result, input, limit, compile_result, 0)
-
- list_size = [] # list of file sizes
- list_config = [] #list of configurations
- list_size_config = [] #list of file size with corresponding optimization
- list_N_size_config=[]
-
- def store_size_list(self, binary_size):
- """stores file size in a list"""
- self.list_size.append(binary_size)
-
- def store_config_list(self,cfg):
- """stores configurations in a list"""
- self.list_config.append(cfg)
-
- counter = 0
- def save_final_config(self,configuration):
- """saves list of file size with corresponding optimization in a file"""
- for list in self.list_size:
- dict_size_config = {self.list_size[self.counter]['binary_size']: self.list_config[self.counter]}
- self.list_size_config.append(dict_size_config)
- self.list_size_config.sort()
- self.counter += 1
- self.extract_topN_resuls(10)
- print "ALL file sizes along with corresponding configurations writtent to size_config.json"
- self.manipulator().save_to_file(self.list_size_config,
- 'size_config.json')
-
- def extract_topN_resuls(self,N):
- """extracts top N results w.r.t size,N currently set to 10"""
- counter=0
- for list in self.list_size_config:
- if counter < N:
- self.list_N_size_config.append(list)
- print "Top "+str(N)+" file sizes along with corresponding configurations writtent to TopN_size_config.json"
- self.manipulator().save_to_file(self.list_size_config,
- 'TopN_size_config.json')
-
-
-
-if __name__ == '__main__':
-
- binary_name = sys.argv[1]
- num_flags = int(sys.argv[2])
-
- for i in range(num_flags):
- tuning_flags.append("flag" + str(i))
-
- print tuning_flags
-
- argparser = opentuner.default_argparser()
- ClangFlagsTuner.main(argparser.parse_args())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/select_top_results.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/select_top_results.py
deleted file mode 100644
index 7ee878e5f8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/tutorials/select_top_results.py
+++ /dev/null
@@ -1,101 +0,0 @@
-
-
-import argparse
-import sys
-import os
-
-
-log_index = 9
-linear_index = 10
-quad_index = 11
-
-top_k = 10
-skip_lines = 1
-
-
-def dump_results(sorted_list, k, result_dir, sub_dir):
-
- ref_dir = result_dir + "/" + sub_dir
- if not os.path.exists(ref_dir):
- os.mkdir(ref_dir)
-
- for i in range(min(k, len(sorted_list)) ):
- file_name = sorted_list[i][1]
- file_name = ref_dir + "/" + file_name + "_rank_" + str(i)
- f = open(file_name, "w+")
- f.write(str(sorted_list[i][2]) + "\t")
- f.write(str(sorted_list[i][3]) + "\t")
- f.write(str(sorted_list[i][4]) + "\n")
- f.write(sorted_list[i][0])
- f.close()
-
-
-
-
-def select_top_results(result_dir):
-
- if not os.path.exists(result_dir):
- print("Path does not exist")
- sys.exit(0)
-
- file_names = os.listdir(result_dir)
- print file_names
-
- results_arr = []
-
- for file_name in file_names:
-
- if file_name == "confidence_summary.txt":
- continue
-
- # Skip sub-directories
- if os.path.isdir(result_dir + "/" + file_name):
- continue
-
- log_result = 0.0
- linear_result = 0.0
- quad_result = 0.0
- file_str = ""
-
- index = 0
- f = open(result_dir + "/" + file_name)
- for x in f:
- if index >= skip_lines:
- words = x.split()
- log_result += float(words[log_index])
- linear_result += float(words[linear_index])
- quad_result += float(words[quad_index])
- file_str += x
-
- index += 1
-
-
- file_result = (file_str, file_name, log_result, linear_result, quad_result)
- results_arr.append(file_result)
-
-
- sorted_list = sorted(results_arr, key = lambda tup: tup[2])
- dump_results(sorted_list, top_k, result_dir, "log")
-
- sorted_list = sorted(results_arr, key = lambda tup: tup[3])
- dump_results(sorted_list, top_k, result_dir, "linear")
-
- sorted_list = sorted(results_arr, key = lambda tup: tup[4])
- dump_results(sorted_list, top_k, result_dir, "quad")
-
-
-#def select_top_configuration(result_dir):
-
-
-if __name__ == "__main__":
-
- argparser = argparse.ArgumentParser(description='runs best configs to get high confidence on accuracy')
- argparser.add_argument('--result-dir', help='Directory containing OpenTuner configurations')
-
- args = argparser.parse_args()
- result_dir = args.result_dir
-
- select_top_results(result_dir)
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/adddeps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/adddeps.py
deleted file mode 100644
index ede22a8fcd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/adddeps.py
+++ /dev/null
@@ -1,6 +0,0 @@
-# we would prefer a symbolic link, but it does not work on windows
-import os
-target = os.path.join(os.path.dirname(__file__),
- '../../opentuner/utils/adddeps.py')
-execfile(target, dict(__file__=target))
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/cla_func.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/cla_func.py
deleted file mode 100644
index f4787a2f23..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/cla_func.py
+++ /dev/null
@@ -1,144 +0,0 @@
-import numpy as np
-import math
-
-
-class Op:
- def __init__(self):
- self.M = []
- self.name = [];
- self.mutation_partners = [];
- self.anti_operator = [];
-
- self.mutation_partners_no = []
- self.anti_operator_no = []
-
- # create all operators
- self.create_operators()
-
- # check unitarity of all operators
- self.check_unitarity()
-
- # determine the indices of the mutation partners
- self.determine_index_of_mutation_partners()
-
- # determine the indices of the anti-operators
- self.determine_index_of_anti_operators()
-
-
- def create_operators(self):
-
- # example with +Z
- #self.M.append(np.matrix([[1.0, 2.0], [2.0+2.0j, 3.0]]))
- # watch out: python needs 1.0 instead of just 1 to assume float variables
- #self.name.append('asd');
- #self.mutation_partners.append(['+z','+w']);
- #self.anti_operator.append('+w');
-
- # Operators
- alpha = math.pi / 3.0;
- da = math.pi / 10.0;
-
- # operator 1 +z
- self.M.append(np.matrix(
- [[math.cos(da / 2.0) - 1j * math.sin(da / 2.0), 0.0],
- [0.0, math.cos(da / 2.0) + 1j * math.sin(da / 2.0)]]))
- self.name.append('+z');
- self.mutation_partners.append(['-z', '+w', '-w']);
- self.anti_operator.append('-z');
-
- # operator 2 -z
- self.M.append(np.matrix(
- [[math.cos(-da / 2.0) - 1j * math.sin(-da / 2.0), 0.0],
- [0.0, math.cos(-da / 2.0) + 1j * math.sin(-da / 2.0)]]))
- self.name.append('-z');
- self.mutation_partners.append(['+z', '+w', '-w']);
- self.anti_operator.append('+z');
-
- # operator 3 +w
- self.M.append(np.matrix([
- [math.cos(da / 2.0) - 1j * math.cos(alpha) * math.sin(da / 2.0),
- -math.sin(alpha) * math.sin(da / 2.0)],
- [math.sin(alpha) * math.sin(da / 2.0),
- math.cos(da / 2.0) + 1j * math.cos(alpha) * math.sin(da / 2.0)]]))
- self.name.append('+w');
- self.mutation_partners.append(['+z', '-z', '-w']);
- self.anti_operator.append('-w');
-
- # operator 4 -w
- self.M.append(np.matrix([
- [math.cos(-da / 2.0) - 1j * math.cos(alpha) * math.sin(-da / 2.0),
- -math.sin(alpha) * math.sin(-da / 2.0)],
- [math.sin(alpha) * math.sin(-da / 2.0),
- math.cos(-da / 2.0) + 1j * math.cos(alpha) * math.sin(-da / 2.0)]]))
- self.name.append('-w');
- self.mutation_partners.append(['+z', '-z', '+w']);
- self.anti_operator.append('+w');
-
-
- def check_unitarity(self):
- # this function checks if all defined operators are unitary
- # in case one isn't unitary the program stops
- for k in range(len(self.M)):
- if (np.trace(self.M[k] * self.M[k].getH()) - 2 != 0):
- print "Operator " + self.name[k] + " (no. " + str(
- k) + ") isn't unitary!"
- exit()
-
- def determine_index_of_mutation_partners(self):
- # create a field for each operator with an array of possible other gates for the mutation step
- for k in range(len(self.M)):
- hlp = []
- for m in range(len(self.mutation_partners[k])):
- # go through all possible partners and find them among the operators
- for n in range(len(self.M)):
- if self.mutation_partners[k][m] is self.name[n]:
- hlp.append(n)
- self.mutation_partners_no.append(hlp)
-
- def determine_index_of_anti_operators(self):
- # determine the Anti operator index
- for k in range(len(self.M)):
- found_operator = False
- for n in range(len(self.M)):
- # go through all possible partners and find them among the operators
- if self.anti_operator[k] is self.name[n]:
- self.anti_operator_no.append(n);
- found_operator = True
-
- if found_operator == False:
- print "Couldn't find the anti-operator for operator " + self.name[
- k] + " (no " + str(k) + ")"
-
- def __str__(self):
- # just a test to play around
- hlpstr = ''
- for k in range(len(self.M)):
- hlpstr = hlpstr + self.name[k] + " " + str(
- self.anti_operator_no[k]) + "\n"
-
- return "Operator Class:\n" + hlpstr
-
-
-def calc_fidelity(sequence, Op, Ugoal):
- # Op will be function that return operator matrix
- # Ugoal 2x2 unitary matrix
- # sequence = [1 2 3 4];
- # return = fidelity
-
- # example:
- # sequence = [1 4 2 4 5];
- # Uapprox = Op(1) * Op(4) * Op(2) * Op(4) * Op(5);
-
- # create identity matrix
- Uapprox = np.eye(len(Ugoal))
-
- for k in range(len(sequence)):
- Uapprox = Op.M[sequence[k]] * Uapprox
-
- # M.getH() returns the complex conjugate of self
- result = (1.0 / len(Ugoal)) * abs(np.trace(Ugoal * Uapprox.getH()))
-
- return result
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/input_generator.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/input_generator.py
deleted file mode 100644
index 009af836f4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/input_generator.py
+++ /dev/null
@@ -1,100 +0,0 @@
-import numpy as np
-import math
-import random
-
-
-def generate_random_Ugoal_HARD(N, **kwargs):
- # N is the length of random matrix multiplication yielding Ugoal
- # N ~ 100 should be enough
- # This method is hard because it creates Ugoal over the whole space
- # Ugoal 2x2 unitary matrix
-
- # create identity matrix
- Ugoal = np.eye(2)
-
- # create all N random angles in 2*pi*[0,1)
- seq_angle = 2.0 * math.pi * np.random.rand(1, N)
-
- # determine random operator
- help2 = np.random.randint(3, size=(1, N))
-
- for k in range(N):
- hlp = seq_angle[0][k];
- if help2[0][k] == 0:
- Ugoal = X_Mat(hlp) * Ugoal
- elif help2[0][k] == 1:
- Ugoal = Y_Mat(hlp) * Ugoal
- else:
- Ugoal = Z_Mat(hlp) * Ugoal
-
- return Ugoal
-
-
-def generate_random_Ugoal_EASY(N, alpha):
- # N is the length of random matrix multiplication yielding Ugoal
- # N ~ 100 should be enough
- # alpha is the used angle between rotation axes
- # This method is easy because it creates Ugoal over the whole space
- # Ugoal 2x2 unitary matrix
-
- # create identity matrix
- Ugoal = np.eye(2)
-
- # create all N random angles in 2*pi*[0,1)
- seq_angle = 2.0 * math.pi * np.random.rand(1, N)
-
- # determine random operator
- help2 = np.random.randint(2, size=(1, N))
-
- for k in range(N):
- hlp = seq_angle[0][k];
- if help2[0][k] == 0:
- Ugoal = Z_Mat(hlp) * Ugoal
- else:
- Ugoal = W_Mat(hlp, alpha) * Ugoal
-
- return Ugoal
-
-
-def generate_random_Ugoal_RANDOM(**kwargs):
- # Random guess with the following parametrization for U
- # U = @(q1, q2, q3) [
- # [ cos(q1)*exp( i*q2 ), sin(q1)*exp( i*q3 )];
- # [-sin(q1)*exp(-i*q3 ), cos(q1)*exp(-i*q2 )]
- # ];
-
- # create random angles
- q1 = random.uniform(0.0, 0.5 * math.pi)
- q2 = random.uniform(0.0, 2.0 * math.pi)
- q3 = random.uniform(0.0, 2.0 * math.pi)
-
- return np.matrix([
- [math.cos(q1) * my_cexp(q2), math.sin(q1) * my_cexp(q3)],
- [-math.sin(q1) * my_cexp(-q3), math.cos(q1) * my_cexp(-q2)]])
-
-
-def my_cexp(x):
- return math.cos(x) + 1j * math.sin(x)
-
-
-def X_Mat(a):
- return np.matrix([[math.cos(a / 2.0), -1j * math.sin(a / 2.0)],
- [-1j * math.sin(a / 2.0), math.cos(a / 2.0)]])
-
-
-def Y_Mat(a):
- return np.matrix([[math.cos(a / 2.0), -math.sin(a / 2.0)],
- [math.sin(a / 2.0), math.cos(a / 2.0)]])
-
-
-def Z_Mat(a):
- return np.matrix([[math.cos(-a / 2.0) + 1j * math.sin(-a / 2.0), 0],
- [0, math.cos(a / 2.0) + 1j * math.sin(a / 2.0)]])
-
-
-def W_Mat(a, alpha):
- return np.matrix([[math.cos(a / 2) - 1j * math.cos(alpha) * math.sin(a / 2.0),
- -math.sin(a / 2.0) * math.sin(alpha)],
- [math.sin(a / 2.0) * math.sin(alpha),
- math.cos(a / 2.0) + 1j * math.cos(alpha) * math.sin(
- a / 2.0)]])
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/problem_description.pdf b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/problem_description.pdf
deleted file mode 100644
index e8d09de95a8a6416bf88f10a4d6e4a0fca92670d..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001
literal 24984
zcmcG#1zeO(_cyMDz@kWrbSPa*cSv`4=hDs6Dcuc<NJt4tmy{q~N~h8(-J#Tf7Ym>J
zzMtRyd*Ao}S$40vX6BliYvy~-obNdUnY^$l9V0ypfUI*fIupQ3%s^~wXaV5i0ZM~x
zOu=TvY>*s9ps2YO800_<6tyw{gM>jww#FbnK7b?G0c2neaD6$gEmeXq_5cMBOKq~W
z7x3{8=nhk=J3qQcOn?=PI4Zqhi-4bYCkj%vksh9*H}7nI!8X^Z4lnQcxHBqDbG9w7
zY-*<qp%b;7vtbOkz!lY%&(ZF}MgPTi^O4B6y7>B8>88eat_~!<{2@&4CwpY%0jY=n
zabuhGJ9=GA+M3*m6$gU%MYrgQbf>2E2(#NcijJ>1@Rw1mp4ngR5(g85L(WQ2#CQ>N
zJST|OUIX}oE*AZ<GfxyJE?f}Kb-0(0`%%PLUFoVefJY~-XcaF9MYpd)t158|%g+$h
zr8wSH=GTXo`Ut02v8tZ!F-N+($5;%gem1FXs0tg5&_bX<PP{E;RF$P_XxDO3auN#Q
zotQiBpLF)LbVjd6nmwb6EN#|bdsrDxRCv;BB!z6kh?QvGGg%&tCb&~#=371^(}3Ey
zRWhtbQFytrZ2GY8-h|9n)Tevoh_&?%2%P({OM7gR-D|6z?(&qZX<vpvp{NF5e0D3n
z?MC4igYx8%ir2T{O`T5bex0DB4fmJ1qGcm7r<kbWzRFkhPw-u@94YJJWX7l!ZH*96
zA2!D_h~o*ZbWtF%24KU|*=86riQ|vm1*{8&k>1&<e0zKK@x2jEj=+kLW<hk~uxfS_
zVz;Yr`Y_h7j)hc{5e>ZIK1S<yPSD1L9QVNWRu?sW8R*e^2tz<B$QS^w&k7TIF8p<J
z-t103t@`fE1PVk0vV1b{)^8DSlfP!%GqIH5LSad{@Cf9PRz7-@#}E2u-l8ivOYV(M
z*F#pr(Su0uON-wW2cIo&TzfGdNKMs}*Pv>Fo*D&EWIUsoi)>kU=eivkSfKIET98lr
z>v;s9kpVU<(FJNRJTg{%5Dq0mey0(-eu!>y(Qblwi|P|OLcCJEHMPq3Jc<Qy#Lo|T
zr7^>~`q$jW-)NFGF#59L-@AK4`h*oRyvZjqG%kMB!r4YEw>RXfd<)r4l$vcrj}(;U
zWp!*S)bvONNui}>d$5=R8FZ{C`;x3O)L_7~Z(Kx7a3EsjYjR}SSG$-Zu-oHJSOS`<
zH`uZX`$CMFgk|s22*jf?H9{lpSS1W5%EWI8-2W74f~ysYCh$6PPRA<h_<2$cM-*JE
z1OZr@=Q)<;nubAlrGzf*LNW`s$q|ts4s0wa9b;Y~ut#!oGC+p*j8iKh6z$2u2bt%v
zp*^jxnRpLAV7H5ve;p>RJ|6|^cQ7i}-$E>^6BojH+Ftp9-Fj{JzD2?SQQ0%RTh88B
z6kK&<d|)&I!L<$vy-oAH0sZ^U5rN3JCiU<`@cFSz<YHijpPuadT7-@wRSd^<D?gHV
zspN#2RtTE0i@~{<@F=O}iMp90gX`lp0tA1ie)UmoXI0qSSu!z5V%VkK{@vvhl~y^T
z*zKk{b}dg49D>YGKZ#&F7a~bc&_9@+=rmpODUH)>l_|#v9Nc|ki;=AFu#1QL;wc&T
z{no<*apK|^XhK0;&P#Ph<lR<QJhwSilk5G6?;tmspANR?)G??kMR!z(qSx3fJW
zX9vUJgx(I}yOK$tSBMPxh`P-!t}AQYn9PAVKWUE>#l#5e;F`X*l4q&zO+Q#p6Hf_?
zFhnEF)}nQJG}ESVhsuZ&Nxhy=oLwF8h`MaEe>8dt)*^-Rz5oauc|heG;FC9@Jb7@K
z@hV2)4j^)5jz`<J91M>fJfbOdMJYz9*_DlUO1Kiln$Ldrfi-iFs(a~k=~LjK=5PqW
z>K+zdL*9Y=`?YRHL~aX%AgO?Pd)!S*+`CD4xZ#o+xbf3Uh1PghU$6}~Q%?I&;E!(T
zFoW7O9p3p!F_u=CRI;^%UzSa%b+o%|zyyqere_RTFYeicIceo}-CSQBT)>mvv3k0P
zjY6%d9Tc=jmj3Fo`4H{PA<Qku4+gvMX^Z*i>Q?EvOUYBy5j5c0YdWv!HMV%rZ$BMe
zd&wgaWb=ToAXHf?#T+<9pL6QQ0aI3X{$x@x)}Z>Sesi$(`xF81M==Pl;Ti0(u(d3a
zFTb5hSC1EE&EGY$I|PM)dqTL%nWjzT{wetUVZbP*BHH}m$sw)0_ai|Jd)bU`{i>M3
z$Pmx00zJioTvY3mYS?Md`V-u?6rg7tPAt-+aF{t6#G29+GCMd&p}7I7P^;XjSH8Eb
z92YT%U*hmZ_Z5Px4_HJ7Fdw`WS9*_?5^}lKzTCgF&2Mx09B_Y5=<8z13x0ZRd;}&$
z&mM48T-;1&bIJ#M9K!{UZs6;-CH4f$PB%*~S#Hlh{<eHSbXCb_wAAJ9rffkWOx>^a
zZd_-tjhHFSQ?Z&)oXar9*Nb#;qdEy{yF<tKF-!P4+c%<R4{&S3BOE9c4Vtk`e8jUk
z=^8dWj#2%$L{G{?w|mF&JA|r0^r#aO;$%rod|ZTO6&#KFiY$Xp_w@7?7-RQk9(~p%
zLq?n<n&@1k)xKn+VVPj4rI*XasJH3CI^~dE^kICP67;@Iw}RAesD}<_`0I2nTvW9V
zQh)(X&-`;c$vOB&)5-U*`dW&;*Mb^g><Cc1d#E3jH|5X6AU`U0e}&M?EUEFhhogun
zr7~!$9?`$zBOv-fEi(siiUEN2iYm}n$NXto_RG~^D|olz=!%$UJy90p(zWRen)Bc$
zOc5120g;;czPgZ{^tq9eUhS@LZV`o#6tNz@Ex_Kg=tWI7q=m!P=Yg|O7avMzm12i~
zI(r{=ApZ$DUwC2yynL|Irrgp#`oz1U!zv(>>X=*XGGXFVz*3=VOAU8yuPB`v_56cK
z($Z%IGyC|<eW-8&1erw+A7M*t&P|ZcmsJsa6=m{JntILklOzV|?}+3)Ymc)yt5Pr5
zSL&CH#3j9*SYz&~EVF%kNN6j^p7-$)$e7vRqc%=0c}00tKiSyf34R2TQa<7^eKFeH
zl4PJ%ZeGc4{SKTvft((`v!zCl6w0D+>oz@ajAQMW3^acdy@f)4ViALeM-mVQ=r~zQ
zASSiPEYqlvcH+UXabCn#*ocls0aNLJkQAJ&4Q^DR+eHMt_bkFuNp&|?a$U$DrmYvE
z6^wB{!~Og`P$ya>hcMW$DH2{M)_Z-<>Z3909r%2V(l5rZvWGEs!qHp#q=V1|XW4N9
z-$)P>1+}X3V8M$`fx9Caq<+&g0^k#3@D6pRVl~lrEea;S?2s|?MHQwY72iu`o)04<
z<+s=*V4g;;bxb!hRFKElqEB<!t>8`eYl8A#aJ(^2V0v`+iB8U&SGBz2B9It%_}%MT
zUn#Y5_zbql44a92-Fc1ZAEjGVdmkXRQmmy&Ohnh(rIsb;rfBDl*hG}E?=dV-zV?#W
z;=wl&Sl~9o^l_&_ibG^#&rlPtz5NbXf3~nJ@BWARvC<b^mI4V0oNq&5)B~v95wl$o
zbuzcuBPgTej(wH95|r45F=9Kg+`R*u(M9uH??iV$LoJ;mtfGlNu_fSv|5&%co+8sh
z++hfxpERt45>bCHuh>kb_)%#5JXt0Rv=5F;u;PB3sW1D4haEjnLLhAchK{(AO1`Rc
zlbuxl$mg+q70G}$mj^4>w`OG#8h)kuRe#z&%JZ*dcY`-RRI`R#&o|jigjq}2bhljq
zaVB1N5|kv&0f})M8Ka#xRbQU*6blUC@XWxJW4Jd7j`;OgP@9x|@SPehAeavA%<UQI
zKsQQkCeF+A-(h;ZPVNv}0~Q^YP)cIsO3X>=>pEI6h10%sMP1*kYkld>0^@3WUg&8N
zAX7x)RIwTkZ`CNn60&4saMnZc_{idvLB8pIXp8#1x+Bh~HHQtbx7)a!EqiY;>&HAT
z1u@|#Q~9dgRPOBM)-<9i=q%4z!<1P!d79+Uyb9(J^Xu{fWZomLl5fU1Wxjmr&aYjI
zGfnPbc_<zCxGi<?G5wY|VfE`nC57<B@Di%<sCNY?9@&<a(dM@MVmixI3*7~5qi7FB
z9_V@xf9+_^ues8ZG}J!KAtAFMI6Y#*a2-wcj%9jV#*z5KEb`WvTmPF3nxL4o;zz7m
zWz95nQMHddwb-2^xY~=6#{h?F2a4R*&0KD{6=pBClv#Qem(IDOPF7ZQx%^-(Zixy{
z`byldqqd*TO)?bYm)IV<-x3J&8FfIqQ$q7tR__2#h&AvaJUMmgf-Ee#dh`lM@Y%51
zv5bz-XuF-BeZYaZjsnd%@&jPHi-1jl`-A$4sUnGh&BuCb(*(UP8res2LfI^l{n#Y+
z$TKHhf@RDv8Qx5@J1Mifu5aokG;|QY7@F+;L|)I%^NtJebFd3hQ;LC*`P-E^(emMg
z@X4+2k^6mIGQ*?eSyM|znHIn@+2d2qk2Xo(Vg4Q%Aa+GCnWv7ZhACVzFpFt&=+UVf
z?Z?-*)`YjL8is6>zM@(=^6`Qu#9@U?{G1i-W!A?u@;3_YOd62rUw%aj)g*ZYe&_6a
zEZbI2v)(I|q1{Te-Od&uSjENshBF-%kTG;Njjf+e1w<KfAqnggsO&O*I&0Tpv`R)w
zQT|zg{P+dK3P_k$$#AUIK3JCW7XI`*tIrmy@i7{+Z(gNTF*w$x6Vg_>tbc`#@T{n`
zb7IqM*D{lR_uec%LF7<pN>-%&Va@`OYuAj-$FkMgQ0M7I)UX@3y&JXC3lyp6a93>p
zlb&DTKa|%WOpZ=(UI=YtY`M@xcwyS4amtr0aTdijbi0*z6sKo+pv!Q=&aN*7<|;}^
z**qOwePG8Xv;ER6&!wr<aE%~hrf{U;uJFnBp)EM}r0*O?_=7;Aj&Iwi3Ve$CD~(Tm
z$R$`sJL*e$a2rM1MFQD(^>B(?Xi@d`lJ6-yPO`l}zqNV%Aj-az%4}<9dr&L-iw$Au
zctS1amxvNPE37YUz|8nwr5y7L#>~XalOn*h1wiM1oYI%XVxjT1Lw<SryQ})2&ucKo
zN|rJ$<?MPYXS}K&e0{g6aWc9{Fru8m;pR`7vOj&1orh_9MaEBF?KP(V!i_WI%zg`%
zPiL=dZ|U5+qob%?1F+#aM2&ix{7L=Q)GLqUp^$U(C?Xr@^9NqmT}lgZZudpMx*$uY
zB4jc1d(TLHaCFD6zlX7?)RQ|{eCC%yGGA~K1U|;Jmj0@aSVo!Lmlipb@u8JQf@lAh
z6<kgKO7leizM1zDBj4Z`B;}zdXZ?KZ3!7j&k0tFvk$~~AxX)R43)}j`)M$N+8*R3{
z)GZPe@}G2T6PIX|z2);XoQ)j2!-O@(ZZYB2F^0u0Z6>t17K@kSPYc8CM)7q(Q_cdf
z<)MruM_IA}U!^PA-J!d2b?q<q=n6+ZI;O9Y88}IWpCsmIwAad5#>%1R(i!FWxQPj~
zU6l8WkAI7k%5G9Dn=GiHbid_rW+843RI7Ne_qq$@UH^$KBh_N6cJyq!&*<IW0qu;_
zvU#sU&&U~<u0m2i%gS(XIKOttH0=3~&u#IZb#9Q2@eisGNkQEMP^SR&%EG|L3iy%7
z^jF^j)4y^aENfUs6SSiptx9=6Ic1w=VxO-*yLc}M!<|%)N{VkC5FoHHkAR3(S_XT0
z@^XDUIX`W!w*Jv=bdhRjQ;%;dDz9aES0;%AaY--UG&R+5O<js%r(N20W%=XIq@*|U
z@o?vT3ofh(u79TP%O5YU>%pzf$C<~)A4I|tVYIuxYuDfdexzyMksc+<0V=|4;!jIJ
zq$1<?b{rAp!s{$Cqn$4zY~s+A2KUJM)7R{^?=1xk_0`mMNbxpurDcaGQsZ>8M^R+?
zhlI*q#L2iyV-$*inpXg%W=pu&-omTTRFNe+Uu~T&|2ToeudkOS!m#L4Qxip%x7R(0
zn3&C;uH7gV78aRFN*7p6Pb9#9#gkUK+n2AtS}mE_@&fa9cluyi4b1l0OTsF(j&W}4
zl9p5QghE4bolWjn6AGii`&4%+ECUp%5@q;%G}+j`%-m~W(52>|o?DFFrnv_m=7w>f
ztI!)TE_$i(t|$_*rmZkEI=WLSOae~*OVJ`#ip5xlHI|_6P#Bw_W`3`OgcSIY38)~z
zLHE(_>q+Z!I#HBM%Cng69Ms%rw(P+1M^s7)Nx3562_tmcj_lEfws1^oaAn)FW%6yB
z-b1a<hS0JDEAfYkgCeco&kvJKeG*mh4LdZ@70%|&<}Z3%{aY)V2dbAWDclD-VEOG4
zMcoY^*X~b;w{C@>1V8N8<KvsUw3841?2|?0!e0}t_t1yHhG1Lo?)|=6*81tCF%ATn
z`vO94qg6*}{mpo@A~vWQPdRRBg+G0{JM)G115!sgP6XhM2}5-}+is6{GI<-gaNtYg
z6T@Jn=B6ia^$AHxD&ub7`g#Q3c*H13jJ)1xgLh8x3b|3*n=&rn@wCYW<!VPe*k?C7
zBwI==st&eAVy0`r$<liG1)Ve?iz>;9gxj59YdFEBhDW_;S*cf*$S=>{e;}tDZb=Hq
z=W(23$nA**_c)ksFnPX47@Y)YX(vYc=@yDaH$j|csV+>u1S=C+D1|W1Y03e~2hIIz
zJi3I(In=pi2%&~pDGe}^Upm7h63mm+F5KA8JkxY1<qrbij`fUgEw#z0H*iUgtDscA
zr#Kl^?5ulHDi}+)@ojwnV?@PttDwh3!(PG%Y$f#27{iDlkFLJKxb`c(X~aA9oe_<U
zEu1B1wwNmLNS&qQ(?z}LlBX%j=h^syuT)4sb1E7XupVpfSTdgQXWxql?4%x5&pr!4
zHHAGs?bQftzISv*(<H3*0N-!J#KjBV#mEB#)1E5~mAkX`MJ4t_?!8^z1BVCqXU%nw
z<BlWn3V5r`_pw&N4_O9dEu4TZWghz^Vgl<cjv}B{<xMdr-+OG?IYBBW)VO<o040BE
zO<WPy)aH`clxQr}+^tgcKCv+q<Ljw{l@#?#nE2y%<?#b_^J4StK><<|#iRbuWA~AI
z+)v~0Ds^nvo$g!9H^B~iqR!;*Idvhwr25P)-W}IB0=mE2DqNf-QSokAFn3m`dwlRs
zE$;A+r8sM3S<gXp4o$O#&sYX!E|d8>jxgp@<CcnXIL%w7y47MthH#8;7+7ki2zGiU
z5B-B2l>}R=RX^q_)l~|u8;_wF=)gBl`*n!`Gy?@0IN81VnoPzGf>kK&RLSW@JY0-2
zGw6z&miqc;6St8I(2A=9T_>5y$?<f|Dk^Oa2mNmQla`!jrbod&j>Kap`$F4im>#C6
z=Z7I|2e+!?ZP&=y5!1P3I8th^$jDPEQ;@jn*lF1m9fX>8cn8a=1t$EN6?n36OQHEJ
zAf9<8XTh3jMM7(l;Yt(Fw4-y%1DSSx`+YJ$`)LMETpDMFf<Lt|f77u2nWxNHuD~Zp
z%pH=o%D3ka>U&aQ?9D%WJj}bjk~*e0E1_3Dh^J}vOdsnaA@%N~*<QL~@HqwW1^#E|
ziXzQScMSLGxS;brPebx&42N9wer-=+Z>?e^<$F-9He}raYgEQzDTI7PQlG%yGf(Yv
zVi4b<0vT=)9d(tNP0vRcIME1<Mt`BS3zKSBpHOvp=P0IUQ#uibx{2zJs-E=50eyNC
z*BV={hFDxoB}lOE7TIHFlXqddmYy<j`W1O@QN!t_D4*@0ESMMe>o90Jp1+vLpVAj_
zV9*_?7Ns`ZE3gWzqjDsj7#fR4m_gO>E2N_&>-V2wrjRw7MSoEWQ|BE^RP|W5Q#jEt
zy`wPgONVO-Rk8R`KK)nlbcWW>hza=JrHJY+j;Yt12)eD6i$rGO&$A0Oom6GI5W^K+
z-z`5Jzru`#DbB*UE3M8@D*0YCj!iMQRQVo$!!r%dXz*?ZF^f|$!I1xzG)2WlwgO%<
z;ONC`C3(7ULtUPNxpImYpWBe5Fz6Gzh!_+z^z$mZjWE(KS=KUb!z4hM?L6u#^^WQO
zR&{(`dsC(k{)99ddVKj)3@0pajz=-aLR~$DO$xCq902;}3oU37ZT-veTE`~H6oh^$
z3^|2>@s9OoI^L%dgV>pL>M!ZVQIB??CVWJS{Km>E<Jga@N?|c+Bgq_HXVzXRl7w%S
z(b^@lhR_wthbCYFh;A`T0Tw;he|mq&t)6I6HAmj;R(Ytl$GZrXC8lXz-8-K2=bO5@
z44ui{mzWo0-E2IWx5Tx$59$I`zRJ?}*;m>hX(!&LzSG;<FN5(=kxh$@j6UU7&w&l#
zeK=WPE*&WcMiJhalsa7WB>IEZPmQ0{Es{F>k`r}CPv1rNU#XHybLpvmbPAiPqop2A
z>;1H^VuPVUxO_Aw((c#o_FCBhNnta*x8t$nf;<nChSSXB@PTaUaP{s_MAIW?TQMVA
zM=KpDYY)GD%bCf;vVD0*9I02GlfnLIK#h|F!E065g<^G%F8KB7S7wseC<%YDVy?{-
zs2Rh=$jb85jA8j50a13d0|Dg?OhG^ykg>Uepsg#h76YW1m6Mg2gM~>4093REL*gOC
z43Jt5AR90-Gc;HN$p$&vIyo4D93f#8c?Vk~B@kE(2nn<h1C>FpV90YKu3#~x>qnWG
zzCR*r3wfOB`}2^<3Q!1g4{6bn7;0510w7sG&8+`_2Pk0;d0)Z58z}qDfpYvD)$f5`
z+kZDl3VoUHu_Q=ykf|iTo=Zn!pp?0>Be50$`XY(|rtcRWVn|C)&^fsAu8hB0_|brn
z0ocIG*7RpfKc-s-1U4{+ln}H0n!g{D1s(JEAzU}`d)@&+Ie8gkMn1luvkF~?A2k)(
z0oU((W6poS3Q&=W0l)-Bfg(_mkqN+fy?&4+1Ay^IE(hRxb|85i0LGt5b^zmbqmWz{
z0OR#L-)R5FIRB=1Xd6Gr_J7v<U%O=dX~IJw`aQUJGq(3L_{Yf3#>e+7%E!#8#30Xb
z+dv3j2_q0wh#r;~R*0Ag(I8?;yI2t3$Pu=bSU&N#c-2$C;+(XcoSZ-^@tN*;so7z*
zSJ`<iiBkOddC#S0Ve{goW~tt5G{;EIlHr=!(LQTTX{@Ngi^0=3%GxQ)Ku)ml>!<qo
zk>-v?S;gW$SV3N^R9`HR-(}-fF#J1b7<(9dL84=%J|9x=FR$~BQZIb8=UxxDWfoH1
zMWNL$P6IsEzV)PBLKZ48f3lYAW&1Iaf(DKtsLTQ7g;XWvq-dm^9Lzxu5UDekHa9f`
z(@8*pW^M$8RuKj{8abHTfo&a#nV^VN{Ej7P5dboG4vt_UGXn>R5|B3dkz{0G2CA7G
zgUuYZ0PO70<@@>h^X^A3;AiR2>tAa9rN+;iH(Ow5hr;L|KL4tPe|Q$!=6^iT%D@0s
zV>XaLu(^#XG4r*S0xCHff}uJSD(w*2`l+(6OaBnVAI|}ytMIcC$c)(t*f^U1x)(M#
zF#$o;DnyTJ0T_VR<~B}_05+hVm6PN5G!qNZ(cIeH%D@4@$^Zmefvh1K6C#GpKtl%u
zBaj1hA$4vH<6l9ZiIwvP<aJ`3ZCZ)%2JN^3W{CAMzhw73+a7AoHUEr^YFL0T`RE1V
z@o`0}(+lU*7(OehmDl{K>HSVK(pO_i=K?8Ij;XlQNgln%qIDmt9QAWKZTWLe8*{5R
z<bdWE=bGJiu!+-p!-?Nqhz>ipnB9`UROd?5jUqdVS?;L=oRweukUcyJEju@Va0{Q<
z@;w1J*C3ipLxE%7Ex1uw>RWu_e6I|>sxt!smikA&0AJn^**tY*k;)91jF0|e?V5ZV
z2+HaRgCPJ*Oqsk=^0D_D2wsBtBDLt1xS;MXWPpCN#gk}KQW~U-*s|`Eu{@h^%r6>9
z`P9B$G&pI1dKjkJt7}5!{whV+8t*@)NJCOcLPGXmN)h9=ng%M{s@RxAPZAI_fA<qy
z@3;R=te6=85G$q|qIFYbSg$*{1~UY<H%p*-(0f+)-%w{`fQZe1!25UUkJNSfUj+<0
zg72eX;(+ROh*<oqmLN5+N5FQy`*B?F9FV+grTn``R#tXmR%RArmg{|um4O-BJ~1?x
zftZPd{YTG`H)3HVW@Uk>8A#vXOIcW$h?%dgiSMnkvT^(&uFw_~|503_rT=fl6=(;M
zLR(`ah+IM>+4v9XWn=^zSlO8wKsH7o$ll4^*}w`a>kxzBr&I&&%mK^{Ksz(@oAS-d
z@uO9GYXh(u(8$)>+Tc0~U}FS2S{XQ+0oa&<j$nvZ;A97WF9I6=sQkSmBLgE4^gZWV
zQ&@paz(CWVWvoE+A7zjx7#TQ$*57Y!epFy&2HO8jvH~5y7lMD}Kp<*tU}_3-xS@Ei
zm-jy<Mpal$T1ojY5~Bz*b%H3g|J-;O|I5Y$naJNe4-?Z*5rM!40(1uGgOJa^ge9~t
zD=XBdVPar{0F3qe`(0=lAz){Lv<JO109aTUAU*u``yFu55+;c81O5IIK`1;~prX#o
z#!Ae_1_AB0&_JJHg4XAR0uNG~^SX^+BWAtcx4*Xw{bl5Uw(<QJYKt)uv#_%fvv9CN
zMh8hludI-sAx~YWIT(qVSRr3%84C*}4e6U5B2<uk7B(gT3&;0*%*>E-=sR;l+hQbU
zhJ4xBAY#PH`Td@g^$!t(zSlpC5VZ9FQiLGOYGML3frtPn(B8=wB0$jn48Y0u{SJ!R
z8yLIM;V;uhK}bYMS^WmYpftwGn(iMk#(Dilf2)KUnQlVtmllRh40I_VpI?jhdzt=7
z{YN>|HSPIX0}85Zpg?M|u>H*a+v|U{1FgdZS?n8if7M`R{v-c)9RRd<=(0n4V}LF$
z1lDX2U_k0$r=fshfldarZDt4veu9Mq0v4#2XJ%%;ei8~;C{(`p4oUq1knhv^PbLwx
z^#2whO#d1k*HZYO8c0GyT180d2R!~eG#LL2HaJ*+V*`@@VfPZVGu+&A{$kwy!@OLR
zq?_~d-5mI76GEmPGH;4NsKE^-){evw`XVSye7$i(x#zXslraE1n7e9yUl_XUi_Ud{
zAcVqSGk8V@$VLkNf#m-(;8{3-zX&A#^F>(LAj^7lxXiyz%Io3W9PZD7{x2+BNLPZw
zHzXOdng4>ue_>{vtT)Zn7+w^JnHuOK+k9_7c<3&gRGqPzO4Yktw=o#uP$hz98t=rR
z5D*Tnn)Z5Qwz?7&O`7c?$~Nj}eNum*@8l8lgyMN5nTkwe>}P?s#PHp0Y@3!QenLt%
z-#*WW$e41fIg*K{*-EAEBqWI3uxq_>Cz_5hsuvp8gwGELC-qGew)Gc69>@))+MwZZ
z(biKxg8b;Ty`XqS^O%hXX<ACti#}AXjbrsld7H#{T5%<T3UUI{dv+l;S_xC6@%bD`
z=WonRsHFWTWHm%&qy!~zXteKS?H?2x<9Al_+s?e6+JB&LEF8aApM&A&>H|0+TOjoN
zUB>_R_g9_&p)SOhV1;7oCq{@NM%sVbL#|8zZ?Ofjm7ty}=w<=7`<}kuM!<FeHkRLI
zhTt1zKX0H9fn0ACfLsBPnQ?T27+hd;sDBFL%la(~*WCE#c>m<i*8(VOU=8u8{f~_M
zhn&(eK`aq=i0Q-5$^;QqD+5zUVio{Uz!7RsLNg(5G6Oqt5OiPseha1Q&@$+g0N{5H
zFJNP81#!_qWEc#xR)t7C8vrO{?&t{Nt=A$BF{L4eP>lj{(^1}_jz1aYcjoz<cm6a_
zsjgoMBBCH>h}8gH${Uk$L%;u3i2u|lj6y%%Ev$cVPir|TGYkbxuM!CpL++0IZFeG=
z{VBg>k|e>QX|q14$`yQ4@JbUWo&-mli7i<^ZYWbBn<F$g7nKJ0i=fmXZ5p0Jh<LDk
zVP*}bpC77#{o1ySE1*%~nFCOJn{$=0FYx^gEOEdE#lD6}U(kp&Pa$dCr}y?Ul2nZ0
zpDecna^H4NfrcfjxM$w23~;9OP2M}`cJ_3`>$7c6E_#1^K0vjB>7gR<L3xsY4X|1d
z|M1Lzq4cE^I?kS;0Dht(6024Miv)g8Co=eM2m%Vh-EVCMWUtI56{57t(Wv74AG9l^
zyDFx}89sbKUGdWM>i*Y<AF0gA$Uw^pFO!&YZ@s_$JirvyH+YWlxi5PL3idbZQdo8l
zz_wq%9lQ7DI@>Ob-?Tl?mT1#AP!bGRCMsKQMmI4YC}}DLdm<)OKo!&=nTaT-VHK0(
z^Kb<}pY7wYsu0g)b+X59M18kt3o_kpmx+6NkGp8BJl<hG5Dy-W@3nALO;%CiUCZ`h
z(GHWZJ9t-Skgv6YY&|2%8o)<<ijrJ%WqkT2DcNYgLagtlmbYD6ba6zHF!=S=uK1HD
zA3m$~&o;T{%JMC1tvy?--abEG`}nYV^c|6qYlp>YqI-bhVI*skYv=*OF=FmhZ``}b
z`gLw@4WgVua9t>xx6$8&7kuaHG?_asPQN`8l4Lzr0In~3Y`s>tiSIia-BB)B-Y?oW
z%<)cHlrQ4bqna&L?m1AOV}sqr`wE0h_!|X!!)f`;sQc~9`Yn(DGk3o!j~txL|8M1y
znfbrUqY_XBs0P&hN#w36<4-zg2q8+p37pY)Cia^U{$OH1sooC<(M_`XgY^9*d_RcZ
zwRZ{lo$md?HUDI@P+|(9s5cbDcUt(968`W_ednatgcSIlCEipV%-06ZZ^dza|3i!Z
zd&R-T#B$Rd$Wif7RuaQ)f}Ki1WaM+FCl^DWwBTK*<_ptQD+G>&0S9`7n$d}a^4xAA
zJtfYYdvo>_Ce!Z@?CFz80Y$x2LA1QFf_A4^6&}qkg~`Ol^`YTfdqWqBxwYfXGM}y5
z6HgbD#vSj2vybPEIBln{jEJvpA4uops;o=Dqc?99O|`;j9naIwk~pjh0>H*HbV;2K
z;EB(tco#>~iZ=)>40Ck3==zXkrVljT_nv-+<2wMz0=0I0gl}Iyvs_<KDSkWUwVS=T
zEY~9aj)ZwMUYU{BzQ%CsIpgAkR|PDQny<yXlxgLQ6*iRUB(oG5TN4r___HvLC}i-7
zvJpq~SVbd65|@yWmDZ)rvD*EJ+XdfgB-dyPwelqU%|zFmvdCnO7AoZ8#<AXsw1|{^
zGo<jEL?d64(R?V;Qn^kzkxl#!3pt*0KCYLd3$sGWD0RKo3!66^)l}2fraZ4Gia+N%
zX&Jp$anf??uTe<uVUV6HCzHy;rgsEr)=FX#6RnQRoxSeH)d^!FVf>h<C<l7pEzQ9E
z^lJiVn>r5c@G}XJVKT>|DI6uygA8jkZfaqXJnXF}WM&kf0-duIm4l8m-gSlb;J{CF
zcLd|KV=}tLAa=`Re*>sw^y(71MjkS`^6sG=&8%O@5!K2i_0y-v7ZPy?^LC-rhrL<2
zqys!i3_S~N-DS*|2cSG_dm3PTWr(nG8G|9IZPhI?#+%YRT<Uj!mELRPEn{*b;B7fU
zu>s4#(`S2UWhZ$_{P<^^Pj(F~>yMf%f;tBPZzERKFbR_A__<F8?h?QAJxntp?K-qK
zHGX}J(3(~A!G2}%v5;5`>2pmSZ|8W~V;Su?^CWYPzGwX0As*p5^h8`@Avj2C_AJpj
zB<Z7eM7566{=tz%F#9A7SN5QBIl(9^G%;)!Z|c}{j-*Svoxa_}TS)_Vj~=WQ+cCRi
zYywu}IURZ~wCQ+EuRd$OPj$!ZOPecIa#dgS7S~c7=X<#<T)fN??$St*fj)TKZir+`
zJl0Ouc0GS7H5-k+K>TrSW{aA51V(3w-Q#@ElVCfc=OYmT^akin^~q$C^b~&6?$P{_
zixYPAX>DceRGw5$q;?yR)OngO9ul4bj%b$$dv*FH?YHv44^!`r%a2IlCm&VEUc883
z$Wk8;(|zl{DBlqIR-@3QqTP^9i;gX2)B-sMpe6B2C}h6@@imJFvXD5*4ywopU@dkC
zQFw-^UJ;d3f~uNO<nv>oF3=psE~EniBz>PWT7>by@i7cb=E_DbW7_s%R7B@(uw-2F
zr3qc&BYXEndt<ySlcM!!RcWlv%rC8F58o-B7JEo+A5d=}d@ZZUTr9Znjf3h)x-SkF
z(iso{kHA%{HJeti<D;7HKt7jwTwaaS*8j<t2PpxdBb*qewACWdJtO^kLY9_C_t*qx
z8dSdd49SXBb~$|OF>zx>ft2p#QQiSO=4Vj$eN;UWz?v=b#SWz%5$Yn!%a24Y7pCXZ
zUU+J~T=R#B!}`N>p^rsf>sa8<E~PIFsXy@q6Z3`#P_!^KF41lr&ttU;ZhbQ=OQoq!
zpHY>H$C)-dH4MI8UI2Kr%)ZCU=9eN>Y#qW^WDds}+D_$3mnKS5OghXY+D~*q)y3ZC
z+zk6+tJAP{L?!J@+hp29-kc34=2+_Gf{?q97Sz<SC|S6sBk_|AKH*{dmM9X5`GxX(
z`AIP3>fuX|gD6D4qR|y)^Mn}#Hoa7iv~6tpkKQ5|EhN6A=VP6zG#q}HZJB->t+g{i
zio{D%e22V6A4PLcu{AWQ#N>05H$r6wU4sR9b8c@J30qgWq0rT7<B0hbjXF5@UQSin
zP9uh6+Osg7(+sh?;@iecSJm0YAHu+p1Li%wSAnP!Ag&>T?giw_*~MyJy4)1xCzoVx
zITY*cc%b8+MGwoly9vT2(aAEOjB7606=F+C>G+<C67Z|4*{`rGD4Pz1*_my5D9eyh
zwxA&fSM8(Wf9OqpRu;3ZDl^=OXD3D5PVq%E@cAC`HxI<3yVDAdrPRDM);hRd82YBI
zeU5ZBk%Xv@;#$X3))i}5Tp1ni*@_K%X7g`9I~1CTBUFIG-NI`}3}Nw<)!%;3lV6|d
zJl!Fr_t+_H;6qqaC>?2#7OQ<(Q1ICim;TA}{!#3-c0iCGC~_A$IybmR&}=%M#8yb1
zVn&X&XpCkU;|U?$XYK+9r9rpi4YUzKne9O<>RlBOjZ~bkxc-VfZ|qv=RU?V|-PO5<
ztwV#bLXr}SgX#FD_&k&Xh&xcmxL}kp{NkPPTZ~g5d6W4mA}Z#5JC>bsro;WtR-Jjm
z_pNJ#lo1`|n%b8*SDA+4fy*lp8k@g3`kGo+5WxzG37y<acxo;*!*5CFlf!z2DWRSZ
zOex&uw*uT=$^{YLkU$0jX_5_sy=rVt32z2kaWPEI%u{m2GdKM$Zu^MLC>hPf1NZ>n
zy40E52zEQ4w@&5(BgO<OgUZOWIr}YMU*EY1HKo~v(}x#KsWcrJ9t`%%-(M|H4SG~>
zwft(L+|0cUPRvwB2hZJ9MyeyR>RIJRQem`ar3It#6XVG+Ax!hsS0wdF-oCG&6>cd@
zTGvcu5gH{K=!A*!iac{kyTVG3o!8ZlO(iqzixMlfptsZ%6vwt2>a}q|?n>UQZLdgq
z;*D=nbSxQ$x{vP~O3=T*+pgKD-)T@7vPpaw6hZujV6j&KyV<Qq-Md3}kX}yqLudd}
z%voJ#JbL4dH{Q0u*P0J<ZNeX|wLiLJCT_eR<Q+K^b}eoZ?7w8#ZYf+-VCFwjRs~CJ
z&U_@-@X}Q3Q*ni{mOFL|dDwK$UsZkNTvr-0<8Ct5VoOb7-&<h^lhUjq4in1VM1=uN
zmc~Y}Gvy@=>0EefyqTtRhzbDZxN!$`5x6?yh&i}~Y&Oa^-O@`39>(bs?|s(Mlz*Py
z#M@M;?@}hbFjZEcN9>~|AmELLl5lA9{OT5a7yR<y*t0kMBLCH%{TH6~YlD^!;^c!6
zMTo=snqoqIL4To`P-b|8VnS|jP)x|({}II$015$xfg(T&pd?TRs0=g!8eW^RmLTwT
zBm!#ELeken9cToJa9G*ee2;Sgje#Jj{c2zZgt&&G-eJg}GYDt`Ic@?p1)2fP-0aLi
zHb8Tr1<(>`1+)g*0By`|KtNldE#&uy2jIGsYrnwvE+A$tlr{e_b+7GNp#5(z^9}Da
z&;jTObOc#L!gkOxfq`H%2M`DdcCiII0iA&^Kv$p}&>iRwa<Kiw$b~N54R*|YV=4c=
zk;}}$0SPMqI*pL3imZ&GhTF(U${c|D;HCZ>GL1K?0;q+4_Y1?qcoGre0>buX?*jJ2
zeD721L^)GhI9XCrHmb%lz+^sp6!K(DHM@)YQy4Gbr(*Ei&SC5N3yPV6?RyzRl5aiS
z`(A!{`QhEw`Rs|$)s@cea@?nIFwGObA6TpfKP@EO+k5(nR3@mmCm{R7(sP#H`tEa%
z+x*^eS78k350f*cKZ_J9cS#m$c0`v~FN#G44aK4|1CZeTzI>Q!r^=r+PeSaWvR6uE
zo|-bDdPbUU8P7&(RFNpa*eyY2&dG?wD7o|D)XKq8p{{@L<#K^`rp%CunIrlFUGKRO
zn{n<crN4G!uhJo-Jr|vZN$AoyJPDNZNYob<W<os+BV>5ba&t_Lzk$AP7_qgRbiyyU
zQIL|GS8fP8b#JUTv4JcSD!`vC9*$SpY%}Yob6WOIxy4(dzVKDe&}Tu*A0Vwnmk}kQ
z#0(^1br78I!Vw|1tpw1C#+o5Hh$xs*?F;nA0QZG1DQ|rgd`gK(Cl(bVIwtPd5QAAH
zR{iB3snFxWM=ZFLBpxhx7{#M>MHdHLNwLHSv|D(}1uF(c!K%J+^L_<rs*hiawnf#D
zR${p$w+2qOILv#!$x*FUP0f_ZbjoB>2%pex5Y-k8rLrvuS$`#ch^Z+vKBRMaGDbhf
zwLvg8icTTnD@@ZJ>q%Uz&)i~C@X<XbMX?NAby$vR#cOS8?2wnRw~7L4mW@i!dbs-%
zfAxi*$Wz0HDxkVD*(JVn*j&ab{S4I$lsk8JP+P@yDdYB%gUelkvUaVA$)Y4(%MF?!
zRi)i3d>bWpRUI9yM#(0FBBMJhUQ3(7PPATSOp-<K-%{ZarP$6<;K4-_#C#@MnkI?A
zz3SUv#+n58p}B9O)(b`aecgqlo%3P5QE*KT^20W#r?4C&&eM%5^_bioJ*UwJoWriN
zcC}hv=?i$tn$+KRbe#xkBmGScAL9EP5gI;3Y@E?hwN|!Eptet*32p#~2X)WE@~w7%
z+!E+t4_w7$;)A6y7SQyOOcm|gLG>f(GmzFT-~$4@SF*5UI=zF}*vG%HlX9)+g<*^a
z_mzWfLuo{1%}d<LNw;anD8b63NyDGzDp<r*8UxD!$>!@%g7kFex9S7uNZH{_z98qG
z8lFZyRabi+B=qQ!Go~<yMSkwIjEIc!YJlu!sOh{$Z<n>8u<+nEBD|Z%KxOMbQ})AL
z0a^+6Gq<{cX3xh4yBc-w96i{1O9^d->PgPsTkwY;%XW6u)NE%RBn>j+!WK{3F|e}H
z;`h{P`I*taAUlDUM@hfvB>77vg)h-8MfY+ZP}ab$a(bpsF_<2c=J(F=<)0pMDejRh
z)6czTI8vS|ZfN0Qi_?jZjgj7<Y%N<K(mW>i&|iHDquC)xNa-p=W2PCnixFWs@%cd3
zBQZZ*2&F>iu)QN}CAWhve-;1U$fw0OVUN|_kNrZ}lRR_l6N}6~TWR+P4b87Rkrc!%
zETUz@$mADkMbKW%_2m!--lbo0Ja>iVSle=se4&JLJlir4SV2roOlh=!EY*2gTPtwe
ziI`Czi9_KSH`3Lh?pc!&0Oa!yc0BX~<*ZUgC5ezvfF)7lk=q*GBVxZ5ojYjj{V`^t
zIE*RdrQoHyAQSj0RWux+&E3t05p`AXK7~8BL_JLK8+0SQZz>{)WV#gNy))fdvmO`)
zeg*MF`DkEL2d$Y@<q6RGcYJ+mUZH!dB-YE`I-=V^BpvoaEh`T>-Ms((qijd&A&csf
z^25aaKDm&Lmv>-ZbfCcD8<~{S^@l99O?dCdwq0U`dI{Xka>a{zZrzm?A@`B}N%3vn
zc{tN=M=&mZJP$+VtS*k;-1|s$CHXl1SrE6}6~|#Z?6*_eq^E!awTl6x_8ysyRtb|M
zv6L`i85m)8tLXg$xUDz-W@5c%G!5{w84udS%kU8lsNh2V+8E$%TWmOFz?yd#Ng^uM
zv2jQyFIs$I-hsl@x#VQOD4~6uGwoGAP4{}3etLPEZmnKv^7Vm5`UydL=vznBy>Q6_
zZ11-{4JrGw$H&nTpVZ1fF~(c;m`xA$44vRMxn@7VIz5p*S8j<k){k5;Z%Gddrw%9m
zXuP-_mZ<L5`M`>PgrKmH6u+ahWEQt)NOzsE6#nx{8aEwlEz)ZVc$zz4!=4z+u7F7N
z#gj8s@7aK^te_8?vYv5QFwQTUh?O3yUk*wfJZ_2gW8HP|_kA0OXU7S@e6BtyasC7$
zm{Qw@kVaGPUe4Z{Fm|?r&q{Qp{})db!Mmm4t{C0^&73D|)8S}Xt|vl`YXH9>WQ4ma
z+15s^p<1J~Te0Umw(qMOUvE^b<CM_Y#YN@uBvQ7v-okk~M?W}2ILxuakmzo#6B?AG
zw(fO~>1v#@T`<|&<)O_A*CJkE;Bt^`B)0F+i%(b4r`W6aq3(5+c0%cdTG5PJdDuuf
z*+tJ74edqgmO)8Qk`Lm(Z^}JM`}YHPa!%`m?$m!nK#16>t{>Vm1(#kSR;!E3JKL;O
zgPHuCbsHJ-D4SsUinK-6=9{=a)yvG6FoQUA+`VL@boE{+sj6y95(}|BwR$l?ROrR<
z-1AW@d%=fD<&n?p6C;OBp3fy1o|&rJ6dRf^#%>S*Jl?(|KVoP;ReZ#uCVPw1IIqdb
z0mN}YPG|N_yX3icsFmND2Ue6iG4fZC!#P&M6O=;KBCJIM4)jCUw}`fQ*#dVwLSMNZ
zmIS`_z8Aftm|P$9K0cWzEfvL+ZXC6<#%%t;T~SOPY^55OwClED2kNFuDy9in2hAUR
z^($3bD@%f3PBcXs<<%o*CJ0z#B`!vLo;e3l8eWZgl8nDw`IJJCr0JnDEIkNdqHM^f
z@9{o&xl+KhU~7)`7(??8MMy9oPFtMN)G!{8q5%ss-x3!}maQK`5r$mnwx(65p^t<8
z#5S5UQ8@;AY>}ltUaMad)@H6h_oO|B2YzRk#->oQGS4SGkC<pa7#EbyE~au1PI@Mu
zJtjmBS#gwGA`#68q2s&@()`xdzM6^fbgToCxp_6B;u*_KtThibG3;n;Bz&=?b*p2_
zx9}Nw>QWXrX%COzEFqQnC^(#^>KzW-9udGT){#_sP)m!Qg$gYT7mgt)1x(2LO%;2n
z>!vTjb`!s+NGTI`*`WZ+5lDEVn_0j$pt8V?!%DsKoUmyobKod{Ls*TGmS}&6DoeId
zqqaN!C@0de+Dq@OIl8O%5yPxLt|jWrFTT7N%<GJ>Znue>jTkU|$vEz>R$+{r>PvHT
z%CBgk#MfcxycYWqodG*NNwhP|{icm;i2P}`jHiT*QukZxCGy#-k(A5cR?meliIpr2
zI@5K$H9q0}`^FnGA>iThY^|5eN2f{vtS5rw7g~B{3=27872K&_ux!Oqk93{(JkQEd
zlD4gM*O<u`jxyi@_)*BhxCQ(mS38)3yUqhf+Qaku={?-7l0g^2>%(8g>QF|j%F>6T
zpJ}I*GO@5vqUXQL%*l?-smC4BeI7NUpYQil+u64@hr?<It9oFZKO<XEM0?=*RGLR*
zPE-8}86xIINQUpCDZ0E$lGQ8U`Sa6cYQV!})3C~z)_Dd%<!)A<oMC|9(>7W{bZxm^
zCv|7kD;4_9c2xs3oIQysk5}Sw)<>tWT=BfETvwsWsHqG%ZS_3Zy3Bfn$Cm%f|Fxhv
zVbKfA6hw#8x>01w_FFS;m$!8%Z^K_al}H2#UHTK0jtYiHQ})&Q>Kwoe3E}Lad5xrS
zEj+GZzt{zqT7Ri^YwAHVHC)|tz3r@CaqOQJx`Z*hwptx$eyfje4k0{*mc4X(V2G|p
zAb};>#YaKXB!84)|C2r%^Q5UuR@2illqPl}e#1?d`&mwxhL;K);KQ_KvG=TkW?3=!
z7f)g;7`&}SGe<FL7hW^*x3Tu8;=N8|c%|xfVq<-9KT2Ld?R253CMvz_gLrgQSP#e5
z@-wkMb{71|G{(~7Kp&9j7iKZCEh6cjdM9k9_U0?oo%^=$zIwPF=~7Z+EpB;G`L~%Z
zQkDT^Vl^uiQ!dVsXK84sRMgM6FHf?EE^<V6wz<?#x6@iObriJ%E|F34S~mYm6mNvy
z{wq)XH^*asMLr;rel`Xc2vz(=6Cq*5pOHw&;qRL?5poFd22Eu7N16yZ3H~>cNcA59
zO~^@@8(~R_Yo2IyO#rWX-ydx6JJb6`@~#iv{WF0BL&%!jAN=e(5Pc&E0loiU2O${Q
zSeb6}vjiV`3{%|BS(%n?#hMI;aChtYymw<mTB6*Y+Es=dKyRuT%TvZ85hT)}Cy1mm
z@~EYIUk;<+T9}KxLCk6K>j(*PC8K=%scKwPYp3v2(#D$EYp#!xvfQmsX&-6(rcwqp
z`;J#kY{;3p<GqgE(sx$R8uS3>$|0F~ZW}x{^`r0VI?TJPG&j>ME;-+PTx6(!3?xb<
z%3caq-fy)itn8i*)C+0ev%oKL+aXy=4`pD&t0sT)l{Py!FLGvOl(j0-roJe1rZn%|
zzBNAl;cJ{R^^Hxo*RKi(V=SE~ET&{u+$#(cJBP#x8w|RKW?$#vj!jq}yL-fsqEq!U
za`QUL1WhRm>B{)mHuT8$JW5r(`kXL)^*F$Lt1#bbeAJ`-W8^Z{hKW|-yY~tg>OAS1
zb#SjdjfH)lkSqHF6wo9E-V7j{-O~(bk|*yGoTNm^d8M5YtgakcNqM_SIP{BvAA-3&
z*Z!;ie7GKgNO>1!YL!4qWy=xxb;2WlmgE=w7W5X4k!~EhR3`kX()^;PL(bov5Y0ex
z(Q3M+3adgfgo>^4-41B{Ds4|VaLAvB=Mb6lH<cCl6ny@0cy!7Q`<%eNQ?Xc$9J@$+
z9s`sp{*D`|U$$U*d2Zgfzpj9k5=leqf$ltzHV{s{#_oaZbc-=rv3qK_vKQe3@%Dh5
z1XrbO=&<?nGt4&+Ewa4$5C!_OaTOEE=^I`))xV52?Z)Gnqi#|&WJ{_gNUVF9oaX)N
z1kt3V@^yC3ySVZpjrt9`w~YZvEQCk*46qF4q+9UcVcDQcbf=?BFdtU-+E1B&s80Rt
ze7Z|+<*2Gw-X1Zqu|wFFT5~Y^&bm-_<@JMm(|#jb7Mxfs_NU>X$8sl0TU1*eos(H&
z0S0^EaGE@fWb{{Nwu~k|Zx?DoZz!mXT@Rbb7$4Fb&?RX)HrkiS=h2w-)m#qk-!`$;
z^}6E1v6!N0ZW0zZ7eD7Rj$0x1LTDjoZ)17OVsm6zy*`Q|<c?|+CPf@5QNS>`cKiTG
z*nVi1PZf^q4e;Q?NUSc<A1=9*qB|V-QB8=R!>9Y-8U%)Vx16l<+cr)5ckOwlciVER
ze0shbdsd$8GNxiVpSZSm+*$ACw&)T^!6Rz2Bqbj^6Q;y^8k?6DA0y<OisWDv@peo_
z4A(mE?xn_EI60I<Tg$CsgybH2txl5<d}T9jz}N_OF|lvww|N=J@Zp}@O6%n;uXt>>
z8>@(-L~IkZZcetkZ5&DZzFbiFM5=y7<hbg%k9R(TC#`$&t;L}?<;-fSKE9Y)Yo>K`
zw5-H-?ZJxHHuj=>_n_txv&d^v4-}M*?XO4l8cl9?kr+6u`I-LYAE+zR781YqoLlmI
z;PP_NuhauFMP-HsX@6Kr(z){>Pn16W?2E9D!IP52lq{KyB9ij%0N;j}@8^;5ckZJ1
zH#(ekL>_-mpB57gWd$m)P-^S*wMh>T^k6MGe#7SBS~O9%C%F5;hdg@#a3KvY?{~|P
zOPyk8K-b8@-RKQ`t}}{Thf?tPS-ayq8nul?3f=6g(n#8U7&J+;b@Z|oJC3C1+EnUw
zAD(}F)6+9TQz6qnG~yT2@C03kAq8i~k!QeVZaL{K(#sXf1ELZ$<%#GI<7%tPY}W3M
zdN>B&pPZbLHHJ1!8@~lHTwPwp@eq)p1)L}v<IZ2nqD>I$n1;Y}+2el+e8nk2B-l%0
z=(RCM*C^`1Xt!i2T-3ka&B2X5;gg~Ftdc|MGU1>eF1U5sfdlT@ED+KAk$cSMsdlK}
zCvJ;cZk}iHxz5is4!5Vc=WCY=ozpRO2T(u4&t8aiJv+^pVvZUJ^6>y-*LK;8CTU0H
zeS5Pg=$${hl}BE1?9wz$rq4E@U=o@4EkDD|zoJb9P<IaN<<I}XteH+tU=b`B7>O^p
zcvwWUu2%kt_fq2O*(okkR<4=u)SWsR$%qdO_f~ftmiy-^>G)k|t7gX=q{cawnf145
zI}f%$qYcos+zmtvkcd8qX`A)k!zX*w+?A?ivlM5Liftwdr+Kfp?-J$+#}E+u-0<7i
z%}=lQl&nHJFgzML>;glm$Qq4V<V&B9z{R)XSmH=I3kXyv8t=N=4)}OJPp#`R^CECi
zK&W&YlGH+Y%!?W}%$q@#7aHBZXP@fD?c8wr>3ICp2~z#yBKJl8r+OFL*;>c5zT;Cp
zl0(}I(e{$RngZA77XQter2pwG<WD;Qa?p^Ci5+sd@wXYk3OQW=+Z2HA{Wm<uf0zP{
z5R>6=9mbr0YY6;0lBDv-UVZ(~6F2ti8^=8USH9F6TeXp~zJZ~>u`%WmJzbA9-M)-W
zH|@vKdy+E1j}o*n5_#t_Cg&urheAoU_1`GY`frWd5U$t|CJD_xMD&V&cuv@Q#jEK1
zG&z>BIIo4XDWjZ{nKw&V;IW2MbJzR#%FSd|9L&6OB>9TX?2Ht=O?lL1qzcXXEIH`n
zYt-19FH+X2YK(JDlZ;h!a(xrr$W{`u)TmC^sZN)lQIMOO!X!SG9%^0T3&Dnkg*Bpt
z^~Zt5c>Fm0;|Z%L|Ko!DS0j|B4F|`;9}HH~H)kIU`y9eu<r}DtyQr#;HJ$5ji$$)a
z6CTeR&XUnK(ZIZFo8B>W@QhUQ36Qy%VM^I@=~X^%)dtos*Qg8byK&UJ)tCa4IbOYf
zYiihENH*;1YEtro$$qAC<`Rkb0>jM}2m98<rPI{kKn;3E@ZUh~Uz}n5g*JA`e~EBH
z&L^|}g4^%&Z>)bH4SM%0G{FdY?jNAW$N~ZDe>wj4>$u?GodSdg4*obs__NY=9N~|%
zf{<!|92EQm!vFE_kC1<H`89R_9>T0_>^C6n6x(l!A%Nd*`huvQ+_I_#t66>AUt};y
zIq#F!=2*C<0-BuR1dFA%bcI31=NIb0Fy5M!Hns$LO7{nc@}$I%{H)go{e+td-FH#<
zi8v~ft=El}EbqI&==7XD^CV(Mb*;Y@u)F&}NXb}6c{ucq<bETcoznl;+PMZdQAGi`
zgW4HdMsPp{0cS0Cwn3A<yGb_LB}peFlt$CElmg*3kS3cp=CL8!khF}|VL+@;h7Rwl
zJe<xr3RGY$6pP41d_e~p9R-KMf(|e~T06)O9P!-kbGKFa(frz;oO|xM_uO-n-JI{%
z^O(jJZueZCHf~*O=bSeu?O(NhYpJcFdrM>b)bBIa9Q|v@#`~_Fu#bDOb8lC5aAjBO
zX7{6W&Rjjn&Ng#(&zL9Ij~Tr?G2hj+HshJtQ0|)d)w8WnzBaLbiZy%9r$2Q&e7*|T
z^wtfV$44&;nrNW=xYv8A$>X{*cEWSvhU0hMyKwky@wP_K7q=wqfv0|6cZAss&u%`t
z62JMzsc%<U!)|wR>#nBjZ4YKG&zpA6-qn%HddD$-QFGR-&G#?M+H`bk=jz<c*_mwH
zYH|JA6W5|GZBzG`mcF#*ZOzNyg&)6tdxq^7^8K@xdVX@#`Q@V*IHveJ_#3CUS5Ml!
zaNCMH&xIWg$J)+Pm$Ew^S-#}^!)?v$pKl}$j_OT$-jnrfjyGoSlAJBgt8y=&p3qWs
z=k>ISEuJ8*pJ!?R?C<IyG`s7n&Pz8AKGfYZlk9HKdFx#MKkeV}JF{lme_d6$%j@it
zaxb?1W}QE3&-F!m@7go&@V0$XZcOUOC-ya8JMcoq;;N3igI7;}zG3{$diSmK%+aMq
z`rNud-~IaI(H}f;d(@@H(_7lFuHN{Hv1(y{cxTzxD~G<U{IIMx^l)%bW&Opm$HY0>
z8JmV?u}oSG&f-3{c%Wp{TjK7k*QE7>)}W;@6PJItBPxZeP|QmHapV({2nDR<0%Hj(
zi93YqP;qTSm|N?q@Yja@T!1XJ<yZ=RaH5W&&G_QlpnS+~l7J#sYo4RrU6h9vGJGZ)
zj{@RMWTKH+lFwL4xd{)~SWJ<GtRjW2q_?DukVG*O4gpB5<8-u^);~=I6Hs=W787BD
z*0Cg*Oay{_IpF9$2H`I&Sq+KJQ&c*g)}{42F;PX)9LG_Jp%_LB8roz{O!8&4vE(%9
zPj)2a?Qa%K>SUjEP<%&)ycDfNR9{DZC5KQkOyG`vm10V=l0h>o1pNWC78wOrOVfg&
zg?fCXHPHdTv680ENHFyE6d&ZFqAPd<`qIM`z!eWfKp+L&r<I&t<|bxWi%F@6+$A_2
zqDiWE<Rjn^L}d7}Ll}DpZvh>?1n}5kjT6oYurY|%YY_{>tw1c#u)M*jMFt)r3#HP8
zkV8`k9IXsTE#E`aL7Ib_iirwX2)4k=P&`=>)(}P=XC@qCyhh%u0fz$grQ8<?`F#;$
zR*gW+6@yY5YQ=lpk!|*xuuwQ`gJx|6=g}80Df$1Ij7ISS$9lX|^rtZ5>vG=X@)^O1
z2mJkR##4!ioFM^!Ux|1Uf=+{<0Y4uU6Hp!xIU0`xKrR=`Q?NKmn+4vKLhuW}P0l}i
zMC!SSgm6Cb5nrszN@lbHA?Qm*B+?d7h?RgeF9U<i#5fE)6%!IVh-XhcA|YGPg=~l#
zFg{GQWDz7h1di(8PHZ{E2)#zm(7+xF@V=Ey;~jD+#y{|Fp)cbLhD&%2d=tk>1RD=d
zqXH+7a}cq}h**Y|(~m@m1T{$0yJ+<0LxG%5OA5kE?meNvegzJD!G8u+(|-&_CLshd
z1ZT*>dtuLuV2wr~=z~PBj5TAt8NY~FuZ+=~a7l$ih5&K|o>C<<V8BW#LkM#+N|}L%
zvKW<&HDm5msmHM<cvPs60hftCL=`f`!K|xf&;zE@lzL_cQ!fe`&EOIpl?)83eP(Dq
zCQuY@k%=DKFVdUU{?UlRM{+3IGROops8WVZ26zUnl+m;vgNBN>h~@B|r&@-gLY2&j
z&umc2AVli%B32Jns!Gp<4C*-mmociOut&obsQaK97H%q3dPc(#UuZK2q^?qrK^%r%
z6n!u>ixGZ>3;}09WE>`L=vp(IFsQC<%Na4$sgMD)4FJ4aHk=+ibPd5IsACBbC?NFx
z_N0UlK%NBr;{$@48>$r`2MIXdN0bP;c8Q=|v7ktR5dzNlaXE>_7CXbz!2a8dXnT=)
srU?}}k;&{}S<Yd1(srZCc6W`X_^caDs!MX>GkOkg<MQ)ol@{gv3zsc^V*mgE
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/testwrapper.sh b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/testwrapper.sh
deleted file mode 100755
index e08593a876..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/testwrapper.sh
+++ /dev/null
@@ -1,11 +0,0 @@
-#!/bin/bash
-COUNT=50
-for Z in `seq $COUNT`
-do
- for T in `./unitary.py --list-techniques $@`;
- do
- echo $Z/$COUNT $T
- ./unitary.py --technique=$T $@
- done
-done
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/unitary.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/unitary.py
deleted file mode 100755
index cfa5fe1141..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/examples/unitary/unitary.py
+++ /dev/null
@@ -1,136 +0,0 @@
-#!/usr/bin/env python
-#
-# This is a quantum control example motivated by the experimental need
-# to synthesize unitary matrices in SU(2) in optimal time, given an
-# explicit and finite control set generating the whole space, and an
-# admissible error.
-#
-# See problem_description.pdf for additional details.
-#
-# Contributed by Clarice D. Aiello <clarice@mit.edu>
-#
-
-import adddeps # fix sys.path
-
-import argparse
-import logging
-import math
-import random
-import sys
-
-try:
- import numpy as np
-except:
- print >> sys.stderr, '''
-
-ERROR: import numpy failed, please install numpy
-
-Possible things to try:
- ../../venv/bin/pip install numpy
- ../../venv/bin/easy_install numpy
- sudo apt-get install python-numpy
-
-'''
- raise
-
-import opentuner
-
-from math import sqrt
-import cla_func
-from input_generator import (generate_random_Ugoal_HARD,
- generate_random_Ugoal_EASY,
- generate_random_Ugoal_RANDOM)
-
-from opentuner.search.manipulator import (ConfigurationManipulator,
- SwitchParameter,
- IntegerParameter,
- FloatParameter)
-
-
-def generate_random_Ugoal_FIXED(**kwargs):
- Ag = -1 / sqrt(10);
- Bg = sqrt(2) / sqrt(10);
- Cg = -sqrt(3) / sqrt(10);
- Dg = -sqrt(4) / sqrt(10);
- return cla_func.np.matrix(
- [[Ag + Cg * 1j, Bg + Dg * 1j], [-Bg + Dg * 1j, Ag - Cg * 1j]])
-
-
-log = logging.getLogger(__name__)
-
-generators = {
- 'hard': generate_random_Ugoal_HARD,
- 'easy': generate_random_Ugoal_EASY,
- 'random': generate_random_Ugoal_RANDOM,
- 'fixed': generate_random_Ugoal_FIXED,
-}
-
-parser = argparse.ArgumentParser(parents=opentuner.argparsers())
-parser.add_argument('--seq-len', type=int, default=10,
- help='maximum length for generated sequence')
-parser.add_argument('--goal-type', choices=generators.keys(), default='hard',
- help='method used to generate goal')
-parser.add_argument('--goal-n', type=int, default=100,
- help='argument to ugoal generator')
-parser.add_argument('--goal-alpha', type=float,
- default=random.random() * math.pi,
- help='argument to ugoal generator')
-
-
-class Unitary(opentuner.measurement.MeasurementInterface):
- def __init__(self, *pargs, **kwargs):
- super(Unitary, self).__init__(*pargs, **kwargs)
-
- self.op = cla_func.Op()
- self.num_operators = len(self.op.M)
- self.Ugoal = generators[args.goal_type](N=args.goal_n,
- alpha=args.goal_alpha)
-
-
- def run(self, desired_result, input, limit):
- cfg = desired_result.configuration.data
-
- sequence = [cfg[i] for i in xrange(self.args.seq_len)
- if cfg[i] < self.num_operators]
- # sequence can be shorter than self.args.seq_len with null operator
-
- if len(sequence) > 0:
- accuracy = cla_func.calc_fidelity(sequence, self.op, self.Ugoal)
- # ~.99 is acceptable
- else:
- accuracy = 0.0
-
- return opentuner.resultsdb.models.Result(time=0.0,
- accuracy=accuracy,
- size=len(sequence))
-
- def manipulator(self):
- manipulator = ConfigurationManipulator()
- for d in xrange(self.args.seq_len):
- # we add 1 to num_operators allow a ignored 'null' operator
- manipulator.add_parameter(SwitchParameter(d, self.num_operators + 1))
- return manipulator
-
- def save_final_config(self, configuration):
- '''
- called at the end of autotuning with the best resultsdb.models.Configuration
- '''
- cfg = configuration.data
- sequence = [cfg[i] for i in xrange(self.args.seq_len)
- if cfg[i] < self.num_operators]
- print "Final sequence", sequence
-
- def objective(self):
- # we could have also chosen to store 1.0 - accuracy in the time field
- # and use the default MinimizeTime() objective
- return opentuner.search.objective.MaximizeAccuracyMinimizeSize()
-
-
-if __name__ == '__main__':
- args = parser.parse_args()
- Unitary.main(args)
-
-
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/gen-venv-bootstrap.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/gen-venv-bootstrap.py
deleted file mode 100755
index ff159bb108..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/gen-venv-bootstrap.py
+++ /dev/null
@@ -1,39 +0,0 @@
-#!./venv/bin/python
-
-extra = '''
-
-default_target_dir = 'venv'
-
-pip_install_packages = filter(len, open('requirements.txt').readlines())
-
-import os
-import subprocess
-import sys
-
-def adjust_options(options, args):
- if len(args)==0:
- os.chdir(os.path.dirname(__file__))
- args.append(default_target_dir)
-
-def after_install(options, home_dir):
- from os.path import join
- pip = join(home_dir, 'bin/pip')
- if not os.path.exists(pip):
- # on windows
- pip = join(home_dir, 'Scripts/pip.exe')
- if not os.path.exists(pip):
- print "error", pip, "is missing"
- if sys.version_info < (2, 7):
- subprocess.call([pip, 'install', 'importlib'])
- for prog in pip_install_packages:
- subprocess.call([pip, 'install', prog])
-
-'''
-
-import os
-import virtualenv
-
-os.chdir(os.path.dirname(__file__))
-output = virtualenv.create_bootstrap_script(extra)
-f = open('venv-bootstrap.py', 'w').write(output)
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/install_reqs.sh b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/install_reqs.sh
deleted file mode 100644
index e671a5f2a1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/install_reqs.sh
+++ /dev/null
@@ -1,3 +0,0 @@
-pip2 install sqlalchemy
-pip2 install psutil
-pip2 install opentuner
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/misc/livedisplay.gnuplot b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/misc/livedisplay.gnuplot
deleted file mode 100644
index 1d4f130213..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/misc/livedisplay.gnuplot
+++ /dev/null
@@ -1,10 +0,0 @@
-
-set terminal x11
-set xlabel "Autotuning Seconds"
-set ylabel "Runtime Seconds"
-set xrange [0:600]
-
-plot "/tmp/livedisplay.dat" u 1:2 w lp lw 3 title "Best Execution Time", \
- "/tmp/livedisplaydetails.dat" w p lw 2 title "Tests (excluding timeouts)"
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/misc/livedisplay.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/misc/livedisplay.py
deleted file mode 100755
index 5aa3d552d8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/misc/livedisplay.py
+++ /dev/null
@@ -1,57 +0,0 @@
-#!/usr/bin/env python
-import os
-import argparse
-import subprocess
-import time
-
-parser = argparse.ArgumentParser()
-parser.add_argument('--gnuplot-filename', default='livedisplay.gnuplot')
-parser.add_argument('--data', default='/tmp/livedisplay.dat')
-parser.add_argument('--details', default='/tmp/livedisplaydetails.dat')
-parser.add_argument('--xrange', type=float, default=300)
-parser.add_argument('--yrange', type=float, default=.05)
-parser.add_argument('--yrange2', type=float, default=1.0)
-parser.add_argument('--remote')
-args = parser.parse_args()
-
-if args.remote:
- if os.path.exists(args.data):
- os.unlink(args.data)
- if os.path.exists(args.details):
- os.unlink(args.details)
- syncproc = subprocess.Popen(
- ["ssh", args.remote, "tail -f -n10000 " + args.data],
- stdout=open(args.data, "w"))
- syncproc2 = subprocess.Popen(
- ["ssh", args.remote, "tail -f -n10000 " + args.details],
- stdout=open(args.details, "w"))
-
-while '\n' not in open(args.data).read():
- time.sleep(1)
-while '\n' not in open(args.details).read():
- time.sleep(1)
-
-p1 = subprocess.Popen(["gnuplot"], stdin=subprocess.PIPE)
-p1.stdin.write(open(args.gnuplot_filename).read())
-print >> p1.stdin, 'set title "Zoomed out"'
-print >> p1.stdin, "set xrange [0:%f]" % args.xrange
-print >> p1.stdin, "set yrange [0:%f]" % args.yrange2
-p1.stdin.flush()
-
-time.sleep(1)
-
-p2 = subprocess.Popen(["gnuplot"], stdin=subprocess.PIPE)
-p2.stdin.write(open(args.gnuplot_filename).read())
-print >> p2.stdin, 'set title "Zoomed in"'
-print >> p2.stdin, "set xrange [0:%f]" % args.xrange
-print >> p2.stdin, "set yrange [0:%f]" % args.yrange
-p2.stdin.flush()
-
-procs = [p1, p2]
-
-while True:
- time.sleep(1)
- for p in procs:
- print >> p.stdin, "replot"
- p.stdin.flush()
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/__init__.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/__init__.py
deleted file mode 100644
index 09a5dead02..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/__init__.py
+++ /dev/null
@@ -1,41 +0,0 @@
-
-import measurement
-import resultsdb
-import search
-import tuningrunmain
-from opentuner.measurement import MeasurementInterface
-from opentuner.resultsdb.models import Configuration
-from opentuner.resultsdb.models import DesiredResult
-from opentuner.resultsdb.models import Result
-from opentuner.resultsdb.models import TuningRun
-from opentuner.search.manipulator import ConfigurationManipulator
-from opentuner.search.manipulator import EnumParameter
-from opentuner.search.manipulator import FloatParameter
-from opentuner.search.manipulator import IntegerParameter
-from opentuner.search.manipulator import LogFloatParameter
-from opentuner.search.manipulator import LogIntegerParameter
-from opentuner.search.manipulator import PermutationParameter
-from opentuner.search.manipulator import ScheduleParameter
-from opentuner.search.manipulator import SwitchParameter
-from opentuner.tuningrunmain import init_logging
-
-
-def argparsers():
- """
- return a list of ArguementParser to be used as parents to the user's
- """
- return [
- measurement.driver.argparser,
- measurement.interface.argparser,
- search.driver.argparser,
- search.plugin.argparser,
- search.technique.argparser,
- #stats.argparser,
- tuningrunmain.argparser,
- ]
-
-
-def default_argparser():
- import argparse
- return argparse.ArgumentParser(parents=argparsers())
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/api.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/api.py
deleted file mode 100644
index 19a2f60935..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/api.py
+++ /dev/null
@@ -1,87 +0,0 @@
-from datetime import datetime
-from opentuner import tuningrunmain
-
-
-class TuningRunManager(tuningrunmain.TuningRunMain):
- """
- This class manages a tuning run in a "slave" configuration, where main()
- is controlled by an another program.
- """
- def __init__(self, measurement_interface, args, **kwargs):
- super(TuningRunManager, self).__init__(measurement_interface, args, **kwargs)
- self.init()
- self.tuning_run.state = 'RUNNING'
- self.commit(force=True)
- self.search_driver.external_main_begin()
-
- def get_next_desired_result(self):
- """
- Returns a opentuner.resultsdb.DesiredResult that should be tested next.
- """
- dr = self.measurement_driver.query_pending_desired_results().first()
- if dr is None:
- self.search_driver.external_main_generation()
- dr = self.measurement_driver.query_pending_desired_results().first()
- if dr is None:
- return None
- self.measurement_driver.claim_desired_result(dr)
- dr.limit = self.measurement_driver.run_time_limit(dr)
- return dr
-
- def get_desired_results(self):
- """
- Returns a list of all opentuner.resultsdb.DesiredResult that should be tested next.
- """
- drs = self.measurement_driver.query_pending_desired_results().all()
- if len(drs) == 0:
- self.search_driver.external_main_generation()
- drs = self.measurement_driver.query_pending_desired_results().all()
- if len(drs) == 0:
- return []
- for dr in drs:
- self.measurement_driver.claim_desired_result(dr)
- dr.limit = self.measurement_driver.run_time_limit(dr)
-
- return drs
-
- def report_result(self, desired_result, result, result_input=None):
- """
- Report a measured result. desired_result should have been returned by
- get_next_desired_result().
- """
- self.measurement_driver.report_result(desired_result, result, result_input)
-
- def get_best_configuration(self):
- """
- The best configuration found so far. From the current tuning run only.
- """
- try:
- return self.search_driver.best_result.configuration.data
- except AttributeError:
- return None
-
- def get_best_result(self):
- """
- The best result found so far. From the current tuning run only.
- """
- try:
- return self.search_driver.best_result
- except AttributeError:
- return None
-
- def finish(self):
- """
- Called at the end of the tuning process to call hooks and close database
- connections.
- """
- self.search_driver.external_main_end()
- self.measurement_interface.save_final_config(
- self.search_driver.best_result.configuration)
- self.tuning_run.final_config = self.search_driver.best_result.configuration
- self.tuning_run.state = 'COMPLETE'
- self.tuning_run.end_date = datetime.now()
- self.commit(force=True)
- self.session.close()
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/driverbase.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/driverbase.py
deleted file mode 100644
index 5486889c0d..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/driverbase.py
+++ /dev/null
@@ -1,48 +0,0 @@
-from opentuner.resultsdb.models import *
-
-
-class DriverBase(object):
- """
- shared base class between MeasurementDriver and SearchDriver
- """
-
- def __init__(self,
- session,
- tuning_run,
- objective,
- tuning_run_main,
- args,
- **kwargs):
- self.args = args
- self.objective = objective
- self.session = session
- self.tuning_run_main = tuning_run_main
- self.tuning_run = tuning_run
- self.program = tuning_run.program
-
- def results_query(self,
- generation=None,
- objective_ordered=False,
- config=None):
- q = self.session.query(Result)
- q = q.filter_by(tuning_run=self.tuning_run)
-
- if config:
- q = q.filter_by(configuration=config)
-
- if generation is not None:
- subq = (self.session.query(DesiredResult.result_id)
- .filter_by(tuning_run=self.tuning_run,
- generation=generation))
- q = q.filter(Result.id.in_(subq.subquery()))
-
- if objective_ordered:
- q = self.objective.result_order_by(q)
-
- return q
-
- def requests_query(self):
- q = self.session.query(DesiredResult).filter_by(tuning_run=self.tuning_run)
- return q
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/#interface.py# b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/#interface.py#
deleted file mode 100644
index 4fe23da5d9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/#interface.py#
+++ /dev/null
@@ -1,359 +0,0 @@
-
-import abc
-import argparse
-import errno
-import hashlib
-import logging
-import os
-import re
-import signal
-import subprocess
-import threading
-import time
-from multiprocessing.pool import ThreadPool
-
-try:
- import resource
-except ImportError:
- resource = None
-
-try:
- import fcntl
-except ImportError:
- fcntl = None
-
-import opentuner
-from opentuner import resultsdb
-from opentuner.resultsdb.models import *
-
-log = logging.getLogger(__name__)
-
-argparser = argparse.ArgumentParser(add_help=False)
-argparser.add_argument('--parallel-compile', action='store_true',
- default=False,
- help="present if compiling can be done in parallel")
-
-the_io_thread_pool = None
-
-
-class MeasurementInterface(object):
- """
- abstract base class for compile and measurement
- """
- __metaclass__ = abc.ABCMeta
-
- def __init__(self,
- args=None,
- project_name=None,
- program_name='unknown',
- program_version='unknown',
- manipulator=None,
- objective=None,
- input_manager=None):
- self.args = args
- self._project = project_name
- self._program = program_name
- self._version = program_version
- self._objective = objective
- self._manipulator = manipulator
- self._input_manager = input_manager
-
- self.pids = []
- self.pid_lock = threading.Lock()
- self.parallel_compile = args.parallel_compile
- # If parallel_compile is False then compile_and_run() will be invoked
- # sequentially otherwise the driver first invokes compile() in parallel
- # followed by run_precompiled() sequentially
-
- def compile(self, config_data, id):
- """
- Compile in PARALLEL according to the configuration in config_data
- (obtained from desired_result.configuration) Should use id parameter
- to determine output location of executable Return value will be passed
- to run_precompiled as compile_result, useful for storing error/timeout
- information
- """
- if self.parallel_compile:
- raise RuntimeError('MeasurementInterface.compile() not implemented for',
- 'parallel compilation')
- pass
-
- def run_precompiled(self, desired_result, input, limit, compile_result, id):
- """
- Run the given desired_result SEQUENTIALLY on input and produce a Result()
- Abort early if limit (in seconds) is reached Assume that the executable
- to be measured has already been compiled to an executable corresponding to
- identifier id by compile() The compile_result is the return result of compile(),
- and it will be None if compile() was not called
- """
- if self.parallel_compile:
- raise RuntimeError('MeasurementInterface.run_precompiled() not implemented',
- 'for parallel compilation')
- pass
-
- def cleanup(self, id):
- """
- Clean up any temporary files associated with the executable
- """
- pass
-
- def pre_process(self):
- """
- The process before each iteration This method will be called
- once per iteration before all threads are launched
- """
- pass
-
- def post_process(self):
- """
- The process after each iteration This method will be called
- once per iteration after all threads are committed
- """
- pass
-
- def extra_convergence_criteria(self, result):
- """
- The extra convergence criteria which returns True if the
- current result is acceptable by the user
- """
- return False
-
- #@abc.abstractmethod
- def compile_and_run(self, desired_result, input, limit):
- """
- Compile and run the given desired_result on input and produce a
- Result(), abort early if limit (in seconds) is reached This function
- is only used for sequential execution flow
-
- FIXME: Shoud uncomment @abc.abstractmethod Now comment out for
- compatiability
- """
- return self.run(desired_result, input, limit)
-
- def run(self, desired_result, input, limit):
- """
- run the given desired_result on input and produce a Result(),
- abort early if limit (in seconds) is reached
- """
- return opentuner.resultdb.models.Result()
-
- def save_final_config(self, config):
- """
- called at the end of autotuning with the best resultsdb.models.Configuration
- """
- try:
- config_str = repr(config.data)
- if len(config_str) > 256:
- config_str = config_str[:256] + '...'
- log.info('final configuration: %s', config_str)
- log.info('you may want to implement save_final_config(), to store this')
- except:
- log.error('error printing configuration', exc_info=True)
-
- def db_program_version(self, session):
- """return a version identifier for the program being tuned"""
- return resultsdb.models.ProgramVersion.get(
- session=session,
- project=self.project_name(),
- name=self.program_name(),
- version=self.program_version(),
- parameter_info=self.manipulator().parameters_to_json(),
- )
-
- def set_driver(self, measurement_driver):
- self.driver = measurement_driver
-
- def project_name(self):
- if self._project is not None:
- return self._project
- autoname = re.sub('(Measurement?)Interface$', '', self.__class__.__name__)
- if autoname:
- return autoname
- else:
- return 'unknown'
-
- def program_name(self):
- return self._program
-
- def program_version(self):
- return self._version
-
- def file_hash(self, filename):
- """helper used to generate program versions"""
- return hashlib.sha256(open(filename).read()).hexdigest()
-
- def manipulator(self):
- """
- called once to create the search.manipulator.ConfigurationManipulator
- """
- if self._manipulator is None:
- msg = ('MeasurementInterface.manipulator() must be implemented or a '
- '"manipulator=..." must be provided to the constructor')
- log.error(msg)
- raise Exception(msg)
- return self._manipulator
-
- def objective(self):
- """
- called once to create the search.objective.SearchObjective
- """
- if self._objective is None:
- from ..search.objective import MinimizeTime
-
- return MinimizeTime()
- return self._objective
-
- def input_manager(self):
- """
- called once to create the measurement.inputmanager.InputManager
- """
- if self._objective is None:
- from .inputmanager import FixedInputManager
-
- return FixedInputManager()
- return self._input_manager
-
- def seed_configurations(self):
- """
- Extra seed configuration objects to add to those given on the command line.
- Configuration objects (typically dictionaries) not database objects.
- """
- return []
-
- def kill_all(self):
- self.pid_lock.acquire()
- for pid in self.pids:
- goodkillpg(pid)
- self.pids = []
- self.pid_lock.release()
-
- def call_program(self, cmd, limit=None, memory_limit=None, **kwargs):
- """
- call cmd and kill it if it runs for longer than limit
-
- returns dictionary like
- {'returncode': 0,
- 'stdout': '', 'stderr': '',
- 'timeout': False, 'time': 1.89}
- """
- the_io_thread_pool_init(self.args.parallelism)
- if limit is float('inf'):
- limit = None
- if type(cmd) in (str, unicode):
- kwargs['shell'] = True
- killed = False
- t0 = time.time()
- p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE,
- preexec_fn=preexec_setpgid_setrlimit(memory_limit),
- **kwargs)
- # Add p.pid to list of processes to kill in case of keyboardinterrupt
- self.pid_lock.acquire()
- self.pids.append(p.pid)
- self.pid_lock.release()
-
- try:
- stdout_result = the_io_thread_pool.apply_async(p.stdout.read)
- stderr_result = the_io_thread_pool.apply_async(p.stderr.read)
- while p.returncode is None:
- if limit is None:
- goodwait(p)
- elif limit and time.time() > t0 + limit:
- killed = True
- goodkillpg(p.pid)
- goodwait(p)
- else:
- # still waiting...
- sleep_for = limit - (time.time() - t0)
- if not stdout_result.ready():
- stdout_result.wait(sleep_for)
- elif not stderr_result.ready():
- stderr_result.wait(sleep_for)
- else:
- #TODO(jansel): replace this with a portable waitpid
- time.sleep(0.001)
- p.poll()
- except:
- if p.returncode is None:
- goodkillpg(p.pid)
- raise
- finally:
- # No longer need to kill p
- self.pid_lock.acquire()
- if p.pid in self.pids:
- self.pids.remove(p.pid)
- self.pid_lock.release()
-
- t1 = time.time()
- return {'time': float('inf') if killed else (t1 - t0),
- 'timeout': killed,
- 'returncode': p.returncode,
- 'stdout': stdout_result.get(),
- 'stderr': stderr_result.get()}
-
- def prefix_hook(self, session):
- pass
-
- @classmethod
- def main(cls, args, *pargs, **kwargs):
- from opentuner.tuningrunmain import TuningRunMain
-
- return TuningRunMain(cls(args, *pargs, **kwargs), args).main()
-
-
-class DefaultMeasurementInterface(MeasurementInterface):
- def run(self, desired_result, input, limit):
- raise RuntimeError('MeasurementInterface.run() not implemented')
-
-
-def preexec_setpgid_setrlimit(memory_limit):
- if resource is not None:
- def _preexec():
- os.setpgid(0, 0)
- try:
- resource.setrlimit(resource.RLIMIT_CORE, (0, 0))
- except ValueError:
- pass # No permission
- if memory_limit:
- try:
- (soft, hard) = resource.getrlimit(resource.RLIMIT_AS)
- resource.setrlimit(resource.RLIMIT_AS, (min(soft, memory_limit),
- min(hard, memory_limit)))
- except ValueError:
- pass # No permission
- return _preexec
-
-
-def the_io_thread_pool_init(parallelism=1):
- global the_io_thread_pool
- if the_io_thread_pool is None:
- the_io_thread_pool = ThreadPool(2 * parallelism)
- # make sure the threads are started up
- the_io_thread_pool.map(int, range(2 * parallelism))
-
-
-def goodkillpg(pid):
- """
- wrapper around kill to catch errors
- """
- log.debug("killing pid %d", pid)
- try:
- if hasattr(os, 'killpg'):
- os.killpg(pid, signal.SIGKILL)
- else:
- os.kill(pid, signal.SIGKILL)
- except:
- log.error('error killing process %s', pid, exc_info=True)
-
-
-def goodwait(p):
- """
- python doesn't check if its system calls return EINTR, retry if it does
- """
- while True:
- try:
- rv = p.wait()
- return rv
- except OSError, e:
- if e.errno != errno.EINTR:
- raise
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/.#interface.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/.#interface.py
deleted file mode 120000
index 68c6820130..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/.#interface.py
+++ /dev/null
@@ -1 +0,0 @@
-hashim@hashim-VirtualBox.2708:1511328915
\ No newline at end of file
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/__init__.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/__init__.py
deleted file mode 100644
index c289e8d6f5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/__init__.py
+++ /dev/null
@@ -1,6 +0,0 @@
-
-import driver
-import interface
-from interface import MeasurementInterface
-from driver import MeasurementDriver
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/driver.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/driver.py
deleted file mode 100644
index d00886920a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/driver.py
+++ /dev/null
@@ -1,271 +0,0 @@
-import argparse
-import logging
-import time
-import socket
-import os
-from multiprocessing.pool import ThreadPool
-from datetime import datetime
-
-from sqlalchemy.exc import SQLAlchemyError
-from sqlalchemy.orm.exc import NoResultFound
-
-from opentuner.driverbase import DriverBase
-from opentuner.resultsdb.models import *
-
-log = logging.getLogger(__name__)
-
-argparser = argparse.ArgumentParser(add_help=False)
-argparser.add_argument('--machine-class',
- help="name of the machine class being run on")
-
-
-class MeasurementDriver(DriverBase):
- """
- manages the measurement process, reading DesiredResults and creating Results
- """
-
- def __init__(self,
- measurement_interface,
- input_manager,
- **kwargs):
- super(MeasurementDriver, self).__init__(**kwargs)
-
- if not self.args.machine_class:
- self.args.machine_class = 'default'
-
- self.interface = measurement_interface
- self.input_manager = input_manager
- self.commit = self.tuning_run_main.commit
- self.upper_limit_multiplier = 10.0
- self.default_limit_multiplier = 2.0
-
- self.laptime = time.time()
- self.machine = self.get_machine()
-
- def get_machine(self):
- """
- get (or create) the machine we are currently running on
- """
- hostname = socket.gethostname()
- try:
- self.session.flush()
- return self.session.query(Machine).filter_by(name=hostname).one()
- except sqlalchemy.orm.exc.NoResultFound:
- m = Machine(name=hostname,
- cpu=_cputype(),
- cores=_cpucount(),
- memory_gb=_memorysize() / (
- 1024.0 ** 3) if _memorysize() else 0,
- machine_class=self.get_machine_class())
- self.session.add(m)
- return m
-
- def get_machine_class(self):
- """
- get (or create) the machine class we are currently running on
- """
- return MachineClass.get(self.session, name=self.args.machine_class)
-
- def run_time_limit(self, desired_result, default=3600.0 * 24 * 365 * 10):
- """return a time limit to apply to a test run (in seconds)"""
- best = self.results_query(objective_ordered=True).first()
- if best is None:
- if desired_result.limit:
- return desired_result.limit
- else:
- return default
-
- if desired_result.limit:
- return min(desired_result.limit, self.upper_limit_multiplier * best.time)
- else:
- return self.default_limit_multiplier * best.time
-
- def report_result(self, desired_result, result, input=None):
- result.configuration = desired_result.configuration
- result.input = input
- result.machine = self.machine
- result.tuning_run = self.tuning_run
- result.collection_date = datetime.now()
- self.session.add(result)
- desired_result.result = result
- desired_result.state = 'COMPLETE'
- self.input_manager.after_run(desired_result, input)
- result.collection_cost = self.lap_timer()
- self.session.flush() # populate result.id
- log.debug(
- 'Result(id=%d, cfg=%d, time=%.4f, accuracy=%.2f, collection_cost=%.2f)',
- result.id,
- result.configuration.id,
- result.time,
- result.accuracy if result.accuracy is not None else float('NaN'),
- result.collection_cost)
- self.commit()
-
- def run_desired_result(self, desired_result, compile_result=None,
- exec_id=None):
- """
- create a new Result using input manager and measurment interface
- Optional compile_result paramater can be passed to run_precompiled as
- the return value of compile()
- Optional exec_id paramater can be passed to run_precompiled in case of
- locating a specific executable
- """
- desired_result.limit = self.run_time_limit(desired_result)
-
- input = self.input_manager.select_input(desired_result)
- self.session.add(input)
- self.session.flush()
-
- log.debug('running desired result %s on input %s', desired_result.id,
- input.id)
-
- self.input_manager.before_run(desired_result, input)
-
- if self.interface.parallel_compile:
- result = self.interface.run_precompiled(desired_result, input,
- desired_result.limit,
- compile_result, exec_id)
- else:
- result = self.interface.compile_and_run(desired_result, input,
- desired_result.limit)
-
- self.report_result(desired_result, result, input)
-
- def lap_timer(self):
- """return the time elapsed since the last call to lap_timer"""
- t = time.time()
- r = t - self.laptime
- self.laptime = t
- return r
-
- def claim_desired_result(self, desired_result):
- """
- claim a desired result by changing its state to running
- return True if the result was claimed for this process
- """
- self.commit()
- try:
- self.session.refresh(desired_result)
- if desired_result.state == 'REQUESTED':
- desired_result.state = 'RUNNING'
- desired_result.start_date = datetime.now()
- self.commit()
- return True
- except SQLAlchemyError:
- self.session.rollback()
- return False
-
- def query_pending_desired_results(self):
- q = (self.session.query(DesiredResult)
- .filter_by(tuning_run=self.tuning_run,
- state='REQUESTED')
- .order_by(DesiredResult.generation,
- DesiredResult.priority.desc()))
- return q
-
- def process_all(self):
- """
- process all desired_results in the database
- """
- self.lap_timer() # reset timer
- q = self.query_pending_desired_results()
-
- if self.interface.parallel_compile:
- desired_results = []
- thread_args = []
-
- def compile_result(args):
- interface, data, result_id = args
- return interface.compile(data, result_id)
-
- for dr in q.all():
- if self.claim_desired_result(dr):
- desired_results.append(dr)
- thread_args.append((self.interface, dr.configuration.data, dr.id))
- if len(desired_results) == 0:
- return
- thread_pool = ThreadPool(len(desired_results))
- # print 'Compiling %d results' % len(thread_args)
- try:
- # Use map_async instead of map because of bug where keyboardinterrupts are ignored
- # See http://stackoverflow.com/questions/1408356/keyboard-interrupts-with-pythons-multiprocessing-pool
- compile_results = thread_pool.map_async(compile_result,
- thread_args).get(9999999)
- except Exception:
- # Need to kill other processes because only one thread receives
- # exception
- self.interface.kill_all()
- raise
- # print 'Running %d results' % len(thread_args)
- for dr, compile_result in zip(desired_results, compile_results):
- # Make sure compile was successful
- self.run_desired_result(dr, compile_result, dr.id)
- try:
- self.interface.cleanup(dr.id)
- except RuntimeError, e:
- print e
- # print 'Done!'
- thread_pool.close()
- else:
- for dr in q.all():
- if self.claim_desired_result(dr):
- self.run_desired_result(dr)
-
-
-def _cputype():
- try:
- return re.search(r"model name\s*:\s*([^\n]*)",
- open("/proc/cpuinfo").read()).group(1)
- except:
- pass
- try:
- # for OS X
- import subprocess
-
- return subprocess.Popen(["sysctl", "-n", "machdep.cpu.brand_string"],
- stdout=subprocess.PIPE).communicate()[0].strip()
- except:
- log.warning("failed to get cpu type")
- return "unknown"
-
-
-def _cpucount():
- try:
- return int(os.sysconf("SC_NPROCESSORS_ONLN"))
- except:
- pass
- try:
- return int(os.sysconf("_SC_NPROCESSORS_ONLN"))
- except:
- pass
- try:
- return int(os.environ["NUMBER_OF_PROCESSORS"])
- except:
- pass
- try:
- return int(os.environ["NUM_PROCESSORS"])
- except:
- log.warning("failed to get the number of processors")
- return 1
-
-
-def _memorysize():
- try:
- return int(os.sysconf("SC_PHYS_PAGES") * os.sysconf("SC_PAGE_SIZE"))
- except:
- pass
- try:
- return int(os.sysconf("_SC_PHYS_PAGES") * os.sysconf("_SC_PAGE_SIZE"))
- except:
- pass
- try:
- # for OS X
- import subprocess
-
- return int(subprocess.Popen(["sysctl", "-n", "hw.memsize"],
- stdout=subprocess.PIPE)
- .communicate()[0].strip())
- except:
- log.warning("failed to get total memory")
- return 1024 ** 3
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/inputmanager.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/inputmanager.py
deleted file mode 100644
index 7acaeaa0cf..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/inputmanager.py
+++ /dev/null
@@ -1,76 +0,0 @@
-import abc
-import opentuner
-from opentuner.resultsdb.models import *
-
-
-class InputManager(object):
- """
- abstract base class for compile and measurement
- """
- __metaclass__ = abc.ABCMeta
-
- def set_driver(self, measurement_driver):
- self.driver = measurement_driver
- self.session = measurement_driver.session
- self.program = measurement_driver.tuning_run.program
-
- @abc.abstractmethod
- def select_input(self, desired_result):
- """
- select the input to be used to test desired_result
- """
- return opentuner.resultsdb.models.Input()
-
-
- def before_run(self, desired_result, input):
- """hook called before an input is used"""
- pass
-
- def after_run(self, desired_result, input):
- """hook called after an input is used"""
- pass
-
- def get_input_class(self):
- return None
-
-
-class FixedInputManager(InputManager):
- """
- an input manage that produces a single input for all tests
- """
-
- def __init__(self,
- input_class_name='fixed',
- size=-1,
- path=None,
- extra=None):
- self.input_class_name = input_class_name
- self.size = size
- self.path = path
- self.extra = extra
- self.the_input = None
- super(FixedInputManager, self).__init__()
-
-
- def get_input_class(self):
- return InputClass.get(self.session,
- program=self.program,
- name=self.input_class_name,
- size=self.size)
-
- def create_input(self, desired_result):
- """create the fixed input database object, result will be cached"""
- return Input(input_class=self.get_input_class(),
- path=self.path,
- extra=self.extra)
-
- def select_input(self, desired_result):
- if self.the_input is None:
- self.the_input = self.create_input(desired_result)
- return self.the_input
-
-
-
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/interface.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/interface.py
deleted file mode 100644
index 1749024882..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/measurement/interface.py
+++ /dev/null
@@ -1,366 +0,0 @@
-import abc
-import argparse
-import errno
-import hashlib
-import logging
-import os
-import re
-import signal
-import subprocess
-import threading
-import time
-from multiprocessing.pool import ThreadPool
-
-try:
- import resource
-except ImportError:
- resource = None
-
-try:
- import fcntl
-except ImportError:
- fcntl = None
-
-import opentuner
-from opentuner import resultsdb
-from opentuner.resultsdb.models import *
-
-log = logging.getLogger(__name__)
-
-argparser = argparse.ArgumentParser(add_help=False)
-argparser.add_argument('--parallel-compile', action='store_true',
- default=False,
- help="present if compiling can be done in parallel")
-
-the_io_thread_pool = None
-
-
-class MeasurementInterface(object):
- """
- abstract base class for compile and measurement
- """
- __metaclass__ = abc.ABCMeta
-
- def __init__(self,
- args=None,
- project_name=None,
- program_name='unknown',
- program_version='unknown',
- manipulator=None,
- objective=None,
- input_manager=None):
- self.args = args
- self._project = project_name
- self._program = program_name
- self._version = program_version
- self._objective = objective
- self._manipulator = manipulator
- self._input_manager = input_manager
-
- self.pids = []
- self.pid_lock = threading.Lock()
- self.parallel_compile = args.parallel_compile
- # If parallel_compile is False then compile_and_run() will be invoked
- # sequentially otherwise the driver first invokes compile() in parallel
- # followed by run_precompiled() sequentially
-
- def compile(self, config_data, id):
- """
- Compile in PARALLEL according to the configuration in config_data
- (obtained from desired_result.configuration) Should use id parameter
- to determine output location of executable Return value will be passed
- to run_precompiled as compile_result, useful for storing error/timeout
- information
- """
- if self.parallel_compile:
- raise RuntimeError('MeasurementInterface.compile() not implemented for',
- 'parallel compilation')
- pass
-
- def run_precompiled(self, desired_result, input, limit, compile_result, id):
- """
- Run the given desired_result SEQUENTIALLY on input and produce a Result()
- Abort early if limit (in seconds) is reached Assume that the executable
- to be measured has already been compiled to an executable corresponding to
- identifier id by compile() The compile_result is the return result of compile(),
- and it will be None if compile() was not called
- """
- if self.parallel_compile:
- raise RuntimeError('MeasurementInterface.run_precompiled() not implemented',
- 'for parallel compilation')
- pass
-
- def cleanup(self, id):
- """
- Clean up any temporary files associated with the executable
- """
- pass
-
- def pre_process(self):
- """
- The process before each iteration This method will be called
- once per iteration before all threads are launched
- """
- pass
-
- def post_process(self):
- """
- The process after each iteration This method will be called
- once per iteration after all threads are committed
- """
- pass
-
- def extra_convergence_criteria(self, result):
- """
- The extra convergence criteria which returns True if the
- current result is acceptable by the user
- """
- return False
-
- #@abc.abstractmethod
- def compile_and_run(self, desired_result, input, limit):
- """
- Compile and run the given desired_result on input and produce a
- Result(), abort early if limit (in seconds) is reached This function
- is only used for sequential execution flow
-
- FIXME: Shoud uncomment @abc.abstractmethod Now comment out for
- compatiability
- """
- return self.run(desired_result, input, limit)
-
- def run(self, desired_result, input, limit):
- """
- run the given desired_result on input and produce a Result(),
- abort early if limit (in seconds) is reached
- """
- return opentuner.resultdb.models.Result()
-
- def save_final_config(self, config):
- """
- called at the end of autotuning with the best resultsdb.models.Configuration
- """
- try:
- config_str = repr(config.data)
- if len(config_str) > 256:
- config_str = config_str[:256] + '...'
- log.info('final configuration: %s', config_str)
- log.info('you may want to implement save_final_config(), to store this')
- except:
- log.error('error printing configuration', exc_info=True)
-
- def db_program_version(self, session):
- """return a version identifier for the program being tuned"""
- return resultsdb.models.ProgramVersion.get(
- session=session,
- project=self.project_name(),
- name=self.program_name(),
- version=self.program_version(),
- parameter_info=self.manipulator().parameters_to_json(),
- )
-
- def set_driver(self, measurement_driver):
- self.driver = measurement_driver
-
- def project_name(self):
- if self._project is not None:
- return self._project
- autoname = re.sub('(Measurement?)Interface$', '', self.__class__.__name__)
- if autoname:
- return autoname
- else:
- return 'unknown'
-
- def program_name(self):
- return self._program
-
- def program_version(self):
- return self._version
-
- def file_hash(self, filename):
- """helper used to generate program versions"""
- return hashlib.sha256(open(filename).read()).hexdigest()
-
- def manipulator(self):
- """
- called once to create the search.manipulator.ConfigurationManipulator
- """
- if self._manipulator is None:
- msg = ('MeasurementInterface.manipulator() must be implemented or a '
- '"manipulator=..." must be provided to the constructor')
- log.error(msg)
- raise Exception(msg)
- return self._manipulator
-
- def objective(self):
- """
- called once to create the search.objective.SearchObjective
- """
- if self._objective is None:
- from ..search.objective import MinimizeSize
-
- return MinimizeSize()
- return self._objective
-
- def input_manager(self):
- """
- called once to create the measurement.inputmanager.InputManager
- """
- if self._objective is None:
- from .inputmanager import FixedInputManager
-
- return FixedInputManager()
- return self._input_manager
-
- def seed_configurations(self):
- """
- Extra seed configuration objects to add to those given on the command line.
- Configuration objects (typically dictionaries) not database objects.
- """
- return []
-
- def kill_all(self):
- self.pid_lock.acquire()
- for pid in self.pids:
- goodkillpg(pid)
- self.pids = []
- self.pid_lock.release()
-
- def call_program(self, cmd, limit=None, memory_limit=None, **kwargs):
- """
- call cmd and kill it if it runs for longer than limit
-
- returns dictionary like
- {'returncode': 0,
- 'stdout': '', 'stderr': '',
- 'timeout': False, 'time': 1.89}
- """
- the_io_thread_pool_init(self.args.parallelism)
- if limit is float('inf'):
- limit = None
- if type(cmd) in (str, unicode):
- kwargs['shell'] = True
- killed = False
- t0 = time.time()
- p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE,
- preexec_fn=preexec_setpgid_setrlimit(memory_limit),
- **kwargs)
- # Add p.pid to list of processes to kill in case of keyboardinterrupt
- self.pid_lock.acquire()
- self.pids.append(p.pid)
- self.pid_lock.release()
-
- try:
- stdout_result = the_io_thread_pool.apply_async(p.stdout.read)
- stderr_result = the_io_thread_pool.apply_async(p.stderr.read)
- while p.returncode is None:
- if limit is None:
- goodwait(p)
- elif limit and time.time() > t0 + limit:
- killed = True
- goodkillpg(p.pid)
- goodwait(p)
- else:
- # still waiting...
- sleep_for = limit - (time.time() - t0)
- if not stdout_result.ready():
- stdout_result.wait(sleep_for)
- elif not stderr_result.ready():
- stderr_result.wait(sleep_for)
- else:
- #TODO(jansel): replace this with a portable waitpid
- time.sleep(0.001)
- p.poll()
- except:
- if p.returncode is None:
- goodkillpg(p.pid)
- raise
- finally:
- # No longer need to kill p
- self.pid_lock.acquire()
- if p.pid in self.pids:
- self.pids.remove(p.pid)
- self.pid_lock.release()
-
- # TODO-autotune: Extract the file size and use it
- # FIXIT: Appropriately update the file size
- t1 = time.time()
- return {'time': float('inf') if killed else (t1 - t0),
- 'timeout': killed,
- 'returncode': p.returncode,
- 'stdout': stdout_result.get(),
- 'stderr': stderr_result.get(),
- }
-
- def getFileSize(self,filename):
- fileinfo=os.stat(filename)
- file_size=fileinfo.st_size
- return {'binary_size': file_size}
-
- def prefix_hook(self, session):
- pass
-
- @classmethod
- def main(cls, args, *pargs, **kwargs):
- from opentuner.tuningrunmain import TuningRunMain
-
- return TuningRunMain(cls(args, *pargs, **kwargs), args).main()
-
-
-class DefaultMeasurementInterface(MeasurementInterface):
- def run(self, desired_result, input, limit):
- raise RuntimeError('MeasurementInterface.run() not implemented')
-
-
-def preexec_setpgid_setrlimit(memory_limit):
- if resource is not None:
- def _preexec():
- os.setpgid(0, 0)
- try:
- resource.setrlimit(resource.RLIMIT_CORE, (0, 0))
- except ValueError:
- pass # No permission
- if memory_limit:
- try:
- (soft, hard) = resource.getrlimit(resource.RLIMIT_AS)
- resource.setrlimit(resource.RLIMIT_AS, (min(soft, memory_limit),
- min(hard, memory_limit)))
- except ValueError:
- pass # No permission
- return _preexec
-
-
-def the_io_thread_pool_init(parallelism=1):
- global the_io_thread_pool
- if the_io_thread_pool is None:
- the_io_thread_pool = ThreadPool(2 * parallelism)
- # make sure the threads are started up
- the_io_thread_pool.map(int, range(2 * parallelism))
-
-
-def goodkillpg(pid):
- """
- wrapper around kill to catch errors
- """
- log.debug("killing pid %d", pid)
- try:
- if hasattr(os, 'killpg'):
- os.killpg(pid, signal.SIGKILL)
- else:
- os.kill(pid, signal.SIGKILL)
- except:
- log.error('error killing process %s', pid, exc_info=True)
-
-
-def goodwait(p):
- """
- python doesn't check if its system calls return EINTR, retry if it does
- """
- while True:
- try:
- rv = p.wait()
- return rv
- except OSError, e:
- if e.errno != errno.EINTR:
- raise
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/resultsdb/__init__.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/resultsdb/__init__.py
deleted file mode 100644
index a0150a1577..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/resultsdb/__init__.py
+++ /dev/null
@@ -1,6 +0,0 @@
-
-from connect import connect
-
-import models
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/resultsdb/connect.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/resultsdb/connect.py
deleted file mode 100644
index 1a04d05447..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/resultsdb/connect.py
+++ /dev/null
@@ -1,66 +0,0 @@
-from sqlalchemy import create_engine
-from sqlalchemy.orm import scoped_session, sessionmaker
-from models import Base, _Meta
-import logging
-import time
-from pprint import pprint
-
-log = logging.getLogger(__name__)
-
-DB_VERSION = "0.0"
-
-if False: # profiling of queries
- import atexit
- from sqlalchemy import event
- from collections import Counter
- from sqlalchemy.engine import Engine
- the_query_totals = Counter()
-
- @event.listens_for(Engine, "before_cursor_execute")
- def before_cursor_execute(conn, cursor, statement,
- parameters, context, executemany):
- context._query_start_time = time.time()
-
- @event.listens_for(Engine, "after_cursor_execute")
- def after_cursor_execute(conn, cursor, statement,
- parameters, context, executemany):
- total = time.time() - context._query_start_time
- the_query_totals[statement] += total
-
- @atexit.register
- def report():
- pprint(the_query_totals.most_common(10))
-
-
-def connect(dbstr):
- engine = create_engine(dbstr, echo = False)
- connection = engine.connect()
-
- #handle case that the db was initialized before a version table existed yet
- if engine.dialect.has_table(connection, "program"):
- # if there are existing tables
- if not engine.dialect.has_table(connection, "_meta"):
- # if no version table, assume outdated db version and error
- connection.close()
- raise Exception("Your opentuner database is currently out of date. Save a back up and reinitialize")
-
- # else if we have the table already, make sure version matches
- if engine.dialect.has_table(connection, "_meta"):
- Session = scoped_session(sessionmaker(autocommit=False,
- autoflush=False,
- bind=engine))
- version = _Meta.get_version(Session)
- if not DB_VERSION == version:
- raise Exception('Your opentuner database version {} is out of date with the current version {}'.format(version, DB_VERSION))
-
- Base.metadata.create_all(engine)
-
- Session = scoped_session(sessionmaker(autocommit=False,
- autoflush=False,
- bind=engine))
- # mark database with current version
- _Meta.add_version(Session, DB_VERSION)
- Session.commit()
-
- return engine, Session
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/resultsdb/models.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/resultsdb/models.py
deleted file mode 100644
index dd88ae8e51..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/resultsdb/models.py
+++ /dev/null
@@ -1,319 +0,0 @@
-from sqlalchemy.ext.declarative import declarative_base
-from sqlalchemy.ext.declarative import declared_attr
-from sqlalchemy import create_engine
-from sqlalchemy.orm import relationship, backref
-from sqlalchemy import (
- Column, Integer, String, DateTime, Boolean, Enum,
- Float, PickleType, ForeignKey, Text, func, Index)
-import sqlalchemy
-import re
-
-from cPickle import dumps, loads
-from gzip import zlib
-class CompressedPickler(object):
- @classmethod
- def dumps(cls, obj, protocol=2):
- s = dumps(obj, protocol)
- sz = zlib.compress(s, 9)
- if len(sz) < len(s):
- return sz
- else:
- return s
-
- @classmethod
- def loads(cls, string):
- try:
- s = zlib.decompress(string)
- except:
- s = string
- return loads(s)
-
-class Base(object):
- @declared_attr
- def __tablename__(cls):
- """convert camel case to underscores"""
- return re.sub(r'([a-z])([A-Z])', r'\1_\2', cls.__name__).lower()
-
- id = Column(Integer, primary_key=True, index=True)
-
-
-Base = declarative_base(cls=Base)
-
-class _Meta(Base):
- """ meta table to track current version """
- db_version = Column(String(128))
-
- @classmethod
- def has_version(cls, session, version):
- try:
- session.flush()
- session.query(_Meta).filter_by(db_version=version).one()
- return True
- except sqlalchemy.orm.exc.NoResultFound:
- return False
-
- @classmethod
- def get_version(cls, session):
- try:
- session.flush()
- x = session.query(_Meta).one()
- return x.db_version
- except sqlalchemy.orm.exc.NoResultFound:
- return None
-
- @classmethod
- def add_version(cls, session, version):
- if not cls.has_version(session, version):
- session.add(_Meta(db_version=version))
-
-
-class Program(Base):
- project = Column(String(128))
- name = Column(String(128))
-
- @classmethod
- def get(cls, session, project, name):
- try:
- session.flush()
- return session.query(Program).filter_by(project=project, name=name).one()
- except sqlalchemy.orm.exc.NoResultFound:
- t = Program(project=project, name=name)
- session.add(t)
- return t
-
-
-class ProgramVersion(Base):
- program_id = Column(ForeignKey(Program.id))
- program = relationship(Program, backref='versions')
- version = Column(String(128))
- parameter_info = Column(Text)
-
- @property
- def name(self):
- return self.program.name
-
- @property
- def project(self):
- return self.program.project
-
- @classmethod
- def get(cls, session, project, name, version, parameter_info=None):
- program = Program.get(session, project, name)
- try:
- session.flush()
- if parameter_info is None:
- return session.query(ProgramVersion).filter_by(program=program,
- version=version).one()
- else:
- return session.query(ProgramVersion).filter_by(program=program,
- version=version,
- parameter_info=parameter_info).one()
- except sqlalchemy.orm.exc.NoResultFound:
- t = ProgramVersion(program=program, version=version, parameter_info=parameter_info)
- session.add(t)
- return t
-
-
-class Configuration(Base):
- program_id = Column(ForeignKey(Program.id))
- program = relationship(Program)
- hash = Column(String(64))
- data = Column(PickleType(pickler=CompressedPickler))
-
- @classmethod
- def get(cls, session, program, hashv, datav):
- try:
- session.flush()
- return (session.query(Configuration)
- .filter_by(program=program, hash=hashv).one())
- except sqlalchemy.orm.exc.NoResultFound:
- t = Configuration(program=program, hash=hashv, data=datav)
- session.add(t)
- return t
-
-
-Index('ix_configuration_custom1', Configuration.program_id, Configuration.hash)
-
-
-class MachineClass(Base):
- name = Column(String(128))
-
- @classmethod
- def get(cls, session, name):
- try:
- session.flush()
- return session.query(MachineClass).filter_by(name=name).one()
- except sqlalchemy.orm.exc.NoResultFound:
- t = MachineClass(name=name)
- session.add(t)
- return t
-
-
-class Machine(Base):
- name = Column(String(128))
-
- cpu = Column(String(128))
- cores = Column(Integer)
- memory_gb = Column(Float)
-
- machine_class_id = Column(ForeignKey(MachineClass.id))
- machine_class = relationship(MachineClass, backref='machines')
-
-
-class InputClass(Base):
- program_id = Column(ForeignKey(Program.id))
- program = relationship(Program, backref='inputs')
-
- name = Column(String(128))
- size = Column(Integer)
-
- @classmethod
- def get(cls, session, program, name='default', size=-1):
- try:
- session.flush()
- return session.query(InputClass).filter_by(program=program,
- name=name,
- size=size).one()
- except sqlalchemy.orm.exc.NoResultFound:
- t = InputClass(program=program, name=name, size=size)
- session.add(t)
- return t
-
-
-class Input(Base):
- #state = Column(Enum('ANY_MACHINE', 'SINGLE_MACHINE', 'DELETED'),
- # default='ANY_MACHINE', name='t_input_state')
-
- input_class_id = Column(ForeignKey(InputClass.id))
- input_class = relationship(InputClass, backref='inputs')
-
- #optional, set only for state='SINGLE_MACHINE'
- #machine_id = Column(ForeignKey(MachineClass.id))
- #machine = relationship(MachineClass, backref='inputs')
-
- #optional, for use by InputManager
- path = Column(Text)
- extra = Column(PickleType(pickler=CompressedPickler))
-
-
-class TuningRun(Base):
- uuid = Column(String(32), index=True, unique=True)
-
- program_version_id = Column(ForeignKey(ProgramVersion.id))
- program_version = relationship(ProgramVersion, backref='tuning_runs')
-
- machine_class_id = Column(ForeignKey(MachineClass.id))
- machine_class = relationship(MachineClass, backref='tuning_runs')
-
- input_class_id = Column(ForeignKey(InputClass.id))
- input_class = relationship(InputClass, backref='tuning_runs')
-
- name = Column(String(128), default='unnamed')
- args = Column(PickleType(pickler=CompressedPickler))
- objective = Column(PickleType(pickler=CompressedPickler))
-
- state = Column(Enum('QUEUED', 'RUNNING', 'COMPLETE', 'ABORTED',
- name='t_tr_state'),
- default='QUEUED')
- start_date = Column(DateTime, default=func.now())
- end_date = Column(DateTime)
-
- final_config_id = Column(ForeignKey(Configuration.id))
- final_config = relationship(Configuration)
-
- #__mapper_args__ = {'primary_key': uuid}
-
- @property
- def program(self):
- return self.program_version.program
-
-
-class Result(Base):
- #set by MeasurementDriver:
- configuration_id = Column(ForeignKey(Configuration.id))
- configuration = relationship(Configuration)
-
- machine_id = Column(ForeignKey(Machine.id))
- machine = relationship(Machine, backref='results')
-
- input_id = Column(ForeignKey(Input.id))
- input = relationship(Input, backref='results')
-
- tuning_run_id = Column(ForeignKey(TuningRun.id), index=True)
- tuning_run = relationship(TuningRun, backref='results')
-
- collection_date = Column(DateTime, default=func.now())
- collection_cost = Column(Float)
-
- #set by MeasurementInterface:
- state = Column(Enum('OK', 'TIMEOUT', 'ERROR',
- name='t_result_state'),
- default='OK')
- time = Column(Float)
- accuracy = Column(Float)
- energy = Column(Float)
- size = Column(Float)
- confidence = Column(Float)
- #extra = Column(PickleType)
-
- #set by SearchDriver
- was_new_best = Column(Boolean)
-
-
-Index('ix_result_custom1', Result.tuning_run_id, Result.was_new_best)
-
-
-class DesiredResult(Base):
- #set by the technique:
- configuration_id = Column(ForeignKey(Configuration.id))
- configuration = relationship(Configuration)
- limit = Column(Float)
-
- #set by the search driver
- priority = Column(Float)
- tuning_run_id = Column(ForeignKey(TuningRun.id))
- tuning_run = relationship(TuningRun, backref='desired_results')
- generation = Column(Integer)
- requestor = Column(String(128))
- request_date = Column(DateTime, default=func.now())
-
- #set by the measurement driver
- state = Column(Enum('UNKNOWN', 'REQUESTED', 'RUNNING',
- 'COMPLETE', 'ABORTED',
- name="t_dr_state"),
- default='UNKNOWN')
- result_id = Column(ForeignKey(Result.id), index=True)
- result = relationship(Result, backref='desired_results')
- start_date = Column(DateTime)
-
- #input_id = Column(ForeignKey(Input.id))
- #input = relationship(Input, backref='desired_results')
-
-
-Index('ix_desired_result_custom1', DesiredResult.tuning_run_id,
- DesiredResult.generation)
-
-Index('ix_desired_result_custom2', DesiredResult.tuning_run_id,
- DesiredResult.configuration_id)
-
-
-# track bandit meta-technique information if a bandit meta-technique is used for a tuning run.
-class BanditInfo(Base):
- tuning_run_id = Column(ForeignKey(TuningRun.id))
- tuning_run = relationship(TuningRun, backref='bandit_info')
- # the bandit exploration/exploitation tradeoff
- c = Column(Float)
- # the bandit window
- window = Column(Integer)
-
-class BanditSubTechnique(Base):
- bandit_info_id = Column(ForeignKey(BanditInfo.id))
- bandit_info = relationship(BanditInfo, backref='subtechniques')
- name = Column(String(128))
-
-
-if __name__ == '__main__':
- #test:
- engine = create_engine('sqlite:///:memory:', echo=True)
- Base.metadata.create_all(engine)
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/__init__.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/__init__.py
deleted file mode 100644
index bb4ce57bb2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/__init__.py
+++ /dev/null
@@ -1,6 +0,0 @@
-
-import driver
-import objective
-import plugin
-import technique
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/bandittechniques.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/bandittechniques.py
deleted file mode 100644
index 29816c03de..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/bandittechniques.py
+++ /dev/null
@@ -1,316 +0,0 @@
-import abc
-import copy
-import logging
-import math
-import random
-from collections import deque
-
-from .metatechniques import MetaSearchTechnique
-from .technique import register, SearchTechnique, all_techniques, get_random_generator_technique
-
-log = logging.getLogger(__name__)
-
-
-class BanditQueue(object):
- def __init__(self, keys, C=0.05, window=500, **kwargs):
- """
- C is exploration/exploitation tradeoff
- window is how long to remember past results
- """
- super(BanditQueue, self).__init__(**kwargs)
- self.C = C
- self.history = deque()
- self.keys = keys
- self.use_counts = dict(((k, 0) for k in keys))
- self.window = window
- self.request_count = 0
-
- @abc.abstractmethod
- def exploitation_term(self, key):
- """
- value 0 to 1.0 to represent quality of technique
- """
- return 0.0
-
- def exploration_term(self, key):
- """
- value represent how unsure we are (optimal bandit solution)
- """
- if self.use_counts[key] > 0:
- return math.sqrt((2.0 * math.log(len(self.history), 2.0))
- / self.use_counts[key])
- else:
- return float('inf')
-
- def bandit_score(self, key):
- return (self.exploitation_term(key) +
- self.C * self.exploration_term(key))
-
- def ordered_keys(self):
- """select the next technique to use"""
-
- keys = list(self.keys)
- random.shuffle(keys) # break ties randomly
- keys.sort(key=self.bandit_score)
-
- self.request_count += 1
- if log.isEnabledFor(logging.DEBUG) and (self.request_count % 1000) == 0:
- log.debug(str([
- (t, self.exploitation_term(t), self.C * self.exploration_term(t))
- for t in keys]))
-
- return reversed(keys)
-
- def on_result(self, key, value):
- self.history.append((key, value))
- self.on_push_history(key, value)
- if len(self.history) > self.window:
- self.on_pop_history(*self.history.popleft())
-
- def on_push_history(self, key, value):
- self.use_counts[key] += 1
-
- def on_pop_history(self, key, value):
- self.use_counts[key] -= 1
-
-
-class AUCBanditQueue(BanditQueue):
- """
- Area Under the Receiving Operator Curve (AUC) credit assignment
-
- See:
- Comparison-based adaptive strategy selection with bandits in differential
- evolution. Fialho et al.
- """
-
- def __init__(self, *args, **kwargs):
- super(AUCBanditQueue, self).__init__(*args, **kwargs)
- self.debug = kwargs.get('debug', False)
- self.auc_sum = dict(((t, 0) for t in self.keys))
- self.auc_decay = dict(((t, 0) for t in self.keys))
-
- def exploitation_term_slow(self, key):
- """
- value 0 to 1.0 to represent quality of key
-
- computes the area under the curve where finding a new
- global best results in adding 1 to a cumulative total
- """
- score = 0.0
- pos = 0
- for t, value in self.history:
- if t is key:
- pos += 1
- if value:
- score += pos
- if pos:
- return score * 2.0 / (pos * (pos + 1.0))
- else:
- return 0.0
-
- def exploitation_term_fast(self, key):
- """
- value 0 to 1.0 to represent quality of key
-
- optimized O(1) implementation exploitation_term_slow()
- """
- score = self.auc_sum[key]
- pos = self.use_counts[key]
- if pos:
- return score * 2.0 / (pos * (pos + 1.0))
- else:
- return 0.0
-
- def exploitation_term(self, key):
- v1 = self.exploitation_term_fast(key)
- if self.debug:
- v2 = self.exploitation_term_slow(key)
- assert v1 == v2
- return v1
-
- def on_push_history(self, key, value):
- super(AUCBanditQueue, self).on_push_history(key, value)
- if value:
- self.auc_sum[key] += self.use_counts[key]
- self.auc_decay[key] += 1
-
- def on_pop_history(self, key, value):
- super(AUCBanditQueue, self).on_pop_history(key, value)
- self.auc_sum[key] -= self.auc_decay[key]
- if value:
- self.auc_decay[key] -= 1
-
-
-
-class AUCBanditMetaTechnique(MetaSearchTechnique):
- def __init__(self, techniques, bandit_kwargs=dict(), **kwargs):
- super(AUCBanditMetaTechnique, self).__init__(techniques, **kwargs)
- self.bandit = AUCBanditQueue([t.name for t in techniques], **bandit_kwargs)
- self.name_to_technique = dict(((t.name, t) for t in self.techniques))
-
- def select_technique_order(self):
- """select the next technique to use"""
- return (self.name_to_technique[k] for k in self.bandit.ordered_keys())
-
- def on_technique_result(self, technique, result):
- self.bandit.on_result(technique.name, result.was_new_best)
-
- def on_technique_no_desired_result(self, technique):
- """treat not providing a configuration as not a best"""
- self.bandit.on_result(technique.name, 0)
-
- @classmethod
- def generate_technique(cls, manipulator=None, num_techniques=5, retry_count=3, generator_weight=10, *args, **kwargs):
- """
- Generate a bandit by randomly selecting existing techniques or composable techniques.
- If a composable technique is selected, the operators are then chosen
-
- :param manipulator: a ConfigurationManipulator used to enumerate parameters
- :param num_techniques: max number of subtechniques in the bandit
- :param retry_count: number of times to try getting a new technique before giving up
- :param generator_weight: weight to increase probability of choosing to generate a technique
- """
- techniques, generators = all_techniques()
-
- # get set of parameters to consider
- paramset = set()
- for p in manipulator.params:
- paramset.add(type(p))
-
- # filter techniques to get rid of metatechniques
- basetechniques = [t for t in techniques if not isinstance(t, MetaSearchTechnique)]
- bandit_techniques = []
- for i in range(num_techniques):
- for j in range(retry_count):
- # pick a technique or generate a composable
- if random.random() < float(len(basetechniques)) / (len(basetechniques) + generator_weight*len(generators)):
- candidate = copy.deepcopy(random.choice(basetechniques))
- else:
- # pick a random generator
- candidate = get_random_generator_technique(generators, manipulator=manipulator)
- if not (candidate.name in [t.name for t in bandit_techniques]):
- bandit_techniques.append(candidate)
- break
-
- # make a bandit of the output list
- return cls(bandit_techniques, name="GeneratedBandit", *args, **kwargs)
-
-
-class AUCBanditMutationTechnique(SearchTechnique):
- def __init__(self, bandit_kwargs=dict(), **kwargs):
- super(AUCBanditMutationTechnique, self).__init__(**kwargs)
- self.bandit = None
- self.bandit_kwargs = bandit_kwargs
- self.pending_results = []
-
- def handle_requested_result(self, result):
- for i in xrange(len(self.pending_results)):
- cfg, name, index = self.pending_results[i]
- if result.configuration == cfg:
- self.bandit.on_result((name, index), result.was_new_best)
- del self.pending_results[i]
- return
- log.warning("unexpected result")
-
- def desired_configuration(self):
- """
- use bandit to pick a single manipulator and apply it
- """
- seed = self.get_seed()
- if self.bandit is None:
- self.init_bandit(seed)
-
- cfg = self.manipulator.copy(seed)
- hash1 = self.manipulator.hash_config(cfg)
- params = self.manipulator.parameters_dict(cfg)
- for name, index in self.bandit.ordered_keys():
- if name in params:
- param = params[name]
- fns = param.manipulators(cfg)
- fn = fns[index % len(fns)]
- fn(cfg)
- hash2 = self.manipulator.hash_config(cfg)
- if hash1 != hash2:
- cfg = self.driver.get_configuration(cfg)
- self.pending_results.append((cfg, name, index))
- log.debug("applied %s[%s] manipulator function", name, index)
- return cfg
-
- return None
-
-
- def init_bandit(self, cfg):
- options = []
- for param in self.manipulator.parameters(cfg):
- for i in xrange(len(param.manipulators(cfg))):
- options.append((param.name, i))
- # TODO(jansel): remove assumption that set of parameters are fixed
- self.bandit = AUCBanditQueue(options, **self.bandit_kwargs)
-
- def get_seed(self):
- """seed mutation with global best"""
- if (self.driver.best_result is not None and
- self.driver.best_result.state == 'OK'):
- return self.driver.best_result.configuration.data
- else:
- return self.manipulator.random()
-
-
-import evolutionarytechniques
-import differentialevolution
-import simplextechniques
-import patternsearch
-import simulatedannealing
-from pso import PSO, HybridParticle
-import globalGA
-register(AUCBanditMutationTechnique())
-
-register(AUCBanditMetaTechnique([
- differentialevolution.DifferentialEvolutionAlt(),
- evolutionarytechniques.UniformGreedyMutation(),
- evolutionarytechniques.NormalGreedyMutation(mutation_rate=0.3),
- simplextechniques.RandomNelderMead(),
- ], name = "AUCBanditMetaTechniqueA"))
-register(AUCBanditMetaTechnique([
- differentialevolution.DifferentialEvolutionAlt(),
- evolutionarytechniques.UniformGreedyMutation(),
- ], name = "AUCBanditMetaTechniqueB"))
-register(AUCBanditMetaTechnique([
- differentialevolution.DifferentialEvolutionAlt(),
- patternsearch.PatternSearch(),
- ], name = "AUCBanditMetaTechniqueC"))
-register(AUCBanditMetaTechnique([
- PSO(crossover = 'op3_cross_OX3'),
- PSO(crossover = 'op3_cross_OX1'),
- PSO(crossover = 'op3_cross_CX'),
- PSO(crossover = 'op3_cross_PMX'),
- PSO(crossover = 'op3_cross_PX'),
- evolutionarytechniques.GA(crossover = 'op3_cross_OX3', mutation_rate=0.01, crossover_rate=0.8),
- evolutionarytechniques.GA(crossover = 'op3_cross_OX1', mutation_rate=0.01, crossover_rate=0.8),
- evolutionarytechniques.GA(crossover = 'op3_cross_CX', mutation_rate=0.01, crossover_rate=0.8),
- evolutionarytechniques.GA(crossover = 'op3_cross_PX', mutation_rate=0.01, crossover_rate=0.8),
- evolutionarytechniques.GA(crossover = 'op3_cross_PMX', mutation_rate=0.01, crossover_rate=0.8),
- evolutionarytechniques.UniformGreedyMutation(name='ga-base', mutation_rate=0.01)
- ], name = "PSO_GA_Bandit"))
-register(AUCBanditMetaTechnique([
- differentialevolution.DifferentialEvolutionAlt(),
- simulatedannealing.PseudoAnnealingSearch()
- ], name = "test"))
-register(AUCBanditMetaTechnique([
- differentialevolution.DifferentialEvolutionAlt(),
- evolutionarytechniques.UniformGreedyMutation(),
- evolutionarytechniques.NormalGreedyMutation(mutation_rate=0.3),
- simplextechniques.RandomNelderMead(),
- simulatedannealing.PseudoAnnealingSearch()
- ], name = "test2"))
-register(AUCBanditMetaTechnique([
- PSO(crossover='op3_cross_OX1'),
- PSO(crossover='op3_cross_PMX'),
- PSO(crossover='op3_cross_PX'),
- evolutionarytechniques.GA(crossover='op3_cross_OX1', crossover_rate=0.5),
- evolutionarytechniques.GA(crossover='op3_cross_PMX', crossover_rate=0.5),
- evolutionarytechniques.GA(crossover='op3_cross_PX', crossover_rate=0.5),
- differentialevolution.DifferentialEvolutionAlt(),
- globalGA.NormalGreedyMutation( crossover_rate=0.5, crossover_strength=0.2, name='GGA')
- ], name='PSO_GA_DE'))
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/composableevolutionarytechniques.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/composableevolutionarytechniques.py
deleted file mode 100644
index e511744f30..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/composableevolutionarytechniques.py
+++ /dev/null
@@ -1,512 +0,0 @@
-import random
-import time
-import sys
-import json
-from fn import _
-from technique import all_techniques
-from technique import register
-from technique import register_generator
-from technique import SequentialSearchTechnique
-from manipulator import *
-from opentuner.search.manipulator import Parameter
-
-
-class PopulationMember(object):
- """
- An extendable object representing a population member for ComposableEvolutionaryTechniques.
- Must have the field "config" which is a configuration
- """
- def __init__(self, config):
- self.config = config
- self.timestamp = time.time()
-
- def touch(self):
- """
- Update the timestamp on a PopulationMember
- """
- self.timestamp = time.time()
-
-
-class ComposableEvolutionaryTechnique(SequentialSearchTechnique):
- """
- An abstract base class for a technique that is composable with operators
- """
- __metaclass__ = abc.ABCMeta
-
- # operator_map - from param-type to dict with operator name + list of arguments TODO
- # min_parent - minimum number of parents returned. Limits which operators can be used
- def __init__(self,
- operator_map = {},
- population_size = 50,
- initial_configs = None,
- *pargs,
- **kwargs):
- """
-
- :param operator_map:
- :param population_size:
- :param initial_configs:
- :param pargs:
- :param kwargs:
- :return:
- """
- super(ComposableEvolutionaryTechnique, self).__init__(*pargs, **kwargs)
- # generate a name based on operators if no name
-
- self.initial_configurations = initial_configs
- self.population_size = population_size
- self.operator_map = operator_map # map from parameter type to an operator function
-
- def set_operator_map(self, operator_map):
- self.operator_map = operator_map
-
- @classmethod
- def get_hyper_parameters(cls):
- return ['population_size']
-
- def default_generated_name(self):
- """
- Gets the default name for this technique based on its operator map
-
- Name is in the format
- classname paramname;opname;[arg1,arg2,[[kwarg1,v1][kwarg2,v2]]] paramname2;opname2;...
- """
- # TODO - include technique hyperparameters
- parts = [self.base_name()]
- for param in sorted(self.operator_map, cmp=lambda x,y: cmp(x.__name__, y.__name__)):
- subparts = [param.__name__]
- operator_info = self.operator_map[param]
- subparts.append(operator_info['op_name'])
- args = list(operator_info['args'])
- kwargs = operator_info['kwargs']
- args.append([(k,kwargs[k]) for k in sorted(kwargs)])
- subparts.append(json.dumps(args, separators=(',', ':')))
- parts.append(';'.join(subparts))
- return ' '.join(parts)
-
-
- def make_population_member(self, config):
- """
- Given a configuration, returns an object representing a single member of the
- population with the given configuration. Meta-data about the configuration,
- such as last selection time as a parent, can be attached to the object.
-
- This can be overriden to return a custom population member for use in
- :py:meth:`get_parents` and :py:meth:`update_population`
-
- :param config: the configuration that this population member will represent
- :return: a population member reresenting the input configuration.
- """
- return PopulationMember(config)
-
- def select_parameters(self, params):
- """
- Given all the available parameters, return a subset of parameters to operate
- on when generating a new configuration.
-
- Can override this to operate on only a subset of parameters.
-
- :param params: a list of all the available parameters
- :return: a subset of params
- """
- return params
-
- @abc.abstractmethod
- def minimum_number_of_parents(self):
- """
- Return the minimum number of parents ever returned by :py:meth:`get_parents`.
- This limits which operators can be composed with the technique. Operators
- requiring more input configurations than the minimum number of parents will
- result in an error.
-
- :return: the minimum number of parents ever generated.
- """
- return 1
-
- @abc.abstractmethod
- def get_parents(self, population):
- """
- Given the current population, return a list of configurations that will be
- used to generate a new configuration via operators. Returning less parents
- than guaranteed by :py:meth:`minimum_number_of_parents` results in an error.
-
- The parents will be passed to operators in order. If there are more parents
- than required by an operator, extra parents will be discarded.
-
- Note that operators mutate the first configuration passed in.
-
- :param population: the current population in the technique
- :return: a list of parent configurations to generate a new configuration from
- """
- pass
-
- @abc.abstractmethod
- def update_population(self, config, population):
- """
- Update the population given the newest configuration and current population
- in the technique. should return the new population
-
- :param config: the newest generated configuration
- :param population: the current population in this iteration of the technique
- :return: the updated population
- """
- pass
-
- def get_initial_population(self):
- """
- Returns an initial population by passing initial configurations into
- :py:meth:`make_population_member`
-
- :return: an initial list of objects returned by :py:meth:`make_population_member`.
- """
- init_configs = self.initial_configurations
- if not init_configs:
- init_configs = [self.manipulator.random() for i in range(self.population_size)]
- return [PopulationMember(config) for config in init_configs]
-
- def lt(self, cfg_a, cfg_b):
- """
- Return whether cfg_a has a better objective function score than cfg_b
-
- :param cfg_a: first configuration
- :param cfg_b: second configuration
- :return: True if cfg_a is better than cfg_b
- """
- def config(cfg):
- return self.driver.get_configuration(cfg)
- return self.objective.lt(config(cfg_a), config(cfg_b))
-
- def lte(self, cfg_a, cfg_b):
- """
- Return whether cfg_a's objective function score is at least as good as cfg_b's
- score
-
- :param cfg_a: first configuration
- :param cfg_b: second configuration
- :return: True if cfg_a is at least as good as cfg_b
- """
- def config(cfg):
- return self.driver.get_configuration(cfg)
- return self.objective.lte(config(cfg_a), config(cfg_b))
-
- def get_global_best_configuration(self):
- """
- Return the current global best configuration in the search
-
- :return: the current global best configuration
- """
- if (self.driver.best_result is not None and
- self.driver.best_result.state == 'OK'):
- return self.manipulator.copy(self.driver.best_result.configuration.data)
- else:
- return self.manipulator.random()
-
- def get_default_operator(self, param_type):
- """
- Given a parameter type, return a dictionary with information about the
- operator to be used for the parameter. The returned dictionary must contain
- the following 3 key, value pairs
-
- 1. 'op_name' - the string name of the operator
- 2. 'args' - an iterable of the non-configuration arguments in order
- 3. 'kwargs' - a dictionary from any optional arguments to their values
-
- :return: a dictionary containing information about the operator to apply for the input parameter type
- """
- return {'op_name': 'op1_nop', 'args': [], 'kwargs': {}}
-
- # HELPER METHODS FOR BUILDING OPERATOR MAP
- @classmethod
- def add_to_map(cls, operator_map, param_type, operator_name, *args, **kwargs):
- """
- A helper method for adding parameter to operator mappings into the operator
- map.
-
- :param operator_map: the operator map to add to
- :param param_type: the parameter type to use the this operator on
- :param operator_name: the string name of the operator method
- :param *args: any non-configuration arguments to the operator
- :param **kwargs: any keyword arguemnts for the operator
- """
- if(isinstance(param_type, Parameter)):
- ptype = type(param_type)
- elif (type(param_type) == str):
- ptype = reduce(getattr, param_type.split("."), sys.modules[__name__])
- else:
- ptype = param_type;
-
- operator_map[ptype] = {'op_name': operator_name, 'args':args, 'kwargs':kwargs}
-
-
- def main_generator(self):
- """
- The primary body of the search technique.
- Initializes an initial population and then yields configurations by applying
- operators to get_parents.
- """
- min_parents = self.minimum_number_of_parents();
- # convert a manipulator configuration to a db.models.Configuration
- def get_driver_configuration(cfg):
- return self.driver.get_configuration(cfg)
-
- # initialize the population
- population = self.get_initial_population()
-
- # measure initial population
- for p in population:
- yield get_driver_configuration(p.config)
-
- while True:
- # get parents
- parents = self.get_parents(population)
- if len(parents) < min_parents:
- log.error("%s: Number of parents returned %d is less than the guaranteed"
- + " minimum returned by minimum_number_of_parents() %d. ",
- self.name, len(parents), min_parents)
- # fail and let other techniques work forever
- while True:
- yield None
-
- params = self.select_parameters(self.manipulator.params)
- config = self.get_new_config(parents, params)
- yield get_driver_configuration(config)
-
- population = self.update_population(config, population)
-
- # safety check that population has all been tested
- for p in population:
- if not self.driver.has_results(get_driver_configuration(p.config)):
- yield get_driver_configuration(p.config)
-
- def get_new_config(self, parents, params):
- """
- Return a new configuration to test, given a list of parent configurations
- This mutates the first parent
-
- :param parents: A list of parent configurations
- :params: A list of parameters to operate on
- :return: The mutated configuration (first parent)
- """
- for param in params:
- self.apply_operator(param, parents) #TODO
- return parents[0]
-
- def apply_operator(self, param, parents):
- """
- Apply the appropriate operator for param to parents.
- If an operator takes less input configurations than the number of parents,
- only the first parents are passed in. If operator takes more input configs
- than minimum_number_of_parents, logs an error and doesn't do anything
- """
- x = self.get_operator(type(param))
-
- operator_name = x['op_name']
- if not self.is_valid_operator(type(param), operator_name):
- # do nothing
- return
-
- # operator is already in valid form and starts with op1, op2, op3, op4, or opn
- num_parents_required = operator_name[2]
- if num_parents_required == 'n':
- args = parents[0] + [parents[1:]]
- else:
- num_parents_required = int(num_parents_required)
- args = parents[:num_parents_required]
- args.extend(x['args'])
-
- kwargs = x['kwargs']
-
- getattr(param, operator_name)(*args, **kwargs)
-
- def get_operator(self, param_type):
- if param_type in self.operator_map:
- return self.operator_map[param_type]
- return self.get_default_operator(param_type)
-
- def is_valid_operator(self, param_type, operator_name):
- if not hasattr(param_type, operator_name):
- log.error("%s: %s is not a valid operator for Parameter type %s",
- self.name, operator_name, param_type.__name__)
- return False
-
- if operator_name[:3] not in ['op1','op2','op3','op4','opn']:
- log.error("%s: %s is not a valid operator for Parameter type %s",
- self.name, operator_name, param_type.__name__)
- return False
-
- num_parents_required = operator_name[2]
- if num_parents_required == 'n':
- return True
-
- num_parents_required = int(num_parents_required)
- minimum_number_of_parents = self.minimum_number_of_parents()
-
- if num_parents_required > minimum_number_of_parents:
- log.error("%s: %s for Parameter type %s requires more input configs "
- + "than minimum number of parents, %d, produced by this technique",
- self.name, operator_name, param_type.__name__, minimum_number_of_parents)
- return False
-
- return True
-
- @classmethod
- def generate_technique(cls, manipulator=None, *args, **kwargs):
- """
- generate a composable technique with random operators
- """
- from manipulator import composable_operators
- # randomly select a composable technique to generate
- t = cls(*args, **kwargs)
- if manipulator is None:
- return t
-
- paramset = set()
- for p in manipulator.params:
- paramset.add(type(p))
-
- # add some random operator for each param
- operator_map = {}
- for param in paramset:
- operators = composable_operators(param, t.minimum_number_of_parents())
- # TODO - sometimes use "default" operator (don't choose an operator?
- # TODO - lower chance of picking op1_nop?
- ComposableEvolutionaryTechnique.add_to_map(operator_map, param, random.choice(operators))
-
- t.set_operator_map(operator_map)
- t.use_default_generated_name()
- return t
-
-
-class RandomThreeParentsComposableTechnique(ComposableEvolutionaryTechnique):
- """
- based on DifferentialEvolution
- """
-
- def __init__(self, cr = 0.9, must_mutate_count=1, information_sharing=1, *pargs, **kwargs):
- super(RandomThreeParentsComposableTechnique, self).__init__(*pargs, **kwargs)
- self.cr = cr
- self.must_mutate_count = must_mutate_count
- self.information_sharing = information_sharing
-
- @classmethod
- def get_hyper_parameters(cls):
- return ['population_size', 'cr', 'must_mutate_count', 'information_sharing']
-
- def minimum_number_of_parents(self):
- return 4
-
- def get_parents(self, population):
- self.use_f = random.random()
- population.sort(key=_.timestamp) # sort population by timestamp
-
- # copy oldest
- cfg = self.manipulator.copy(population[0].config)
-
- shuffled_population = map(_.config, population[1:])
- # mix in the global best configuration
- shuffled_population += ([self.get_global_best_configuration()]
- * self.information_sharing)
- random.shuffle(shuffled_population)
-
- # return oldest configuration +_3 other configurations
- return [cfg] + shuffled_population[0:3]
-
- def update_population(self, config, population):
- # replace the oldest configuration if the new one is better.
- population.sort(key=_.timestamp)
- if self.lt(config, population[0].config):
- population[0].config = config
-
- # mark that oldest configuration is updated
- population[0].touch()
-
- return population
-
- def select_parameters(self, params):
- """
- randomly select a subset of parameters to operate on
- """
- random.shuffle(params)
- ret_list = params[:self.must_mutate_count]
- for param in params[self.must_mutate_count:]:
- if random.random() < self.cr:
- ret_list.append(param)
- return ret_list
-
- def get_default_operator(self, param_type):
- return {'op_name': 'op4_set_linear', 'args': [1.0, self.use_f, -self.use_f], 'kwargs': {}}
-
-class GreedyComposableTechnique(ComposableEvolutionaryTechnique):
- """
- Always mixes in global best as parents
- """
- def __init__(self,
- mutation_rate = 0.1,
- must_mutate_count = 1,
- population_size = 10,
- *pargs, **kwargs):
- super(GreedyComposableTechnique, self).__init__(*pargs, **kwargs)
- self.mutation_rate = mutation_rate
- self.must_mutate_count = must_mutate_count
- self.population_size = population_size
-
-
- @classmethod
- def get_hyper_parameters(cls):
- return ['mutation_rate', 'must_mutate_count']
-
- def minimum_number_of_parents(self):
- # specify that we will return at least 4 cfgs from get_parents
- # this maximizes # of operators we can use
- return 4
-
- def get_parents(self, population):
- population.sort(key=_.timestamp) # sort population by timestamp
-
- # get a 50-50 mix of base and best cfgs as many operators do nothing given identical input cfgs
- cfg = self.manipulator.copy(population[0].config)
- # duplicate to get a total of 4 configurations to fulfill the promise in minimum_number_of_parents
- cfgs = [self.get_global_best_configuration(), cfg]*2
- # return a random 50-50 mix of the current configuration and global best to pass into operators
- random.shuffle(cfgs)
- return cfgs
-
- def update_population(self, config, population):
- # replace the oldest configuration if the new one is better.
- population.sort(key=_.timestamp)
- if self.lt(config, population[0].config):
- population[0].config = config
-
- # mark that oldest configuration is updated
- population[0].touch()
-
- return population
-
- def select_parameters(self, params):
- random.shuffle(params)
- ret_list = params[:self.must_mutate_count]
- for param in params[self.must_mutate_count:]:
- if random.random() < self.mutation_rate:
- ret_list.append(param)
- return ret_list
-
- def get_default_operator(self, param_type):
- return {'op_name': 'op1_randomize', 'args': [], 'kwargs':{}}
-
-
-register(RandomThreeParentsComposableTechnique(name='ComposableDiffEvolution',
- population_size=30))
-register_generator(RandomThreeParentsComposableTechnique)
-register_generator(GreedyComposableTechnique)
-
-
-op_map = {}
-ComposableEvolutionaryTechnique.add_to_map(op_map,
- PermutationParameter,
- "op3_cross", xchoice='op3_cross_CX')
-ComposableEvolutionaryTechnique.add_to_map(op_map,
- "FloatArray",
- "op3_cross", strength=0.4)
-register(RandomThreeParentsComposableTechnique(name='ComposableDiffEvolutionCX',
- operator_map=op_map,
- population_size=30))
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/differentialevolution.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/differentialevolution.py
deleted file mode 100644
index cecffc460c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/differentialevolution.py
+++ /dev/null
@@ -1,148 +0,0 @@
-import random
-import time
-import logging
-from fn import _
-from technique import register
-from technique import SearchTechnique
-
-log = logging.getLogger(__name__)
-log.setLevel(logging.WARNING)
-
-
-class PopulationMember(object):
- def __init__(self, config, submitted=True):
- self.config = config
- self.submitted = submitted
- self.timestamp = time.time()
- self.candidate_replacement = None
-
- def touch(self):
- self.timestamp = time.time()
-
-
-class DifferentialEvolution(SearchTechnique):
- """
- based on http://cci.lbl.gov/cctbx_sources/scitbx/differential_evolution.py
- """
-
- def __init__(self,
- population_size=30,
- cr=0.9, # crossover rate
- n_cross=1, # force at least 1 to crossover
- information_sharing=1, # number token sharing pop members
- duplicate_retries=5, # how many times to retry on duplicate
- *pargs, **kwargs):
-
- self.population_size = population_size
- self.cr = cr
- self.n_cross = n_cross
- self.information_sharing = information_sharing
- self.population = None
- self.duplicate_retries = duplicate_retries
- self.limit = None
- super(DifferentialEvolution, self).__init__(*pargs, **kwargs)
-
- @classmethod
- def get_hyper_parameters(cls):
- return ['population_size', 'cr', 'n_cross', 'information_sharing']
-
- def initial_population(self):
- self.population = [PopulationMember(
- self.driver.get_configuration(
- self.manipulator.random()), submitted=False)
- for z in xrange(self.population_size)]
-
- def oldest_pop_member(self):
- # since tests are run in parallel, exclude things with a replacement pending
- pop_without_replacements = filter(lambda x: x.candidate_replacement is None,
- self.population)
- if not pop_without_replacements:
- # everything has a pending replacement
- return None
- pop_without_replacements.sort(key=_.timestamp)
- return pop_without_replacements[0]
-
- def desired_configuration(self):
- """
- return a cfg that we should test,
- """
- if not self.population:
- # first time called
- self.initial_population()
-
- # make sure initial population is completely submitted
- for p in self.population:
- if not p.submitted:
- p.submitted = True
- if p is self.population[-1]:
- log.info('initial population testing done')
- return p.config
-
- # pp is member of population to be replaced
- oldest_pop_member = self.oldest_pop_member()
- if not oldest_pop_member:
- return None
-
- config = None
- for retry in xrange(self.duplicate_retries):
- config = self.driver.get_configuration(
- self.create_new_configuration(oldest_pop_member))
- if not self.driver.has_results(config):
- break
- # new configuration would have been a duplicate, try again
-
- oldest_pop_member.touch() # move to back of the line for next replacement
- oldest_pop_member.candidate_replacement = config
- self.limit = self.driver.objective.limit_from_config(
- oldest_pop_member.config)
- return oldest_pop_member.candidate_replacement
-
- def create_new_configuration(self, parent_pop_member):
- cfg = self.manipulator.copy(parent_pop_member.config.data)
- cfg_params = self.manipulator.proxy(cfg)
-
- # pick 3 random parents, not pp
- shuffled_pop = list(set(self.population) - set([parent_pop_member]))
-
- # share information with other techniques
- if self.driver.best_result:
- shuffled_pop += ([PopulationMember(self.driver.best_result.configuration)]
- * self.information_sharing)
-
- random.shuffle(shuffled_pop)
- x1, x2, x3 = map(_.config.data, shuffled_pop[0:3])
-
- use_f = random.random() / 2.0 + 0.5
-
- params = self.manipulator.param_names(cfg, x1, x2, x3)
- random.shuffle(params)
- for i, k in enumerate(params):
- if i < self.n_cross or random.random() < self.cr:
- # cfg = x1 + use_f*(x2 - x3)
- cfg_params[k].op4_set_linear(x1, x2, x3, 1.0, use_f, -use_f)
-
- return cfg
-
- def handle_requested_result(self, result):
- """called when new results are added"""
- for p in self.population:
- if p.candidate_replacement == result.configuration:
- if self.objective.lt(p.candidate_replacement, p.config):
- # candidate replacement was better, replace it!
- p.config = p.candidate_replacement
- log.info('better point')
- p.candidate_replacement = None
-
-
-class DifferentialEvolutionAlt(DifferentialEvolution):
- def __init__(self, cr=0.2, **kwargs):
- kwargs['cr'] = cr
- super(DifferentialEvolutionAlt, self).__init__(**kwargs)
-
-
-register(DifferentialEvolution())
-register(DifferentialEvolutionAlt())
-register(DifferentialEvolution(population_size=100, cr=0.2,
- name='DifferentialEvolution_20_100'))
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/driver.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/driver.py
deleted file mode 100644
index 7924e36e6f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/driver.py
+++ /dev/null
@@ -1,301 +0,0 @@
-import argparse
-import copy
-import logging
-import os
-import sys
-
-from datetime import datetime
-from fn import _
-from opentuner.driverbase import DriverBase
-from opentuner.resultsdb.models import Configuration
-from opentuner.resultsdb.models import DesiredResult
-from opentuner.resultsdb.models import Result
-from opentuner.resultsdb.models import BanditInfo
-from opentuner.resultsdb.models import BanditSubTechnique
-from opentuner.search import plugin
-from opentuner.search import technique
-from opentuner.search.bandittechniques import AUCBanditMetaTechnique
-
-log = logging.getLogger(__name__)
-
-argparser = argparse.ArgumentParser(add_help=False)
-argparser.add_argument('--test-limit', type=int, default=5000,
- help='stop tuning after given tests count')
-argparser.add_argument('--stop-after', type=float,
- help='stop tuning after given seconds')
-argparser.add_argument('--parallelism', type=int, default=8,
- help='how many tests to support at once')
-argparser.add_argument('--pipelining', type=int, default=0,
- help='how long a delay (in generations) before results are available')
-argparser.add_argument('--bail-threshold', type=int, default=5000,
- help='abort if no requests have been made in X generations')
-argparser.add_argument('--no-dups', action='store_true',
- help='don\'t print out warnings for duplicate requests')
-argparser.add_argument('--seed-configuration', action='append', default=[],
- metavar='FILENAME', help="""
- Start search at a given configuration. Can be
- specified multiple times. Configurations are loaded
- with ConfigurationManipulator.load_from_file()
- and file format is detected from extension.""")
-
-
-class SearchDriver(DriverBase):
- """
- controls the search process managing root_technique and creating
- DesiredResults
- """
-
- def __init__(self, manipulator, extra_seeds=None, extra_criteria=None, **kwargs):
- super(SearchDriver, self).__init__(**kwargs)
- if extra_seeds is None:
- extra_seeds = []
- self.manipulator = manipulator
- self.wait_for_results = self.tuning_run_main.results_wait
- self.commit = self.tuning_run_main.commit
- self.extra_criteria = extra_criteria
-
- self.generation = 0
- self.test_count = 0
- self.plugins = plugin.get_enabled(self.args)
- self.pending_result_callbacks = list() # (DesiredResult, function) tuples
- # deepcopy is required to have multiple tuning runs in a single process
- if self.args.list_techniques:
- techniques, generators = technique.all_techniques()
- for t in techniques:
- print t.name
- sys.exit(0)
-
- if self.args.generate_bandit_technique:
- # generate a bandit
- self.root_technique = AUCBanditMetaTechnique.generate_technique(manipulator)
- else:
- self.root_technique = copy.deepcopy(technique.get_root(self.args))
-
- if isinstance(self.root_technique, AUCBanditMetaTechnique):
- self.session.flush()
- info = BanditInfo(tuning_run=self.tuning_run,
- c=self.root_technique.bandit.C,
- window=self.root_technique.bandit.window,)
- self.session.add(info)
- for t in self.root_technique.techniques:
- subtechnique = BanditSubTechnique(bandit_info=info,
- name=t.name)
- self.session.add(subtechnique)
-
-
- self.objective.set_driver(self)
- self.pending_config_ids = set()
- self.best_result = None
- self.new_results = []
-
- for t in self.plugins:
- t.set_driver(self)
- self.root_technique.set_driver(self)
- self.seed_cfgs = list(extra_seeds)
- for cfg_filename in reversed(self.args.seed_configuration):
- if os.path.exists(cfg_filename):
- self.seed_cfgs.append(manipulator.load_from_file(cfg_filename))
- else:
- log.error('no such file for --seed-configuration %s', cfg_filename)
-
- self.plugins.sort(key=_.priority)
-
- def add_plugin(self, p):
- if p in self.plugins:
- return
- self.plugins.append(p)
- self.plugins.sort(key=_.priority)
- p.set_driver(self)
-
- def convergence_criteria(self):
- """returns true if the tuning process should stop"""
- if self.args.stop_after:
- elapsed = (datetime.now() - self.tuning_run.start_date)
- try:
- elapsed = elapsed.total_seconds()
- except: # python 2.6
- elapsed = elapsed.days * 86400 + elapsed.seconds
- if elapsed > self.args.stop_after:
- return True
- if self.test_count > self.args.test_limit:
- return True
- if self.extra_criteria:
- if self.extra_criteria(self.new_results):
- return True
- return False
-
- def register_result_callback(self, desired_result, callback):
- if desired_result.result is not None:
- callback(desired_result.result)
- else:
- self.pending_result_callbacks.append((desired_result, callback))
-
- def result_callbacks(self):
- pending = self.pending_result_callbacks
- self.pending_result_callbacks = list()
- for dr, callback in pending:
- if dr.result is not None:
- callback(dr.result)
- continue
- elif self.generation - dr.generation > self.args.pipelining:
- # see if we can find a result
- results = self.results_query(config=dr.configuration).all()
- log.warning("Result callback %d (requestor=%s) pending for "
- "%d generations %d results available",
- dr.id, dr.requestor, self.generation - dr.generation,
- len(results))
- if len(results):
- dr.result = results[0]
- callback(dr.result)
- continue
- # try again later
- self.pending_result_callbacks.append((dr, callback))
-
- def has_results(self, config):
- return self.results_query(config=config).count() > 0
-
- def run_generation_techniques(self):
- tests_this_generation = 0
- self.plugin_proxy.before_techniques()
- for z in xrange(self.args.parallelism):
-
- if self.seed_cfgs:
- config = self.get_configuration(self.seed_cfgs.pop())
- dr = DesiredResult(configuration=config,
- requestor='seed',
- generation=self.generation,
- request_date=datetime.now(),
- tuning_run=self.tuning_run)
- else:
- dr = self.root_technique.desired_result()
- if dr is None or dr is False:
- log.debug("no desired result, skipping to testing phase")
- break
- self.session.flush() # populate configuration_id
- duplicates = (self.session.query(DesiredResult)
- .filter_by(tuning_run=self.tuning_run,
- configuration_id=dr.configuration_id)
- .filter(DesiredResult.id != dr.id)
- .order_by(DesiredResult.request_date)
- .limit(1).all())
- self.session.add(dr)
- if len(duplicates):
- if not self.args.no_dups:
- log.warning("duplicate configuration request #%d %s/%s %s",
- self.test_count,
- dr.requestor,
- duplicates[0].requestor,
- 'OLD' if duplicates[0].result else 'PENDING')
- self.session.flush()
- desired_result_id = dr.id
-
- def callback(result):
- dr = self.session.query(DesiredResult).get(desired_result_id)
- dr.result = result
- dr.state = 'COMPLETE'
- dr.start_date = datetime.now()
-
- self.register_result_callback(duplicates[0], callback)
- else:
- log.debug("desired result id=%d, cfg=%d", dr.id, dr.configuration_id)
- dr.state = 'REQUESTED'
- self.test_count += 1
- tests_this_generation += 1
- self.plugin_proxy.after_techniques()
- return tests_this_generation
-
- def process_new_results(self):
- self.new_results = []
- for result in (self.results_query()
- .filter_by(was_new_best=None)
- .order_by(Result.collection_date)):
- self.plugin_proxy.on_result(result)
- self.new_results.append(result)
- if self.best_result is None:
- self.best_result = result
- result.was_new_best = True
- elif self.objective.lt(result, self.best_result):
- self.best_result = result
- result.was_new_best = True
- self.plugin_proxy.on_new_best_result(result)
- else:
- result.was_new_best = False
- self.result_callbacks()
-
- def run_generation_results(self, offset=0):
- self.commit()
- self.plugin_proxy.before_results_wait()
- self.wait_for_results(self.generation + offset)
- self.plugin_proxy.after_results_wait()
- self.process_new_results()
-
- @property
- def plugin_proxy(self):
- """
- forward any method calls on the returned object to all plugins
- """
- plugins = self.plugins
-
- class PluginProxy(object):
- def __getattr__(self, method_name):
- def plugin_method_proxy(*args, **kwargs):
- rv = []
- for plugin in plugins:
- rv.append(getattr(plugin, method_name)(*args, **kwargs))
- return filter(lambda x: x is not None, rv)
-
- return plugin_method_proxy
-
- return PluginProxy()
-
- def get_configuration(self, cfg):
- """called by SearchTechniques to create Configuration objects"""
- self.manipulator.normalize(cfg)
- hashv = self.manipulator.hash_config(cfg)
- config = Configuration.get(self.session,self.program, hashv, cfg)
- return config
-
- def main(self):
- self.plugin_proxy.set_driver(self)
- self.plugin_proxy.before_main()
-
- no_tests_generations = 0
-
- # prime pipeline with tests
- for z in xrange(self.args.pipelining):
- self.run_generation_techniques()
- self.generation += 1
-
- while not self.convergence_criteria():
- if self.run_generation_techniques() > 0:
- no_tests_generations = 0
- elif no_tests_generations <= self.args.bail_threshold:
- no_tests_generations += 1
- else:
- break
- self.run_generation_results(offset=-self.args.pipelining)
- self.generation += 1
-
- self.plugin_proxy.after_main()
-
- def external_main_begin(self):
- self.plugin_proxy.set_driver(self)
- self.plugin_proxy.before_main()
-
- def external_main_generation(self):
- if self.generation > 0:
- self.plugin_proxy.after_results_wait()
- self.process_new_results()
- self.run_generation_techniques()
- self.commit()
- self.plugin_proxy.before_results_wait()
-
- def external_main_end(self):
- self.plugin_proxy.after_main()
-
-
-
-
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/evolutionarytechniques.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/evolutionarytechniques.py
deleted file mode 100644
index e663ac1345..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/evolutionarytechniques.py
+++ /dev/null
@@ -1,153 +0,0 @@
-import abc
-import copy
-import random
-from technique import SearchTechnique
-from opentuner.search import technique
-
-class EvolutionaryTechnique(SearchTechnique):
- def __init__(self,
- mutation_rate = 0.1,
- crossover_rate = 0.0,
- must_mutate_count = 1,
- *pargs, **kwargs):
- super(EvolutionaryTechnique, self).__init__(*pargs, **kwargs)
- self.mutation_rate = mutation_rate
- self.crossover_rate = crossover_rate
- self.must_mutate_count = must_mutate_count
-
- @classmethod
- def get_hyper_parameters(cls):
- return ['mutation_rate', 'crossover_rate', 'must_mutate_count']
-
-
- def desired_configuration(self):
- """
- return a (cfg, priority) that we should test,
- through random mutation and crossover
- """
- #TODO: set limit value
-
- parents = self.selection()
- parents = map(copy.deepcopy, parents)
- parent_hashes = map(self.manipulator.hash_config, parents)
-
- if len(parents) > 1:
- cfg = self.crossover(parents)
- else:
- cfg = parents[0]
-
- for z in xrange(10): #retries
- self.mutation(cfg)
- if self.manipulator.hash_config(cfg) in parent_hashes:
- continue # try again
- return cfg
-
- def mutation(self, cfg):
- """
- mutate cfg in place
- """
- params = self.manipulator.parameters(cfg)
- random.shuffle(params)
- for param in params[:self.must_mutate_count]:
- self.mutate_param(cfg, param)
- for param in params[self.must_mutate_count:]:
- if random.random() < self.mutation_rate:
- self.mutate_param(cfg, param)
-
- def mutate_param(self, cfg, param):
- """
- mutate single parameter of cfg in place
- """
- param.op1_randomize(cfg)
-
- def crossover(self):
- raise Exception('Not implemented')
-
- def selection(self):
- """return a list of parent configurations to use"""
- if random.random() < self.crossover_rate:
- return [self.select(),
- self.select()]
- else:
- return [self.select()]
-
- @abc.abstractmethod
- def select(self):
- """return a single random parent configuration"""
- return None
-
-class GreedySelectionMixin(object):
- """
- EvolutionaryTechnique mixin for greedily selecting the best known
- configuration
- """
- def select(self):
- """return a single random parent configuration"""
- if (self.driver.best_result is not None and
- self.driver.best_result.state == 'OK'):
- return self.driver.best_result.configuration.data
- else:
- return self.manipulator.random()
-
-class NormalMutationMixin(object):
- """
- Mutate primitive parameters according to normal distribution
- """
-
- def __init__(self, sigma = 0.1, *pargs, **kwargs):
- super(NormalMutationMixin, self).__init__(*pargs, **kwargs)
- self.sigma = sigma
-
- def mutate_param(self, cfg, param):
- """
- mutate single parameter of cfg in place
- """
- if param.is_primitive():
- param.op1_normal_mutation(cfg, self.sigma)
- else:
- random.choice(param.manipulators(cfg))(cfg)
-
-
-class CrossoverMixin(object):
- def __init__(self, crossover, *pargs, **kwargs):
- super(CrossoverMixin, self).__init__(*pargs, **kwargs)
- self.crossover_op = crossover
- self.name = 'ga-'+crossover.replace("op3_cross_","")
-
- def crossover(self, cfgs):
- """
- Crossover the first permtation parameter, if found, of two parents and
- return one offspring cfg
- """
- cfg1, cfg2, = cfgs
- new = self.manipulator.copy(cfg1)
- params = self.manipulator.parameters(cfg1)
- for param in params:
- if param.is_permutation() and param.size>6:
- getattr(param, self.crossover_op)(new, cfg1, cfg2, d=param.size/3)
- return new
-
-
-class UniformGreedyMutation(GreedySelectionMixin, EvolutionaryTechnique):
- pass
-
-class NormalGreedyMutation(NormalMutationMixin, GreedySelectionMixin, EvolutionaryTechnique):
- pass
-
-class GA(CrossoverMixin, UniformGreedyMutation):
- pass
-
-technique.register(GA(crossover = 'op3_cross_OX3', mutation_rate=0.10, crossover_rate=0.8))
-technique.register(GA(crossover = 'op3_cross_OX1', mutation_rate=0.10,crossover_rate=0.8))
-technique.register(GA(crossover = 'op3_cross_PX', mutation_rate=0.10, crossover_rate=0.8))
-technique.register(GA(crossover = 'op3_cross_CX', mutation_rate=0.10, crossover_rate=0.8))
-technique.register(GA(crossover = 'op3_cross_PMX', mutation_rate=0.10, crossover_rate=0.8))
-technique.register(UniformGreedyMutation(name='ga-base', mutation_rate=0.10))
-
-technique.register(UniformGreedyMutation(name='UniformGreedyMutation05', mutation_rate=0.05))
-technique.register(UniformGreedyMutation(name='UniformGreedyMutation10', mutation_rate=0.10))
-technique.register(UniformGreedyMutation(name='UniformGreedyMutation20', mutation_rate=0.20))
-technique.register(NormalGreedyMutation(name='NormalGreedyMutation05', mutation_rate=0.05))
-technique.register(NormalGreedyMutation(name='NormalGreedyMutation10', mutation_rate=0.10))
-technique.register(NormalGreedyMutation(name='NormalGreedyMutation20', mutation_rate=0.20))
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/globalGA.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/globalGA.py
deleted file mode 100644
index e9b1f71174..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/globalGA.py
+++ /dev/null
@@ -1,125 +0,0 @@
-import abc
-import copy
-import random
-from technique import SearchTechnique
-from opentuner.search import technique
-
-class GlobalEvolutionaryTechnique(SearchTechnique):
- def __init__(self,
- mutation_rate = 0.1,
- crossover_rate = 0.0,
- must_mutate_count = 1,
- crossover_strength = 0.1,
- *pargs, **kwargs):
- super(GlobalEvolutionaryTechnique, self).__init__(*pargs, **kwargs)
- self.mutation_rate = mutation_rate
- self.crossover_rate = crossover_rate
- self.must_mutate_count = must_mutate_count
- self.crossover_strength = crossover_strength
-
- @classmethod
- def get_hyper_parameters(cls):
- return ['mutation_rate', 'crossover_rate', 'must_mutate_count', 'crossover_strength']
-
- def desired_configuration(self):
- """
- return a (cfg, priority) that we should test,
- through random mutation and crossover
- """
- #TODO: set limit value
-
- parents = self.selection()
- parents = map(copy.deepcopy, parents)
- parent_hashes = map(self.manipulator.hash_config, parents)
-
- if len(parents) > 1:
- cfg = self.crossover(parents)
- else:
- cfg = parents[0]
-
- for z in xrange(10): #retries
- self.mutation(cfg)
- if self.manipulator.hash_config(cfg) in parent_hashes:
- continue # try again
- return cfg
-
- def mutation(self, cfg):
- """
- mutate cfg in place
- """
- params = self.manipulator.parameters(cfg)
- random.shuffle(params)
- for param in params[:self.must_mutate_count]:
- self.mutate_param(cfg, param)
- for param in params[self.must_mutate_count:]:
- if random.random() < self.mutation_rate:
- self.mutate_param(cfg, param)
-
- def mutate_param(self, cfg, param):
- """
- mutate single parameter of cfg in place
- """
- param.op1_randomize(cfg)
-
- def crossover(self, cfgs):
- cfg1, cfg2, = cfgs
- new = self.manipulator.copy(cfg1)
- params = self.manipulator.parameters(cfg1)
- random.shuffle(params)
- d = int(self.crossover_strength*len(params))
- for param in params[:d]:
- param.set_value(new, param.get_value(cfg2))
- return new
-
- def selection(self):
- """return a list of parent configurations to use"""
- if random.random() < self.crossover_rate:
- return [self.select(),
- self.select()]
- else:
- return [self.select()]
-
- @abc.abstractmethod
- def select(self):
- """return a single random parent configuration"""
- return None
-
-class GreedySelectionMixin(object):
- """
- EvolutionaryTechnique mixin for greedily selecting the best known
- configuration
- """
- def select(self):
- """return a single random parent configuration"""
- if (self.driver.best_result is not None and
- self.driver.best_result.state == 'OK'):
- return self.driver.best_result.configuration.data
- else:
- return self.manipulator.random()
-
-class NormalMutationMixin(object):
- """
- Mutate primitive parameters according to normal distribution
- """
-
- def __init__(self, sigma = 0.1, *pargs, **kwargs):
- super(NormalMutationMixin, self).__init__(*pargs, **kwargs)
- self.sigma = sigma
-
- def mutate_param(self, cfg, param):
- """
- mutate single parameter of cfg in place
- """
- if param.is_primitive():
- param.op1_normal_mutation(cfg, self.sigma)
- else:
- random.choice(param.manipulators(cfg))(cfg)
-
-
-class UniformGreedyMutation(GreedySelectionMixin, GlobalEvolutionaryTechnique):
- pass
-
-class NormalGreedyMutation(NormalMutationMixin, GreedySelectionMixin, GlobalEvolutionaryTechnique):
- pass
-
-technique.register(NormalGreedyMutation( crossover_rate=0.5, crossover_strength=0.2, name='GGA'))
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/manipulator.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/manipulator.py
deleted file mode 100755
index decd476bf3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/manipulator.py
+++ /dev/null
@@ -1,1853 +0,0 @@
-# vim: tabstop=2 shiftwidth=2 softtabstop=2 expandtab autoindent smarttab
-import abc
-import collections
-import copy
-import hashlib
-import json
-import logging
-import math
-import os
-import pickle
-import random
-from fn import _
-import argparse
-from datetime import datetime
-import numpy
-import inspect
-import sys
-
-log = logging.getLogger(__name__)
-argparser = argparse.ArgumentParser(add_help=False)
-argparser.add_argument('--list-params', '-lp',
- help='list available parameter classes')
-
-
-class ConfigurationManipulatorBase(object):
- """
- abstract interface for objects used by search techniques to mutate
- configurations
- """
- __metaclass__ = abc.ABCMeta
-
- # List of file formats, which can be extended by subclasses. Used in
- # write_to_file() and load_from_file(). Objects in list must define
- # load(fd) and dump(cfg, fd).
- FILE_FORMATS = {'default': json, 'json': json,
- 'pickle': pickle, 'pk': pickle}
-
- def validate(self, config):
- """is the given config valid???"""
- return all(map(_.validate(config), self.parameters(config)))
-
- def normalize(self, config):
- """mutate config into canonical form"""
- for param in self.parameters(config):
- param.normalize(config)
-
- def set_search_driver(self, search_driver):
- """called exactly once during setup"""
- pass
-
- def copy(self, config):
- """produce copy of config"""
- return copy.deepcopy(config)
-
- def parameters_dict(self, config):
- """convert self.parameters() to a dictionary by name"""
- return dict([(p.name, p) for p in self.parameters(config)])
-
- def param_names(self, *args):
- """return union of parameter names in args"""
- return sorted(reduce(set.union,
- [set(map(_.name, self.parameters(cfg)))
- for cfg in args]))
-
- def linear_config(self, a, cfg_a, b, cfg_b, c, cfg_c):
- """return a configuration that is a linear combination of 3 other configs"""
- dst = self.copy(cfg_a)
- dst_params = self.proxy(dst)
- for k in self.param_names(dst, cfg_a, cfg_b, cfg_c):
- dst_params[k].op4_set_linear(cfg_a, cfg_b, cfg_c, a, b, c)
- return dst
-
- def _get_serializer(self, filename, format=None):
- """
- Extract the correct file format serializer from self.FILE_FORMATS.
- Guess the format by extension if one is not given.
- """
- if format is None:
- format = os.path.splitext(filename)[1].lower().replace('.', '')
- if format not in self.FILE_FORMATS:
- serializer = self.FILE_FORMATS['default']
- if len(self.FILE_FORMATS) > 1:
- log.warning('Unknown file format "%s", using "%s" instead', format,
- serializer.__name__)
- else:
- serializer = self.FILE_FORMATS[format]
- return serializer
-
- def save_to_file(self, cfg, filename, format=None):
- """
- Write cfg to filename. Guess the format by extension if one is not given.
- """
- with open(filename, 'a+') as fd:
- self._get_serializer(filename, format).dump(cfg, fd)
-
- def load_from_file(self, filename, format=None):
- """
- Read cfg from filename. Guess the format by extension if one is not given.
- """
- with open(filename, 'rb') as fd:
- return self._get_serializer(filename, format).load(fd)
-
- def proxy(self, cfg):
- return ManipulatorProxy(self, cfg)
-
- @abc.abstractmethod
- def random(self):
- """produce a random initial configuration"""
- return
-
- @abc.abstractmethod
- def parameters(self, config):
- """return a list of of Parameter objects"""
- return list()
-
- @abc.abstractmethod
- def hash_config(self, config):
- """produce unique hash value for the given config"""
- return
-
-
-class ConfigurationManipulator(ConfigurationManipulatorBase):
- """
- a configuration manipulator using a fixed set of parameters and storing
- configs in a dict-like object
- """
-
- def __init__(self, params=None, config_type=dict, seed_config=None, **kwargs):
- if params is None:
- params = []
- self.params = list(params)
- self.config_type = config_type
- self.search_driver = None
- self._seed_config = seed_config
- super(ConfigurationManipulator, self).__init__(**kwargs)
- for p in self.params:
- p.parent = self
-
- def add_parameter(self, p):
- p.set_parent(self)
- self.params.append(p)
-
- #TODO sub parameters should be recursed on
- # not currently an issue since no doubly-nested sub-parameters
- sub_params = p.sub_parameters()
- for sp in sub_params:
- sp.set_parent(p)
- self.params.extend(sub_params)
-
- def set_search_driver(self, search_driver):
- self.search_driver = search_driver
-
- def seed_config(self):
- """produce a fixed seed configuration"""
- if self._seed_config:
- cfg = copy.deepcopy(self._seed_config)
- else:
- cfg = self.config_type()
- for p in self.params:
- if not isinstance(p.name, str) or '/' not in p.name:
- cfg[p.name] = p.seed_value()
- return cfg
-
- def random(self):
- """produce a random configuration"""
- cfg = self.seed_config()
- for p in self.parameters(cfg):
- p.op1_randomize(cfg)
- return cfg
-
- def parameters(self, config):
- """return a list of Parameter objects"""
- if type(config) is not self.config_type:
- log.error("wrong type, expected %s got %s",
- str(self.config_type),
- str(type(config)))
- raise TypeError()
- return self.params
-
- def parameters_to_json(self):
- """
- output information about the parameters in this manipulator in json format:
- [ConfigurationManipulator,{pinfo:count,pinfo:count ...}]
- where pinfo has a similar form to describe the parameter's sub-parameters:
- [param_name,{pinfo:count,pinfo:count ...}]
- """
- def param_info_to_json(param, sub_parameters):
- """
- recursively output information about a parameter and its subparameters in a json format:
-
- [parameter_name, {subparam_info:count,subparam_info:count,...}]
- or if no subparams
- [parameter_name,{}]
-
- where subparam_info are sorted alphabetically. Note we can't directly use json since
- sets/dictionaries aren't always ordered by key
- """
- sub_parameter_counts = {}
- # build the string
- if isinstance(param, str):
- param_name = param
- else:
- param_name = param.__class__.__name__
- out = ['[', param_name, ',{']
-
- if len(sub_parameters) > 0:
- # count sub params
- for sp in sub_parameters:
- spout = param_info_to_json(sp, sp.sub_parameters())
- sub_parameter_counts[spout] = sub_parameter_counts.get(spout, 0) + 1
- # add the count map in sorted order
- for sp in sorted(sub_parameter_counts):
- out.append(sp)
- out.append(':')
- out.append(str(sub_parameter_counts[sp]))
- out.append(',')
- out.pop() # remove trailing comma
-
- out.append('}]')
- return ''.join(out)
-
- # filter out subparameters to avoid double counting
- params = [p for p in self.params if p.parent is self]
- return param_info_to_json(self, params)
-
- def hash_config(self, config):
- """produce unique hash value for the given config"""
- m = hashlib.sha256()
- params = list(self.parameters(config))
- params.sort(key=_.name)
- for i, p in enumerate(params):
- m.update(str(p.name))
- m.update(p.hash_value(config))
- m.update(str(i))
- m.update("|")
- return m.hexdigest()
-
- def search_space_size(self):
- """estimate the size of the search space, not precise"""
- return reduce(_ * _, [x.search_space_size() for x in self.params])
-
- def difference(self, cfg1, cfg2):
- cfg = self.copy(cfg1)
- for param in self.parameters(cfg1):
- if param.is_primitive(cfg1):
- # TODO: check range
- param.set_value(cfg, param.get_value(cfg1) - param.get_value(cfg2))
- else:
- pass
- return cfg
-
- def applySVs(self, cfg, sv_map, args, kwargs):
- """
- Apply operators to each parameter according to given map. Updates cfg.
- Parameters with no operators specified are not updated.
- cfg: configuration data
- sv_map: python dict that maps string parameter name to class method name
- arg_map: python dict that maps string parameter name to class method
- arguments
- """
- # TODO: check consistency between sv_map and cfg
- param_dict = self.parameters_dict(cfg)
- for pname in self.param_names(cfg):
- param = param_dict[pname]
- getattr(param, sv_map[pname])(cfg, *args[pname], **kwargs[pname])
-
-
-class Parameter(object):
- """
- abstract base class for parameters in a ConfigurationManipulator
- """
- __metaclass__ = abc.ABCMeta
-
- def __init__(self, name):
- self.name = name
- self.parent = None
- super(Parameter, self).__init__()
-
- def _to_storage_type(self, val):
- """hook to support transformation applied while stored"""
- return val
-
- def _from_storage_type(self, sval):
- """hook to support transformation applied while stored"""
- return sval
-
- def _read_node(self, config):
- """hook to support different storage structures"""
- node = config
- if not isinstance(self.name, str):
- return node, self.name
- name_parts = self.name.split('/')
- for part in name_parts[:-1]:
- if isinstance(node, list):
- part = int(part)
- node = node[part]
- part = name_parts[-1]
- if isinstance(node, list):
- part = int(part)
- return node, part
-
- def _get(self, config):
- """hook to support different storage structures"""
- node, part = self._read_node(config)
- return self._from_storage_type(node[part])
-
- def _set(self, config, v):
- """hook to support different storage structures"""
- node, part = self._read_node(config)
- node[part] = self._to_storage_type(v)
-
- def set_parent(self, manipulator):
- self.parent = manipulator
-
- def validate(self, config):
- """is the given config valid???"""
- return True
-
- def is_primitive(self, ignored=None):
- return isinstance(self, PrimitiveParameter)
-
- def is_permutation(self, ignored=None):
- return isinstance(self, PermutationParameter)
-
- def manipulators(self, config):
- """
- a list of manipulator functions to change this value in the config
- manipulators must be functions that take a config and change it in place
-
- default implementation just has op1_randomize as only operation
- """
- return [self.op1_randomize]
-
- def normalize(self, config):
- """
- mutate this parameter into a canonical form
- """
- pass
-
- def sub_parameters(self):
- """
- additional parameters added with this parameter
- """
- return []
-
- @abc.abstractmethod
- def op1_randomize(self, cfg):
- """
- Set this parameter's value in a configuration to a random value
-
- :param config: the configuration to be changed
- """
- pass
-
- @abc.abstractmethod
- def seed_value(self):
- """some legal value of this parameter (for creating initial configs)"""
- return
-
- @abc.abstractmethod
- def copy_value(self, src, dst):
- """copy the value of this parameter from src to dst config"""
- pass
-
- @abc.abstractmethod
- def same_value(self, cfg1, cfg2):
- """test if cfg1 and cfg2 have the same value of this parameter"""
- return
-
- @abc.abstractmethod
- def hash_value(self, config):
- """produce unique hash for this value in the config"""
- return
-
- @abc.abstractmethod
- def op4_set_linear(self, cfg, cfg_a, cfg_b, cfg_c, a, b, c):
- """
- Sets the parameter value in a configuration to a linear combination of 3
- other configurations: :math:`a*cfg_a + b*cfg_b + c*cfg_c`
-
- :param cfg: the configuration to be changed
- :param cfg_a: a parent configuration
- :param cfg_b: a parent configuration
- :param cfg_c: a parent configuration
- :param a: weight for cfg_a
- :param b: weight for cfg_b
- :param c: weight for cfg_c
- """
- pass
-
- def search_space_size(self):
- return 1
-
- def op1_nop(self, cfg):
- """
- The 'null' operator. Does nothing.
-
- :param cfg: the configuration to be changed
- """
- pass
-
- # Stochastic variators
- def op3_swarm(self, cfg, cfg1, cfg2, c, c1, c2, *args, **kwargs):
- """
- Stochastically 'move' the parameter value in a configuration towards those
- in two parent configurations. This is done by calling :py:meth:`opn_stochastic_mix`
-
- :param cfg: the configuration to be changed
- :param cfg1: a parent configuration
- :param cfg2: a parent configuration
- :param c: weight of original configuration
- :param c1: weight for cfg1
- :param c2: weight for cfg2
- """
- # default to probabilistic treatment
- self.opn_stochastic_mix(cfg, [cfg, cfg1, cfg2], [c, c1, c2])
-
- def opn_stochastic_mix(self, cfg, cfgs, ratio, *args, **kwargs):
- """
- Stochastically recombine a list of parent values into a single result.
-
- This randomly copies a value from a list of parents configurations according
- to a list of weights.
-
- :param cfg: the configuration to be changed
- :param cfgs: a list of parent configurations
- :param ratio: a list of floats representing the weight of each configuration
- in cfgs
-
- """
- assert len(cfgs) == len(ratio)
- r = random.random()
- c = numpy.array(ratio, dtype=float) / sum(ratio)
- for i in range(len(c)):
- if r < sum(c[:i + 1]):
- self.copy_value(cfg, cfgs[i])
- break
-
-
-class PrimitiveParameter(Parameter):
- """
- An abstract interface implemented by parameters that represent a single
- dimension in a cartesian space in a legal range
- """
- __metaclass__ = abc.ABCMeta
-
- def __init__(self, name, value_type=float, **kwargs):
- self.value_type = value_type
- super(PrimitiveParameter, self).__init__(name, **kwargs)
-
- def hash_value(self, config):
- """produce unique hash for this value in the config"""
- self.normalize(config)
- return hashlib.sha256(repr(self.get_value(config))).hexdigest()
-
- def copy_value(self, src, dst):
- """copy the value of this parameter from src to dst config"""
- self.set_value(dst, self.get_value(src))
-
- def same_value(self, cfg1, cfg2):
- """test if cfg1 and cfg2 have the same value of this parameter"""
- return self.get_value(cfg1) == self.get_value(cfg2)
-
- def is_integer_type(self):
- """true if self.value_type can only represent integers"""
- return self.value_type(0) == self.value_type(0.1)
-
- def get_unit_value(self, config):
- """get_value scaled such that range is between 0.0 and 1.0"""
- low, high = self.legal_range(config)
- if self.is_integer_type():
- # account for rounding
- low -= 0.4999
- high += 0.4999
- val = self.get_value(config)
- if low < high:
- return float(val - low) / float(high - low)
- else:
- if low > high:
- log.warning('invalid range for parameter %s, %s to %s',
- self.name, low, high)
- # only a single legal value!
- return 0.0
-
- def set_unit_value(self, config, unit_value):
- """set_value scaled such that range is between 0.0 and 1.0"""
- assert 0.0 <= unit_value <= 1.0
- low, high = self.legal_range(config)
- if self.is_integer_type():
- # account for rounding
- low -= 0.4999
- high += 0.4999
- if low < high:
- val = unit_value * float(high - low) + low
- if self.is_integer_type():
- val = round(val)
- val = max(low, min(val, high))
- self.set_value(config, self.value_type(val))
-
- def op1_normal_mutation(self, cfg, sigma=0.1, *args, **kwargs):
- """
- apply normally distributed noise to this parameter's value in a
- configuration
-
- :param cfg: The configuration to be changed
- :param sigma: the std. deviation of the normally distributed noise on a unit
- scale
- """
- v = self.get_unit_value(cfg)
- v += random.normalvariate(0.0, sigma)
- # handle boundary cases by reflecting off the edge
- if v < 0.0:
- v *= -1.0
- if v > 1.0:
- v = 1.0 - (v % 1)
- self.set_unit_value(cfg, v)
-
- def op4_set_linear(self, cfg, cfg_a, cfg_b, cfg_c, a, b, c):
- """
- set the parameter value in a configuration to a linear combination of 3
- other configurations: :math:`a*cfg_a + b*cfg_b + c*cfg_c`
-
- :param cfg: The configuration to be changed
- :param cfg_a: a parent configuration
- :param cfg_b: a parent configuration
- :param cfg_c: a parent configuration
- :param a: weight for cfg_a
- :param b: weight for cfg_b
- :param c: weight for cfg_c
- """
- va = self.get_unit_value(cfg_a)
- vb = self.get_unit_value(cfg_b)
- vc = self.get_unit_value(cfg_c)
- v = a * va + b * vb + c * vc
- v = max(0.0, min(v, 1.0))
-
- self.set_unit_value(cfg, v)
-
- def manipulators(self, config):
- """
- a list of manipulator functions to change this value in the config
- manipulators must be functions that take a config and change it in place
-
- for primitive params default implementation is uniform random and normal
- """
- return [self.op1_randomize, self.op1_normal_mutation]
-
- @abc.abstractmethod
- def set_value(self, config, value):
- """assign this value in the given configuration"""
- pass
-
- @abc.abstractmethod
- def get_value(self, config):
- """retrieve this value from the given configuration"""
- return 0
-
- @abc.abstractmethod
- def legal_range(self, config):
- """return the legal range for this parameter, inclusive"""
- return 0, 1
-
-
-class NumericParameter(PrimitiveParameter):
- """
- A parameter representing a number with a minimum and maximum value
- """
- def __init__(self, name, min_value, max_value, **kwargs):
- """min/max are inclusive"""
- assert min_value <= max_value
- super(NumericParameter, self).__init__(name, **kwargs)
- # after super call so self.value_type is initialized
- self.min_value = self.value_type(min_value)
- self.max_value = self.value_type(max_value)
-
- def seed_value(self):
- """some legal value of this parameter (for creating initial configs)"""
- return self.min_value
-
- def set_value(self, config, value):
- assert value >= self.min_value
- assert value <= self.max_value
- self._set(config, value)
-
- def get_value(self, config):
- return self._get(config)
-
- def legal_range(self, config):
- return self.min_value, self.max_value
-
- def op1_randomize(self, config):
- """
- Set this parameter's value in a configuration to a random value in its legal
- range
-
- :param config: the configuration to be changed
- """
- if self.is_integer_type():
- self.set_value(config, random.randint(*self.legal_range(config)))
- else:
- self.set_value(config, random.uniform(*self.legal_range(config)))
-
- def op1_scale(self, cfg, k):
- """
- Scale this parameter's value in a configuration by a constant factor
-
- :param cfg: the configuration to be changed
- :param k: the constant factor to scale the parameter value by
- """
- v = self.get_value(cfg) * k
- v = max(self.min_value, min(self.max_value, v))
- self.set_value(cfg, v)
-
- def op3_difference(self, cfg, cfg1, cfg2):
- """
- Set this parameter's value in a configuration to the difference between this
- parameter's values in 2 other configs (cfg2 - cfg1)
-
- :param cfg: the configuration to be changed
- :param cfg1: The configuration whose parameter value is being subtracted
- :param cfg2: The configuration whose parameter value is subtracted from
- """
- v = self.get_value(cfg2) - self.get_value(cfg1)
- v = max(self.min_value, min(self.max_value, v))
- self.set_value(cfg, v)
-
- def opn_sum(self, cfg, *cfgs):
- """
- Set this parameter's value in a configuration to the sum of it's values in a
- list of configurations
-
- :param cfg: the configuration to be changed
- :param cfgs: a list of configurations to sum
- """
- v = sum([self.get_value(c) for c in cfgs])
- v = max(self.min_value, min(self.max_value, v))
- self.set_value(cfg, v)
-
- def search_space_size(self):
- if self.value_type is float:
- return 2 ** 32
- else:
- return self.max_value - self.min_value + 1 # inclusive range
-
-
-class IntegerParameter(NumericParameter):
- """
- A parameter representing an integer value in a legal range
- """
- def __init__(self, name, min_value, max_value, **kwargs):
- """min/max are inclusive"""
- kwargs['value_type'] = int
- super(IntegerParameter, self).__init__(name, min_value, max_value, **kwargs)
-
- def op3_swarm(self, cfg, cfg1, cfg2, c=1, c1=0.5,
- c2=0.5, velocity=0, sigma=0.2, *args, **kwargs):
- """
- Simulates a single update step in particle swarm optimization by updating
- the current position and returning a new velocity.
-
- The new velocity is given by
-
- .. math:: c*velocity + r1*c1*(cfg1-cfg) + r2*c2*(cfg2-cfg)
-
- where r1 and r2 are random values between 0 and 1.
-
- The new current position is the new velocity with gaussian noise added.
-
- :param cfg: the configuration to be changed. Represents the current position
- :param cfg1: a configuration to shift towards. Should be the local best
- position
- :param cfg2: a configuration to shift towards. Should be the global best
- position
- :param c: the weight of the current velocity
- :param c1: weight of cfg1
- :param c2: weight of cfg2
- :param velocity: the old velocity
- :param sigma: standard deviation of the gaussian noise, on a unit-scale
- :return: the new velocity, a float
-
- """
- vmin, vmax = self.legal_range(cfg)
- k = vmax - vmin
- # calculate the new velocity
- v = velocity * c + (self.get_value(cfg1) - self.get_value(
- cfg)) * c1 * random.random() + (self.get_value(
- cfg2) - self.get_value(cfg)) * c2 * random.random()
- # Map velocity to continuous space with sigmoid
- s = k / (1 + numpy.exp(-v)) + vmin
- # Add Gaussian noise
- p = random.gauss(s, sigma * k)
- # Discretize and bound
- p = int(min(vmax, max(round(p), vmin)))
- self.set_value(cfg, p)
- return v
-
-
-class FloatParameter(NumericParameter):
- def __init__(self, name, min_value, max_value, **kwargs):
- """min/max are inclusive"""
- kwargs['value_type'] = float
- super(FloatParameter, self).__init__(name, min_value, max_value, **kwargs)
-
- def op3_swarm(self, cfg, cfg1, cfg2, c=1, c1=0.5,
- c2=0.5, velocity=0, *args, **kwargs):
- """
-
- Simulates a single update step in particle swarm optimization by updating
- the current position and returning a new velocity.
-
- The new velocity is given by
-
- .. math:: c*velocity + r1*c1*(cfg1-cfg) + r2*c2*(cfg2-cfg)
-
- where r1 and r2 are random values between 0 and 1
-
- The new current position is the old current position offset by the new
- velocity:
-
- :param cfg: the configuration to be changed. Represents the current position
- :param cfg1: a configuration to shift towards. Should be the local best
- position
- :param cfg2: a configuration to shift towards. Should be the global best
- position
- :param c: the weight of the current velocity
- :param c1: weight of cfg1
- :param c2: weight of cfg2
- :param velocity: the old velocity
- :return: the new velocity, a float
-
- """
- vmin, vmax = self.legal_range(cfg)
- v = velocity * c + (self.get_value(cfg1) - self.get_value(
- cfg)) * c1 * random.random() + (self.get_value(
- cfg2) - self.get_value(cfg)) * c2 * random.random()
- p = self.get_value(cfg) + v
- p = min(vmax, max(p, vmin))
- self.set_value(cfg, p)
- return v
-
-
-class ScaledNumericParameter(NumericParameter):
- """
- A Parameter that is stored in configurations normally, but has a scaled
- value when accessed using 'get_value'.
- Because search techniques interact with Parameters through get_value, these
- parameters are searched on a different scale (e.g. log scale).
- """
-
- @abc.abstractmethod
- def _scale(self, v):
- """
- called on a value when getting it from it's configuration. Transforms the
- actual value to the scale it is searched on
- """
- return v
-
- @abc.abstractmethod
- def _unscale(self, v):
- """
- called on a value when storing it. Transforms a value from it's search scale
- to it's actual value
- """
- return v
-
- def set_value(self, config, value):
- NumericParameter.set_value(self, config, self._unscale(value))
-
- def get_value(self, config):
- return self._scale(NumericParameter.get_value(self, config))
-
- def legal_range(self, config):
- return map(self._scale, NumericParameter.legal_range(self, config))
-
-
-class LogIntegerParameter(ScaledNumericParameter, FloatParameter):
- """
- an integer value that is searched on a log scale, but stored without scaling
- """
-
- def _scale(self, v):
- return math.log(v + 1.0 - self.min_value, 2.0)
-
- def _unscale(self, v):
- v = 2.0 ** v - 1.0 + self.min_value
- v = int(round(v))
- return v
-
- def legal_range(self, config):
- low, high = NumericParameter.legal_range(self, config)
- # increase the bounds account for rounding
- return self._scale(low - 0.4999), self._scale(high + 0.4999)
-
-
-class LogFloatParameter(ScaledNumericParameter, FloatParameter):
- """
- a float parameter that is searched on a log scale, but stored without scaling
- """
-
- def _scale(self, v):
- return math.log(v + 1.0 - self.min_value, 2.0)
-
- def _unscale(self, v):
- v = 2.0 ** v - 1.0 + self.min_value
- return v
-
-
-class PowerOfTwoParameter(ScaledNumericParameter, IntegerParameter):
- """
- An integer power of two, with a min and max value. Searched by the exponent
- """
-
- def __init__(self, name, min_value, max_value, **kwargs):
- kwargs['value_type'] = int
- assert min_value >= 1
- assert math.log(min_value, 2) % 1 == 0 # must be power of 2
- assert math.log(max_value, 2) % 1 == 0 # must be power of 2
- super(PowerOfTwoParameter, self).__init__(name, min_value, max_value,
- **kwargs)
-
- def _scale(self, v):
- return int(math.log(v, 2))
-
- def _unscale(self, v):
- return 2 ** int(v)
-
- def legal_range(self, config):
- return int(math.log(self.min_value, 2)), int(math.log(self.max_value, 2))
-
- def search_space_size(self):
- return int(math.log(super(PowerOfTwoParameter, self).search_space_size(), 2))
-
-
-##################
-
-class ComplexParameter(Parameter):
- """
- A non-cartesian parameter that can't be manipulated directly, but has a set
- of user defined manipulation functions
- """
-
- def copy_value(self, src, dst):
- """copy the value of this parameter from src to dst config"""
- self._set(dst, copy.deepcopy(self._get(src)))
-
- def same_value(self, cfg1, cfg2):
- """test if cfg1 and cfg2 have the same value of this parameter"""
- return self._get(cfg1) == self._get(cfg2)
-
- def hash_value(self, config):
- """produce unique hash for this value in the config"""
- self.normalize(config)
- return hashlib.sha256(repr(self._get(config))).hexdigest()
-
- def get_value(self, config):
- return self._get(config)
-
- def set_value(self, config, value):
- self._set(config, value)
-
- def op4_set_linear(self, cfg, cfg_a, cfg_b, cfg_c, a, b, c):
- """
- set this value to :math:`a*cfg_a + b*cfg_b + c*cfg_c`
-
- this operation is not possible in general with complex parameters but
- we make an attempt to "fake" it for common use cases
-
- basically a call to randomize unless after normalization,
- a = 1.0, b == -c, and cfg_b == cfg_c, in which case nothing is done
-
- :param cfg: the configuration to be changed
- :param cfg_a: a parent configuration
- :param cfg_b: a parent configuration
- :param cfg_c: a parent configuration
- :param a: weight for cfg_a
- :param b: weight for cfg_b
- :param c: weight for cfg_c
- """
- # attempt to normalize order, we prefer a==1.0
- if a != 1.0 and b == 1.0: # swap a and b
- a, cfg_a, b, cfg_b = b, cfg_b, a, cfg_a
- if a != 1.0 and c == 1.0: # swap a and c
- a, cfg_a, c, cfg_c = c, cfg_c, a, cfg_a
-
- # attempt to normalize order, we prefer b==-c
- if b < c: # swap b and c
- b, cfg_b, c, cfg_c = c, cfg_c, b, cfg_b
- if b != -c and a == -c: # swap a and c
- a, cfg_a, c, cfg_c = c, cfg_c, a, cfg_a
-
- if a == 1.0 and b == -c:
- self.copy_value(cfg_a, cfg)
- self.add_difference(cfg, b, cfg_b, cfg_c) # TODO inline this logic?
- else:
- # TODO: should handle more cases
- self.op1_randomize(cfg)
-
- def add_difference(self, cfg_dst, scale, cfg_b, cfg_c):
- """
- add the difference cfg_b-cfg_c to cfg_dst
-
- this is the key operation used in differential evolution
- and some simplex techniques
-
- this operation is not possible in general with complex parameters but
- we make an attempt to "fake" it
- """
- if not self.same_value(cfg_b, cfg_c):
- self.op1_randomize(cfg_dst)
-
- @abc.abstractmethod
- def op1_randomize(self, config):
- """
- randomize this value without taking into account the current position
- :param config: the configuration to be changed
- """
- pass
-
- @abc.abstractmethod
- def seed_value(self):
- """some legal value of this parameter (for creating initial configs)"""
- return
-
-
-class BooleanParameter(ComplexParameter):
- def manipulators(self, config):
- return [self.op1_flip]
-
- def get_value(self, config):
- return self._get(config)
-
- def set_value(self, config, value):
- self._set(config, value)
-
- def op1_randomize(self, config):
- """
- Set this parameter's value in a configuration randomly
-
- :param config: the configuration to be changed
- """
- self._set(config, self.seed_value())
-
- def seed_value(self):
- return random.choice((True, False))
-
- def op1_flip(self, config):
- """
- Flip this parameter's value in a configuration
-
- :param config: the configuration to be changed
- """
- self._set(config, not self._get(config))
-
- def search_space_size(self):
- return 2
-
- def op3_swarm(self, cfg, cfg1, cfg2, c=1, c1=0.5,
- c2=0.5, velocity=0, *args, **kwargs):
- """
- Simulates a single update step in particle swarm optimization by updating
- the current position and returning a new velocity.
-
- The new velocity is given by
-
- .. math:: c*velocity + r1*c1*(cfg1-cfg) + r2*c2*(cfg2-cfg)
-
- where r1 and r2 are random values between 0 and 1
-
- The new current position is randomly chosen based on the new velocity
-
- :param cfg: the configuration to be changed. Represents the current position
- :param cfg1: a configuration to shift towards. Should be the local best position
- :param cfg2: a configuration to shift towards. Should be the global best position
- :param c: the weight of the current velocity
- :param c1: weight of cfg1
- :param c2: weight of cfg2
- :param velocity: the old velocity
- :param args:
- :param kwargs:
- :return: the new velocity, a float
-
- """
- v = velocity * c + (self.get_value(cfg1) - self.get_value(
- cfg)) * c1 * random.random() + (self.get_value(
- cfg2) - self.get_value(cfg)) * c2 * random.random()
- # Map velocity to continuous space with sigmoid
- s = 1 / (1 + numpy.exp(-v))
- # Decide position randomly
- p = (s - random.random()) > 0
- self.set_value(cfg, p)
- return v
-
-
-class SwitchParameter(ComplexParameter):
- """
- A parameter representing an unordered collection of options with no implied
- correlation between the choices. The choices are range(option_count)
- """
-
- def __init__(self, name, option_count):
- self.option_count = option_count
- super(SwitchParameter, self).__init__(name)
-
- def op1_randomize(self, config):
- """
- Set this parameter's value in a configuration to a random value
-
- :param config: the configuration to be changed
- """
- self._set(config, random.randrange(self.option_count))
-
- def seed_value(self):
- return random.randrange(self.option_count)
-
- def search_space_size(self):
- return max(1, self.option_count)
-
-
-class EnumParameter(ComplexParameter):
- """
- same as a SwitchParameter but choices are taken from an arbitrarily typed list
- """
-
- def __init__(self, name, options):
- super(EnumParameter, self).__init__(name)
- self.options = list(options)
-
- def op1_randomize(self, config):
- """
- Set this parameter's value in a configuration to a random value
-
- :param config: the configuration to be changed
- """
- self._set(config, random.choice(self.options))
-
- def seed_value(self):
- return random.choice(self.options)
-
- def search_space_size(self):
- return max(1, len(self.options))
-
-
-class PermutationParameter(ComplexParameter):
- """
- A parameter representing a permutation (or ordering) as a list of items
- """
- def __init__(self, name, items):
- super(PermutationParameter, self).__init__(name)
- self._items = list(items)
- self.size = len(items)
-
- def op1_randomize(self, config):
- """
- Set this parameter's value in a configuration to a random value
-
- :param config: the configuration to be changed
- """
- random.shuffle(self._get(config))
- self.normalize(config)
-
- def op1_small_random_change(self, config, p=0.25):
- """
- Iterates through the list and probabilistically swaps each element with the
- next element
-
- :param p: probability of swapping an element with the next element
- :param config: the configuration to be changed
- """
- cfg_item = self._get(config)
- for i in xrange(1, len(cfg_item)):
- if random.random() < p:
- # swap
- cfg_item[i - 1], cfg_item[i] = cfg_item[i], cfg_item[i - 1]
- self.normalize(config)
-
- def seed_value(self):
- return list(self._items) # copy
-
- def manipulators(self, config):
- return [self.op1_randomize, self.op1_small_random_change]
-
- def get_value(self, config):
- return self._get(config)
-
- def set_value(self, config, value):
- self._set(config, value)
-
- def search_space_size(self):
- return math.factorial(max(1, len(self._items)))
-
- def op3_cross(self, cfg, cfg1, cfg2, xchoice='op3_cross_OX1', strength=0.3,
- *args, **kwargs):
- """
- Calls the crossover operator specified by xchoice
- Passes argument d = strength*(size of the permutation)
-
- :param cfg: the configuration to be changed
- :param cfg1: a parent configuration
- :param cfg2: a parent configuration
- :param xchoice: string specifying which crossover operator to use (should start with op3_cross prefix)
- :param strength: the strength of the crossover
- """
- dd = int(round(self.size * strength))
- if dd < 1:
- log.warning('Crossover length too small. Cannot create new solution.')
- if dd >= self.size:
- log.warning('Crossover length too big. Cannot create new solution.')
- getattr(self, xchoice)(cfg, cfg1, cfg2, d=dd, *args, **kwargs)
-
- def op3_swarm(self, cfg, cfg1, cfg2, xchoice='op3_cross_OX1', c=0.5,
- c1=0.5, c2=0.5, strength=0.3, velocity=0, *args, **kwargs):
- """
- Replacement for particle swarm optimization iterative step for permutations.
- Given a target cfg and 2 parent cfgs, probabilistically performs an
- op3_cross with one of the 2 parents.
-
- :param cfg: the configuration to be changed. Represents the current position
- :param cfg1: a configuration to shift towards. Should be the local best
- position
- :param cfg2: a configuration to shift towards. Should be the global best
- position
- :param xchoice: which crossover operator should be used
- :param c: the probability of not performing a crossover
- :param c1: the probability of performing a crossover with cfg1 (if a
- crossover is performed)
- :param c2: unused
- :param strength: the strength of the crossover
- :param velocity: the old velocity - unused
- """
- if random.uniform(0, 1) > c:
- if random.uniform(0, 1) < c1:
- # Select crossover operator
- self.op3_cross(cfg, cfg, cfg1, xchoice, strength)
- else:
- self.op3_cross(cfg, cfg, cfg2, xchoice, strength)
-
- # swap-based operators
- def op2_random_swap(self, cfg, cfg1, *args, **kwargs):
- """
- Swap a random pair of items in cfg1 and save the result into cfg
-
- :param cfg: the configuration to be changed
- :param cfg1: the configuration whose PermutationParameter's elements are
- swapped and copied into cfg
- """
- p = self.get_value(cfg1)[:]
- r = random.randint(0, len(p) - 1)
- s = random.randint(0, len(p) - 1)
- v1 = p[r]
- v2 = p[s]
- p[r] = v2
- p[s] = v1
- self.set_value(cfg, p)
-
- def op2_random_invert(self, cfg, cfg1, strength=0.3, *args, **kwargs):
- """
- Reverse the ordering of a random subsection of size d in cfg1 and save the
- result in cfg where d = strength*total-size
-
- :param cfg: the configuration to be changed
- :param cfg1: the configuration whose PermutationParameter is inverted
- :param strength: the size of the reversed subsection as a fraction of the
- total size
- """
- p = self.get_value(cfg1)[:]
- d = int(round(len(p) * strength))
- r = random.randint(0, len(p) - d)
- subpath = p[r:r + d][:]
- subpath.reverse()
- p[r:r + d] = subpath
- self.set_value(cfg, p)
-
- # Crossover operators
- def op3_cross_PX(self, cfg, cfg1, cfg2, d=0):
- """
- Partition crossover (Whitley 2009?)
-
- Chooses a random cut point and reorders elements in cfg1 up to the cut point
- according to their order in cfg2.
-
- Saves the result in cfg
-
- :param cfg: the configuration to be changed
- :param cfg1: the first parent configuration. The "base" configuration
- :param cfg2: the second parent configuration. Is "crossed into" cfg1
- :param d: unused
- """
- p1 = self.get_value(cfg1)
- p2 = self.get_value(cfg2)
- c1 = random.randint(2, len(p1))
- self.set_value(cfg, sorted(p1[:c1], key=lambda x: p2.index(x)) + p1[c1:])
-
- def op3_cross_PMX(self, cfg, cfg1, cfg2, d=0):
- """
- Partially-mapped crossover Goldberg & Lingle (1985)
-
- Replaces a random section of cfg1 with the corresponding section in cfg2.
- Displaced elements in cfg1 are moved to the old position of the elements
- displacing them
-
- :param cfg: the configuration to be changed
- :param cfg1: the first parent configuration. The "base" configuration
- :param cfg2: the second parent configuration. Is "crossed into" cfg1
- :param d: the size of the crossover
- """
- if d == 0:
- d = max(1, int(round(self.size * 0.3))) # default to 1/3 of permutation size
- p1 = self.get_value(cfg1)[:]
- p2 = self.get_value(cfg2)[:]
-
- r = random.randint(0, len(p1) - d)
-
- c1 = p1[r:r + d]
- c2 = p2[r:r + d]
-
- # get new permutation by crossing over a section of p2 onto p1
- pnew = self.get_value(cfg1)[:]
- pnew[r:r + d] = c2
- # fix conflicts by taking displaced elements in crossed over section
- # displaced = (elements x in c1 where x does not have corresponding value in c2)
- # and putting them where the value that displaced them was
-
- #candidates for displacement
- candidate_indices = set(range(r) + range(r+d, len(p1)))
- # Check through displaced elements to find values to swap conflicts to
- while c1 != []:
- n = c1[0]
- #try to match up a value in c1 to the equivalent value in c2
- while c2[0] in c1:
- if n == c2[0]:
- # already match up
- break
- # find position idx of c2[0] in c1
- link_idx = c1.index(c2[0])
- # get value of c2 at idx
- link = c2[link_idx]
- # remove c2[idx] and c1[idx] since they match up when we swap c2[0] with c2[idx] (this avoids an infinite loop)
- del c2[link_idx]
- del c1[link_idx]
- # swap new value into c2[0]
- c2[0] = link
-
- if n != c2[0]:
- # first check if we can swap in the crossed over section still
- if n in c2:
- c2[c2.index(n)] = c2[0]
- else:
- # assign first instance of c2[0] outside of the crossed over section in pnew to c1[0]
- for idx in candidate_indices:
- if pnew[idx] == c2[0]:
- pnew[idx] = c1[0]
- candidate_indices.remove(idx) # make sure we don't override this value now
- break
- # remove first elements
- del c1[0]
- del c2[0]
- self.set_value(cfg, pnew)
-
- def op3_cross_CX(self, cfg, cfg1, cfg2, d=0):
- """
- Implementation of a cyclic crossover.
-
- Repeatedly replaces elements of cfg1 with the element at the same index in
- cfg2. This is done until a cycle is reached and cfg1 is valid again. The
- initial replacement is random.
-
- Saves the result in cfg.
-
- :param cfg: the configuration to be changed
- :param cfg1: the first parent configuration. The "base" configuration
- :param cfg2: the second parent configuration. Is "crossed into" cfg1
- :param d: unused
- """
- p1 = self.get_value(cfg1)
- p2 = self.get_value(cfg2)
- p = p1[:]
-
- s = random.randint(0, len(p1) - 1)
- i = s
- indices = set()
-
- while len(indices) < len(p1): # should never exceed this
- indices.add(i)
- val = p1[i]
- i = p2.index(val)
- # deal with duplicate values
- while i in indices:
- if i == s:
- break
- i = p2[i+1:].index(val) + i + 1
- if i == s:
- break
-
- for j in indices:
- p[j] = p2[j]
-
- self.set_value(cfg, p)
-
- def op3_cross_OX1(self, cfg, cfg1, cfg2, d=0):
- """
- Ordered Crossover (Davis 1985)
-
- Exchanges a subpath from cfg2 into cfg1 while maintaining the order of the
- remaining elements in cfg1.
-
- Saves the result in cfg.
-
- :param cfg: the configuration to be changed
- :param cfg1: the first parent configuration. The "base" configuration
- :param cfg2: the second parent configuration. Is "crossed into" cfg1
- :param d: size of the exchanged subpath
- """
- if d == 0:
- d = max(1, int(round(self.size * 0.3))) # default to 1/3 of permutation size
- p1 = self.get_value(cfg1)
- p2 = self.get_value(cfg2)
- c1 = p1[:]
- c2 = p2[:]
- # Randomly find cut points
- r = random.randint(0, len(
- p1) - d) # Todo: treat path as circle i.e. allow cross-boundary cuts
- [c1.remove(i) for i in p2[r:int(r + d)]]
- self.set_value(cfg, c1[:r] + p2[r:r + d] + c1[r:])
-
- def op3_cross_OX3(self, cfg, cfg1, cfg2, d=0):
- """
- Ordered crossover variation 3 (Deep 2010)
-
- Same as op3_cross_OX1, except the parents have different cut points for
- their subpaths
-
- :param cfg: the configuration to be changed
- :param cfg1: the first parent configuration. The "base" configuration
- :param cfg2: the second parent configuration. Is "crossed into" cfg1
- :param d: size of the exchanged subpath
- """
- if d == 0:
- d = max(1, int(round(self.size * 0.3))) # default to 1/3 of permutation size
- p1 = self.get_value(cfg1)
- p2 = self.get_value(cfg2)
- c1 = p1[:]
- c2 = p2[:]
- # Randomly find cut points
- # Todo: treat path as circle i.e. allow cross-boundary cuts
- r1 = random.randint(0, len(p1) - d)
- r2 = random.randint(0, len(p1) - d)
- [c1.remove(i) for i in p2[r2:r2 + d]]
- self.set_value(cfg, c1[:r1] + p2[r2:r2 + d] + c1[r1:])
-
- def search_space_size(self):
- return math.factorial(max(1, len(self._items)))
-
-
-class ScheduleParameter(PermutationParameter):
- def __init__(self, name, items, deps):
- super(ScheduleParameter, self).__init__(name, items)
- self.deps = dict((k, set(v)) for k, v in deps.items())
- log.debug("ScheduleParameter(%s, %s, %s)", repr(name), repr(items),
- repr(deps))
- self._expand_deps()
-
- def _expand_deps(self):
- """expand self.deps to include recursive dependencies"""
- fixed_point = False
- while not fixed_point:
- fixed_point = True
- for k in self.deps.keys():
- oldlen = len(self.deps[k])
- for dep in list(self.deps[k]):
- if dep in self.deps:
- self.deps[k].update(self.deps[dep])
- if oldlen != len(self.deps[k]):
- fixed_point = False
-
- # verify schedule is valid
- items = set(self._items)
- for k, v in self.deps.items():
- if k in v:
- raise Exception("ScheduleParameter('%s') cycle: %s depends on itself" %
- (self.name, k))
-
- if v - items:
- raise Exception("ScheduleParameter('%s'): %s is unknown" %
- (self.name, v - items))
-
- if set(self.deps.keys()) - items:
- raise Exception("ScheduleParameter('%s'): %s is unknown" %
- (self.name, set(self.deps.keys()) - items))
-
- def is_topologically_sorted(self, values):
- used = set()
- for v in values:
- if v in self.deps and self.deps[v].union(used):
- return False
- used.add(v)
- return True
-
- def topologically_sorted_depth_first(self, values):
- """faster but not stable enough"""
- if self.is_topologically_sorted(values):
- return values
- sorted_values = []
- used = set()
- deps = dict((k, sorted(v, key=values.index, reverse=True))
- for k, v in self.deps.items())
-
- def visit(v):
- if v in used:
- return
- if v in deps:
- for dv in deps[v]:
- visit(dv)
- used.add(v)
- sorted_values.append(v)
-
- for v in reversed(values):
- visit(v)
- return list(reversed(sorted_values))
-
- def topologically_sorted(self, values):
- if self.is_topologically_sorted(values):
- return values
- deps = copy.deepcopy(self.deps)
- queue = collections.deque(reversed(values))
- sorted_values = []
- while queue:
- v = queue.popleft()
- if v in deps and deps[v]:
- queue.append(v)
- else:
- for k, d in deps.items():
- d.discard(v)
- if not d:
- del deps[k]
- sorted_values.append(v)
-
- return list(reversed(sorted_values))
-
- def normalize(self, cfg):
- self._set(cfg, self.topologically_sorted(self._get(cfg)))
-
-
-class SelectorParameter(ComplexParameter):
- def __init__(self, name, choices, max_cutoff,
- order_class=PermutationParameter,
- offset_class=LogIntegerParameter):
- super(SelectorParameter, self).__init__(name)
- self.choices = choices
- self.max_cutoff = max_cutoff
- self.order_param = order_class('{0}/order'.format(name), choices)
- self.offset_params = [
- offset_class('{0}/offsets/{1}'.format(name, i), 0, max_cutoff)
- for i in xrange(len(choices) - 1)]
-
- def sub_parameters(self):
- return [self.order_param] + self.offset_params
-
- def seed_value(self):
- return {'order': self.order_param.seed_value(),
- 'offsets': [co.seed_value() for co in self.offset_params]}
-
- def op1_randomize(self, config):
- random.choice(self.sub_parameters()).op1_randomize(config)
-
- def selector_iter(self, config):
- """
- yield (cutoff, choice) pairs
- cutoff will be None on the first value
- """
- order = config[self.name]['order']
- yield (None, order[0])
- cutoff = 0
- for n, offset in enumerate(config[self.name]['offsets']):
- if offset > 0:
- cutoff += offset
- yield cutoff, order[n + 1]
-
-
-class ParameterArray(ComplexParameter):
- """
- Represents an array of Parameters
- """
- def __init__(self, name, count, element_type, *args, **kwargs):
- super(ParameterArray, self).__init__(name)
- self.count = count
-
- self.sub_params = [
- element_type('{0}/{1}'.format(name, i), *args[i], **kwargs[i])
- for i in xrange(count)]
-
- def sub_parameters(self):
- return self.sub_params
-
- def seed_value(self):
- return [p.seed_value() for p in self.sub_params]
-
- def op1_randomize(self, config):
- """
- randomly selects a sub-parameter and randomizes it
-
- :param config: the configuration to be changed
- """
- random.choice(self.sub_parameters()).op1_randomize(config)
-
-
-class BooleanParameterArray(ParameterArray):
- """
- Represents an array of BooleanParameters - currently unimplimented
- """
- def __init__(self, name, count):
- super(BooleanParameterArray, self).__init__(name, count, BooleanParameter)
-
- def op3_swarm(self, cfg, cfg1, cfg2, *args, **kwargs):
- # TODO
- pass
-
- def op3_cross(self, cfg, cfg1, cfg2, *args, **kwargs):
- # TODO
- pass
-
-
-class IntegerParameterArray(ParameterArray):
- """
- Represents an array of IntegerParameters - currently unimplemented
- """
- def __init__(self, name, min_values, max_values):
- assert len(min_values) == len(max_values)
- super(IntegerParameterArray, self).__init__(name, len(min_values),
- IntegerParameter,
- min_value=min_values,
- max_value=max_values)
-
- def op3_swarm(self, cfg, cfg1, cfg2, *args, **kwargs):
- # TODO
- pass
-
- def op3_cross(self, cfg, cfg1, cfg2, *args, **kwargs):
- # TODO
- pass
-
-
-class Array(ComplexParameter):
- """
- An interface for parameters representing an array of values.
- """
- # TODO: constraints? (upper & lower bound etc)
- def __init__(self, name, size):
- super(Array, self).__init__(name)
- self.size = size
-
- def op3_cross(self, cfg, cfg1, cfg2, strength=0.3, *args, **kwargs):
- """
- Crosses two arrays by replacing a random subsection of cfg1 with the
- corresponding subsection of cfg2.The size of the chunk is a fixed fraction
- of the total length, given by the strength
-
- Behaves like a specialized 2-point crossover, where the first cut point is
- random and the second cut is a set distance after.
-
- :param cfg: the configuration to be changed
- :param cfg1: the configuration being inserted into
- :param cfg2: the configuration being inserted
- :param strength: the size of the crossover, as a fraction of total array
- length
- """
- d = int(round(self.size * strength))
- if d < 1:
- log.debug('Crossover length too small. Cannot create new solution.')
- if d >= self.size:
- log.debug('Crossover length too big. Cannot create new solution.')
- p1 = self.get_value(cfg1)
- p2 = self.get_value(cfg2)
- r = random.randint(0, len(
- p1) - d) # Todo: treat path as circle i.e. allow cross-boundary cuts
- p = numpy.concatenate([p1[:r], p2[r:r + d], p1[r + d:]])
- self.set_value(cfg, p)
-
- def op3_swarm(self, cfg, cfg1, cfg2, c=1, c1=0.5,
- c2=0.5, velocity=0, strength=0.3, *args, **kwargs):
- """
- Replacement for a particle swarm optimization iterative step for arrays.
- Given a target cfg and 2 parent cfgs, probabilistically performs an
- :py:meth:`op3_cross` with one of the 2 parents.
-
- :param cfg: the configuration to be changed. Represents the cfg position
- :param cfg1: a configuration to shift towards. Should be the local best
- position
- :param cfg2: a configuration to shift towards. Should be the global best
- position
- :param c: the probability of not performing a crossover
- :param c1: the probability of performing a crossover with cfg1 (if a
- crossover is performed)
- :param c2: unused
- :param velocity: the old velocity - unused
- :param strength: the strength of the crossover
- """
- if random.uniform(0, 1) > c:
- if random.uniform(0, 1) < c1:
- # Select crossover operator
- self.op3_cross(cfg, cfg, cfg1, strength)
- else:
- self.op3_cross(cfg, cfg, cfg2, strength)
-
- def get_value(self, config):
- return self._get(config)
-
- def set_value(self, config, value):
- self._set(config, value)
-
-
-class BooleanArray(Array):
- """
- Represents an array of boolean values which are either 0 or 1
- """
- def op3_swarm_parallel(self, cfg, cfg1, cfg2, c=1,
- c1=0.5, c2=0.5, velocities=0):
- """
- Simulates a single particle swarm optimization step for each element in the
- array by updating each position and returning an array of new velocities.
-
- The new velocities are given by
-
- .. math:: c*velocity + r1*c1*(cfg1-cfg) + r2*c2*(cfg2-cfg)
-
- where r1 and r2 are random values between 0 and 1. In each iteration, r1 and
- r2 are constant across array elements
-
- The new cfg positions are randomly chosen based on the new velocities
-
- :param cfg: the configuration to be changed. This represents the current
- position
- :param cfg1: a configuration to shift towards. Should be the local best
- position
- :param cfg2: a configuration to shift towards. Should be the global best
- position
- :param c: the weight of the current velocities
- :param c1: weight of cfg1
- :param c2: weight of cfg2
- :param velocities: the current velocities
- :return: a numpy array of new velocities
- """
- vs = velocities * c + (self.get_value(cfg1) - self.get_value(
- cfg)) * c1 * random.random() + (self.get_value(
- cfg2) - self.get_value(cfg)) * c2 * random.random()
- # Map velocity to continuous space with sigmoid
- ss = 1 / (1 + numpy.exp(-vs))
- # Decide position randomly
- ps = (ss - numpy.random.rand(1, self.size)) > 0
- self.set_value(cfg, ps)
- return vs
-
- def op1_randomize(self, config):
- """
- Set this parameter's value in a configuration randomly
-
- :param config: the configuration to be changed
- """
- value = numpy.random.rand(1, self.size) > 0.5
- self._set(config, value)
-
- def seed_value(self):
- return numpy.random.rand(1, self.size) > 0.5
-
-
-class FloatArray(Array):
- """
- Represents an array of float values
- """
- def __init__(self, name, size, fmax, fmin):
- super(FloatArray, self).__init__(name, size)
- self.fmax = fmax
- self.fmin = fmin
-
- def op1_randomize(self, config):
- """
- Set this parameter's value in a configuration randomly
-
- :param config: the configuration to be changed
- """
- value = numpy.random.rand(1, self.size) * (
- self.fmax - self.fmin) + self.fmin
- self._set(config, value)
-
- def seed_value(self):
- value = numpy.random.rand(1, self.size) * (
- self.fmax - self.fmin) + self.fmin
- return value
-
- def op3_swarm_parallel(self, cfg, cfg1, cfg2, c=1,
- c1=0.5, c2=0.5, velocities=0):
- """
- Simulates a single particle swarm optimization step for each element in the
- array by updating the each position and returning an array of new velocities
-
- The new velocity is given by
-
- .. math:: c*velocity + r1*c1*(cfg1-cfg) + r2*c2*(cfg2-cfg)
-
- where r1 and r2 are random values between 0 and 1. In each iteration, r1 and
- r2 are constant across array elements
-
- The new cfg positions are randomly chosen based on the new velocities
-
- :param cfg: the configuration to be changed. This represents the current
- position
- :param cfg1: a configuration to shift towards. Should be the local best
- position
- :param cfg2: a configuration to shift towards. Should be the global best
- position
- :param c: the weight of the cfg velocities
- :param c1: weight of cfg1
- :param c2: weight of cfg2
- :param velocities: the cfg velocities
- :return: a numpy array of new velocities
- """
- vs = velocities * c + (self.get_value(cfg1) - self.get_value(
- cfg)) * c1 * random.random() + (self.get_value(
- cfg2) - self.get_value(cfg)) * c2 * random.random()
- p = self.get_value(cfg) + vs
- p[p > self.fmax] = self.fmax
- p[p < self.fmin] = self.fmin
- self.set_value(cfg, p)
- return vs
-
-
-##################
-
-class ManipulatorProxy(object):
- """
- wrapper around configuration manipulator and config pair
- """
-
- def __init__(self, manipulator, cfg):
- self.cfg = cfg
- self.manipulator = manipulator
- self.params = manipulator.parameters_dict(self.cfg)
-
- def keys(self):
- return self.params.keys()
-
- def __getitem__(self, k):
- return ParameterProxy(self.params[k], self.cfg)
-
-
-class ParameterProxy(object):
- """
- wrapper aint parameter and config pair, adds config
- as first argument to all method calls to parameter
- """
-
- def __init__(self, param, cfg):
- self.cfg = cfg
- self.param = param
-
- def __getattr__(self, key):
- """equivalent of self.param.key(self.cfg, ...)"""
- member = getattr(self.param, key)
-
- def param_method_proxy(*args, **kwargs):
- return member(self.cfg, *args, **kwargs)
-
- if callable(member):
- return param_method_proxy
- else:
- # we should only hit this for key == 'name'
- return member
-
-
-# Inspection Methods
-def operators(param, num_parents):
- """
- Return a list of operators for the given parameter that take the specified
- number of input configurations
-
- :param param: a Parameter class
- :param num_parents: a String specifying number of inputs required by the operator.
- should be one of '1', '2', '3', '4', or 'n'
- """
- ops = []
- methods = inspect.getmembers(param, inspect.ismethod)
- for m in methods:
- name, obj = m
- if is_operator(name, num_parents):
- ops.append(name)
- return ops
-
-def composable_operators(param, min_num_parents):
- """
- Return a list of operators for the given parameter that can be programatically composed
- with a composable technique generating min_num_parents.
-
- Programatically composable operators have no non-cfg arguments
-
- :param param: a Parameter class
- :param min_num_parents: the minimum number of parents passed to the operator
- """
- if min_num_parents < 1:
- return []
-
- allowed_num_parents = ['n']
- for i in range(1,5):
- if i > min_num_parents:
- break
- allowed_num_parents.append(str(i))
-
- ops = []
- methods = inspect.getmembers(param, inspect.ismethod)
- for m in methods:
- name, obj = m
- argspec = inspect.getargspec(obj)
- numargs = len(argspec.args) - (len(argspec.defaults) if argspec.defaults else 0)
- for num_parents in allowed_num_parents:
- if is_operator(name, num_parents):
- if num_parents == 'n':
- if numargs == 3: # self, cfg, cfgs
- ops.append(name)
- else:
- if numargs == (1 + int(num_parents)):
- ops.append(name)
- break
- return ops
-
-
-def is_operator(name, num_parents):
- """
- Tells whether a method is an operator taking in the specified number of inputs
- from the method name
-
- :param name: the method name
- :param num_parents: a String specifying number of inputs required by the operator.
- should be one of '1', '2', '3', '4', or 'n'
- """
- return ('op' + num_parents + '_') == name[:4]
-
-def all_operators():
- """
- Return a dictionary mapping from parameter names to lists of operator function
- names
- """
- ops = {}
- for p in all_params():
- name, obj = p
- all_ops = []
- for num in ['1', '2', '3', '4', 'n']:
- all_ops += operators(obj, num)
- ops[name] = all_ops
- return ops
-
-def all_params():
- params = inspect.getmembers(sys.modules[__name__], lambda x: inspect.isclass(
- x) and x.__module__ == __name__ and issubclass(x, Parameter))
- return params
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/metatechniques.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/metatechniques.py
deleted file mode 100644
index 2e33e7961a..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/metatechniques.py
+++ /dev/null
@@ -1,186 +0,0 @@
-import abc
-import logging
-from collections import deque, defaultdict
-from fn import _
-
-from .technique import SearchTechniqueBase
-
-log = logging.getLogger(__name__)
-
-class MetaSearchTechnique(SearchTechniqueBase):
- """
- a technique made up of a collection of other techniques
- """
- def __init__(self, techniques, log_freq = 500, *pargs, **kwargs):
- super(MetaSearchTechnique, self).__init__(*pargs, **kwargs)
- self.techniques = techniques
- self.request_count = 0
- self.log_freq = log_freq
- self.logging_use_counters = defaultdict(int)
- self.unique_names()
-
- def unique_names(self):
- names = set()
- for t in self.techniques:
- while t.name in names:
- t.name += '~'
- t.name = intern(t.name)
- names.add(t.name)
-
- def set_driver(self, driver):
- super(MetaSearchTechnique, self).set_driver(driver)
- for t in self.techniques:
- t.set_driver(driver)
- self.driver = driver
-
- def desired_result(self):
- techniques = self.select_technique_order()
- for technique in techniques:
- dr = technique.desired_result()
- if dr is not None:
- if dr is False:
- # technique is waiting for results
- continue
- self.driver.register_result_callback(dr,
- lambda result: self.on_technique_result(technique, result))
- if self.log_freq:
- self.logging_use_counters[technique.name] += 1
- self.debug_log()
- self.request_count += 1
- return dr
- else:
- self.on_technique_no_desired_result(technique)
- return None
-
- def on_technique_no_desired_result(self, technique):
- """called if a sub-technique returns None"""
- pass
-
- def on_technique_result(self, technique, result):
- """callback for results of sub-techniques"""
- pass
-
- @abc.abstractmethod
- def select_technique_order(self):
- """select the order of next techniques to try"""
- return []
-
- def debug_log(self):
- if self.log_freq and sum(self.logging_use_counters.values())>self.log_freq:
- log.info("%s: %s", self.name,
- str(sorted(self.logging_use_counters.items(), key = _[1]*-1)))
- self.logging_use_counters = defaultdict(int)
-
-class RoundRobinMetaSearchTechnique(MetaSearchTechnique):
- """evenly switch between all source techniques"""
- def __init__(self, techniques, **kwargs):
- techniques = deque(techniques)
- super(RoundRobinMetaSearchTechnique, self).__init__(techniques, **kwargs)
-
- def select_technique_order(self):
- rv = list(self.techniques)
- self.techniques.rotate(1)
- return rv
-
-class RecyclingMetaTechnique(MetaSearchTechnique):
- """
- periodically restart techniques that are not performing well compared to
- global best
- """
- def __init__(self,
- techniques_generators,
- window = 100,
- factor = 5.0,
- **kwargs):
- if 'log_freq' not in kwargs:
- kwargs['log_freq'] = None
- techniques = deque((g(seed_cfg = None) for g in techniques_generators))
- self.rename_i = 0
- for t in techniques:
- self.rename_technique(t)
- super(RecyclingMetaTechnique, self).__init__(techniques, **kwargs)
- self.best_results = defaultdict(lambda: None)
- self.factor = factor
- self.last_check = 0
- self.old_best_results = defaultdict(lambda: None)
- self.technique_generators = deque(techniques_generators)
- self.window = window
-
- def rename_technique(self, technique):
- technique.name += ".R%d" % self.rename_i
- self.rename_i += 1
-
- def on_technique_result(self, technique, result):
- """callback for results of sub-techniques"""
- if (self.best_results[technique] is None or
- self.driver.objective.lt(result, self.best_results[technique])):
- self.best_results[technique] = result
-
- def technique_cmp(self, a, b):
- # a1 = self.old_best_results[a]
- # a2 = self.best_results[a]
- # b1 = self.old_best_results[b]
- # b2 = self.best_results[b]
- # if a1 is None and b1 is None:
- # return 0
- # if a1 is None:
- # return -1
- # if b1 is None:
- # return 1
- # return self.driver.objective.project_compare(a1, a2, b1, b2, self.factor)
-
- # not ready techniques go to the back
- if not a.is_ready() or not b.is_ready():
- return cmp(b.is_ready(), a.is_ready())
-
- a = self.best_results[a]
- b = self.best_results[b]
- if a is None and b is None:
- return 0
- if a is None:
- return -1
- if b is None:
- return 1
- return self.driver.objective.compare(a, b)
-
- def recycle_techniques(self):
- techniques = list(self.techniques)
- techniques.sort(cmp=self.technique_cmp)
- worst = techniques[-1]
-
- if (not worst.is_ready()
- or (self.old_best_results[worst] is not None
- and self.driver.objective.lt(self.driver.best_result,
- self.best_results[worst]))):
- techniques_new = deque()
- tn = None
- for t, gen in zip(self.techniques, self.technique_generators):
- if t is worst:
- tn = gen(seed_cfg=self.driver.best_result.configuration.data)
- self.rename_technique(tn)
- tn.set_driver(self.driver)
- log.info("%s replacing %s with %s", self.name, t.name, tn.name)
- techniques_new.append(tn)
- else:
- techniques_new.append(t)
- self.techniques = techniques_new
- else:
- log.debug("%s: not replacing techniques", self.name)
-
- self.old_best_results = self.best_results
- self.best_results = defaultdict(lambda: None)
- for t in self.techniques:
- self.best_results[t] = self.old_best_results[t]
-
- def select_technique_order(self):
- """
- round robin between techniques
- """
- if self.last_check + self.window < self.request_count:
- self.last_check = self.request_count
- self.recycle_techniques()
- rv = list(self.techniques)
- self.techniques.rotate(1)
- self.technique_generators.rotate(1)
- return rv
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/objective.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/objective.py
deleted file mode 100644
index b46a2f54b2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/objective.py
+++ /dev/null
@@ -1,338 +0,0 @@
-import abc
-import logging
-
-from fn import _
-
-import opentuner
-from opentuner.resultsdb.models import *
-
-log = logging.getLogger(__name__)
-
-
-class SearchObjective(object):
- """
- delegates the comparison of results and configurations
- """
- __metaclass__ = abc.ABCMeta
-
- @abc.abstractmethod
- def result_order_by_terms(self):
- """return database columns required to order by the objective"""
- return []
-
- @abc.abstractmethod
- def result_compare(self, result1, result2):
- """cmp() compatible comparison of resultsdb.models.Result"""
- return
-
- def config_compare(self, config1, config2):
- """cmp() compatible comparison of resultsdb.models.Configuration"""
- return self.result_compare(self.driver.results_query(config=config1).one(),
- self.driver.results_query(config=config2).one())
-
- @abc.abstractmethod
- def result_relative(self, result1, result2):
- """return None, or a relative goodness of resultsdb.models.Result"""
- return
-
- def config_relative(self, config1, config2):
- """return None, or a relative goodness of resultsdb.models.Configuration"""
- return self.result_relative(self.driver.results_query(config=config1).one(),
- self.driver.results_query(config=config2).one())
-
-
- def __init__(self):
- self.driver = None
-
- def set_driver(self, driver):
- self.driver = driver
-
- def result_order_by(self, q):
- return q.order_by(*self.result_order_by_terms())
-
- def compare(self, a, b):
- """cmp() compatible compare"""
- if isinstance(a, Configuration):
- return self.config_compare(a, b)
- if isinstance(a, Result):
- return self.result_compare(a, b)
- assert False
-
- def relative(self, a, b):
- if isinstance(a, Configuration):
- return self.config_relative(a, b)
- if isinstance(a, Result):
- return self.result_relative(a, b)
- assert None
-
- def lt(self, a, b):
- return self.compare(a, b) < 0
-
- def lte(self, a, b):
- return self.compare(a, b) <= 0
-
- def gt(self, a, b):
- return self.compare(a, b) > 0
-
- def gte(self, a, b):
- return self.compare(a, b) >= 0
-
- def min(self, *l):
- if len(l) == 1:
- l = l[0]
- rv = l[0]
- for i in l[1:]:
- if self.lt(i, rv):
- rv = i
- return rv
-
- def max(self, *l):
- if len(l) == 1:
- l = l[0]
- rv = l[0]
- for i in l[1:]:
- if self.gt(i, rv):
- rv = i
- return rv
-
- def limit_from_config(self, config):
- """
- a time limit to kill a result after such that it can be compared to config
- """
- results = self.driver.results_query(config=config)
- if results.count() == 0:
- return None
- else:
- return max(map(_.time, self.driver.results_query(config=config)))
-
-
- def project_compare(self, a1, a2, b1, b2, factor=1.0):
- """
- linearly project both a and b forward to see how they will compare in the
- future
- """
- a3 = Result()
- b3 = Result()
- a3.time = _project(a1.time, a2.time, factor)
- a3.accuracy = _project(a1.accuracy, a2.accuracy, factor)
- a3.energy = _project(a1.energy, a2.energy, factor)
- a3.confidence = _project(a1.confidence, a2.confidence, factor)
- return self.result_compare(a3, b3)
-
- def display(self, result):
- """
- produce a string version of a resultsdb.models.Result()
- """
- rv = []
- for k in ('time', 'accuracy', 'energy', 'size', 'confidence'):
- v = getattr(result, k)
- if v is not None:
- rv.append('%s=%.4f' % (k, float(v)))
- return ', '.join(rv)
-
- def filter_acceptable(self, query):
- """Return a Result() query that only returns acceptable results"""
- return query
-
- def is_acceptable(self, result):
- """Test if a Result() meets thresholds"""
- return True
-
- def stats_quality_score(self, result, worst_result, best_result):
- """return a score for statistics"""
- if not self.is_acceptable(result):
- return worst_result.time
- else:
- return result.time
-
-
-def _project(a1, a2, factor):
- if a1 is None or a2 is None:
- return None
- return a2 + factor * (a2 - a1)
-
-
-class MinimizeTime(SearchObjective):
- """
- minimize Result().time
- """
-
- def result_order_by_terms(self):
- """return database columns required to order by the objective"""
- return [Result.time]
-
- def result_compare(self, result1, result2):
- """cmp() compatible comparison of resultsdb.models.Result"""
- return cmp(result1.time, result2.time)
-
- def config_compare(self, config1, config2):
- """cmp() compatible comparison of resultsdb.models.Configuration"""
- return cmp(min(map(_.time, self.driver.results_query(config=config1))),
- min(map(_.time, self.driver.results_query(config=config2))))
-
- def result_relative(self, result1, result2):
- """return None, or a relative goodness of resultsdb.models.Result"""
- if result2.time == 0:
- return float('inf') * result1.time
- return result1.time / result2.time
-
-class MinimizeSize(SearchObjective):
-
- def result_order_by_terms(self):
- """return database columns required to order by the objective"""
- return [Result.size]
-
- def result_compare(self, result1, result2):
- """cmp() compatible comparison of resultsdb.models.Result"""
- return cmp(result1.size, result2.size)
-
- def result_relative(self, result1, result2):
- """return None, or a relative goodness of resultsdb.models.Result"""
- if result2.size == 0:
- return float('inf') * result1.size
- return result1.size / result2.size
-
-
-class MinimizeSizeMinimizeTime(SearchObjective):
- """
- minimize Result.size() and Result.time()
- """
- def result_order_by_terms(self):
- """return database columns required to order by the objective"""
- return [Result.time, Result.size]
-
-
- def result_compare(self, result1, result2):
- """cmp() compatible comparison of resultsdb.models.Result"""
- return cmp((result1.time, result1.size),(result2.time,result2.size))
-
- def result_relative(self, result1, result2):
- """return None, or a relative goodness of resultsdb.models.Result"""
- log.warning('result_relative() not yet implemented for %s',
- self.__class__.__name__)
-
-class MaximizeAccuracy(SearchObjective):
- """
- maximize Result().accuracy
- """
-
- def result_order_by_terms(self):
- """return database columns required to order by the objective"""
- return [-Result.accuracy]
-
- def result_compare(self, result1, result2):
- """cmp() compatible comparison of resultsdb.models.Result"""
- # note opposite order
- return cmp(result2.accuracy, result1.accuracy)
-
- def result_relative(self, result1, result2):
- """return None, or a relative goodness of resultsdb.models.Result"""
- # note opposite order
- if result1.accuracy == 0:
- return float('inf') * result2.accuracy
- return result2.accuracy / result1.accuracy
-
- def stats_quality_score(self, result, worst_result, best_result):
- """return a score for statistics"""
- if not self.is_acceptable(result):
- return worst_result.time
- else:
- return result.time
-
- def stats_raw_score(self, result):
- return result.accuracy
-
-
-class MaximizeAccuracyMinimizeSize(MaximizeAccuracy):
- """
- maximize Result().accuracy, break ties with Result().size
- """
-
- def result_order_by_terms(self):
- """return database columns required to order by the objective"""
- return [-Result.accuracy, Result.size]
-
- def result_compare(self, result1, result2):
- """cmp() compatible comparison of resultsdb.models.Result"""
- return cmp((-result1.accuracy, result1.size),
- (-result2.accuracy, result2.size))
-
- def display(self, result):
- """
- produce a string version of a resultsdb.models.Result()
- """
- return "accuracy=%.8f, size=%.1f" % (result.accuracy, result.size)
-
- def result_relative(self, result1, result2):
- """return None, or a relative goodness of resultsdb.models.Result"""
- # unimplemented for now
- log.warning('result_relative() not yet implemented for %s',
- self.__class__.__name__)
- return None
-
-
-class ThresholdAccuracyMinimizeTime(SearchObjective):
- """
- if accuracy >= target:
- minimize time
- else:
- maximize accuracy
- """
-
- def __init__(self, accuracy_target, low_accuracy_limit_multiplier=10.0):
- self.accuracy_target = accuracy_target
- self.low_accuracy_limit_multiplier = low_accuracy_limit_multiplier
- super(ThresholdAccuracyMinimizeTime, self).__init__()
-
- def result_order_by_terms(self):
- """return database columns required to order by the objective"""
-
- return ["min(accuracy, %f) desc" % self.accuracy_target,
- opentuner.resultsdb.models.Result.time]
-
- def result_compare(self, result1, result2):
- """cmp() compatible comparison of resultsdb.models.Result"""
- return cmp((-min(self.accuracy_target, result1.accuracy),
- result1.time),
- (-min(self.accuracy_target, result2.accuracy), result2.time))
-
- def config_compare(self, config1, config2):
- """cmp() compatible comparison of resultsdb.models.Configuration"""
- return self.result_compare(
- self.driver.results_query(config=config1, objective_ordered=True)[0],
- self.driver.results_query(config=config2, objective_ordered=True)[0])
-
- def limit_from_config(self, config):
- """
- a time limit to kill a result after such that it can be compared to config
- """
- results = self.driver.results_query(config=config)
- if results.count() == 0:
- return None
- if self.accuracy_target > min(map(_.accuracy, results)):
- m = self.low_accuracy_limit_multiplier
- else:
- m = 1.0
- return m * max(map(_.time, results))
-
-
- def filter_acceptable(self, query):
- """Return a Result() query that only returns acceptable results"""
- return query.filter(opentuner.resultsdb.models.Result.accuracy
- >= self.accuracy_target)
-
- def is_acceptable(self, result):
- """Test if a Result() meets thresholds"""
- return result.accuracy >= self.accuracy_target
-
- def result_relative(self, result1, result2):
- """return None, or a relative goodness of resultsdb.models.Result"""
- # unimplemented for now
- log.warning('result_relative() not yet implemented for %s',
- self.__class__.__name__)
- return None
-
-
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/patternsearch.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/patternsearch.py
deleted file mode 100644
index 7b526e7897..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/patternsearch.py
+++ /dev/null
@@ -1,72 +0,0 @@
-
-
-from opentuner.search import technique
-
-class PatternSearch(technique.SequentialSearchTechnique):
- def main_generator(self):
-
- objective = self.objective
- driver = self.driver
- manipulator = self.manipulator
-
- # start at a random position
- center = driver.get_configuration(manipulator.random())
- self.yield_nonblocking(center)
-
- # initial step size is arbitrary
- step_size = 0.1
-
- while True:
- points = list()
- for param in manipulator.parameters(center.data):
- if param.is_primitive():
- # get current value of param, scaled to be in range [0.0, 1.0]
- unit_value = param.get_unit_value(center.data)
-
- if unit_value > 0.0:
- # produce new config with param set step_size lower
- down_cfg = manipulator.copy(center.data)
- param.set_unit_value(down_cfg, max(0.0, unit_value - step_size))
- down_cfg = driver.get_configuration(down_cfg)
- self.yield_nonblocking(down_cfg)
- points.append(down_cfg)
-
- if unit_value < 1.0:
- # produce new config with param set step_size higher
- up_cfg = manipulator.copy(center.data)
- param.set_unit_value(up_cfg, min(1.0, unit_value + step_size))
- up_cfg = driver.get_configuration(up_cfg)
- self.yield_nonblocking(up_cfg)
- points.append(up_cfg)
-
- else: # ComplexParameter
- for mutate_function in param.manipulators(center.data):
- cfg = manipulator.copy(center.data)
- mutate_function(cfg)
- cfg = driver.get_configuration(cfg)
- self.yield_nonblocking(cfg)
- points.append(cfg)
-
-
- yield None # wait for all results
-
- #sort points by quality, best point will be points[0], worst is points[-1]
- points.sort(cmp=objective.compare)
-
- if (objective.lt(driver.best_result.configuration, center)
- and driver.best_result.configuration != points[0]):
- # another technique found a new global best, switch to that
- center = driver.best_result.configuration
- elif objective.lt(points[0], center):
- # we found a better point, move there
- center = points[0]
- else:
- # no better point, shrink the pattern
- step_size /= 2.0
-
-# register our new technique in global list
-technique.register(PatternSearch())
-
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/plugin.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/plugin.py
deleted file mode 100644
index ad8481837c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/plugin.py
+++ /dev/null
@@ -1,152 +0,0 @@
-import abc
-import argparse
-import logging
-import time
-
-from datetime import datetime
-from fn import _
-
-log = logging.getLogger(__name__)
-display_log = logging.getLogger(__name__ + ".DisplayPlugin")
-
-argparser = argparse.ArgumentParser(add_help=False)
-argparser.add_argument('--results-log',
- help="file to store log of the best configuration times")
-argparser.add_argument('--results-log-details',
- help="file to store log of the non-best configuration times")
-argparser.add_argument('--quiet', action='store_true',
- help="print less information")
-argparser.add_argument('--display-frequency', default=10, type=int,
- help="how often for DisplayPlugin to print")
-
-class SearchPlugin(object):
- @property
- def priority(self):
- """control order the plugin hooks gets run in, lower runs first"""
- return 0
-
- def set_driver(self, driver):
- """called before all other methods"""
- self.driver = driver
-
- def before_main(self): pass
- def after_main(self): pass
-
- def before_techniques(self): pass
- def after_techniques(self): pass
-
- def before_results_wait(self): pass
- def after_results_wait(self): pass
-
- def on_result(self, result):
- """
- called once for every new result
- """
- pass
-
- def on_result_for_technique(self, result, technique):
- """
- called right before a result is given to a technique
- (result may be requested by multiple techniques)
- """
- pass
-
- def on_new_best_result(self, result):
- """
- called whenever the global best result changes
- """
- pass
-
-class DisplayPlugin(SearchPlugin):
- __metaclass__ = abc.ABCMeta
- def __init__(self, display_period=5):
- super(DisplayPlugin, self).__init__()
- self.last = time.time()
- self.start = time.time()
- self.display_period = display_period
-
- def after_results_wait(self):
- t = time.time()
- if t - self.display_period > self.last:
- # call display every 5 seconds
- self.last = t
- self.display(t)
-
- def after_main(self):
- self.display()
-
- @abc.abstractmethod
- def display(self, t=None):
- pass
-
-
-class LogDisplayPlugin(DisplayPlugin):
- def display(self, t=None):
- if not t:
- t = time.time()
- count = self.driver.results_query().count()
- best = self.driver.results_query(objective_ordered = True).first()
- if best is None:
- log.warning("no results yet")
- return
- requestor = ','.join(map(_.requestor, best.desired_results))
- display_log.info("tests=%d, best %s, cost %s, found by %s",
- count,
- cfg_repr(best.configuration),
- self.driver.objective.display(best),
- requestor,
- )
-
-class FileDisplayPlugin(SearchPlugin):
- def __init__(self, out, details, *args, **kwargs):
- super(FileDisplayPlugin, self).__init__(*args, **kwargs)
- self.last_best = float('inf')
- self.start_date = datetime.now()
- if out:
- self.out = open(out, "w")
- else:
- self.out = None
- if out == details:
- self.details = self.out
- self.out = None
- elif details:
- self.details = open(details, "w")
- else:
- self.details = None
-
- def on_result(self, result):
- if self.out and result.time < self.last_best:
- self.last_best = result.time
- print >>self.out, \
- (result.collection_date - self.start_date).total_seconds(), \
- result.time
- self.out.flush()
- if self.details:
- print >>self.details, \
- (result.collection_date - self.start_date).total_seconds(), \
- result.time
- self.details.flush()
-
-def get_enabled(args):
- plugins = []
- if not args.quiet:
- plugins.append(LogDisplayPlugin(args.display_frequency))
- if args.results_log or args.results_log_details:
- plugins.append(FileDisplayPlugin(args.results_log,
- args.results_log_details))
- return plugins
-
-def cfg_repr(cfg):
- try:
- s = repr(cfg.data)
- if len(s) < 100:
- return s
- except:
- pass
- return "#{0}".format(cfg.id)
-
-
-
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/pso.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/pso.py
deleted file mode 100644
index 3b8c37a778..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/pso.py
+++ /dev/null
@@ -1,81 +0,0 @@
-# -*- coding: utf-8 -*-
-# vim: tabstop=2 shiftwidth=2 softtabstop=2 expandtab autoindent smarttab
-from manipulator import *
-from opentuner.search import technique
-import random
-import math
-
-class PSO(technique.SequentialSearchTechnique ):
- """ Particle Swarm Optimization """
- def __init__(self, crossover, N = 30, init_pop=None, *pargs, **kwargs):
- """
- crossover: name of crossover operator function
- """
- super(PSO, self).__init__(*pargs, **kwargs)
- self.crossover = crossover
- self.name = 'pso-'+crossover.replace("op3_cross_","")
- self.init_pop = init_pop
- self.N = N
-
- def main_generator(self):
-
- objective = self.objective
- driver = self.driver
- m = self.manipulator
- def config(cfg):
- return driver.get_configuration(cfg)
-
- population = self.init_pop
- if not population:
- population = [HybridParticle(m, self.crossover) for i in range(self.N)]
-
- for p in population:
- yield driver.get_configuration(p.position)
-
- while True:
- for particle in population:
- g = driver.best_result.configuration.data
- old=m.copy(particle.position)
- particle.move(g)
- yield config(particle.position)
- # update individual best
- if objective.lt(config(particle.position), config(particle.best)):
- particle.best = particle.position
-
-class HybridParticle(object):
- def __init__(self, m, crossover_choice, omega=0.5, phi_l=0.5, phi_g=0.5):
-
- """
- m: a configuraiton manipulator
- omega: influence of the particle's last velocity, a float in range [0,1] ; omega=1 means even speed
- phi_l: influence of the particle's distance to its historial best position, a float in range [0,1]
- phi_g: influence of the particle's distance to the global best position, a float in range [0,1]
- """
-
- self.manipulator = m
- self.position = self.manipulator.random()
- self.best = self.position
- self.omega = omega
- self.phi_l = phi_l
- self.phi_g = phi_g
- self.crossover_choice = crossover_choice
- self.velocity = {}
- for p in self.manipulator.params:
- # Velocity as a continous value
- self.velocity[p.name]=0
-
- def move(self, global_best):
- """
- Update parameter values using corresponding operators.
- TODO: introduce operator choice map
- """
- m = self.manipulator
- for p in m.params:
- self.velocity[p.name] = p.op3_swarm(self.position, global_best, self.best, c=self.omega, c1=self.phi_g, c2=self.phi_l, xchoice=self.crossover_choice, velocity=self.velocity[p.name])
-
-
-technique.register(PSO(crossover = 'op3_cross_OX3'))
-technique.register(PSO(crossover = 'op3_cross_OX1'))
-technique.register(PSO(crossover = 'op3_cross_PMX'))
-technique.register(PSO(crossover = 'op3_cross_PX'))
-technique.register(PSO(crossover = 'op3_cross_CX'))
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/simplextechniques.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/simplextechniques.py
deleted file mode 100644
index 3cfec0eebb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/simplextechniques.py
+++ /dev/null
@@ -1,457 +0,0 @@
-import abc
-import logging
-import math
-from collections import defaultdict
-from fn import _
-from fn.iters import map, filter
-from .manipulator import Parameter
-from .metatechniques import RecyclingMetaTechnique
-from .technique import SequentialSearchTechnique, register
-
-log = logging.getLogger(__name__)
-
-
-class SimplexTechnique(SequentialSearchTechnique):
- """
- Base class with utility functions common
- to simplex type methods
- """
-
- def __init__(self, seed_cfg=None, *args, **kwargs):
- super(SimplexTechnique, self).__init__(*args, **kwargs)
- self.centroid = None
- self.last_simplex_points = None
- self.seed_cfg = seed_cfg
- self.simplex_points = []
-
- def calculate_centroid(self):
- """
- average of all the PrimitiveParameters in self.simplex_points
- ComplexParameters are copied from self.simplex_points[0]
- """
- sums = defaultdict(float)
- counts = defaultdict(int)
-
- for config in self.simplex_points:
- cfg = config.data
- for param in self.manipulator.parameters(cfg):
- if param.is_primitive():
- sums[param.name] += param.get_unit_value(cfg)
- counts[param.name] += 1
-
- centroid = self.manipulator.copy(self.simplex_points[0].data)
- for param in self.manipulator.parameters(centroid):
- if param.is_primitive():
- param.set_unit_value(centroid,
- sums[param.name] / float(counts[param.name]))
-
- return centroid
-
- def cfg_to_str(self, cfg):
- params = list(filter(Parameter.is_primitive,
- self.manipulator.parameters(cfg)))
- params.sort(key=_.name)
- return str(tuple(map(lambda x: x.get_unit_value(cfg), params)))
-
- def debug_log(self):
- for i, config in enumerate(self.simplex_points):
- log.debug("simplex_points[%d] = %s", i, self.cfg_to_str(config.data))
- if self.centroid:
- log.debug("centroid = %s", self.cfg_to_str(self.centroid))
-
- def linear_point(self, p1, p2, scale):
- """
- return a point on the line passing between p1 and p2 at position scale
- such that p1 + scale*(p1 - p2)
- """
- return self.manipulator.linear_config(1.0, p1, scale, p1, -scale, p2)
-
- def convergence_criterea(self):
- """True will cause the simplex method to stop"""
- if self.rounds_since_novel_request > 3 * len(self.simplex_points) + 1:
- return True
- if self.last_simplex_points == self.simplex_points:
- return True
- self.last_simplex_points = list(self.simplex_points)
- return False
-
- def initial_simplex_seed(self):
- """
- return a point to base the initial simplex on
- """
- if self.seed_cfg is not None:
- return self.seed_cfg
- return self.manipulator.random()
-
- @abc.abstractmethod
- def initial_simplex(self):
- """
- return a initial list of configurations
- """
- return []
-
-
-class RandomInitialMixin(object):
- """
- start with random initial simplex
- """
-
- def initial_simplex(self):
- # we implicitly assume number of parameters is fixed here, however
- # it will work if it isn't (simplex size is undefined)
- cfg0 = self.initial_simplex_seed()
- params = self.manipulator.parameters(cfg0)
- return [cfg0] + [self.manipulator.random()
- for p in params
- if p.is_primitive()]
-
-
-class RightInitialMixin(object):
- """
- start with random initial right triangle like simplex
- """
-
- def __init__(self, initial_unit_edge_length=0.1, *args, **kwargs):
- assert initial_unit_edge_length <= 0.5
- self.initial_unit_edge_length = initial_unit_edge_length
- super(RightInitialMixin, self).__init__(*args, **kwargs)
-
- def initial_simplex(self):
- cfg0 = self.initial_simplex_seed()
- simplex = [cfg0]
- params = self.manipulator.parameters(cfg0)
- params = filter(lambda x: x.is_primitive(), params)
- for p in params:
- simplex.append(self.manipulator.copy(cfg0))
- v = p.get_unit_value(simplex[-1])
- if v <= 0.5:
- v += self.initial_unit_edge_length
- else:
- v -= self.initial_unit_edge_length
- p.set_unit_value(simplex[-1], v)
- return simplex
-
-
-class RegularInitialMixin(object):
- """
- start with random initial regular simplex (all edges equal length)
- """
-
- def __init__(self, initial_unit_edge_length=0.1, *args, **kwargs):
- assert initial_unit_edge_length <= 0.5
- self.initial_unit_edge_length = initial_unit_edge_length
- super(RegularInitialMixin, self).__init__(*args, **kwargs)
-
- def initial_simplex(self):
- cfg0 = self.initial_simplex_seed()
- simplex = [cfg0]
- params = self.manipulator.parameters(cfg0)
- params = list(filter(lambda x: x.is_primitive(), params))
- if len(params) == 0:
- return simplex
-
- q = (((math.sqrt(len(params) + 1.0) - 1.0) / (len(params) * math.sqrt(2.0)))
- * self.initial_unit_edge_length)
- p = q + ((1.0 / math.sqrt(2.0)) * self.initial_unit_edge_length)
-
- base = [x.get_unit_value(cfg0) for x in params]
- for j in xrange(len(base)):
- if max(p, q) + base[j] > 1.0:
- #flip this dimension as we would overflow our [0,1] bounds
- base[j] *= -1.0
-
- for i in xrange(len(params)):
- simplex.append(self.manipulator.copy(cfg0))
- params[i].set_unit_value(simplex[-1], abs(base[i] + p))
- for j in xrange(i + 1, len(params)):
- params[j].set_unit_value(simplex[-1], abs(base[i] + q))
-
- return simplex
-
-
-class NelderMead(SimplexTechnique):
- """
- Nelder-Mead downhill simplex method.
-
- Based on description of method on page 82 of
- 'Noisy Optimization With Evolution Strategies' by Dirk V. Arnold.
-
- We set alpha=2.0 by default instead of the often recommended alpha=1.0 to
- avoid a common degenerate case, where the volume of the simplex becomes zero.
- This is easiest to see with a single parameter. Let the simplex points
- be x0,x1. Let the centroid be c=(x0+x1)/2.0 and the reflection point be:
- reflection = c + alpha*(c-x1) = (x0+x1)*(1+alpha)/2 - x1
- The problem is, if we set alpha = 1.0, then the x1's cancel out and the
- reflection point becomes just reflection=x0, which also happens to be the
- second best point, meaning we will use it. So in a single step of the
- algorithm the simplex becomes singular.
- """
-
- def __init__(self,
- alpha=2.0,
- gamma=2.0,
- beta=0.5,
- sigma=0.5,
- *args, **kwargs):
- self.alpha = alpha
- self.gamma = gamma
- self.beta = beta
- self.sigma = sigma
- super(NelderMead, self).__init__(*args, **kwargs)
-
- @classmethod
- def get_hyper_parameters(cls):
- return ['alpha', 'gamma', 'beta', 'sigma']
-
-
- def main_generator(self):
- objective = self.objective
- driver = self.driver
-
- # test the entire initial simplex
- self.simplex_points = list(map(driver.get_configuration,
- self.initial_simplex()))
-
- if len(self.simplex_points) <= 1:
- log.warning("only 1 point in simplex, will not use %s", self.name)
- return
-
- log.debug("initial points")
- for p in self.simplex_points:
- self.yield_nonblocking(p)
- yield None # wait until results are ready
-
- while not self.convergence_criterea():
- # next steps assume this ordering
- self.simplex_points.sort(cmp=objective.compare)
- # set limit from worst point
- self.limit = objective.limit_from_config(self.simplex_points[-1])
- self.centroid = self.calculate_centroid()
- if log.isEnabledFor(logging.DEBUG):
- self.debug_log()
-
- reflection = self.reflection_point()
- yield reflection
-
- if objective.lt(reflection, self.simplex_points[0]):
- #expansion case
- expansion = self.expansion_point(reflection)
- yield expansion
-
- if objective.lt(expansion, reflection):
- log.debug("using expansion point")
- self.simplex_points[-1] = expansion
- else:
- log.debug("using reflection point (considered expansion)")
- self.simplex_points[-1] = reflection
-
- elif objective.lt(reflection, self.simplex_points[1]):
- #reflection case
- log.debug("using reflection point")
- self.simplex_points[-1] = reflection
- else:
- # contraction case
- if objective.lte(reflection, self.simplex_points[-1]):
- # outside contraction
- contract_base = reflection
- else:
- # inside contraction
- contract_base = self.simplex_points[-1]
-
- contraction = self.contraction_point(contract_base)
- yield contraction
-
- if objective.lte(contraction, contract_base):
- log.debug("using contraction point")
- self.simplex_points[-1] = contraction
- else:
- #reduction case
- log.debug("performing shrink reduction")
- self.perform_shrink_reduction()
- for p in self.simplex_points:
- self.yield_nonblocking(p)
- yield None # wait until results are ready
-
- def reflection_point(self):
- """
- reflect worst point across centroid
- """
- return self.driver.get_configuration(
- self.linear_point(self.centroid,
- self.simplex_points[-1].data,
- self.alpha))
-
- def expansion_point(self, reflection):
- """
- reflect worst point across centroid more (by default 2x as much)
- """
- return self.driver.get_configuration(
- self.linear_point(self.centroid,
- reflection.data,
- -self.gamma))
-
- def contraction_point(self, contract_base):
- """
- reflect worst point across centroid less
- """
- return self.driver.get_configuration(
- self.linear_point(self.centroid,
- contract_base.data,
- -self.beta))
-
- def perform_shrink_reduction(self):
- """
- shrink the simplex in size by sigma=1/2 (default), moving it closer to the
- best point
- """
- for i in xrange(1, len(self.simplex_points)):
- self.simplex_points[i] = self.driver.get_configuration(
- self.linear_point(self.simplex_points[0].data,
- self.simplex_points[i].data,
- -self.sigma))
-
-
-class Torczon(SimplexTechnique):
- """
- Torczon multi-directional search algorithm.
-
- Based on description of method on page 85 of
- 'Noisy Optimization With Evolution Strategies' by Dirk V. Arnold.
- """
-
- def __init__(self,
- alpha=1.0,
- gamma=2.0,
- beta=0.5,
- *args, **kwargs):
- self.alpha = alpha
- self.gamma = gamma
- self.beta = beta
- super(Torczon, self).__init__(*args, **kwargs)
-
- @classmethod
- def get_hyper_parameters(cls):
- return ['alpha', 'gamma', 'beta']
-
-
- def main_generator(self):
- objective = self.objective
- driver = self.driver
-
- # test the entire initial simplex
- self.simplex_points = list(map(driver.get_configuration,
- self.initial_simplex()))
- if len(self.simplex_points) <= 1:
- log.warning("only 1 point in simplex, will not use %s", self.name)
- return
-
- log.debug("initial points")
- for p in self.simplex_points:
- self.yield_nonblocking(p)
- yield None # wait until results are ready
- self.simplex_points.sort(cmp=objective.compare)
-
- while not self.convergence_criterea():
- # set limit from worst point
- self.limit = objective.limit_from_config(self.simplex_points[-1])
-
- if log.isEnabledFor(logging.DEBUG):
- self.debug_log()
-
- reflected = self.reflected_simplex()
- yield None # wait until results are ready
- reflected.sort(cmp=objective.compare)
-
- # this next condition implies reflected[0] < simplex_points[0] since
- # reflected is sorted and contains simplex_points[0] (saves a db query)
- if reflected[0] is not self.simplex_points[0]:
- expanded = self.expanded_simplex()
- yield None # wait until results are ready
- expanded.sort(cmp=objective.compare)
-
- if objective.lt(expanded[0], reflected[0]):
- log.debug("expansion performed")
- self.simplex_points = expanded
- else:
- log.debug("reflection performed")
- self.simplex_points = reflected
- else:
- contracted = self.contracted_simplex()
- yield None # wait until results are ready
- contracted.sort(cmp=objective.compare)
-
- log.debug("contraction performed")
- self.simplex_points = contracted
-
- def scaled_simplex(self, scale):
- """
- assumes self.simplex_points[0] is best point and returns a new simplex
- reflected across self.simplex_points[0] by scale
- """
- simplex = list(self.simplex_points) # shallow copy
- for i in xrange(1, len(simplex)):
- simplex[i] = self.driver.get_configuration(
- self.linear_point(simplex[0].data, simplex[i].data, scale))
- self.yield_nonblocking(simplex[i])
- return simplex
-
- def reflected_simplex(self):
- return self.scaled_simplex(self.alpha)
-
- def expanded_simplex(self):
- return self.scaled_simplex(self.gamma)
-
- def contracted_simplex(self):
- return self.scaled_simplex(-self.beta)
-
-
-class RandomNelderMead(RandomInitialMixin, NelderMead):
- pass
-
-
-class RightNelderMead(RightInitialMixin, NelderMead):
- pass
-
-
-class RegularNelderMead(RegularInitialMixin, NelderMead):
- pass
-
-
-class RandomTorczon(RandomInitialMixin, Torczon):
- pass
-
-
-class RightTorczon(RightInitialMixin, Torczon):
- pass
-
-
-class RegularTorczon(RegularInitialMixin, Torczon):
- pass
-
-
-class MultiNelderMead(RecyclingMetaTechnique):
- def __init__(self):
- super(MultiNelderMead, self).__init__([RightNelderMead, RandomNelderMead,
- RegularNelderMead])
-
-
-class MultiTorczon(RecyclingMetaTechnique):
- def __init__(self):
- super(MultiTorczon, self).__init__([RightTorczon, RandomTorczon,
- RegularTorczon])
-
-
-register(RandomNelderMead())
-register(RegularNelderMead())
-register(RightNelderMead())
-register(MultiNelderMead())
-register(RandomTorczon())
-register(RegularTorczon())
-register(RightTorczon())
-register(MultiTorczon())
-
-
-
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/simulatedannealing.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/simulatedannealing.py
deleted file mode 100644
index 45b315f2e6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/simulatedannealing.py
+++ /dev/null
@@ -1,133 +0,0 @@
-from opentuner.search import technique
-import math
-import random
-#Default interval steps for cooling schedules
-DEFAULT_INTERVAL = 100
-
-#Pseudo-annealing - no relative energy input into acceptance function
-class PseudoAnnealingSearch(technique.SequentialSearchTechnique):
- def __init__(self,
- temps = [30,0], #temperature schedule
- intervals = [], #duration schedule
- loop = True, #Do we loop the schedule if we reach the end?
- *pargs, **kwargs):
- #fill intervals sufficiently
- ext_intervals = list(intervals)
- for i in range(len(temps)-len(intervals)-1):
- ext_intervals.append(DEFAULT_INTERVAL)
-
- #create temperature schedule (list of temps)
- cool_schedule = [temps[0]]
- for i in range(len(temps)-1):
- step = (float(temps[i+1]) - temps[i])/ext_intervals[i]
- for j in range(ext_intervals[i]):
- cool_schedule.append(max(cool_schedule[-1] + step,0))
-
- self.cool_schedule = cool_schedule
- self.loop = loop
- self.scaling = 50 #scaling of acceptance function
-
- super(PseudoAnnealingSearch,self).__init__(*pargs,**kwargs)
-
-
- def main_generator(self):
- objective = self.objective
- driver = self.driver
- manipulator = self.manipulator
-
- #Start in a random spot
- state = driver.get_configuration(manipulator.random())
- yield state
- #schedule counter
- counter = 0
- max_time = len(self.cool_schedule)-1
- #Check whether relative objective implemented
- has_rel = objective.relative(state,state) is not None
- has_rel=False
-
- while True:
- #Determine temperature
- temp = self.cool_schedule[min(counter,max_time)]
- #scale stepsize with temp and time (arbitrary)
- step_size = math.exp(-(20 + counter/100)/(temp+ 1))
-
- #get candidate neighbors using manipulator
- points = list()
- points.append(state)
- for param in manipulator.parameters(state.data):
- if param.is_primitive():
- # get current value of param, scaled to be in range [0.0, 1.0]
- unit_value = param.get_unit_value(state.data)
- if unit_value > 0.0:
- # produce new config with param set step_size lower
- down_cfg = manipulator.copy(state.data)
- param.set_unit_value(down_cfg, max(0.0, unit_value - step_size*random.random()))
- down_cfg = driver.get_configuration(down_cfg)
- self.yield_nonblocking(down_cfg)
- points.append(down_cfg)
-
- if unit_value < 1.0:
- # produce new config with param set step_size higher
- up_cfg = manipulator.copy(state.data)
- param.set_unit_value(up_cfg, min(1.0, unit_value + step_size*random.random()))
- up_cfg = driver.get_configuration(up_cfg)
- self.yield_nonblocking(up_cfg)
- points.append(up_cfg)
- else: # ComplexParameter
- for mutate_function in param.manipulators(state.data):
- cfg = manipulator.copy(state.data)
- mutate_function(cfg)
- cfg = driver.get_configuration(cfg)
- self.yield_nonblocking(cfg)
- points.append(cfg)
- yield None # wait for all results
-
- #Relative comparison implemented
- if has_rel:
- while True:
- if len(points) == 0:
- state = driver.best_result.configuration
- break
- candidate = points.pop(random.randint(0,len(points)-1))
- #compare to global best
- if random.random() < AcceptanceFunction(1, objective.relative(candidate,driver.best_result.configuration), temp, self.scaling):
- state = candidate
- break
- #No relative compare
- else:
- #sort points by "energy" (quality)
- points.sort(cmp=objective.compare)
-
- #Make decision about changing state
- #probability picking next-best state is exp^(-1/temp)
- #repeat and cycle to get state p-dist resembling this
- sel = 0
- while AcceptanceFunction(0,1,temp,1)>random.random():
- sel += 1
- state = points[sel%len(points)]
-
- #switch to the global best if temperature is low (i.e. we aren't moving much)
- if AcceptanceFunction(0,1,temp,1)< .0001 and objective.lt(driver.best_result.configuration, state):
- state = driver.best_result.configuration
-
- #update counter
- counter +=1
- if counter>max_time and self.loop:
- counter=counter-max_time
-
-
-#Acceptance probability function for annealing
-def AcceptanceFunction(e,e_new,temp,scaling):
- #Standard acceptance probability function using relative "goodness"
- if e>=e_new:
- return 1
- if temp == 0:
- return 0
- if scaling*(e_new-e)/temp > 10:
- #for practical purposes, probability is too low.
- return 0
- return math.exp(scaling*(e-e_new)/temp)
-
-
-#register technique
-technique.register(PseudoAnnealingSearch())
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/technique.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/technique.py
deleted file mode 100644
index 849391df9b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/search/technique.py
+++ /dev/null
@@ -1,358 +0,0 @@
-import abc
-import argparse
-import logging
-import os
-import random
-import sys
-
-from importlib import import_module
-from datetime import datetime
-from fn import _
-
-from opentuner.resultsdb.models import *
-from plugin import SearchPlugin
-
-log = logging.getLogger(__name__)
-#log.setLevel(logging.DEBUG)
-
-argparser = argparse.ArgumentParser(add_help=False)
-argparser.add_argument('--technique','-t', action='append',
- help="which technique to use")
-argparser.add_argument('--list-techniques','-lt', action='store_true',
- help="list techniques available and exit")
-argparser.add_argument('--generate-bandit-technique','-gbt', action='store_true',
- help="randomly generate a bandit to use")
-
-class SearchTechniqueBase(object):
- """
- abstract base class for search techniques, with minimal interface
- """
- __metaclass__ = abc.ABCMeta
-
- def __init__(self, name = None):
- super(SearchTechniqueBase, self).__init__()
- if name:
- self.name = name
- else:
- self.name = self.default_name()
-
- def is_ready(self):
- """test if enough data has been gathered to use this technique"""
- return True
-
- def default_name(self):
- """name of this SearchTechnique uses for display/accounting"""
- return self.__class__.__name__
-
- def handle_requested_result(self, result):
- """called for each new Result(), requested by this technique"""
- pass
-
- @abc.abstractmethod
- def set_driver(self, driver):
- """called at start of tuning process"""
- return
-
- @abc.abstractmethod
- def desired_result(self):
- """
- return at most count resultsdb.models.DesiredResult objects based on past
- performance
- """
- return
-
-class SearchTechnique(SearchPlugin, SearchTechniqueBase):
- """
- a search search technique with basic utility functions
- """
-
- def __init__(self, *pargs, **kwargs):
- super(SearchTechnique, self).__init__(*pargs, **kwargs)
- self.driver = None
- self.manipulator = None
- self.objective = None
- self.request_count = 0
-
- def set_driver(self, driver):
- super(SearchTechnique, self).set_driver(driver)
- self.manipulator = driver.manipulator
- self.objective = driver.objective
- driver.add_plugin(self)
-
- def desired_result(self):
- """
- create and return a resultsdb.models.DesiredResult
- returns None if no desired results and False if waiting for results
- """
- cfg = self.desired_configuration()
- if cfg is None:
- return None
- if cfg is False:
- return False
- if type(cfg) is Configuration:
- config = cfg
- else:
- config = self.driver.get_configuration(cfg)
- desired = DesiredResult(configuration=config,
- requestor=self.name,
- generation=self.driver.generation,
- request_date=datetime.now(),
- tuning_run=self.driver.tuning_run)
- if hasattr(self, 'limit'):
- desired.limit = self.limit
- self.driver.register_result_callback(desired, self.handle_requested_result)
- self.request_count += 1
- return desired
-
- @abc.abstractmethod
- def desired_configuration(self):
- """
- return a cfg that we should test
- given a ConfigurationManipulator and SearchDriver
- return None if there are no configurations to test
- return False if waiting for results
- """
- return dict()
-
- def handle_requested_result(self, result):
- """called for each new Result(), regardless of who requested it"""
- pass
-
- def default_generated_name(self):
- """ The default generated name for this technique """
- return self.base_name()
-
- def use_default_generated_name(self):
- """ set the name of this technique to the default generated name """
- self.name = self.default_generated_name()
-
- def base_name(self):
- """
- Return the base name of this technique with form
- classname;hyperparam1,v1;hyperparam2,v2 ...
- where hyperparams are taken in order from get_hyper_parameters()
-
- Should only be called after this technique has finished initializing.
- """
- out = [self.__class__.__name__]
- for hyper_parameter in self.get_hyper_parameters():
- # get hyperparam,v as a string and append
- try:
- out.append(hyper_parameter + ',' + str(getattr(self, hyper_parameter)))
- except AttributeError:
- log.error("Uninitialized hyper-parameter %s for technique %s.",
- hyper_parameter, self.__class__.__name__)
-
- return ';'.join(out)
-
- @classmethod
- def get_hyper_parameters(cls):
- """
- return a list of hyper-parameters names for this technique
-
- Name strings must match the corresponding attribute with the hyper-parameter
- value on technique instances. Names should also match the key word argument
- used when initializing an instance. Hyperparameters should only take literal
- values.
-
- For example, given hyper parameter "mutation_rate", then the __init__ method
- should have 'mutation_rate' as a key word argument and later have the line
- self.mutation_rate = mutation_rate
- """
- return []
-
- @classmethod
- def generate_technique(cls, manipulator=None, *args, **kwargs):
- """ return a new technique based off this instance """
- t = cls(*args, **kwargs)
- t.use_default_generated_name()
- return t
-
-class PureRandom(SearchTechnique):
- """
- request configurations completely randomly
- """
- def desired_configuration(self):
- return self.manipulator.random()
-
-class AsyncProceduralSearchTechnique(SearchTechnique):
- def __init__(self, *pargs, **kwargs):
- super(AsyncProceduralSearchTechnique, self).__init__(*pargs, **kwargs)
- self.gen = None
- self.done = False
- self.latest_results = []
-
- def call_main_generator(self):
- """passthrough (used in subclasses)"""
- return self.main_generator()
-
- def desired_configuration(self):
- if self.gen is None:
- log.debug("%s: creating generator", self.name)
- self.gen = self.call_main_generator()
- if not self.done:
- try:
- return self.gen.next()
- except StopIteration:
- log.debug("%s: generator finished", self.name)
- self.done = True
- return None
-
- @abc.abstractmethod
- def main_generator(self):
- """
- custom generator to conduct this search, should:
- yield config
- to request tests and call driver.get_results() to read the results
-
- in AsyncProceduralSearchTechnique results are ready at an undefined
- time (`yield False` to stall and wait for them)
-
- in SequentialSearchTechnique results are ready after the yield
- """
- pass
-
- def is_ready(self):
- return not self.done
-
-class SequentialSearchTechnique(AsyncProceduralSearchTechnique):
- def __init__(self, novelty_threshold=50, reset_threshold=500, *pargs, **kwargs):
- super(SequentialSearchTechnique, self).__init__(*pargs, **kwargs)
- self.pending_tests = []
- self.novelty_threshold = novelty_threshold
- self.rounds_since_novel_request = 0
- self.reset_threshold = reset_threshold
-
- def yield_nonblocking(self, cfg):
- """
- within self.main_generator() act like `yield cfg`, but don't wait for the
- results until the following yield (spawn/sync style)
- """
- if cfg:
- self.pending_tests.append(cfg)
-
- def call_main_generator(self):
- """insert waits for results after every yielded item"""
- subgen = self.main_generator()
- self.rounds_since_novel_request = 0
- while True:
- self.rounds_since_novel_request += 1
- if (self.rounds_since_novel_request % self.novelty_threshold) == 0:
- log.warning("%s has not requested a new result for %d rounds",
- self.name, self.rounds_since_novel_request)
- if (self.rounds_since_novel_request > self.reset_threshold):
- log.warning("%s is being reset", self.name)
- subgen = self.main_generator()
- self.rounds_since_novel_request = 0
- yield None # give other techniques a shot
- try:
- p = subgen.next()
- if p:
- self.pending_tests.append(p)
- except StopIteration:
- return
- finally:
- for p in self.pending_tests:
- if not self.driver.has_results(p):
- self.rounds_since_novel_request = 0
- yield p
-
- # wait for all pending_tests to have results
- c = 0
- while self.pending_tests:
- log.debug("%s: waiting for %d pending tests",
- self.name, len(self.pending_tests))
- c += 1
- if (c % 100) == 0:
- log.error("%s: still waiting for %d pending tests (c=%d)",
- self.name, len(self.pending_tests), c)
-
- self.pending_tests = filter(lambda x: not self.driver.has_results(x),
- self.pending_tests)
- if self.pending_tests:
- self.rounds_since_novel_request = 0
- yield False # wait
-
-#list of all techniques
-the_registry = list()
-
-#list of technique generators
-the_generator_registry = list()
-
-def register(t):
- the_registry.append(t)
-
-def register_generator(cls, generator_weight=1.0, *args, **kwargs):
- """
- register a technique generator - a tuple of (technique class, args, kwargs)
- where args and kwargs will be passed into the generate_technique classmethod -
- with specified probability weight when randomly choosing a generator
-
- :param cls: a technique class to use as a generator
- :param generator_weight: probability weighting when randomly choosing a generator
- :param args: arguments to pass into generate_technique class method
- :param kwargs: arguments to pass into generate_technique class method
- """
- the_generator_registry.append(((cls, args, kwargs), generator_weight))
-
-register(PureRandom())
-
-def get_random_generator_technique(generators=None, manipulator=None):
- """
- Takes in a sequence of ((generator, args, kwargs), weight) tuples.
- Returns a random generated technique info tuple
-
- :param generators: optional argument to avoid repeated getting of generators
- :param manipulator: manipulator to pass to generate_technique class method.
- """
- if generators is None:
- techniques, generators = all_techniques()
- g, args, kwargs = weighted_choice(generators)
- return g.generate_technique(manipulator, *args, **kwargs)
-
-
-def weighted_choice(choices):
- """ takes in a sequence of (choice, weight) tuples and randomly returns one """
- total = sum(w for c, w in choices)
- r = random.uniform(0, total)
- upto = 0
- for c, w in choices:
- upto += w
- if upto > r:
- return c
- return random.choice([c for c, w in choices])
-
-
-def all_techniques():
- #import all modules in search to ensure techniques are Registered
- for f in sorted(os.listdir(os.path.dirname(__file__))):
- m = re.match(r'^(.*)[.]py$', f)
- if m:
- import_module('opentuner.search.'+m.group(1))
-
- return the_registry, the_generator_registry
-
-def get_enabled(args):
- techniques, generators = all_techniques()
- if args.list_techniques:
- for t in techniques:
- print t.name
- sys.exit(0)
-
- if not args.technique:
- # no techniques specified, default technique
- args.technique = ['AUCBanditMetaTechniqueA']
-
- for unknown in set(args.technique) - set(map(_.name, techniques)):
- log.error('unknown technique %s', unknown)
- raise Exception('Unknown technique: --technique={}'.format(unknown))
-
- return [t for t in techniques if t.name in args.technique]
-
-def get_root(args):
- from metatechniques import RoundRobinMetaSearchTechnique
- enabled = get_enabled(args)
- if len(enabled) == 1:
- return enabled[0]
- return RoundRobinMetaSearchTechnique(get_enabled(args))
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/tuningrunmain.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/tuningrunmain.py
deleted file mode 100644
index 9bcf1b5270..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/tuningrunmain.py
+++ /dev/null
@@ -1,224 +0,0 @@
-# vim: tabstop=2 shiftwidth=2 softtabstop=2 expandtab autoindent smarttab
-import argparse
-import copy
-import inspect
-import logging
-import math
-import os
-import socket
-import sys
-import time
-import uuid
-from datetime import datetime
-
-from opentuner import resultsdb
-from opentuner.search.driver import SearchDriver
-from opentuner.measurement.driver import MeasurementDriver
-
-log = logging.getLogger(__name__)
-
-argparser = argparse.ArgumentParser(add_help=False)
-argparser.add_argument('--label',
- help="name for the TuningRun")
-argparser.add_argument('--print-search-space-size', action='store_true',
- help="Print out the estimated size of the search space and exit")
-argparser.add_argument('--database',
- help=("database to store tuning results in, see: "
- "http://docs.sqlalchemy.org/en/rel_0_8/core/engines.html#database-urls"))
-argparser.add_argument('--print-params','-pp',action='store_true',
- help='show parameters of the configuration being tuned')
-
-
-class CleanStop(Exception):
- pass
-
-
-class LogFormatter(logging.Formatter):
- def format(self, record):
- record.relativeCreated /= 1000.0
- try:
- # python 2.7
- return super(LogFormatter, self).format(record)
- except:
- # python 2.6
- return _OldFormatter.format(self, record)
-
-
-_OldFormatter = logging.Formatter
-logging.Formatter = LogFormatter
-
-try:
- # python 2.7
- from logging.config import dictConfig
-except:
- # python 2.6
- from .utils.dictconfig import dictConfig
-
-the_logging_config = {
- 'version': 1,
- 'disable_existing_loggers': False,
- 'formatters': {'console': {'format': '[%(relativeCreated)6.0fs] '
- '%(levelname)7s %(name)s: '
- '%(message)s'},
- 'file': {'format': '[%(asctime)-15s] '
- '%(levelname)7s %(name)s: '
- '%(message)s '
- '@%(filename)s:%(lineno)d'}},
- 'handlers': {'console': {'class': 'logging.StreamHandler',
- 'formatter': 'console',
- 'level': 'INFO'},
- 'file': {'class': 'logging.FileHandler',
- 'filename': 'opentuner.log',
- 'formatter': 'file',
- 'level': 'WARNING'}},
- 'loggers': {'': {'handlers': ['console', 'file'],
- 'level': 'INFO',
- 'propagate': True}}}
-
-
-def init_logging():
- dictConfig(the_logging_config)
- global init_logging
- init_logging = lambda: None
-
-
-class TuningRunMain(object):
- def __init__(self,
- measurement_interface,
- args,
- search_driver=SearchDriver,
- measurement_driver=MeasurementDriver):
- init_logging()
-
- manipulator = measurement_interface.manipulator()
- if args.print_search_space_size:
- print "10^{%.2f}" % math.log(manipulator.search_space_size(), 10)
- sys.exit(0)
- # show internal parameter representation
- if args.print_params:
- cfg = manipulator.seed_config()
- d = manipulator.parameters_dict(cfg)
- params_dict ={}
- for k in d:
- cls = d[k].__class__.__name__
- p = (k, d[k].search_space_size())
- if cls in params_dict:
- params_dict[cls].append(p)
- else:
- params_dict[cls] = [p]
- for k in params_dict:
- print k, params_dict[k]
- print
- sys.exit(0)
-
- input_manager = measurement_interface.input_manager()
- objective = measurement_interface.objective()
-
- if not args.database:
- #args.database = 'sqlite://' #in memory
- if not os.path.isdir('opentuner.db'):
- os.mkdir('opentuner.db')
- args.database = 'sqlite:///' + os.path.join('opentuner.db',
- socket.gethostname() + '.db')
-
- if '://' not in args.database:
- args.database = 'sqlite:///' + args.database
-
- if not args.label:
- args.label = 'unnamed'
-
- #self.fake_commit = ('sqlite' in args.database)
- self.fake_commit = True
-
- self.args = args
-
- self.engine, self.Session = resultsdb.connect(args.database)
- self.session = self.Session()
- self.tuning_run = None
- self.search_driver_cls = search_driver
- self.measurement_driver_cls = measurement_driver
- self.measurement_interface = measurement_interface
- self.input_manager = input_manager
- self.manipulator = manipulator
- self.objective = objective
- self.objective_copy = copy.copy(objective)
- self.last_commit_time = time.time()
-
- def init(self):
- if self.tuning_run is None:
- program_version = (self.measurement_interface
- .db_program_version(self.session))
- self.session.flush()
- self.measurement_interface.prefix_hook(self.session)
- self.tuning_run = (
- resultsdb.models.TuningRun(
- uuid=uuid.uuid4().hex,
- name=self.args.label,
- args=self.args,
- start_date=datetime.now(),
- program_version=program_version,
- objective=self.objective_copy,
- ))
- self.session.add(self.tuning_run)
-
- driver_kwargs = {
- 'args': self.args,
- 'input_manager': self.input_manager,
- 'manipulator': self.manipulator,
- 'measurement_interface': self.measurement_interface,
- 'objective': self.objective,
- 'session': self.session,
- 'tuning_run_main': self,
- 'tuning_run': self.tuning_run,
- 'extra_seeds': self.measurement_interface.seed_configurations(),
- 'extra_criteria': self.measurement_interface.extra_convergence_criteria
- }
-
- self.search_driver = self.search_driver_cls(**driver_kwargs)
-
- self.measurement_driver = self.measurement_driver_cls(**driver_kwargs)
- self.measurement_interface.set_driver(self.measurement_driver)
- self.input_manager.set_driver(self.measurement_driver)
-
- self.tuning_run.machine_class = self.measurement_driver.get_machine_class()
- self.tuning_run.input_class = self.input_manager.get_input_class()
-
- def commit(self, force=False):
- if (force or not self.fake_commit or
- time.time() - self.last_commit_time > 30):
- self.session.commit()
- self.last_commit_time = time.time()
- else:
- self.session.flush()
-
- def main(self):
- self.init()
- try:
- self.tuning_run.state = 'RUNNING'
- self.commit(force=True)
- self.search_driver.main()
- if self.search_driver.best_result:
- self.measurement_interface.save_final_config(
- self.search_driver.best_result.configuration)
- self.tuning_run.final_config = self.search_driver.best_result.configuration
- self.tuning_run.state = 'COMPLETE'
- except:
- self.tuning_run.state = 'ABORTED'
- raise
- finally:
- self.tuning_run.end_date = datetime.now()
- self.commit(force=True)
- self.session.close()
-
- def results_wait(self, generation):
- """called by search_driver to wait for results"""
- #single process version:
- self.measurement_interface.pre_process()
- self.measurement_driver.process_all()
- self.measurement_interface.post_process()
-
-def main(interface, args, *pargs, **kwargs):
- if inspect.isclass(interface):
- interface = interface(args=args, *pargs, **kwargs)
- return TuningRunMain(interface, args).main()
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/__init__.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/__init__.py
deleted file mode 100644
index e69de29bb2..0000000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/adddeps.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/adddeps.py
deleted file mode 100644
index e2fc74064b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/adddeps.py
+++ /dev/null
@@ -1,13 +0,0 @@
-
-import sys
-from os.path import normpath, realpath, dirname, join, isfile
-
-project_root = normpath(join(dirname(realpath(__file__)), '../..'))
-
-if 'venv' not in ','.join(sys.path):
- venv_activate = join(project_root, 'venv/bin/activate_this.py')
- if isfile(venv_activate):
- execfile(venv_activate, dict(__file__=venv_activate))
-
-sys.path.insert(0, project_root)
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/compactdb.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/compactdb.py
deleted file mode 100755
index 25a70d2d3b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/compactdb.py
+++ /dev/null
@@ -1,54 +0,0 @@
-#!/usr/bin/env python
-
-if __name__ == '__main__':
- import adddeps
-
-import argparse
-import logging
-import sys
-
-import opentuner
-from opentuner.resultsdb.models import *
-
-log = logging.getLogger('opentuner.utils.compactdb')
-
-argparser = argparse.ArgumentParser()
-argparser.add_argument('database')
-argparser.add_argument('--level', type=int, default=2)
-
-
-def main(args):
- if '://' not in args.database:
- args.database = "sqlite:///" + args.database
- engine, Session = opentuner.resultsdb.connect(args.database)
- session = Session()
-
- config_count = session.query(Configuration).count()
- # result_count = session.query(Result).count()
- # desired_result_count = session.query(DesiredResult).count()
-
- if args.level >= 1:
- q = (session.query(Configuration)
- .filter(~Configuration.id.in_(session.query(Result.configuration_id)
- .filter_by(was_new_best=True)
- .subquery()))
- .filter(Configuration.data != None))
-
- log.info("%s: compacted %d of %d Configurations",
- args.database,
- q.update({'data': None}, False),
- config_count)
- session.commit()
-
- if args.level >= 2:
- session.execute('VACUUM;')
- session.commit()
-
- log.info('done')
-
-
-if __name__ == '__main__':
- opentuner.tuningrunmain.init_logging()
- sys.exit(main(argparser.parse_args()))
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/dictconfig.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/dictconfig.py
deleted file mode 100644
index 7b835a4108..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/dictconfig.py
+++ /dev/null
@@ -1,544 +0,0 @@
-# This is a copy of the Python logging.config.dictconfig module,
-# reproduced with permission. It is provided here for backwards
-# compatibility for Python versions prior to 2.7.
-#
-# Copyright 2009-2010 by Vinay Sajip. All Rights Reserved.
-#
-# Permission to use, copy, modify, and distribute this software and its
-# documentation for any purpose and without fee is hereby granted,
-# provided that the above copyright notice appear in all copies and that
-# both that copyright notice and this permission notice appear in
-# supporting documentation, and that the name of Vinay Sajip
-# not be used in advertising or publicity pertaining to distribution
-# of the software without specific, written prior permission.
-# VINAY SAJIP DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING
-# ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL
-# VINAY SAJIP BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR
-# ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER
-# IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT
-# OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
-
-import logging.handlers
-import re
-import sys
-import types
-
-
-IDENTIFIER = re.compile('^[a-z_][a-z0-9_]*$', re.I)
-
-def valid_ident(s):
- m = IDENTIFIER.match(s)
- if not m:
- raise ValueError('Not a valid Python identifier: %r' % s)
- return True
-
-#
-# This function is defined in logging only in recent versions of Python
-#
-try:
- from logging import _checkLevel
-except ImportError:
- def _checkLevel(level):
- if isinstance(level, int):
- rv = level
- elif str(level) == level:
- if level not in logging._levelNames:
- raise ValueError('Unknown level: %r' % level)
- rv = logging._levelNames[level]
- else:
- raise TypeError('Level not an integer or a '
- 'valid string: %r' % level)
- return rv
-
-# The ConvertingXXX classes are wrappers around standard Python containers,
-# and they serve to convert any suitable values in the container. The
-# conversion converts base dicts, lists and tuples to their wrapped
-# equivalents, whereas strings which match a conversion format are converted
-# appropriately.
-#
-# Each wrapper should have a configurator attribute holding the actual
-# configurator to use for conversion.
-
-class ConvertingDict(dict):
- """A converting dictionary wrapper."""
-
- def __getitem__(self, key):
- value = dict.__getitem__(self, key)
- result = self.configurator.convert(value)
- #If the converted value is different, save for next time
- if value is not result:
- self[key] = result
- if type(result) in (ConvertingDict, ConvertingList,
- ConvertingTuple):
- result.parent = self
- result.key = key
- return result
-
- def get(self, key, default=None):
- value = dict.get(self, key, default)
- result = self.configurator.convert(value)
- #If the converted value is different, save for next time
- if value is not result:
- self[key] = result
- if type(result) in (ConvertingDict, ConvertingList,
- ConvertingTuple):
- result.parent = self
- result.key = key
- return result
-
- def pop(self, key, default=None):
- value = dict.pop(self, key, default)
- result = self.configurator.convert(value)
- if value is not result:
- if type(result) in (ConvertingDict, ConvertingList,
- ConvertingTuple):
- result.parent = self
- result.key = key
- return result
-
-class ConvertingList(list):
- """A converting list wrapper."""
- def __getitem__(self, key):
- value = list.__getitem__(self, key)
- result = self.configurator.convert(value)
- #If the converted value is different, save for next time
- if value is not result:
- self[key] = result
- if type(result) in (ConvertingDict, ConvertingList,
- ConvertingTuple):
- result.parent = self
- result.key = key
- return result
-
- def pop(self, idx=-1):
- value = list.pop(self, idx)
- result = self.configurator.convert(value)
- if value is not result:
- if type(result) in (ConvertingDict, ConvertingList,
- ConvertingTuple):
- result.parent = self
- return result
-
-class ConvertingTuple(tuple):
- """A converting tuple wrapper."""
- def __getitem__(self, key):
- value = tuple.__getitem__(self, key)
- result = self.configurator.convert(value)
- if value is not result:
- if type(result) in (ConvertingDict, ConvertingList,
- ConvertingTuple):
- result.parent = self
- result.key = key
- return result
-
-class BaseConfigurator(object):
- """
- The configurator base class which defines some useful defaults.
- """
-
- CONVERT_PATTERN = re.compile(r'^(?P<prefix>[a-z]+)://(?P<suffix>.*)$')
-
- WORD_PATTERN = re.compile(r'^\s*(\w+)\s*')
- DOT_PATTERN = re.compile(r'^\.\s*(\w+)\s*')
- INDEX_PATTERN = re.compile(r'^\[\s*(\w+)\s*\]\s*')
- DIGIT_PATTERN = re.compile(r'^\d+$')
-
- value_converters = {
- 'ext' : 'ext_convert',
- 'cfg' : 'cfg_convert',
- }
-
- # We might want to use a different one, e.g. importlib
- importer = __import__
-
- def __init__(self, config):
- self.config = ConvertingDict(config)
- self.config.configurator = self
-
- def resolve(self, s):
- """
- Resolve strings to objects using standard import and attribute
- syntax.
- """
- name = s.split('.')
- used = name.pop(0)
- try:
- found = self.importer(used)
- for frag in name:
- used += '.' + frag
- try:
- found = getattr(found, frag)
- except AttributeError:
- self.importer(used)
- found = getattr(found, frag)
- return found
- except ImportError:
- e, tb = sys.exc_info()[1:]
- v = ValueError('Cannot resolve %r: %s' % (s, e))
- v.__cause__, v.__traceback__ = e, tb
- raise v
-
- def ext_convert(self, value):
- """Default converter for the ext:// protocol."""
- return self.resolve(value)
-
- def cfg_convert(self, value):
- """Default converter for the cfg:// protocol."""
- rest = value
- m = self.WORD_PATTERN.match(rest)
- if m is None:
- raise ValueError("Unable to convert %r" % value)
- else:
- rest = rest[m.end():]
- d = self.config[m.groups()[0]]
- #print d, rest
- while rest:
- m = self.DOT_PATTERN.match(rest)
- if m:
- d = d[m.groups()[0]]
- else:
- m = self.INDEX_PATTERN.match(rest)
- if m:
- idx = m.groups()[0]
- if not self.DIGIT_PATTERN.match(idx):
- d = d[idx]
- else:
- try:
- n = int(idx) # try as number first (most likely)
- d = d[n]
- except TypeError:
- d = d[idx]
- if m:
- rest = rest[m.end():]
- else:
- raise ValueError('Unable to convert '
- '%r at %r' % (value, rest))
- #rest should be empty
- return d
-
- def convert(self, value):
- """
- Convert values to an appropriate type. dicts, lists and tuples are
- replaced by their converting alternatives. Strings are checked to
- see if they have a conversion format and are converted if they do.
- """
- if not isinstance(value, ConvertingDict) and isinstance(value, dict):
- value = ConvertingDict(value)
- value.configurator = self
- elif not isinstance(value, ConvertingList) and isinstance(value, list):
- value = ConvertingList(value)
- value.configurator = self
- elif not isinstance(value, ConvertingTuple) and\
- isinstance(value, tuple):
- value = ConvertingTuple(value)
- value.configurator = self
- return value
-
- def configure_custom(self, config):
- """Configure an object with a user-supplied factory."""
- c = config.pop('()')
- if not hasattr(c, '__call__') and hasattr(types, 'ClassType') and type(c) != types.ClassType:
- c = self.resolve(c)
- props = config.pop('.', None)
- # Check for valid identifiers
- kwargs = dict([(k, config[k]) for k in config if valid_ident(k)])
- result = c(**kwargs)
- if props:
- for name, value in props.items():
- setattr(result, name, value)
- return result
-
- def as_tuple(self, value):
- """Utility function which converts lists to tuples."""
- if isinstance(value, list):
- value = tuple(value)
- return value
-
-class DictConfigurator(BaseConfigurator):
- """
- Configure logging using a dictionary-like object to describe the
- configuration.
- """
-
- def configure(self):
- """Do the configuration."""
-
- config = self.config
- if 'version' not in config:
- raise ValueError("dictionary doesn't specify a version")
- if config['version'] != 1:
- raise ValueError("Unsupported version: %s" % config['version'])
- incremental = config.pop('incremental', False)
- EMPTY_DICT = {}
- logging._acquireLock()
- try:
- if incremental:
- handlers = config.get('handlers', EMPTY_DICT)
- # incremental handler config only if handler name
- # ties in to logging._handlers (Python 2.7)
- if sys.version_info[:2] == (2, 7):
- for name in handlers:
- if name not in logging._handlers:
- raise ValueError('No handler found with '
- 'name %r' % name)
- else:
- try:
- handler = logging._handlers[name]
- handler_config = handlers[name]
- level = handler_config.get('level', None)
- if level:
- handler.setLevel(_checkLevel(level))
- except StandardError as e:
- raise ValueError('Unable to configure handler '
- '%r: %s' % (name, e))
- loggers = config.get('loggers', EMPTY_DICT)
- for name in loggers:
- try:
- self.configure_logger(name, loggers[name], True)
- except StandardError as e:
- raise ValueError('Unable to configure logger '
- '%r: %s' % (name, e))
- root = config.get('root', None)
- if root:
- try:
- self.configure_root(root, True)
- except StandardError as e:
- raise ValueError('Unable to configure root '
- 'logger: %s' % e)
- else:
- disable_existing = config.pop('disable_existing_loggers', True)
-
- logging._handlers.clear()
- del logging._handlerList[:]
-
- # Do formatters first - they don't refer to anything else
- formatters = config.get('formatters', EMPTY_DICT)
- for name in formatters:
- try:
- formatters[name] = self.configure_formatter(
- formatters[name])
- except StandardError as e:
- raise ValueError('Unable to configure '
- 'formatter %r: %s' % (name, e))
- # Next, do filters - they don't refer to anything else, either
- filters = config.get('filters', EMPTY_DICT)
- for name in filters:
- try:
- filters[name] = self.configure_filter(filters[name])
- except StandardError as e:
- raise ValueError('Unable to configure '
- 'filter %r: %s' % (name, e))
-
- # Next, do handlers - they refer to formatters and filters
- # As handlers can refer to other handlers, sort the keys
- # to allow a deterministic order of configuration
- handlers = config.get('handlers', EMPTY_DICT)
- for name in sorted(handlers):
- try:
- handler = self.configure_handler(handlers[name])
- handler.name = name
- handlers[name] = handler
- except StandardError as e:
- raise ValueError('Unable to configure handler '
- '%r: %s' % (name, e))
- # Next, do loggers - they refer to handlers and filters
-
- #we don't want to lose the existing loggers,
- #since other threads may have pointers to them.
- #existing is set to contain all existing loggers,
- #and as we go through the new configuration we
- #remove any which are configured. At the end,
- #what's left in existing is the set of loggers
- #which were in the previous configuration but
- #which are not in the new configuration.
- root = logging.root
- existing = list(root.manager.loggerDict)
- #The list needs to be sorted so that we can
- #avoid disabling child loggers of explicitly
- #named loggers. With a sorted list it is easier
- #to find the child loggers.
- existing.sort()
- #We'll keep the list of existing loggers
- #which are children of named loggers here...
- child_loggers = []
- #now set up the new ones...
- loggers = config.get('loggers', EMPTY_DICT)
- for name in loggers:
- if name in existing:
- i = existing.index(name)
- prefixed = name + "."
- pflen = len(prefixed)
- num_existing = len(existing)
- i = i + 1 # look at the entry after name
- while (i < num_existing) and\
- (existing[i][:pflen] == prefixed):
- child_loggers.append(existing[i])
- i = i + 1
- existing.remove(name)
- try:
- self.configure_logger(name, loggers[name])
- except StandardError as e:
- raise ValueError('Unable to configure logger '
- '%r: %s' % (name, e))
-
- #Disable any old loggers. There's no point deleting
- #them as other threads may continue to hold references
- #and by disabling them, you stop them doing any logging.
- #However, don't disable children of named loggers, as that's
- #probably not what was intended by the user.
- for log in existing:
- logger = root.manager.loggerDict[log]
- if log in child_loggers:
- logger.level = logging.NOTSET
- logger.handlers = []
- logger.propagate = True
- elif disable_existing:
- logger.disabled = True
-
- # And finally, do the root logger
- root = config.get('root', None)
- if root:
- try:
- self.configure_root(root)
- except StandardError as e:
- raise ValueError('Unable to configure root '
- 'logger: %s' % e)
- finally:
- logging._releaseLock()
-
- def configure_formatter(self, config):
- """Configure a formatter from a dictionary."""
- if '()' in config:
- factory = config['()'] # for use in exception handler
- try:
- result = self.configure_custom(config)
- except TypeError as te:
- if "'format'" not in str(te):
- raise
- #Name of parameter changed from fmt to format.
- #Retry with old name.
- #This is so that code can be used with older Python versions
- #(e.g. by Django)
- config['fmt'] = config.pop('format')
- config['()'] = factory
- result = self.configure_custom(config)
- else:
- fmt = config.get('format', None)
- dfmt = config.get('datefmt', None)
- result = logging.Formatter(fmt, dfmt)
- return result
-
- def configure_filter(self, config):
- """Configure a filter from a dictionary."""
- if '()' in config:
- result = self.configure_custom(config)
- else:
- name = config.get('name', '')
- result = logging.Filter(name)
- return result
-
- def add_filters(self, filterer, filters):
- """Add filters to a filterer from a list of names."""
- for f in filters:
- try:
- filterer.addFilter(self.config['filters'][f])
- except StandardError as e:
- raise ValueError('Unable to add filter %r: %s' % (f, e))
-
- def configure_handler(self, config):
- """Configure a handler from a dictionary."""
- formatter = config.pop('formatter', None)
- if formatter:
- try:
- formatter = self.config['formatters'][formatter]
- except StandardError as e:
- raise ValueError('Unable to set formatter '
- '%r: %s' % (formatter, e))
- level = config.pop('level', None)
- filters = config.pop('filters', None)
- if '()' in config:
- c = config.pop('()')
- if not hasattr(c, '__call__') and hasattr(types, 'ClassType') and type(c) != types.ClassType:
- c = self.resolve(c)
- factory = c
- else:
- klass = self.resolve(config.pop('class'))
- #Special case for handler which refers to another handler
- if issubclass(klass, logging.handlers.MemoryHandler) and\
- 'target' in config:
- try:
- config['target'] = self.config['handlers'][config['target']]
- except StandardError as e:
- raise ValueError('Unable to set target handler '
- '%r: %s' % (config['target'], e))
- elif issubclass(klass, logging.handlers.SMTPHandler) and\
- 'mailhost' in config:
- config['mailhost'] = self.as_tuple(config['mailhost'])
- elif issubclass(klass, logging.handlers.SysLogHandler) and\
- 'address' in config:
- config['address'] = self.as_tuple(config['address'])
- factory = klass
- kwargs = dict([(k, config[k]) for k in config if valid_ident(k)])
- try:
- result = factory(**kwargs)
- except TypeError as te:
- if "'stream'" not in str(te):
- raise
- #The argument name changed from strm to stream
- #Retry with old name.
- #This is so that code can be used with older Python versions
- #(e.g. by Django)
- kwargs['strm'] = kwargs.pop('stream')
- result = factory(**kwargs)
- if formatter:
- result.setFormatter(formatter)
- if level is not None:
- result.setLevel(_checkLevel(level))
- if filters:
- self.add_filters(result, filters)
- return result
-
- def add_handlers(self, logger, handlers):
- """Add handlers to a logger from a list of names."""
- for h in handlers:
- try:
- logger.addHandler(self.config['handlers'][h])
- except StandardError as e:
- raise ValueError('Unable to add handler %r: %s' % (h, e))
-
- def common_logger_config(self, logger, config, incremental=False):
- """
- Perform configuration which is common to root and non-root loggers.
- """
- level = config.get('level', None)
- if level is not None:
- logger.setLevel(_checkLevel(level))
- if not incremental:
- #Remove any existing handlers
- for h in logger.handlers[:]:
- logger.removeHandler(h)
- handlers = config.get('handlers', None)
- if handlers:
- self.add_handlers(logger, handlers)
- filters = config.get('filters', None)
- if filters:
- self.add_filters(logger, filters)
-
- def configure_logger(self, name, config, incremental=False):
- """Configure a non-root logger from a dictionary."""
- logger = logging.getLogger(name)
- self.common_logger_config(logger, config, incremental)
- propagate = config.get('propagate', None)
- if propagate is not None:
- logger.propagate = propagate
-
- def configure_root(self, config, incremental=False):
- """Configure a root logger from a dictionary."""
- root = logging.getLogger()
- self.common_logger_config(root, config, incremental)
-
-dictConfigClass = DictConfigurator
-
-def dictConfig(config):
- """Configure logging using a dictionary."""
- dictConfigClass(config).configure()
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/stats.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/stats.py
deleted file mode 100755
index 99449c8a90..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/stats.py
+++ /dev/null
@@ -1,468 +0,0 @@
-#!/usr/bin/env python
-
-if __name__ == '__main__':
- import adddeps
-
-import argparse
-import csv
-import hashlib
-import itertools
-import logging
-import math
-import os
-import sqlalchemy.orm.exc
-import subprocess
-import sys
-
-from collections import defaultdict
-from fn import _
-from fn import Stream
-from fn.iters import repeat
-from pprint import pprint
-
-import opentuner
-from opentuner import resultsdb
-from opentuner.resultsdb.models import *
-
-log = logging.getLogger('opentuner.utils.stats')
-
-argparser = argparse.ArgumentParser()
-argparser.add_argument('--label')
-argparser.add_argument('--stats', action='store_true',
- help="run in stats mode")
-argparser.add_argument('--by-request-count', action='store_true',
- help='report stats by request count')
-argparser.add_argument('--stats-quanta', type=float, default=10,
- help="step size in seconds for binning with --stats")
-argparser.add_argument('--stats-dir', default='stats',
- help="directory to output --stats to")
-argparser.add_argument('--stats-input', default="opentuner.db")
-argparser.add_argument('--min-runs', type=int, default=1,
- help="ignore series with less then N runs")
-
-PCTSTEPS = map(_/20.0, xrange(21))
-
-def mean(vals):
- n = 0.0
- d = 0.0
- for v in vals:
- if v is not None:
- n += v
- d += 1.0
- if d == 0.0:
- return None
- return n/d
-
-def median(vals):
- vals = sorted(vals)
- a = (len(vals)-1)/2
- b = (len(vals))/2
- return (vals[a]+vals[b])/2.0
-
-def percentile(vals, pct):
- vals = sorted(vals)
- pos = (len(vals)-1) * pct
- a = int(math.floor(pos))
- b = min(len(vals) - 1, a + 1)
- return (1.0-(pos-a))*vals[a] + (pos-a)*vals[b]
-
-def variance(vals):
- vals = filter(lambda x: x is not None, vals)
- avg = mean(vals)
- if avg is None:
- return None
- if avg in (float('inf'), float('-inf')):
- return avg
- return mean(map((_ - avg) ** 2, vals))
-
-def stddev(vals):
- var = variance(vals)
- if var is None:
- return None
- return math.sqrt(var)
-
-def hash_args(x):
- d = dict(vars(x))
- for k in ('database', 'results_log', 'results_log_details'):
- d[k] = None
- return hashlib.sha256(str(sorted(d.items()))).hexdigest()[:20]
-
-def run_label(tr, short = False):
- techniques = ','.join(tr.args.technique)
- if not tr.name or tr.name=='unnamed':
- if short:
- return techniques
- else:
- return "%s_%s" % (techniques, hash_args(tr.args)[:6])
- else:
- return tr.name
-
-def run_dir(base, tr):
- return os.path.join(base,
- tr.program.project,
- tr.program.name.split('/')[-1],
- tr.program_version.version[:16])
-
-class StatsMain(object):
- def __init__(self, args):
- self.args = args
- path = args.stats_input
- self.dbs = list()
- for f in os.listdir(path):
- if 'journal' in f:
- continue
- try:
- e, sm = resultsdb.connect('sqlite:///'+os.path.join(path, f))
- self.dbs.append(sm())
- except:
- log.error('failed to load database: %s',
- os.path.join(path, f),
- exc_info=True)
-
- def main(self):
- dir_label_runs = defaultdict(lambda: defaultdict(list))
- for session in self.dbs:
- q = (session.query(resultsdb.models.TuningRun)
- .filter_by(state='COMPLETE')
- .order_by('name'))
-
- if self.args.label:
- q = q.filter(TuningRun.name.in_(
- map(str.strip,self.args.label.split(','))))
-
- for tr in q:
- d = run_dir(self.args.stats_dir, tr)
- d = os.path.normpath(d)
- dir_label_runs[d][run_label(tr)].append((tr, session))
-
- summary_report = defaultdict(lambda: defaultdict(list))
- for d, label_runs in dir_label_runs.iteritems():
- if not os.path.isdir(d):
- os.makedirs(d)
- session = label_runs.values()[0][0][1]
- objective = label_runs.values()[0][0][0].objective
- all_run_ids = map(_[0].id, itertools.chain(*label_runs.values()))
- q = (session.query(Result)
- .filter(Result.tuning_run_id.in_(all_run_ids))
- .filter(Result.time < float('inf'))
- .filter_by(was_new_best=True, state='OK'))
- total = q.count()
- if total == 0:
- continue
- q = objective.filter_acceptable(q)
- acceptable = q.count()
- q = q.order_by(*objective.result_order_by_terms())
- best = q.limit(1).one()
- worst = q.offset(acceptable-1).limit(1).one()
-
- map(len, label_runs.values())
-
- log.info("%s -- best %.4f / worst %.f4 "
- "-- %d of %d acceptable -- %d techniques with %d to %d runs",
- d,
- best.time,
- worst.time,
- acceptable,
- total,
- len(label_runs.values()),
- min(map(len, label_runs.values())),
- max(map(len, label_runs.values())))
-
- for label, runs in sorted(label_runs.items()):
- if len(runs) < self.args.min_runs:
- print len(runs) ,self.args.min_runs
- continue
- log.debug('%s/%s has %d runs %s',d, label, len(runs), runs[0][0].args.technique)
- self.combined_stats_over_time(d, label, runs, objective, worst, best)
-
- final_scores = list()
- for run, session in runs:
- try:
- final = (session.query(Result)
- .filter_by(tuning_run=run,
- configuration=run.final_config)
- .limit(1)
- .one())
- except sqlalchemy.orm.exc.NoResultFound:
- continue
- final_scores.append(objective.stats_quality_score(final, worst, best))
- final_scores.sort()
- if final_scores:
- norm = objective.stats_quality_score(best, worst, best)
- if norm > 0.00001:
- summary_report[d][run_label(run, short=True)] = (
- percentile(final_scores, 0.5) / norm,
- percentile(final_scores, 0.1) / norm,
- percentile(final_scores, 0.9) / norm,
- )
- else:
- summary_report[d][run_label(run, short=True)] = (
- percentile(final_scores, 0.5) + norm + 1.0,
- percentile(final_scores, 0.1) + norm + 1.0,
- percentile(final_scores, 0.9) + norm + 1.0,
- )
-
-
- with open(self.args.stats_dir+ "/summary.dat", 'w') as o:
- # make summary report
- keys = sorted(reduce(set.union,
- [set(x.keys()) for x in summary_report.values()],
- set()))
- print >>o, '#####',
- for k in keys:
- print >>o, k,
- print >>o
- for d, label_vals in sorted(summary_report.items()):
- print >>o, d.split('/')[-2],
- for k in keys:
- if k in label_vals:
- print >>o, '-', label_vals[k][0], label_vals[k][1], label_vals[k][2],
- else:
- print >>o, '-', '-', '-', '-',
- print >>o
-
- if keys:
- plotcmd = ["""1 w lines lt 1 lc rgb "black" notitle""",
- """'summary.dat' using 3:4:5:xtic(1) ti "%s" """ % keys[0]]
- for n, k in enumerate(keys[1:]):
- plotcmd.append("""'' using %d:%d:%d ti "%s" """ % (
- 4*n + 7,
- 4*n + 8,
- 4*n + 9,
- k))
- self.gnuplot_summary_file(self.args.stats_dir, 'summary', plotcmd)
-
-
-
- for d, label_runs in dir_label_runs.iteritems():
- labels = [k for k,v in label_runs.iteritems()
- if len(v)>=self.args.min_runs]
- self.gnuplot_file(d,
- "medianperfe",
- ['"%s_percentiles.dat" using 1:12:4:18 with errorbars title "%s"' % (l,l) for l in labels])
- self.gnuplot_file(d,
- "meanperfe",
- ['"%s_percentiles.dat" using 1:21:4:18 with errorbars title "%s"' % (l,l) for l in labels])
- self.gnuplot_file(d,
- "medianperfl",
- ['"%s_percentiles.dat" using 1:12 with lines title "%s"' % (l,l) for l in labels])
- self.gnuplot_file(d,
- "meanperfl",
- ['"%s_percentiles.dat" using 1:21 with lines title "%s"' % (l,l) for l in labels])
-
- # print
- # print "10% Scores", d
- # pprint(self.technique_scores(d, labels, '0.1'))
- # print
- # print "90% Scores", d
- # pprint(self.technique_scores(d, labels, '0.9'))
- # print
- # print "Mean Scores", d
- # pprint(self.technique_scores(d, labels, 'mean'))
- print
- print "Median Scores", d
- pprint(self.technique_scores(d, labels, '0.5'))
-
-
- def technique_scores(self, directory, labels, ykey, xkey='#sec', factor=10.0):
- max_duration = None
- min_value = float('inf')
- for label in labels:
- try:
- dr = csv.DictReader(open(os.path.join(directory,label+"_percentiles.dat")), delimiter=' ', lineterminator='\n')
- lastrow = list(dr)[-1]
- max_duration = max(max_duration, float(lastrow[xkey]))
- min_value = min(min_value, float(lastrow[ykey]))
- except:
- log.exception("failed computing score")
-
- scores = list()
-
- for label in labels:
- try:
- dr = csv.DictReader(open(os.path.join(directory,label+"_percentiles.dat")), delimiter=' ', lineterminator='\n')
- score = 0.0
- lastsec = 0.0
- value = float('inf')
- for row in dr:
- duration = float(row[xkey]) - lastsec
- lastsec = float(row[xkey])
- value = float(row[ykey])
- score += duration * (value - min_value)
- score += (factor*max_duration - lastsec) * (value - min_value)
- scores.append((score, label))
- except:
- log.exception("failed computing score")
-
- return sorted(scores)
-
-
- def combined_stats_over_time(self,
- output_dir,
- label,
- runs,
- objective,
- worst,
- best,
- ):
- """
- combine stats_over_time() vectors for multiple runs
- """
-
- #extract_fn = lambda dr: objective.stats_quality_score(dr.result, worst, best)
- extract_fn = _.result.time
- combine_fn = min
- no_data = 999
-
- log.debug("writing stats for %s to %s", label, output_dir)
- by_run = [self.stats_over_time(session, run, extract_fn, combine_fn, no_data)
- for run, session in runs]
- max_len = max(map(len, by_run))
-
- by_run_streams = [Stream() << x << repeat(x[-1], max_len-len(x))
- for x in by_run]
- by_quanta = zip(*by_run_streams[:])
-
- def data_file(suffix, headers, value_function):
- with open(os.path.join(output_dir, label+suffix), 'w') as fd:
- out = csv.writer(fd, delimiter=' ', lineterminator='\n')
- out.writerow(['#sec'] + headers)
- for quanta, values in enumerate(by_quanta):
- sec = quanta*self.args.stats_quanta
- out.writerow([sec] + value_function(values))
-
- #data_file('_details.dat',
- # map(lambda x: 'run%d'%x, xrange(max_len)),
- # list)
- #self.gnuplot_file(output_dir,
- # label+'_details',
- # [('"'+label+'_details.dat"'
- # ' using 1:%d'%i +
- # ' with lines'
- # ' title "Run %d"'%i)
- # for i in xrange(max_len)])
-
- data_file('_mean.dat',
- ['#sec', 'mean', 'stddev'],
- lambda values: [mean(values), stddev(values)])
- self.gnuplot_file(output_dir,
- label+'_mean',
- ['"'+label+'_mean.dat" using 1:2 with lines title "Mean"'])
-
- def extract_percentiles(values):
- values = sorted(values)
- return ([values[int(round(p*(len(values)-1)))] for p in PCTSTEPS]
- + [mean(values)])
- data_file("_percentiles.dat", PCTSTEPS + ['mean'], extract_percentiles)
- self.gnuplot_file(output_dir,
- label+'_percentiles',
- reversed([
- '"'+label+'_percentiles.dat" using 1:2 with lines title "0%"',
- # '"" using 1:3 with lines title "5%"',
- '"" using 1:4 with lines title "10%"',
- # '"" using 1:5 with lines title "25%"',
- '"" using 1:6 with lines title "20%"',
- # '"" using 1:7 with lines title "35%"',
- '"" using 1:8 with lines title "30%"',
- # '"" using 1:9 with lines title "45%"',
- '"" using 1:10 with lines title "40%"',
- # '"" using 1:11 with lines title "55%"',
- '"" using 1:12 with lines title "50%"',
- # '"" using 1:13 with lines title "65%"',
- '"" using 1:14 with lines title "70%"',
- # '"" using 1:15 with lines title "75%"',
- '"" using 1:16 with lines title "80%"',
- # '"" using 1:17 with lines title "85%"',
- '"" using 1:18 with lines title "90%"',
- # '"" using 1:19 with lines title "95%"',
- '"'+label+'_percentiles.dat" using 1:20 with lines title "100%"',
- ]))
-
- def gnuplot_file(self, output_dir, prefix, plotcmd):
- with open(os.path.join(output_dir, prefix+'.gnuplot'), 'w') as fd:
- print >>fd, 'set terminal postscript eps enhanced color'
- print >>fd, 'set output "%s"' % (prefix+'.eps')
- print >>fd, 'set ylabel "Execution Time (seconds)"'
- print >>fd, 'set xlabel "Autotuning Time (seconds)"'
- print >>fd, 'plot', ',\\\n'.join(plotcmd)
-
- try:
- subprocess.call(['gnuplot', prefix+'.gnuplot'], cwd=output_dir, stdin=None)
- except OSError:
- log.error("command gnuplot not found")
-
- def gnuplot_summary_file(self, output_dir, prefix, plotcmd):
- with open(os.path.join(output_dir, prefix+'.gnuplot'), 'w') as fd:
- print >>fd, 'set terminal postscript eps enhanced color'
- print >>fd, 'set output "%s"' % (prefix+'.eps')
- print >>fd, '''
-set boxwidth 0.9
-set style fill solid 1.00 border 0
-set style histogram errorbars gap 2 lw 1
-set style data histograms
-set xtics rotate by -45
-set bars 0.5
-set yrange [0:20]
-
-set yrange [0:10]
-set key out vert top left
-set size 1.5,1
-set ytics 1
-
-'''
- print >>fd, 'plot', ',\\\n'.join(plotcmd)
- subprocess.call(['gnuplot', prefix+'.gnuplot'], cwd=output_dir, stdin=None)
-
-
- def stats_over_time(self,
- session,
- run,
- extract_fn,
- combine_fn,
- no_data = None):
- """
- return reduce(combine_fn, map(extract_fn, data)) for each quanta of the
- tuning run
- """
- value_by_quanta = [ no_data ]
- start_date = run.start_date
-
- subq = (session.query(Result.id)
- .filter_by(tuning_run = run, was_new_best = True, state='OK'))
-
- q = (session.query(DesiredResult)
- .join(Result)
- .filter(DesiredResult.state=='COMPLETE',
- DesiredResult.tuning_run == run,
- DesiredResult.result_id.in_(subq.subquery()))
- .order_by(DesiredResult.request_date))
-
- first_id = None
- for dr in q:
- if first_id is None:
- first_id = dr.id
- td = (dr.request_date - start_date)
- duration = td.seconds + (td.days * 24 * 3600.0)
- if self.args.by_request_count:
- quanta = dr.id - first_id
- else:
- quanta = int(duration / self.args.stats_quanta)
- while len(value_by_quanta) <= quanta:
- value_by_quanta.append(value_by_quanta[-1])
-
- if value_by_quanta[-1] is no_data:
- value_by_quanta[-1] = extract_fn(dr)
- else:
- value_by_quanta[-1] = combine_fn(value_by_quanta[-1], extract_fn(dr))
-
- return value_by_quanta
-
-
-
-
-
-if __name__ == '__main__':
- opentuner.tuningrunmain.init_logging()
- sys.exit(StatsMain(argparser.parse_args()).main())
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/stats_matplotlib.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/stats_matplotlib.py
deleted file mode 100644
index 54e9132a66..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/opentuner/utils/stats_matplotlib.py
+++ /dev/null
@@ -1,290 +0,0 @@
-#!usr/bin/python
-
-if __name__ == '__main__':
- import adddeps
-
-import itertools
-import math
-import matplotlib.pyplot as plt
-import numpy
-import os
-import sqlalchemy
-import sqlalchemy.orm.exc
-
-from collections import defaultdict
-from fn import _
-from fn import Stream
-from fn.iters import repeat
-from opentuner import resultsdb
-
-PCTSTEPS = map(_/20.0, xrange(21))
-
-
-def mean(vals):
- """
- Arguments,
- vals: List of floating point numbers
- Returns,
- The mean of the numbers in the input list
- None if all values in the list are None
- """
- filtered_values = [float(x) for x in vals if x is not None]
- if len(filtered_values) == 0:
- return None
- return numpy.mean(numpy.array(filtered_values))
-
-
-def stddev(vals):
- """
- Arguments,
- vals: List of floating point numbers
- Returns,
- The standard deviation of numbers in the input list
- None if all values in the list are None
- """
- filtered_values = [float(x) for x in vals if x is not None]
- if len(filtered_values) == 0:
- return None
- return math.sqrt(numpy.var(numpy.array(filtered_values)))
-
-
-def get_dbs(path, db_type='sqlite:///'):
- """
- Arguments,
- path: Path of directory containing .db files
- Returns,
- A list of (engine, session) pairs to the dbs pointed to by
- the db files
- """
- dbs = list()
- for f in os.listdir(path):
- if 'journal' in f:
- continue
- try:
- db_path = os.path.join(path, f)
- e, sm = resultsdb.connect(db_type + db_path)
- dbs.append(sm())
- except Exception as e:
- print e
- print "Error encountered while connecting to db"
- return dbs
-
-
-def matplotlibplot_file(labels, xlim = None, ylim = None, disp_types=['median']):
- """
- Arguments,
- labels: List of labels that need to be included in the plot
- xlim: Integer denoting the maximum X-coordinate in the plot
- ylim: Integer denoting the maximum Y-coordinate in the plot
- disp_types: List of measures that are to be displayed in the plot
- Returns,
- A figure object representing the required plot
- """
-
- figure = plt.figure()
- values = get_values(labels)
- for label in values:
- (mean_values, percentile_values) = values[label]
- for disp_type in disp_types:
- cols = None
- data = percentile_values
-
- if disp_type == 'median':
- cols = [11]
- elif disp_type == 'mean':
- cols = [1]
- data = mean_values
- elif disp_type == 'all_percentiles':
- cols = range(1,22)
-
- plotted_data = [[] for x in xrange(len(cols))]
-
- x_indices = []
- for data_point in data[1:]:
- x_indices.append(int(data_point[0]))
- for i in range(0, len(cols)):
- plotted_data[i].append(float(data_point[cols[i]]))
- args = []
- for to_plot in plotted_data:
- args.append(x_indices)
- args.append(to_plot)
-
- plt.plot(*args, label='%s(%s)' % (label, disp_type))
-
- if xlim is not None:
- plt.xlim(xlim)
- if ylim is not None:
- plt.ylim(ylim)
-
- plt.xlabel('Autotuning Time (seconds)')
- plt.ylabel('Execution Time (seconds)')
- plt.legend(loc='upper right')
- return figure
-
-
-def run_label(tr):
- techniques = ','.join(tr.args.technique)
- if not tr.name or tr.name == 'unnamed':
- return techniques
- return tr.name
-
-
-def combined_stats_over_time(label,
- runs,
- objective,
- worst,
- best,
- ):
- """
- combine stats_over_time() vectors for multiple runs
- """
-
- extract_fn = _.result.time
- combine_fn = min
- no_data = 999
-
- by_run = [stats_over_time(session, run, extract_fn, combine_fn, no_data)
- for run, session in runs]
- max_len = max(map(len, by_run))
-
- by_run_streams = [Stream() << x << repeat(x[-1], max_len-len(x))
- for x in by_run]
- by_quanta = zip(*by_run_streams[:])
-
- # TODO: Fix this, this variable should be configurable
- stats_quanta = 10
- def get_data(value_function):
- final_values = []
- for quanta, values in enumerate(by_quanta):
- sec = quanta*stats_quanta
- final_values.append([sec] + value_function(values))
- return final_values
-
- mean_values = get_data(lambda values: [mean(values), stddev(values)])
-
- def extract_percentiles(values):
- values = sorted(values)
- return ([values[int(round(p*(len(values)-1)))] for p in PCTSTEPS]
- + [mean(values)])
- percentile_values = get_data(extract_percentiles)
- return mean_values, percentile_values
-
-
-def stats_over_time(session,
- run,
- extract_fn,
- combine_fn,
- no_data = None):
- """
- return reduce(combine_fn, map(extract_fn, data)) for each quanta of the
- tuning run
- """
- value_by_quanta = [ no_data ]
- start_date = run.start_date
-
- subq = (session.query(resultsdb.models.Result.id)
- .filter_by(tuning_run = run, was_new_best = True, state='OK'))
-
- q = (session.query(resultsdb.models.DesiredResult)
- .join(resultsdb.models.Result)
- .filter(resultsdb.models.DesiredResult.state=='COMPLETE',
- resultsdb.models.DesiredResult.tuning_run == run,
- resultsdb.models.DesiredResult.result_id.in_(subq.subquery()))
- .order_by(resultsdb.models.DesiredResult.request_date))
-
- first_id = None
- for dr in q:
- if first_id is None:
- first_id = dr.id
- td = (dr.request_date - start_date)
- duration = td.seconds + (td.days * 24 * 3600.0)
- # TODO: Make this variable configurable
- by_request_count = True
- stats_quanta = 10
- if by_request_count:
- quanta = dr.id - first_id
- else:
- quanta = int(duration / stats_quanta)
- while len(value_by_quanta) <= quanta:
- value_by_quanta.append(value_by_quanta[-1])
-
- if value_by_quanta[-1] is no_data:
- value_by_quanta[-1] = extract_fn(dr)
- else:
- value_by_quanta[-1] = combine_fn(value_by_quanta[-1], extract_fn(dr))
-
- return value_by_quanta
-
-
-def get_all_labels():
- """
- Returns,
- List of labels that are in the complete state
- """
- dbs = get_dbs(os.getcwd())
- all_labels = list()
- for db in dbs:
- all_labels.extend(db.query(resultsdb.models.TuningRun.name)
- .filter_by(state='COMPLETE')
- .distinct()
- .all())
- all_labels = [str(element[0]) for element in all_labels]
- return all_labels
-
-
-def get_values(labels):
- """
- Arguments,
- labels: List of labels whose values are of interest
- Returns,
- A list of (mean, percentile) tuples, corresponding to the
- provided list of labels
- """
- dbs = get_dbs(os.getcwd())
- dir_label_runs = defaultdict(lambda: defaultdict(list))
- for db in dbs:
- q = (db.query(resultsdb.models.TuningRun)
- .filter_by(state='COMPLETE')
- .order_by('name'))
- if labels:
- q = q.filter(resultsdb.models.TuningRun.name.in_(labels))
- for tr in q:
- dir_label_runs[run_label(tr)][run_label(tr)].append((tr, db))
- all_run_ids = list()
- returned_values = {}
- for d, label_runs in dir_label_runs.iteritems():
- all_run_ids = map(_[0].id, itertools.chain(*label_runs.values()))
- session = label_runs.values()[0][0][1]
- objective = label_runs.values()[0][0][0].objective
-
- q = (session.query(resultsdb.models.Result)
- .filter(resultsdb.models.Result.tuning_run_id.in_(all_run_ids))
- .filter(resultsdb.models.Result.time < float('inf'))
- .filter_by(was_new_best=True, state='OK'))
- total = q.count()
- q = objective.filter_acceptable(q)
- acceptable = q.count()
- q = q.order_by(*objective.result_order_by_terms())
- best = q.limit(1).one()
- worst = q.offset(acceptable - 1).limit(1).one()
-
- for label, runs in sorted(label_runs.items()):
- (mean_values, percentile_values) = combined_stats_over_time(label, runs, objective, worst, best)
- returned_values[label] = (mean_values, percentile_values)
- final_scores = list()
- for run, session in runs:
- try:
- final = (session.query(resultsdb.models.Result)
- .filter_by(tuning_run = run,
- configuration = run.final_config)
- .limit(1).one())
- except sqlalchemy.orm.exc.NoResultFound:
- continue
- final_scores.append(objective.stats_quality_score(final, worst, best))
- final_scores.sort()
- return returned_values
-
-if __name__ == '__main__':
- labels = [u'timeouts', u'always_reorder', u'add_store_at', u'all_options']
- get_values(labels)
- print get_all_labels()
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/optional-requirements.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/optional-requirements.txt
deleted file mode 100644
index 9848f674cb..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/optional-requirements.txt
+++ /dev/null
@@ -1,3 +0,0 @@
-django==1.6.1
-matplotlib==1.1.1
-virtualenv==1.9.1
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/requirements.txt b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/requirements.txt
deleted file mode 100644
index fa9cfeca2e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/requirements.txt
+++ /dev/null
@@ -1,5 +0,0 @@
-argparse>=1.2.1
-fn>=0.2.12
-numpy>=1.8.0
-pysqlite>=2.6.3
-SQLAlchemy>=0.8.2
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/setup.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/setup.py
deleted file mode 100755
index 633d4359d9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/setup.py
+++ /dev/null
@@ -1,33 +0,0 @@
-#!/usr/bin/python
-try:
- from setuptools import setup
-except ImportError:
- try:
- from setuptools.core import setup
- except ImportError:
- from distutils.core import setup
-
-try:
- from pypandoc import convert
- read_md = lambda f: convert(f, 'rest')
-except ImportError:
- print("warning: pypandoc module not found, could not convert Markdown to RST")
- read_md = lambda f: open(f, 'r').read()
-
-required = open('requirements.txt').read().splitlines()
-required = [l.strip() for l in required
- if l.strip() and not l.strip().startswith('#')]
-
-setup(
- name='opentuner',
- version='0.8.0',
- url='http://opentuner.org/',
- license='MIT',
- author='Jason Ansel',
- author_email='jansel@jansel.net',
- description='An extensible framework for program autotuning',
- long_description=read_md('README.md'),
- packages=['opentuner', 'opentuner.resultsdb', 'opentuner.utils',
- 'opentuner.measurement', 'opentuner.search'],
- install_requires=required,
-)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/manage.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/manage.py
deleted file mode 100644
index f27b5b8db1..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/manage.py
+++ /dev/null
@@ -1,10 +0,0 @@
-#!/usr/bin/env python
-import os
-import sys
-
-if __name__ == "__main__":
- os.environ.setdefault("DJANGO_SETTINGS_MODULE", "stats_app.settings")
-
- from django.core.management import execute_from_command_line
-
- execute_from_command_line(sys.argv)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/__init__.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/__init__.py
deleted file mode 100644
index e69de29bb2..0000000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/settings.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/settings.py
deleted file mode 100644
index 09505be03e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/settings.py
+++ /dev/null
@@ -1,162 +0,0 @@
-# Django settings for stats_app project.
-import os
-
-DEBUG = True
-TEMPLATE_DEBUG = DEBUG
-
-ADMINS = (
- # ('Your Name', 'your_email@example.com'),
-)
-
-MANAGERS = ADMINS
-DIRECTORY_NAME = os.path.dirname(os.path.realpath(__file__))
-
-DATABASES = {
- 'default': {
- 'ENGINE': 'django.db.backends.sqlite3', # Add 'postgresql_psycopg2', 'mysql', 'sqlite3' or 'oracle'.
- 'NAME': DIRECTORY_NAME + '/db', # Or path to database file if using sqlite3.
- # The following settings are not used with sqlite3:
- 'USER': '',
- 'PASSWORD': '',
- 'HOST': '', # Empty for localhost through domain sockets or '127.0.0.1' for localhost through TCP.
- 'PORT': '', # Set to empty string for default.
- }
-}
-
-# Hosts/domain names that are valid for this site; required if DEBUG is False
-# See https://docs.djangoproject.com/en/1.5/ref/settings/#allowed-hosts
-ALLOWED_HOSTS = []
-
-# Local time zone for this installation. Choices can be found here:
-# http://en.wikipedia.org/wiki/List_of_tz_zones_by_name
-# although not all choices may be available on all operating systems.
-# In a Windows environment this must be set to your system time zone.
-TIME_ZONE = 'America/Chicago'
-
-# Language code for this installation. All choices can be found here:
-# http://www.i18nguy.com/unicode/language-identifiers.html
-LANGUAGE_CODE = 'en-us'
-
-SITE_ID = 1
-
-# If you set this to False, Django will make some optimizations so as not
-# to load the internationalization machinery.
-USE_I18N = True
-
-# If you set this to False, Django will not format dates, numbers and
-# calendars according to the current locale.
-USE_L10N = True
-
-# If you set this to False, Django will not use timezone-aware datetimes.
-USE_TZ = True
-
-# Absolute filesystem path to the directory that will hold user-uploaded files.
-# Example: "/var/www/example.com/media/"
-MEDIA_ROOT = ''
-
-# URL that handles the media served from MEDIA_ROOT. Make sure to use a
-# trailing slash.
-# Examples: "http://example.com/media/", "http://media.example.com/"
-MEDIA_URL = ''
-
-# Absolute path to the directory static files should be collected to.
-# Don't put anything in this directory yourself; store your static files
-# in apps' "static/" subdirectories and in STATICFILES_DIRS.
-# Example: "/var/www/example.com/static/"
-STATIC_ROOT = ''
-
-# URL prefix for static files.
-# Example: "http://example.com/static/", "http://static.example.com/"
-STATIC_URL = '/static/'
-
-# Additional locations of static files
-STATICFILES_DIRS = (
- # Put strings here, like "/home/html/static" or "C:/www/django/static".
- # Always use forward slashes, even on Windows.
- # Don't forget to use absolute paths, not relative paths.
- DIRECTORY_NAME + '/static',
-)
-
-# List of finder classes that know how to find static files in
-# various locations.
-STATICFILES_FINDERS = (
- 'django.contrib.staticfiles.finders.FileSystemFinder',
- 'django.contrib.staticfiles.finders.AppDirectoriesFinder',
-# 'django.contrib.staticfiles.finders.DefaultStorageFinder',
-)
-
-# Make this unique, and don't share it with anybody.
-SECRET_KEY = 't!!j*1gt0(5n%6nj-lirzja-9uj6s86s#0@kp2@8v&x#+c2+c-'
-
-# List of callables that know how to import templates from various sources.
-TEMPLATE_LOADERS = (
- 'django.template.loaders.filesystem.Loader',
- 'django.template.loaders.app_directories.Loader',
-# 'django.template.loaders.eggs.Loader',
-)
-
-MIDDLEWARE_CLASSES = (
- 'django.middleware.common.CommonMiddleware',
- 'django.contrib.sessions.middleware.SessionMiddleware',
- 'django.middleware.csrf.CsrfViewMiddleware',
- 'django.contrib.auth.middleware.AuthenticationMiddleware',
- 'django.contrib.messages.middleware.MessageMiddleware',
- # Uncomment the next line for simple clickjacking protection:
- # 'django.middleware.clickjacking.XFrameOptionsMiddleware',
-)
-
-ROOT_URLCONF = 'stats_app.urls'
-
-# Python dotted path to the WSGI application used by Django's runserver.
-WSGI_APPLICATION = 'stats_app.wsgi.application'
-
-TEMPLATE_DIRS = (
- # Put strings here, like "/home/html/django_templates" or "C:/www/django/templates".
- # Always use forward slashes, even on Windows.
- # Don't forget to use absolute paths, not relative paths.
- DIRECTORY_NAME + '/templates',
-)
-
-INSTALLED_APPS = (
- 'django.contrib.auth',
- 'django.contrib.contenttypes',
- 'django.contrib.sessions',
- 'django.contrib.sites',
- 'django.contrib.messages',
- 'django.contrib.staticfiles',
- # Uncomment the next line to enable the admin:
- 'django.contrib.admin',
- # Uncomment the next line to enable admin documentation:
- 'django.contrib.admindocs',
-)
-
-SESSION_SERIALIZER = 'django.contrib.sessions.serializers.JSONSerializer'
-
-# A sample logging configuration. The only tangible logging
-# performed by this configuration is to send an email to
-# the site admins on every HTTP 500 error when DEBUG=False.
-# See http://docs.djangoproject.com/en/dev/topics/logging for
-# more details on how to customize your logging configuration.
-LOGGING = {
- 'version': 1,
- 'disable_existing_loggers': False,
- 'filters': {
- 'require_debug_false': {
- '()': 'django.utils.log.RequireDebugFalse'
- }
- },
- 'handlers': {
- 'mail_admins': {
- 'level': 'ERROR',
- 'filters': ['require_debug_false'],
- 'class': 'django.utils.log.AdminEmailHandler'
- }
- },
- 'loggers': {
- 'django.request': {
- 'handlers': ['mail_admins'],
- 'level': 'ERROR',
- 'propagate': True,
- },
- }
-}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/static/charts.css b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/static/charts.css
deleted file mode 100644
index e32e2832ae..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/static/charts.css
+++ /dev/null
@@ -1,11 +0,0 @@
-img.center {
-display: block;
-margin-left: auto;
-margin-right: auto;
-
-padding: 8px;
-border: solid;
-border-color: #dddddd #aaaaaa #aaaaaa #dddddd;
-border-width: 1px 2px 2px 1px;
-background-color: white;
-}
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/templates/charts.html b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/templates/charts.html
deleted file mode 100644
index d38bb4c0c6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/templates/charts.html
+++ /dev/null
@@ -1,41 +0,0 @@
-<!DOCTYPE html>
-<head>
-<link rel="stylesheet" type="text/css" href="{{ STATIC_URL }}charts.css" media="screen" />
-<html lang="en">
- <title>{% block title %}Graph{% endblock %}</title>
-</head>
-
-<body>
- <div id="graphForm">
- <script src="http://ajax.googleapis.com/ajax/libs/jquery/1.10.2/jquery.min.js"> </script>
- <script>
- function callback() {{
- var values = $('#graphForm form').serialize();
- $('#graphForm img').attr("src", "graph.png?" + values);
- }}
- </script>
- <p style="text-align:center">
- <img src="graph.png" id="graph" />
- </p>
- <form method = "GET" action="" style="text-align:center">
- <h3>X Limits:</h3>
- <input type="range" name="xlim" min="0" max="10000">
- <br><h3>Y Limits:</h3>
- <input type="range" name="ylim" min="0" max="20">
- <br>
- <h3>Labels:</h3>
- {0}
- <br>
- <h3>Measure:</h3>
- <b>Mean:</b>
- <input type="checkbox" name="disp_type" value="mean">
- <b>Median:</b>
- <input type="checkbox" name="disp_type" value="median">
- <b>All percentiles:</b>
- <input type="checkbox" name="disp_type" value="all_percentiles">
- <br>
- <input type="button" value="Graph!" onclick="callback()">
- </form>
- </div>
-</body>
-</html>
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/urls.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/urls.py
deleted file mode 100644
index 15743290ec..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/urls.py
+++ /dev/null
@@ -1,20 +0,0 @@
-from django.conf.urls import patterns, include, url
-
-# Uncomment the next two lines to enable the admin:
-from django.contrib import admin
-import views.charts
-admin.autodiscover()
-
-urlpatterns = patterns('',
- # Examples:
- # url(r'^$', 'stats_app.views.home', name='home'),
- # url(r'^stats_app/', include('stats_app.foo.urls')),
-
- # Uncomment the admin/doc line below to enable admin documentation:
- # url(r'^admin/doc/', include('django.contrib.admindocs.urls')),
-
- # Uncomment the next line to enable the admin:
- url(r'^admin/', include(admin.site.urls)),
- url(r'^graph.png$', views.charts.display_graph, name='graph'),
- url(r'^$', views.charts.display_full_page, name='graph_page'),
-)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/views.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/views.py
deleted file mode 100644
index 7cb32b3655..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/views.py
+++ /dev/null
@@ -1,5 +0,0 @@
-from django.http import HttpResponse
-
-
-def index(request):
- return HttpResponse("Hello, world. You're at the stats application index.")
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/views/__init__.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/views/__init__.py
deleted file mode 100644
index e69de29bb2..0000000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/views/charts.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/views/charts.py
deleted file mode 100644
index c3a2ebff32..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/views/charts.py
+++ /dev/null
@@ -1,67 +0,0 @@
-import datetime
-import django
-from django.shortcuts import render
-from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
-from matplotlib.dates import DateFormatter
-from matplotlib.figure import Figure
-import random
-
-from opentuner.utils import stats_matplotlib as stats
-
-
-def display_graph(request):
- """
- Handles request to display graph with provided parameters
- """
- request_dict = dict(request.GET.iterlists())
-
- xlim = request_dict.get('xlim', None)
- if xlim:
- xlim = int(xlim[0])
- else:
- xlim = 5000
- xlim = [0, xlim]
-
- ylim = request_dict.get('ylim', None)
- if ylim:
- ylim = int(ylim[0])
- else:
- ylim = 10
- ylim = [0, ylim]
-
- labels = request_dict.get('labels', None)
-
- disp_types = request_dict.get('disp_type', None)
- if not disp_types:
- disp_types = ['median']
-
- fig = stats.matplotlibplot_file(labels, xlim=xlim, ylim=ylim, disp_types=disp_types)
- canvas = FigureCanvas(fig)
- response = django.http.HttpResponse(content_type='image/png')
- canvas.print_png(response)
- return response
-
-
-def display_full_page(request):
- """
- Handles request to display the full page
- """
- all_labels = stats.get_all_labels()
- label_list = get_label_list(all_labels)
- html = render(request, 'charts.html')
- content = html.content
- content = content.format(label_list)
- html.content = content
- return html
-
-
-def get_label_list(all_labels):
- """
- Returns list of html form inputs corresponding to the different
- labels in the provided db file
- """
- label_list = ''
- for label in all_labels:
- label_list += '<b>%s</b>:<input type="checkbox" name="labels" value="%s">' % (label, label)
- return label_list
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/wsgi.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/wsgi.py
deleted file mode 100644
index 90f54d8e3d..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/stats_app/stats_app/wsgi.py
+++ /dev/null
@@ -1,32 +0,0 @@
-"""
-WSGI config for stats_app project.
-
-This module contains the WSGI application used by Django's development server
-and any production WSGI deployments. It should expose a module-level variable
-named ``application``. Django's ``runserver`` and ``runfcgi`` commands discover
-this application via the ``WSGI_APPLICATION`` setting.
-
-Usually you will have the standard Django WSGI application here, but it also
-might make sense to replace the whole Django WSGI application with a custom one
-that later delegates to the Django one. For example, you could introduce WSGI
-middleware here, or combine a Django application with an application of another
-framework.
-
-"""
-import os
-
-# We defer to a DJANGO_SETTINGS_MODULE already in the environment. This breaks
-# if running multiple sites in the same mod_wsgi process. To fix this, use
-# mod_wsgi daemon mode with each site in its own daemon process, or use
-# os.environ["DJANGO_SETTINGS_MODULE"] = "stats_app.settings"
-os.environ.setdefault("DJANGO_SETTINGS_MODULE", "stats_app.settings")
-
-# This application object is used by any WSGI server configured to use this
-# file. This includes Django's development server, if the WSGI_APPLICATION
-# setting points here.
-from django.core.wsgi import get_wsgi_application
-application = get_wsgi_application()
-
-# Apply WSGI middleware here.
-# from helloworld.wsgi import HelloWorldApplication
-# application = HelloWorldApplication(application)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/tests/test_manipulator.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/tests/test_manipulator.py
deleted file mode 100644
index c6548510fd..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/tests/test_manipulator.py
+++ /dev/null
@@ -1,270 +0,0 @@
-import unittest
-import opentuner
-import mock
-import random
-import numpy
-from opentuner.search import manipulator
-
-def faked_random(nums):
- f = fake_random(nums)
- def inner(*args, **kwargs):
- return f.next()
- return inner
-
-def fake_random(nums):
- i = 0
- while True:
- yield nums[i]
- i = (i+1) % len(nums)
-
-
-class PermutationOperatorTests(unittest.TestCase):
-
- def setUp(self):
- """
- Set up a few configurations. The values of the PermutationParameter are:
- config1 - 0 1 2 3 4 5 6 7 8 9
- config2 - 4 3 2 1 0 9 8 7 6 5
- config3 - 1 0 4 2 7 9 5 3 6 8
-
- """
- self.manipulator = manipulator.ConfigurationManipulator()
- self.param1 = manipulator.PermutationParameter("param1", [0,1,2,3,4,5,6,7,8,9])
- self.manipulator.add_parameter(self.param1)
-
- self.cfg = self.manipulator.seed_config()
- self.config1 = self.manipulator.seed_config()
- self.config2 = self.manipulator.seed_config()
- self.config3 = self.manipulator.seed_config()
-
- # repeating values
- self.config4 = self.manipulator.seed_config()
- self.config5 = self.manipulator.seed_config()
-
-
- self.param1.set_value(self.config1, [0,1,2,3,4,5,6,7,8,9])
- self.param1.set_value(self.config2, [4,3,2,1,0,9,8,7,6,5])
- self.param1.set_value(self.config3, [1,0,4,2,7,9,5,3,6,8])
-
- # repeating values
- self.param1.set_value(self.config4, [1,2,3,4,2,3,4,3,4,4])
- self.param1.set_value(self.config5, [4,2,4,3,3,1,3,4,2,4])
-
- @mock.patch('random.randint', side_effect=faked_random([1,6]))
- def test_op2_random_swap_1_6(self, randint_func):
- # operator shouuld swap the indices at 1 and 6
- self.param1.op2_random_swap(self.cfg, self.config1)
-
- self.assertEqual(self.param1.get_value(self.cfg),[0,6,2,3,4,5,1,7,8,9])
- self.assertEqual(self.param1.get_value(self.config1),[0,1,2,3,4,5,6,7,8,9])
-
-
- @mock.patch('random.randint', side_effect=faked_random([7,2]))
- def test_op2_random_invert(self, randint_func):
- #should reverse a section of length 3 starting at index given by randint
- self.param1.op2_random_invert(self.cfg, self.config1)
- self.assertEqual(self.param1.get_value(self.cfg),[0,1,2,3,4,5,6,9,8,7])
-
- self.param1.op2_random_invert(self.cfg, self.config1)
- self.assertEqual(self.param1.get_value(self.cfg),[0,1,4,3,2,5,6,7,8,9])
-
-
- @mock.patch('random.randint', side_effect=faked_random([0]))
- def test_op3_cross_PMX_str5(self, randint_func):
- # should perform PMX with a cut at 0 and crossover size 5
- self.param1.op3_cross(self.cfg, self.config1, self.config3,
- xchoice='op3_cross_PMX', strength=0.5)
- self.assertEqual(self.param1.get_value(self.cfg),[1,0,4,2,7,5,6,3,8,9])
-
- @mock.patch('random.randint', side_effect=faked_random([5]))
- @mock.patch('random.uniform', side_effect=faked_random([0.4]))
- def test_op3_swarm_CX_no_cross(self, uniform_func, randint_func):
- # should perform no cross
- self.param1.op3_swarm(self.config1, self.config2, self.config3,
- xchoice='op3_cross_CX', c=0.8)
- self.assertEqual(self.param1.get_value(self.config1),[0,1,2,3,4,5,6,7,8,9])
-
-
- @mock.patch('random.randint', side_effect=faked_random([5]))
- @mock.patch('random.uniform', side_effect=faked_random([0.4]))
- def test_op3_swarm_CX_cross_p1(self, uniform_func, randint_func):
- # should cross the first parent
- self.param1.op3_swarm(self.config1, self.config2, self.config3,
- xchoice='op3_cross_CX', c=0.3, c1=0.5, c2="unused")
- self.assertEqual(self.param1.get_value(self.config1),[0,1,2,3,4,9,6,7,8,5])
-
- @mock.patch('random.randint', side_effect=faked_random([5]))
- @mock.patch('random.uniform', side_effect=faked_random([0.4]))
- def test_op3_swarm_CX_cross_p2(self, uniform_func, randint_func):
- # should cross the second parent
- self.param1.op3_swarm(self.config1, self.config2, self.config3,
- xchoice='op3_cross_CX', c=0.3, c1=0.3, c2="unused")
- self.assertEqual(self.param1.get_value(self.config1),[0,1,2,3,4,9,5,7,6,8])
-
-
- @mock.patch('random.randint', side_effect=faked_random([5]))
- def test_op3_cross_PX_5(self, randint_func):
- # Random cut point = 5 (index = 4)
- self.param1.op3_cross_PX(self.cfg, self.config1, self.config3, 2)
- self.assertEqual(self.param1.get_value(self.cfg),[1,0,4,2,3,5,6,7,8,9])
-
- @mock.patch('random.randint', side_effect=faked_random([2]))
- def test_op3_cross_PMX_0_d4(self, randint_func):
- # cut = 2, d = 4
- self.param1.op3_cross_PMX(self.cfg, self.config2, self.config3, 4)
- self.assertEqual(self.param1.get_value(self.cfg),[1,3,4,2,7,9,8,0,6,5])
-
-
- @mock.patch('random.randint', side_effect=faked_random([0]))
- def test_op3_cross_PMX_0_d5(self, randint_func):
- # cut = 0, d = 5
- self.param1.op3_cross_PMX(self.cfg, self.config1, self.config3, 5)
- self.assertEqual(self.param1.get_value(self.cfg),[1,0,4,2,7,5,6,3,8,9])
-
- @mock.patch('random.randint', side_effect=faked_random([4]))
- def test_op3_cross_PMX_dups(self, randint_func):
- # cut = 4, d = 5
- self.param1.op3_cross_PMX(self.cfg, self.config5, self.config4, 5)
-
- # [4,2,4,3,3,1,3,4,2,4]
- # [1,2,3,4,2,3,4,3,4,4]
- # expected:
- # [1,2,4,3,2,3,4,3,4,4]
-
- self.assertEqual(self.param1.get_value(self.cfg), [1,2,4,3,2,3,4,3,4,4])
-
-
- @mock.patch('random.randint', side_effect=faked_random([5]))
- def test_op3_cross_CX_5(self, randint_func):
- # initial replacement at index 5
- self.param1.op3_cross_CX(self.cfg, self.config1, self.config2, "unused")
- self.assertEqual(self.param1.get_value(self.cfg),[0,1,2,3,4,9,6,7,8,5])
- self.param1.op3_cross_CX(self.cfg, self.config1, self.config3, "unused")
- self.assertEqual(self.param1.get_value(self.cfg),[0,1,2,3,4,9,5,7,6,8])
-
- @mock.patch('random.randint', side_effect=faked_random([0]))
- def test_op3_cross_CX_dups(self, randint_func):
- # initial replacement at index 4
- self.param1.op3_cross_CX(self.cfg, self.config5, self.config4, "unused")
-
- # [4,2,4,3,3,1,3,4,2,4]
- # [1,2,3,4,2,3,4,3,4,4]
- # expected:
- # [1,2,3,4,3,3,4,4,2,4]
-
- self.assertEqual(self.param1.get_value(self.cfg), [1,2,3,4,3,3,4,4,2,4])
-
-
- @mock.patch('random.randint', side_effect=faked_random([3]))
- def test_op3_cross_OX1_3_d4(self, randint_func):
- # cut at 3
- # d = 4
- self.param1.op3_cross_OX1(self.cfg, self.config1, self.config2, 4)
- self.assertEqual(self.param1.get_value(self.cfg),[2,3,4,1,0,9,8,5,6,7])
- self.param1.op3_cross_OX1(self.cfg, self.config1, self.config3, 4)
- self.assertEqual(self.param1.get_value(self.cfg),[0,1,3,2,7,9,5,4,6,8])
-
- @mock.patch('random.randint', side_effect=faked_random([4,2]))
- def test_op3_cross_OX3_2_5_d4(self, randint_func):
- # cuts at 4,2
- # d = 4
- self.param1.op3_cross_OX3(self.cfg, self.config1, self.config2, 4)
- self.assertEqual(self.param1.get_value(self.cfg),[3,4,5,6,2,1,0,9,7,8])
- self.param1.op3_cross_OX3(self.cfg, self.config1, self.config3, 4)
- self.assertEqual(self.param1.get_value(self.cfg),[0,1,3,5,4,2,7,9,6,8])
-
-
-class FloatArrayOperatorTests(unittest.TestCase):
- """
- also tests the operators for Array (since Array is abstract)
- """
-
- def setUp(self):
- """
- Set up a few configurations. The values of the FloatArray are:
- config1 - 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
- config2 - 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9
- config3 - 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9
- """
- self.manipulator = manipulator.ConfigurationManipulator()
- self.param1 = manipulator.FloatArray("param1", 10, 4, 0)
- self.manipulator.add_parameter(self.param1)
-
- self.cfg = self.manipulator.seed_config()
- self.config1 = self.manipulator.seed_config()
- self.config2 = self.manipulator.seed_config()
- self.config3 = self.manipulator.seed_config()
-
- self.param1.set_value(self.config1, numpy.array([1.0,1.1,1.2,1.3,1.4,1.5,1.6,1.7,1.8,1.9]))
- self.param1.set_value(self.config2, numpy.array([2.0,2.1,2.2,2.3,2.4,2.5,2.6,2.7,2.8,2.9]))
- self.param1.set_value(self.config3, numpy.array([3.0,3.1,3.2,3.3,3.4,3.5,3.6,3.7,3.8,3.8]))
-
-
- @mock.patch('random.randint', side_effect=faked_random([3]))
- def test_op3_cross_3_str4(self, randint_func):
- self.param1.op3_cross(self.cfg, self.config1, self.config2, strength=0.4)
-
- val = self.param1.get_value(self.cfg)
- expected = [1.0,1.1,1.2,2.3,2.4,2.5,2.6,1.7,1.8,1.9]
- for i in range(len(val)):
- self.assertAlmostEqual(val[i], expected[i])
-
- @mock.patch('random.randint', side_effect=faked_random([3]))
- @mock.patch('random.uniform', side_effect=faked_random([0.4]))
- def test_op3_swarm_no_cross(self, uniform_func, randint_func):
- #should perform no cross
- self.param1.op3_swarm(self.config1, self.config2, self.config3,
- xchoice='op3_cross_CX', c=0.8)
- val = self.param1.get_value(self.config1)
- expected = [1.0,1.1,1.2,1.3,1.4,1.5,1.6,1.7,1.8,1.9]
- for i in range(len(val)):
- self.assertAlmostEqual(val[i], expected[i])
-
-
- @mock.patch('random.randint', side_effect=faked_random([3]))
- @mock.patch('random.uniform', side_effect=faked_random([0.4]))
- def test_op3_swarm_cross_p1(self, uniform_func, randint_func):
- #should cross the first parent
- self.param1.op3_swarm(self.config1, self.config2, self.config3,
- xchoice='op3_cross_CX', c=0.3, c1=0.5, c2="unused")
- val = self.param1.get_value(self.config1)
- expected = [1.0,1.1,1.2,2.3,2.4,2.5,1.6,1.7,1.8,1.9]
- for i in range(len(val)):
- self.assertAlmostEqual(val[i], expected[i])
-
-
- @mock.patch('random.randint', side_effect=faked_random([3]))
- @mock.patch('random.uniform', side_effect=faked_random([0.4]))
- def test_op3_swarm_cross_p2(self, uniform_func, randint_func):
- #should cross the second parent
- self.param1.op3_swarm(self.config1, self.config2, self.config3,
- xchoice='op3_cross_CX', c=0.3, c1=0.3, c2="unused")
- val = self.param1.get_value(self.config1)
- expected = [1.0,1.1,1.2,3.3,3.4,3.5,1.6,1.7,1.8,1.9]
- self.assertEqual(len(val),len(expected))
- for i in range(len(val)):
- self.assertAlmostEqual(val[i], expected[i])
-
- @mock.patch('random.random', side_effect=faked_random([0.2, 0.4]))
- def test_op3_swarm_parallel(self, random_func):
- # r1 = 0.2, r2 = 0.4, velocities = [-2,0,0,0,0,0,1,1.5,2,3]
- # max and min are 4, 0
- velocities = numpy.array([-2.0,0.0,0,0,0,0,1.0,1.5,2,3.0])
-
- vs = self.param1.op3_swarm_parallel(self.config1, self.config2, self.config3, velocities=velocities)
- vs_expected = [-1.5,.5,.5,.5,.5,.5,1.5,2.0,2.5,3.48]
-
- self.assertEqual(len(vs),len(vs_expected))
-
- for i in range(len(vs)):
- self.assertAlmostEqual(vs[i], vs_expected[i])
-
-
- val = self.param1.get_value(self.config1)
- expected = [0,1.6,1.7,1.8,1.9,2.0,3.1,3.7,4,4]
- self.assertEqual(len(val),len(expected))
- for i in range(len(val)):
- self.assertAlmostEqual(val[i], expected[i])
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/tests/test_technique.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/tests/test_technique.py
deleted file mode 100644
index c6107bace9..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/tests/test_technique.py
+++ /dev/null
@@ -1,77 +0,0 @@
-import unittest
-import opentuner
-import mock
-from opentuner.search.composableevolutionarytechniques import ComposableEvolutionaryTechnique
-from opentuner.search import manipulator
-
-def faked_random(nums):
- f = fake_random(nums)
- def inner(*args, **kwargs):
- return f.next()
- return inner
-
-def fake_random(nums):
- i = 0
- while True:
- yield nums[i]
- i = (i+1) % len(nums)
-
-class EmptyComposableEvolutionaryTechnique(ComposableEvolutionaryTechnique):
- def __init__(self, *pargs, **kwargs):
- super(EmptyComposableEvolutionaryTechnique, self).__init__(*pargs, **kwargs)
-
- def minimum_number_of_parents(self):
- return 4
-
- def get_parents(self, population):
- cfg = self.manipulator.copy(population[0].config)
-
- return [cfg]
-
- def update_population(self, config, population):
- # replace the oldest configuration if the new one is better.
- population[0].config = config
-
- return population
-
-class ComposableSearchTechniqueTests(unittest.TestCase):
-
- def setUp(self):
- self.operator_map = {}
- ComposableEvolutionaryTechnique.add_to_map(self.operator_map,
- manipulator.PermutationParameter,
- "op3_cross", xchoice='op3_cross_CX')
- ComposableEvolutionaryTechnique.add_to_map(self.operator_map,
- "FloatArray",
- "op3_cross", strength=0.4)
- self.technique = EmptyComposableEvolutionaryTechnique(operator_map = self.operator_map)
-
- def test_add_to_map(self):
- op_map = {}
- op_map[manipulator.PermutationParameter] = {'op_name': 'op3_cross',
- 'args': (),
- 'kwargs': {'xchoice': 'op3_cross_CX'}}
- op_map[manipulator.FloatArray] = {'op_name': 'op3_cross',
- 'args': (),
- 'kwargs': {'strength': 0.4}}
- self.assertDictEqual(self.operator_map, op_map)
-
- def test_get_default_oeprator(self):
- default = self.technique.get_default_operator(manipulator.PermutationParameter)
- self.assertDictEqual(default, {'op_name': 'op1_nop', 'args': [], 'kwargs': {}})
-
-
- def test_get_operator(self):
- default = self.technique.get_operator(manipulator.IntegerParameter)
- self.assertDictEqual(default, {'op_name': 'op1_nop', 'args': [], 'kwargs': {}})
-
- default = self.technique.get_operator(manipulator.PermutationParameter)
- self.assertDictEqual(default, {'op_name': 'op3_cross','args': (),'kwargs': {'xchoice': 'op3_cross_CX'}})
-
- @mock.patch('opentuner.search.manipulator.PermutationParameter.op3_cross')
- def test_apply_operator(self, op3_cross_func):
- param_instance = manipulator.PermutationParameter('temp', [1,2,3,4,5])
- self.technique.apply_operator(param_instance, ['p1', 'p2', 'p3', 'p4'])
- op3_cross_func.assert_called_once_with('p1', 'p2', 'p3', xchoice='op3_cross_CX')
-
-#TODO tests for RandomThreeParentsComposableTechnique
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/venv-bootstrap.py b/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/venv-bootstrap.py
deleted file mode 100755
index 6d6ad0113b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/opentuner/venv-bootstrap.py
+++ /dev/null
@@ -1,2611 +0,0 @@
-#!/usr/bin/env python
-## WARNING: This file is generated
-#!/usr/bin/env python
-"""Create a "virtual" Python installation
-"""
-
-# If you change the version here, change it in setup.py
-# and docs/conf.py as well.
-__version__ = "1.9.1" # following best practices
-virtualenv_version = __version__ # legacy, again
-
-import base64
-import sys
-import os
-import codecs
-import optparse
-import re
-import shutil
-import logging
-import tempfile
-import zlib
-import errno
-import glob
-import distutils.sysconfig
-from distutils.util import strtobool
-import struct
-import subprocess
-
-if sys.version_info < (2, 6):
- print('ERROR: %s' % sys.exc_info()[1])
- print('ERROR: this script requires Python 2.6 or greater.')
- sys.exit(101)
-
-try:
- set
-except NameError:
- from sets import Set as set
-try:
- basestring
-except NameError:
- basestring = str
-
-try:
- import ConfigParser
-except ImportError:
- import configparser as ConfigParser
-
-join = os.path.join
-py_version = 'python%s.%s' % (sys.version_info[0], sys.version_info[1])
-
-is_jython = sys.platform.startswith('java')
-is_pypy = hasattr(sys, 'pypy_version_info')
-is_win = (sys.platform == 'win32')
-is_cygwin = (sys.platform == 'cygwin')
-is_darwin = (sys.platform == 'darwin')
-abiflags = getattr(sys, 'abiflags', '')
-
-user_dir = os.path.expanduser('~')
-if is_win:
- default_storage_dir = os.path.join(user_dir, 'virtualenv')
-else:
- default_storage_dir = os.path.join(user_dir, '.virtualenv')
-default_config_file = os.path.join(default_storage_dir, 'virtualenv.ini')
-
-if is_pypy:
- expected_exe = 'pypy'
-elif is_jython:
- expected_exe = 'jython'
-else:
- expected_exe = 'python'
-
-
-REQUIRED_MODULES = ['os', 'posix', 'posixpath', 'nt', 'ntpath', 'genericpath',
- 'fnmatch', 'locale', 'encodings', 'codecs',
- 'stat', 'UserDict', 'readline', 'copy_reg', 'types',
- 're', 'sre', 'sre_parse', 'sre_constants', 'sre_compile',
- 'zlib']
-
-REQUIRED_FILES = ['lib-dynload', 'config']
-
-majver, minver = sys.version_info[:2]
-if majver == 2:
- if minver >= 6:
- REQUIRED_MODULES.extend(['warnings', 'linecache', '_abcoll', 'abc'])
- if minver >= 7:
- REQUIRED_MODULES.extend(['_weakrefset'])
- if minver <= 3:
- REQUIRED_MODULES.extend(['sets', '__future__'])
-elif majver == 3:
- # Some extra modules are needed for Python 3, but different ones
- # for different versions.
- REQUIRED_MODULES.extend(['_abcoll', 'warnings', 'linecache', 'abc', 'io',
- '_weakrefset', 'copyreg', 'tempfile', 'random',
- '__future__', 'collections', 'keyword', 'tarfile',
- 'shutil', 'struct', 'copy', 'tokenize', 'token',
- 'functools', 'heapq', 'bisect', 'weakref',
- 'reprlib'])
- if minver >= 2:
- REQUIRED_FILES[-1] = 'config-%s' % majver
- if minver == 3:
- import sysconfig
- platdir = sysconfig.get_config_var('PLATDIR')
- REQUIRED_FILES.append(platdir)
- # The whole list of 3.3 modules is reproduced below - the current
- # uncommented ones are required for 3.3 as of now, but more may be
- # added as 3.3 development continues.
- REQUIRED_MODULES.extend([
- #"aifc",
- #"antigravity",
- #"argparse",
- #"ast",
- #"asynchat",
- #"asyncore",
- "base64",
- #"bdb",
- #"binhex",
- #"bisect",
- #"calendar",
- #"cgi",
- #"cgitb",
- #"chunk",
- #"cmd",
- #"codeop",
- #"code",
- #"colorsys",
- #"_compat_pickle",
- #"compileall",
- #"concurrent",
- #"configparser",
- #"contextlib",
- #"cProfile",
- #"crypt",
- #"csv",
- #"ctypes",
- #"curses",
- #"datetime",
- #"dbm",
- #"decimal",
- #"difflib",
- #"dis",
- #"doctest",
- #"dummy_threading",
- "_dummy_thread",
- #"email",
- #"filecmp",
- #"fileinput",
- #"formatter",
- #"fractions",
- #"ftplib",
- #"functools",
- #"getopt",
- #"getpass",
- #"gettext",
- #"glob",
- #"gzip",
- "hashlib",
- #"heapq",
- "hmac",
- #"html",
- #"http",
- #"idlelib",
- #"imaplib",
- #"imghdr",
- "imp",
- "importlib",
- #"inspect",
- #"json",
- #"lib2to3",
- #"logging",
- #"macpath",
- #"macurl2path",
- #"mailbox",
- #"mailcap",
- #"_markupbase",
- #"mimetypes",
- #"modulefinder",
- #"multiprocessing",
- #"netrc",
- #"nntplib",
- #"nturl2path",
- #"numbers",
- #"opcode",
- #"optparse",
- #"os2emxpath",
- #"pdb",
- #"pickle",
- #"pickletools",
- #"pipes",
- #"pkgutil",
- #"platform",
- #"plat-linux2",
- #"plistlib",
- #"poplib",
- #"pprint",
- #"profile",
- #"pstats",
- #"pty",
- #"pyclbr",
- #"py_compile",
- #"pydoc_data",
- #"pydoc",
- #"_pyio",
- #"queue",
- #"quopri",
- #"reprlib",
- "rlcompleter",
- #"runpy",
- #"sched",
- #"shelve",
- #"shlex",
- #"smtpd",
- #"smtplib",
- #"sndhdr",
- #"socket",
- #"socketserver",
- #"sqlite3",
- #"ssl",
- #"stringprep",
- #"string",
- #"_strptime",
- #"subprocess",
- #"sunau",
- #"symbol",
- #"symtable",
- #"sysconfig",
- #"tabnanny",
- #"telnetlib",
- #"test",
- #"textwrap",
- #"this",
- #"_threading_local",
- #"threading",
- #"timeit",
- #"tkinter",
- #"tokenize",
- #"token",
- #"traceback",
- #"trace",
- #"tty",
- #"turtledemo",
- #"turtle",
- #"unittest",
- #"urllib",
- #"uuid",
- #"uu",
- #"wave",
- #"weakref",
- #"webbrowser",
- #"wsgiref",
- #"xdrlib",
- #"xml",
- #"xmlrpc",
- #"zipfile",
- ])
-
-if is_pypy:
- # these are needed to correctly display the exceptions that may happen
- # during the bootstrap
- REQUIRED_MODULES.extend(['traceback', 'linecache'])
-
-class Logger(object):
-
- """
- Logging object for use in command-line script. Allows ranges of
- levels, to avoid some redundancy of displayed information.
- """
-
- DEBUG = logging.DEBUG
- INFO = logging.INFO
- NOTIFY = (logging.INFO+logging.WARN)/2
- WARN = WARNING = logging.WARN
- ERROR = logging.ERROR
- FATAL = logging.FATAL
-
- LEVELS = [DEBUG, INFO, NOTIFY, WARN, ERROR, FATAL]
-
- def __init__(self, consumers):
- self.consumers = consumers
- self.indent = 0
- self.in_progress = None
- self.in_progress_hanging = False
-
- def debug(self, msg, *args, **kw):
- self.log(self.DEBUG, msg, *args, **kw)
- def info(self, msg, *args, **kw):
- self.log(self.INFO, msg, *args, **kw)
- def notify(self, msg, *args, **kw):
- self.log(self.NOTIFY, msg, *args, **kw)
- def warn(self, msg, *args, **kw):
- self.log(self.WARN, msg, *args, **kw)
- def error(self, msg, *args, **kw):
- self.log(self.ERROR, msg, *args, **kw)
- def fatal(self, msg, *args, **kw):
- self.log(self.FATAL, msg, *args, **kw)
- def log(self, level, msg, *args, **kw):
- if args:
- if kw:
- raise TypeError(
- "You may give positional or keyword arguments, not both")
- args = args or kw
- rendered = None
- for consumer_level, consumer in self.consumers:
- if self.level_matches(level, consumer_level):
- if (self.in_progress_hanging
- and consumer in (sys.stdout, sys.stderr)):
- self.in_progress_hanging = False
- sys.stdout.write('\n')
- sys.stdout.flush()
- if rendered is None:
- if args:
- rendered = msg % args
- else:
- rendered = msg
- rendered = ' '*self.indent + rendered
- if hasattr(consumer, 'write'):
- consumer.write(rendered+'\n')
- else:
- consumer(rendered)
-
- def start_progress(self, msg):
- assert not self.in_progress, (
- "Tried to start_progress(%r) while in_progress %r"
- % (msg, self.in_progress))
- if self.level_matches(self.NOTIFY, self._stdout_level()):
- sys.stdout.write(msg)
- sys.stdout.flush()
- self.in_progress_hanging = True
- else:
- self.in_progress_hanging = False
- self.in_progress = msg
-
- def end_progress(self, msg='done.'):
- assert self.in_progress, (
- "Tried to end_progress without start_progress")
- if self.stdout_level_matches(self.NOTIFY):
- if not self.in_progress_hanging:
- # Some message has been printed out since start_progress
- sys.stdout.write('...' + self.in_progress + msg + '\n')
- sys.stdout.flush()
- else:
- sys.stdout.write(msg + '\n')
- sys.stdout.flush()
- self.in_progress = None
- self.in_progress_hanging = False
-
- def show_progress(self):
- """If we are in a progress scope, and no log messages have been
- shown, write out another '.'"""
- if self.in_progress_hanging:
- sys.stdout.write('.')
- sys.stdout.flush()
-
- def stdout_level_matches(self, level):
- """Returns true if a message at this level will go to stdout"""
- return self.level_matches(level, self._stdout_level())
-
- def _stdout_level(self):
- """Returns the level that stdout runs at"""
- for level, consumer in self.consumers:
- if consumer is sys.stdout:
- return level
- return self.FATAL
-
- def level_matches(self, level, consumer_level):
- """
- >>> l = Logger([])
- >>> l.level_matches(3, 4)
- False
- >>> l.level_matches(3, 2)
- True
- >>> l.level_matches(slice(None, 3), 3)
- False
- >>> l.level_matches(slice(None, 3), 2)
- True
- >>> l.level_matches(slice(1, 3), 1)
- True
- >>> l.level_matches(slice(2, 3), 1)
- False
- """
- if isinstance(level, slice):
- start, stop = level.start, level.stop
- if start is not None and start > consumer_level:
- return False
- if stop is not None and stop <= consumer_level:
- return False
- return True
- else:
- return level >= consumer_level
-
- #@classmethod
- def level_for_integer(cls, level):
- levels = cls.LEVELS
- if level < 0:
- return levels[0]
- if level >= len(levels):
- return levels[-1]
- return levels[level]
-
- level_for_integer = classmethod(level_for_integer)
-
-# create a silent logger just to prevent this from being undefined
-# will be overridden with requested verbosity main() is called.
-logger = Logger([(Logger.LEVELS[-1], sys.stdout)])
-
-def mkdir(path):
- if not os.path.exists(path):
- logger.info('Creating %s', path)
- os.makedirs(path)
- else:
- logger.info('Directory %s already exists', path)
-
-def copyfileordir(src, dest):
- if os.path.isdir(src):
- shutil.copytree(src, dest, True)
- else:
- shutil.copy2(src, dest)
-
-def copyfile(src, dest, symlink=True):
- if not os.path.exists(src):
- # Some bad symlink in the src
- logger.warn('Cannot find file %s (bad symlink)', src)
- return
- if os.path.exists(dest):
- logger.debug('File %s already exists', dest)
- return
- if not os.path.exists(os.path.dirname(dest)):
- logger.info('Creating parent directories for %s' % os.path.dirname(dest))
- os.makedirs(os.path.dirname(dest))
- if not os.path.islink(src):
- srcpath = os.path.abspath(src)
- else:
- srcpath = os.readlink(src)
- if symlink and hasattr(os, 'symlink') and not is_win:
- logger.info('Symlinking %s', dest)
- try:
- os.symlink(srcpath, dest)
- except (OSError, NotImplementedError):
- logger.info('Symlinking failed, copying to %s', dest)
- copyfileordir(src, dest)
- else:
- logger.info('Copying to %s', dest)
- copyfileordir(src, dest)
-
-def writefile(dest, content, overwrite=True):
- if not os.path.exists(dest):
- logger.info('Writing %s', dest)
- f = open(dest, 'wb')
- f.write(content.encode('utf-8'))
- f.close()
- return
- else:
- f = open(dest, 'rb')
- c = f.read()
- f.close()
- if c != content.encode("utf-8"):
- if not overwrite:
- logger.notify('File %s exists with different content; not overwriting', dest)
- return
- logger.notify('Overwriting %s with new content', dest)
- f = open(dest, 'wb')
- f.write(content.encode('utf-8'))
- f.close()
- else:
- logger.info('Content %s already in place', dest)
-
-def rmtree(dir):
- if os.path.exists(dir):
- logger.notify('Deleting tree %s', dir)
- shutil.rmtree(dir)
- else:
- logger.info('Do not need to delete %s; already gone', dir)
-
-def make_exe(fn):
- if hasattr(os, 'chmod'):
- oldmode = os.stat(fn).st_mode & 0xFFF # 0o7777
- newmode = (oldmode | 0x16D) & 0xFFF # 0o555, 0o7777
- os.chmod(fn, newmode)
- logger.info('Changed mode of %s to %s', fn, oct(newmode))
-
-def _find_file(filename, dirs):
- for dir in reversed(dirs):
- files = glob.glob(os.path.join(dir, filename))
- if files and os.path.isfile(files[0]):
- return True, files[0]
- return False, filename
-
-def _install_req(py_executable, unzip=False, distribute=False,
- search_dirs=None, never_download=False):
-
- if search_dirs is None:
- search_dirs = file_search_dirs()
-
- if not distribute:
- egg_path = 'setuptools-*-py%s.egg' % sys.version[:3]
- found, egg_path = _find_file(egg_path, search_dirs)
- project_name = 'setuptools'
- bootstrap_script = EZ_SETUP_PY
- tgz_path = None
- else:
- # Look for a distribute egg (these are not distributed by default,
- # but can be made available by the user)
- egg_path = 'distribute-*-py%s.egg' % sys.version[:3]
- found, egg_path = _find_file(egg_path, search_dirs)
- project_name = 'distribute'
- if found:
- tgz_path = None
- bootstrap_script = DISTRIBUTE_FROM_EGG_PY
- else:
- # Fall back to sdist
- # NB: egg_path is not None iff tgz_path is None
- # iff bootstrap_script is a generic setup script accepting
- # the standard arguments.
- egg_path = None
- tgz_path = 'distribute-*.tar.gz'
- found, tgz_path = _find_file(tgz_path, search_dirs)
- bootstrap_script = DISTRIBUTE_SETUP_PY
-
- if is_jython and os._name == 'nt':
- # Jython's .bat sys.executable can't handle a command line
- # argument with newlines
- fd, ez_setup = tempfile.mkstemp('.py')
- os.write(fd, bootstrap_script)
- os.close(fd)
- cmd = [py_executable, ez_setup]
- else:
- cmd = [py_executable, '-c', bootstrap_script]
- if unzip and egg_path:
- cmd.append('--always-unzip')
- env = {}
- remove_from_env = ['__PYVENV_LAUNCHER__']
- if logger.stdout_level_matches(logger.DEBUG) and egg_path:
- cmd.append('-v')
-
- old_chdir = os.getcwd()
- if egg_path is not None and os.path.exists(egg_path):
- logger.info('Using existing %s egg: %s' % (project_name, egg_path))
- cmd.append(egg_path)
- if os.environ.get('PYTHONPATH'):
- env['PYTHONPATH'] = egg_path + os.path.pathsep + os.environ['PYTHONPATH']
- else:
- env['PYTHONPATH'] = egg_path
- elif tgz_path is not None and os.path.exists(tgz_path):
- # Found a tgz source dist, let's chdir
- logger.info('Using existing %s egg: %s' % (project_name, tgz_path))
- os.chdir(os.path.dirname(tgz_path))
- # in this case, we want to be sure that PYTHONPATH is unset (not
- # just empty, really unset), else CPython tries to import the
- # site.py that it's in virtualenv_support
- remove_from_env.append('PYTHONPATH')
- elif never_download:
- logger.fatal("Can't find any local distributions of %s to install "
- "and --never-download is set. Either re-run virtualenv "
- "without the --never-download option, or place a %s "
- "distribution (%s) in one of these "
- "locations: %r" % (project_name, project_name,
- egg_path or tgz_path,
- search_dirs))
- sys.exit(1)
- elif egg_path:
- logger.info('No %s egg found; downloading' % project_name)
- cmd.extend(['--always-copy', '-U', project_name])
- else:
- logger.info('No %s tgz found; downloading' % project_name)
- logger.start_progress('Installing %s...' % project_name)
- logger.indent += 2
- cwd = None
- if project_name == 'distribute':
- env['DONT_PATCH_SETUPTOOLS'] = 'true'
-
- def _filter_ez_setup(line):
- return filter_ez_setup(line, project_name)
-
- if not os.access(os.getcwd(), os.W_OK):
- cwd = tempfile.mkdtemp()
- if tgz_path is not None and os.path.exists(tgz_path):
- # the current working dir is hostile, let's copy the
- # tarball to a temp dir
- target = os.path.join(cwd, os.path.split(tgz_path)[-1])
- shutil.copy(tgz_path, target)
- try:
- call_subprocess(cmd, show_stdout=False,
- filter_stdout=_filter_ez_setup,
- extra_env=env,
- remove_from_env=remove_from_env,
- cwd=cwd)
- finally:
- logger.indent -= 2
- logger.end_progress()
- if cwd is not None:
- shutil.rmtree(cwd)
- if os.getcwd() != old_chdir:
- os.chdir(old_chdir)
- if is_jython and os._name == 'nt':
- os.remove(ez_setup)
-
-def file_search_dirs():
- here = os.path.dirname(os.path.abspath(__file__))
- dirs = ['.', here,
- join(here, 'virtualenv_support')]
- if os.path.splitext(os.path.dirname(__file__))[0] != 'virtualenv':
- # Probably some boot script; just in case virtualenv is installed...
- try:
- import virtualenv
- except ImportError:
- pass
- else:
- dirs.append(os.path.join(os.path.dirname(virtualenv.__file__), 'virtualenv_support'))
- return [d for d in dirs if os.path.isdir(d)]
-
-def install_setuptools(py_executable, unzip=False,
- search_dirs=None, never_download=False):
- _install_req(py_executable, unzip,
- search_dirs=search_dirs, never_download=never_download)
-
-def install_distribute(py_executable, unzip=False,
- search_dirs=None, never_download=False):
- _install_req(py_executable, unzip, distribute=True,
- search_dirs=search_dirs, never_download=never_download)
-
-_pip_re = re.compile(r'^pip-.*(zip|tar.gz|tar.bz2|tgz|tbz)$', re.I)
-def install_pip(py_executable, search_dirs=None, never_download=False):
- if search_dirs is None:
- search_dirs = file_search_dirs()
-
- filenames = []
- for dir in search_dirs:
- filenames.extend([join(dir, fn) for fn in os.listdir(dir)
- if _pip_re.search(fn)])
- filenames = [(os.path.basename(filename).lower(), i, filename) for i, filename in enumerate(filenames)]
- filenames.sort()
- filenames = [filename for basename, i, filename in filenames]
- if not filenames:
- filename = 'pip'
- else:
- filename = filenames[-1]
- easy_install_script = 'easy_install'
- if is_win:
- easy_install_script = 'easy_install-script.py'
- # There's two subtle issues here when invoking easy_install.
- # 1. On unix-like systems the easy_install script can *only* be executed
- # directly if its full filesystem path is no longer than 78 characters.
- # 2. A work around to [1] is to use the `python path/to/easy_install foo`
- # pattern, but that breaks if the path contains non-ASCII characters, as
- # you can't put the file encoding declaration before the shebang line.
- # The solution is to use Python's -x flag to skip the first line of the
- # script (and any ASCII decoding errors that may have occurred in that line)
- cmd = [py_executable, '-x', join(os.path.dirname(py_executable), easy_install_script), filename]
- # jython and pypy don't yet support -x
- if is_jython or is_pypy:
- cmd.remove('-x')
- if filename == 'pip':
- if never_download:
- logger.fatal("Can't find any local distributions of pip to install "
- "and --never-download is set. Either re-run virtualenv "
- "without the --never-download option, or place a pip "
- "source distribution (zip/tar.gz/tar.bz2) in one of these "
- "locations: %r" % search_dirs)
- sys.exit(1)
- logger.info('Installing pip from network...')
- else:
- logger.info('Installing existing %s distribution: %s' % (
- os.path.basename(filename), filename))
- logger.start_progress('Installing pip...')
- logger.indent += 2
- def _filter_setup(line):
- return filter_ez_setup(line, 'pip')
- try:
- call_subprocess(cmd, show_stdout=False,
- filter_stdout=_filter_setup)
- finally:
- logger.indent -= 2
- logger.end_progress()
-
-def filter_ez_setup(line, project_name='setuptools'):
- if not line.strip():
- return Logger.DEBUG
- if project_name == 'distribute':
- for prefix in ('Extracting', 'Now working', 'Installing', 'Before',
- 'Scanning', 'Setuptools', 'Egg', 'Already',
- 'running', 'writing', 'reading', 'installing',
- 'creating', 'copying', 'byte-compiling', 'removing',
- 'Processing'):
- if line.startswith(prefix):
- return Logger.DEBUG
- return Logger.DEBUG
- for prefix in ['Reading ', 'Best match', 'Processing setuptools',
- 'Copying setuptools', 'Adding setuptools',
- 'Installing ', 'Installed ']:
- if line.startswith(prefix):
- return Logger.DEBUG
- return Logger.INFO
-
-
-class UpdatingDefaultsHelpFormatter(optparse.IndentedHelpFormatter):
- """
- Custom help formatter for use in ConfigOptionParser that updates
- the defaults before expanding them, allowing them to show up correctly
- in the help listing
- """
- def expand_default(self, option):
- if self.parser is not None:
- self.parser.update_defaults(self.parser.defaults)
- return optparse.IndentedHelpFormatter.expand_default(self, option)
-
-
-class ConfigOptionParser(optparse.OptionParser):
- """
- Custom option parser which updates its defaults by by checking the
- configuration files and environmental variables
- """
- def __init__(self, *args, **kwargs):
- self.config = ConfigParser.RawConfigParser()
- self.files = self.get_config_files()
- self.config.read(self.files)
- optparse.OptionParser.__init__(self, *args, **kwargs)
-
- def get_config_files(self):
- config_file = os.environ.get('VIRTUALENV_CONFIG_FILE', False)
- if config_file and os.path.exists(config_file):
- return [config_file]
- return [default_config_file]
-
- def update_defaults(self, defaults):
- """
- Updates the given defaults with values from the config files and
- the environ. Does a little special handling for certain types of
- options (lists).
- """
- # Then go and look for the other sources of configuration:
- config = {}
- # 1. config files
- config.update(dict(self.get_config_section('virtualenv')))
- # 2. environmental variables
- config.update(dict(self.get_environ_vars()))
- # Then set the options with those values
- for key, val in config.items():
- key = key.replace('_', '-')
- if not key.startswith('--'):
- key = '--%s' % key # only prefer long opts
- option = self.get_option(key)
- if option is not None:
- # ignore empty values
- if not val:
- continue
- # handle multiline configs
- if option.action == 'append':
- val = val.split()
- else:
- option.nargs = 1
- if option.action == 'store_false':
- val = not strtobool(val)
- elif option.action in ('store_true', 'count'):
- val = strtobool(val)
- try:
- val = option.convert_value(key, val)
- except optparse.OptionValueError:
- e = sys.exc_info()[1]
- print("An error occured during configuration: %s" % e)
- sys.exit(3)
- defaults[option.dest] = val
- return defaults
-
- def get_config_section(self, name):
- """
- Get a section of a configuration
- """
- if self.config.has_section(name):
- return self.config.items(name)
- return []
-
- def get_environ_vars(self, prefix='VIRTUALENV_'):
- """
- Returns a generator with all environmental vars with prefix VIRTUALENV
- """
- for key, val in os.environ.items():
- if key.startswith(prefix):
- yield (key.replace(prefix, '').lower(), val)
-
- def get_default_values(self):
- """
- Overridding to make updating the defaults after instantiation of
- the option parser possible, update_defaults() does the dirty work.
- """
- if not self.process_default_values:
- # Old, pre-Optik 1.5 behaviour.
- return optparse.Values(self.defaults)
-
- defaults = self.update_defaults(self.defaults.copy()) # ours
- for option in self._get_all_options():
- default = defaults.get(option.dest)
- if isinstance(default, basestring):
- opt_str = option.get_opt_string()
- defaults[option.dest] = option.check_value(opt_str, default)
- return optparse.Values(defaults)
-
-
-def main():
- parser = ConfigOptionParser(
- version=virtualenv_version,
- usage="%prog [OPTIONS] DEST_DIR",
- formatter=UpdatingDefaultsHelpFormatter())
-
- parser.add_option(
- '-v', '--verbose',
- action='count',
- dest='verbose',
- default=0,
- help="Increase verbosity")
-
- parser.add_option(
- '-q', '--quiet',
- action='count',
- dest='quiet',
- default=0,
- help='Decrease verbosity')
-
- parser.add_option(
- '-p', '--python',
- dest='python',
- metavar='PYTHON_EXE',
- help='The Python interpreter to use, e.g., --python=python2.5 will use the python2.5 '
- 'interpreter to create the new environment. The default is the interpreter that '
- 'virtualenv was installed with (%s)' % sys.executable)
-
- parser.add_option(
- '--clear',
- dest='clear',
- action='store_true',
- help="Clear out the non-root install and start from scratch")
-
- parser.set_defaults(system_site_packages=False)
- parser.add_option(
- '--no-site-packages',
- dest='system_site_packages',
- action='store_false',
- help="Don't give access to the global site-packages dir to the "
- "virtual environment (default)")
-
- parser.add_option(
- '--system-site-packages',
- dest='system_site_packages',
- action='store_true',
- help="Give access to the global site-packages dir to the "
- "virtual environment")
-
- parser.add_option(
- '--unzip-setuptools',
- dest='unzip_setuptools',
- action='store_true',
- help="Unzip Setuptools or Distribute when installing it")
-
- parser.add_option(
- '--relocatable',
- dest='relocatable',
- action='store_true',
- help='Make an EXISTING virtualenv environment relocatable. '
- 'This fixes up scripts and makes all .pth files relative')
-
- parser.add_option(
- '--distribute', '--use-distribute', # the second option is for legacy reasons here. Hi Kenneth!
- dest='use_distribute',
- action='store_true',
- help='Use Distribute instead of Setuptools. Set environ variable '
- 'VIRTUALENV_DISTRIBUTE to make it the default ')
-
- parser.add_option(
- '--no-setuptools',
- dest='no_setuptools',
- action='store_true',
- help='Do not install distribute/setuptools (or pip) '
- 'in the new virtualenv.')
-
- parser.add_option(
- '--no-pip',
- dest='no_pip',
- action='store_true',
- help='Do not install pip in the new virtualenv.')
-
- parser.add_option(
- '--setuptools',
- dest='use_distribute',
- action='store_false',
- help='Use Setuptools instead of Distribute. Set environ variable '
- 'VIRTUALENV_SETUPTOOLS to make it the default ')
-
- # Set this to True to use distribute by default, even in Python 2.
- parser.set_defaults(use_distribute=False)
-
- default_search_dirs = file_search_dirs()
- parser.add_option(
- '--extra-search-dir',
- dest="search_dirs",
- action="append",
- default=default_search_dirs,
- help="Directory to look for setuptools/distribute/pip distributions in. "
- "You can add any number of additional --extra-search-dir paths.")
-
- parser.add_option(
- '--never-download',
- dest="never_download",
- action="store_true",
- help="Never download anything from the network. Instead, virtualenv will fail "
- "if local distributions of setuptools/distribute/pip are not present.")
-
- parser.add_option(
- '--prompt',
- dest='prompt',
- help='Provides an alternative prompt prefix for this environment')
-
- if 'extend_parser' in globals():
- extend_parser(parser)
-
- options, args = parser.parse_args()
-
- global logger
-
- if 'adjust_options' in globals():
- adjust_options(options, args)
-
- verbosity = options.verbose - options.quiet
- logger = Logger([(Logger.level_for_integer(2 - verbosity), sys.stdout)])
-
- if options.python and not os.environ.get('VIRTUALENV_INTERPRETER_RUNNING'):
- env = os.environ.copy()
- interpreter = resolve_interpreter(options.python)
- if interpreter == sys.executable:
- logger.warn('Already using interpreter %s' % interpreter)
- else:
- logger.notify('Running virtualenv with interpreter %s' % interpreter)
- env['VIRTUALENV_INTERPRETER_RUNNING'] = 'true'
- file = __file__
- if file.endswith('.pyc'):
- file = file[:-1]
- popen = subprocess.Popen([interpreter, file] + sys.argv[1:], env=env)
- raise SystemExit(popen.wait())
-
- # Force --distribute on Python 3, since setuptools is not available.
- if majver > 2:
- options.use_distribute = True
-
- if os.environ.get('PYTHONDONTWRITEBYTECODE') and not options.use_distribute:
- print(
- "The PYTHONDONTWRITEBYTECODE environment variable is "
- "not compatible with setuptools. Either use --distribute "
- "or unset PYTHONDONTWRITEBYTECODE.")
- sys.exit(2)
- if not args:
- print('You must provide a DEST_DIR')
- parser.print_help()
- sys.exit(2)
- if len(args) > 1:
- print('There must be only one argument: DEST_DIR (you gave %s)' % (
- ' '.join(args)))
- parser.print_help()
- sys.exit(2)
-
- home_dir = args[0]
-
- if os.environ.get('WORKING_ENV'):
- logger.fatal('ERROR: you cannot run virtualenv while in a workingenv')
- logger.fatal('Please deactivate your workingenv, then re-run this script')
- sys.exit(3)
-
- if 'PYTHONHOME' in os.environ:
- logger.warn('PYTHONHOME is set. You *must* activate the virtualenv before using it')
- del os.environ['PYTHONHOME']
-
- if options.relocatable:
- make_environment_relocatable(home_dir)
- return
-
- create_environment(home_dir,
- site_packages=options.system_site_packages,
- clear=options.clear,
- unzip_setuptools=options.unzip_setuptools,
- use_distribute=options.use_distribute,
- prompt=options.prompt,
- search_dirs=options.search_dirs,
- never_download=options.never_download,
- no_setuptools=options.no_setuptools,
- no_pip=options.no_pip)
- if 'after_install' in globals():
- after_install(options, home_dir)
-
-def call_subprocess(cmd, show_stdout=True,
- filter_stdout=None, cwd=None,
- raise_on_returncode=True, extra_env=None,
- remove_from_env=None):
- cmd_parts = []
- for part in cmd:
- if len(part) > 45:
- part = part[:20]+"..."+part[-20:]
- if ' ' in part or '\n' in part or '"' in part or "'" in part:
- part = '"%s"' % part.replace('"', '\\"')
- if hasattr(part, 'decode'):
- try:
- part = part.decode(sys.getdefaultencoding())
- except UnicodeDecodeError:
- part = part.decode(sys.getfilesystemencoding())
- cmd_parts.append(part)
- cmd_desc = ' '.join(cmd_parts)
- if show_stdout:
- stdout = None
- else:
- stdout = subprocess.PIPE
- logger.debug("Running command %s" % cmd_desc)
- if extra_env or remove_from_env:
- env = os.environ.copy()
- if extra_env:
- env.update(extra_env)
- if remove_from_env:
- for varname in remove_from_env:
- env.pop(varname, None)
- else:
- env = None
- try:
- proc = subprocess.Popen(
- cmd, stderr=subprocess.STDOUT, stdin=None, stdout=stdout,
- cwd=cwd, env=env)
- except Exception:
- e = sys.exc_info()[1]
- logger.fatal(
- "Error %s while executing command %s" % (e, cmd_desc))
- raise
- all_output = []
- if stdout is not None:
- stdout = proc.stdout
- encoding = sys.getdefaultencoding()
- fs_encoding = sys.getfilesystemencoding()
- while 1:
- line = stdout.readline()
- try:
- line = line.decode(encoding)
- except UnicodeDecodeError:
- line = line.decode(fs_encoding)
- if not line:
- break
- line = line.rstrip()
- all_output.append(line)
- if filter_stdout:
- level = filter_stdout(line)
- if isinstance(level, tuple):
- level, line = level
- logger.log(level, line)
- if not logger.stdout_level_matches(level):
- logger.show_progress()
- else:
- logger.info(line)
- else:
- proc.communicate()
- proc.wait()
- if proc.returncode:
- if raise_on_returncode:
- if all_output:
- logger.notify('Complete output from command %s:' % cmd_desc)
- logger.notify('\n'.join(all_output) + '\n----------------------------------------')
- raise OSError(
- "Command %s failed with error code %s"
- % (cmd_desc, proc.returncode))
- else:
- logger.warn(
- "Command %s had error code %s"
- % (cmd_desc, proc.returncode))
-
-
-def create_environment(home_dir, site_packages=False, clear=False,
- unzip_setuptools=False, use_distribute=False,
- prompt=None, search_dirs=None, never_download=False,
- no_setuptools=False, no_pip=False):
- """
- Creates a new environment in ``home_dir``.
-
- If ``site_packages`` is true, then the global ``site-packages/``
- directory will be on the path.
-
- If ``clear`` is true (default False) then the environment will
- first be cleared.
- """
- home_dir, lib_dir, inc_dir, bin_dir = path_locations(home_dir)
-
- py_executable = os.path.abspath(install_python(
- home_dir, lib_dir, inc_dir, bin_dir,
- site_packages=site_packages, clear=clear))
-
- install_distutils(home_dir)
-
- if not no_setuptools:
- if use_distribute:
- install_distribute(py_executable, unzip=unzip_setuptools,
- search_dirs=search_dirs, never_download=never_download)
- else:
- install_setuptools(py_executable, unzip=unzip_setuptools,
- search_dirs=search_dirs, never_download=never_download)
-
- if not no_pip:
- install_pip(py_executable, search_dirs=search_dirs, never_download=never_download)
-
- install_activate(home_dir, bin_dir, prompt)
-
-def is_executable_file(fpath):
- return os.path.isfile(fpath) and os.access(fpath, os.X_OK)
-
-def path_locations(home_dir):
- """Return the path locations for the environment (where libraries are,
- where scripts go, etc)"""
- # XXX: We'd use distutils.sysconfig.get_python_inc/lib but its
- # prefix arg is broken: http://bugs.python.org/issue3386
- if is_win:
- # Windows has lots of problems with executables with spaces in
- # the name; this function will remove them (using the ~1
- # format):
- mkdir(home_dir)
- if ' ' in home_dir:
- import ctypes
- GetShortPathName = ctypes.windll.kernel32.GetShortPathNameW
- size = max(len(home_dir)+1, 256)
- buf = ctypes.create_unicode_buffer(size)
- try:
- u = unicode
- except NameError:
- u = str
- ret = GetShortPathName(u(home_dir), buf, size)
- if not ret:
- print('Error: the path "%s" has a space in it' % home_dir)
- print('We could not determine the short pathname for it.')
- print('Exiting.')
- sys.exit(3)
- home_dir = str(buf.value)
- lib_dir = join(home_dir, 'Lib')
- inc_dir = join(home_dir, 'Include')
- bin_dir = join(home_dir, 'Scripts')
- if is_jython:
- lib_dir = join(home_dir, 'Lib')
- inc_dir = join(home_dir, 'Include')
- bin_dir = join(home_dir, 'bin')
- elif is_pypy:
- lib_dir = home_dir
- inc_dir = join(home_dir, 'include')
- bin_dir = join(home_dir, 'bin')
- elif not is_win:
- lib_dir = join(home_dir, 'lib', py_version)
- multiarch_exec = '/usr/bin/multiarch-platform'
- if is_executable_file(multiarch_exec):
- # In Mageia (2) and Mandriva distros the include dir must be like:
- # virtualenv/include/multiarch-x86_64-linux/python2.7
- # instead of being virtualenv/include/python2.7
- p = subprocess.Popen(multiarch_exec, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
- stdout, stderr = p.communicate()
- # stdout.strip is needed to remove newline character
- inc_dir = join(home_dir, 'include', stdout.strip(), py_version + abiflags)
- else:
- inc_dir = join(home_dir, 'include', py_version + abiflags)
- bin_dir = join(home_dir, 'bin')
- return home_dir, lib_dir, inc_dir, bin_dir
-
-
-def change_prefix(filename, dst_prefix):
- prefixes = [sys.prefix]
-
- if is_darwin:
- prefixes.extend((
- os.path.join("/Library/Python", sys.version[:3], "site-packages"),
- os.path.join(sys.prefix, "Extras", "lib", "python"),
- os.path.join("~", "Library", "Python", sys.version[:3], "site-packages"),
- # Python 2.6 no-frameworks
- os.path.join("~", ".local", "lib","python", sys.version[:3], "site-packages"),
- # System Python 2.7 on OSX Mountain Lion
- os.path.join("~", "Library", "Python", sys.version[:3], "lib", "python", "site-packages")))
-
- if hasattr(sys, 'real_prefix'):
- prefixes.append(sys.real_prefix)
- if hasattr(sys, 'base_prefix'):
- prefixes.append(sys.base_prefix)
- prefixes = list(map(os.path.expanduser, prefixes))
- prefixes = list(map(os.path.abspath, prefixes))
- # Check longer prefixes first so we don't split in the middle of a filename
- prefixes = sorted(prefixes, key=len, reverse=True)
- filename = os.path.abspath(filename)
- for src_prefix in prefixes:
- if filename.startswith(src_prefix):
- _, relpath = filename.split(src_prefix, 1)
- if src_prefix != os.sep: # sys.prefix == "/"
- assert relpath[0] == os.sep
- relpath = relpath[1:]
- return join(dst_prefix, relpath)
- assert False, "Filename %s does not start with any of these prefixes: %s" % \
- (filename, prefixes)
-
-def copy_required_modules(dst_prefix):
- import imp
- # If we are running under -p, we need to remove the current
- # directory from sys.path temporarily here, so that we
- # definitely get the modules from the site directory of
- # the interpreter we are running under, not the one
- # virtualenv.py is installed under (which might lead to py2/py3
- # incompatibility issues)
- _prev_sys_path = sys.path
- if os.environ.get('VIRTUALENV_INTERPRETER_RUNNING'):
- sys.path = sys.path[1:]
- try:
- for modname in REQUIRED_MODULES:
- if modname in sys.builtin_module_names:
- logger.info("Ignoring built-in bootstrap module: %s" % modname)
- continue
- try:
- f, filename, _ = imp.find_module(modname)
- except ImportError:
- logger.info("Cannot import bootstrap module: %s" % modname)
- else:
- if f is not None:
- f.close()
- # special-case custom readline.so on OS X, but not for pypy:
- if modname == 'readline' and sys.platform == 'darwin' and not (
- is_pypy or filename.endswith(join('lib-dynload', 'readline.so'))):
- dst_filename = join(dst_prefix, 'lib', 'python%s' % sys.version[:3], 'readline.so')
- else:
- dst_filename = change_prefix(filename, dst_prefix)
- copyfile(filename, dst_filename)
- if filename.endswith('.pyc'):
- pyfile = filename[:-1]
- if os.path.exists(pyfile):
- copyfile(pyfile, dst_filename[:-1])
- finally:
- sys.path = _prev_sys_path
-
-
-def subst_path(prefix_path, prefix, home_dir):
- prefix_path = os.path.normpath(prefix_path)
- prefix = os.path.normpath(prefix)
- home_dir = os.path.normpath(home_dir)
- if not prefix_path.startswith(prefix):
- logger.warn('Path not in prefix %r %r', prefix_path, prefix)
- return
- return prefix_path.replace(prefix, home_dir, 1)
-
-
-def install_python(home_dir, lib_dir, inc_dir, bin_dir, site_packages, clear):
- """Install just the base environment, no distutils patches etc"""
- if sys.executable.startswith(bin_dir):
- print('Please use the *system* python to run this script')
- return
-
- if clear:
- rmtree(lib_dir)
- ## FIXME: why not delete it?
- ## Maybe it should delete everything with #!/path/to/venv/python in it
- logger.notify('Not deleting %s', bin_dir)
-
- if hasattr(sys, 'real_prefix'):
- logger.notify('Using real prefix %r' % sys.real_prefix)
- prefix = sys.real_prefix
- elif hasattr(sys, 'base_prefix'):
- logger.notify('Using base prefix %r' % sys.base_prefix)
- prefix = sys.base_prefix
- else:
- prefix = sys.prefix
- mkdir(lib_dir)
- fix_lib64(lib_dir)
- stdlib_dirs = [os.path.dirname(os.__file__)]
- if is_win:
- stdlib_dirs.append(join(os.path.dirname(stdlib_dirs[0]), 'DLLs'))
- elif is_darwin:
- stdlib_dirs.append(join(stdlib_dirs[0], 'site-packages'))
- if hasattr(os, 'symlink'):
- logger.info('Symlinking Python bootstrap modules')
- else:
- logger.info('Copying Python bootstrap modules')
- logger.indent += 2
- try:
- # copy required files...
- for stdlib_dir in stdlib_dirs:
- if not os.path.isdir(stdlib_dir):
- continue
- for fn in os.listdir(stdlib_dir):
- bn = os.path.splitext(fn)[0]
- if fn != 'site-packages' and bn in REQUIRED_FILES:
- copyfile(join(stdlib_dir, fn), join(lib_dir, fn))
- # ...and modules
- copy_required_modules(home_dir)
- finally:
- logger.indent -= 2
- mkdir(join(lib_dir, 'site-packages'))
- import site
- site_filename = site.__file__
- if site_filename.endswith('.pyc'):
- site_filename = site_filename[:-1]
- elif site_filename.endswith('$py.class'):
- site_filename = site_filename.replace('$py.class', '.py')
- site_filename_dst = change_prefix(site_filename, home_dir)
- site_dir = os.path.dirname(site_filename_dst)
- writefile(site_filename_dst, SITE_PY)
- writefile(join(site_dir, 'orig-prefix.txt'), prefix)
- site_packages_filename = join(site_dir, 'no-global-site-packages.txt')
- if not site_packages:
- writefile(site_packages_filename, '')
-
- if is_pypy or is_win:
- stdinc_dir = join(prefix, 'include')
- else:
- stdinc_dir = join(prefix, 'include', py_version + abiflags)
- if os.path.exists(stdinc_dir):
- copyfile(stdinc_dir, inc_dir)
- else:
- logger.debug('No include dir %s' % stdinc_dir)
-
- platinc_dir = distutils.sysconfig.get_python_inc(plat_specific=1)
- if platinc_dir != stdinc_dir:
- platinc_dest = distutils.sysconfig.get_python_inc(
- plat_specific=1, prefix=home_dir)
- if platinc_dir == platinc_dest:
- # Do platinc_dest manually due to a CPython bug;
- # not http://bugs.python.org/issue3386 but a close cousin
- platinc_dest = subst_path(platinc_dir, prefix, home_dir)
- if platinc_dest:
- # PyPy's stdinc_dir and prefix are relative to the original binary
- # (traversing virtualenvs), whereas the platinc_dir is relative to
- # the inner virtualenv and ignores the prefix argument.
- # This seems more evolved than designed.
- copyfile(platinc_dir, platinc_dest)
-
- # pypy never uses exec_prefix, just ignore it
- if sys.exec_prefix != prefix and not is_pypy:
- if is_win:
- exec_dir = join(sys.exec_prefix, 'lib')
- elif is_jython:
- exec_dir = join(sys.exec_prefix, 'Lib')
- else:
- exec_dir = join(sys.exec_prefix, 'lib', py_version)
- for fn in os.listdir(exec_dir):
- copyfile(join(exec_dir, fn), join(lib_dir, fn))
-
- if is_jython:
- # Jython has either jython-dev.jar and javalib/ dir, or just
- # jython.jar
- for name in 'jython-dev.jar', 'javalib', 'jython.jar':
- src = join(prefix, name)
- if os.path.exists(src):
- copyfile(src, join(home_dir, name))
- # XXX: registry should always exist after Jython 2.5rc1
- src = join(prefix, 'registry')
- if os.path.exists(src):
- copyfile(src, join(home_dir, 'registry'), symlink=False)
- copyfile(join(prefix, 'cachedir'), join(home_dir, 'cachedir'),
- symlink=False)
-
- mkdir(bin_dir)
- py_executable = join(bin_dir, os.path.basename(sys.executable))
- if 'Python.framework' in prefix:
- # OS X framework builds cause validation to break
- # https://github.com/pypa/virtualenv/issues/322
- if os.environ.get('__PYVENV_LAUNCHER__'):
- os.unsetenv('__PYVENV_LAUNCHER__')
- if re.search(r'/Python(?:-32|-64)*$', py_executable):
- # The name of the python executable is not quite what
- # we want, rename it.
- py_executable = os.path.join(
- os.path.dirname(py_executable), 'python')
-
- logger.notify('New %s executable in %s', expected_exe, py_executable)
- pcbuild_dir = os.path.dirname(sys.executable)
- pyd_pth = os.path.join(lib_dir, 'site-packages', 'virtualenv_builddir_pyd.pth')
- if is_win and os.path.exists(os.path.join(pcbuild_dir, 'build.bat')):
- logger.notify('Detected python running from build directory %s', pcbuild_dir)
- logger.notify('Writing .pth file linking to build directory for *.pyd files')
- writefile(pyd_pth, pcbuild_dir)
- else:
- pcbuild_dir = None
- if os.path.exists(pyd_pth):
- logger.info('Deleting %s (not Windows env or not build directory python)' % pyd_pth)
- os.unlink(pyd_pth)
-
- if sys.executable != py_executable:
- ## FIXME: could I just hard link?
- executable = sys.executable
- shutil.copyfile(executable, py_executable)
- make_exe(py_executable)
- if is_win or is_cygwin:
- pythonw = os.path.join(os.path.dirname(sys.executable), 'pythonw.exe')
- if os.path.exists(pythonw):
- logger.info('Also created pythonw.exe')
- shutil.copyfile(pythonw, os.path.join(os.path.dirname(py_executable), 'pythonw.exe'))
- python_d = os.path.join(os.path.dirname(sys.executable), 'python_d.exe')
- python_d_dest = os.path.join(os.path.dirname(py_executable), 'python_d.exe')
- if os.path.exists(python_d):
- logger.info('Also created python_d.exe')
- shutil.copyfile(python_d, python_d_dest)
- elif os.path.exists(python_d_dest):
- logger.info('Removed python_d.exe as it is no longer at the source')
- os.unlink(python_d_dest)
- # we need to copy the DLL to enforce that windows will load the correct one.
- # may not exist if we are cygwin.
- py_executable_dll = 'python%s%s.dll' % (
- sys.version_info[0], sys.version_info[1])
- py_executable_dll_d = 'python%s%s_d.dll' % (
- sys.version_info[0], sys.version_info[1])
- pythondll = os.path.join(os.path.dirname(sys.executable), py_executable_dll)
- pythondll_d = os.path.join(os.path.dirname(sys.executable), py_executable_dll_d)
- pythondll_d_dest = os.path.join(os.path.dirname(py_executable), py_executable_dll_d)
- if os.path.exists(pythondll):
- logger.info('Also created %s' % py_executable_dll)
- shutil.copyfile(pythondll, os.path.join(os.path.dirname(py_executable), py_executable_dll))
- if os.path.exists(pythondll_d):
- logger.info('Also created %s' % py_executable_dll_d)
- shutil.copyfile(pythondll_d, pythondll_d_dest)
- elif os.path.exists(pythondll_d_dest):
- logger.info('Removed %s as the source does not exist' % pythondll_d_dest)
- os.unlink(pythondll_d_dest)
- if is_pypy:
- # make a symlink python --> pypy-c
- python_executable = os.path.join(os.path.dirname(py_executable), 'python')
- if sys.platform in ('win32', 'cygwin'):
- python_executable += '.exe'
- logger.info('Also created executable %s' % python_executable)
- copyfile(py_executable, python_executable)
-
- if is_win:
- for name in 'libexpat.dll', 'libpypy.dll', 'libpypy-c.dll', 'libeay32.dll', 'ssleay32.dll', 'sqlite.dll':
- src = join(prefix, name)
- if os.path.exists(src):
- copyfile(src, join(bin_dir, name))
-
- if os.path.splitext(os.path.basename(py_executable))[0] != expected_exe:
- secondary_exe = os.path.join(os.path.dirname(py_executable),
- expected_exe)
- py_executable_ext = os.path.splitext(py_executable)[1]
- if py_executable_ext == '.exe':
- # python2.4 gives an extension of '.4' :P
- secondary_exe += py_executable_ext
- if os.path.exists(secondary_exe):
- logger.warn('Not overwriting existing %s script %s (you must use %s)'
- % (expected_exe, secondary_exe, py_executable))
- else:
- logger.notify('Also creating executable in %s' % secondary_exe)
- shutil.copyfile(sys.executable, secondary_exe)
- make_exe(secondary_exe)
-
- if '.framework' in prefix:
- if 'Python.framework' in prefix:
- logger.debug('MacOSX Python framework detected')
- # Make sure we use the the embedded interpreter inside
- # the framework, even if sys.executable points to
- # the stub executable in ${sys.prefix}/bin
- # See http://groups.google.com/group/python-virtualenv/
- # browse_thread/thread/17cab2f85da75951
- original_python = os.path.join(
- prefix, 'Resources/Python.app/Contents/MacOS/Python')
- if 'EPD' in prefix:
- logger.debug('EPD framework detected')
- original_python = os.path.join(prefix, 'bin/python')
- shutil.copy(original_python, py_executable)
-
- # Copy the framework's dylib into the virtual
- # environment
- virtual_lib = os.path.join(home_dir, '.Python')
-
- if os.path.exists(virtual_lib):
- os.unlink(virtual_lib)
- copyfile(
- os.path.join(prefix, 'Python'),
- virtual_lib)
-
- # And then change the install_name of the copied python executable
- try:
- mach_o_change(py_executable,
- os.path.join(prefix, 'Python'),
- '@executable_path/../.Python')
- except:
- e = sys.exc_info()[1]
- logger.warn("Could not call mach_o_change: %s. "
- "Trying to call install_name_tool instead." % e)
- try:
- call_subprocess(
- ["install_name_tool", "-change",
- os.path.join(prefix, 'Python'),
- '@executable_path/../.Python',
- py_executable])
- except:
- logger.fatal("Could not call install_name_tool -- you must "
- "have Apple's development tools installed")
- raise
-
- if not is_win:
- # Ensure that 'python', 'pythonX' and 'pythonX.Y' all exist
- py_exe_version_major = 'python%s' % sys.version_info[0]
- py_exe_version_major_minor = 'python%s.%s' % (
- sys.version_info[0], sys.version_info[1])
- py_exe_no_version = 'python'
- required_symlinks = [ py_exe_no_version, py_exe_version_major,
- py_exe_version_major_minor ]
-
- py_executable_base = os.path.basename(py_executable)
-
- if py_executable_base in required_symlinks:
- # Don't try to symlink to yourself.
- required_symlinks.remove(py_executable_base)
-
- for pth in required_symlinks:
- full_pth = join(bin_dir, pth)
- if os.path.exists(full_pth):
- os.unlink(full_pth)
- os.symlink(py_executable_base, full_pth)
-
- if is_win and ' ' in py_executable:
- # There's a bug with subprocess on Windows when using a first
- # argument that has a space in it. Instead we have to quote
- # the value:
- py_executable = '"%s"' % py_executable
- # NOTE: keep this check as one line, cmd.exe doesn't cope with line breaks
- cmd = [py_executable, '-c', 'import sys;out=sys.stdout;'
- 'getattr(out, "buffer", out).write(sys.prefix.encode("utf-8"))']
- logger.info('Testing executable with %s %s "%s"' % tuple(cmd))
- try:
- proc = subprocess.Popen(cmd,
- stdout=subprocess.PIPE)
- proc_stdout, proc_stderr = proc.communicate()
- except OSError:
- e = sys.exc_info()[1]
- if e.errno == errno.EACCES:
- logger.fatal('ERROR: The executable %s could not be run: %s' % (py_executable, e))
- sys.exit(100)
- else:
- raise e
-
- proc_stdout = proc_stdout.strip().decode("utf-8")
- proc_stdout = os.path.normcase(os.path.abspath(proc_stdout))
- norm_home_dir = os.path.normcase(os.path.abspath(home_dir))
- if hasattr(norm_home_dir, 'decode'):
- norm_home_dir = norm_home_dir.decode(sys.getfilesystemencoding())
- if proc_stdout != norm_home_dir:
- logger.fatal(
- 'ERROR: The executable %s is not functioning' % py_executable)
- logger.fatal(
- 'ERROR: It thinks sys.prefix is %r (should be %r)'
- % (proc_stdout, norm_home_dir))
- logger.fatal(
- 'ERROR: virtualenv is not compatible with this system or executable')
- if is_win:
- logger.fatal(
- 'Note: some Windows users have reported this error when they '
- 'installed Python for "Only this user" or have multiple '
- 'versions of Python installed. Copying the appropriate '
- 'PythonXX.dll to the virtualenv Scripts/ directory may fix '
- 'this problem.')
- sys.exit(100)
- else:
- logger.info('Got sys.prefix result: %r' % proc_stdout)
-
- pydistutils = os.path.expanduser('~/.pydistutils.cfg')
- if os.path.exists(pydistutils):
- logger.notify('Please make sure you remove any previous custom paths from '
- 'your %s file.' % pydistutils)
- ## FIXME: really this should be calculated earlier
-
- fix_local_scheme(home_dir)
-
- if site_packages:
- if os.path.exists(site_packages_filename):
- logger.info('Deleting %s' % site_packages_filename)
- os.unlink(site_packages_filename)
-
- return py_executable
-
-
-def install_activate(home_dir, bin_dir, prompt=None):
- home_dir = os.path.abspath(home_dir)
- if is_win or is_jython and os._name == 'nt':
- files = {
- 'activate.bat': ACTIVATE_BAT,
- 'deactivate.bat': DEACTIVATE_BAT,
- 'activate.ps1': ACTIVATE_PS,
- }
-
- # MSYS needs paths of the form /c/path/to/file
- drive, tail = os.path.splitdrive(home_dir.replace(os.sep, '/'))
- home_dir_msys = (drive and "/%s%s" or "%s%s") % (drive[:1], tail)
-
- # Run-time conditional enables (basic) Cygwin compatibility
- home_dir_sh = ("""$(if [ "$OSTYPE" "==" "cygwin" ]; then cygpath -u '%s'; else echo '%s'; fi;)""" %
- (home_dir, home_dir_msys))
- files['activate'] = ACTIVATE_SH.replace('__VIRTUAL_ENV__', home_dir_sh)
-
- else:
- files = {'activate': ACTIVATE_SH}
-
- # suppling activate.fish in addition to, not instead of, the
- # bash script support.
- files['activate.fish'] = ACTIVATE_FISH
-
- # same for csh/tcsh support...
- files['activate.csh'] = ACTIVATE_CSH
-
- files['activate_this.py'] = ACTIVATE_THIS
- if hasattr(home_dir, 'decode'):
- home_dir = home_dir.decode(sys.getfilesystemencoding())
- vname = os.path.basename(home_dir)
- for name, content in files.items():
- content = content.replace('__VIRTUAL_PROMPT__', prompt or '')
- content = content.replace('__VIRTUAL_WINPROMPT__', prompt or '(%s)' % vname)
- content = content.replace('__VIRTUAL_ENV__', home_dir)
- content = content.replace('__VIRTUAL_NAME__', vname)
- content = content.replace('__BIN_NAME__', os.path.basename(bin_dir))
- writefile(os.path.join(bin_dir, name), content)
-
-def install_distutils(home_dir):
- distutils_path = change_prefix(distutils.__path__[0], home_dir)
- mkdir(distutils_path)
- ## FIXME: maybe this prefix setting should only be put in place if
- ## there's a local distutils.cfg with a prefix setting?
- home_dir = os.path.abspath(home_dir)
- ## FIXME: this is breaking things, removing for now:
- #distutils_cfg = DISTUTILS_CFG + "\n[install]\nprefix=%s\n" % home_dir
- writefile(os.path.join(distutils_path, '__init__.py'), DISTUTILS_INIT)
- writefile(os.path.join(distutils_path, 'distutils.cfg'), DISTUTILS_CFG, overwrite=False)
-
-def fix_local_scheme(home_dir):
- """
- Platforms that use the "posix_local" install scheme (like Ubuntu with
- Python 2.7) need to be given an additional "local" location, sigh.
- """
- try:
- import sysconfig
- except ImportError:
- pass
- else:
- if sysconfig._get_default_scheme() == 'posix_local':
- local_path = os.path.join(home_dir, 'local')
- if not os.path.exists(local_path):
- os.mkdir(local_path)
- for subdir_name in os.listdir(home_dir):
- if subdir_name == 'local':
- continue
- os.symlink(os.path.abspath(os.path.join(home_dir, subdir_name)), \
- os.path.join(local_path, subdir_name))
-
-def fix_lib64(lib_dir):
- """
- Some platforms (particularly Gentoo on x64) put things in lib64/pythonX.Y
- instead of lib/pythonX.Y. If this is such a platform we'll just create a
- symlink so lib64 points to lib
- """
- if [p for p in distutils.sysconfig.get_config_vars().values()
- if isinstance(p, basestring) and 'lib64' in p]:
- logger.debug('This system uses lib64; symlinking lib64 to lib')
- assert os.path.basename(lib_dir) == 'python%s' % sys.version[:3], (
- "Unexpected python lib dir: %r" % lib_dir)
- lib_parent = os.path.dirname(lib_dir)
- top_level = os.path.dirname(lib_parent)
- lib_dir = os.path.join(top_level, 'lib')
- lib64_link = os.path.join(top_level, 'lib64')
- assert os.path.basename(lib_parent) == 'lib', (
- "Unexpected parent dir: %r" % lib_parent)
- if os.path.lexists(lib64_link):
- return
- os.symlink('lib', lib64_link)
-
-def resolve_interpreter(exe):
- """
- If the executable given isn't an absolute path, search $PATH for the interpreter
- """
- if os.path.abspath(exe) != exe:
- paths = os.environ.get('PATH', '').split(os.pathsep)
- for path in paths:
- if os.path.exists(os.path.join(path, exe)):
- exe = os.path.join(path, exe)
- break
- if not os.path.exists(exe):
- logger.fatal('The executable %s (from --python=%s) does not exist' % (exe, exe))
- raise SystemExit(3)
- if not is_executable(exe):
- logger.fatal('The executable %s (from --python=%s) is not executable' % (exe, exe))
- raise SystemExit(3)
- return exe
-
-def is_executable(exe):
- """Checks a file is executable"""
- return os.access(exe, os.X_OK)
-
-############################################################
-## Relocating the environment:
-
-def make_environment_relocatable(home_dir):
- """
- Makes the already-existing environment use relative paths, and takes out
- the #!-based environment selection in scripts.
- """
- home_dir, lib_dir, inc_dir, bin_dir = path_locations(home_dir)
- activate_this = os.path.join(bin_dir, 'activate_this.py')
- if not os.path.exists(activate_this):
- logger.fatal(
- 'The environment doesn\'t have a file %s -- please re-run virtualenv '
- 'on this environment to update it' % activate_this)
- fixup_scripts(home_dir)
- fixup_pth_and_egg_link(home_dir)
- ## FIXME: need to fix up distutils.cfg
-
-OK_ABS_SCRIPTS = ['python', 'python%s' % sys.version[:3],
- 'activate', 'activate.bat', 'activate_this.py']
-
-def fixup_scripts(home_dir):
- # This is what we expect at the top of scripts:
- shebang = '#!%s/bin/python' % os.path.normcase(os.path.abspath(home_dir))
- # This is what we'll put:
- new_shebang = '#!/usr/bin/env python%s' % sys.version[:3]
- if is_win:
- bin_suffix = 'Scripts'
- else:
- bin_suffix = 'bin'
- bin_dir = os.path.join(home_dir, bin_suffix)
- home_dir, lib_dir, inc_dir, bin_dir = path_locations(home_dir)
- for filename in os.listdir(bin_dir):
- filename = os.path.join(bin_dir, filename)
- if not os.path.isfile(filename):
- # ignore subdirs, e.g. .svn ones.
- continue
- f = open(filename, 'rb')
- try:
- try:
- lines = f.read().decode('utf-8').splitlines()
- except UnicodeDecodeError:
- # This is probably a binary program instead
- # of a script, so just ignore it.
- continue
- finally:
- f.close()
- if not lines:
- logger.warn('Script %s is an empty file' % filename)
- continue
- if not lines[0].strip().startswith(shebang):
- if os.path.basename(filename) in OK_ABS_SCRIPTS:
- logger.debug('Cannot make script %s relative' % filename)
- elif lines[0].strip() == new_shebang:
- logger.info('Script %s has already been made relative' % filename)
- else:
- logger.warn('Script %s cannot be made relative (it\'s not a normal script that starts with %s)'
- % (filename, shebang))
- continue
- logger.notify('Making script %s relative' % filename)
- script = relative_script([new_shebang] + lines[1:])
- f = open(filename, 'wb')
- f.write('\n'.join(script).encode('utf-8'))
- f.close()
-
-def relative_script(lines):
- "Return a script that'll work in a relocatable environment."
- activate = "import os; activate_this=os.path.join(os.path.dirname(os.path.realpath(__file__)), 'activate_this.py'); execfile(activate_this, dict(__file__=activate_this)); del os, activate_this"
- # Find the last future statement in the script. If we insert the activation
- # line before a future statement, Python will raise a SyntaxError.
- activate_at = None
- for idx, line in reversed(list(enumerate(lines))):
- if line.split()[:3] == ['from', '__future__', 'import']:
- activate_at = idx + 1
- break
- if activate_at is None:
- # Activate after the shebang.
- activate_at = 1
- return lines[:activate_at] + ['', activate, ''] + lines[activate_at:]
-
-def fixup_pth_and_egg_link(home_dir, sys_path=None):
- """Makes .pth and .egg-link files use relative paths"""
- home_dir = os.path.normcase(os.path.abspath(home_dir))
- if sys_path is None:
- sys_path = sys.path
- for path in sys_path:
- if not path:
- path = '.'
- if not os.path.isdir(path):
- continue
- path = os.path.normcase(os.path.abspath(path))
- if not path.startswith(home_dir):
- logger.debug('Skipping system (non-environment) directory %s' % path)
- continue
- for filename in os.listdir(path):
- filename = os.path.join(path, filename)
- if filename.endswith('.pth'):
- if not os.access(filename, os.W_OK):
- logger.warn('Cannot write .pth file %s, skipping' % filename)
- else:
- fixup_pth_file(filename)
- if filename.endswith('.egg-link'):
- if not os.access(filename, os.W_OK):
- logger.warn('Cannot write .egg-link file %s, skipping' % filename)
- else:
- fixup_egg_link(filename)
-
-def fixup_pth_file(filename):
- lines = []
- prev_lines = []
- f = open(filename)
- prev_lines = f.readlines()
- f.close()
- for line in prev_lines:
- line = line.strip()
- if (not line or line.startswith('#') or line.startswith('import ')
- or os.path.abspath(line) != line):
- lines.append(line)
- else:
- new_value = make_relative_path(filename, line)
- if line != new_value:
- logger.debug('Rewriting path %s as %s (in %s)' % (line, new_value, filename))
- lines.append(new_value)
- if lines == prev_lines:
- logger.info('No changes to .pth file %s' % filename)
- return
- logger.notify('Making paths in .pth file %s relative' % filename)
- f = open(filename, 'w')
- f.write('\n'.join(lines) + '\n')
- f.close()
-
-def fixup_egg_link(filename):
- f = open(filename)
- link = f.readline().strip()
- f.close()
- if os.path.abspath(link) != link:
- logger.debug('Link in %s already relative' % filename)
- return
- new_link = make_relative_path(filename, link)
- logger.notify('Rewriting link %s in %s as %s' % (link, filename, new_link))
- f = open(filename, 'w')
- f.write(new_link)
- f.close()
-
-def make_relative_path(source, dest, dest_is_directory=True):
- """
- Make a filename relative, where the filename is dest, and it is
- being referred to from the filename source.
-
- >>> make_relative_path('/usr/share/something/a-file.pth',
- ... '/usr/share/another-place/src/Directory')
- '../another-place/src/Directory'
- >>> make_relative_path('/usr/share/something/a-file.pth',
- ... '/home/user/src/Directory')
- '../../../home/user/src/Directory'
- >>> make_relative_path('/usr/share/a-file.pth', '/usr/share/')
- './'
- """
- source = os.path.dirname(source)
- if not dest_is_directory:
- dest_filename = os.path.basename(dest)
- dest = os.path.dirname(dest)
- dest = os.path.normpath(os.path.abspath(dest))
- source = os.path.normpath(os.path.abspath(source))
- dest_parts = dest.strip(os.path.sep).split(os.path.sep)
- source_parts = source.strip(os.path.sep).split(os.path.sep)
- while dest_parts and source_parts and dest_parts[0] == source_parts[0]:
- dest_parts.pop(0)
- source_parts.pop(0)
- full_parts = ['..']*len(source_parts) + dest_parts
- if not dest_is_directory:
- full_parts.append(dest_filename)
- if not full_parts:
- # Special case for the current directory (otherwise it'd be '')
- return './'
- return os.path.sep.join(full_parts)
-
-
-
-############################################################
-## Bootstrap script creation:
-
-def create_bootstrap_script(extra_text, python_version=''):
- """
- Creates a bootstrap script, which is like this script but with
- extend_parser, adjust_options, and after_install hooks.
-
- This returns a string that (written to disk of course) can be used
- as a bootstrap script with your own customizations. The script
- will be the standard virtualenv.py script, with your extra text
- added (your extra text should be Python code).
-
- If you include these functions, they will be called:
-
- ``extend_parser(optparse_parser)``:
- You can add or remove options from the parser here.
-
- ``adjust_options(options, args)``:
- You can change options here, or change the args (if you accept
- different kinds of arguments, be sure you modify ``args`` so it is
- only ``[DEST_DIR]``).
-
- ``after_install(options, home_dir)``:
-
- After everything is installed, this function is called. This
- is probably the function you are most likely to use. An
- example would be::
-
- def after_install(options, home_dir):
- subprocess.call([join(home_dir, 'bin', 'easy_install'),
- 'MyPackage'])
- subprocess.call([join(home_dir, 'bin', 'my-package-script'),
- 'setup', home_dir])
-
- This example immediately installs a package, and runs a setup
- script from that package.
-
- If you provide something like ``python_version='2.5'`` then the
- script will start with ``#!/usr/bin/env python2.5`` instead of
- ``#!/usr/bin/env python``. You can use this when the script must
- be run with a particular Python version.
- """
- filename = __file__
- if filename.endswith('.pyc'):
- filename = filename[:-1]
- f = codecs.open(filename, 'r', encoding='utf-8')
- content = f.read()
- f.close()
- py_exe = 'python%s' % python_version
- content = (('#!/usr/bin/env %s\n' % py_exe)
- + '## WARNING: This file is generated\n'
- + content)
- return content.replace('##EXT' 'END##', extra_text)
-
-
-
-default_target_dir = 'venv'
-
-pip_install_packages = filter(len, open('requirements.txt').readlines())
-
-import os
-import subprocess
-import sys
-
-def adjust_options(options, args):
- if len(args)==0:
- os.chdir(os.path.dirname(__file__))
- args.append(default_target_dir)
-
-def after_install(options, home_dir):
- from os.path import join
- pip = join(home_dir, 'bin/pip')
- if not os.path.exists(pip):
- # on windows
- pip = join(home_dir, 'Scripts/pip.exe')
- if not os.path.exists(pip):
- print "error", pip, "is missing"
- if sys.version_info < (2, 7):
- subprocess.call([pip, 'install', 'importlib'])
- for prog in pip_install_packages:
- subprocess.call([pip, 'install', prog])
-
-
-
-def convert(s):
- b = base64.b64decode(s.encode('ascii'))
- return zlib.decompress(b).decode('utf-8')
-
-##file site.py
-SITE_PY = convert("""
-eJzFPf1z2zaWv/OvwMqToZTIdOK0vR2nzo2TOK3v3MTbpLO5dT1aSoIs1hTJEqRl7c3d337vAwAB
-kpLtTXdO04klEnh4eHhfeHgPHQwGJ0Uhs7lY5fM6lULJuJwtRRFXSyUWeSmqZVLO94u4rDbwdHYT
-X0slqlyojYqwVRQET7/yEzwVn5eJMijAt7iu8lVcJbM4TTciWRV5Wcm5mNdlkl2LJEuqJE6Tf0CL
-PIvE06/HIDjLBMw8TWQpbmWpAK4S+UJcbKplnolhXeCcX0Tfxi9HY6FmZVJU0KDUOANFlnEVZFLO
-AU1oWSsgZVLJfVXIWbJIZrbhOq/TuSjSeCbF3//OU6OmYRiofCXXS1lKkQEyAFMCrALxgK9JKWb5
-XEZCvJGzGAfg5w2xAoY2xjVTSMYsF2meXcOcMjmTSsXlRgyndUWACGUxzwGnBDCokjQN1nl5o0aw
-pLQea3gkYmYPfzLMHjBPHL/LOYDjxyz4JUvuxgwbuAfBVUtmm1IukjsRI1j4Ke/kbKKfDZOFmCeL
-BdAgq0bYJGAElEiT6UFBy/G9XqHXB4SV5coYxpCIMjfml9QjCs4qEacK2LYukEaKMH8np0mcATWy
-WxgOIAJJg75x5omq7Dg0O5EDgBLXsQIpWSkxXMVJBsz6UzwjtP+aZPN8rUZEAVgtJX6rVeXOf9hD
-AGjtEGAc4GKZ1ayzNLmR6WYECHwG7Eup6rRCgZgnpZxVeZlIRQAAtY2Qd4D0WMSl1CRkzjRyOyb6
-E02SDBcWBQwFHl8iSRbJdV2ShIlFApwLXPH+48/i3embs5MPmscMMJbZ6xXgDFBooR2cYABxUKvy
-IM1BoKPgHP+IeD5HIbvG8QGvpsHBvSsdDGHuRdTu4yw4kF0vrh4G5liBMqGxAur339BlrJZAn/+5
-Z72D4GQbVWji/G29zEEms3glxTJm/kLOCL7XcF5HRbV8BdygEE4FpFK4OIhggvCAJC7NhnkmRQEs
-liaZHAVAoSm19VcRWOFDnu3TWrc4ASCUQQYvnWcjGjGTMNEurFeoL0zjDc1MNwnsOq/ykhQH8H82
-I12UxtkN4aiIofjbVF4nWYYIIS8E4V5IA6ubBDhxHolzakV6wTQSIWsvbokiUQMvIdMBT8q7eFWk
-cszii7p1txqhwWQlzFqnzHHQsiL1SqvWTLWX9w6jLy2uIzSrZSkBeD31hG6R52MxBZ1N2BTxisWr
-WufEOUGPPFEn5AlqCX3xO1D0RKl6Je1L5BXQLMRQwSJP03wNJDsKAiH2sJExyj5zwlt4B/8CXPw3
-ldVsGQTOSBawBoXIbwOFQMAkyExztUbC4zbNym0lk2SsKfJyLksa6mHEPmDEH9gY5xp8yCtt1Hi6
-uMr5KqlQJU21yUzY4mVhxfrxFc8bpgGWWxHNTNOGTiucXlos46k0LslULlAS9CK9sssOYwY9Y5It
-rsSKrQy8A7LIhC1Iv2JBpbOoJDkBAIOFL86Sok6pkUIGEzEMtCoI/ipGk55rZwnYm81ygAqJzfcM
-7A/g9g8Qo/UyAfrMAAJoGNRSsHzTpCrRQWj0UeAbfdOfxwdOPVto28RDLuIk1VY+zoIzenhaliS+
-M1lgr7EmhoIZZhW6dtcZ0BHFfDAYBIFxhzbKfM1VUJWbI2AFYcaZTKZ1goZvMkFTr3+ogEcRzsBe
-N9vOwgMNYTp9ACo5XRZlvsLXdm6fQJnAWNgj2BMXpGUkO8geJ75C8rkqvTBN0XY77CxQDwUXP5++
-P/ty+kkci8tGpY3b+uwKxjzNYmBrsgjAVK1hG10GLVHxJaj7xHsw78QUYM+oN4mvjKsaeBdQ/1zW
-9BqmMfNeBqcfTt6cn05++XT68+TT2edTQBDsjAz2aMpoHmtwGFUEwgFcOVeRtq9Bpwc9eHPyyT4I
-JomafPcNsBs8GV7LCpi4HMKMxyJcxXcKGDQcU9MR4thpABY8HI3Ea3H49OnLQ4JWbIoNAAOz6zTF
-hxNt0SdJtsjDETX+jV36Y1ZS2n+7PPrmShwfi/C3+DYOA/ChmqbMEj+ROH3eFBK6VvBnmKtREMzl
-AkTvRqKADp+SXzziDrAk0DLXdvq3PMnMe+ZKdwjSH0PqAThMJrM0VgobTyYhEIE69HygQ8TONUrd
-EDoWG7frSKOCn1LCwmbYZYz/9KAYT6kfosEoul1MIxDX1SxWklvR9KHfZII6azIZ6gFBmEliwOFi
-NRQK0wR1VpmAX0uchzpsqvIUfyJ81AIkgLi1Qi2Ji6S3TtFtnNZSDZ1JARGHwxYZUdEmivgRXJQh
-WOJm6UajNjUNz0AzIF+agxYtW5TDzx74O6CuzCYON3q892KaIab/wTsNwgFczhDVvVItKKwdxcXp
-hXj5/HAf3RnYc84tdbzmaKGTrJb24QJWy8gDI8y9jLy4dFmgnsWnR7thriK7Ml1WWOglLuUqv5Vz
-wBYZ2Fll8TO9gZ05zGMWwyqCXid/gFWo8Rtj3Ify7EFa0HcA6q0Iill/s/R7HAyQmQJFxBtrIrXe
-9bMpLMr8NkFnY7rRL8FWgrJEi2kcm8BZOI/J0CSChgAvOENKrWUI6rCs2WElvBEk2ot5o1gjAneO
-mvqKvt5k+Tqb8E74GJXucGRZFwVLMy82aJZgT7wHKwRI5rCxa4jGUMDlFyhb+4A8TB+mC5SlvQUA
-AkOvaLvmwDJbPZoi7xpxWIQxeiVIeEuJ/sKtGYK2WoYYDiR6G9kHRksgJJicVXBWNWgmQ1kzzWBg
-hyQ+151HvAX1AbSoGIHZHGpo3MjQ7/IIlLM4d5WS0w8t8pcvX5ht1JLiK4jYFCeNLsSCjGVUbMCw
-JqATjEfG0RpigzU4twCmVpo1xf4nkRfsjcF6XmjZBj8AdndVVRwdHKzX60hHF/Ly+kAtDr7983ff
-/fk568T5nPgHpuNIiw61RQf0Dj3a6HtjgV6blWvxY5L53EiwhpK8MnJFEb8f6mSei6P9kdWfyMWN
-mcZ/jSsDCmRiBmUqA20HDUZP1P6T6KUaiCdknW3b4Yj9Em1SrRXzrS70qHLwBMBvmeU1muqGE5R4
-BtYNduhzOa2vQzu4ZyPND5gqyunQ8sD+iyvEwOcMw1fGFE9QSxBboMV3SP8zs01M3pHWEEheNFGd
-3fOmX4sZ4s4fLu/W13SExswwUcgdKBF+kwcLoG3clRz8aNcW7Z7j2pqPZwiMpQ8M82rHcoiCQ7jg
-WoxdqXO4Gj1ekKY1q2ZQMK5qBAUNTuKUqa3BkY0MESR6N2azzwurWwCdWpFDEx8wqwAt3HE61q7N
-Co4nhDxwLF7QEwku8lHn3XNe2jpNKaDT4lGPKgzYW2i00znw5dAAGItB+cuAW5ptysfWovAa9ADL
-OQaEDLboMBO+cX3Awd6gh506Vn9bb6ZxHwhcpCHHoh4EnVA+5hFKBdJUDP2e21jcErc72E6LQ0xl
-lolEWm0Rrrby6BWqnYZpkWSoe51FimZpDl6x1YrESM1731mgfRA+7jNmWgI1GRpyOI2OydvzBDDU
-7TB8dl1joMGNwyBGq0SRdUMyLeEfcCsovkHBKKAlQbNgHipl/sT+AJmz89VftrCHJTQyhNt0mxvS
-sRgajnm/J5CMOhoDUpABCbvCSK4jq4MUOMxZIE+44bXcKt0EI1IgZ44FITUDuNNLb4ODTyI8ASEJ
-Rch3lZKFeCYGsHxtUX2Y7v5DudQEIYZOA3IVdPTi2I1sOFGN41aUw2doP75BZyVFDhw8BZfHDfS7
-bG6Y1gZdwFn3FbdFCjQyxWEGIxfVK0MYN5j8p2OnRUMsM4hhKG8g70jHjDQK7HJr0LDgBoy35u2x
-9GM3YoF9h2GuDuXqDvZ/YZmoWa5Cipm0YxfuR3NFlzYW2/NkOoA/3gIMRlceJJnq+AVGWf6JQUIP
-etgH3ZsshkXmcblOspAUmKbfsb80HTwsKT0jd/CJtlMHMFGMeB68L0FA6OjzAMQJNQHsymWotNvf
-BbtzigMLl7sPPLf58ujlVZe4420RHvvpX6rTu6qMFa5WyovGQoGr1TXgqHRhcnG20YeX+nAbtwll
-rmAXKT5++iKQEBzXXcebx029YXjE5t45eR+DOui1e8nVmh2xCyCCWhEZ5SB8PEc+HNnHTm7HxB4B
-5FEMs2NRDCTNJ/8MnF0LBWPszzcZxtHaKgM/8Pq7byY9kVEXye++GdwzSosYfWI/bHmCdmROKtg1
-21LGKbkaTh8KKmYN69g2xYj1OW3/NI9d9ficGi0b++5vgR8DBUPqEnyE5+OGbN2p4sd3p7bC03Zq
-B7DObtV89mgRYG+fT3+DHbLSQbXbOEnpXAEmv7+PytVs7jle0a89PEg7FYxDgr79l7p8DtwQcjRh
-1J2OdsZOTMC5ZxdsPkWsuqjs6RyC5gjMywtwjz+7ULUFM4z7nI8XDntUkzfjPmfia9Qqfv4QDWSB
-eTQY9JF9Kzv+f8zy+b9mkg+cijm5/gOt4SMB/VEzYePB0LTx8GH1L7trdw2wB5inLW7nDrewOzSf
-VS6Mc8cqSYmnqLueijWlK1BsFU+KAMqc/b4eOLiM+tD7bV2WfHRNKrCQ5T4ex44FZmoZz6/XxOyJ
-gw+yQkxssxnFqp28nrxPjYQ6+mxnEjb7hn45W+YmZiWz26SEvqBwh+GPH386DftNCMZxodPDrcjD
-/QaE+wimDTVxwsf0YQo9pss/L1XtrYtPUJMRYCLCmmy99sEPBJs4Qv8a3BMR8g5s+Zgdd+izpZzd
-TCSlDiCbYlcnKP4WXyMmNqPAz/9S8YKS2GAms7RGWrHjjdmHizqb0flIJcG/0qnCmDpECQEc/luk
-8bUYUuc5hp40N1J06jYutfdZlDkmp4o6mR9cJ3Mhf6/jFLf1crEAXPDwSr+KeHiKQIl3nNPASYtK
-zuoyqTZAgljl+uyP0h+chtMNT3ToIcnHPExATIg4Ep9w2vieCTc35DLBAf/EAyeJ+27s4CQrRPQc
-3mf5BEedUI7vmJHqnsvT46A9Qg4ABgAU5j8Y6cid/0bSK/eAkdbcJSpqSY+UbqQhJ2cMoQxHGOng
-3/TTZ0SXt7Zgeb0dy+vdWF63sbzuxfLax/J6N5auSODC2qCVkYS+wFX7WKM338aNOfEwp/Fsye0w
-9xNzPAGiKMwG28gUp0B7kS0+3yMgpLadA2d62OTPJJxUWuYcAtcgkfvxEEtv5k3yutOZsnF0Z56K
-cWe35RD5fQ+iiFLFptSd5W0eV3HkycV1mk9BbC264wbAWLTTiThWmt1OphzdbVmqwcV/ff7x4wds
-jqAGJr2BuuEiomHBqQyfxuW16kpTs/krgB2ppZ+IQ900wL0HRtZ4lD3+5x1leCDjiDVlKOSiAA+A
-srpsMzf3KQxbz3WSlH7OTM6HTcdikFWDZlJbiHRycfHu5PPJgEJ+g/8duAJjaOtLh4uPaWEbdP03
-t7mlOPYBodaxrcb4uXPyaN1wxP021oDt+PCtB4cPMdi9YQJ/lv9SSsGSAKEiHfx9DKEevAf6qm1C
-hz6GETvJf+7JGjsr9p0je46L4oh+37FDewD/sBP3GBMggHahhmZn0GymWkrfmtcdFHWAPtDX++ot
-WHvr1d7J+BS1k+hxAB3K2mbb3T/vnIaNnpLVm9Mfzj6cn725OPn8o+MCoiv38dPBoTj96Yug/BA0
-YOwTxZgaUWEmEhgWt9BJzHP4r8bIz7yuOEgMvd6dn+uTmhWWumDuM9qcCJ5zGpOFxkEzjkLbhzr/
-CDFK9QbJqSmidB2qOcL90orrWVSu86OpVGmKzmqtt166VszUlNG5dgTSB41dUjAITjGDV5TFXpld
-YckngLrOqgcpbaNtYkhKQcFOuoBz/mVOV7xAKXWGJ01nregvQxfX8CpSRZrATu5VaGVJd8P0mIZx
-9EN7wM149WlApzuMrBvyrLdigVbrVchz0/1HDaP9XgOGDYO9g3lnktJDKAMbk9tEiI34JCeUd/DV
-Lr1eAwULhgd9FS6iYboEZh/D5losE9hAAE8uwfriPgEgtFbCPxA4cqIDMsfsjPDtar7/l1ATxG/9
-6689zasy3f+bKGAXJDiVKOwhptv4HWx8IhmJ04/vRyEjR6m54i81lgeAQ0IBUEfaKX+JT9AnQyXT
-hc4v8fUBvtB+Ar1udS9lUeru/a5xiBLwRA3Ja3iiDP1CTPeysMc4lVELNFY+WMywgtBNQzCfPfFp
-KdNU57ufvTs/Bd8RizFQgvjc7RSG43gJHqHr5DuucGyBwgN2eF0iG5fowlKSxTzymvUGrVHkqLeX
-l2HXiQLD3V6dKHAZJ8pFe4jTZlimnCBCVoa1MMvKrN1qgxR22xDFUWaYJSYXJSWw+jwBvExPY94S
-wV4JSz1MBJ5PkZOsMhmLaTIDPQoqFxTqGIQEiYv1jMR5ecYx8LxUpgwKHhabMrleVni6AZ0jKsHA
-5j+dfDk/+0BlCYcvG6+7hznHtBMYcxLJMaYIYrQDvrhpf8hVk0kfz+pXCAO1D/xpv+LslGMeoNOP
-A4v4p/2K69COnZ0gzwAUVF20xQM3AE63PrlpZIFxtftg/LgpgA1mPhiKRWLZi070cOfX5UTbsmVK
-KO5jXj7iAGdR2JQ03dlNSWt/9BwXBZ5zzYf9jeBtn2yZzxS63nTebEt+cz8dKcSSWMCo29ofw2SH
-dZrq6TjMto1baFurbeyvmRMrddrNMhRlIOLQ7TxymaxfCevmzIFeGnUHmPheo2sksVeVD37NBtrD
-8DCxxO7sU0xHKmMhI4CRDKlrf2rwodAigAKh7N+hI7nj0dNDb46ONbh/jlp3gW38ERShzsWlGo+8
-BE6EL7+z48ivCC3Uo0cidDyVTGa5zRPDz3qJXuULf469MkBBTBS7Ms6u5ZBhjQ3MZz6xt4RgSdt6
-pL5MrvoMizgD5/RuC4d35aL/4MSg1mKETrsbuWmrI5882KC3FGQnwXzwZbwG3V/U1ZBXcss5dG8t
-3Xao90PE7ENoqk/fhyGGY34Pt6xPA7iXGhoWeni/bzmF5bUxjqy1j62qptC+0B7srIStWaXoWMYp
-TjS+qPUCGoN73Jj8gX2qE4Xs7546MScmZIHy4C5Ib24D3aAVThhwuRJXjiaUDt9U0+h3c3krUzAa
-YGSHWO3wm612GEU2nNKbB/bV2F1sLjb9uNGbBrMjU46BnpkqYP2iTFYHiE5vxGcXZg0yuNS/6i1J
-nN2Ql/z2r2dj8fbDz/DvG/kRTCkWP47F3wAN8TYvYX/J1bt0rQJWclS8ccxrhRWSBI2OKvgGCnTb
-Ljw647GILjHxa0usphSYVVuu+NoTQJEnSBXtjZ9gCifgt6nsanmjxlPsW5SBfok02F7sggUiB7pl
-tKxWKdoLJ0rSrObl4Pzs7emHT6dRdYccbn4OnCiKn5CF09FnxCWeh42FfTKr8cmV4zj/KNOix2/W
-m05TOIObThHCvqSwG02+UiO2m4u4xMiBKDbzfBZhS2B5rtWr1uBIj5z95b2G3rOyCGs40qdojTeP
-j4Ea4te2IhpAQ+qj50Q9CaF4ikVj/Dga9JvisaDQNvx5erOeu5FxXf1DE2xj2sx66He3unDJdNbw
-LCcRXsd2GUxBaJrEajWduYWCHzOhb0QBLUfnHHIR12klZAaSS5t8upoCNL1b28cSwqzC5owK3ihM
-k67jjXKSkGIlBjjqgKrr8UCGIoawB/8pvmF7gEWHouZaaIBOiNL+KXe6qnq2ZAnmLRFRryfxYJ1k
-L918Hk1hHpR3yLPGkYV5otvIGF3LSs+fHwxHly+aTAeKSs+8yt5ZAVbPZZM9UJ3F06dPB+Lf7/d+
-GJUozfMbcMsAdq/Xck6vt1huPTm7Wl3P3ryJgB9nS3kJD64oem6f1xmFJnd0pQWR9q+BEeLahJYZ
-TfuWXeagXckHzdyCD6y05fglS+jeIwwtSVS2+vooDDsZaSKWBMUQxmqWJCGHKWA9NnmNRXkYZtT8
-Iu+A4xMEM8a3eELGW+0lepiUQGu5x6JzLAYEeEC5ZTwaVTVTWRrgObnYaDQnZ1lSNfUkz93DU30X
-QGWvM9J8JeI1SoaZR4sYTn2nx6qNh53vZFFvx5LPLt2AY2uW/Po+3IG1QdLyxcJgCg/NIs1yWc6M
-OcUVS2ZJ5YAx7RAOd6ZbnMj6REEPSgNQ72QV5lai7ds/2XVxMf1I58j7ZiSdPlTZm7E4OBRnrQTD
-KGrGpzCUJaTlW/NlBKN8oLC29gS8scSfdFAViwm8CzzcusY60xdzcP5Gc1sHwKHLoKyCtOzo6Qjn
-BjILn5l2y3Ua+KEtOuF2m5RVHacTff/DBB22iT1Y13jaeridlZ7WWwEnPwcPeF+n7oPjYLJskJ6Y
-emtKM47FQocoIrfEzK/GKnL08g7ZVwKfAikzn5jCaBNEurTsaitOdc6mo+IR1DNTxbTFMzflM53K
-ExfzMeU5mbqHLV60waV9kYV4fSyGL8bi29ZGaFZs8GInQPnJPHoyD32fjLpeHh02dqa78WxB2Ark
-5dWjp5smU5pe2Jdzfn9fnXSIG8AVyM4ikfP9JwqxY5y/FqqG0sxrO6fQjLEkfc9mPelq7KZGhUrR
-puDVrxuF4qgW43/aQUyZt9YDXBGLQssWyFbxm8STVvKfvbcNEwM1ev7Koucy6Tucwm94Wwq81wR1
-HZ2th5Y6rd6C7dmT69pJPoJqGjYcf69H9ShRaueId1rh8WQjcS7rP4KHQ7pZhpjmWetY+F/JPJy0
-v+1wsYPld9/swtNVML1lEj0Lurt2gZe6XbDQLLf59Ie6PEbp6/pVAuNAaUQHvD5z+SP5a0eYD8y3
-uuQ2L3iF1yvSWS/allS6/gfvSfkeLXQIaBNO6VmwFuCS1As8mr2l2yJPFKWR4aUv3xy+GJtaWwak
-J/AyevlMX6pI3cx1Ar6zOtabIHip+x1G/+YASyq/t33V2RbQtI5btyv5g4UUjxpFE0uHxnLcX1nR
-rFks8BbChpjspNorNd6D2zAFh8FcJ5qD5wM7u6gPXVdjNNK7TbVtEeCtwUP72SY5D+raKFJEepew
-bVOeuxTno0VB9+q3ILgXR85fxvwGfaq6OLKxKmNT8Cxx6OZH4qe66a3kYnuCxrW6CXdNn/vvmrtu
-EdiZm/SAztz9ik2XBrrvdivaRwOOE2hCPKjooNH4/cbEtQNjnZXSH/PWHyS/2wlnusWs3AfG5MBg
-BJ3YU2NvzP4qnrnfMcVqn684dgt0e52N1rQ7NqPN8Q/xFDidBJ/bmn3KEZprDuSNB91ZN+Gs04m8
-vlaTGO9LnNBulTKkOtsQs/95T9fdyVhtzLYFrwECEIabdC6rm64OjAG6ku9t5gQj574XQUNTGq6T
-16uSOZsEvUcCcBGHHqm/CW1zYu4glRgxVnVZlLCtHOjbfTnzpS9ZuAFqImGrWN0Y1E2Psb7slRQr
-pVuZol4OeLbSZoAIbMQ7pmEyse+AV543FxckY8sMMqtXsoyr5tIe/4w9Ea+dEaiMGxfXiXM1Utni
-EhexxPKGgxRGmuz3Z7BD83anO24qGFlt93B2oh46dvqYSxAcY2S4OLmzF/a5F0XN6bJo1zu0zRqu
-s5cUwTKY2+dIR+qgE7/VN2Lxra0cEkf/0uEfkHe3ltHP67bqjL1bi4bzzFUI3SuQsAafjHPfzYYd
-DujeYdjaodrxfX1hGaXjYW5pbKmoffJehdOMNmpCMZiCeU8oxk+zf2QoxoP/wFCMvocSDI3GR+uB
-3sT7e2I2rB7cSx0bRoA+EyASHgm3rgQ0pnLoprEXuUruBvaKZtaVTm2cMQ/Ikd3bvggEX96o3Jxf
-73K1XaEYX7ro8Q/nH9+cnBMtJhcnb//z5AdKc8Jzh5atenCsKsv3mdr7XkK1G7fSqSl9gzfY9ty5
-ylVBGkLnfedUvwdCfwVY34K2FZn7eluHTiVNtxMgvnvaLajbVHYv5I5fpqs23ISUVuZzoJ9ymqr5
-5Zz1m0fmyIvFoTnSMu+bUwgto50g7baFcxJGu+pE+6v6Xs0tAeSRTVumFcDDB+Qve/ZgalBshJsd
-lPb/OINyrbF+z9xJA1I4k87diHQtIoOq/P9DRwnKLsa9HTuKY3vbNbXjcxZlr3HHQ9SZjAxBvAK6
-QXd+rrDPZbqFCkHACk/f/MeIGP2nTybtOf4TJS73qVR3H5XNlf2Fa6ad278meFpf2Ru0FKf88Hkl
-NF7UqXsCb/t0OpDTR8c6+cKpDQHNdwB0bsRTAXujv8QKcboRIWwctUuG6aZER339nYM82k0He0Or
-52J/WyGnW8goxIvtDeetWknd45B7qHt6qNqUyzkWGPMet1VoitcEmc8FBV2Z5TkfeBitt/3w9fby
-xZGN0iO/42tHkVB+1sAx7JdOfuPOaxqd7sQs5ZgS4HCv5tT36hZXDlT2CbbtbTpFHlv2PyZhgCEN
-vPf9ITPTw7vMftDG1LLeEUxJDJ+oEU3LKYvRuNsno+50G7XVBcIlPg8A0lGBAAvBdHSjk3K54bzp
-4XO9G5zWdMGte1QTOlJB6Vc+R3AP4/s1+LW7U2nug7oziqY/N2hzoF5yEG72HbjVyAuFbDcJ7ak3
-fLDFBeAq5/7+Lx7Qv5sYaLsf7vKrbauXvZV17MtiLimm2LRIZB5HYGRAbw5JW2MBghF0vNiloaPL
-UM3ckC/Q8aP8VLy+mjYY5MxOtAdgjULwf2RtvCc=
-""")
-
-##file ez_setup.py
-EZ_SETUP_PY = convert("""
-eJzNWmmP20YS/a5fwSgYSIJlDu9DhrzIJg5gIMgGuYCFPavpc8SYIhWS8li7yH/f181DJDWcJIt8
-WAbOzJDN6qpXVa+qWvr8s+O52ufZbD6f/z3Pq7IqyNEoRXU6VnmelkaSlRVJU1IlWDR7K41zfjIe
-SVYZVW6cSjFcq54WxpGwD+RBLMr6oXk8r41fTmWFBSw9cWFU+6ScySQV6pVqDyHkIAyeFIJVeXE2
-HpNqbyTV2iAZNwjn+gW1oVpb5Ucjl/VOrfzNZjYzcMkiPxji3zt930gOx7yolJa7i5Z63fDWcnVl
-WSF+PUEdgxjlUbBEJsz4KIoSIKi9L6+u1e9YxfPHLM0Jnx2SosiLtZEXGh2SGSStRJGRSnSLLpau
-9aYMq3hulLlBz0Z5Oh7Tc5I9zJSx5Hgs8mORqNfzo3KCxuH+fmzB/b05m/2oYNK4Mr2xkiiM4oTf
-S2UKK5KjNq/xqtby+FAQ3vejqYJh1oBXnsvZV2++/uKnb37c/fzm+x/e/uNbY2vMLTNgtj3vHv30
-/TcKV/VoX1XHze3t8XxMzDq4zLx4uG2Cory9KW/xX7fb7dy4UbuYDb7vNu7dbHbg/o6TikDgf7TH
-Fpc3XmJzar88nh3TNcXDw2JjLKLIcRiRsWU7vsUjL6JxHNBQOj4LRMDIYv2MFK+VQsOYRMSzXOH5
-liMpjXwhXGnHnh26PqMTUpyhLn7gh6Ef84gEPJLM86zQIjG3Qid0eBw/L6XTxYMBJOJ2EHOHiiCw
-JXEdEgjfEZ6MnCmL3KEulLo2syQL3TgmgeuHcRz6jPBY+sQK7OhZKZ0ubkQihrs8EIw7juOF0g5j
-GXISBLEkbEKKN9QlcCzPJ44nuCdsQVkYSmG5MSGeCGQo/GelXHBh1CF25EOPiBMmJXW4DX0sl7rU
-Zt7TUtgoXqgrHer7bswD+DWUoUd4GNsOBJHYiiYsYuN4gT1ccCAZhNzhjpTC9iwrdgNPOsSb8DSz
-raEyDHA4hPrcJZbjB54fwD/MdiPLIqEVW8+L6bTxQ44X4aOYRlYYOsyPie+SyHNd4nM+iUwtxm/F
-cOEFhEXAMg5ZFPt+6AhfRD7CUdCIhc+LCTptIoFMIkJaAQBymAg824M0B0YC8Alvg1SG2DiUCIIc
-tl2O95FGTiRCSnzqE2jExfNiLp7igRvLmFoQ5jHP8eLQcj0umCOYxZxJT9lDbAKPxZ50qQxJiCh0
-BYtcYVEH7g69mDrPi+mwoZLEjm1ZlMNNHDkBSYJzF44PPCsKJsSMeEZaVuBRGRDi0JBbUAvIeghs
-K7JD5kw5asQzgR3YsSMEc33phQJeswPGA2I7kOqEU1JGPCPtCAQF8uUSoUIcP2YxpEibhzSM5ARb
-sRHPCEvw0Asih8VxRCUNgXRkIXot+Dy0p5ztDp1EqJB2IDmHYb7v217k2SwEf/E4igN/SsqIrahF
-Y9u1CSPUdSyAAZ4LpecxH0QR2vJZKZ1FCBKJPQPuSSpdZBSVsRcwC1CB9cRUwHhDiyLF1iB+12Gc
-xix0KJMe6MsJpBMROcVW/tAiIWLJIwvqICERsdIV4HQ/BGHwyA6mPO0PLSISXMUlqoodWrYQADdE
-cfIpQ8EjwRTL+CMfRdyVAQjBY4yQKLQ9BA53Q8oYd7nPJ6QEQ4uQMBGqfGTbASpRFHmhAxGomL4X
-I7WniDMYVTfmB0T6IQW+6B6QDYEFQzzPRYL5ZIobgqFF1JERCX0HxR60S10UaQuu5sKXaCV8d0JK
-OKI7Cz6SMeHMJYHtC9+2faQhWooIFDgZL+GoEpBIxr6HKsDB5ZakQcikLR24AY+cqQwIhxZ5qLEE
-fCvRMiABPdezbVtyEbk2/oVTukSjbshSvZATA5GYo36oEASBR66lGivreSmdRYwSNwI3oOfwIpdZ
-KmYRbQCbobJMloFoaJEdOnYIkoOjY85s3/Jji/gRdQXyPPanPB0PLYLuzLPQzNgKYerFgfCYpMKK
-YCuzpjwdj5gBQYbGDrXVjSIegJ2IEFYA8mKB6031d42UziIp4FpX+MQOqe0wuIn5nk1D1F5UfjFV
-SeJhPWIEaWNLxZrEERzEZMcuKltI/dhBjwMpv816EwHGm3JWFedNPXDtSblPE9rOW+jdZ+ITExg1
-3uo7b9RI1KzFw/66GRfS2H0kaYJuX+xwawmddhnmwbWhBoDVRhuQSKO9r2bGdjyoH6qLJ5gtKowL
-SoR+0dyLT/VdzHftMshpVn627aS8a0XfXeSpC3MXpsHXr9V0UlZcFJjrloMV6porkxoLmvnwBlMY
-wRjGPzOM5Xd5WSY07Y1/GOnw9+Fvq/mVsJvOzMGj1eAvpY/4lFRLp75fwLlFpuGqAR0Nh3pRM15t
-R8PculNrR0kptr2Bbo1JcYdRdZuXJjsV+K0Opu4FLlJy3tr+rHESxsYvTlV+AA4M0+UZo2jGbzuz
-eycFaq4/kA/wJYbnj4CKKIAAnjLtSKp9Pc7fN0rfG+U+P6VcTbOkxrovrZ3Ms9OBisKo9qQyMAh3
-grUsNQFnCl1DYurtlDplXL8ijPsBEPeGGmmXj/uE7dvdBbRWRxO1PGNxu1iZULJG6V5tqeT0jjH2
-ohgckDwmmLnpJRIEXyMi6wDXKmc58EgLQfj5oj72eCt76mnY9XbN2YQWUzVaamlUaFUaQPSJBcsz
-XtbYtGocCQJFgQpEVFolVQLXZQ+984za4439eSb0eUJ9NsJrvQBqnioMnzwfUVo2hw2iEabPcor8
-hJ1ErUqdZ8Q4iLIkD6I+4Lgk3f29jpeCJKUwfjiXlTi8+aTwympHZAapcK8+2SBUUYsyXoWgMqY+
-9TDbCNU/H0m5q1kI9m+NxfHDw64QZX4qmCgXimHU9oecn1JRqlOSHoGOH9c5gazjiIMGtuXqwiQq
-5LaXpOnlZYPYKAXbtFuPEu3CAW2SmEBWFNXSWqtNeiTXEHW306v+6Q5tj/l2jWN2mpi3SkbtIBD7
-WNYAIP3wCYbvXmoJqQ9I8+h6h4Foswmu5fyi8evt/EUD1epVI7uvwlDAz/XKL/NMpgmrAM2mz/59
-z/9Ztp//uL9E/0S8L19vb8pVl8ttDuujzPfZkPDnjGSLSqVUlyLgDHV8p3OkOa5T2XLKMoSyaXyX
-CkRIu/xKnsohlcogIAFbWg1lUpQA4lSqdFhAwrl1vfHyp57yC3Mk7332Plt+eSoKSAOd1wJuilHd
-WqFqXWJZmKR4KN9Zd8/XrCd991WCwEzoSdXRb/Pq6xzs3AsUUpazJtvS4ZvrfkK+G6XznXrlc4Ci
-CT//MKiZ/RCti+dTmfpXV1CVz8i4Qen86ok6qTOTXHjeSHNWdxmaEWsbkqo+9NVdw/9p3axZVx3r
-t3Xz98qmuqd2va6ZNZXfX8rgRKnL6wLX1jdVJ1h1IunFiKZuDGtD+6lBgfJBHUTWHvGY1kHbtqBb
-o8dPL29KtNM3peqm5/1cGJ1q14EPuf1yoDAzXgy7vpJ8FNB+iy675vlf8iRbtlWhXVqLKwumxOnW
-91sU6LZbVuzTvo68K6tyWYtdbVQyfPExT1QAHQVRJbBVp+ySbUDR6tKhyCFIoVG2KKX5w2CV6q+V
-X4bvqgsrzUdSZEuF88u/7qo/9Gi4siHn8qkov9EhoT4MWYqPIlN/wJwjlJ3tRXpUrdzbOtp67UQX
-Kug3VPyrj2uWCooZWH5tgKpm6tYB6ZwJAIlXkIeqmQXpikdFsQQTalnqt/u0rknZnDVbgo2btuWy
-I1TmbTSbs9kSjCg2CmEt5kDYXnVQPBd1rdnDvVCiesyLD82ma+NYF4ycVqT5qE0xhWaJG5CpYhEg
-wHQjrhdA8iUTm8wpRFOA+gaYq7/SiwiK9VXI9Ej3qkfSUbZW2XT1GpoEHaxVoobFphdKhTi+qn8s
-R+3UMDpbGtalrpzrLUalTKdcww8mfuZHkS2vln1ufI8+/vaxSCqQD3wMfHUHDQ7/sFaf9j0q76kO
-gBUqDUGNLC+Kkw6OVIyEab/3w0M11pXQ61tObK/mk7OpuRoGmGrGWK6GGtcsoq2puWI9f6RzwIkH
-prajnqy7lzDfqTlvM6YAbLDRu7A0L8VydUURZbXRQvvPm2rWkhYUTNUvLW3N/sil6vcBkb5ED/Jx
-PVWxLzX37XOfg+oa+wbdUrOqLRBP9cejz5efa47reaDj6iuJlzXPzwx6+Lauu6zhZDAYDLTPVGr0
-xgGWHw4w1By0he0JDWlmrPZqfKQhTlELNM6rF+oA5W6lw/RRLAod1sJQZfx3Q0VZqnAe1Sql9nUN
-waJThqHuw7IzS6TlsMHvmbbbNWjtdsYWU55lWqa9+NNd/z9B8Jpc1ahLyzwVyNWJabft41FM6l79
-qkcvxCH/qPlWe6L+GoMealE5KlBv+ju8O2q+J7vsJql+HTYrvWGq3+1cz3d/YEbDz2ea+dEgtpmO
-9v85JJ9Ls07w70q5iuan8q5Nt7vhGK7BtlYIfFilqj8cx3SkqCdPR6ja5S8CoFNfa37BZbCldqAO
-8/kPV23RfN0yyhwk+KALUaFOdBGEaJIuAT1/Qt5i+T3aqXn7hRvzeB4OlPP6qzTX3zYxV4vmpPLY
-1ad2hCkv9PyTfmqoFKGnJK1e1ke/EPmgJsWzYuR+FBfN/KN6rfaouBN7AUT33JfuWv2pViwvXbUW
-0tZCXTQXBV1cnnUnx+rdu+bUWbZF9cmTZ9kVu3oErEv0u7n646bY4N8aXIHxoek064as3chE8T2U
-y9Vd97JZwuKudB7VUDGf15NCXaT7wMADGCGrdmLQXxHatnfNB1HVSavuL/uT9E53DLtdE/UdJI2M
-taFhedW0RC0Ar8bGHkiFaXALPc1SkILtl/P3Wf8rPu+z5bt//Xb3YvXbXLcnq/4Yo9/ucdETjI1C
-rr9klRpCscBn8+skbRmxVhX/f7fRgk3dei/t1R3GMA3kC/20fojRFY82d0+bv3hsYkI27VGneg+A
-GcxocdxuF7udStjdbtF9sJEqiVBT5/BrR5fD9u939h3eefkSYNWp0itfvdzpljubu6fqouaIi0y1
-qL7+C1AkCcw=
-""")
-
-##file distribute_from_egg.py
-DISTRIBUTE_FROM_EGG_PY = convert("""
-eJw9j8tqAzEMRfcG/4MgmxQyptkGusonZBmGoGTUGYFfWPKE6dfXTkM3gqt7rh47OKP3NMF3SQFW
-LlrRU1zhybpAxoKBlIqcrNnBdRjQP3GTocYfzmNrrCPQPN9iwzpxSQfQhWBi0cL3qtRtYIG/4Mv0
-KApY5hooqrOGQ05FQTaxptF9Fnx16Rq0XofjaE1XGXVxHIWK7j8P8EY/rHndLqQ1a0pe3COFgHFy
-hLLdWkDbi/DeEpCjNb3u/zccT2Ob8gtnwVyI
-""")
-
-##file distribute_setup.py
-DISTRIBUTE_SETUP_PY = convert("""
-eJztPGtz2ziS3/UrcHK5SOUkxs7MzV25TlOVmTizrs0mKdvZ/ZC4aIiEJI75GpC0ov311403SEp2
-LrMfruq8O7ZENBqNfncDzMm/1ft2W5WT6XT6S1W1TctpTdIM/marrmUkK5uW5jltMwCaXK3JvurI
-jpYtaSvSNYw0rO3qtqryBmBxlJOaJg90w4JGDkb1fk5+75oWAJK8Sxlpt1kzWWc5oocvgIQWDFbl
-LGkrvie7rN2SrJ0TWqaEpqmYgAsibFvVpFrLlTT+i4vJhMDPmleFQ30sxklW1BVvkdrYUivg/Ufh
-bLBDzv7ogCxCSVOzJFtnCXlkvAFmIA126hw/A1Ra7cq8oumkyDiv+JxUXHCJloTmLeMlBZ5qILvj
-uVg0Aai0Ik1FVnvSdHWd77NyM8FN07rmVc0znF7VKAzBj/v7/g7u76PJ5BbZJfibiIURIyO8g88N
-biXhWS22p6QrqKw3nKauPCNUioliXtXoT822a7PcfNubgTYrmP68LgvaJlszxIoa6THfKXe/wo5q
-yhs2mRgB4hqNllxebSaTlu8vrJCbDJVTDn+6ubyOb65uLyfsa8JgZ1fi+SVKQE4xEGRJ3lclc7Dp
-fXQr4HDCmkZqUsrWJJa2ESdFGr6gfNPM5BT8wa+ALIT9R+wrS7qWrnI2n5F/F0MGjgM7eemgjxJg
-eCiwkeWSnE0OEn0CdgCyAcmBkFOyBiFJgsir6Ic/lcgT8kdXtaBr+LgrWNkC69ewfAmqasHgEWKq
-wRsAMQWSHwDMD68Cu6QmCxEy3ObMH1N4Avgf2D6MD4cdtgXT02YakFMEHMApmP6Q2vRnS4FgHXxQ
-KzZ3felUTdTUFIwyhE8f43+8vrqdkx7TyAtXZm8u377+9O42/vvl9c3Vh/ew3vQs+in64cepGfp0
-/Q4fb9u2vnj5st7XWSRFFVV881L5yOZlA34sYS/Tl9ZtvZxObi5vP328/fDh3U389vVfL9/0FkrO
-z6cTF+jjX3+Lr96//YDj0+mXyd9YS1Pa0sXfpbe6IOfR2eQ9uNkLx8InZvS0mdx0RUHBKshX+Jn8
-pSrYogYKxffJ6w4o5+7nBStolssn77KElY0CfcOkfxF48QEQBBI8tKPJZCLUWLmiEFzDCv7OtW+K
-ke3LcDbTRsG+QoxKhLaKcCDhxWBb1OBSgQfa30TFQ4qfwbPjOPiRaEd5GQaXFgkoxWkTzNVkCVjl
-abxLARHow4a1yS5VGIzbEFBgzFuYE7pTBRQVREgnF1U1K/W2LEys9qH27E2OkrxqGIYja6GbShGL
-mzaBwwCAg5FbB6Jq2m6j3wFeETbHhzmol0Pr57O72XAjEosdsAx7X+3IruIPLsc0tEOlEhqGrSGO
-KzNI3hhlD2aufymr1vNogY7wsFygkMPHF65y9DyMXe8GdBgyB1huBy6N7HgFH9OOa9Vxc5vIoaOH
-hTEBzdAzkwJcOFgFoavqkfUnoXJmbVJBGNWu+5UHoPyNfLjOSlh9TJ+k+lncMuRGvGg5Y0bblOGs
-ugzA2WYTwn9zYuynrWIE+3+z+T9gNkKGIv6WBKQ4gugXA+HYDsJaQUh5W04dMqPFH/h7hfEG1UY8
-WuA3+MUdRH+Kksr9Sb3XusdZ0+Wtr1pAiARWTkDLAwyqaRsxbGngNIOc+uqDSJbC4Neqy1MxS/BR
-Wutmg9apbCSFLamkO1T5+9yk4fGKNkxv23mcspzu1arI6L6SKPjABu7FabOo96dpBP9Hzo6mNvBz
-SiwVmGaoLxAD1xVo2MjD87vZ89mjjAYINntxSoQD+z9Ea+/nAJes1j3hjgSgyCKRfPDAjLfh2ZxY
-+at83C/UnKpkpctUnTLEoiBYCsOR8u4VRWrHy17S1uPA0kncRrkhd7BEA+j4CBOW5/8xB+HEa/rA
-lre8Y8b3FlQ4gKaDSnIn0nmho3TVVDmaMfJiYpdwNA1A8G/ocm9Hm1hyiaGvDeqHTQwmJfLIRqTV
-yN+iSrucNVjafTG7CSxX+oBDP+19cUTjrecDSOXc0oa2LQ89QDCUOHWi/mhZgLMVB8frAjHkl+x9
-EOUcbDVlIA4VWmamjM7f4y0OM89jRqT6CuHUsuTn5RTqMrXebISw/j58jCqV/7Uq13mWtP7iDPRE
-1jOJ8CfhDDxKX3SuXg25j9MhFEIWFO04FN/hAGJ6K3y72FjqtkmcdlL48/IUiqisEaKmj1BCiOrq
-Szkd4sPuT0LLoMVEShk7YN5tsbMhWkKqkwGfeFdifInIx5yBgEbx6W4HJUXFkdQE00JN6DrjTTsH
-4wQ0o9MDQLzXTocsPjn7CqIR+C/llzL8teMcVsn3EjE55TNA7kUAFmEWi5nFUJml0LI2fOWPsbwZ
-sRDQQdIzOsfCP/c8xR1OwdgselHVw6EC+1vs4VlR5JDNjOq1yXZg1fdV+7bqyvS7zfZJMsdIHKRC
-xxxWnHBGW9b3VzFuTligybJExDoSqL83bImfkdilQpZyxFCkv7FtSWOvIrSa5icYX14lol4SrVnF
-+ayV3caSFkxmjfeK9nvICkVytsIW6iPNMw+7Nr2yK1aMg0lTYcvGLQhc2LIUWbFo45jeKaiBmMLI
-vcePe4KNlxCcRLLVq7MylZET+8qUBC+DWUTuJU/ucUWvOAAHwzjTWaSp5PQqLI3kHgUHzXS1B9EV
-TqoyFf3ZmmKsX7E1+htsxSZtR3PbJRb7a7HUaiMthn9JzuCFIyHUjkMlvhKBiGFrXvXIeY5118Qx
-x9Fw6aB4NTa33fwzRnXAfpSXH0dYp23+iR5QSV824rmXrqIgIRhqLDIFpI8MWHogC9egKsHkCaKD
-fal+r2OuvdRZop1dIM9fP1YZanWNppsacmySM4jqpn4x1iOcfDOd45Z8ny2JUlwKB8Mn5JrR9KUI
-rgQjDORnQDpZgck9zPFUYIdKiOFQ+hbQ5KTiHNyFsL4eMtit0GptLxmez7RMwGsV1j/YKcQMgSeg
-DzTtJVWSjYJoyaw5me5W0wGQygsQmR0bOE0lCVhrJMcAAnQN34MH/CPxDhZ14W07V0gY9pILS1Ay
-1tUgOOwG3Neq+hquuzJBd6a8oBh2x0XTd05evHjYzY5kxvJIwtYoarq2jDfatdzI58eS5j4s5s1Q
-ao8lzEjtY1bJBtag+e/+1LRpBgP9lSJcByQ9fG4WeQYOAwuYDs+r8XRIlC9YKD0jtbET3lIAeHZO
-3593WIZKebRGeKJ/Up3VMkO6jzNoVASjad04pKv1rt5qTRdkxegdQjSEOTgM8AFla4P+P0R0o8lD
-Vwt/sZa5NSvlliC265C01k4AMc1UhAAXCg4vVmgBYu16kLVnncCm4YSlJsmy7gS8HyLZa66OtMNe
-+xBuI1axw6qJnfURobFKiPQESDQxasTCTdiNeXsFC9wFY2FUOTzN0/EkcT3moYTSTxzxwHqu23FG
-jNfCM3LNt1FpfreAFHFHhKRpGXBNUlCynY76+BQieBB9ePcmOm3wDA/PhyP8NWgrXyM6GTgxaxLt
-TLlDjVH1l7Fwxq/h2KgiXz+0tBbVIyTiYHSx2/EP65wmbAtmxHSXvJchZA32OYdgPvGfygeIsd5h
-AuR0ahPO3MMKusaaxvNsmOnq+xFOE3qcFKBaHbdH6m+Ic+dut+cF9iMXWHj0A4lefOCHV6AnDy5b
-1n7pZTlg+6+iOnDvELjr9hgw6SnB36pHVAGWM3kAXXUtZtPolHZ0b01WV1D9TNBhzpxIy1HE9+Sp
-5jt8sEFCGR4QHXuw0pq8yDSYJN2smjEnI6ezqqeu+DmIGZYXYAe07+HmxKdmVJVOAPOO5KwNGoJq
-b3x6n59GzRS/UdNCtz047zUW1eEB3rvAjw73NIZj8lAw3llfv4etQHp1tOtqBliGucKYVoJPlocC
-wFZNrOLEgRZ9cGNvNaVOAyLo7cR354c8Td+5H4Izrp6uIVE3J+JIgOKKEwARxNzfMT1xYySW+VgI
-AQY8kAOPXhRARVytfg/Nceos0o30GopNqOhkZHyqgeH5NkX4t8zxXK5LLyjlSJ32lBseEbfmju5Z
-DF2QYNX+UTAJjE4FqvDZZzKy2LQbVaHcsSN1JNRYPwgLfPG0Ljx0NWIuafsGt9cjZeABNS+HLnDU
-90jwI56n78N/RfnLQD6Y5edOJlcx/tIkWSqlvywfM16VaGy9vN4turEc3kJ5R2rGi6xp9M04WUaf
-Ygf0IatroGl6ZBtD+lRuN+rEBcDhPE+KqzWJ3WFxOXoSwYSgnxf12NluHalaDqrHT6WpHhlOI7Cv
-M0/v7ykz7/m7Z7mTycyvWUwEttnliYprEA6TB9TqDL+N1QoHbUVm85e//bZASWI8A6nKz99gK9kg
-Gz8a9A8FqOcGeaunTqA/ULgA8cWD4Zv/6CgrZk94mSc5d8yi/zTTcljhlVBKW8arKDVoL8yIdqwJ
-r4PQ+ots1x6MrSNnkAqz6EnHNWfr7Guoo44NdCbiijCljl8p3zxe9PyRTcbVZUYN+Fl/gJCdsq9O
-DIda6/zizmR1YniuLz2ysisYp/I6pNsjQlB5nVjmf4sFh93KGyFyG/1yAbYBOCJYlbcN9tNRj5cY
-1CSekQZUW9VKOGJmnWdtGOA6y2D2edE7h3SYoBnoLqZw9Q/DJFVYqEoqRg+Xc1BOeYfzZ8mf8V6Z
-R27zWUAid4d0fiutlkpgb9cwHohTFHs5WR2LYsd6tDc1toqZPWIdUisH6tpX+JuEisNT54xVX08d
-M+CD1wCO9eJOyI4FYFUJkDCSdDj5Nqikc8MprZhkSsNYgYHdPQoetn3E1x2ajF+8qDtYyIbhhpxw
-hJkyTN41EWaR/hm3j/FaHnRjehKJy+u96okzEepxfCnctq+zXqpzu6/ZgF/YjHXOyl5/vPpXEmyp
-s0VqfxlQT1813Xtu7osgbskk2wbjgjohKWuZuk+I8RzvIJigiHqb9jNsc/647JMX6aG+drsvqDhF
-mVwadF03a0ZWUbwQpynSN6J6Ct+YfRXE1rx6zFKWyndVsrWCd9+KaZzWSKquIhZze5qjG61uPeSH
-kjHKxqWgsAFD532CAZE8BBq7hDv0bfJ+PtCyherocAXlZWZgo1KOjXuRUW1pZBMRK1MVRMR9uQOb
-KhfynqMVnkcHWvvhLt+oVPVkRRrgGPO3I00f5yrsYZIOJVEjpBzPqRSJ4aGUFHXO75Z8Q1p6MC89
-0lvv8cafN+yuu7phzizRrMXBuvSQ4pDb8f4l64vWLwi+V55DeiEmFTUQyZxDgZx2ZbK1mZ190g+e
-12rE2zhGO1mWinfIJIToSeiXjCRUndWkoPwBbzJUhIrjZ2onrLqNKp6K9BzfaQkWiX8RHhIJvFaU
-s4VqTSzYV/GaGSTQi4KWEMPT4M4geXUICWdJxTWkes9HJJwXP9xhwiIpAFcyNvDKCaV6+OzO9EGw
-Xegms5/9N2vuILnS0yYah7jzNPrSlBGJcxG8YflanhgspxHU+QXDuxjNEqOVPepSl9fF2bqCkAe3
-4l4FBxFKeeHXRF7b0ne39f7sHRH09vjKX7UrsZIvqhRfDpSRBc84BIDbk7CHoBpJBuotOn2gSGkT
-kXvcQGDu2uCbeoB0zQQhg6vrQKjiAHyEyWpHAfp4mQTTXBBR4JuX4v4N8FOQLFqfGg+eLSj7gOi0
-2pMNaxWucOZfSlGJX1LVe/c7VH1QW6h7lpKh8gq/BlCMt5cxXQ6APtyZjEOLZZBp6AGM+vl6Yuoc
-WEl4WohVCsQr09Ww6vz3PN6JJsyjR90RauiaoVRZ76aEhYxoDeVuGqo1fCep6VoKbkX46ygg3tHD
-XtGPP/6XTIuSrAD5ifoMCDz7z7MzJ/vL15GSvUYqtd+kK9cM3QEjDbLfpdm1b7eZSf6bhK/m5EeH
-RWhkOJ/xEDCczxHPq9loXZIUtYCJsCUhASN7LtfnGyINJeZxAC6pD8dOXQaIHth+qTUwwhsUoL9I
-c4AEBDNMxAU2eSNbMwiSQnF5BnAZEzZmi7or5IFZYp95Pa1zxj0ixfnnaBNFS9xn0OA6gpBysgXi
-rIwV3tkQsBPnqs8ATLawsyOAuvnqmOz/4iqxVFGcnAP3cyi4z4fFtrio3Svkx65+CGRxutqEoIRT
-5VvwlUW8RMZ670G5L4aF6k1pGwLE31/MSyL2bVfwpoF6uVbHLGK6NZV+e8gUY6o89r2js7L0aooZ
-iooIK35Nn+elDhjjT4cytKnsHui71g35qF8L/glDNOSjjPeuZ8lL8Tf7pmXFJcbWcydpcgjXTk03
-KLymggtomrVgWpLZPS5/xBEZS+WhE0Sakjkdp8YDF4jELUb1Lnj0QUAJNFy5AgkU0TSNJQ5b72qC
-8WJr0y4Dl9nwkIo7PcugabH114IrEJBr2uWqPLd3Z7csr5c6PUIbF8wWL5wruZPwGOtnwXOo1Rfz
-FnjX0ZDt3YAMMJNp6SPly+mn63dTS6KmfPTur6Rf/3MDmNTgjVgRmNXN1speCxxXbLUDJai5ztzU
-jlyh60S2Av6onMMYFcUu6qYEjqeuGmnxCw0qKDjGAzedrUZdHft3CoTPvqTNXkFpldL/TsLSV1PZ
-/zn6ipR/wVrbr/fUM4zhy8vHvBF4rExcM8RaLRbtwDhGPsSxepHeZMCCOzDhfwBqDMd7
-""")
-
-##file activate.sh
-ACTIVATE_SH = convert("""
-eJytVVFvokAQfudXTLEPtTlLeo9tvMSmJpq02hSvl7u2wRUG2QR2DSxSe7n/frOACEVNLlceRHa+
-nfl25pvZDswCnoDPQ4QoTRQsENIEPci4CsBMZBq7CAsuLOYqvmYKTTj3YxnBgiXBudGBjUzBZUJI
-BXEqgCvweIyuCjeG4eF2F5x14bcB9KQiQQWrjSddI1/oQIx6SYYeoFjzWIoIhYI1izlbhJjkKO7D
-M/QEmKfO9O7WeRo/zr4P7pyHwWxkwitcgwpQ5Ej96OX+PmiFwLeVjFUOrNYKaq1Nud3nR2n8nI2m
-k9H0friPTGVsUdptaxGrTEfpNVFEskxpXtUkkCkl1UNF9cgLBkx48J4EXyALuBtAwNYIjF5kcmUU
-abMKmMq1ULoiRbgsDEkTSsKSGFCJ6Z8vY/2xYiSacmtyAfCDdCNTVZoVF8vSTQOoEwSnOrngBkws
-MYGMBMg8/bMBLSYKS7pYEXP0PqT+ZmBT0Xuy+Pplj5yn4aM9nk72JD8/Wi+Gr98sD9eWSMOwkapD
-BbUv91XSvmyVkICt2tmXR4tWmrcUCsjWOpw87YidEC8i0gdTSOFhouJUNxR+4NYBG0MftoCTD9F7
-2rTtxG3oPwY1b2HncYwhrlmj6Wq924xtGDWqfdNxap+OYxplEurnMVo9RWks+rH8qKEtx7kZT5zJ
-4H7oOFclrN6uFe+d+nW2aIUsSgs/42EIPuOhXq+jEo3S6tX6w2ilNkDnIpHCWdEQhFgwj9pkk7FN
-l/y5eQvRSIQ5+TrL05lewxWpt/Lbhes5cJF3mLET1MGhcKCF+40tNWnUulxrpojwDo2sObdje3Bz
-N3QeHqf3D7OjEXMVV8LN3ZlvuzoWHqiUcNKHtwNd0IbvPGKYYM31nPKCgkUILw3KL+Y8l7aO1ArS
-Ad37nIU0fCj5NE5gQCuC5sOSu+UdI2NeXg/lFkQIlFpdWVaWZRfvqGiirC9o6liJ9FXGYrSY9mI1
-D/Ncozgn13vJvsznr7DnkJWXsyMH7e42ljdJ+aqNDF1bFnKWFLdj31xtaJYK6EXFgqmV/ymD/ROG
-+n8O9H8f5vsGOWXsL1+1k3g=
-""")
-
-##file activate.fish
-ACTIVATE_FISH = convert("""
-eJyVVWFv2jAQ/c6vuBoqQVWC9nVSNVGVCaS2VC2rNLWVZZILWAs2s52wVvvxsyEJDrjbmgpK7PP5
-3bt3d22YLbmGlGcIq1wbmCPkGhPYcLMEEsGciwGLDS+YwSjlekngLFVyBe73GXSXxqw/DwbuTS8x
-yyKpFr1WG15lDjETQhpQuQBuIOEKY5O9tlppLqxHKSDByjVAPwEy+mXtCq5MzjIUBTCRgEKTKwFG
-gpBqxTLYXgN2myspVigMaYF92tZSowGZJf4mFExxNs9Qb614CgZtmH0BpEOn11f0cXI/+za8pnfD
-2ZjA1sg9zlV/8QvcMhxbNu0QwgYokn/d+n02nt6Opzcjcnx1vXcIoN74O4ymWQXmHURfJw9jenc/
-vbmb0enj6P5+cuVhqlKm3S0u2XRtRbA2QQAhV7VhBF0rsgUX9Ur1rBUXJgVSy8O751k8mzY5OrKH
-RW3eaQhYGTr8hrXO59ALhxQ83mCsDLAid3T72CCSdJhaFE+fXgicXAARUiR2WeVO37gH3oYHzFKo
-9k7CaPZ1UeNwH1tWuXA4uFKYYcEa8vaKqXl7q1UpygMPhFLvlVKyNzsSM3S2km7UBOl4xweUXk5u
-6e3wZmQ9leY1XE/Ili670tr9g/5POBBpGIJXCCF79L1siarl/dbESa8mD8PL61GpzqpzuMS7tqeB
-1YkALrRBloBMbR9yLcVx7frQAgUqR7NZIuzkEu110gbNit1enNs82Rx5utq7Z3prU78HFRgulqNC
-OTwbqJa9vkJFclQgZSjbKeBgSsUtCtt9D8OwAbIVJuewQdfvQRaoFE9wd1TmCuRG7OgJ1bVXGHc7
-z5WDL/WW36v2oi37CyVBak61+yPBA9C1qqGxzKQqZ0oPuocU9hpud0PIp8sDHkXR1HKkNlzjuUWA
-a0enFUyzOWZA4yXGP+ZMI3Tdt2OuqU/SO4q64526cPE0A7ZyW2PMbWZiZ5HamIZ2RcCKLXhcDl2b
-vXL+eccQoRzem80mekPDEiyiWK4GWqZmwxQOmPM0eIfgp1P9cqrBsewR2p/DPMtt+pfcYM+Ls2uh
-hALufTAdmGl8B1H3VPd2af8fQAc4PgqjlIBL9cGQqNpXaAwe3LrtVn8AkZTUxg==
-""")
-
-##file activate.csh
-ACTIVATE_CSH = convert("""
-eJx9VG1P2zAQ/u5fcYQKNgTNPtN1WxlIQ4KCUEGaxuQ6yYVYSuzKdhqVX7+zk3bpy5YPUXL3PPfc
-ne98DLNCWshliVDV1kGCUFvMoJGugMjq2qQIiVSxSJ1cCofD1BYRnOVGV0CfZ0N2DD91DalQSjsw
-tQLpIJMGU1euvPe7QeJlkKzgWixlhnAt4aoUVsLnLBiy5NtbJWQ5THX1ZciYKKWwkOFaE04dUm6D
-r/zh7pq/3D7Nnid3/HEy+wFHY/gEJydg0aFaQrBFgz1c5DG1IhTs+UZgsBC2GMFBlaeH+8dZXwcW
-VPvCjXdlAvCfQsE7al0+07XjZvrSCUevR5dnkVeKlFYZmUztG4BdzL2u9KyLVabTU0bdfg7a0hgs
-cSmUg6UwUiQl2iHrcbcVGNvPCiLOe7+cRwG13z9qRGgx2z6DHjfm/Op2yqeT+xvOLzs0PTKHDz2V
-tkckFHoQfQRXoGJAj9el0FyJCmEMhzgMS4sB7KPOE2ExoLcSieYwDvR+cP8cg11gKkVJc2wRcm1g
-QhYFlXiTaTfO2ki0fQoiFM4tLuO4aZrhOzqR4dIPcWx17hphMBY+Srwh7RTyN83XOWkcSPh1Pg/k
-TXX/jbJTbMtUmcxZ+/bbqOsy82suFQg/BhdSOTRhMNBHlUarCpU7JzBhmkKmRejKOQzayQe6MWoa
-n1wqWmuh6LZAaHxcdeqIlVLhIBJdO9/kbl0It2oEXQj+eGjJOuvOIR/YGRqvFhttUB2XTvLXYN2H
-37CBdbW2W7j2r2+VsCn0doVWcFG1/4y1VwBjfwAyoZhD
-""")
-
-##file activate.bat
-ACTIVATE_BAT = convert("""
-eJx9UdEKgjAUfW6wfxjiIH+hEDKUFHSKLCMI7kNOEkIf9P9pTJ3OLJ/03HPPPed4Es9XS9qqwqgT
-PbGKKOdXL4aAFS7A4gvAwgijuiKlqOpGlATS2NeMLE+TjJM9RkQ+SmqAXLrBo1LLIeLdiWlD6jZt
-r7VNubWkndkXaxg5GO3UaOOKS6drO3luDDiO5my3iA0YAKGzPRV1ack8cOdhysI0CYzIPzjSiH5X
-0QcvC8Lfaj0emsVKYF2rhL5L3fCkVjV76kShi59NHwDniAHzkgDgqBcwOgTMx+gDQQqXCw==
-""")
-
-##file deactivate.bat
-DEACTIVATE_BAT = convert("""
-eJxzSE3OyFfIT0vj4ipOLVEI8wwKCXX0iXf1C7Pl4spMU0hJTcvMS01RiPf3cYmHyQYE+fsGhCho
-cCkAAUibEkTEVhWLMlUlLk6QGixStlyaeCyJDPHw9/Pw93VFsQguim4ZXAJoIUw5DhX47XUM8UCx
-EchHtwsohN1bILUgw61c/Vy4AJYPYm4=
-""")
-
-##file activate.ps1
-ACTIVATE_PS = convert("""
-eJylWdmS40Z2fVeE/oHT6rCloNUEAXDThB6wAyQAEjsB29GBjdgXYiWgmC/zgz/Jv+AEWNVd3S2N
-xuOKYEUxM+/Jmzfvcm7W//zXf/+wUMOoXtyi1F9kbd0sHH/hFc2iLtrK9b3FrSqyxaVQwr8uhqJd
-uHaeg9mqzRdR8/13Pyy8qPLdJh0+LMhi0QCoXxYfFh9WtttEnd34H8p6/f1300KauwrULws39e18
-0ZaLNm9rgN/ZVf3h++/e124Vlc0vKsspHy+Yyi5+XbzPhijvCtduoiL/kA1ukWV27n0o7Sb8LIFj
-CvWR5GQgUJdp1Pw8TS9+rPy6SDv/+e3d+0+4qw8f3v20+PliV37efEYBAB9FTKC+RHn/Cfxn3rdv
-00Fube5O+iyCtHDs9BfPfz3q4sfFv9d91Ljhfy7ei0VO+nVTtdOkv/jpt0l2AX6iG1jXgKnnDuD4
-ke2k/i8fzzz5UedkVcP4pwF+Wvz2FJl+3vt598urXf5Y6LNA5WcFOP7r0sW7b9a+W/xcu0Xpv5zk
-Kfq3P9Dz9di/fCxS72MXVU1rpx9L4Bxl85Wmn5a+zP76Zuh3pL9ROWr87PN+//GHIl+oOtvn9XSU
-qH+p0gQBFnx1uV+JLH5O5zv+PXW+WepXVVHZT0+oQezkIATcIm+ivPV/z5J/+cYj3ir4w0Lx09vC
-e5n/y5/Y5LPPfdrqb88ga/PabxZRVfmp39l588m/6u+/e+OpP+dF7n1WZpJ9//Z4v372fDDz9eHB
-7Juvs/BLMHzrxL9+9twXpJfhd1/DrpQ5Euu/vlss3wp9HXC/54C/Ld69m6zwdx3tC0d8daSv0V8B
-n4b9YYF53sJelJV/ix6LZspw/sJtqyl5LJ5r/23htA1Imfm/gt9R7dqVB1LjhydAX4Gb+zksQF59
-9+P7H//U+376afFuvh2/T6P85Xr/5c8C6OXyFY4BGuN+EE0+GeR201b+wkkLN5mmBY5TfMw8ngqL
-CztXxCSXKMCYrRIElWkEJlEPYsSOeKBVZCAQTKBhApMwRFQzmCThE0YQu2CdEhgjbgmk9GluHpfR
-/hhwJCZhGI5jt5FsAkOrObVyE6g2y1snyhMGFlDY1x+BoHpCMulTj5JYWNAYJmnKpvLxXgmQ8az1
-4fUGxxcitMbbhDFcsiAItg04E+OSBIHTUYD1HI4FHH4kMREPknuYRMyhh3AARWMkfhCketqD1CWJ
-mTCo/nhUScoQcInB1hpFhIKoIXLo5jLpwFCgsnLCx1QlEMlz/iFEGqzH3vWYcpRcThgWnEKm0QcS
-rA8ek2a2IYYeowUanOZOlrbWSJUC4c7y2EMI3uJPMnMF/SSXdk6E495VLhzkWHps0rOhKwqk+xBI
-DhJirhdUCTamMfXz2Hy303hM4DFJ8QL21BcPBULR+gcdYxoeiDqOFSqpi5B5PUISfGg46gFZBPo4
-jdh8lueaWuVSMTURfbAUnLINr/QYuuYoMQV6l1aWxuZVTjlaLC14UzqZ+ziTGDzJzhiYoPLrt3uI
-tXkVR47kAo09lo5BD76CH51cTt1snVpMOttLhY93yxChCQPI4OBecS7++h4p4Bdn4H97bJongtPk
-s9gQnXku1vzsjjmX4/o4YUDkXkjHwDg5FXozU0fW4y5kyeYW0uJWlh536BKr0kMGjtzTkng6Ep62
-uTWnQtiIqKnEsx7e1hLtzlXs7Upw9TwEnp0t9yzCGgUJIZConx9OHJArLkRYW0dW42G9OeR5Nzwk
-yk1mX7du5RGHT7dka7N3AznmSif7y6tuKe2N1Al/1TUPRqH6E2GLVc27h9IptMLkCKQYRqPQJgzV
-2m6WLsSipS3v3b1/WmXEYY1meLEVIU/arOGVkyie7ZsH05ZKpjFW4cpY0YkjySpSExNG2TS8nnJx
-nrQmWh2WY3cP1eISP9wbaVK35ZXc60yC3VN/j9n7UFoK6zvjSTE2+Pvz6Mx322rnftfP8Y0XKIdv
-Qd7AfK0nexBTMqRiErvCMa3Hegpfjdh58glW2oNMsKeAX8x6YJLZs9K8/ozjJkWL+JmECMvhQ54x
-9rsTHwcoGrDi6Y4I+H7yY4/rJVPAbYymUH7C2D3uiUS3KQ1nrCAUkE1dJMneDQIJMQQx5SONxoEO
-OEn1/Ig1eBBUeEDRuOT2WGGGE4bNypBLFh2PeIg3bEbg44PHiqNDbGIQm50LW6MJU62JHCGBrmc9
-2F7WBJrrj1ssnTAK4sxwRgh5LLblhwNAclv3Gd+jC/etCfyfR8TMhcWQz8TBIbG8IIyAQ81w2n/C
-mHWAwRzxd3WoBY7BZnsqGOWrOCKwGkMMNfO0Kci/joZgEocLjNnzgcmdehPHJY0FudXgsr+v44TB
-I3jnMGnsK5veAhgi9iXGifkHMOC09Rh9cAw9sQ0asl6wKMk8mpzFYaaDSgG4F0wisQDDBRpjCINg
-FIxhlhQ31xdSkkk6odXZFpTYOQpOOgw9ugM2cDQ+2MYa7JsEirGBrOuxsQy5nPMRdYjsTJ/j1iNw
-FeSt1jY2+dd5yx1/pzZMOQXUIDcXeAzR7QlDRM8AMkUldXOmGmvYXPABjxqkYKO7VAY6JRU7kpXr
-+Epu2BU3qFFXClFi27784LrDZsJwbNlDw0JzhZ6M0SMXE4iBHehCpHVkrQhpTFn2dsvsZYkiPEEB
-GSEAwdiur9LS1U6P2U9JhGp4hnFpJo4FfkdJHcwV6Q5dV1Q9uNeeu7rV8PAjwdFg9RLtroifOr0k
-uOiRTo/obNPhQIf42Fr4mtThWoSjitEdAmFW66UCe8WFjPk1YVNpL9srFbond7jrLg8tqAasIMpy
-zkH0SY/6zVAwJrEc14zt14YRXdY+fcJ4qOd2XKB0/Kghw1ovd11t2o+zjt+txndo1ZDZ2T+uMVHT
-VSXhedBAHoJIID9xm6wPQI3cXY+HR7vxtrJuCKh6kbXaW5KkVeJsdsjqsYsOwYSh0w5sMbu7LF8J
-5T7U6LJdiTx+ca7RKlulGgS5Z1JSU2Llt32cHFipkaurtBrvNX5UtvNZjkufZ/r1/XyLl6yOpytL
-Km8Fn+y4wkhlqZP5db0rooqy7xdL4wxzFVTX+6HaxuQJK5E5B1neSSovZ9ALB8091dDbbjVxhWNY
-Ve5hn1VnI9OF0wpvaRm7SZuC1IRczwC7GnkhPt3muHV1YxUJfo+uh1sYnJy+vI0ZwuPV2uqWJYUH
-bmBsi1zmFSxHrqwA+WIzLrHkwW4r+bad7xbOzJCnKIa3S3YvrzEBK1Dc0emzJW+SqysQfdEDorQG
-9ZJlbQzEHQV8naPaF440YXzJk/7vHGK2xwuP+Gc5xITxyiP+WQ4x18oXHjFzCBy9kir1EFTAm0Zq
-LYwS8MpiGhtfxiBRDXpxDWxk9g9Q2fzPPAhS6VFDAc/aiNGatUkPtZIStZFQ1qD0IlJa/5ZPAi5J
-ySp1ETDomZMnvgiysZSBfMikrSDte/K5lqV6iwC5q7YN9I1dBZXUytDJNqU74MJsUyNNLAPopWK3
-tzmLkCiDyl7WQnj9sm7Kd5kzgpoccdNeMw/6zPVB3pUwMgi4C7hj4AMFAf4G27oXH8NNT9zll/sK
-S6wVlQwazjxWKWy20ZzXb9ne8ngGalPBWSUSj9xkc1drsXkZ8oOyvYT3e0rnYsGwx85xZB9wKeKg
-cJKZnamYwiaMymZvzk6wtDUkxmdUg0mPad0YHtvzpjEfp2iMxvORhnx0kCVLf5Qa43WJsVoyfEyI
-pzmf8ruM6xBr7dnBgzyxpqXuUPYaKahOaz1LrxNkS/Q3Ae5AC+xl6NbxAqXXlzghZBZHmOrM6Y6Y
-ctAkltwlF7SKEsShjVh7QHuxMU0a08/eiu3x3M+07OijMcKFFltByXrpk8w+JNnZpnp3CfgjV1Ax
-gUYCnWwYow42I5wHCcTzLXK0hMZN2DrPM/zCSqe9jRSlJnr70BPE4+zrwbk/xVIDHy2FAQyHoomT
-Tt5jiM68nBQut35Y0qLclLiQrutxt/c0OlSqXAC8VrxW97lGoRWzhOnifE2zbF05W4xuyhg7JTUL
-aqJ7SWDywhjlal0b+NLTpERBgnPW0+Nw99X2Ws72gOL27iER9jgzj7Uu09JaZ3n+hmCjjvZpjNst
-vOWWTbuLrg+/1ltX8WpPauEDEvcunIgTxuMEHweWKCx2KQ9DU/UKdO/3za4Szm2iHYL+ss9AAttm
-gZHq2pkUXFbV+FiJCKrpBms18zH75vax5jSo7FNunrVWY3Chvd8KKnHdaTt/6ealwaA1x17yTlft
-8VBle3nAE+7R0MScC3MJofNCCkA9PGKBgGMYEwfB2QO5j8zUqa8F/EkWKCzGQJ5EZ05HTly1B01E
-z813G5BY++RZ2sxbQS8ZveGPJNabp5kXAeoign6Tlt5+L8i5ZquY9+S+KEUHkmYMRFBxRrHnbl2X
-rVemKnG+oB1yd9+zT+4c43jQ0wWmQRR6mTCkY1q3VG05Y120ZzKOMBe6Vy7I5Vz4ygPB3yY4G0FP
-8RxiMx985YJPXsgRU58EuHj75gygTzejP+W/zKGe78UQN3yOJ1aMQV9hFH+GAfLRsza84WlPLAI/
-9G/5JdcHftEfH+Y3/fHUG7/o8bv98dzzy3e8S+XCvgqB+VUf7sH0yDHpONdbRE8tAg9NWOzcTJ7q
-TuAxe/AJ07c1Rs9okJvl1/0G60qvbdDzz5zO0FuPFQIHNp9y9Bd1CufYVx7dB26mAxwa8GMNrN/U
-oGbNZ3EQ7inLzHy5tRg9AXJrN8cB59cCUBeCiVO7zKM0jU0MamhnRThkg/NMmBOGb6StNeD9tDfA
-7czsAWopDdnGoXUHtA+s/k0vNPkBcxEI13jVd/axp85va3LpwGggXXWw12Gwr/JGAH0b8CPboiZd
-QO1l0mk/UHukud4C+w5uRoNzpCmoW6GbgbMyaQNkga2pQINB18lOXOCJzSWPFOhZcwzdgrsQnne7
-nvjBi+7cP2BbtBeDOW5uOLGf3z94FasKIguOqJl+8ss/6Kumns4cuWbqq5592TN/RNIbn5Qo6qbi
-O4F0P9txxPAwagqPlftztO8cWBzdN/jz3b7GD6JHYP/Zp4ToAMaA74M+EGSft3hEGMuf8EwjnTk/
-nz/P7SLipB/ogQ6xNX0fDqNncMCfHqGLCMM0ZzFa+6lPJYQ5p81vW4HkCvidYf6kb+P/oB965g8K
-C6uR0rdjX1DNKc5pOSTquI8uQ6KXxYaKBn+30/09tK4kMpJPgUIQkbENEPbuezNPPje2Um83SgyX
-GTCJb6MnGVIpgncdQg1qz2bvPfxYD9fewCXDomx9S+HQJuX6W3VAL+v5WZMudRQZk9ZdOk6GIUtC
-PqEb/uwSIrtR7/edzqgEdtpEwq7p2J5OQV+RLrmtTvFwFpf03M/VrRyTZ73qVod7v7Jh2Dwe5J25
-JqFOU2qEu1sP+CRotklediycKfLjeIZzjJQsvKmiGSNQhxuJpKa+hoWUizaE1PuIRGzJqropwgVB
-oo1hr870MZLgnXF5ZIpr6mF0L8aSy2gVnTAuoB4WEd4d5NPVC9TMotYXERKlTcwQ2KiB/C48AEfH
-Qbyq4CN8xTFnTvf/ebOc3isnjD95s0QF0nx9s+y+zMmz782xL0SgEmRpA3x1w1Ff9/74xcxKEPdS
-IEFTz6GgU0+BK/UZ5Gwbl4gZwycxEw+Kqa5QmMkh4OzgzEVPnDAiAOGBFaBW4wkDmj1G4RyElKgj
-NlLCq8zsp085MNh/+R4t1Q8yxoSv8PUpTt7izZwf2BTHZZ3pIZpUIpuLkL1nNL6sYcHqcKm237wp
-T2+RCjgXweXd2Zp7ZM8W6dG5bZsqo0nrJBTx8EC0+CQQdzEGnabTnkzofu1pYkWl4E7XSniECdxy
-vLYavPMcL9LW5SToJFNnos+uqweOHriUZ1ntIYZUonc7ltEQ6oTRtwOHNwez2sVREskHN+bqG3ua
-eaEbJ8XpyO8CeD9QJc8nbLP2C2R3A437ISUNyt5Yd0TbDNcl11/DSsOzdbi/VhCC0KE6v1vqVNkq
-45ZnG6fiV2NwzInxCNth3BwL0+8814jE6+1W1EeWtpWbSZJOJNYXmWRXa7vLnAljE692eHjZ4y5u
-y1u63De0IzKca7As48Z3XshVF+3XiLNz0JIMh/JOpbiNLlMi672uO0wYzOCZjRxcxj3D+gVenGIE
-MvFUGGXuRps2RzMcgWIRolHXpGUP6sMsQt1hspUBnVKUn/WQj2u6j3SXd9Xz0QtEzoM7qTu5y7gR
-q9gNNsrlEMLdikBt9bFvBnfbUIh6voTw7eDsyTmPKUvF0bHqWLbHe3VRHyRZnNeSGKsB73q66Vsk
-taxWYmwz1tYVFG/vOQhlM0gUkyvIab3nv2caJ1udU1F3pDMty7stubTE4OJqm0i0ECfrJIkLtraC
-HwRWKzlqpfhEIqYH09eT9WrOhQyt8YEoyBlnXtAT37WHIQ03TIuEHbnRxZDdLun0iok9PUC79prU
-m5beZzfQUelEXnhzb/pIROKx3F7qCttYIFGh5dXNzFzID7u8vKykA8Uejf7XXz//S4nKvW//ofS/
-QastYw==
-""")
-
-##file distutils-init.py
-DISTUTILS_INIT = convert("""
-eJytV1uL4zYUfvevOE0ottuMW9q3gVDa3aUMXXbLMlDKMBiNrSTqOJKRlMxkf33PkXyRbGe7Dw2E
-UXTu37lpxLFV2oIyifAncxmOL0xLIfcG+gv80x9VW6maw7o/CANSWWBwFtqeWMPlGY6qPjV8A0bB
-C4eKSTgZ5LRgFeyErMEeOBhbN+Ipgeizhjtnhkn7DdyjuNLPoCS0l/ayQTG0djwZC08cLXozeMss
-aG5EzQ0IScpnWtHSTXuxByV/QCmxE7y+eS0uxWeoheaVVfqSJHiU7Mhhi6gULbOHorshkrEnKxpT
-0n3A8Y8SMpuwZx6aoix3ouFlmW8gHRSkeSJ2g7hU+kiHLDaQw3bmRDaTGfTnty7gPm0FHbIBg9U9
-oh1kZzAFLaue2R6htPCtAda2nGlDSUJ4PZBgCJBGVcwKTAMz/vJiLD+Oin5Z5QlvDPdulC6EsiyE
-NFzb7McNTKJzbJqzphx92VKRFY1idenzmq3K0emRcbWBD0ryqc4NZGmKOOOX9Pz5x+/l27tP797c
-f/z0d+4NruGNai8uAM0bfsYaw8itFk8ny41jsfpyO+BWlpqfhcG4yxLdi/0tQqoT4a8Vby382mt8
-p7XSo7aWGdPBc+b6utaBmCQ7rQKQoWtAuthQCiold2KfJIPTT8xwg9blPumc+YDZC/wYGdAyHpJk
-vUbHbHWAp5No6pK/WhhLEWrFjUwtPEv1Agf8YmnsuXUQYkeZoHm8ogP16gt2uHoxcEMdf2C6pmbw
-hUMsWGhanboh4IzzmsIpWs134jVPqD/c74bZHdY69UKKSn/+KfVhxLgUlToemayLMYQOqfEC61bh
-cbhwaqoGUzIyZRFHPmau5juaWqwRn3mpWmoEA5nhzS5gog/5jbcFQqOZvmBasZtwYlG93k5GEiyw
-buHhMWLjDarEGpMGB2LFs5nIJkhp/nUmZneFaRth++lieJtHepIvKgx6PJqIlD9X2j6pG1i9x3pZ
-5bHuCPFiirGHeO7McvoXkz786GaKVzC9DSpnOxJdc4xm6NSVq7lNEnKdVlnpu9BNYoKX2Iq3wvgh
-gGEUM66kK6j4NiyoneuPLSwaCWDxczgaolEWpiMyDVDb7dNuLAbriL8ig8mmeju31oNvQdpnvEPC
-1vAXbWacGRVrGt/uXN/gU0CDDwgooKRrHfTBb1/s9lYZ8ZqOBU0yLvpuP6+K9hLFsvIjeNhBi0KL
-MlOuWRn3FRwx5oHXjl0YImUx0+gLzjGchrgzca026ETmYJzPD+IpuKzNi8AFn048Thd63OdD86M6
-84zE8yQm0VqXdbbgvub2pKVnS76icBGdeTHHXTKspUmr4NYo/furFLKiMdQzFjHJNcdAnMhltBJK
-0/IKX3DVFqvPJ2dLE7bDBkH0l/PJ29074+F0CsGYOxsb7U3myTUncYfXqnLLfa6sJybX4g+hmcjO
-kMRBfA1JellfRRKJcyRpxdS4rIl6FdmQCWjo/o9Qz7yKffoP4JHjOvABcRn4CZIT2RH4jnxmfpVG
-qgLaAvQBNfuO6X0/Ux02nb4FKx3vgP+XnkX0QW9pLy/NsXgdN24dD3LxO2Nwil7Zlc1dqtP3d7/h
-kzp1/+7hGBuY4pk0XD/0Ao/oTe/XGrfyM773aB7iUhgkpy+dwAMalxMP0DrBcsVw/6p25+/hobP9
-GBknrWExDhLJ1bwt1NcCNblaFbMKCyvmX0PeRaQ=
-""")
-
-##file distutils.cfg
-DISTUTILS_CFG = convert("""
-eJxNj00KwkAMhfc9xYNuxe4Ft57AjYiUtDO1wXSmNJnK3N5pdSEEAu8nH6lxHVlRhtDHMPATA4uH
-xJ4EFmGbvfJiicSHFRzUSISMY6hq3GLCRLnIvSTnEefN0FIjw5tF0Hkk9Q5dRunBsVoyFi24aaLg
-9FDOlL0FPGluf4QjcInLlxd6f6rqkgPu/5nHLg0cXCscXoozRrP51DRT3j9QNl99AP53T2Q=
-""")
-
-##file activate_this.py
-ACTIVATE_THIS = convert("""
-eJyNU01v2zAMvetXEB4K21jmDOstQA4dMGCHbeihlyEIDMWmG62yJEiKE//7kXKdpN2KzYBt8euR
-fKSyLPs8wiEo8wh4wqZTGou4V6Hm0wJa1cSiTkJdr8+GsoTRHuCotBayiWqQEYGtMCgfD1KjGYBe
-5a3p0cRKiAe2NtLADikftnDco0ko/SFEVgEZ8aRC5GLux7i3BpSJ6J1H+i7A2CjiHq9z7JRZuuQq
-siwTIvpxJYCeuWaBpwZdhB+yxy/eWz+ZvVSU8C4E9FFZkyxFsvCT/ZzL8gcz9aXVE14Yyp2M+2W0
-y7n5mp0qN+avKXvbsyyzUqjeWR8hjGE+2iCE1W1tQ82hsCZN9UzlJr+/e/iab8WfqsmPI6pWeUPd
-FrMsd4H/55poeO9n54COhUs+sZNEzNtg/wanpjpuqHJaxs76HtZryI/K3H7KJ/KDIhqcbJ7kI4ar
-XL+sMgXnX0D+Te2Iy5xdP8yueSlQB/x/ED2BTAtyE3K4SYUN6AMNfbO63f4lBW3bUJPbTL+mjSxS
-PyRfJkZRgj+VbFv+EzHFi5pKwUEepa4JslMnwkowSRCXI+m5XvEOvtuBrxHdhLalG0JofYBok6qj
-YdN2dEngUlbC4PG60M1WEN0piu7Nq7on0mgyyUw3iV1etLo6r/81biWdQ9MWHFaePWZYaq+nmp+t
-s3az+sj7eA0jfgPfeoN1
-""")
-
-MH_MAGIC = 0xfeedface
-MH_CIGAM = 0xcefaedfe
-MH_MAGIC_64 = 0xfeedfacf
-MH_CIGAM_64 = 0xcffaedfe
-FAT_MAGIC = 0xcafebabe
-BIG_ENDIAN = '>'
-LITTLE_ENDIAN = '<'
-LC_LOAD_DYLIB = 0xc
-maxint = majver == 3 and getattr(sys, 'maxsize') or getattr(sys, 'maxint')
-
-
-class fileview(object):
- """
- A proxy for file-like objects that exposes a given view of a file.
- Modified from macholib.
- """
-
- def __init__(self, fileobj, start=0, size=maxint):
- if isinstance(fileobj, fileview):
- self._fileobj = fileobj._fileobj
- else:
- self._fileobj = fileobj
- self._start = start
- self._end = start + size
- self._pos = 0
-
- def __repr__(self):
- return '<fileview [%d, %d] %r>' % (
- self._start, self._end, self._fileobj)
-
- def tell(self):
- return self._pos
-
- def _checkwindow(self, seekto, op):
- if not (self._start <= seekto <= self._end):
- raise IOError("%s to offset %d is outside window [%d, %d]" % (
- op, seekto, self._start, self._end))
-
- def seek(self, offset, whence=0):
- seekto = offset
- if whence == os.SEEK_SET:
- seekto += self._start
- elif whence == os.SEEK_CUR:
- seekto += self._start + self._pos
- elif whence == os.SEEK_END:
- seekto += self._end
- else:
- raise IOError("Invalid whence argument to seek: %r" % (whence,))
- self._checkwindow(seekto, 'seek')
- self._fileobj.seek(seekto)
- self._pos = seekto - self._start
-
- def write(self, bytes):
- here = self._start + self._pos
- self._checkwindow(here, 'write')
- self._checkwindow(here + len(bytes), 'write')
- self._fileobj.seek(here, os.SEEK_SET)
- self._fileobj.write(bytes)
- self._pos += len(bytes)
-
- def read(self, size=maxint):
- assert size >= 0
- here = self._start + self._pos
- self._checkwindow(here, 'read')
- size = min(size, self._end - here)
- self._fileobj.seek(here, os.SEEK_SET)
- bytes = self._fileobj.read(size)
- self._pos += len(bytes)
- return bytes
-
-
-def read_data(file, endian, num=1):
- """
- Read a given number of 32-bits unsigned integers from the given file
- with the given endianness.
- """
- res = struct.unpack(endian + 'L' * num, file.read(num * 4))
- if len(res) == 1:
- return res[0]
- return res
-
-
-def mach_o_change(path, what, value):
- """
- Replace a given name (what) in any LC_LOAD_DYLIB command found in
- the given binary with a new name (value), provided it's shorter.
- """
-
- def do_macho(file, bits, endian):
- # Read Mach-O header (the magic number is assumed read by the caller)
- cputype, cpusubtype, filetype, ncmds, sizeofcmds, flags = read_data(file, endian, 6)
- # 64-bits header has one more field.
- if bits == 64:
- read_data(file, endian)
- # The header is followed by ncmds commands
- for n in range(ncmds):
- where = file.tell()
- # Read command header
- cmd, cmdsize = read_data(file, endian, 2)
- if cmd == LC_LOAD_DYLIB:
- # The first data field in LC_LOAD_DYLIB commands is the
- # offset of the name, starting from the beginning of the
- # command.
- name_offset = read_data(file, endian)
- file.seek(where + name_offset, os.SEEK_SET)
- # Read the NUL terminated string
- load = file.read(cmdsize - name_offset).decode()
- load = load[:load.index('\0')]
- # If the string is what is being replaced, overwrite it.
- if load == what:
- file.seek(where + name_offset, os.SEEK_SET)
- file.write(value.encode() + '\0'.encode())
- # Seek to the next command
- file.seek(where + cmdsize, os.SEEK_SET)
-
- def do_file(file, offset=0, size=maxint):
- file = fileview(file, offset, size)
- # Read magic number
- magic = read_data(file, BIG_ENDIAN)
- if magic == FAT_MAGIC:
- # Fat binaries contain nfat_arch Mach-O binaries
- nfat_arch = read_data(file, BIG_ENDIAN)
- for n in range(nfat_arch):
- # Read arch header
- cputype, cpusubtype, offset, size, align = read_data(file, BIG_ENDIAN, 5)
- do_file(file, offset, size)
- elif magic == MH_MAGIC:
- do_macho(file, 32, BIG_ENDIAN)
- elif magic == MH_CIGAM:
- do_macho(file, 32, LITTLE_ENDIAN)
- elif magic == MH_MAGIC_64:
- do_macho(file, 64, BIG_ENDIAN)
- elif magic == MH_CIGAM_64:
- do_macho(file, 64, LITTLE_ENDIAN)
-
- assert(len(what) >= len(value))
- do_file(open(path, 'r+b'))
-
-
-if __name__ == '__main__':
- main()
-
-## TODO:
-## Copy python.exe.manifest
-## Monkeypatch distutils.sysconfig
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/__init__.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/__init__.py
deleted file mode 100644
index e69de29bb2..0000000000
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/backup/benchmarks.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/backup/benchmarks.py
deleted file mode 100644
index 7a19da0684..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/backup/benchmarks.py
+++ /dev/null
@@ -1,493 +0,0 @@
-
-# NOTES: Batch13 (Baseline) Batch14 - With ErrorSens (10, 25, 35)
-
-# Batch 9: No Error Sens. Min : P3
-# Batch 10: No Error Sens + More Runs for Loss1 and Loss2. Min: P3
-# Batch 11: Error Sens: Skipping 30% elems in each : Min: P3. More runs in Loss1 (4000) and Loss2 (2000)
-# Batch 12: Error Sens: 10, 25, 35, for Loss1, 2, 3, respectively, Min: P3. 1000 Runs for All
-# Batch 13: No Error Sens: Equal Runs (1000) for all. Min: P1
-# Batch 14: Reruning Batch12 with bugFix!
-# Batch 15: MAJOR CHANGE: 3 different skip levels for each Loss1,Loss2,Loss3
-
-# Batch 18: Batch13 (Basline) + ParetoCurve (1500 Runs) - BUGGY IGNORE!!!
-
-# Batch 19: (Basline) + ParetoCurve + 2 runs in Tuning Phase (1500 Runs)
-
-# Batch 20: 3-Skip levels + + 2 runs + 1500 Runs + EnergyBandSize now % of Max (Compare against Batch19
-
-
-# Batch 200: AlgoTuner - 1000 images - 1500 runs (IGNORE)
-# Batch 201: AlgoTuner - 2000 images - 1500 runs
-# Batch 202: AlgoTuner - 2000 images - 500 runs
-# Batch 203: AlgoTuner - 2000 images - 3000 runs
-
-
-#---- CHANGES: i) Reshufled inputs ii) 3K images for tuning
-# Batch 210: 3K images, 1000 runs (1500 resnet), no FP32 used in tuning
-# Batch 211: Same as Batch-210 + uses tensorConvPerfCuda*Half*
-
-
-#batch_id = "batch210"
-#batch_id = "batch211"
-#batch_id = "batch210"
-
-
-batch_id = "batch310"
-
-
-class Benchmark:
- def __init__(self):
- self.tuner_binary = ""
- self.promise_binary = ""
- self.tuner_accuracy = 0
- self.promise_accuracy = 0
- self.num_flags = 0
- self.num_layers = 0
- self.autotuner_runs = 0
- self.error_range_1 = 0
- self.error_range_2 = 0
- self.result_dir_1 = ""
- self.result_dir_2 = ""
- self.promise_result_dir_1 = ""
- self.promise_result_dir_2 = ""
-
-
-
-bench_tuner_data = {}
-
-# FIXIT: Fix Variable Names below
-Alexnet1 = Benchmark()
-Alexnet1.tuner_binary = "alexnet_cifar10_tuner"
-Alexnet1.fp16_binary = "alexnet_half"
-Alexnet1.promise_binary = "alexnet_promise"
-Alexnet1.validation_binary = "alexnet_valid"
-Alexnet1.num_flags = 21
-Alexnet1.num_layers = 6
-Alexnet1.error_range_1 = 10
-Alexnet1.error_range_2 = 13
-Alexnet1.start_promise_range = 1
-Alexnet1.skip_layers = 0
-#Alexnet1.skip_layer_str = "0"
-Alexnet1.skip_layer_str = "5_0"
-
-Alexnet1.base_dir = "../build_tuner/tuner_results/alexnet_cifar10/"
-Alexnet1.result_dir_1 = "../build_tuner/tuner_results/alexnet_cifar10/loss_1/" + batch_id
-Alexnet1.result_dir_2 = "../build_tuner/tuner_results/alexnet_cifar10/loss_2/" + batch_id
-Alexnet1.result_dir_3 = "../build_tuner/tuner_results/alexnet_cifar10/loss_3/" + batch_id
-
-Alexnet1.tensor_desc_file = "tuner_results/alexnet_cifar10/alexnet_tensors.txt"
-Alexnet1.layer_file = "tuner_results/alexnet_cifar10/alexnet_layers.txt"
-Alexnet1.cost_file = "../build_tuner/tuner_results/alexnet_cifar10/op_cost.txt"
-Alexnet1.layer_knobs = "../opentuner/data/alexnet/knobs.txt"
-
-#Alexnet1.loss1_result_file = "tuner_results/alexnet2_cifar10/alexnet_layers.txt"
-Alexnet1.loss1_result_file = "tuner_results/alexnet_cifar10/loss_1/promise_tuned_confs/promise_confs.txt"
-Alexnet1.loss2_result_file = "tuner_results/alexnet_cifar10/loss_2/promise_tuned_confs/promise_confs.txt"
-
-Alexnet1.autotuner_runs = 1000
-Alexnet1.tuner_accuracy = 79.9
-#Alexnet1.promise_accuracy = 79.9
-Alexnet1.promise_accuracy = 78.86
-Alexnet1.validation_accuracy = 79.19
-
-bench_tuner_data["alexnet_cifar10"] = Alexnet1
-
-
-Alexnet2 = Benchmark()
-Alexnet2.tuner_binary = "alexnet2_cifar10_tuner"
-Alexnet2.fp16_binary = "alexnet2_half"
-Alexnet2.promise_binary = "alexnet2_promise"
-Alexnet2.validation_binary = "alexnet2_valid"
-Alexnet2.num_flags = 23
-Alexnet2.num_layers = 7
-Alexnet2.error_range_1 = 10
-Alexnet2.error_range_2 = 13
-Alexnet2.start_promise_range = 1
-#Alexnet2.skip_layer_str = "0"
-Alexnet2.skip_layer_str = "6_1_0"
-
-Alexnet2.base_dir = "../build_tuner/tuner_results/alexnet2_cifar10/"
-Alexnet2.result_dir_1 = "../build_tuner/tuner_results/alexnet2_cifar10/loss_1/" + batch_id
-Alexnet2.result_dir_2 = "../build_tuner/tuner_results/alexnet2_cifar10/loss_2/" + batch_id
-Alexnet2.result_dir_3 = "../build_tuner/tuner_results/alexnet2_cifar10/loss_3/" + batch_id
-Alexnet2.tensor_desc_file = "tuner_results/alexnet2_cifar10/alexnet2_tensors.txt"
-Alexnet2.layer_file = "tuner_results/alexnet2_cifar10/alexnet2_layers.txt"
-Alexnet2.cost_file = "../build_tuner/tuner_results/alexnet2_cifar10/op_cost.txt"
-Alexnet2.layer_knobs = "../opentuner/data/alexnet2/knobs.txt"
-#Alexnet2.loss1_result_file = "tuner_results/alexnet2_cifar10/loss_1/promise_tuned_confs/promise_confs.txt"
-#Alexnet2.loss2_result_file = "tuner_results/alexnet2_cifar10/loss_2/promise_tuned_confs/promise_confs.txt"
-Alexnet2.autotuner_runs = 1000
-Alexnet2.tuner_accuracy = 84.19
-#Alexnet2.promise_accuracy = 84.19
-Alexnet2.promise_accuracy = 84.7
-Alexnet2.validation_accuracy = 85.15
-
-bench_tuner_data["alexnet2_cifar10"] = Alexnet2
-
-
-
-Alexnet3 = Benchmark()
-Alexnet3.tuner_binary = "vgg16_cifar10_tuner"
-Alexnet3.fp16_binary = "vgg16_cifar10_half"
-Alexnet3.promise_binary = "./vgg16_cifar10_promise"
-Alexnet3.validation_binary = "vgg16_cifar10_valid"
-Alexnet3.num_flags = 50
-Alexnet3.num_layers = 15
-Alexnet3.error_range_1 = 9
-Alexnet3.error_range_2 = 11
-Alexnet3.start_promise_range = 1
-#Alexnet3.skip_layer_str = "0"
-Alexnet3.skip_layer_str = "14_3_4_1_6"
-
-Alexnet3.base_dir = "../build_tuner/tuner_results/vgg16_cifar10/"
-Alexnet3.result_dir_1 = "../build_tuner/tuner_results/vgg16_cifar10/loss_1/" + batch_id
-Alexnet3.result_dir_2 = "../build_tuner/tuner_results/vgg16_cifar10/loss_2/" + batch_id
-Alexnet3.result_dir_3 = "../build_tuner/tuner_results/vgg16_cifar10/loss_3/" + batch_id
-
-Alexnet3.tensor_desc_file = "tuner_results/vgg16_cifar10/vgg16_tensors.txt"
-Alexnet3.layer_file = "tuner_results/vgg16_cifar10/vgg16_layers.txt"
-Alexnet3.cost_file = "../build_tuner/tuner_results/vgg16_cifar10/op_cost.txt"
-Alexnet3.layer_knobs = "../opentuner/data/vgg16_cifar10/knobs.txt"
-
-Alexnet3.loss1_result_file = "tuner_results/vgg16_cifar10/loss_1/promise_tuned_confs/promise_confs.txt"
-Alexnet3.loss2_result_file = "tuner_results/vgg16_cifar10/loss_2/promise_tuned_confs/promise_confs.txt"
-
-Alexnet3.autotuner_runs = 1000
-Alexnet3.tuner_accuracy = 90.19
-#Alexnet3.promise_accuracy = 90.19
-Alexnet3.promise_accuracy = 88.53
-Alexnet3.validation_accuracy = 89.05
-
-bench_tuner_data["vgg16_cifar10"] = Alexnet3
-
-
-
-Alexnet4 = Benchmark()
-Alexnet4.tuner_binary = "resnet18_cifar10_tuner"
-Alexnet4.fp16_binary = "resnet18_half"
-Alexnet4.promise_binary = "resnet18_promise"
-Alexnet4.validation_binary = "resnet18_valid"
-Alexnet4.num_flags = 73
-Alexnet4.num_layers = 22
-Alexnet4.error_range_1 = 7
-Alexnet4.error_range_2 = 9
-Alexnet4.start_promise_range = 1
-#Alexnet4.skip_layer_str = "0"
-Alexnet4.skip_layer_str = "0_1_2_14_15_17_18_21"
-Alexnet4.base_dir = "../build_tuner/tuner_results/resnet18_cifar10/"
-Alexnet4.result_dir_1 = "../build_tuner/tuner_results/resnet18_cifar10/loss_1/" + batch_id
-Alexnet4.result_dir_2 = "../build_tuner/tuner_results/resnet18_cifar10/loss_2/" + batch_id
-Alexnet4.result_dir_3 = "../build_tuner/tuner_results/resnet18_cifar10/loss_3/" + batch_id
-Alexnet4.tensor_desc_file = "tuner_results/resnet18_cifar10/resnet_tensors.txt"
-Alexnet4.layer_file = "tuner_results/resnet18_cifar10/resnet_layers.txt"
-Alexnet4.cost_file = "../build_tuner/tuner_results/resnet18_cifar10/op_cost.txt"
-Alexnet4.layer_knobs = "../opentuner/data/resnet/knobs.txt"
-
-Alexnet4.loss1_result_file = "tuner_results/resnet18_cifar10/loss_1/promise_tuned_confs/promise_confs.txt"
-Alexnet4.loss2_result_file = "tuner_results/resnet18_cifar10/loss_2/promise_tuned_confs/promise_confs.txt"
-
-Alexnet4.autotuner_runs = 1500
-Alexnet4.tuner_accuracy = 89.6
-#Alexnet4.promise_accuracy = 89.59 - 1000 images
-Alexnet4.promise_accuracy = 89.5
-Alexnet4.validation_accuracy = 89.65
-
-bench_tuner_data["resnet18_cifar10"] = Alexnet4
-
-
-
-
-
-Alexnet5 = Benchmark()
-Alexnet5.tuner_binary = "vgg16_cifar100_tuner"
-Alexnet5.fp16_binary = "vgg16_cifar100_half"
-Alexnet5.promise_binary = "vgg16_cifar100_promise"
-Alexnet5.validation_binary = "vgg16_cifar100_valid"
-Alexnet5.num_flags = 50
-Alexnet5.num_layers = 15
-Alexnet5.error_range_1 = 9
-Alexnet5.error_range_2 = 11
-Alexnet5.start_promise_range = 1
-Alexnet5.skip_layer_str = "0_1_2_3_4"
-Alexnet5.base_dir = "../build_tuner/tuner_results/vgg16_cifar100/"
-Alexnet5.result_dir_1 = "../build_tuner/tuner_results/vgg16_cifar100/loss_1/" + batch_id
-Alexnet5.result_dir_2 = "../build_tuner/tuner_results/vgg16_cifar100/loss_2/" + batch_id
-Alexnet5.result_dir_3 = "../build_tuner/tuner_results/vgg16_cifar100/loss_3/" + batch_id
-
-Alexnet5.tensor_desc_file = "../build_tuner/tuner_results/vgg16_cifar100/vgg16_tensors.txt"
-Alexnet5.layer_file = "../build_tuner/tuner_results/vgg16_cifar100/vgg16_layers.txt"
-Alexnet5.cost_file = "../build_tuner/tuner_results/vgg16_cifar100/op_cost.txt"
-Alexnet5.layer_knobs = "../opentuner/data/vgg16_cifar100/knobs.txt"
-
-Alexnet5.loss1_result_file = "tuner_results/vgg_cifar100/loss_1/promise_tuned_confs/promise_confs.txt"
-Alexnet5.loss2_result_file = "tuner_results/vgg_cifar100/loss_2/promise_tuned_confs/promise_confs.txt"
-Alexnet5.autotuner_runs = 1000
-Alexnet5.tuner_accuracy = 67.95
-#Alexnet5.promise_accuracy = 66.8
-Alexnet5.promise_accuracy = 67.86
-Alexnet5.validation_accuracy = 68.65
-
-bench_tuner_data["vgg16_cifar100"] = Alexnet5
-
-
-
-Alexnet6 = Benchmark()
-Alexnet6.tuner_binary = "lenet_keras"
-Alexnet6.fp16_binary = "lenet_half"
-Alexnet6.promise_binary = "lenet_promise"
-Alexnet6.validation_binary = "lenet_promise"
-
-Alexnet6.num_flags = 14
-Alexnet6.num_layers = 4
-Alexnet6.error_range_1 = 16
-Alexnet6.error_range_2 = 20
-Alexnet6.start_promise_range = 1
-Alexnet6.skip_layer_str = "0"
-
-Alexnet6.base_dir = "../build_tuner/tuner_results/lenet_keras/"
-Alexnet6.result_dir = "../build_tuner/tuner_results/lenet_keras/loss_123/" + batch_id
-Alexnet6.result_dir_1 = "../build_tuner/tuner_results/lenet_keras/loss_1/" + batch_id
-Alexnet6.result_dir_2 = "../build_tuner/tuner_results/lenet_keras/loss_2/" + batch_id
-Alexnet6.result_dir_3 = "../build_tuner/tuner_results/lenet_keras/loss_3/" + batch_id
-
-Alexnet6.tensor_desc_file = "tuner_results/lenet_keras/lenet_tensors.txt"
-Alexnet6.layer_file = "tuner_results/lenet_keras/lenet_layers.txt"
-Alexnet6.cost_file = "../build_tuner/tuner_results/lenet_keras/op_cost.txt"
-Alexnet6.layer_knobs = "../autotuner/data/lenet/knobs.txt"
-
-#Alexnet6.loss1_result_file = "tuner_results/vgg_cifar100/loss_1/promise_tuned_confs/promise_confs.txt"
-#Alexnet6.loss2_result_file = "tuner_results/vgg_cifar100/loss_2/promise_tuned_confs/promise_confs.txt"
-Alexnet6.autotuner_runs = 1000
-Alexnet6.tuner_accuracy = 98.9
-Alexnet6.promise_accuracy = 99.7
-Alexnet6.validation_accuracy = 99
-
-bench_tuner_data["lenet_keras"] = Alexnet6
-
-
-
-
-Alexnet7 = Benchmark()
-Alexnet7.tuner_binary = "mobilenet_cifar10"
-Alexnet7.fp16_binary = "mobilenet_half"
-Alexnet7.promise_binary = "mobilenet_promise"
-Alexnet7.validation_binary = "mobilenet_valid"
-Alexnet7.num_flags = 85
-Alexnet7.num_layers = 15
-Alexnet7.error_range_1 = 7
-Alexnet7.error_range_2 = 8
-Alexnet7.start_promise_range = 1
-#Alexnet7.skip_layer_str = "0"
-Alexnet7.skip_layer_str = "1_14_0_6_2"
-Alexnet7.base_dir = "../build_tuner/tuner_results/mobilenet/"
-Alexnet7.result_dir_1 = "../build_tuner/tuner_results/mobilenet/loss_1/" + batch_id
-Alexnet7.result_dir_2 = "../build_tuner/tuner_results/mobilenet/loss_2/" + batch_id
-Alexnet7.result_dir_3 = "../build_tuner/tuner_results/mobilenet/loss_3/" + batch_id
-
-Alexnet7.tensor_desc_file = "tuner_results/mobilenet/mobilenet_ops.txt"
-Alexnet7.layer_file = "tuner_results/mobilenet/mobilenet_layer_comp.txt"
-Alexnet7.cost_file = "../build_tuner/tuner_results/mobilenet/op_cost.txt"
-Alexnet7.layer_knobs = "../opentuner/data/mobilenet/knobs.txt"
-
-#--- Files below needed for VALIDATION experiment
-Alexnet7.loss1_result_file = "tuner_results/mobilenet/loss_1/batch1/promise_tuner/high_confidence/promise_confs.txt"
-Alexnet7.loss2_result_file = "tuner_results/mobilenet/loss_2/batch1/promise_tuner/high_confidence/promise_confs.txt"
-Alexnet7.autotuner_runs = 1000
-Alexnet7.tuner_accuracy = 84.8
-#Alexnet7.promise_accuracy = 84.8
-Alexnet7.promise_accuracy = 83.73
-Alexnet7.validation_accuracy = 84.4
-
-bench_tuner_data["mobilenet_cifar10"] = Alexnet7
-
-
-
-Alexnet8 = Benchmark()
-Alexnet8.tuner_binary = "mobilenet_cifar10_shallow"
-Alexnet8.fp16_binary = "mobilenet_shallow_half"
-Alexnet8.promise_binary = "mobilenet_shallow_promise"
-Alexnet8.validation_binary = "mobilenet_shallow_valid"
-Alexnet8.num_flags = 42
-Alexnet8.num_layers = 8
-Alexnet8.error_range_1 = 10
-Alexnet8.error_range_2 = 12
-Alexnet8.start_promise_range = 1
-#Alexnet8.skip_layer_str = "0"
-Alexnet8.skip_layer_str = "7_0_1"
-Alexnet8.base_dir = "../build_tuner/tuner_results/mobilenet_shallow/"
-Alexnet8.result_dir_1 = "../build_tuner/tuner_results/mobilenet_shallow/loss_1/" + batch_id
-Alexnet8.result_dir_2 = "../build_tuner/tuner_results/mobilenet_shallow/loss_2/" + batch_id
-Alexnet8.result_dir_3 = "../build_tuner/tuner_results/mobilenet_shallow/loss_3/" + batch_id
-
-Alexnet8.tensor_desc_file = "../build_tuner/tuner_results/mobilenet_shallow/mobilenet_shallow_ops.txt"
-Alexnet8.layer_file = "../build_tuner/tuner_results/mobilenet_shallow/mobilenet_shallow_layer_comp.txt"
-Alexnet8.cost_file = "../build_tuner/tuner_results/mobilenet_shallow/op_cost.txt"
-Alexnet8.layer_knobs = "../opentuner/data/mobilenet_shallow/knobs.txt"
-
-Alexnet8.loss1_result_file = "../build_tuner/tuner_results/mobilenet_shallow/loss_1/batch2/promise_tuner/high_confidence/promise_selected_confs.txt"
-Alexnet8.loss2_result_file = "../build_tuner/tuner_results/mobilenet_shallow/loss_2/batch2/promise_tuner/high_confidence/promise_selected_confs.txt"
-
-Alexnet8.autotuner_runs = 1000
-Alexnet8.tuner_accuracy = 87.6
-#Alexnet8.promise_accuracy = 87.59
-Alexnet8.promise_accuracy = 87.76
-Alexnet8.validation_accuracy = 88.5
-
-bench_tuner_data["mobilenet_shallow"] = Alexnet8
-
-
-
-"""
-Alexnet9 = Benchmark()
-Alexnet9.tuner_binary = "fc4_clipped"
-Alexnet9.promise_binary = ""
-Alexnet9.validation_binary = ""
-Alexnet9.num_flags = 12
-Alexnet9.num_layers = 4
-Alexnet9.error_range_1 = 12
-Alexnet9.error_range_2 = 16
-Alexnet9.start_promise_range = 3
-Alexnet9.skip_layer_str = "0"
-Alexnet9.base_dir = "../build_tuner/tuner_results/fc4/"
-Alexnet9.result_dir_1 = "../build_tuner/tuner_results/fc4/loss1/batch1"
-Alexnet9.result_dir_2 = "../build_tuner/tuner_results/fc4/loss2/batch1"
-Alexnet9.tensor_desc_file = ""
-Alexnet9.layer_file = ""
-
-Alexnet9.loss1_result_file = ""
-Alexnet9.loss2_result_file = ""
-
-Alexnet9.autotuner_runs = 1000
-Alexnet9.tuner_accuracy = 93.8
-Alexnet9.promise_accuracy = 0.0
-Alexnet9.validation_accuracy = 0.0
-
-bench_tuner_data["fc4"] = Alexnet9
-
-
-
-
-Pipeline1 = Benchmark()
-Pipeline1.tuner_binary = "pipeline_GEOM"
-Pipeline1.promise_binary = "pipeline_GEOM_promise"
-Pipeline1.validation_binary = "pipeline_GEOM_valid"
-Pipeline1.num_flags = 9
-Pipeline1.num_layers = 4
-Pipeline1.error_range_1 = 10
-Pipeline1.error_range_2 = 15
-Pipeline1.start_promise_range = 2
-Pipeline1.skip_layer_str = "1_2"
-Pipeline1.result_dir_1 = "tuner_results/pipeline_GEOM/loss_30/batch1"
-Pipeline1.result_dir_2 = "tuner_results/pipeline_GEOM/loss_20/batch1"
-Pipeline1.tensor_desc_file = "tuner_results/pipeline_GEOM/pipeline_GEOM_tensors.txt"
-Pipeline1.layer_file = "tuner_results/pipeline_GEOM/pipeline_GEOM_layers.txt"
-#Pipeline1.loss1_result_file = "tuner_results/vgg_cifar100/loss_1/promise_tuned_confs/promise_confs.txt"
-#Pipeline1.loss2_result_file = "tuner_results/vgg_cifar100/loss_2/promise_tuned_confs/promise_confs.txt"
-Pipeline1.autotuner_runs = 300
-Pipeline1.tuner_accuracy = 95
-Pipeline1.promise_accuracy = 95
-Pipeline1.validation_accuracy = 95
-
-bench_tuner_data["pipeline_GEOM"] = Pipeline1
-
-
-Pipeline2 = Benchmark()
-Pipeline2.tuner_binary = "pipeline_GEMO"
-Pipeline2.promise_binary = "pipeline_GEMO_promise"
-Pipeline2.validation_binary = "pipeline_GEMO_valid"
-Pipeline2.num_flags = 9
-Pipeline2.num_layers = 4
-Pipeline2.error_range_1 = 10
-Pipeline2.error_range_2 = 15
-Pipeline2.start_promise_range = 2
-Pipeline2.skip_layer_str = "1_3"
-Pipeline2.result_dir_1 = "tuner_results/pipeline_GEMO/loss_30/batch1"
-Pipeline2.result_dir_2 = "tuner_results/pipeline_GEMO/loss_20/batch1"
-Pipeline2.tensor_desc_file = "tuner_results/pipeline_GEMO/pipeline_GEMO_tensors.txt"
-Pipeline2.layer_file = "tuner_results/pipeline_GEMO/pipeline_GEMO_layers.txt"
-#Pipeline1.loss1_result_file = "tuner_results/vgg_cifar100/loss_1/promise_tuned_confs/promise_confs.txt"
-#Pipeline1.loss2_result_file = "tuner_results/vgg_cifar100/loss_2/promise_tuned_confs/promise_confs.txt"
-Pipeline2.autotuner_runs = 300
-Pipeline2.tuner_accuracy = 95
-Pipeline2.promise_accuracy = 95
-Pipeline2.validation_accuracy = 95
-
-bench_tuner_data["pipeline_GEMO"] = Pipeline2
-
-
-
-
-Pipeline3 = Benchmark()
-Pipeline3.tuner_binary = "pipeline_GSME"
-Pipeline3.promise_binary = "pipeline_GSME_promise"
-Pipeline3.validation_binary = "pipeline_GSME_valid"
-Pipeline3.num_flags = 9
-Pipeline3.num_layers = 4
-Pipeline3.error_range_1 = 10
-Pipeline3.error_range_2 = 15
-Pipeline3.start_promise_range = 2
-Pipeline3.skip_layer_str = "1_3"
-Pipeline3.result_dir_1 = "tuner_results/pipeline_GSME/loss_30/batch1"
-Pipeline3.result_dir_2 = "tuner_results/pipeline_GSME/loss_20/batch1"
-Pipeline3.tensor_desc_file = "tuner_results/pipeline_GSME/pipeline_GSME_tensors.txt"
-Pipeline3.layer_file = "tuner_results/pipeline_GSME/pipeline_GSME_layers.txt"
-#Pipeline1.loss1_result_file = "tuner_results/vgg_cifar100/loss_1/promise_tuned_confs/promise_confs.txt"
-#Pipeline1.loss2_result_file = "tuner_results/vgg_cifar100/loss_2/promise_tuned_confs/promise_confs.txt"
-Pipeline3.autotuner_runs = 300
-Pipeline3.tuner_accuracy = 95
-Pipeline3.promise_accuracy = 95
-Pipeline3.validation_accuracy = 95
-
-bench_tuner_data["pipeline_GSME"] = Pipeline3
-
-
-Pipeline4 = Benchmark()
-Pipeline4.tuner_binary = "pipeline_GEO"
-Pipeline4.promise_binary = "pipeline_GEO_promise"
-Pipeline4.validation_binary = "pipeline_GEO_valid"
-Pipeline4.num_flags = 7
-Pipeline4.num_layers = 3
-Pipeline4.error_range_1 = 10
-Pipeline4.error_range_2 = 15
-Pipeline4.start_promise_range = 2
-Pipeline4.skip_layer_str = "1_2"
-Pipeline4.result_dir_1 = "tuner_results/pipeline_GEO/loss_30/batch1"
-Pipeline4.result_dir_2 = "tuner_results/pipeline_GEO/loss_20/batch1"
-Pipeline4.tensor_desc_file = "tuner_results/pipeline_GEO/pipeline_GEO_tensors.txt"
-Pipeline4.layer_file = "tuner_results/pipeline_GEO/pipeline_GEO_layers.txt"
-#Pipeline1.loss1_result_file = "tuner_results/vgg_cifar100/loss_1/promise_tuned_confs/promise_confs.txt"
-#Pipeline1.loss2_result_file = "tuner_results/vgg_cifar100/loss_2/promise_tuned_confs/promise_confs.txt"
-Pipeline4.autotuner_runs = 300
-Pipeline4.tuner_accuracy = 95
-Pipeline4.promise_accuracy = 95
-Pipeline4.validation_accuracy = 95
-
-bench_tuner_data["pipeline_GEO"] = Pipeline4
-
-
-Pipeline5 = Benchmark()
-Pipeline5.tuner_binary = "pipeline_GSM"
-Pipeline5.promise_binary = "pipeline_GSM_promise"
-Pipeline5.validation_binary = "pipeline_GSM_valid"
-Pipeline5.num_flags = 6
-Pipeline5.num_layers = 3
-Pipeline5.error_range_1 = 10
-Pipeline5.error_range_2 = 15
-Pipeline5.start_promise_range = 2
-Pipeline5.skip_layer_str = "1_1"
-Pipeline5.result_dir_1 = "tuner_results/pipeline_GSM/loss_30/batch1"
-Pipeline5.result_dir_2 = "tuner_results/pipeline_GSM/loss_20/batch1"
-Pipeline5.tensor_desc_file = "tuner_results/pipeline_GSM/pipeline_GSM_tensors.txt"
-Pipeline5.layer_file = "tuner_results/pipeline_GSM/pipeline_GSM_layers.txt"
-#Pipeline1.loss1_result_file = "tuner_results/vgg_cifar100/loss_1/promise_tuned_confs/promise_confs.txt"
-#Pipeline1.loss2_result_file = "tuner_results/vgg_cifar100/loss_2/promise_tuned_confs/promise_confs.txt"
-Pipeline5.autotuner_runs = 300
-Pipeline5.tuner_accuracy = 95
-Pipeline5.promise_accuracy = 95
-Pipeline5.validation_accuracy = 95
-
-bench_tuner_data["pipeline_GSM"] = Pipeline5
-
-"""
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/benchmarks.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/benchmarks.py
deleted file mode 100644
index 0662ddaa76..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/benchmarks.py
+++ /dev/null
@@ -1,599 +0,0 @@
-
-# NOTES: Batch13 (Baseline) Batch14 - With ErrorSens (10, 25, 35)
-
-# Batch 9: No Error Sens. Min : P3
-# Batch 10: No Error Sens + More Runs for Loss1 and Loss2. Min: P3
-# Batch 11: Error Sens: Skipping 30% elems in each : Min: P3. More runs in Loss1 (4000) and Loss2 (2000)
-# Batch 12: Error Sens: 10, 25, 35, for Loss1, 2, 3, respectively, Min: P3. 1000 Runs for All
-# Batch 13: No Error Sens: Equal Runs (1000) for all. Min: P1
-# Batch 14: Reruning Batch12 with bugFix!
-# Batch 15: MAJOR CHANGE: 3 different skip levels for each Loss1,Loss2,Loss3
-
-# Batch 18: Batch13 (Basline) + ParetoCurve (1500 Runs) - BUGGY IGNORE!!!
-
-# Batch 19: (Basline) + ParetoCurve + 2 runs in Tuning Phase (1500 Runs)
-
-# Batch 20: 3-Skip levels + + 2 runs + 1500 Runs + EnergyBandSize now % of Max (Compare against Batch19
-
-
-# Batch 200: AlgoTuner - 1000 images - 1500 runs (IGNORE)
-# Batch 201: AlgoTuner - 2000 images - 1500 runs
-# Batch 202: AlgoTuner - 2000 images - 500 runs
-# Batch 203: AlgoTuner - 2000 images - 3000 runs
-
-
-#---- CHANGES: i) Reshufled inputs ii) 3K images for tuning
-# Batch 210: 3K images, 1000 runs (1500 resnet), no FP32 used in tuning
-# Batch 211: Same as Batch-210 + uses tensorConvPerfCuda*Half*
-
-
-#batch_id = "batch210"
-#batch_id = "batch211"
-#batch_id = "batch210"
-
-# NOTE: Testing new devtuner script
-#batch_id = "batch311"
-
-# NOTE: batch with 3K runs each - new devtuner script
-#batch_id = "batch312"
-
-
-# NOTE: Trying out piped execution
-#batch_id = "batch313"
-
-# NOTE: Trying out piped execution with 3K each - to measure time is the goal
-#batch_id = "batch314"
-
-# NOTE: Trying out VGG16 Imagenet with new error slack approach
-# batch_id = "batch_315"
-
-# NOTE: Using Batch with 2K images in VGG16_imagenet
-#-- batch_id = "batch316"
-
-# Running all non imagenet DNNs for Yifan - ENDED UP TESTING RUN
-#-- batch_id = "batch321"
-
-
-# Running all non imagenet DNNs for Yifan - Long Running
-#--- batch_id = "batch322"
-
-
-# Re-Running VGG16_imagenet after issues with python setup
-#-- batch_id = "batch323"
-
-
-# Re-Running all CIFAR-10 benchmarks after using AUTOMATIC KNOBS
-#-- batch_id = "batch324"
-
-# After Fixing Yasmin's code first batch of runs on CIFAR-10 DNNs
-# NOTE: First batch with 33% sampling - 2K runs for each threshold
-#-- batch_id = "batch325"
-
-
-# After Fixing Yasmin's code second batch of runs on CIFAR-10 DNNs
-# NOTE: Second batch with 33% sampling - 5K runs for each threshold
-# NOTE: First batch with dumping CPU runtime configs
-#-- batch_id = "batch327"
-
-
-# IMP: Increased SAMPLING Knobs ---- Adding interpolation-based Knobs - 8K iterations
-#---- batch_id = "batch328"
-
-
-# IMP: Increased SAMPLING Knobs ---- Adding interpolation-based Knobs -- 12K
-#-- batch_id = "batch329"
-
-# IMP: Increased SAMPLING Knobs ---- Adding interpolation-based Knobs -- 12K - NEW: 5K images calibration set
-#-- batch_id = "batch330"
-
-
-# IMP: Increased SAMPLING Knobs -- 20K iterations - NEW: 5K images calibration set -- fixed bugs
-#-- batch_id = "batch331"
-
-# testing install-time tuner
-#batch_id = "batch340"
-
-# First run of install time tuner
-batch_id = "batch341"
-
-# Install Timer Tuner with FP32 SEEDING
-batch_id = "batch342"
-
-# Install Timer Tuner with FP32 SEEDING
-# FIRST time reducing sampling knobs 239 above
-# Fixed bugs --- added FP32 to search space
-batch_id = "batch343"
-
-
-# testing pareto-only validation
-batch_id = "batch344"
-
-
-# First Install-time tuning run with different iterations per DNN benchmark
-batch_id = "batch345"
-
-
-# First Install-time tuning run with 10K iterations per DNN benchmark
-batch_id = "batch346"
-
-
-
-class Benchmark:
- def __init__(self):
- self.tuner_binary = ""
- self.promise_binary = ""
- self.tuner_accuracy = 0
- self.promise_accuracy = 0
- self.num_flags = 0
- self.num_layers = 0
- self.autotuner_runs = 0
- self.error_range_1 = 0
- self.error_range_2 = 0
- self.result_dir_1 = ""
- self.result_dir_2 = ""
- self.promise_result_dir_1 = ""
- self.promise_result_dir_2 = ""
-
-
-
-bench_tuner_data = {}
-
-
-LeNet = Benchmark()
-LeNet.tuner_binary = "lenet_keras"
-LeNet.fp16_binary = "lenet_half"
-LeNet.promise_binary = "lenet_promise"
-LeNet.piped_binary = "lenet_piped"
-LeNet.validation_binary = "lenet_promise"
-
-LeNet.num_flags = 14
-LeNet.num_layers = 4
-LeNet.error_range_1 = 16
-LeNet.error_range_2 = 20
-LeNet.start_promise_range = 1
-LeNet.skip_layer_str = "0"
-
-LeNet.base_dir = "tuner_results/lenet_keras/"
-
-LeNet.tensor_desc_file = "autotuner/data/lenet/lenet_tensors.txt"
-LeNet.layer_file = "autotuner/data/lenet/lenet_layers.txt"
-LeNet.cost_file = "autotuner/data/lenet/op_cost.txt"
-LeNet.layer_knobs = "autotuner/data/lenet/dev_knobs.txt"
-
-LeNet.autotuner_runs = 2000
-LeNet.tuner_accuracy = 98.9
-LeNet.promise_accuracy = 99.7
-LeNet.validation_accuracy = 99
-
-bench_tuner_data["lenet_keras"] = LeNet
-
-
-
-
-
-
-# FIXIT: Fix Variable Names below
-Alexnet = Benchmark()
-Alexnet.tuner_binary = "alexnet_cifar10_tuner"
-Alexnet.fp16_binary = "alexnet_half"
-Alexnet.promise_binary = "alexnet_promise"
-Alexnet.piped_binary = "alexnet_piped"
-Alexnet.validation_binary = "alexnet_valid"
-Alexnet.num_flags = 21
-Alexnet.num_layers = 6
-Alexnet.error_range_1 = 10
-Alexnet.error_range_2 = 13
-Alexnet.start_promise_range = 1
-Alexnet.skip_layers = 0
-Alexnet.skip_layer_str = "5_0"
-
-Alexnet.base_dir = "tuner_results/alexnet_cifar10/"
-
-Alexnet.tensor_desc_file = "autotuner/data/alexnet/alexnet_tensors.txt"
-Alexnet.layer_file = "autotuner/data/alexnet/alexnet_layers.txt"
-Alexnet.cost_file = "autotuner/data/alexnet/op_cost.txt"
-Alexnet.layer_knobs = "autotuner/data/alexnet/dev_knobs.txt"
-
-Alexnet.autotuner_runs = 4000
-Alexnet.tuner_accuracy = 79.9
-Alexnet.promise_accuracy = 78.86
-Alexnet.validation_accuracy = 79.19
-
-bench_tuner_data["alexnet_cifar10"] = Alexnet
-
-
-Alexnet2 = Benchmark()
-Alexnet2.tuner_binary = "alexnet2_cifar10_tuner"
-Alexnet2.fp16_binary = "alexnet2_half"
-Alexnet2.promise_binary = "alexnet2_promise"
-Alexnet2.piped_binary = "alexnet2_piped"
-Alexnet2.validation_binary = "alexnet2_valid"
-Alexnet2.num_flags = 23
-Alexnet2.num_layers = 7
-Alexnet2.error_range_1 = 10
-Alexnet2.error_range_2 = 13
-Alexnet2.start_promise_range = 1
-Alexnet2.skip_layer_str = "6_1_0"
-
-Alexnet2.base_dir = "tuner_results/alexnet2_cifar10/"
-
-Alexnet2.tensor_desc_file = "autotuner/data/alexnet2/alexnet2_tensors.txt"
-Alexnet2.layer_file = "autotuner/data/alexnet2/alexnet2_layers.txt"
-Alexnet2.cost_file = "autotuner/data/alexnet2/op_cost.txt"
-Alexnet2.layer_knobs = "autotuner/data/alexnet2/dev_knobs.txt"
-Alexnet2.autotuner_runs = 4000
-Alexnet2.tuner_accuracy = 84.19
-Alexnet2.promise_accuracy = 84.7
-Alexnet2.validation_accuracy = 85.15
-
-bench_tuner_data["alexnet2_cifar10"] = Alexnet2
-
-
-
-VGG16_10 = Benchmark()
-VGG16_10.tuner_binary = "vgg16_cifar10_tuner"
-VGG16_10.fp16_binary = "vgg16_cifar10_half"
-VGG16_10.promise_binary = "./vgg16_cifar10_promise"
-VGG16_10.piped_binary = "./vgg16_cifar10_piped"
-VGG16_10.validation_binary = "vgg16_cifar10_valid"
-VGG16_10.num_flags = 50
-VGG16_10.num_layers = 15
-VGG16_10.error_range_1 = 9
-VGG16_10.error_range_2 = 11
-VGG16_10.start_promise_range = 1
-VGG16_10.skip_layer_str = "14_3_4_1_6"
-
-VGG16_10.base_dir = "tuner_results/vgg16_cifar10/"
-
-VGG16_10.tensor_desc_file = "autotuner/data/vgg16_cifar10/vgg16_tensors.txt"
-VGG16_10.layer_file = "autotuner/data/vgg16_cifar10/vgg16_layers.txt"
-VGG16_10.cost_file = "autotuner/data/vgg16_cifar10/op_cost.txt"
-VGG16_10.layer_knobs = "autotuner/data/vgg16_cifar10/dev_knobs.txt"
-
-VGG16_10.autotuner_runs = 8000
-VGG16_10.tuner_accuracy = 90.19
-
-VGG16_10.promise_accuracy = 88.53
-VGG16_10.validation_accuracy = 89.05
-
-bench_tuner_data["vgg16_cifar10"] = VGG16_10
-
-
-
-
-VGG16_100 = Benchmark()
-VGG16_100.tuner_binary = "vgg16_cifar100_tuner"
-VGG16_100.fp16_binary = "vgg16_cifar100_half"
-VGG16_100.promise_binary = "vgg16_cifar100_promise"
-VGG16_100.piped_binary = "vgg16_cifar100_piped"
-VGG16_100.validation_binary = "vgg16_cifar100_valid"
-VGG16_100.num_flags = 50
-VGG16_100.num_layers = 15
-VGG16_100.error_range_1 = 9
-VGG16_100.error_range_2 = 11
-VGG16_100.start_promise_range = 1
-VGG16_100.skip_layer_str = "0_1_2_3_4"
-
-VGG16_100.base_dir = "tuner_results/vgg16_cifar100/"
-
-VGG16_100.tensor_desc_file = "autotuner/data/vgg16_cifar100/vgg16_tensors.txt"
-VGG16_100.layer_file = "autotuner/data/vgg16_cifar100/vgg16_layers.txt"
-VGG16_100.cost_file = "autotuner/data/vgg16_cifar100/op_cost.txt"
-VGG16_100.layer_knobs = "autotuner/data/vgg16_cifar100/dev_knobs.txt"
-
-VGG16_100.autotuner_runs = 5000
-VGG16_100.tuner_accuracy = 67.95
-
-VGG16_100.promise_accuracy = 67.86
-VGG16_100.validation_accuracy = 68.65
-
-bench_tuner_data["vgg16_cifar100"] = VGG16_100
-
-
-
-
-VGG16_imagenet = Benchmark()
-VGG16_imagenet.tuner_binary = ""
-VGG16_imagenet.fp16_binary = ""
-VGG16_imagenet.promise_binary = "vgg16_imagenet_promise"
-VGG16_imagenet.piped_binary = "vgg16_imagenet_piped"
-VGG16_imagenet.validation_binary = "vgg16_imagenet_promise"
-VGG16_imagenet.num_flags = 53
-VGG16_imagenet.num_layers = 16
-
-VGG16_imagenet.base_dir = "tuner_results/vgg16_imagenet/"
-VGG16_imagenet.tensor_desc_file = "autotuner/data/vgg16_imagenet/vgg16_tensors.txt"
-VGG16_imagenet.layer_file = "autotuner/data/vgg16_imagenet/vgg16_layers.txt"
-VGG16_imagenet.cost_file = "autotuner/data/vgg16_imagenet/op_cost.txt"
-VGG16_imagenet.layer_knobs = "autotuner/data/vgg16_imagenet/dev_knobs.txt"
-
-VGG16_imagenet.autotuner_runs = 5000
-VGG16_imagenet.tuner_accuracy = 0.0
-VGG16_imagenet.promise_accuracy = 69.62
-VGG16_imagenet.validation_accuracy = 69.62
-
-#-- bench_tuner_data["vgg16_imagenet"] = VGG16_imagenet
-
-
-
-
-ResNet = Benchmark()
-ResNet.tuner_binary = "resnet18_cifar10_tuner"
-ResNet.fp16_binary = "resnet18_half"
-ResNet.promise_binary = "resnet18_promise"
-ResNet.piped_binary = "resnet18_piped"
-ResNet.validation_binary = "resnet18_valid"
-ResNet.num_flags = 73
-ResNet.num_layers = 22
-ResNet.error_range_1 = 7
-ResNet.error_range_2 = 9
-ResNet.start_promise_range = 1
-
-ResNet.skip_layer_str = "0_1_2_14_15_17_18_21"
-ResNet.base_dir = "tuner_results/resnet18_cifar10/"
-
-ResNet.tensor_desc_file = "autotuner/data/resnet/resnet_tensors.txt"
-ResNet.layer_file = "autotuner/data/resnet/resnet_layers.txt"
-ResNet.cost_file = "autotuner/data/resnet/op_cost.txt"
-ResNet.layer_knobs = "autotuner/data/resnet/dev_knobs.txt"
-
-ResNet.autotuner_runs = 8000
-ResNet.tuner_accuracy = 89.6
-
-ResNet.promise_accuracy = 89.5
-ResNet.validation_accuracy = 89.65
-
-bench_tuner_data["resnet18_cifar10"] = ResNet
-
-
-
-
-
-
-ResNet50 = Benchmark()
-ResNet50.tuner_binary = ""
-ResNet50.fp16_binary = ""
-ResNet50.promise_binary = "resnet50_imagenet_promise"
-ResNet50.piped_binary = "resnet50_imagenet_piped"
-ResNet50.validation_binary = "resnet50_valid"
-ResNet50.num_flags = 1 # FIXIT
-ResNet50.num_layers = 54
-
-ResNet50.base_dir = "tuner_results/resnet50_imagenet/"
-
-ResNet50.tensor_desc_file = "autotuner/data/resnet50_imagenet/resnet50_tensors.txt"
-ResNet50.layer_file = "autotuner/data/resnet50_imagenet/resnet50_layers.txt"
-ResNet50.cost_file = "autotuner/data/resnet50_imagenet/op_cost.txt"
-ResNet50.layer_knobs = "autotuner/data/resnet50_imagenet/dev_knobs.txt"
-
-ResNet50.autotuner_runs = 5000
-ResNet50.tuner_accuracy = 89.6
-
-ResNet50.promise_accuracy = 77
-ResNet50.validation_accuracy = 20 # FIXIT
-
-#--- bench_tuner_data["resnet50_imagenet"] = ResNet50
-
-
-
-
-
-
-
-
-
-MobileNet = Benchmark()
-MobileNet.tuner_binary = "mobilenet_cifar10"
-MobileNet.fp16_binary = "mobilenet_half"
-MobileNet.promise_binary = "mobilenet_promise"
-MobileNet.piped_binary = "mobilenet_piped"
-MobileNet.validation_binary = "mobilenet_valid"
-MobileNet.num_flags = 85
-MobileNet.num_layers = 15
-MobileNet.error_range_1 = 7
-MobileNet.error_range_2 = 8
-MobileNet.start_promise_range = 1
-
-MobileNet.skip_layer_str = "1_14_0_6_2"
-MobileNet.base_dir = "tuner_results/mobilenet/"
-
-MobileNet.tensor_desc_file = "autotuner/data/mobilenet/mobilenet_ops.txt"
-MobileNet.layer_file = "autotuner/data/mobilenet/mobilenet_layer_comp.txt"
-MobileNet.cost_file = "autotuner/data/mobilenet/op_cost.txt"
-MobileNet.layer_knobs = "autotuner/data/mobilenet/dev_knobs.txt"
-
-MobileNet.autotuner_runs = 8000
-MobileNet.tuner_accuracy = 84.8
-
-MobileNet.promise_accuracy = 83.73
-MobileNet.validation_accuracy = 84.4
-
-bench_tuner_data["mobilenet_cifar10"] = MobileNet
-
-
-
-MobileNet_SH = Benchmark()
-MobileNet_SH.tuner_binary = "mobilenet_cifar10_shallow"
-MobileNet_SH.fp16_binary = "mobilenet_shallow_half"
-MobileNet_SH.promise_binary = "mobilenet_shallow_promise"
-MobileNet_SH.piped_binary = "mobilenet_shallow_piped"
-MobileNet_SH.validation_binary = "mobilenet_shallow_valid"
-MobileNet_SH.num_flags = 42
-MobileNet_SH.num_layers = 8
-MobileNet_SH.error_range_1 = 10
-MobileNet_SH.error_range_2 = 12
-MobileNet_SH.start_promise_range = 1
-
-MobileNet_SH.skip_layer_str = "7_0_1"
-MobileNet_SH.base_dir = "tuner_results/mobilenet_shallow/"
-
-MobileNet_SH.tensor_desc_file = "autotuner/data/mobilenet_shallow/mobilenet_shallow_ops.txt"
-MobileNet_SH.layer_file = "autotuner/data/mobilenet_shallow/mobilenet_shallow_layer_comp.txt"
-MobileNet_SH.cost_file = "autotuner/data/mobilenet_shallow/op_cost.txt"
-MobileNet_SH.layer_knobs = "autotuner/data/mobilenet_shallow/dev_knobs.txt"
-
-
-MobileNet_SH.autotuner_runs = 1000
-MobileNet_SH.tuner_accuracy = 87.6
-
-MobileNet_SH.promise_accuracy = 87.76
-MobileNet_SH.validation_accuracy = 88.5
-
-#-- bench_tuner_data["mobilenet_shallow"] = MobileNet_SH
-
-
-
-"""
-Alexnet9 = Benchmark()
-FC4.tuner_binary = "fc4_clipped"
-FC4.promise_binary = ""
-FC4.validation_binary = ""
-FC4.num_flags = 12
-FC4.num_layers = 4
-FC4.error_range_1 = 12
-FC4.error_range_2 = 16
-FC4.start_promise_range = 3
-FC4.skip_layer_str = "0"
-FC4.base_dir = "../build_tuner/tuner_results/fc4/"
-FC4.result_dir_1 = "../build_tuner/tuner_results/fc4/loss1/batch1"
-FC4.result_dir_2 = "../build_tuner/tuner_results/fc4/loss2/batch1"
-FC4.tensor_desc_file = ""
-FC4.layer_file = ""
-
-FC4.loss1_result_file = ""
-FC4.loss2_result_file = ""
-
-FC4.autotuner_runs = 1000
-FC4.tuner_accuracy = 93.8
-FC4.promise_accuracy = 0.0
-FC4.validation_accuracy = 0.0
-
-bench_tuner_data["fc4"] = FC4
-
-
-
-
-Pipeline1 = Benchmark()
-Pipeline1.tuner_binary = "pipeline_GEOM"
-Pipeline1.promise_binary = "pipeline_GEOM_promise"
-Pipeline1.validation_binary = "pipeline_GEOM_valid"
-Pipeline1.num_flags = 9
-Pipeline1.num_layers = 4
-Pipeline1.error_range_1 = 10
-Pipeline1.error_range_2 = 15
-Pipeline1.start_promise_range = 2
-Pipeline1.skip_layer_str = "1_2"
-Pipeline1.result_dir_1 = "tuner_results/pipeline_GEOM/loss_30/batch1"
-Pipeline1.result_dir_2 = "tuner_results/pipeline_GEOM/loss_20/batch1"
-Pipeline1.tensor_desc_file = "tuner_results/pipeline_GEOM/pipeline_GEOM_tensors.txt"
-Pipeline1.layer_file = "tuner_results/pipeline_GEOM/pipeline_GEOM_layers.txt"
-#Pipeline1.loss1_result_file = "tuner_results/vgg_cifar100/loss_1/promise_tuned_confs/promise_confs.txt"
-#Pipeline1.loss2_result_file = "tuner_results/vgg_cifar100/loss_2/promise_tuned_confs/promise_confs.txt"
-Pipeline1.autotuner_runs = 300
-Pipeline1.tuner_accuracy = 95
-Pipeline1.promise_accuracy = 95
-Pipeline1.validation_accuracy = 95
-
-bench_tuner_data["pipeline_GEOM"] = Pipeline1
-
-
-Pipeline2 = Benchmark()
-Pipeline2.tuner_binary = "pipeline_GEMO"
-Pipeline2.promise_binary = "pipeline_GEMO_promise"
-Pipeline2.validation_binary = "pipeline_GEMO_valid"
-Pipeline2.num_flags = 9
-Pipeline2.num_layers = 4
-Pipeline2.error_range_1 = 10
-Pipeline2.error_range_2 = 15
-Pipeline2.start_promise_range = 2
-Pipeline2.skip_layer_str = "1_3"
-Pipeline2.result_dir_1 = "tuner_results/pipeline_GEMO/loss_30/batch1"
-Pipeline2.result_dir_2 = "tuner_results/pipeline_GEMO/loss_20/batch1"
-Pipeline2.tensor_desc_file = "tuner_results/pipeline_GEMO/pipeline_GEMO_tensors.txt"
-Pipeline2.layer_file = "tuner_results/pipeline_GEMO/pipeline_GEMO_layers.txt"
-#Pipeline1.loss1_result_file = "tuner_results/vgg_cifar100/loss_1/promise_tuned_confs/promise_confs.txt"
-#Pipeline1.loss2_result_file = "tuner_results/vgg_cifar100/loss_2/promise_tuned_confs/promise_confs.txt"
-Pipeline2.autotuner_runs = 300
-Pipeline2.tuner_accuracy = 95
-Pipeline2.promise_accuracy = 95
-Pipeline2.validation_accuracy = 95
-
-bench_tuner_data["pipeline_GEMO"] = Pipeline2
-
-
-
-
-Pipeline3 = Benchmark()
-Pipeline3.tuner_binary = "pipeline_GSME"
-Pipeline3.promise_binary = "pipeline_GSME_promise"
-Pipeline3.validation_binary = "pipeline_GSME_valid"
-Pipeline3.num_flags = 9
-Pipeline3.num_layers = 4
-Pipeline3.error_range_1 = 10
-Pipeline3.error_range_2 = 15
-Pipeline3.start_promise_range = 2
-Pipeline3.skip_layer_str = "1_3"
-Pipeline3.result_dir_1 = "tuner_results/pipeline_GSME/loss_30/batch1"
-Pipeline3.result_dir_2 = "tuner_results/pipeline_GSME/loss_20/batch1"
-Pipeline3.tensor_desc_file = "tuner_results/pipeline_GSME/pipeline_GSME_tensors.txt"
-Pipeline3.layer_file = "tuner_results/pipeline_GSME/pipeline_GSME_layers.txt"
-#Pipeline1.loss1_result_file = "tuner_results/vgg_cifar100/loss_1/promise_tuned_confs/promise_confs.txt"
-#Pipeline1.loss2_result_file = "tuner_results/vgg_cifar100/loss_2/promise_tuned_confs/promise_confs.txt"
-Pipeline3.autotuner_runs = 300
-Pipeline3.tuner_accuracy = 95
-Pipeline3.promise_accuracy = 95
-Pipeline3.validation_accuracy = 95
-
-bench_tuner_data["pipeline_GSME"] = Pipeline3
-
-
-Pipeline4 = Benchmark()
-Pipeline4.tuner_binary = "pipeline_GEO"
-Pipeline4.promise_binary = "pipeline_GEO_promise"
-Pipeline4.validation_binary = "pipeline_GEO_valid"
-Pipeline4.num_flags = 7
-Pipeline4.num_layers = 3
-Pipeline4.error_range_1 = 10
-Pipeline4.error_range_2 = 15
-Pipeline4.start_promise_range = 2
-Pipeline4.skip_layer_str = "1_2"
-Pipeline4.result_dir_1 = "tuner_results/pipeline_GEO/loss_30/batch1"
-Pipeline4.result_dir_2 = "tuner_results/pipeline_GEO/loss_20/batch1"
-Pipeline4.tensor_desc_file = "tuner_results/pipeline_GEO/pipeline_GEO_tensors.txt"
-Pipeline4.layer_file = "tuner_results/pipeline_GEO/pipeline_GEO_layers.txt"
-#Pipeline1.loss1_result_file = "tuner_results/vgg_cifar100/loss_1/promise_tuned_confs/promise_confs.txt"
-#Pipeline1.loss2_result_file = "tuner_results/vgg_cifar100/loss_2/promise_tuned_confs/promise_confs.txt"
-Pipeline4.autotuner_runs = 300
-Pipeline4.tuner_accuracy = 95
-Pipeline4.promise_accuracy = 95
-Pipeline4.validation_accuracy = 95
-
-bench_tuner_data["pipeline_GEO"] = Pipeline4
-
-
-Pipeline5 = Benchmark()
-Pipeline5.tuner_binary = "pipeline_GSM"
-Pipeline5.promise_binary = "pipeline_GSM_promise"
-Pipeline5.validation_binary = "pipeline_GSM_valid"
-Pipeline5.num_flags = 6
-Pipeline5.num_layers = 3
-Pipeline5.error_range_1 = 10
-Pipeline5.error_range_2 = 15
-Pipeline5.start_promise_range = 2
-Pipeline5.skip_layer_str = "1_1"
-Pipeline5.result_dir_1 = "tuner_results/pipeline_GSM/loss_30/batch1"
-Pipeline5.result_dir_2 = "tuner_results/pipeline_GSM/loss_20/batch1"
-Pipeline5.tensor_desc_file = "tuner_results/pipeline_GSM/pipeline_GSM_tensors.txt"
-Pipeline5.layer_file = "tuner_results/pipeline_GSM/pipeline_GSM_layers.txt"
-#Pipeline1.loss1_result_file = "tuner_results/vgg_cifar100/loss_1/promise_tuned_confs/promise_confs.txt"
-#Pipeline1.loss2_result_file = "tuner_results/vgg_cifar100/loss_2/promise_tuned_confs/promise_confs.txt"
-Pipeline5.autotuner_runs = 300
-Pipeline5.tuner_accuracy = 95
-Pipeline5.promise_accuracy = 95
-Pipeline5.validation_accuracy = 95
-
-bench_tuner_data["pipeline_GSM"] = Pipeline5
-
-"""
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/buildRtConfig.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/buildRtConfig.py
deleted file mode 100644
index 5a6a9e0f03..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/buildRtConfig.py
+++ /dev/null
@@ -1,583 +0,0 @@
-
-
-import os
-import sys
-import utils
-from benchmarks import bench_tuner_data
-from swing_selection import loadLayerDesc
-from benchmarks import batch_id
-
-
-op_mapping = {}
-op_mapping["conv"] = "conv"
-op_mapping["depthwise_conv"] = "group_conv"
-op_mapping["dense"] = "mul"
-op_mapping["batchnorm"] = "batchnorm"
-op_mapping["pool"] = "pool_max"
-op_mapping["pool_mean"] = "pool_mean"
-op_mapping["activation"] = "relu"
-op_mapping["tanh"] = "tanh"
-op_mapping["add"] = "add"
-
-
-approx_map = {}
-
-
-def initializeApproxMap(knobs_file_path):
-
- f = open(knobs_file_path, "r")
-
- for x in f:
- toks = x.split("\t")
- approx_type = toks[0].split(",")[0]
- knob_id = toks[0].split(",")[1]
- approx_str = approx_type + " " + knob_id
- approx_map[knob_id] = approx_str
-
-
- print (approx_map)
-
-
-
-
-
-
-class Config:
- def __init__(self):
- self.avg_accuracy = 0
- self.avg_loss = 0
- self.speedup = 1
- self.fname = ""
- self.flags = []
-
-
-
-
-def isLayer(layer_comp):
- if layer_comp[0] == "dense" or layer_comp[0] == "conv":
- return True
- else:
- return False
-
-
-
-def getOpMapping(op_name):
-
- if op_name not in op_mapping:
- print ("ERROR: OP not found!! = ", op_name, "\n")
- sys.exit(0)
-
- return op_mapping[op_name]
-
-
-
-def getApproxMapping(flag, layer_comp):
-
- flag_str = str(flag)
- if flag_str not in approx_map:
- print ("ERROR: OP not found!! = ", flag_str, "\n")
- sys.exit(0)
-
- if "dense" in layer_comp and flag > 7:
- if flag == 12:
- return "fp16 1"
- else:
- return "fp32 1"
-
-
- return approx_map[flag_str]
-
-
-
-def skipFile(fname):
-
- skip_files = {}
- skip_files["confidence_summary.txt"] = 1
- skip_files["promise_confs.txt"] = 1
-
- if "accuracy" in fname: # *_accuracy files should be skipped
- return True
-
- if "norms" in fname: # *_accuracy files should be skipped
- return True
-
- if ".#" in fname: # *_accuracy files should be skipped
- return True
-
- #if "_promise" in fname: # *_accuracy files should be skipped
- # return True
-
- if not fname[-1].isdigit():
- return True
-
- if fname in skip_files:
- return True
- else:
- return False
-
-
-
-def parseTopLine(x):
-
- toks = x.split()
-
- speedup = 1.0
- accuracy = 0.0
- for tok in toks:
- if "avg_accuracy" in tok:
- avg_accuracy = float(tok.split("=")[1])
- if "speedup" in tok:
- speedup = float(tok.split("=")[1])
-
-
- return avg_accuracy, speedup
-
-
-
-def loadConfigData(result_dir, baseline_accuracy, sub_dir = "high_confidence"):
-
- config_arr = []
-
- #result_dir += "/promise_tuner/high_confidence/"
- #result_dir += "/algo_tuner/high_confidence/"
- result_dir += "/algo_tuner/" + sub_dir + "/"
- file_names = os.listdir(result_dir)
-
-
- for fname in file_names:
- if not skipFile(fname):
-
- fpath = result_dir + fname
- config = Config()
- f = open(fpath, "r")
-
- it = 0
- for x in f:
- if x.strip == "":
- continue
- if it == 0:
- avg_accuracy, speedup = parseTopLine(x)
- config.avg_accuracy = avg_accuracy
- config.avg_loss = baseline_accuracy - avg_accuracy
- config.speedup = speedup
- config.fname = fname
- #print ("acc = " + str(avg_accuracy) + "\n")
- else:
- flag = int(x.strip())
- config.flags.append(flag)
- it += 1
-
- config_arr.append(config)
-
-
- return config_arr
-
-
-
-
-def loadConfigsFromDir(result_dir, baseline_accuracy):
-
- config_arr = []
- file_names = os.listdir(result_dir)
-
- for fname in file_names:
- if not skipFile(fname):
-
- fpath = result_dir + '/' + fname
- config = Config()
- f = open(fpath, "r")
-
- it = 0
- for x in f:
- if x.strip == "":
- continue
- if it == 0:
- avg_accuracy, speedup = parseTopLine(x)
- config.avg_accuracy = avg_accuracy
- config.avg_loss = baseline_accuracy - avg_accuracy
- config.speedup = speedup
- config.fname = fname
- #print ("acc = " + str(avg_accuracy) + "\n")
- else:
- flag = int(x.strip())
- config.flags.append(flag)
- it += 1
-
- config_arr.append(config)
-
-
- return config_arr
-
-
-
-
-
-
-
-def loadPromiseConfigs(result_dir, baseline_accuracy, sub_dir = "promise_test"):
-
- config_arr = []
- result_dir += "/algo_tuner/" + sub_dir + "/"
- file_names = os.listdir(result_dir)
-
- for fname in file_names:
- if "_promise" in fname:
-
- fpath = result_dir + fname
- config = Config()
- f = open(fpath, "r")
-
- it = 0
- for x in f:
- if x.strip == "":
- continue
-
- if it == 0:
- avg_accuracy, speedup = parseTopLine(x)
- config.avg_accuracy = avg_accuracy
- config.avg_loss = baseline_accuracy - avg_accuracy
- config.speedup = speedup
- config.fname = fname
- #print ("acc = " + str(avg_accuracy) + "\n")
- else:
- flag = int(x.strip())
- config.flags.append(flag)
-
- it += 1
-
- config_arr.append(config)
-
-
- return config_arr
-
-
-
-
-
-def getFP(flag):
-
- if flag < 11:
- return "fp16"
- else:
- return "fp32"
-
-
-
-def getHardwareTarget(flag):
-
- if flag <= 7:
- return "promise"
- else:
- return "gpu"
-
- return "gpu"
-
-
-def handlePromiseConfs(flag, layer_comp):
-
- approx_tech = getApproxMapping(flag, layer_comp)
- config_str = ""
- if flag <= 7:
- config_str += approx_tech + " "
-
- return config_str
-
-
-def handleGPUApproxs(flag, layer_comp):
-
- approx_tech = getApproxMapping(flag, layer_comp)
- config_str = ""
- if flag > 7:
- utils.debug_print ("flag = " + str(flag))
- config_str += getOpMapping(layer_comp[0]) + " " + approx_tech + " "
- for op in layer_comp[1:]:
- utils.debug_print (layer_comp[1:])
- utils.debug_print (op)
-
- op_name = getOpMapping(op)
- config_str += str(op_name) + " " + getFP(flag) + " 1 "
-
- return config_str
-
-
-def generateBaselineConfig(layer_comp):
-
- config_str = ""
- config_str += "gpu "
- for op in layer_comp:
- op_name = getOpMapping(op)
- config_str += str(op_name) + " fp16 1 "
-
- return config_str
-
-
-
-
-
-def buildConfigStr(config, layer_desc, hardware_target):
-
- index = 1
- it = 0
- flags = config.flags
- config_str = ""
-
- for layer_comp in layer_desc:
- config_str += str(index) + " "
- #-- print ("laye_comp = ", layer_comp)
-
- if isLayer(layer_comp):
- flag = flags[it]
- it += 1
-
- utils.debug_print ("flag* = " + str(flag))
- # Add Target Target - GPU, PROMISE
- #config_str += getHardwareTarget(flag) + " "
-
- config_str += hardware_target + " "
-
- utils.debug_print ("config_str = " + str(config_str))
-
- config_str += handlePromiseConfs(flag, layer_comp)
- config_str += handleGPUApproxs(flag, layer_comp)
-
- else: # if a non-Layer Operation
- config_str += generateBaselineConfig(layer_comp)
-
-
- config_str += "\n"
- index += 1
-
-
- config_str += str(index) + " " + hardware_target + " softmax fp32 1\n"
-
- return config_str
-
-
-
-# Adjusts for expected loss on unseen dataset
-def adjustDevTimeLoss(loss):
-
- # Adjusts for negative and low loss values
- if loss < 0.3:
- loss += 0.4
- else:
- loss = loss * 1.33 # 33% extra error for unseen data
-
- if loss < 0.0:
- loss = 0.1
-
- return loss
-
-
-
-def adjustConfigLosses(configurations):
-
- for config in configurations:
- config.avg_loss = adjustDevTimeLoss(config.avg_loss)
-
-
-
-
-def dumpConfig(layer_desc, config_arrs, result_dir):
-
- f = open(result_dir + "/tuner_pareto_confs_" + batch_id + ".txt", "w+")
- it = 1
- for config in config_arrs:
- f.write("+++++\n")
- f.write("conf" + str(it) + " " + str(config.speedup) + " 0 " + \
- str(config.avg_accuracy) + " " + str(config.avg_loss) + "\n")
-
- config_str = buildConfigStr(config, layer_desc)
-
- f.write(config_str)
- f.write("-----\n")
-
- it += 1
-
-
-def dumpBaseLineConfig(conf_id, perf_improv, energy_red, \
- baseline_acc, hardware_target, bench_layer_composition, f_out):
-
- f_out.write("+++++\n")
- f_out.write("conf" + str(conf_id) + " " + str(perf_improv) + " " + str(energy_red) + " " + \
- str(baseline_acc) + " " + str(0) + "\n")
-
- config_str = genFP32Config(bench_layer_composition, hardware_target)
-
- f_out.write(config_str)
- f_out.write("-----\n")
-
-
-
-def genFP32Config(layer_comp, hardware_target):
-
- it = 1
- config_str = ""
- for layer in layer_comp:
- config_str += str(it) + " "
- config_str += hardware_target + " "
-
- for op in layer:
- op_name = getOpMapping(op)
- config_str += str(op_name) + " fp32 1 "
-
- config_str += "\n"
-
- it += 1
-
- config_str += str(it) + " " + hardware_target + " softmax fp32 1\n"
-
- return config_str
-
-
-
-# ***** Exported Interface --- Generates file used by HPVM RT controller ******/
-def dumpDevConfigsToRTFile(configurations, config_out_path, \
- bench_layer_composition, baseline_acc, hardware_target):
-
- f = open(config_out_path, "w+")
-
- dumpBaseLineConfig(1, 1.0, 0, baseline_acc, hardware_target, bench_layer_composition, f)
-
- it = 2
- for config in configurations:
- f.write("+++++\n")
- f.write("conf" + str(it) + " " + str(config.speedup) + " 0 " + \
- str(config.avg_accuracy) + " " + str(config.avg_loss) + "\n")
-
- config_str = buildConfigStr(config, bench_layer_composition, hardware_target)
-
- f.write(config_str)
- f.write("-----\n")
-
- it += 1
-
-
-
-def prependBaseline(Bench):
-
- f1 = open(Bench.base_dir + "/tuner_confs_base.txt", "r")
- baseline_str = f1.read()
- f1.close()
-
- f2 = open(Bench.base_dir + "/tuner_pareto_confs_" + batch_id + ".txt", "r")
- config_str = f2.read()
- f2.close()
-
- f3 = open(Bench.base_dir + "/tuner_pareto_confs_" + batch_id + ".txt", "w+")
- f3.write(baseline_str)
- f3.write(config_str)
- f3.close()
-
-
-
-def generateConf(Bench):
-
- layer_desc = loadLayerDesc(Bench.layer_file)
-
- utils.debug_print ("layer_desc = ", layer_desc)
-
- #config_arr1 = loadConfigData(Bench.result_dir_1, Bench.promise_accuracy)
- #config_arr2 = loadConfigData(Bench.result_dir_2, Bench.promise_accuracy)
- #config_arr3 = loadConfigData(Bench.result_dir_3, Bench.promise_accuracy)
-
- result_dir1 = Bench.result_dir_1 + "/algo_tuner/pareto/"
- result_dir2 = Bench.result_dir_2 + "/algo_tuner/pareto/"
- result_dir3 = Bench.result_dir_3 + "/algo_tuner/pareto/"
-
- config_arr1 = loadConfigsFromDir(result_dir1, Bench.promise_accuracy)
- config_arr2 = loadConfigsFromDir(result_dir2, Bench.promise_accuracy)
- config_arr3 = loadConfigsFromDir(result_dir3, Bench.promise_accuracy)
-
- config_arrs = config_arr1 + config_arr2 + config_arr3
-
- dumpConfig(layer_desc, config_arrs, Bench.base_dir)
-
- prependBaseline(Bench)
-
-
-
-
-def dumpBaselineConfs(Bench):
-
- layer_desc = loadLayerDesc(Bench.layer_file)
-
- #-- print("layer_desc = ", layer_desc)
-
- f = open(Bench.base_dir + "/tuner_confs_base.txt", "w+")
-
- f.write("+++++\n")
- f.write("conf" + str(1) + " " + str(1) + " 0 " + str(Bench.promise_accuracy) + " " + str(0) + "\n")
-
- config = Config()
- flags = []
- for i in range(Bench.num_layers):
- flags.append(11)
-
- config.flags = flags
- config_str = buildConfigStr(config, layer_desc)
-
- f.write(config_str)
- f.write("-----\n")
-
-
- f.close()
-
-
- #f.write("+++++\n")
- #f.write("conf" + str(2) + " " + str(1.5) + " 0 " + str(Bench.promise_accuracy) + " " + str(0) + "\n")
-
- #config = Config()
- #flags = []
- #for i in range(Bench.num_layers):
- # flags.append(12)
-
- #config.flags = flags
- #config_str = buildConfigStr(config, layer_desc)
-
- #f.write(config_str)
- #f.write("-----\n")
-
-
-
-
-
-
-if __name__ == "__main__":
-
- Bench = bench_tuner_data["alexnet_cifar10"]
- generateConf(Bench)
- dumpBaselineConfs(Bench)
-
- Bench = bench_tuner_data["alexnet2_cifar10"]
- generateConf(Bench)
- dumpBaselineConfs(Bench)
-
- Bench = bench_tuner_data["vgg16_cifar10"]
- generateConf(Bench)
- dumpBaselineConfs(Bench)
-
- Bench = bench_tuner_data["vgg16_cifar100"]
- generateConf(Bench)
- dumpBaselineConfs(Bench)
-
- Bench = bench_tuner_data["resnet18_cifar10"]
- generateConf(Bench)
- dumpBaselineConfs(Bench)
-
- Bench = bench_tuner_data["lenet_keras"]
- generateConf(Bench)
- dumpBaselineConfs(Bench)
-
- Bench = bench_tuner_data["mobilenet_cifar10"]
- generateConf(Bench)
- dumpBaselineConfs(Bench)
-
- Bench = bench_tuner_data["mobilenet_shallow"]
- generateConf(Bench)
- dumpBaselineConfs(Bench)
-
-
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/compareResults.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/compareResults.py
deleted file mode 100644
index 6ee7466242..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/compareResults.py
+++ /dev/null
@@ -1,66 +0,0 @@
-
-
-
-import os
-from benchmarks import bench_tuner_data
-from buildRtConfig import loadConfigData
-from buildRtConfig import loadConfigsFromDir
-
-
-
-def compareBench(batch_ids, Bench):
-
- losses = ["1", "2", "3"]
- for loss in losses:
- print ("\n Loss = ", loss, " % \n")
- for id in batch_ids:
- result_dir = Bench.base_dir + "/loss_" + loss + "/batch" + id
- #config_arr = loadConfigData(result_dir, Bench.promise_accuracy, "high_confidence")
-
- #result_dir += "/algo_tuner/high_confidence/"
- result_dir += "/promise_tuner3/high_confidence/"
-
- config_arr = loadConfigsFromDir(result_dir, Bench.promise_accuracy)
-
- count = len(config_arr)
- if len(config_arr) > 0:
- max_speedup = max(config.speedup for config in config_arr)
- else:
- max_speedup = 1.0
- print ("Bench = ", Bench.promise_binary, " BatchID = ", id, " Loss = ", loss, " Count = ", count, " MaxS = ", max_speedup)
-
-
-
-
-if __name__ == "__main__":
-
-
- batch_ids = []
-
- #batch_ids.append("13") # No Error Sens - baseline
- #batch_ids.append("14") # Ops Skipped 10% for Loss1, 25% Loss2, 40% Loss3
- #batch_ids.append("15") # 3 differnet levels for each of Loss1, Loss2, Loss3
- #batch_ids.append("19") # Baseline + Pareto
- #batch_ids.append("20") # Batch18 + Pareto
-
- #batch_ids.append("101") # Algo-specific tuning
-
- #batch_ids.append("201") # Algo-specific tuning
-
- #---- batch_ids.append("202") # Algo-specific tuning
- #batch_ids.append("212") # Algo-specific tuning
- #batch_ids.append("211") # Algo-specific tuning
-
-
- batch_ids.append("220") # Algo-specific tuning
-
-
- compareBench(batch_ids, bench_tuner_data["lenet_keras"])
- compareBench(batch_ids, bench_tuner_data["alexnet_cifar10"])
- compareBench(batch_ids, bench_tuner_data["mobilenet_cifar10"])
-
- compareBench(batch_ids, bench_tuner_data["alexnet2_cifar10"])
- compareBench(batch_ids, bench_tuner_data["vgg16_cifar10"])
- compareBench(batch_ids, bench_tuner_data["vgg16_cifar100"])
- compareBench(batch_ids, bench_tuner_data["resnet18_cifar10"])
- compareBench(batch_ids, bench_tuner_data["mobilenet_shallow"])
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/compute_confs.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/compute_confs.py
deleted file mode 100644
index f82c09095c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/compute_confs.py
+++ /dev/null
@@ -1,56 +0,0 @@
-
-
-from swing_selection import compute_swing_selection
-from swing_selection2 import compute_swing_selection2
-
-
-def computeBenchSwings(Bench):
-
- dir_prefix = "../build_tuner/"
-
- loss_confs = []
- conf_ranks = []
- # Swing selection for 1% and 2% results
- #Bench = bench_tuner_data[bench_name]
- tuned_result_dir = dir_prefix + Bench.result_dir_1 + "/high_confidence/"
- layer_file = Bench.layer_file
- layer_swings, file_names = compute_swing_selection(tuned_result_dir, layer_file)
- loss_confs.append(layer_swings)
- conf_ranks.append(file_names)
- print (file_names)
-
- tuned_result_dir = dir_prefix + Bench.result_dir_2 + "/high_confidence/"
- layer_swings, file_names = compute_swing_selection(tuned_result_dir, layer_file)
- loss_confs.append(layer_swings)
- conf_ranks.append(file_names)
- print (file_names)
-
-
- return loss_confs, conf_ranks
-
-
-
-
-
-def computePSNRBenchSwings(Bench):
-
- loss_confs = []
- conf_ranks = []
- # Swing selection for 1% and 2% results
- #Bench = bench_tuner_data[bench_name]
- tuned_result_dir = Bench.result_dir_1 + "/high_confidence/"
- layer_file = Bench.layer_file
- layer_swings, file_names = compute_swing_selection2(tuned_result_dir, layer_file)
- loss_confs.append(layer_swings)
- conf_ranks.append(file_names)
- print (file_names)
-
- tuned_result_dir = Bench.result_dir_2 + "/high_confidence/"
- layer_swings, file_names = compute_swing_selection2(tuned_result_dir, layer_file)
- loss_confs.append(layer_swings)
- conf_ranks.append(file_names)
- print (file_names)
-
-
- return loss_confs, conf_ranks
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/error_sensitivity.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/error_sensitivity.py
deleted file mode 100644
index 1864771642..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/error_sensitivity.py
+++ /dev/null
@@ -1,378 +0,0 @@
-
-
-import subprocess
-import os
-import operator
-from benchmarks import bench_tuner_data
-from swing_selection import loadLayerDesc
-import math
-
-
-def constructTunerFile(num_flags, tensor_id, error_level, default_error):
-
- f = open("opentuner_flags", "w+")
-
- for i in range(num_flags):
- if i == tensor_id:
- f.write(str(error_level) + "\n")
- else:
- f.write(str(default_error) + "\n")
-
- f.close()
-
-
-
-def runAndTestError(binary_name, gold_acc):
-
- num_runs = 10
-
- binary_name = "./" + binary_name
- FNULL = open(os.devnull, 'wb')
- p = subprocess.Popen([binary_name, str(num_runs)], stdout = FNULL)
- p.wait()
-
- f = open("run_accuracies.txt")
-
- total_err = 0.0
- for x in f:
- acc = float(x.strip())
- total_err += (gold_acc - acc)
-
- avg_err = total_err / num_runs
-
- return avg_err
-
-
-
-def roundDecimal(val):
-
- new_val = int(val * 10000)
- new_val = float(new_val) / 10000
-
- return new_val
-
-
-
-
-def test_sensitivity(Bench):
-
- tensor_errors = []
-
- error_levels = [6, 9, 12, 15]
- num_flags = Bench.num_flags
-
- for tensor_id in range(num_flags):
- total_error = 0
- for error_level in error_levels:
- constructTunerFile(num_flags, tensor_id, error_level, 0)
- error = runAndTestError(Bench.tuner_binary, Bench.tuner_accuracy)
- #print (tensor_id, error_level, error)
- total_error += error
-
- avg_error = total_error / len(error_levels)
-
- tensor_errors.append([tensor_id, avg_error])
-
-
- print ("\n\n*** Per-Tensor Avg Errors \n\n")
-
- f_name = Bench.base_dir + "/tensor_errors_multiple.txt"
- f = open(f_name, "w+")
- for i in range(len(tensor_errors)):
- print (i, tensor_errors[i][1])
- f.write(str(i) + "\t" + str(tensor_errors[i][1]) + "\n")
-
- f.close()
-
- f_name = Bench.base_dir + "/tensor_errors_ranked_1000.txt"
- f2 = open(f_name, "w+")
- tensor_errors.sort(key=operator.itemgetter(1))
-
-
- for i in range(len(tensor_errors)):
- print (i, tensor_errors[i][1])
-
- f2.write(str(tensor_errors[i][0]) + "\t" + str(tensor_errors[i][1]) + "\n")
-
-
- f2.close()
-
-
-
-def test_sensitivity2(Bench):
-
- num_flags = Bench.num_flags
-
- constructTunerFile(num_flags, 0, 6, 6)
- error = runAndTestError(Bench.tuner_binary, Bench.tuner_accuracy)
-
- ref_acc = Bench.tuner_accuracy - error
- print ("*** Gold accuracy = ", Bench.tuner_accuracy, " Ref accuracy = ", ref_acc, " *** \n\n")
-
-
- tensor_errors = []
-
- error_levels = [6, 9, 12, 15]
-
- for tensor_id in range(num_flags):
- total_error = 0
- for error_level in error_levels:
- constructTunerFile(num_flags, tensor_id, error_level, 6)
- error = runAndTestError(Bench.tuner_binary, ref_acc)
- print (tensor_id, error_level, error)
- total_error += error
-
- avg_error = total_error / len(error_levels)
-
- tensor_errors.append([tensor_id, avg_error])
-
-
- print ("\n\n*** Per-Tensor Avg Errors \n\n")
-
- f_name = Bench.base_dir + "/tensor_composite_errors.txt"
- f = open(f_name, "w+")
- for i in range(len(tensor_errors)):
- print (i, tensor_errors[i][1])
- f.write(str(i) + "\t" + str(tensor_errors[i][1]) + "\n")
-
- f.close()
-
- f_name = Bench.base_dir + "/tensor_composite_errors_ranked.txt"
- f2 = open(f_name, "w+")
- tensor_errors.sort(key=operator.itemgetter(1))
-
-
- for i in range(len(tensor_errors)):
- print (i, tensor_errors[i][1])
-
- f2.write(str(tensor_errors[i][0]) + "\t" + str(tensor_errors[i][1]) + "\n")
-
-
- f2.close()
-
-
-
-def test_sensitivity3(Bench):
-
- tensor_errors = []
-
- error_levels = [2, 5, 8, 11, 14, 17]
- num_flags = Bench.num_flags
-
- for tensor_id in range(num_flags):
- total_error = 0
- errors = []
- for error_level in error_levels:
- constructTunerFile(num_flags, tensor_id, error_level, 0)
- error = runAndTestError(Bench.tuner_binary, Bench.tuner_accuracy)
- print (tensor_id, error_level, error)
- errors.append(error)
-
- tensor_errors.append([tensor_id, errors])
-
-
- print ("\n\n*** Per-Tensor Avg Errors \n\n")
-
- f_name = Bench.base_dir + "/tensor_errors_multiple.txt"
- f = open(f_name, "w+")
- for i in range(len(tensor_errors)):
- print (i, tensor_errors[i][1])
- f.write(str(i))
- for j in range(len(tensor_errors[i][1])):
- val = roundDecimal(tensor_errors[i][1][j])
- f.write("\t" + str(val) )
- f.write("\n")
-
- f.close()
-
-
-
-
-
-def test_sensitivity4(Bench):
-
- num_flags = Bench.num_flags
-
- constructTunerFile(num_flags, 0, 5, 5)
- error = runAndTestError(Bench.tuner_binary, Bench.tuner_accuracy)
-
- ref_acc = Bench.tuner_accuracy - error
- print ("*** Gold accuracy = ", Bench.tuner_accuracy, " Ref accuracy = ", ref_acc, " *** \n\n")
-
-
- tensor_errors = []
- error_levels = [4, 8, 11, 14, 16, 19]
-
- for tensor_id in range(num_flags):
- errors = []
- for error_level in error_levels:
- constructTunerFile(num_flags, tensor_id, error_level, 5)
- error = runAndTestError(Bench.tuner_binary, ref_acc)
- print (tensor_id, error_level, error)
- errors.append(error)
-
- tensor_errors.append([tensor_id, errors])
-
-
-
- print ("\n\n*** Per-Tensor Avg Errors \n\n")
-
- f_name = Bench.base_dir + "/composite_errors.txt"
- f = open(f_name, "w+")
- for i in range(len(tensor_errors)):
- print (i, tensor_errors[i][1])
- f.write(str(i))
- for j in range(len(tensor_errors[i][1])):
- val = roundDecimal(tensor_errors[i][1][j])
- f.write("\t" + str(val) )
- f.write("\n")
-
- f.close()
-
-
-
-
-
-def readTensorErrs(result_dir):
-
- tensor_errs = []
- f = open(result_dir + "/tensor_errors.txt")
-
- for x in f:
- err = float(x.split()[1])
- tensor_errs.append(err)
-
- return tensor_errs
-
-
-
-def readTensorErrs2(result_dir):
-
- tensor_errs = []
- f = open(result_dir + "/tensor_errors_multiple.txt")
-
- for x in f:
- toks = x.split()
- total_err = 0.0
- for tok in toks[2:-1]:
- err = float(tok)
- total_err += err
-
- avg_err = total_err / len(toks[2:-1])
- tensor_errs.append(avg_err)
-
- return tensor_errs
-
-
-def isSkipLayer(layer):
-
- if "dense" in layer or "conv" in layer:
- return False
- else:
- return True
-
-
-def readLayerCosts(cost_file):
-
- f = open(cost_file)
- layer_costs = []
- for x in f:
- cost = float(x.strip())
- layer_costs.append(cost)
-
- return layer_costs
-
-
-
-disable_skipping = False
-
-def select_skip_layers(Bench, percent_to_skip):
-
- if disable_skipping:
- return "0"
-
- result_dir = Bench.base_dir
- layer_file = Bench.layer_file
-
- tensor_errs = readTensorErrs2(result_dir)
- layer_costs = readLayerCosts(Bench.cost_file)
- layer_desc = loadLayerDesc(layer_file)
-
- it = 0
- index = 0
- layer_errs = []
- for layer in layer_desc:
- layer_len = len(layer)
- avg_err = tensor_errs[index]
- index += layer_len
-
- if isSkipLayer(layer):
- continue
-
- cost = (math.sqrt(layer_costs[it])) / 100;
- ERR_IMPACT = avg_err / cost
- #print ("layer, ", it, " avg_err = ", avg_err, " cost = ", cost, " err_impact = ", ERR_IMPACT)
-
- layer_errs.append((ERR_IMPACT, it))
- it += 1
-
- layer_errs.sort(key=operator.itemgetter(0), reverse=True)
-
- to_skip = len(layer_errs)
- to_skip = math.ceil((percent_to_skip / 100.0) * to_skip)
-
- skip_str = ""
- it = 0
- for err in layer_errs:
- if it >= to_skip:
- break
-
- skip_str += str(err[1])
- if it < to_skip - 1:
- skip_str += "_"
-
- it += 1
-
- return skip_str
-
-
-
-
-
-
-if __name__ == "__main__":
-
-
- AlexNet = bench_tuner_data["alexnet_cifar10"]
- skip_str = select_skip_layers(AlexNet, 10)
- print ("AlexNet skip_str = ", skip_str)
-
-
- AlexNet2 = bench_tuner_data["alexnet2_cifar10"]
- skip_str = select_skip_layers(AlexNet2, 15)
- print ("AlexNet2 skip_str = ", skip_str)
-
-
- VGG16 = bench_tuner_data["vgg16_cifar10"]
- skip_str = select_skip_layers(VGG16, 15)
- print ("VGG16 skip_str = ", skip_str)
-
-
- VGG16_100 = bench_tuner_data["vgg16_cifar100"]
- skip_str = select_skip_layers(VGG16_100, 15)
- print ("VGG16_100 skip_str = ", skip_str)
-
-
- ResNet = bench_tuner_data["resnet18_cifar10"]
- skip_str = select_skip_layers(ResNet, 10)
- print ("ResNet skip_str = ", skip_str)
-
-
- MobileNet = bench_tuner_data["mobilenet_cifar10"]
- skip_str = select_skip_layers(MobileNet, 15)
- print ("MobileNet skip_str = ", skip_str)
-
-
- MobileNet_SH = bench_tuner_data["mobilenet_shallow"]
- skip_str = select_skip_layers(MobileNet_SH, 15)
- print ("MobileNet_SH skip_str = ", skip_str)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/genPlots.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/genPlots.py
deleted file mode 100644
index df05ddc52b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/genPlots.py
+++ /dev/null
@@ -1,41 +0,0 @@
-
-
-import matplotlib.pyplot as plt
-import seaborn
-import numpy as np
-
-
-
-
-
-def genScatterPlot(accuracy_losses, speedups, file_path):
-
- plt.scatter(accuracy_losses, speedups)
- plt.xlabel("accuracy_loss")
- plt.ylabel("speedup")
- plt.savefig(file_path)
- plt.close()
-
-
-
-def genScatterPlotFromConfigs(configurations, file_path):
-
- accuracy_losses = []
- speedups = []
-
- for conf in configurations:
- accuracy_losses.append(conf.avg_loss)
- speedups.append(conf.speedup)
-
- genScatterPlot(accuracy_losses, speedups, file_path)
-
-
-
-if __name__ == "__main__":
-
- x = np.array([1, 2, 3])
- y = np.array([1, 2, 3])
-
- print ("type = ", type(plt.scatter(x, y)))
-
- plt.savefig("output.png")
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/global_paths.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/global_paths.py
deleted file mode 100644
index d93e96ef7c..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/global_paths.py
+++ /dev/null
@@ -1,12 +0,0 @@
-
-import os
-import sys
-
-
-
-if "LLVM_SRC_ROOT" not in os.environ:
- print ("ERROR: LLVM_SRC_ROOT not set --- set $LLVM_SRC_ROOT to top of LLVM source tree ")
- sys.exit(-1)
-
-opentuner_src_dir = os.environ["LLVM_SRC_ROOT"] + "/projects/hpvm-tensor-rt/autotuner/opentuner/autotuner/"
-tensorRT_dir = os.environ["LLVM_SRC_ROOT"] + "/projects/hpvm-tensor-rt/"
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/knob_pruning.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/knob_pruning.py
deleted file mode 100644
index dfcab4f36b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/knob_pruning.py
+++ /dev/null
@@ -1,86 +0,0 @@
-
-
-import utils
-import subprocess
-
-
-def createPromiseFile(l_knob, layer_ind, num_layers):
-
- f = open("promise_flags", "w+")
-
- for i in range(num_layers):
- if i == layer_ind:
- f.write(str(l_knob) + "\n")
- else:
- f.write("11\n")
-
- f.close()
-
-
-
-def runBinary(binary_path):
-
- run_cmd = "./" + binary_path
- print (run_cmd)
-
- p = subprocess.Popen(run_cmd, shell=True)
- p.wait()
-
- return utils.readAccuracy("final_accuracy")
-
-
-
-
-
-def getPrunedKnobs(binary_path, layer_file, global_knobs_file, \
- baseline_acc, acc_slack):
-
-
- knobs = utils.getInstallAndDevKnobs(layer_file, \
- global_knobs_file)
-
- pruned_knobs = []
- num_layers = len(knobs)
- layer_ind = 0
- for layer_knobs in knobs:
- pruned_layer_knobs = []
- for l_knob in layer_knobs:
- createPromiseFile(l_knob, layer_ind, num_layers)
- accuracy = runBinary(binary_path)
- acc_loss = baseline_acc - accuracy
- if acc_loss <= acc_slack:
- pruned_layer_knobs.append(l_knob)
- print ("\n + l_knob = ", l_knob, " - layer_ind = ", layer_ind)
- print ("- acc_loss = ", acc_loss, " **** SELECTED *** ")
- else:
- print ("\n -- l_knob = ", l_knob, " - layer_ind = ", layer_ind)
- print ("- acc_loss = ", acc_loss, " --- REJECTED --- ")
-
- pruned_knobs.append(pruned_layer_knobs)
-
- layer_ind += 1
-
-
- print ("*** knobs = ", knobs)
-
- return pruned_knobs
-
-
-
-if __name__ == "__main__":
-
-
- #pruned_knobs = getPrunedKnobs("alexnet2_promise", "../autotuner/data/alexnet2/alexnet2_layers.txt", \
- # "../autotuner/data/global_knobs.txt", 84.5, 3)
-
- pruned_knobs = getPrunedKnobs("lenet_promise", "../autotuner/data/lenet/lenet_layers.txt", \
- "../autotuner/data/global_knobs.txt", 99.7, 1)
-
-
-
- print ("*** pruned_knobs = ", pruned_knobs)
-
-
- utils.dumpKnobsFile(pruned_knobs, "install_knobs.txt")
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/main_driver.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/main_driver.py
deleted file mode 100644
index c4a5e0fac0..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/main_driver.py
+++ /dev/null
@@ -1,69 +0,0 @@
-
-import os
-import sys
-import subprocess
-import shutil
-
-
-from benchmarks import bench_tuner_data, batch_id
-from utils import createResultDirs
-from run_devtime_tuner import DevTimeTuner
-from run_install_tuner import InstallTimeTuner
-
-
-
-
-# Invoke Dev-time Autotuner
-def runDevTimeBenchs():
-
- Bench = bench_tuner_data["mobilenet_cifar10"]
- benchTuner = DevTimeTuner(Bench)
- benchTuner.runDevTuner()
-
- Bench = bench_tuner_data["resnet18_cifar10"]
- benchTuner = DevTimeTuner(Bench)
- benchTuner.runDevTuner()
-
- Bench = bench_tuner_data["alexnet_cifar10"]
- benchTuner = DevTimeTuner(Bench)
- benchTuner.runDevTuner()
-
-
- #for bench_id in bench_tuner_data:
- # Bench = bench_tuner_data[bench_id]
- # benchTuner = DevTimeTuner(Bench)
- # benchTuner.runDevTuner()
-
-
-
-
-
-
-# Invoke Dev-time Autotuner
-def runInstallTimeBenchs():
-
- Bench = bench_tuner_data["alexnet_cifar10"]
- benchTuner = InstallTimeTuner(Bench)
- benchTuner.runDevTuner()
-
-
- Bench = bench_tuner_data["alexnet2_cifar10"]
- benchTuner = InstallTimeTuner(Bench)
- benchTuner.runDevTuner()
-
-
-
-
-
-
-
-if __name__ == "__main__":
-
- createResultDirs(bench_tuner_data)
-
- #runDevTimeBenchs()
-
- runInstallTimeBenchs()
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/pareto_curve.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/pareto_curve.py
deleted file mode 100644
index d90403be23..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/pareto_curve.py
+++ /dev/null
@@ -1,313 +0,0 @@
-
-
-from buildRtConfig import loadConfigData
-from benchmarks import bench_tuner_data
-import os
-import shutil
-
-
-AL_THRESHOLD = 0.1
-SPEEDUP_BAND_SIZE = 0.1
-ENERGY_BAND_SIZE = 10
-
-class Configuration:
- def __init__(self, name, speedup, energy, accuracy, accuracy_loss, flags):
- self.name = name
- self.speedup = speedup
- self.energy = energy
- self.accuracy = accuracy
- self.accuracy_loss = accuracy_loss
- self.flags = flags
- def __repr__(self):
- return repr((self.name, self.speedup, self.energy, self.accuracy, self.accuracy_loss))
-
-configuration_objects = [
- Configuration('conf1', 1.05, 15, 85, 1.2, []),
- Configuration('conf2', 2.51, 12, 83, 1.4, []),
- Configuration('conf3', 2.05, 10, 84, 0.8, []),
-]
-
-def compute_pareto_points(configurations):
- speedupconfigurations = []
- energyconfigurations = []
- #sort configurations based on speedup
- sorted_configurations = sorted(configurations, key=lambda conf: conf.accuracy_loss)
-
- start_idx = 0
- while start_idx < len(sorted_configurations):
- end_idx = start_idx + 1;
- # find end_idx
- while end_idx < len(sorted_configurations) and (sorted_configurations[end_idx].accuracy_loss - sorted_configurations[start_idx].accuracy_loss < AL_THRESHOLD) :
- end_idx += 1
- # find best speedup end energy in this accuracy loss level
- sp = -1.0
- sp_idx = 0
- en = -1.0
- en_idx = 0
- for i in range(start_idx, end_idx):
- if sorted_configurations[i].speedup > sp:
- sp = sorted_configurations[i].speedup
- sp_idx = i
- if sorted_configurations[i].energy > en:
- en = sorted_configurations[i].energy
- en_idx = i
- sp_not_dominated = True
- # if not empty list of configurations
- if speedupconfigurations:
- if speedupconfigurations[-1].speedup >= sp:
- sp_not_dominated = False
- en_not_dominated = True
- # if not empty list of configurations
- if energyconfigurations:
- if energyconfigurations[-1].energy >= en:
- en_not_dominated = False
- if sp_not_dominated:
- speedupconfigurations.append(sorted_configurations[sp_idx])
- if en_not_dominated:
- energyconfigurations.append(sorted_configurations[en_idx])
- # outer while loop variable increment
- start_idx = end_idx
- return [speedupconfigurations, energyconfigurations]
-
-
-def compute_pareto_points_with_margin(configurations, speedup_band_width, energy_band_width):
- speedupconfigurations = []
- energyconfigurations = []
- #sort configurations based on speedup
- sorted_configurations = sorted(configurations, key=lambda conf: conf.accuracy_loss)
-
- idx_to_sp_conf_dict = {}
- idx_to_en_conf_dict = {}
-
- start_idx = 0
- while start_idx < len(sorted_configurations):
- end_idx = start_idx + 1;
- # find end_idx
- while end_idx < len(sorted_configurations) and (sorted_configurations[end_idx].accuracy_loss - sorted_configurations[start_idx].accuracy_loss < AL_THRESHOLD) :
- end_idx += 1
- # find best speedup end energy in this accuracy loss level
- sp = -1.0
- sp_idx = 0
- en = -1.0
- en_idx = 0
- for i in range(start_idx, end_idx):
- if sorted_configurations[i].speedup > sp:
- sp = sorted_configurations[i].speedup
- sp_idx = i
- if sorted_configurations[i].energy < en:
- en = sorted_configurations[i].energy
- en_idx = i
- sp_not_dominated = True
- # if not empty list of configurations
- if speedupconfigurations:
- if speedupconfigurations[-1].speedup >= sp:
- sp_not_dominated = False
- en_not_dominated = True
- # if not empty list of configurations
- if energyconfigurations:
- if energyconfigurations[-1].energy >= en:
- en_not_dominated = False
- if sp_not_dominated:
- speedupconfigurations.append(sorted_configurations[sp_idx])
- idx_to_sp_conf_dict[start_idx] = len(speedupconfigurations)-1
- if en_not_dominated:
- energyconfigurations.append(sorted_configurations[en_idx])
- idx_to_en_conf_dict[start_idx] = len(energyconfigurations)-1
- # outer while loop variable increment
- start_idx = end_idx
-
- # We want to add configurations in a band of a certain width around the curves
- # not possible to do during contruction, because the quality of the curve would
- # deteriorate quickly
-
- AdjustedSpeedupCurve = []
- AdjustedEnergyCurve = []
-
- start_idx = 0
- while start_idx < len(sorted_configurations):
- end_idx = start_idx + 1;
- # find end_idx
- while end_idx < len(sorted_configurations) and (sorted_configurations[end_idx].accuracy_loss - sorted_configurations[start_idx].accuracy_loss < AL_THRESHOLD) :
- end_idx += 1
- for i in range(start_idx, end_idx):
- if sorted_configurations[i].speedup + speedup_band_width >= speedupconfigurations[idx_to_sp_conf_dict[start_idx]].speedup:
- AdjustedSpeedupCurve.append(sorted_configurations[i])
- if sorted_configurations[i].energy + energy_band_width >= energyconfigurations[idx_to_en_conf_dict[start_idx]].energy:
- AdjustedEnergyCurve.append(sorted_configurations[i])
- # outer while loop variable increment
- start_idx = end_idx
-
- return [AdjustedSpeedupCurve, AdjustedEnergyCurve]
-
-
-
-def findParetoConfigs(base_dir, accuracy):
-
- result_dir = base_dir + "/algo_tuner/pareto/"
- try:
- os.mkdir(result_dir)
- except:
- print ("could not create dir")
-
- input_dir = base_dir
- config_arr = loadConfigData(input_dir, accuracy, "high_confidence")
-
- config_list = []
- it = 0
- for config in config_arr:
- config = Configuration(config.fname , config.speedup, 100, config.avg_accuracy, config.avg_loss, config.flags)
- config_list.append(config)
-
-
- if (len(config_list) > 0):
- max_speedup = max(config.speedup for config in config_list)
- else:
- max_speedup = 1.5
-
- SPEEDUP_BAND_SIZE = 0.05 # max_speedup * 1.0 / 12 # 4 # 20% of the max speedup
- ENERGY_BAND_SIZE = 10
-
- print ("max_speedup = ", max_speedup, " BAND_SIZE = ", SPEEDUP_BAND_SIZE)
-
-
- print ("*SPEEDUP_BAND_SIZE = ", SPEEDUP_BAND_SIZE)
-
- ASC, AEC = compute_pareto_points_with_margin(config_list, SPEEDUP_BAND_SIZE, ENERGY_BAND_SIZE)
-
-
- #print (ASC)
- #print (config_list)
-
- if len(ASC) < 5:
- ASC = config_list
-
-
- if len(ASC) > 50:
- ASC, AEC = compute_pareto_points_with_margin(config_list, SPEEDUP_BAND_SIZE/4, ENERGY_BAND_SIZE)
-
-
- print ("len(config_list) = ", len(config_list))
- print ("len(ASC) = ", len(ASC))
-
- for conf in ASC:
- src_path = base_dir + "/algo_tuner/high_confidence/" + conf.name
- dst_path = base_dir + "/algo_tuner/pareto/" + conf.name
- shutil.copy(src_path, dst_path)
-
- return ASC
-
-
-
-def flagsPerLayer(ASC, num_layers):
-
- layer_flags = []
- for i in range(num_layers):
- layer_map = {}
- layer_flags.append(layer_map)
-
-
- for config in ASC:
- config_flags = config.flags
- for i in range(len(config_flags)):
- layer_flags[i][config_flags[i]] = 1
-
-
- print (layer_flags)
-
- return layer_flags
-
-
-
-
-
-
-def dumpBenchPareto(Bench):
-
- result_dir1 = Bench.result_dir_1
- result_dir2 = Bench.result_dir_2
- result_dir3 = Bench.result_dir_3
-
- acc_thresh = Bench.promise_accuracy
-
- ASC1 = findParetoConfigs(result_dir1, acc_thresh)
- ASC2 = findParetoConfigs(result_dir2, acc_thresh)
- ASC3 = findParetoConfigs(result_dir3, acc_thresh)
-
-
- flags1 = flagsPerLayer(ASC1, Bench.num_layers)
- flags2 = flagsPerLayer(ASC2, Bench.num_layers)
- flags3 = flagsPerLayer(ASC3, Bench.num_layers)
-
- return flags1, flags2, flags3
-
-
-
-
-if __name__ == "__main__":
-
- Bench = bench_tuner_data["alexnet_cifar10"]
- dumpBenchPareto(Bench)
-
- Bench = bench_tuner_data["alexnet2_cifar10"]
- dumpBenchPareto(Bench)
-
- Bench = bench_tuner_data["vgg16_cifar10"]
- dumpBenchPareto(Bench)
-
- Bench = bench_tuner_data["vgg16_cifar100"]
- dumpBenchPareto(Bench)
-
- Bench = bench_tuner_data["resnet18_cifar10"]
- dumpBenchPareto(Bench)
-
- Bench = bench_tuner_data["lenet_keras"]
- dumpBenchPareto(Bench)
-
- Bench = bench_tuner_data["mobilenet_cifar10"]
- dumpBenchPareto(Bench)
-
- Bench = bench_tuner_data["mobilenet_shallow"]
- dumpBenchPareto(Bench)
-
-
- #get_pareto_configs("")
-
- #SC, EC = compute_pareto_points(configuration_objects)
- #ASC, AEC = compute_pareto_points_with_margin(configuration_objects, SPEEDUP_BAND_SIZE, ENERGY_BAND_SIZE)
-
- #print(SC)
- #print(EC)
-
- #print(ASC)
- #print(AEC)
-
-
-
-
-
-
- #result_dir = base_dir + "/pareto/"
- #try:
- # os.mkdir(result_dir)
- #except:
- # print "could not create dir"
-
- #input_dir = base_dir + "/full_results/"
- #result_dir = "../build_tuner/tuner_results/alexnet_cifar10/loss_3/batch15"
- #config_arr = loadConfigData(input_dir, accuracy)
-
- #config_list = []
-
- #it = 0
- #for config in config_arr:
- # config = Configuration(config.fname , config.speedup, 100, config.avg_accuracy, config.avg_loss)
- # config_list.append(config)
-
-
- #ASC, AEC = compute_pareto_points_with_margin(config_list, SPEEDUP_BAND_SIZE, ENERGY_BAND_SIZE)
-
- #for conf in ASC:
- # dst_path = conf.name.replace("full_results", "pareto")
- # shutil.copy(conf.name, dst_path)
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/pareto_utils.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/pareto_utils.py
deleted file mode 100644
index ae85160e8f..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/pareto_utils.py
+++ /dev/null
@@ -1,257 +0,0 @@
-
-
-from buildRtConfig import loadConfigsFromDir
-import os
-import shutil
-
-
-AL_THRESHOLD = 0.1
-SPEEDUP_BAND_SIZE = 0.1
-ENERGY_BAND_SIZE = 10
-
-
-class Configuration:
-
- def __init__(self, name, speedup, energy, accuracy, accuracy_loss, flags):
- self.name = name
- self.speedup = speedup
- self.energy = energy
- self.accuracy = accuracy
- self.avg_accuracy = accuracy
- self.accuracy_loss = accuracy_loss
- self.avg_loss = accuracy_loss
- self.flags = flags
-
- def __repr__(self):
- return repr((self.name, self.speedup, self.energy, self.accuracy, self.accuracy_loss))
-
- @staticmethod
- def speedup_points(configurations):
-
- return [
- (conf.speedup, conf.accuracy)
- for conf in configurations
- ]
-
-
-configuration_objects = [
- Configuration('conf1', 1.05, 15, 85, 1.2, []),
- Configuration('conf2', 2.51, 12, 83, 1.4, []),
- Configuration('conf3', 2.05, 10, 84, 0.8, []),
-]
-
-
-
-def is_pareto_efficient(configs, values, value_margins):
- import numpy as np
- from pprint import pprint
-
- np_values = np.array(values)
- np_margins = np.array(value_margins)
- is_efficient = np.ones(np_values.shape[0], dtype=bool)
-
- for i, c in enumerate(np_values):
- if is_efficient[i]:
- # Keep any point with a higher value
- is_efficient[is_efficient] = np.any(np_values[is_efficient] + np_margins >= c, axis=1)
- is_efficient[i] = True # And keep self
-
- return (np.array(configs)[is_efficient]).tolist()
-
-
-
-
-
-def compute_pareto_points(configurations):
- speedupconfigurations = []
- energyconfigurations = []
- #sort configurations based on speedup
- sorted_configurations = sorted(configurations, key=lambda conf: conf.accuracy_loss)
-
- start_idx = 0
- while start_idx < len(sorted_configurations):
- end_idx = start_idx + 1;
- # find end_idx
- while end_idx < len(sorted_configurations) and \
- (sorted_configurations[end_idx].accuracy_loss - \
- sorted_configurations[start_idx].accuracy_loss < AL_THRESHOLD) :
- end_idx += 1
-
- # find best speedup end energy in this accuracy loss level
- sp = -1.0
- sp_idx = 0
- en = -1.0
- en_idx = 0
- for i in range(start_idx, end_idx):
- if sorted_configurations[i].speedup > sp:
- sp = sorted_configurations[i].speedup
- sp_idx = i
- if sorted_configurations[i].energy > en:
- en = sorted_configurations[i].energy
- en_idx = i
- sp_not_dominated = True
- # if not empty list of configurations
- if speedupconfigurations:
- if speedupconfigurations[-1].speedup >= sp:
- sp_not_dominated = False
- en_not_dominated = True
- # if not empty list of configurations
- if energyconfigurations:
- if energyconfigurations[-1].energy >= en:
- en_not_dominated = False
- if sp_not_dominated:
- speedupconfigurations.append(sorted_configurations[sp_idx])
- if en_not_dominated:
- energyconfigurations.append(sorted_configurations[en_idx])
- # outer while loop variable increment
- start_idx = end_idx
- return [speedupconfigurations, energyconfigurations]
-
-
-
-
-def compute_pareto_points_with_margin(configurations, speedup_band_width, energy_band_width):
-
- speedupconfigurations = []
- energyconfigurations = []
- #sort configurations based on speedup
- sorted_configurations = sorted(configurations, key=lambda conf: conf.accuracy_loss)
-
- idx_to_sp_conf_dict = {}
- idx_to_en_conf_dict = {}
-
- start_idx = 0
- while start_idx < len(sorted_configurations):
- end_idx = start_idx + 1;
- # find end_idx
- while end_idx < len(sorted_configurations) and \
- (sorted_configurations[end_idx].accuracy_loss - \
- sorted_configurations[start_idx].accuracy_loss < AL_THRESHOLD) :
- end_idx += 1
-
- # find best speedup end energy in this accuracy loss level
- sp = -1.0
- sp_idx = 0
- en = -1.0
- en_idx = 0
- for i in range(start_idx, end_idx):
- if sorted_configurations[i].speedup > sp:
- sp = sorted_configurations[i].speedup
- sp_idx = i
- if sorted_configurations[i].energy < en:
- en = sorted_configurations[i].energy
- en_idx = i
- sp_not_dominated = True
- # if not empty list of configurations
- if speedupconfigurations:
- if speedupconfigurations[-1].speedup >= sp:
- sp_not_dominated = False
- en_not_dominated = True
- # if not empty list of configurations
- if energyconfigurations:
- if energyconfigurations[-1].energy >= en:
- en_not_dominated = False
- if sp_not_dominated:
- speedupconfigurations.append(sorted_configurations[sp_idx])
- idx_to_sp_conf_dict[start_idx] = len(speedupconfigurations)-1
- if en_not_dominated:
- energyconfigurations.append(sorted_configurations[en_idx])
- idx_to_en_conf_dict[start_idx] = len(energyconfigurations)-1
- # outer while loop variable increment
- start_idx = end_idx
-
- # We want to add configurations in a band of a certain width around the curves
- # not possible to do during contruction, because the quality of the curve would
- # deteriorate quickly
-
- AdjustedSpeedupCurve = []
- AdjustedEnergyCurve = []
-
- start_idx = 0
- while start_idx < len(sorted_configurations):
- end_idx = start_idx + 1;
- # find end_idx
- while end_idx < len(sorted_configurations) and \
- (sorted_configurations[end_idx].accuracy_loss - \
- sorted_configurations[start_idx].accuracy_loss < AL_THRESHOLD) :
- end_idx += 1
-
- for i in range(start_idx, end_idx):
- if sorted_configurations[i].speedup + speedup_band_width >= \
- speedupconfigurations[idx_to_sp_conf_dict[start_idx]].speedup:
- AdjustedSpeedupCurve.append(sorted_configurations[i])
- if sorted_configurations[i].energy + energy_band_width >= \
- energyconfigurations[idx_to_en_conf_dict[start_idx]].energy:
- AdjustedEnergyCurve.append(sorted_configurations[i])
- # outer while loop variable increment
- start_idx = end_idx
-
- return [AdjustedSpeedupCurve, AdjustedEnergyCurve]
-
-
-
-
-#***** Exported Routine *******/
-def dumpParetoConfigsToDir(input_dir, output_dir, gold_accuracy, enable_band):
-
- config_arr = loadConfigsFromDir(input_dir, gold_accuracy)
- config_list = []
- it = 0
-
- for config in config_arr:
- config = Configuration(config.fname , config.speedup, 100, \
- config.avg_accuracy, config.avg_loss, config.flags)
-
- config_list.append(config)
-
-
- if (len(config_list) > 0):
- max_speedup = max(config.speedup for config in config_list)
- else:
- max_speedup = 1.0 # No Speedup since no configuration found
-
-
- #SPEEDUP_BAND_SIZE = 0.05 # max_speedup * 1.0 / 12 # 4 # 20% of the max speedup
-
- if enable_band:
- SPEEDUP_BAND_SIZE = max_speedup * 1.0 / 10 # 10% of the max speedup is the 'BAND SIZE'
- ENERGY_BAND_SIZE = 0 # Unused right now
-
- ASC, AEC = compute_pareto_points_with_margin(config_list, SPEEDUP_BAND_SIZE, ENERGY_BAND_SIZE)
-
- else:
- SPEEDUP_BAND_SIZE = 0 # no pareto band - true pareto curve
- ENERGY_BAND_SIZE = 0 # Unused right now
-
- #ASC, AEC = compute_pareto_points(config_list)
-
- speedup_points = Configuration.speedup_points(config_list)
- ASC = is_pareto_efficient(config_list, speedup_points, [-0.001, -0.001]) # [0.05, 0.05])
-
-
-
- print ("*max_speedup = ", max_speedup)
- print ("*SPEEDUP_BAND_SIZE = ", SPEEDUP_BAND_SIZE)
-
-
- # Prevents very small pareto-curves
- #if len(ASC) < 10 or len(config_list) < 20:
-
- #if len(config_list) < 10:
- # ASC = config_list
-
-
- print ("len(config_list) = ", len(config_list))
- print ("len(ASC) = ", len(ASC))
-
- for conf in ASC:
- src_path = input_dir + '/' + conf.name
- dst_path = output_dir + '/' + conf.name
- shutil.copy(src_path, dst_path)
-
- return ASC
-
-
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/profiling.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/profiling.py
deleted file mode 100644
index 3ed37822a6..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/profiling.py
+++ /dev/null
@@ -1,26 +0,0 @@
-
-import time
-
-profiled_ops = {}
-
-def startProfile(op_id):
- start = time.time()
- return start
-
-
-def stopProfile(op_id, start):
- end = time.time()
- total_time = end - start
-
- profiled_ops[op_id] = total_time
-
-
-def dumpProfiles(file_name):
-
- f = open(file_name, "w+")
- for op_id in profiled_ops:
- f.write(op_id + " : " + str(profiled_ops[op_id]) + "\n")
-
- f.close()
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/remap.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/remap.py
deleted file mode 100644
index 8dc6935752..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/remap.py
+++ /dev/null
@@ -1,291 +0,0 @@
-
-import sys
-import os
-import shutil
-from validation import invokeBinary
-from buildRtConfig import loadConfigData, loadPromiseConfigs
-from benchmarks import bench_tuner_data, batch_id
-from swing_selection import convL1bins, convL2bins
-
-
-
-def readKnobConfig(file_path):
-
- knobs_speedup = {}
- f = open(file_path, "r")
- for x in f:
- toks = x.split("\t")
- ID = int(toks[0].split(",")[1])
-
- speedup = float(toks[2])
- knobs_speedup[ID] = speedup
-
- print ("knobs_speedup = ", knobs_speedup)
-
- return knobs_speedup
-
-
-
-def getPromiseSwing(l1, l2, flag):
-
- if l1 < 0.1 or l2 < 0.1:
- return flag
-
- swing = 1
- for i in range(len(convL1bins)):
- l1_t = convL1bins[i][0]
- l2_t = convL2bins[i][0]
-
- if l1 > l1_t and l2 > l2_t:
- break
- swing += 1
-
- return swing
-
-
-
-def replaceWithPromise(layer_flags, norms_file):
-
- num_layers = len(layer_flags)
-
- f = open(norms_file, "r")
- it = 0
- for x in f:
- op_name = x.split()[0]
- print ("op_name = ", op_name)
- if op_name == "tensorMul":
- break;
-
- l1 = float(x.split()[5])
- l2 = float(x.split()[6])
-
- if it > 0:
- flag = getPromiseSwing(l1, l2, layer_flags[it])
- layer_flags[it] = flag
-
- #print ("l1 = ", l1, " l2 = ", l2)
- it += 1
-
- if it == num_layers:
- break
-
- print (layer_flags)
- return layer_flags
-
-
-
-
-def readCostFile(file_path):
-
- layer_costs = []
- f = open(file_path)
- for x in f:
- cost = float(x.strip())
- layer_costs.append(cost)
-
- print ("len(layer_costs) = ", layer_costs)
- f.close()
-
- return layer_costs
-
-
-
-def getSpeedup(flags, knobs_speedup, layer_costs):
-
- orig_cost = 0.0
- total_cost = 0.0
- it = 0
- for flag_value in flags:
- op_cost = layer_costs[it]
- speedup = knobs_speedup[flag_value]
-
- total_cost += (op_cost * 1.0 / speedup * 1.0)
- orig_cost += op_cost
- it += 1
-
- speedup = (orig_cost * 1.0) / (total_cost * 1.0)
-
- return speedup
-
-
-
-def dumpNewFlags(new_flags, orig_file, promise_flags_file, layer_costs, knobs_speedup):
-
- speedup = getSpeedup(new_flags, knobs_speedup, layer_costs)
-
- top_line = ""
- for x in open(orig_file, "r"):
- top_line = x
- break
-
- f = open(promise_flags_file, "w+")
- f.write(top_line.replace("\n", ""))
- f.write("\tnew_speedup=" + str(speedup) + "\n")
-
-
- for flag in new_flags:
- f.write(str(flag) + "\n")
-
- f.close()
-
-
-
-
-def remapLossConfig(configs_arr, result_dir, sub_dir, layer_costs, knobs_speedup):
-
-
- for conf in configs_arr:
- layer_flags = conf.flags
- fname = conf.fname
- norms_file = result_dir + "/algo_tuner/" + sub_dir + "/" + fname + "_norms"
- orig_file = result_dir + "/algo_tuner/" + sub_dir + "/" + fname
- new_flags = replaceWithPromise(layer_flags, norms_file)
-
- promise_test_dir = result_dir + "/algo_tuner/promise_test/"
- if not os.path.exists(promise_test_dir):
- os.mkdir(promise_test_dir)
-
- promise_flags_file = result_dir + "/algo_tuner/promise_test/" + fname + "_promise"
- dumpNewFlags(new_flags, orig_file, promise_flags_file, layer_costs, knobs_speedup)
-
-
-
-def remapConfigs(Bench):
-
-
- loss1_dir = Bench.result_dir_1
- loss2_dir = Bench.result_dir_2
- loss3_dir = Bench.result_dir_3
-
- loss1_configs = loadConfigData(loss1_dir, 100, "validated")
- loss2_configs = loadConfigData(loss2_dir, 100, "validated")
- loss3_configs = loadConfigData(loss3_dir, 100, "validated")
-
- knobs_speedup = readKnobConfig("../opentuner/data/global_knobs.txt")
- layer_costs = readCostFile(Bench.cost_file)
-
- remapLossConfig(loss1_configs, loss1_dir, "validated", layer_costs, knobs_speedup)
- remapLossConfig(loss2_configs, loss2_dir, "validated", layer_costs, knobs_speedup)
- remapLossConfig(loss3_configs, loss3_dir, "validated", layer_costs, knobs_speedup)
-
-
-
-
-def validateRemapConfigs(Bench):
-
- num_layers = Bench.num_layers
- base_conf = getBaselineConfig(num_layers)
- # Path to binary to run
- binary_path = Bench.promise_binary
- # NOTE: 'target_acc' passed 0.0 since unused for baseline run
- invokeBinary(binary_path, base_conf, 1, 2000, 8000, 0.0)
- gold_acc = readAccuracy("final_accuracy")
-
-
- loss1_dir = Bench.result_dir_1
- loss2_dir = Bench.result_dir_2
- loss3_dir = Bench.result_dir_3
-
- loss1_configs = loadPromiseConfigs(loss1_dir, 100, "promise_test")
- loss2_configs = loadPromiseConfigs(loss2_dir, 100, "promise_test")
- loss3_configs = loadPromiseConfigs(loss3_dir, 100, "promise_test")
-
- runs = 30
- validateAlgoConfigs(binary_path, loss1_dir, loss1_configs, gold_acc, 1.0, runs)
- validateAlgoConfigs(binary_path, loss2_dir, loss2_configs, gold_acc, 2.0, runs)
- validateAlgoConfigs(binary_path, loss3_dir, loss3_configs, gold_acc, 3.0, runs)
-
-
-
-
-
-
-
-
-def copyNormFile(fname, result_dir, sub_dir):
-
- target_dir = result_dir + "/algo_tuner/" + sub_dir
- dest_file = target_dir + "/" + fname + "_norms"
-
- shutil.copy("accuracy_summary", dest_file)
-
-
-
-
-def dumpNorms(binary_path, result_dir, configs_arr):
-
- runs = 1
- for conf in configs_arr:
- layer_swings = conf.flags
- invokeBinary(binary_path, layer_swings, runs, 3000, 5000, 100)
-
- #copyNormFile(conf.fname, result_dir, "high_confidence")
- copyNormFile(conf.fname, result_dir, "validated")
-
-
-
-def computeConfigNorms(Bench):
-
- loss1_dir = Bench.result_dir_1
- loss2_dir = Bench.result_dir_2
- loss3_dir = Bench.result_dir_3
-
- loss1_configs = loadConfigData(loss1_dir, 100, "validated")
- loss2_configs = loadConfigData(loss2_dir, 100, "validated")
- loss3_configs = loadConfigData(loss3_dir, 100, "validated")
-
-
- binary_path = Bench.promise_binary
-
- dumpNorms(binary_path, loss1_dir, loss1_configs)
- dumpNorms(binary_path, loss2_dir, loss2_configs)
- dumpNorms(binary_path, loss3_dir, loss3_configs)
-
-
-
-if __name__ == "__main__":
-
- Bench = bench_tuner_data["alexnet_cifar10"]
- computeConfigNorms(Bench)
- remapConfigs(Bench)
- #validateRemapConfigs(Bench)
-
- Bench = bench_tuner_data["alexnet2_cifar10"]
- computeConfigNorms(Bench)
- remapConfigs(Bench)
- #validateRemapConfigs(Bench)
-
- Bench = bench_tuner_data["vgg16_cifar10"]
- computeConfigNorms(Bench)
- remapConfigs(Bench)
- #validateRemapConfigs(Bench)
-
- Bench = bench_tuner_data["vgg16_cifar100"]
- computeConfigNorms(Bench)
- remapConfigs(Bench)
- #validateRemapConfigs(Bench)
-
- Bench = bench_tuner_data["resnet18_cifar10"]
- computeConfigNorms(Bench)
- remapConfigs(Bench)
- #validateRemapConfigs(Bench)
-
- Bench = bench_tuner_data["mobilenet_shallow"]
- computeConfigNorms(Bench)
- remapConfigs(Bench)
- #validateRemapConfigs(Bench)
-
- Bench = bench_tuner_data["mobilenet_cifar10"]
- computeConfigNorms(Bench)
- remapConfigs(Bench)
- #validateRemapConfigs(Bench)
-
- Bench = bench_tuner_data["lenet_keras"]
- computeConfigNorms(Bench)
- remapConfigs(Bench)
- #validateRemapConfigs(Bench)
-
- #computeConfigNorms(Bench)
- #remapConfigs(Bench)
-
- #validateRemapConfigs(Bench)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_algo_tuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_algo_tuner.py
deleted file mode 100644
index 2df75fbfc4..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_algo_tuner.py
+++ /dev/null
@@ -1,102 +0,0 @@
-
-
-import os
-import subprocess
-from error_sensitivity import select_skip_layers
-
-
-def runAlgoTunerCmd(Bench, dir_prefix, result_dir, acc_threshold, autotuner_runs):
-
- tuner_cmd = "python2 ../opentuner/autotuner/algo_tuner.py "
- tuner_cmd += " --test-limit "
- tuner_cmd += str(autotuner_runs)
- tuner_cmd += " --binary ./"
- tuner_cmd += Bench.promise_binary
- tuner_cmd += " --num-layers "
- tuner_cmd += str(Bench.num_layers)
- tuner_cmd += " --result-dir "
- tuner_cmd += dir_prefix
- tuner_cmd += result_dir + "/algo_tuner/"
- tuner_cmd += " --accuracy "
- tuner_cmd += str(Bench.promise_accuracy - acc_threshold)
- tuner_cmd += " --cost-file "
- tuner_cmd += Bench.cost_file
- tuner_cmd += " --knobs-config "
- tuner_cmd += "../opentuner/data/global_knobs.txt"
- tuner_cmd += " --layer-knobs "
- tuner_cmd += Bench.layer_knobs
-
-
- print (tuner_cmd)
-
- p = subprocess.Popen(tuner_cmd, shell=True)
- p.wait()
-
-
-"""
-
-def promiseTunerLoss1(Bench, dir_prefix):
-
- tuner_runs = int(Bench.autotuner_runs / 3)
-
- skip_layers1 = "0"
- skip_layers2 = "0_" + select_skip_layers(Bench, 30)
- skip_layers3 = "0_" + select_skip_layers(Bench, 50)
-
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_1, 0.85, tuner_runs, skip_layers1)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_1, 0.85, tuner_runs, skip_layers2)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_1, 0.85, tuner_runs, skip_layers3)
-
-
-def promiseTunerLoss2(Bench, dir_prefix):
-
- tuner_runs = int(Bench.autotuner_runs / 3)
-
- skip_layers1 = "0"
- skip_layers2 = "0_" + select_skip_layers(Bench, 20)
- skip_layers3 = "0_" + select_skip_layers(Bench, 40)
-
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_2, 1.7, tuner_runs, skip_layers1)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_2, 1.7, tuner_runs, skip_layers2)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_2, 1.7, tuner_runs, skip_layers3)
-
-
-
-def promiseTunerLoss3(Bench, dir_prefix):
-
- tuner_runs = int (Bench.autotuner_runs / 3)
-
- skip_layers1 = "0"
- skip_layers2 = "0_" + select_skip_layers(Bench, 10)
- skip_layers3 = "0_" + select_skip_layers(Bench, 30)
-
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_3, 2.5, tuner_runs, skip_layers1)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_3, 2.5, tuner_runs, skip_layers2)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_3, 2.5, tuner_runs, skip_layers3)
-
-
-"""
-
-
-BASELINE = True
-
-
-def runAlgoBench(Bench):
-
- # NOTE-IMP: Changing current directory to one with promise binaries
- dir_prefix = "../build_tuner/"
-
-
- if BASELINE:
- tuner_runs = Bench.autotuner_runs
- runAlgoTunerCmd(Bench, dir_prefix, Bench.result_dir_1, 0.85, tuner_runs)
- runAlgoTunerCmd(Bench, dir_prefix, Bench.result_dir_2, 1.7, tuner_runs)
- runAlgoTunerCmd(Bench, dir_prefix, Bench.result_dir_3, 2.5, tuner_runs)
-
- else:
- promiseTunerLoss1(Bench, dir_prefix)
- promiseTunerLoss2(Bench, dir_prefix)
- promiseTunerLoss3(Bench, dir_prefix)
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_algo_tuner2.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_algo_tuner2.py
deleted file mode 100644
index 99867fade3..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_algo_tuner2.py
+++ /dev/null
@@ -1,186 +0,0 @@
-
-
-import os
-import numpy as np
-import subprocess
-from error_sensitivity import select_skip_layers
-from pareto_curve import dumpBenchPareto
-from remap import readCostFile
-
-
-def runAlgoTunerCmd(Bench, dir_prefix, result_dir, acc_threshold, autotuner_runs):
-
- fixed_runs = 100
-
- tuner_cmd = "python2 ../opentuner/autotuner/algo_tuner2.py "
- tuner_cmd += " --test-limit "
- tuner_cmd += str(fixed_runs)
- tuner_cmd += " --binary ./"
- tuner_cmd += Bench.promise_binary
- tuner_cmd += " --num-layers "
- tuner_cmd += str(Bench.num_layers)
- tuner_cmd += " --result-dir "
- tuner_cmd += dir_prefix
- tuner_cmd += result_dir + "/promise_tuner3/"
- tuner_cmd += " --accuracy "
- tuner_cmd += str(Bench.promise_accuracy - acc_threshold)
- tuner_cmd += " --cost-file "
- tuner_cmd += Bench.cost_file
- tuner_cmd += " --layer-file "
- tuner_cmd += Bench.layer_file
- tuner_cmd += " --knobs-config "
- tuner_cmd += "../opentuner/data/global_knobs.txt"
- tuner_cmd += " --layer-knobs "
- tuner_cmd += " local_knobs.txt "
-
-
- print (tuner_cmd)
-
- p = subprocess.Popen(tuner_cmd, shell=True)
- p.wait()
-
-
-
-def is50Knob(flag):
-
- flags50 = {}
- flags50[21] = 1
- flags50[22] = 1
- flags50[26] = 1
- flags50[27] = 1
- flags50[31] = 1
- flags50[32] = 1
-
- if flag in flags50:
- return True
- else:
- return False
-
-
-
-def is25Knob(flag):
-
- flags25 = {}
- flags25[23] = 1
- flags25[24] = 1
- flags25[25] = 1
- flags25[28] = 1
- flags25[29] = 1
- flags25[30] = 1
- flags25[33] = 1
- flags25[34] = 1
- flags25[35] = 1
- flags25[36] = 1
-
- if flag in flags25:
- return True
- else:
- return False
-
-
-
-def addPromiseFlags(flag_map):
-
- flags = []
-
- has_50_flag = False
- has_25_flag = False
-
- for flag in flag_map:
- if is50Knob(flag):
- has_50_flag = True
- if is25Knob(flag):
- has_25_flag = True
-
-
- if has_50_flag:
- flag_map[7] = 1
- flag_map[5] = 1
- flag_map[3] = 1
-
- if has_25_flag:
- flag_map[7] = 1
-
- return flag_map
-
-
-
-def addCostBasedFlags(flag_map, layer_costs, i):
-
- median = np.median(layer_costs)
- max_cost = np.max(layer_costs)
- sorted_vals = np.sort(layer_costs)
-
- print ("**** Median = ", median)
- print ("**** Max_cost = ", max_cost)
- print ("**** Sorted_vals = ", sorted_vals, "\n\n")
-
-
- if (layer_costs[i] > (median * 1.5)):
- flag_map[7] = 1
-
- if (layer_costs[i] > (median * 3)) or layer_costs[i] == max_cost:
- flag_map[7] = 1
- flag_map[5] = 1
- flag_map[3] = 1
-
-
- if (layer_costs[i] < (median / 10)):
- flag_map = {}
- flag_map[12] = 1
-
- return flag_map
-
-
-
-
-
-def constructKnobsFile(flags, layer_costs):
-
- f = open("local_knobs.txt", "w+")
- for i in range(len(flags)):
- flag_map = flags[i]
-
- if i > 0:
- flag_map = addPromiseFlags(flag_map)
- flag = addCostBasedFlags(flag_map, layer_costs, i)
-
- it = 0
- for flag in flag_map:
- f.write(str(flag))
- if it < len(flag_map) - 1:
- f.write(",")
- it += 1
-
- f.write("\n")
-
- f.close()
-
-
-
-
-
-def runPromiseAlgoBench(Bench):
-
- # NOTE-IMP: Changing current directory to one with promise binaries
- dir_prefix = "../build_tuner/"
-
-
- tuner_runs = Bench.autotuner_runs
-
- layer_costs = readCostFile(Bench.cost_file)
-
- flags1, flags2, flags3 = dumpBenchPareto(Bench)
-
- constructKnobsFile(flags1, layer_costs)
- runAlgoTunerCmd(Bench, dir_prefix, Bench.result_dir_1, 0.8, tuner_runs)
-
- constructKnobsFile(flags2, layer_costs)
- runAlgoTunerCmd(Bench, dir_prefix, Bench.result_dir_2, 1.6, tuner_runs)
-
- constructKnobsFile(flags3, layer_costs)
- runAlgoTunerCmd(Bench, dir_prefix, Bench.result_dir_3, 2.2, tuner_runs)
-
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_autotuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_autotuner.py
deleted file mode 100644
index 800bf926a5..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_autotuner.py
+++ /dev/null
@@ -1,403 +0,0 @@
-
-import os
-import sys
-import subprocess
-import shutil
-
-from swing_selection import loadLayerDesc
-from error_sensitivity import test_sensitivity, test_sensitivity2, test_sensitivity3, test_sensitivity4
-from benchmarks import bench_tuner_data, batch_id
-from run_psnr import runPSNRTuner
-from run_ha_tuner import runTunerBench
-from run_hs_tuner import runPromiseBench
-from run_algo_tuner import runAlgoBench
-from run_algo_tuner2 import runPromiseAlgoBench
-from compute_confs import computePSNRBenchSwings, computeBenchSwings
-from validation import runPromiseBenchValidation2, runBenchValidation, runAlgoBenchValidate
-from profiling import startProfile, stopProfile, dumpProfiles
-from utils import createResultDirs
-from benchmarks import batch_id
-from run_devtime_tuner import DevTimeTuner
-
-
-
-def runTunerValidation():
-
- runBenchValidation(bench_tuner_data["mobilenet_shallow"])
-
- #runBenchValidation("mobilenet_cifar10")
-
- #runBenchValidation("alexnet_cifar10")
- #runBenchValidation("vgg16_cifar10")
- #runBenchValidation("alexnet2_cifar10")
- #runBenchValidation("resnet18_cifar10")
- #runBenchValidation("vgg16_cifar100")
-
-
-def computeLayerSwings():
-
-
- computeBenchSwings(bench_tuner_data["mobilenet_shallow"])
-
- #computeBenchSwings("mobilenet_cifar10")
-
- #computeBenchSwings("mobilenet_cifar10")
-
- #computeBenchSwings("lenet_keras")
- #computeBenchSwings("alexnet_cifar10")
- #computeBenchSwings("alexnet2_cifar10")
- #computePSNRBenchSwings("pipeline_GEOM")
- #computePSNRBenchSwings("pipeline_GEMO")
- #computePSNRBenchSwings("pipeline_GEO")
- #computePSNRBenchSwings("pipeline_GSM")
- #computePSNRBenchSwings("pipeline_GSME")
-
-
-
-
-
-def runPromiseTuner():
-
-
- start = startProfile("MobileNet")
- runPromiseBench(bench_tuner_data["mobilenet_cifar10"])
- stopProfile("MobileNet", start)
-
- start = startProfile("Alexnet")
- runPromiseBench(bench_tuner_data["alexnet_cifar10"])
- stopProfile("Alexnet", start)
-
- start = startProfile("Alexnet2")
- runPromiseBench(bench_tuner_data["alexnet2_cifar10"])
- stopProfile("Alexnet2", start)
-
- start = startProfile("VGG16_10")
- runPromiseBench(bench_tuner_data["vgg16_cifar10"])
- stopProfile("VGG16_10", start)
-
- start = startProfile("VGG16_100")
- runPromiseBench(bench_tuner_data["vgg16_cifar100"])
- stopProfile("VGG16_100", start)
-
- start = startProfile("ResNet")
- runPromiseBench(bench_tuner_data["resnet18_cifar10"])
- stopProfile("ResNet", start)
-
- start = startProfile("MobileNet-SH")
- runPromiseBench(bench_tuner_data["mobilenet_shallow"])
- stopProfile("MobileNet-SH", start)
-
- start = startProfile("LeNet")
- runPromiseBench(bench_tuner_data["lenet_keras"])
- stopProfile("LeNet", start)
-
-
-
- #runPSNRPromiseBench("pipeline_GEOM")
- #runPSNRPromiseBench("pipeline_GEMO")
- #runPSNRPromiseBench("pipeline_GEO")
- #runPSNRPromiseBench("pipeline_GSM")
- #runPSNRPromiseBench("pipeline_GSME")
-
- dumpProfiles("time_profile" + batch_id + ".txt")
-
-
-
-
-def runPromiseValidation():
-
-
- start = startProfile("AlexNet")
- runPromiseBenchValidation2(bench_tuner_data["alexnet_cifar10"])
- stopProfile("AlexNet", start)
-
- start = startProfile("AlexNet2")
- runPromiseBenchValidation2(bench_tuner_data["alexnet2_cifar10"])
- stopProfile("AlexNet2", start)
-
- start = startProfile("VGG16_100")
- runPromiseBenchValidation2(bench_tuner_data["vgg16_cifar100"])
- stopProfile("VGG16_100", start)
-
- start = startProfile("VGG16_10")
- runPromiseBenchValidation2(bench_tuner_data["vgg16_cifar10"])
- stopProfile("VGG16_10", start)
- #runPromiseBenchValidation2(bench_tuner_data["lenet_keras"])
-
- start = startProfile("ResNet")
- runPromiseBenchValidation2(bench_tuner_data["resnet18_cifar10"])
- stopProfile("ResNet", start)
-
- start = startProfile("MobileNet_SH")
- runPromiseBenchValidation2(bench_tuner_data["mobilenet_shallow"])
- stopProfile("MobileNet_SH", start)
-
- start = startProfile("MobileNet")
- runPromiseBenchValidation2(bench_tuner_data["mobilenet_cifar10"])
- stopProfile("MobileNet", start)
-
-
- dumpProfiles("validation_prof" + batch_id + ".txt")
-
-
-
-
-def runAutotuner():
-
- runTunerBench(bench_tuner_data["alexnet_cifar10"])
- runTunerBench(bench_tuner_data["alexnet2_cifar10"])
-
- #runTunerBench("mobilenet_shallow")
- #runTunerBench("mobilenet_cifar10")
-
- #runTunerBench("lenet_keras")
- #runTunerBench("resnet18_cifar10")
- #runTunerBench("vgg16_cifar10")
-
- #runPSNRTuner("pipeline_GEOM")
- #runPSNRTuner("pipeline_GEMO")
- #runPSNRTuner("pipeline_GEO")
- #runPSNRTuner("pipeline_GSM")
- #runPSNRTuner("pipeline_GSME")
-
-
-
-
-def runSensAnalysis():
-
- start = startProfile("LeNet")
- test_sensitivity4(bench_tuner_data["lenet_keras"])
- stopProfile("LeNet", start)
-
- """
- start = startProfile("AlexNet")
- test_sensitivity4(bench_tuner_data["alexnet_cifar10"])
- stopProfile("AlexNet", start)
-
- start = startProfile("AlexNet2")
- test_sensitivity4(bench_tuner_data["alexnet2_cifar10"])
- stopProfile("AlexNet2", start)
-
- start = startProfile("ResNet")
- test_sensitivity4(bench_tuner_data["resnet18_cifar10"])
- stopProfile("ResNet", start)
-
- start = startProfile("MobileNet")
- test_sensitivity4(bench_tuner_data["mobilenet_cifar10"])
- stopProfile("MobileNet", start)
-
- start = startProfile("MobileNet_SH")
- test_sensitivity4(bench_tuner_data["mobilenet_shallow"])
- stopProfile("MobileNet_SH", start)
-
- start = startProfile("VGG_10")
- test_sensitivity4(bench_tuner_data["vgg16_cifar10"])
- stopProfile("VGG16_10", start)
-
- start = startProfile("VGG_100")
- test_sensitivity4(bench_tuner_data["vgg16_cifar100"])
- stopProfile("VGG16_100", start)
-
- dumpProfiles("sens_time_prof.txt")
-
- """
-
- start = startProfile("LeNet")
- test_sensitivity3(bench_tuner_data["lenet_keras"])
- stopProfile("LeNet", start)
-
- start = startProfile("AlexNet")
- test_sensitivity3(bench_tuner_data["alexnet_cifar10"])
- stopProfile("AlexNet", start)
-
- start = startProfile("AlexNet2")
- test_sensitivity3(bench_tuner_data["alexnet2_cifar10"])
- stopProfile("AlexNet2", start)
-
- start = startProfile("ResNet")
- test_sensitivity3(bench_tuner_data["resnet18_cifar10"])
- stopProfile("ResNet", start)
-
-
- start = startProfile("MobileNet")
- test_sensitivity3(bench_tuner_data["mobilenet_cifar10"])
- stopProfile("MobileNet", start)
-
- start = startProfile("MobileNet_SH")
- test_sensitivity3(bench_tuner_data["mobilenet_shallow"])
- stopProfile("MobileNet_SH", start)
-
- start = startProfile("VGG_10")
- test_sensitivity3(bench_tuner_data["vgg16_cifar10"])
- stopProfile("VGG16_10", start)
-
- start = startProfile("VGG_100")
- test_sensitivity3(bench_tuner_data["vgg16_cifar100"])
- stopProfile("VGG16_100", start)
-
- dumpProfiles("sens_time_prof.txt")
-
-
- """
- test_sensitivity2(bench_tuner_data["fc4"])
- test_sensitivity2(bench_tuner_data["lenet_keras"])
- test_sensitivity2(bench_tuner_data["mobilenet_cifar10"])
- test_sensitivity2(bench_tuner_data["mobilenet_shallow"])
- test_sensitivity2(bench_tuner_data["resnet18_cifar10"])
- test_sensitivity2(bench_tuner_data["alexnet_cifar10"])
- test_sensitivity2(bench_tuner_data["alexnet2_cifar10"])
- test_sensitivity2(bench_tuner_data["vgg16_cifar10"])
- test_sensitivity2(bench_tuner_data["vgg16_cifar100"])
-
-
- test_sensitivity(bench_tuner_data["fc4"])
- test_sensitivity(bench_tuner_data["lenet_keras"])
- test_sensitivity(bench_tuner_data["mobilenet_cifar10"])
- test_sensitivity(bench_tuner_data["mobilenet_shallow"])
- test_sensitivity(bench_tuner_data["resnet18_cifar10"])
- test_sensitivity(bench_tuner_data["alexnet_cifar10"])
- test_sensitivity(bench_tuner_data["alexnet2_cifar10"])
- test_sensitivity(bench_tuner_data["vgg16_cifar10"])
- test_sensitivity(bench_tuner_data["vgg16_cifar100"])
- """
-
-
-
-
-
-
-def runAlgoTuner():
-
- Bench = bench_tuner_data["alexnet_cifar10"]
- runAlgoBench(Bench)
-
- Bench = bench_tuner_data["mobilenet_shallow"]
- runAlgoBench(Bench)
-
- Bench = bench_tuner_data["mobilenet_cifar10"]
- runAlgoBench(Bench)
-
- Bench = bench_tuner_data["vgg16_cifar10"]
- runAlgoBench(Bench)
-
- Bench = bench_tuner_data["lenet_keras"]
- runAlgoBench(Bench)
-
- Bench = bench_tuner_data["alexnet2_cifar10"]
- runAlgoBench(Bench)
-
- Bench = bench_tuner_data["vgg16_cifar100"]
- runAlgoBench(Bench)
-
- Bench = bench_tuner_data["resnet18_cifar10"]
- runAlgoBench(Bench)
-
-
-
-
-
-def runPromiseAlgoTuner():
-
- Bench = bench_tuner_data["alexnet_cifar10"]
- runPromiseAlgoBench(Bench)
-
- Bench = bench_tuner_data["mobilenet_shallow"]
- runPromiseAlgoBench(Bench)
-
- Bench = bench_tuner_data["mobilenet_cifar10"]
- runPromiseAlgoBench(Bench)
-
- Bench = bench_tuner_data["vgg16_cifar10"]
- runPromiseAlgoBench(Bench)
-
- Bench = bench_tuner_data["lenet_keras"]
- runPromiseAlgoBench(Bench)
-
- Bench = bench_tuner_data["alexnet2_cifar10"]
- runPromiseAlgoBench(Bench)
-
- Bench = bench_tuner_data["vgg16_cifar100"]
- runPromiseAlgoBench(Bench)
-
- Bench = bench_tuner_data["resnet18_cifar10"]
- runPromiseAlgoBench(Bench)
-
-
-
-
-
-
-
-def runAlgoTunerValidation():
-
- Bench = bench_tuner_data["alexnet_cifar10"]
- runAlgoBenchValidate(Bench)
-
- Bench = bench_tuner_data["mobilenet_shallow"]
- runAlgoBenchValidate(Bench)
-
- Bench = bench_tuner_data["mobilenet_cifar10"]
- runAlgoBenchValidate(Bench)
-
- Bench = bench_tuner_data["vgg16_cifar10"]
- runAlgoBenchValidate(Bench)
-
- Bench = bench_tuner_data["lenet_keras"]
- runAlgoBenchValidate(Bench)
-
- Bench = bench_tuner_data["alexnet2_cifar10"]
- runAlgoBenchValidate(Bench)
-
- Bench = bench_tuner_data["vgg16_cifar100"]
- runAlgoBenchValidate(Bench)
-
- Bench = bench_tuner_data["resnet18_cifar10"]
- runAlgoBenchValidate(Bench)
-
-
-
-
-
-# Invoke Dev-time Autotuner
-def runDevTimeBenchs():
-
- Bench = bench_tuner_data["lenet_keras"]
-
- lenetTuner = DevTimeTuner(Bench)
- lenetTuner.runDevTuner()
-
-
-
-
-
-if __name__ == "__main__":
-
- createResultDirs(bench_tuner_data)
-
-
- #-- runAutotuner()
-
-
- #runTunerValidation()
-
- #computeLayerSwings()
-
- #runPromiseTuner()
-
- #runAlgoTuner()
-
-
- runDevTimeBenchs()
-
-
- #--- runPromiseAlgoTuner()
-
-
-
- #runAlgoTunerValidation()
-
- #runPromiseValidation()
-
- #runSensAnalysis()
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_devtime_tuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_devtime_tuner.py
deleted file mode 100644
index 0c701714f2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_devtime_tuner.py
+++ /dev/null
@@ -1,311 +0,0 @@
-
-
-import os
-import sys
-import subprocess
-import shutil
-import time
-from benchmarks import batch_id
-import utils
-import global_paths
-import pareto_utils
-import buildRtConfig
-import genPlots
-
-
-
-class DevTimeTuner:
-
- def __init__(self, Bench):
-
- self.piped_execution = True
- self.autotuner_runs = 8000
-
- self.promise_binary = Bench.promise_binary
-
- if self.piped_execution:
- self.binary_path = Bench.piped_binary
- else:
- self.binary_path = Bench.promise_binary
-
- self.num_layers = Bench.num_layers
- self.gold_accuracy = Bench.promise_accuracy
- self.cost_file = global_paths.tensorRT_dir + "/" + Bench.cost_file
- self.layer_file = global_paths.tensorRT_dir + "/" + Bench.layer_file
- #self.layer_knobs = global_paths.tensorRT_dir + "/" + Bench.layer_knobs
-
- global_knobs_file = global_paths.tensorRT_dir + "/autotuner/data/global_knobs.txt"
- buildRtConfig.initializeApproxMap(global_knobs_file) # Initialize knobs - configfile gen
- utils.createDevKnobs(self.layer_file, global_knobs_file, "dev_knobs.txt")
- self.layer_knobs = "dev_knobs.txt"
-
- self.result_dir = global_paths.tensorRT_dir + "/" + Bench.base_dir + \
- "/loss_123/" + batch_id + "/dev_tuner/"
-
- # NOTE: maintains total iterations completed - across multiple invocations
- self.iterations_completed = 0
- # Start time for timing autotuner runs
- self.start_time = 0
-
-
-
-
- def invokeDevTunerScript(self, accuracy_slack, \
- additional_error_slack, autotuner_runs):
-
- accuracy_threshold = self.gold_accuracy - accuracy_slack
- accuracy_additional_slack = self.gold_accuracy - additional_error_slack
-
- tuner_cmd = "python2 " + global_paths.opentuner_src_dir + "/devtuner.py "
- tuner_cmd += " --test-limit "
- tuner_cmd += str(self.autotuner_runs)
- tuner_cmd += " --binary ./"
- tuner_cmd += self.binary_path
- tuner_cmd += " --num-layers "
- tuner_cmd += str(self.num_layers)
- tuner_cmd += " --result-dir "
- tuner_cmd += self.result_dir
- tuner_cmd += " --accuracy "
- tuner_cmd += str(accuracy_threshold)
- tuner_cmd += " --accuracy-slack "
- tuner_cmd += str(accuracy_additional_slack)
- tuner_cmd += " --cost-file "
- tuner_cmd += self.cost_file
- tuner_cmd += " --knobs-config "
- tuner_cmd += global_paths.tensorRT_dir + "/autotuner/data/global_knobs.txt"
- ### tuner_cmd += "../autotuner/data/global_knobs.txt"
- tuner_cmd += " --layer-knobs "
- tuner_cmd += self.layer_knobs
- tuner_cmd += " --start-id "
- tuner_cmd += str(self.iterations_completed)
-
-
- print (tuner_cmd)
-
- p = subprocess.Popen(tuner_cmd, shell=True)
- p.wait()
-
-
- # Update iterations completed after each completed devtuner.py invocation with N iterations
- self.iterations_completed += self.autotuner_runs
-
-
-
- def checkExistingDir(self):
-
- files_dir = self.result_dir + "/high_confidence/"
- if os.path.exists(files_dir) and len(os.listdir(files_dir)) >= 1:
- print ("result_dir = ", files_dir, " has existing files \n ")
- user_str = input("Enter 'yes' to override - Enter 'no' to skip and exit \n ")
- if user_str != "yes":
- print ("\n\n NOTE:Exiting \n\n")
- sys.exit(0)
-
-
-
- def dumpBestConfig(self, configurations):
-
- best_conf_path = self.result_dir + "/best_config.txt"
- conf_file = open(best_conf_path, "w+")
-
- sorted_configurations = sorted(configurations, key=lambda conf: conf.speedup)
-
- if len(sorted_configurations) > 0:
- best_conf = sorted_configurations[-1]
-
- conf_file.write("speedup = " + str(best_conf.speedup) + \
- " avg_loss = " + str(best_conf.avg_loss) + "\n")
-
- for flag in best_conf.flags:
- conf_file.write(str(flag) + "\n")
-
- conf_file.close()
-
-
-
- def dumpAllConfigurations(self):
-
- input_dir = self.result_dir + "/high_confidence/"
-
- configurations = buildRtConfig.loadConfigsFromDir(input_dir, self.gold_accuracy)
- bench_layer_composition = utils.getLayerComposition(self.layer_file)
-
- sorted_configurations = sorted(configurations, key=lambda conf: conf.speedup)
-
- # Adding an extra loss to tuned configurations - adjusting for unseen data
- buildRtConfig.adjustConfigLosses(sorted_configurations)
-
-
- config_out_path = self.result_dir + "dev_gpu_all_config.txt"
-
- buildRtConfig.dumpDevConfigsToRTFile(sorted_configurations, \
- config_out_path, bench_layer_composition, \
- self.gold_accuracy, "gpu")
-
- config_out_path = self.result_dir + "dev_cpu_all_config.txt"
-
- buildRtConfig.dumpDevConfigsToRTFile(sorted_configurations, \
- config_out_path, bench_layer_composition, \
- self.gold_accuracy, "cpu")
-
-
- plot_file_path = self.result_dir + "dev_all_conf_plot.png"
- genPlots.genScatterPlotFromConfigs(sorted_configurations, plot_file_path)
-
-
- self.dumpBestConfig(sorted_configurations)
-
-
-
- def dumpBandPareto(self):
-
- input_dir = self.result_dir + "/high_confidence/"
- output_dir = self.result_dir + "/pareto/"
- utils.createDir(output_dir)
-
- configurations = pareto_utils.dumpParetoConfigsToDir(input_dir, \
- output_dir, self.gold_accuracy, True)
-
- config_out_path = self.result_dir + "dev_pareto_config.txt"
- bench_layer_composition = utils.getLayerComposition(self.layer_file)
-
- #-- sorted_configurations = sorted(configurations, key=lambda conf: conf.avg_loss)
- sorted_configurations = sorted(configurations, key=lambda conf: conf.speedup)
-
- # Adding an extra loss to tuned configurations - adjusting for unseen data
- buildRtConfig.adjustConfigLosses(sorted_configurations)
-
-
- config_out_path = self.result_dir + "dev_gpu_pareto_config.txt"
-
- buildRtConfig.dumpDevConfigsToRTFile(sorted_configurations, \
- config_out_path, bench_layer_composition, \
- self.gold_accuracy, "gpu")
-
- config_out_path = self.result_dir + "dev_cpu_pareto_config.txt"
-
- buildRtConfig.dumpDevConfigsToRTFile(sorted_configurations, \
- config_out_path, bench_layer_composition, \
- self.gold_accuracy, "cpu")
-
- plot_file_path = self.result_dir + "dev_pareto_plot.png"
- genPlots.genScatterPlotFromConfigs(sorted_configurations, plot_file_path)
-
-
-
- def dumpTruePareto(self):
-
- input_dir = self.result_dir + "/high_confidence/"
- output_dir = self.result_dir + "/true_pareto/"
- utils.createDir(output_dir)
-
- # NOTE: This is a true pareto curve construction
- configurations = pareto_utils.dumpParetoConfigsToDir(input_dir, \
- output_dir, self.gold_accuracy, False)
-
- config_out_path = self.result_dir + "true_pareto_config.txt"
- bench_layer_composition = utils.getLayerComposition(self.layer_file)
-
- sorted_configurations = sorted(configurations, key=lambda conf: conf.avg_loss)
-
- # Adding an extra loss to tuned configurations - adjusting for unseen data
- buildRtConfig.adjustConfigLosses(sorted_configurations)
-
-
- config_out_path = self.result_dir + "true_gpu_pareto_config.txt"
-
- buildRtConfig.dumpDevConfigsToRTFile(sorted_configurations, \
- config_out_path, bench_layer_composition, \
- self.gold_accuracy, "gpu")
-
- config_out_path = self.result_dir + "true_cpu_pareto_config.txt"
-
- buildRtConfig.dumpDevConfigsToRTFile(sorted_configurations, \
- config_out_path, bench_layer_composition, \
- self.gold_accuracy, "cpu")
-
-
- plot_file_path = self.result_dir + "true_pareto_plot.png"
- genPlots.genScatterPlotFromConfigs(sorted_configurations, plot_file_path)
-
-
-
-
- def dumpParetoFiles(self):
-
- self.dumpBandPareto()
- self.dumpTruePareto()
-
-
-
- def dumpReferenceFiles(self):
-
- ref_dir = self.result_dir + "/references/"
- utils.createDir(ref_dir)
-
- sources = {"run_devtime_tuner.py", "benchmarks.py", \
- "buildRtConfig.py", "global_paths.py"}
-
- for src in sources:
- src_path = global_paths.tensorRT_dir + "/autotuner/tuner_driver_src/" + src
- dst_path = ref_dir + "/" + src
- shutil.copy(src_path, dst_path)
-
-
- data_files = {self.cost_file, self.layer_file, self.layer_knobs}
-
- for datafile in data_files:
- shutil.copy(datafile, ref_dir)
-
-
-
- def setBaselineAccuracy(self):
-
- self.gold_accuracy = utils.getBaselineAccuracy(self.promise_binary, self.num_layers)
- print ("NOTE: Baseline Accuracy = ", self.gold_accuracy, "\n\n")
-
-
-
- def startTimer(self):
- self.start_time = time.time()
- print ("\n\n ---Starting DevTuner Timer ----- \n\n")
-
-
-
- def endTimer(self):
- end_time = time.time()
- total_tuning_time = end_time - self.start_time
- time_hrs = total_tuning_time * 1.0 / (60 * 60)
- print ("\n\n --- Time In Hours = ", time_hrs, " \n\n")
-
- time_file_path = self.result_dir + "tuning_time.txt"
-
- f = open(time_file_path, "w+")
- f.write("time_hrs = " + str(time_hrs) + "\n")
- f.close()
-
-
-
-
- def runDevTuner(self):
-
- #self.checkExistingDir()
-
-
- self.startTimer()
-
- self.setBaselineAccuracy()
-
- self.invokeDevTunerScript(0.8, 2.1, self.autotuner_runs)
- self.invokeDevTunerScript(1.5, 2.1, self.autotuner_runs)
- self.invokeDevTunerScript(2.1, 2.1, self.autotuner_runs)
-
- self.dumpParetoFiles()
- self.dumpAllConfigurations()
-
-
- # NOTE: dumping files for checking experimental parameters for each batch
- self.dumpReferenceFiles()
-
-
- self.endTimer()
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_ha_tuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_ha_tuner.py
deleted file mode 100644
index 055d2c4c1b..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_ha_tuner.py
+++ /dev/null
@@ -1,52 +0,0 @@
-
-
-
-import subprocess
-
-
-
-#, bench_name
-def runTunerBench(Bench):
-
- #Bench = bench_tuner_data[bench_name]
-
- #FIXIT: Replace approxhpvm_tuner2 with approxhpvm_tuner
- tuner_cmd = "python ../opentuner/autotuner/approxhpvm_tuner.py "
- tuner_cmd += " --test-limit "
- tuner_cmd += str(Bench.autotuner_runs)
- tuner_cmd += " --binary ./"
- tuner_cmd += Bench.tuner_binary
- tuner_cmd += " --num-flags "
- tuner_cmd += str(Bench.num_flags)
- tuner_cmd += " --error-range "
- tuner_cmd += str(Bench.error_range_2)
- tuner_cmd += " --result-dir "
- tuner_cmd += Bench.result_dir_2
- tuner_cmd += " --accuracy "
- tuner_cmd += str(Bench.tuner_accuracy - 1.70)
-
- print (tuner_cmd)
-
- p = subprocess.Popen(tuner_cmd, shell=True)
- p.wait()
-
-
- tuner_cmd = "python ../opentuner/autotuner/approxhpvm_tuner.py "
- tuner_cmd += " --test-limit "
- tuner_cmd += str(Bench.autotuner_runs)
- tuner_cmd += " --binary ./"
- tuner_cmd += Bench.tuner_binary
- tuner_cmd += " --num-flags "
- tuner_cmd += str(Bench.num_flags)
- tuner_cmd += " --error-range "
- tuner_cmd += str(Bench.error_range_1)
- tuner_cmd += " --result-dir "
- tuner_cmd += Bench.result_dir_1
- tuner_cmd += " --accuracy "
- tuner_cmd += str(Bench.tuner_accuracy - 0.85)
-
- print (tuner_cmd)
-
- p = subprocess.Popen(tuner_cmd, shell=True)
- p.wait()
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_hs_tuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_hs_tuner.py
deleted file mode 100644
index f1a9c8f417..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_hs_tuner.py
+++ /dev/null
@@ -1,185 +0,0 @@
-
-
-import os
-import subprocess
-from error_sensitivity import select_skip_layers
-
-
-def runPromiseTunerCmd(Bench, dir_prefix, result_dir, acc_threshold, autotuner_runs, skip_layers):
-
- tuner_cmd = "python2 ../opentuner/autotuner/promise_tuner3.py "
- tuner_cmd += " --test-limit "
- tuner_cmd += str(autotuner_runs)
- tuner_cmd += " --binary ./"
- tuner_cmd += Bench.promise_binary
- tuner_cmd += " --num-flags "
- tuner_cmd += str(Bench.num_layers)
- tuner_cmd += " --start-range "
- tuner_cmd += str(Bench.start_promise_range)
- tuner_cmd += " --error-range "
- #tuner_cmd += str(10)
- # NOTE: Increasing flags from ApproxTechiqueTuner
- tuner_cmd += str(12)
- tuner_cmd += " --result-dir "
- tuner_cmd += dir_prefix
- tuner_cmd += result_dir + "/promise_tuner/"
- tuner_cmd += " --accuracy "
- tuner_cmd += str(Bench.promise_accuracy - acc_threshold)
- tuner_cmd += " --layer-file "
- tuner_cmd += dir_prefix
- tuner_cmd += Bench.tensor_desc_file
- # NOTE: Cost file is new addition - ***NOT*** present in promisetuner1 and promisetuner2
- tuner_cmd += " --cost-file "
- tuner_cmd += Bench.cost_file
- #tuner_cmd += " --gpu-layers "
- #tuner_cmd += str(Bench.skip_layers)
- tuner_cmd += " --gpu-layers 0 "
- tuner_cmd += " --skip-layers \""
- #tuner_cmd += str(Bench.skip_layer_str) + "\""
- tuner_cmd += str(skip_layers) + "\""
-
- print (tuner_cmd)
-
- p = subprocess.Popen(tuner_cmd, shell=True)
- p.wait()
-
-
-
-
-def promiseTunerLoss1(Bench, dir_prefix):
-
- tuner_runs = Bench.autotuner_runs
-
- skip_layers1 = "0"
- skip_layers2 = "0_" + select_skip_layers(Bench, 30)
- skip_layers3 = "0_" + select_skip_layers(Bench, 50)
-
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_1, 0.85, tuner_runs, skip_layers1)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_1, 0.85, tuner_runs, skip_layers2)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_1, 0.85, tuner_runs, skip_layers3)
-
-
-def promiseTunerLoss2(Bench, dir_prefix):
-
- tuner_runs = Bench.autotuner_runs
-
- skip_layers1 = "0"
- skip_layers2 = "0_" + select_skip_layers(Bench, 20)
- skip_layers3 = "0_" + select_skip_layers(Bench, 40)
-
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_2, 1.7, tuner_runs, skip_layers1)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_2, 1.7, tuner_runs, skip_layers2)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_2, 1.7, tuner_runs, skip_layers3)
-
-
-
-def promiseTunerLoss3(Bench, dir_prefix):
-
- tuner_runs = Bench.autotuner_runs
-
- skip_layers1 = "0"
- skip_layers2 = "0_" + select_skip_layers(Bench, 10)
- skip_layers3 = "0_" + select_skip_layers(Bench, 30)
-
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_3, 2.5, tuner_runs, skip_layers1)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_3, 2.5, tuner_runs, skip_layers2)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_3, 2.5, tuner_runs, skip_layers3)
-
-
-BASELINE = True
-
-
-def runPromiseBench(Bench):
-
- # NOTE-IMP: Changing current directory to one with promise binaries
- dir_prefix = "../build_tuner/"
-
-
- if BASELINE:
- tuner_runs = Bench.autotuner_runs * 2
- skip_layers = "0"
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_1, 0.85, tuner_runs, skip_layers)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_2, 1.7, tuner_runs, skip_layers)
- runPromiseTunerCmd(Bench, dir_prefix, Bench.result_dir_3, 2.5, tuner_runs, skip_layers)
-
- else:
-
- promiseTunerLoss1(Bench, dir_prefix)
-
- promiseTunerLoss2(Bench, dir_prefix)
-
- promiseTunerLoss3(Bench, dir_prefix)
-
-
-
-
- """
- #tuner_cmd = "python ../opentuner/autotuner/promise_tuner2.py "
- tuner_cmd = "python ../opentuner/autotuner/promise_tuner3.py "
- tuner_cmd += " --test-limit "
- tuner_cmd += str(Bench.autotuner_runs)
- tuner_cmd += " --binary ./"
- tuner_cmd += Bench.promise_binary
- tuner_cmd += " --num-flags "
- tuner_cmd += str(Bench.num_layers)
- tuner_cmd += " --start-range "
- tuner_cmd += str(Bench.start_promise_range)
- tuner_cmd += " --error-range "
- #tuner_cmd += str(10)
- tuner_cmd += " --result-dir "
- tuner_cmd += result_dir_prefix
- tuner_cmd += Bench.result_dir_2 + "/promise_tuner/"
- tuner_cmd += " --accuracy "
- tuner_cmd += str(Bench.promise_accuracy - 1.90)
- tuner_cmd += " --layer-file "
- tuner_cmd += result_dir_prefix
- tuner_cmd += Bench.tensor_desc_file
- # NOTE: Cost file is new addition - ***NOT*** present in promisetuner1 and promisetuner2
- tuner_cmd += " --cost-file "
- tuner_cmd += Bench.cost_file
- #tuner_cmd += " --gpu-layers "
- #tuner_cmd += str(Bench.skip_layers)
- tuner_cmd += " --gpu-layers 0 "
- tuner_cmd += " --skip-layers \""
- tuner_cmd += str(Bench.skip_layer_str) + "\""
-
- print (tuner_cmd)
-
- p = subprocess.Popen(tuner_cmd, shell=True)
- p.wait()
-
- #tuner_cmd = "python ../opentuner/autotuner/promise_tuner2.py "
- tuner_cmd = "python ../opentuner/autotuner/promise_tuner3.py "
- tuner_cmd += " --test-limit "
- tuner_cmd += str(Bench.autotuner_runs)
- tuner_cmd += " --binary ./"
- tuner_cmd += Bench.promise_binary
- tuner_cmd += " --num-flags "
- tuner_cmd += str(Bench.num_layers)
- tuner_cmd += " --start-range "
- tuner_cmd += str(Bench.start_promise_range)
- tuner_cmd += " --error-range "
- tuner_cmd += str(10)
- tuner_cmd += " --result-dir "
- tuner_cmd += result_dir_prefix
- tuner_cmd += Bench.result_dir_1 + "/promise_tuner/"
- tuner_cmd += " --accuracy "
- tuner_cmd += str(Bench.promise_accuracy - 0.95)
- tuner_cmd += " --layer-file "
- tuner_cmd += result_dir_prefix
- tuner_cmd += Bench.tensor_desc_file
- tuner_cmd += " --cost-file "
- tuner_cmd += Bench.cost_file
- #tuner_cmd += " --gpu-layers "
- #tuner_cmd += str(Bench.skip_layers)
- tuner_cmd += " --gpu-layers 0 "
- tuner_cmd += " --skip-layers \""
- tuner_cmd += str(Bench.skip_layer_str) + "\""
-
-
- print (tuner_cmd)
-
- p = subprocess.Popen(tuner_cmd, shell=True)
- p.wait()
-
- """
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_install_tuner.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_install_tuner.py
deleted file mode 100644
index 6fe682eb4e..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_install_tuner.py
+++ /dev/null
@@ -1,164 +0,0 @@
-
-
-import os
-import sys
-import subprocess
-import shutil
-import time
-from benchmarks import batch_id
-import utils
-import global_paths
-import pareto_utils
-import buildRtConfig
-import genPlots
-from run_devtime_tuner import DevTimeTuner
-import validation
-import knob_pruning
-
-
-class InstallTimeTuner(DevTimeTuner):
-
- def __init__(self, Bench):
-
- self.knob_pruning = True
- self.piped_execution = True
- self.autotuner_runs = 10000 # Bench.autotuner_runs
- self.validation_runs = 15
- self.abort_after = 3
- self.conf_threshold = 100
-
- self.promise_binary = Bench.promise_binary
-
- if self.piped_execution:
- self.binary_path = Bench.piped_binary
- else:
- self.binary_path = Bench.promise_binary
-
- self.num_layers = Bench.num_layers
- #self.gold_accuracy = Bench.promise_accuracy
- self.setBaselineAccuracy()
-
-
- self.cost_file = global_paths.tensorRT_dir + "/" + Bench.cost_file
- self.layer_file = global_paths.tensorRT_dir + "/" + Bench.layer_file
-
-
- global_knobs_file = global_paths.tensorRT_dir + "/autotuner/data/global_knobs.txt"
- buildRtConfig.initializeApproxMap(global_knobs_file) # Initialize knobs - configfile gen
-
-
- if self.knob_pruning == False:
- utils.createInstallAndDevKnobs(self.layer_file, global_knobs_file, "install_knobs.txt")
-
- elif self.knob_pruning == True:
- pruned_knobs = knob_pruning.getPrunedKnobs(self.promise_binary, self.layer_file, \
- global_knobs_file, self.gold_accuracy, 3)
-
- print ("*** pruned_knobs = ", pruned_knobs)
- utils.dumpKnobsFile(pruned_knobs, "install_knobs.txt")
-
-
-
- self.layer_knobs = "install_knobs.txt"
-
- self.result_dir = global_paths.tensorRT_dir + "/" + Bench.base_dir + \
- "/loss_123/" + batch_id + "/install_tuner/"
-
- # NOTE: maintains total iterations completed - across multiple invocations
- self.iterations_completed = 0
- # Start time for timing autotuner runs
- self.start_time = 0
-
-
-
- def validateAccuracyConfigs(self, configurations, accuracy_slack):
-
- filtered_configs = []
- for config in configurations:
- flags = config.flags
- avg_acc, confidence = validation.getStatisticalConfidence(self.promise_binary, flags, \
- self.gold_accuracy, \
- accuracy_slack, self.validation_runs, \
- self.abort_after)
-
- print ("avg_acc, confidence = ", avg_acc, confidence)
-
-
- if confidence >= self.conf_threshold:
- config.avg_accuracy = avg_acc
- filtered_configs.append(config)
-
- return filtered_configs
-
-
-
-
- def dumpValidatedConfigs(self, accuracy_slack):
-
- #input_dir = self.result_dir + "/high_confidence/"
- input_dir = self.result_dir + "/high_confidence/"
- output_dir = self.result_dir + "/pareto/"
- utils.createDir(output_dir)
-
- # Get Pareto Points with a "BAND" -- enable_band below is 'True'
- configurations = pareto_utils.dumpParetoConfigsToDir(input_dir, \
- output_dir, self.gold_accuracy, True)
-
- print ("**** pareto config count = ", len(configurations), "\n")
- time.sleep(10)
-
- #configurations = buildRtConfig.loadConfigsFromDir(input_dir, self.gold_accuracy)
- bench_layer_composition = utils.getLayerComposition(self.layer_file)
-
-
- filtered_configs = self.validateAccuracyConfigs(configurations, accuracy_slack)
-
-
- sorted_configurations = sorted(filtered_configs, key=lambda conf: conf.speedup)
-
-
- config_out_path = self.result_dir + "install_gpu_all_config.txt"
-
- buildRtConfig.dumpDevConfigsToRTFile(sorted_configurations, \
- config_out_path, bench_layer_composition, \
- self.gold_accuracy, "gpu")
-
-
- plot_file_path = self.result_dir + "install_all_conf_plot.png"
- genPlots.genScatterPlotFromConfigs(sorted_configurations, plot_file_path)
-
-
- self.dumpBestConfig(sorted_configurations)
-
-
-
-
- def runDevTuner(self):
-
- #self.checkExistingDir()
-
-
- self.startTimer()
-
- self.setBaselineAccuracy()
-
- #self.invokeDevTunerScript(0.8, 2.1, self.autotuner_runs)
- #self.invokeDevTunerScript(1.5, 2.1, self.autotuner_runs)
- #self.invokeDevTunerScript(2.1, 2.1, self.autotuner_runs)
- #self.invokeDevTunerScript(0.9, 2.1, self.autotuner_runs)
-
- # NOTE: for purposes of comparison with fedtuning
- self.invokeDevTunerScript(2.1, 2.1, self.autotuner_runs)
-
-
- #--- self.dumpParetoFiles()
- self.dumpValidatedConfigs(2.1)
-
-
- # NOTE: dumping files for checking experimental parameters for each batch
- self.dumpReferenceFiles()
-
-
- self.endTimer()
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_psnr.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_psnr.py
deleted file mode 100644
index 77e70609b8..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/run_psnr.py
+++ /dev/null
@@ -1,143 +0,0 @@
-
-
-import subprocess
-
-
-def gen30dbFile():
-
- f = open("psnr.txt", "w+");
- f.write("30");
- f.close()
-
-
-def gen20dbFile():
-
- f = open("psnr.txt", "w+");
- f.write("20");
- f.close()
-
-
-
-def runPSNRTuner(bench_name):
-
- Bench = bench_tuner_data[bench_name]
-
- # 20DB run
- gen20dbFile()
- tuner_cmd = "python ../opentuner/autotuner/approxhpvm_tuner.py "
- tuner_cmd += " --test-limit "
- tuner_cmd += str(Bench.autotuner_runs)
- tuner_cmd += " --binary ./"
- tuner_cmd += Bench.tuner_binary
- tuner_cmd += " --num-flags "
- tuner_cmd += str(Bench.num_flags)
- tuner_cmd += " --error-range "
- tuner_cmd += str(Bench.error_range_2)
- tuner_cmd += " --result-dir "
- tuner_cmd += Bench.result_dir_2
- tuner_cmd += " --accuracy "
- tuner_cmd += str(Bench.tuner_accuracy)
-
- print (tuner_cmd)
-
- p = subprocess.Popen(tuner_cmd, shell=True)
- p.wait()
-
-
- # 30DB run
- gen30dbFile()
- tuner_cmd = "python ../opentuner/autotuner/approxhpvm_tuner.py "
- tuner_cmd += " --test-limit "
- tuner_cmd += str(Bench.autotuner_runs)
- tuner_cmd += " --binary ./"
- tuner_cmd += Bench.tuner_binary
- tuner_cmd += " --num-flags "
- tuner_cmd += str(Bench.num_flags)
- tuner_cmd += " --error-range "
- tuner_cmd += str(Bench.error_range_1)
- tuner_cmd += " --result-dir "
- tuner_cmd += Bench.result_dir_1
- tuner_cmd += " --accuracy "
- tuner_cmd += str(Bench.tuner_accuracy)
-
- print (tuner_cmd)
-
- p = subprocess.Popen(tuner_cmd, shell=True)
- p.wait()
-
-
-
-def runPSNRPromiseBench(bench_name):
-
- # NOTE-IMP: Changing current directory to one with promise binaries
- #os.chdir("../build_promise/")
- result_dir_prefix = "../build_tuner/"
-
- Bench = bench_tuner_data[bench_name]
-
- # 20db Run
- gen20dbFile()
- tuner_cmd = "python ../opentuner/autotuner/promise_tuner2.py "
- tuner_cmd += " --test-limit "
- tuner_cmd += str(Bench.autotuner_runs)
- tuner_cmd += " --binary ./"
- tuner_cmd += Bench.promise_binary
- tuner_cmd += " --num-flags "
- tuner_cmd += str(Bench.num_layers)
- tuner_cmd += " --start-range "
- tuner_cmd += str(Bench.start_promise_range)
- tuner_cmd += " --error-range "
- tuner_cmd += str(10)
- tuner_cmd += " --result-dir "
- tuner_cmd += result_dir_prefix
- tuner_cmd += Bench.result_dir_2 + "/promise_tuner/"
- tuner_cmd += " --accuracy "
- tuner_cmd += str(Bench.promise_accuracy)
- tuner_cmd += " --layer-file "
- tuner_cmd += result_dir_prefix
- tuner_cmd += Bench.tensor_desc_file
- tuner_cmd += " --gpu-layers 0 "
- tuner_cmd += " --skip-layers \""
- tuner_cmd += str(Bench.skip_layer_str) + "\""
-
-
- print (tuner_cmd)
-
- p = subprocess.Popen(tuner_cmd, shell=True)
- p.wait()
-
- # 30DB run
- gen30dbFile()
- tuner_cmd = "python ../opentuner/autotuner/promise_tuner2.py "
- tuner_cmd += " --test-limit "
- tuner_cmd += str(Bench.autotuner_runs)
- tuner_cmd += " --binary ./"
- tuner_cmd += Bench.promise_binary
- tuner_cmd += " --num-flags "
- tuner_cmd += str(Bench.num_layers)
- tuner_cmd += " --start-range "
- tuner_cmd += str(Bench.start_promise_range)
- tuner_cmd += " --error-range "
- tuner_cmd += str(10)
- tuner_cmd += " --result-dir "
- tuner_cmd += result_dir_prefix
- tuner_cmd += Bench.result_dir_1 + "/promise_tuner/"
- tuner_cmd += " --accuracy "
- tuner_cmd += str(Bench.promise_accuracy)
- tuner_cmd += " --layer-file "
- tuner_cmd += result_dir_prefix
- tuner_cmd += Bench.tensor_desc_file
- tuner_cmd += " --gpu-layers 0 "
- tuner_cmd += " --skip-layers \""
- tuner_cmd += str(Bench.skip_layer_str) + "\""
-
-
- print (tuner_cmd)
-
- p = subprocess.Popen(tuner_cmd, shell=True)
- p.wait()
-
-
-
-
-
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/swing_selection.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/swing_selection.py
deleted file mode 100644
index 399143c357..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/swing_selection.py
+++ /dev/null
@@ -1,304 +0,0 @@
-
-
-import os
-import warnings
-import matplotlib.pyplot as plt
-import matplotlib.cm as cm
-from matplotlib.ticker import MultipleLocator
-import numpy as np
-from scipy.signal import savgol_filter
-import math
-import struct
-
-
-
-def readDataFromText(textFile):
- results = []
- with open(textFile, "r") as f:
- for line in f:
- token = line.split("\t")
- if (len(token) < 7):
- continue
- record = (token[0], float(token[1]), float(token[5]), float(token[6]))
- results.append(record)
- return results
-
-
-convL1bins = [(0.985901, 1.36474), (0.852871, 1.16982), (0.422283, 0.55701), (0.259752, 0.335259), (0.216577, 0.277843), (0.185812, 0.23733), (0.148996, 0.189171), (0.100007, 0.125816), (0.0003127876261714846, 0.014511194080114365)]
-convL2bins = [(0.995298, 1.3643), (0.861066, 1.16279), (0.426857, 0.547827), (0.262645, 0.330186), (0.218984, 0.273731), (0.187878, 0.233872), (0.150619, 0.186512), (0.10106, 0.124477), (0.00035427528200671077, 0.020199092105031013)]
-
-biasL1bins = [(0.3510325849056244, 0.49078235030174255), (0.30895063281059265, 0.4311973750591278), (0.16023841500282288, 0.22283604741096497), (0.099583700299263, 0.1381179839372635), (0.08340170979499817, 0.11503150314092636), (0.07280077040195465, 0.09948030859231949), (0.05857400223612785, 0.07965542376041412), (0.04044099152088165, 0.054193537682294846), (0.0, 0.0)]
-biasL2bins = [(0.4154910147190094, 0.5820578932762146), (0.3656001389026642, 0.5121639370918274), (0.18930286169052124, 0.2637346684932709), (0.11687946319580078, 0.16306844353675842), (0.09796475619077682, 0.13558265566825867), (0.0848352462053299, 0.11619425565004349), (0.06783176958560944, 0.09277229756116867), (0.046059850603342056, 0.062238890677690506), (0.0, 0.0)]
-
-gemmL1bins= [(0.711203, 0.772211), (0.625894, 0.679601), (0.322665, 0.350383), (0.199646, 0.216727), (0.166556, 0.180781), (0.142945, 0.155132), (0.114662, 0.124399), (0.0771065, 0.0835984), (0.00034660729579627514, 0.008546584285795689)]
-gemmL2bins= [(0.715208, 0.768102), (0.629411, 0.675947), (0.324433, 0.348358), (0.200659, 0.21539), (0.167381, 0.179634), (0.143637, 0.154119), (0.115197, 0.123548), (0.0774642, 0.0829647), (0.0003496285935398191, 0.009841435588896275)]
-
-
-
-def findBinByOp(op):
- if op == 'tensorConv':
- return convL1bins, convL2bins
- if op == 'tensorAdd':
- return biasL1bins, biasL2bins
- if op == 'tensorGemm':
- return gemmL1bins, gemmL2bins
-
- return None, None
-
-
-def getSwing(Lx, opLxbin):
- if opLxbin == None:
- return 0
- for i, (minT, maxT) in enumerate(opLxbin):
- if Lx > minT:
- return i
-
- return 9
-
-
-
-def getConfiguration(L_thresholds):
- configuration = []
- for l in L_thresholds:
- # L0 is op_type
- opL1bin, opL2bin = findBinByOp(l[0])
- # NOTE: L2 is L1 error, L3 is L2 error
- sL1 = getSwing(l[2], opL1bin)
- sL2 = getSwing(l[3], opL2bin)
- if sL1 < 7:
- sL1 = sL1 + 1
- if sL2 < 7:
- sL2 = sL2 + 1
- configuration.append((l[0], l[1], l[2], l[3], sL1, sL2, max(sL1, sL2)))
-
- return configuration
-
-
-def displayConfig(config):
- for c in config:
- print(c)
-
-def displayMultipleConfigurations(configurations):
- for f, c in configurations.items():
- print(f)
- displayConfig(c)
- print()
-
-def getConfigFromFile(filename):
- L_requirements = readDataFromText(filename)
- config = getConfiguration(L_requirements)
- return config
-
-
-def getConfigurationsFromDir(dirname):
- configurations = dict()
- for f in os.listdir(dirname):
- configurations[f] = getConfigFromFile(os.path.join(dirname, f))
-
- return configurations
-
-
-def getLayerWiseTarget(config):
- target = []
- for i, op in enumerate(config):
- if (op[0] == 'tensorGemm') or (op[0] == 'tensorConv'):
- t = op[6]
- for j in range(i+1, len(config)):
- if config[j][0] == 'tensorGemm' or config[j][0] == 'tensorConv':
- break
- t = max(t, config[j][6])
- target.append(t)
- t = 0
-
- return target
-
-
-def dumpLayerWiseTarget(file, targets):
- with open(file, "w") as f:
- for name, t in targets.items():
- f.write(name)
- f.write(" ")
- for i in t:
- f.write(str(i))
- f.write(" ")
- f.write("\n")
-
-
-def getTargetsFromConfigurations(configs):
- targets = dict()
- for f, c in configs.items():
- targets[f] = [d[6] for d in c]
-
- return targets
-
-
-def dumpBenchmarkTargets(name, benchmark_dir):
- benchmark_targets = dict()
- error = ['linear', 'log', 'quad']
- for e in error:
- results_dir = os.path.join(benchmark_dir, e)
- configs = getConfigurationsFromDir(results_dir)
- benchmark_targets[e] = getTargetsFromConfigurations(configs)
-
- return benchmark_targets
-
-
-
-def dumpTargets(filename, targets):
- with open(filename, "w") as f:
- for e, file_configs in targets.items():
- for name, config in file_configs.items():
- for c in config:
- f.write(str(c))
- f.write(" ")
- f.write("\n")
-
-
-
-def getLayerSwings(layer_desc, configurations):
-
- layer_swings = []
- for i in range(len(configurations)):
- config_vals = configurations[i]
- if len(config_vals) == 0:
- continue
-
- layer_index = 0
- index = 0
- swing_vals = []
-
- while layer_index < len(layer_desc):
- if len(layer_desc[layer_index]) == 1:
- promise_swing = config_vals[index]
- layer_type = layer_desc[layer_index][0]
- layer_type = layer_type.strip()
- print ("****layer_type = ", layer_type)
- if layer_type != "conv" and layer_type != "dense":
- promise_swing = -9
- if layer_type == "depthwise_conv":
- promise_swing = -9
- index += 1
- else:
- #print ("index = ", index)
- # FIXIT: Doesn't look right
- print (config_vals[index], config_vals[index+1])
- promise_swing = max(config_vals[index], config_vals[index+1])
- stride = len(layer_desc[layer_index])
- index += stride
-
- swing_vals.append(promise_swing)
- layer_index += 1
-
- layer_swings.append(swing_vals)
-
- return layer_swings
-
-
-
-
-def loadLayerDesc(layer_desc_file):
-
- layer_desc = []
- f = open(layer_desc_file)
- for x in f:
- vals = x.split()
- layer_desc.append(vals)
-
- return layer_desc
-
-
-
-def dumpLayerTargets(targets, tuned_result_dir, layer_desc_file):
-
- layer_desc = loadLayerDesc(layer_desc_file)
- print (layer_desc)
-
- file_names = []
- configurations = []
- for e, file_configs in targets.items():
- for name, config in file_configs.items():
- config_vals = []
- for c in config:
- config_vals.append(c)
- print (config_vals)
-
- configurations.append(config_vals)
-
- rank = e + "_" + "_".join(name.split("_")[-2:])
- file_names.append(rank)
-
-
- # NOTE: get PROMISE swing values corresponding to each layer
- layer_swings = getLayerSwings(layer_desc, configurations)
-
- targets_file_path = tuned_result_dir + "/layer_targets.txt"
- f = open(targets_file_path, "w+")
-
- for config in layer_swings:
- index = 0
- for swing in config:
- swing_str = ""
- if swing == 8 or swing == 9:
- layer_size = len(layer_desc[index])
- for i in range(layer_size):
- swing_str += str(swing)
- if i < layer_size - 1:
- swing_str += " "
- elif swing == -9:
- swing_str += "8"
- else:
- swing_str += str(swing)
-
- if index < len(config) - 1:
- swing_str += ","
-
- f.write(swing_str)
- index += 1
-
- f.write("\n")
-
- f.close()
-
- print(layer_swings)
- return layer_swings, file_names
-
-
-
-def replaceFirstLayer(layer_swings):
-
- # Ensuring first conv on GPU
- for conf in layer_swings:
- conf[0] = 9
-
-
-
-def computeLayerTargets(tuned_result_dir, layer_desc_file):
-
- targets_file_path = tuned_result_dir + "/tensor_targets.txt"
- targets = dumpBenchmarkTargets(targets_file_path, tuned_result_dir)
-
- dumpTargets(targets_file_path, targets)
-
- layer_swings, file_names = dumpLayerTargets(targets, tuned_result_dir, layer_desc_file)
-
- replaceFirstLayer(layer_swings)
-
- return layer_swings, file_names
-
-
-# Externally-called function
-def compute_swing_selection(tuned_result_dir, layer_file):
-
- return computeLayerTargets(tuned_result_dir, layer_file)
-
-
-
-
-if __name__ == "__main__":
-
- tuned_result_dir = "./vgg16_cifar10_tuner_1/high_confidence/"
- layer_file = "layer_composition.txt"
-
- tuned_result_dir = "./resnet18_cifar10_tuner_1/high_confidence/"
- layer_file = "layer_composition2.txt"
- computeLayerTargets(tuned_result_dir, layer_file)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/swing_selection2.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/swing_selection2.py
deleted file mode 100644
index 588edad2a2..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/swing_selection2.py
+++ /dev/null
@@ -1,289 +0,0 @@
-
-
-import os
-import warnings
-import matplotlib.pyplot as plt
-import matplotlib.cm as cm
-from matplotlib.ticker import MultipleLocator
-import numpy as np
-from scipy.signal import savgol_filter
-import math
-import struct
-
-
-
-def readDataFromText(textFile):
- results = []
- with open(textFile, "r") as f:
- for line in f:
- token = line.split("\t")
- if (len(token) < 7):
- continue
- record = (token[0], float(token[1]), float(token[5]), float(token[6]))
- results.append(record)
- return results
-
-
-convL1bins = [(0.985901, 1.36474), (0.852871, 1.16982), (0.422283, 0.55701), (0.259752, 0.335259), (0.216577, 0.277843), (0.185812, 0.23733), (0.148996, 0.189171), (0.100007, 0.125816), (0.0003127876261714846, 0.014511194080114365)]
-convL2bins = [(0.995298, 1.3643), (0.18, 0.19), (0.14, 0.16), (0.11, 0.12), (0.08, 0.09), (0.06, 0.07), (0.04, 0.05), (0.029, 0.035), (0.00031427528200671077, 0.020199092105031013)]
-#convL2bins = [(0.995298, 1.3643), (0.18, 0.19), (0.14, 0.16), (0.11, 0.12), (0.08, 0.09), (0.06, 0.07), (0.04, 0.05), (0.001, 0.004), (0.00031427528200671077, 0.020199092105031013)]
-
-biasL1bins = [(0.3510325849056244, 0.49078235030174255), (0.30895063281059265, 0.4311973750591278), (0.16023841500282288, 0.22283604741096497), (0.099583700299263, 0.1381179839372635), (0.08340170979499817, 0.11503150314092636), (0.07280077040195465, 0.09948030859231949), (0.05857400223612785, 0.07965542376041412), (0.04044099152088165, 0.054193537682294846), (0.0, 0.0)]
-biasL2bins = [(0.4154910147190094, 0.5820578932762146), (0.3656001389026642, 0.5121639370918274), (0.18930286169052124, 0.2637346684932709), (0.11687946319580078, 0.16306844353675842), (0.09796475619077682, 0.13558265566825867), (0.0848352462053299, 0.11619425565004349), (0.06783176958560944, 0.09277229756116867), (0.046059850603342056, 0.062238890677690506), (0.0, 0.0)]
-
-gemmL1bins= [(0.711203, 0.772211), (0.625894, 0.679601), (0.322665, 0.350383), (0.199646, 0.216727), (0.166556, 0.180781), (0.142945, 0.155132), (0.114662, 0.124399), (0.0771065, 0.0835984), (0.00034660729579627514, 0.008546584285795689)]
-gemmL2bins= [(0.715208, 0.768102), (0.629411, 0.675947), (0.324433, 0.348358), (0.200659, 0.21539), (0.167381, 0.179634), (0.143637, 0.154119), (0.115197, 0.123548), (0.0774642, 0.0829647), (0.0003496285935398191, 0.009841435588896275)]
-
-
-
-def findBinByOp(op):
- if op == 'tensorConv':
- return convL1bins, convL2bins
- if op == 'tensorAdd':
- return biasL1bins, biasL2bins
- if op == 'tensorGemm':
- return gemmL1bins, gemmL2bins
-
- return None, None
-
-
-def getSwing(Lx, opLxbin):
- if opLxbin == None:
- return 0
- for i, (minT, maxT) in enumerate(opLxbin):
- if Lx > minT:
- return i
-
- return 9
-
-
-
-def getConfiguration(L_thresholds):
- configuration = []
- for l in L_thresholds:
- # L0 is op_type
- opL1bin, opL2bin = findBinByOp(l[0])
- # NOTE: L2 is L1 error, L3 is L2 error
- # only using L2 for image pipelines
- sL2 = getSwing(l[3], opL2bin)
- if sL2 < 7:
- sL2 = sL2 + 1
- configuration.append((l[0], l[1], l[2], l[3], sL2, sL2, sL2))
-
- return configuration
-
-
-def displayConfig(config):
- for c in config:
- print(c)
-
-def displayMultipleConfigurations(configurations):
- for f, c in configurations.items():
- print(f)
- displayConfig(c)
- print()
-
-def getConfigFromFile(filename):
- L_requirements = readDataFromText(filename)
- config = getConfiguration(L_requirements)
- return config
-
-
-def getConfigurationsFromDir(dirname):
- configurations = dict()
- for f in os.listdir(dirname):
- configurations[f] = getConfigFromFile(os.path.join(dirname, f))
-
- return configurations
-
-
-def getLayerWiseTarget(config):
- target = []
- for i, op in enumerate(config):
- if (op[0] == 'tensorGemm') or (op[0] == 'tensorConv'):
- t = op[6]
- target.append(t)
-
- return target
-
-
-def dumpLayerWiseTarget(file, targets):
- with open(file, "w") as f:
- for name, t in targets.items():
- f.write(name)
- f.write(" ")
- for i in t:
- f.write(str(i))
- f.write(" ")
- f.write("\n")
-
-
-def getTargetsFromConfigurations(configs):
- targets = dict()
- for f, c in configs.items():
- targets[f] = [d[6] for d in c]
-
- return targets
-
-
-def dumpBenchmarkTargets(name, benchmark_dir):
- benchmark_targets = dict()
- error = ['linear', 'log', 'quad']
- for e in error:
- results_dir = os.path.join(benchmark_dir, e)
- configs = getConfigurationsFromDir(results_dir)
- benchmark_targets[e] = getTargetsFromConfigurations(configs)
-
- return benchmark_targets
-
-
-def dumpTargets(filename, targets):
- with open(filename, "w") as f:
- for e, file_configs in targets.items():
- for name, config in file_configs.items():
- for c in config:
- f.write(str(c))
- f.write(" ")
- f.write("\n")
-
-
-
-def getLayerSwings(layer_desc, configurations):
-
- layer_swings = []
- for i in range(len(configurations)):
- config_vals = configurations[i]
- layer_index = 0
- index = 0
- swing_vals = []
-
- while layer_index < len(layer_desc):
- if len(layer_desc[layer_index]) == 1:
- promise_swing = config_vals[index]
- layer_type = layer_desc[layer_index]
- if layer_type != "conv" and layer_type != "dense":
- promise_swing = -9
- index += 1
- else:
- print (config_vals[index], config_vals[index+1])
- promise_swing = max(config_vals[index], config_vals[index+1])
- stride = len(layer_desc[layer_index])
- #print ("*stride = ", stride)
- index += stride
-
- swing_vals.append(promise_swing)
- layer_index += 1
-
- layer_swings.append(swing_vals)
-
- return layer_swings
-
-
-
-
-def loadLayerDesc(layer_desc_file):
-
- layer_desc = []
- f = open(layer_desc_file)
- for x in f:
- vals = x.split()
- layer_desc.append(vals)
-
- return layer_desc
-
-
-
-def dumpLayerTargets(targets, tuned_result_dir, layer_desc_file):
-
- layer_desc = loadLayerDesc(layer_desc_file)
- print (layer_desc)
-
- file_names = []
- configurations = []
- for e, file_configs in targets.items():
- for name, config in file_configs.items():
- config_vals = []
- for c in config:
- config_vals.append(c)
- print (config_vals)
-
- configurations.append(config_vals)
-
- rank = e + "_" + "_".join(name.split("_")[-2:])
- file_names.append(rank)
-
-
- # NOTE: get PROMISE swing values corresponding to each layer
- layer_swings = getLayerSwings(layer_desc, configurations)
-
- targets_file_path = tuned_result_dir + "/layer_targets.txt"
- f = open(targets_file_path, "w+")
-
- for config in layer_swings:
- index = 0
- for swing in config:
- swing_str = ""
- if swing == 8 or swing == 9:
- layer_size = len(layer_desc[index])
- for i in range(layer_size):
- swing_str += str(swing)
- if i < layer_size - 1:
- swing_str += " "
- elif swing == -9:
- swing_str += "8"
- else:
- swing_str += str(swing)
-
- if index < len(config) - 1:
- swing_str += ","
-
- f.write(swing_str)
- index += 1
-
- f.write("\n")
-
- f.close()
-
- print(layer_swings)
- return layer_swings, file_names
-
-
-
-def replaceFirstLayer(layer_swings):
-
- # Ensuring first conv on GPU
- for conf in layer_swings:
- conf[0] = 9
-
-
-
-def computeLayerTargets(tuned_result_dir, layer_desc_file):
-
- targets_file_path = tuned_result_dir + "/tensor_targets.txt"
- targets = dumpBenchmarkTargets(targets_file_path, tuned_result_dir)
-
- dumpTargets(targets_file_path, targets)
-
- layer_swings, file_names = dumpLayerTargets(targets, tuned_result_dir, layer_desc_file)
-
- replaceFirstLayer(layer_swings)
-
- return layer_swings, file_names
-
-
-# Externally-called function
-def compute_swing_selection2(tuned_result_dir, layer_file):
-
- return computeLayerTargets(tuned_result_dir, layer_file)
-
-
-
-
-if __name__ == "__main__":
-
- tuned_result_dir = "./vgg16_cifar10_tuner_1/high_confidence/"
- layer_file = "layer_composition.txt"
-
- tuned_result_dir = "./resnet18_cifar10_tuner_1/high_confidence/"
- layer_file = "layer_composition2.txt"
- computeLayerTargets(tuned_result_dir, layer_file)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/utils.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/utils.py
deleted file mode 100644
index 9ff3622d13..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/utils.py
+++ /dev/null
@@ -1,296 +0,0 @@
-
-
-import os
-import sys
-import subprocess
-from benchmarks import batch_id
-from global_paths import tensorRT_dir
-
-
-
-def createDir(dir_path):
-
- try:
- if not os.path.exists(dir_path):
- os.mkdir(dir_path)
- except:
- print ("!ERROR: Could NOT create result directory = ", dir_path)
- sys.exit(-1)
-
-
-def createResultDirs(benchmarks):
-
- for bench_name in benchmarks:
- Bench = benchmarks[bench_name]
-
- print ("Base Directory: ", Bench.base_dir , "BatchId = ", batch_id)
-
- createDir(tensorRT_dir + Bench.base_dir + "/loss_1")
- createDir(tensorRT_dir + Bench.base_dir + "/loss_2")
- createDir(tensorRT_dir + Bench.base_dir + "/loss_3")
- createDir(tensorRT_dir + Bench.base_dir + "/loss_123/" + batch_id)
- createDir(tensorRT_dir + Bench.base_dir + "/loss_1/" + batch_id)
- createDir(tensorRT_dir + Bench.base_dir + "/loss_2/" + batch_id)
- createDir(tensorRT_dir + Bench.base_dir + "/loss_3/" + batch_id)
- createDir(tensorRT_dir + Bench.base_dir + "/loss_123/" + batch_id + "/devtuner/")
- createDir(tensorRT_dir + Bench.base_dir + "/loss_1/" + batch_id + "/devtuner/" )
- createDir(tensorRT_dir + Bench.base_dir + "/loss_2/" + batch_id + "/devtuner/")
- createDir(tensorRT_dir + Bench.base_dir + "/loss_3/" + batch_id + "/devtuner/")
-
-
-
-
-
-def genBaselineConfig(flags_file_path, default_flag, num_layers):
-
- f = open(flags_file_path, "w+")
- for i in range(num_layers):
- f.write(str(default_flag) + "\n")
-
- f.close()
-
-
-
-def readAccuracy(accuray_res_file):
-
- file = open(accuray_res_file, "r")
- accuracy_str = file.read()
- file.close()
- accuracy = 0 # float(accuracy_str)
-
- try:
- accuracy = float(accuracy_str)
- except:
- print("ERROR: Reading Accuray file - Aborting.... \n")
- sys.exit(-1)
-
-
- print ("*Configuration Accuracy = ", accuracy)
- return accuracy
-
-
-
-#**** Exported Function ****/
-def getBaselineAccuracy(binary_path, num_layers):
-
- genBaselineConfig("promise_flags", 11, num_layers)
-
- run_cmd = "./" + binary_path
- print (run_cmd)
-
- p = subprocess.Popen(run_cmd, shell=True)
- p.wait()
-
- return readAccuracy("final_accuracy")
-
-
-
-def getLayerComposition(layer_composition_file):
-
- layer_desc = []
- f = open(layer_composition_file)
- for x in f:
- vals = x.split()
- layer_desc.append(vals)
-
- return layer_desc
-
-
-def debug_print(str):
-
- debug_flag = False
-
- if debug_flag == True:
- print (str)
-
-
-
-def readOpKnobs(global_knobs_file, op_type, analysis_mode):
-
- knobs_file = open(global_knobs_file, "r")
-
- tuning_knobs = []
- for knob in knobs_file:
- toks = knob.split("\t")
- if op_type in toks[-1] and analysis_mode in toks[-2]:
- knob_id = toks[0].split(",")[1]
- tuning_knobs.append(knob_id)
-
- return tuning_knobs
-
-
-
-def readConvKnobs(global_knobs_file, analysis_mode):
-
- return readOpKnobs(global_knobs_file, "conv", analysis_mode)
-
-
-def readFCKnobs(global_knobs_file, analysis_mode):
-
- return readOpKnobs(global_knobs_file, "fc", analysis_mode)
-
-
-def readRedKnobs(global_knobs_file, analysis_mode):
-
- return readOpKnobs(global_knobs_file, "red", analysis_mode)
-
-
-
-
-def createDevKnobs(layer_file, global_knobs_file, out_file):
-
- f = open(layer_file, "r")
-
- conv_knobs = readConvKnobs(global_knobs_file, "dev")
- fc_knobs = readFCKnobs(global_knobs_file, "dev")
- red_knobs = readRedKnobs(global_knobs_file, "dev")
-
- print (conv_knobs, fc_knobs, red_knobs)
-
- f_out = open(out_file, "w+")
-
- for x in f:
- if "conv" in x:
- f_out.write(",".join(conv_knobs) + "\n")
- if "dense" in x:
- f_out.write(",".join(fc_knobs) + "\n")
- if "red" in x:
- f_out.write(",".join(red_knobs) + "\n")
-
- f_out.close()
-
-
-
-def removePromiseKnobs(conv_knobs):
-
- promise_knobs = ["1", "2", "3", "4", "5", "6", "7"]
- conv_knobs2 = []
-
- for knob in conv_knobs:
- if knob not in promise_knobs:
- conv_knobs2.append(knob)
-
- return conv_knobs2
-
-
-
-def createInstallAndDevKnobs(layer_file, global_knobs_file, out_file):
-
- f = open(layer_file, "r")
-
- conv_knobs_dev = readConvKnobs(global_knobs_file, "dev")
- fc_knobs_dev = readFCKnobs(global_knobs_file, "dev")
- red_knobs_dev = readRedKnobs(global_knobs_file, "dev")
-
- conv_knobs_install = readConvKnobs(global_knobs_file, "install")
- fc_knobs_install = readFCKnobs(global_knobs_file, "install")
- red_knobs_install = readRedKnobs(global_knobs_file, "install")
-
-
-
- #conv_knobs_dev.remove("11") # remove FP32 from install-time tuning
- #fc_knobs_dev.remove("11") # remove FP32 from install-time tuning
- #red_knobs_dev.remove("11") # remove FP32 from install-time tuning
-
- conv_knobs = conv_knobs_dev + conv_knobs_install
- fc_knobs = fc_knobs_dev + fc_knobs_install
- red_knobs = red_knobs_dev + red_knobs_install
-
- print (conv_knobs, fc_knobs, red_knobs)
-
- #sys.exit(0)
-
- f_out = open(out_file, "w+")
-
- ind = 0
- for x in f:
- if "conv" in x:
- layer_conv_knobs = conv_knobs
- if ind == 0:
- layer_conv_knobs = removePromiseKnobs(conv_knobs)
- f_out.write(",".join(layer_conv_knobs) + "\n")
- if "dense" in x:
- f_out.write(",".join(fc_knobs) + "\n")
- if "red" in x:
- f_out.write(",".join(red_knobs) + "\n")
-
- ind += 1
-
- f_out.close()
-
-
-
-
-
-def getInstallAndDevKnobs(layer_file, global_knobs_file):
-
- f = open(layer_file, "r")
-
- conv_knobs_dev = readConvKnobs(global_knobs_file, "dev")
- fc_knobs_dev = readFCKnobs(global_knobs_file, "dev")
- red_knobs_dev = readRedKnobs(global_knobs_file, "dev")
-
- conv_knobs_install = readConvKnobs(global_knobs_file, "install")
- fc_knobs_install = readFCKnobs(global_knobs_file, "install")
- red_knobs_install = readRedKnobs(global_knobs_file, "install")
-
-
- conv_knobs = conv_knobs_dev + conv_knobs_install
- fc_knobs = fc_knobs_dev + fc_knobs_install
- red_knobs = red_knobs_dev + red_knobs_install
-
- print (conv_knobs, fc_knobs, red_knobs)
-
-
- bench_knobs = []
-
- ind = 0
- for x in f:
- if "conv" in x:
- layer_conv_knobs = conv_knobs
- if ind == 0:
- layer_conv_knobs = removePromiseKnobs(conv_knobs)
- bench_knobs.append(layer_conv_knobs)
- if "dense" in x:
- bench_knobs.append(fc_knobs)
- if "red" in x:
- bench_knobs.append(red_knobs)
-
- ind += 1
-
-
- return bench_knobs
-
-
-
-
-
-
-def dumpKnobsFile(knobs, out_file):
-
-
- f_out = open(out_file, "w+")
-
- for layer_knobs in knobs:
- f_out.write(",".join(layer_knobs) + "\n")
-
- f_out.close()
-
-
-
-
-
-
-
-
-if __name__ == "__main__":
-
-
- #createDevKnobs("../data/alexnet2/alexnet2_layers.txt", \
- # "../data/global_knobs.txt", "dev_knobs.txt")
-
-
- knobs = getInstallAndDevKnobs("../data/alexnet2/alexnet2_layers.txt", \
- "../data/global_knobs.txt")
-
- print ("*** knobs = ", knobs)
diff --git a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/validation.py b/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/validation.py
deleted file mode 100644
index c334b2b319..0000000000
--- a/hpvm/projects/hpvm-tensor-rt/autotuner/tuner_driver_src/validation.py
+++ /dev/null
@@ -1,802 +0,0 @@
-
-
-import os
-import sys
-import subprocess
-import shutil
-from compute_confs import computePSNRBenchSwings, computeBenchSwings
-from buildRtConfig import loadConfigData
-
-
-
-def getLayerString(layer_swings):
-
- index = 0
- layer_string = ""
- for swing in layer_swings:
- layer_string += str(swing)
- if index < len(layer_swings) - 1:
- layer_string += ","
- return layer_string
-
-
-
-def testValidationRun(Bench, validation_dir, layer_swings, threshold, rank_str):
-
- #### FIXME
- #os.chdir("../build_promise/")
-
- validation_acc = Bench.validation_accuracy
- target_acc = validation_acc - threshold
-
- validation_binary = Bench.validation_binary
-
- # Write to promise_flags
- fout = open("promise_flags", "w+")
- for swing in layer_swings:
- int_swing = int(swing)
- if int_swing > 0:
- fout.write(str(swing) + "\n")
- fout.close()
-
- # Execute Validation Run
- p = subprocess.Popen("./" + validation_binary, shell=True)
- p.wait()
-
- f = open("run_accuracies.txt")
- index = 0.0
- unsuccessful = 0.0
- sum_acc = 0.0
- for x in f:
- x = x.strip()
- acc = float(x)
- if acc < target_acc:
- unsuccessful += 1
- index += 1
- sum_acc += acc
-
- f.close()
-
- confidence = ( (index - unsuccessful) / index) * 100.0
- print ("run_confidence = ", confidence)
- avg_acc = sum_acc / index
-
- out_fname = validation_dir + validation_binary + "_" + str(avg_acc)
- shutil.copy("run_accuracies.txt", out_fname + "_" + rank_str)
-
- layer_string = getLayerString(layer_swings)
- f = open(out_fname, "w")
- f.write("config:\t" + layer_string + "\n")
- f.write("confidence:\t" + str(confidence) + "\n")
- f.close()
-
- return confidence
-
-
-
-
-def testPromiseRun(Bench, layer_swings, threshold):
-
- #### FIXME
- #os.chdir("../build_promise/")
-
- validation_acc = Bench.validation_accuracy
- target_acc = validation_acc - threshold
-
- validation_binary = Bench.validation_binary
-
- # Write to promise_flags
- fout = open("promise_flags", "w+")
- for swing in layer_swings:
- int_swing = int(swing)
- if int_swing > 0:
- fout.write(str(swing) + "\n")
- fout.close()
-
- # Execute Validation Run
- p = subprocess.Popen("./" + validation_binary, shell=True)
- p.wait()
-
- f = open("run_accuracies.txt")
- index = 0.0
- unsuccessful = 0.0
- sum_acc = 0.0
- for x in f:
- x = x.strip()
- acc = float(x)
- if acc < target_acc:
- unsuccessful += 1
- index += 1
- sum_acc += acc
-
- f.close()
-
- confidence = ( (index - unsuccessful) / index) * 100.0
- print ("run_confidence = ", confidence)
- avg_acc = sum_acc / index
-
- return confidence
-
-
-
-
-
-
-
-
-def dumpConfigConfidence(configs, confidence_list,
- result_dir, layer_desc_file):
-
- #### FIXME
- #os.chdir("../build_tuner/")
-
- layer_desc = loadLayerDesc(layer_desc_file)
- print (layer_desc)
-
- f = open(result_dir + "/conf_confidences.txt", "w+")
-
- count = 0
- for config in configs:
- index = 0
- for swing in config:
- swing_str = ""
- if swing == 8 or swing == 9:
- layer_size = len(layer_desc[index])
- for i in range(layer_size):
- swing_str += str(swing)
- if i < layer_size - 1:
- swing_str += " "
- elif swing == -9:
- swing_str += "8"
- else:
- swing_str += str(swing)
-
- if index < len(config) - 1:
- swing_str += ","
- f.write(swing_str)
-
- index += 1
-
- f.write("\t" + str(confidence_list[count]))
- f.write("\n")
- count +=1
-
- f.close()
-
-
-
-
-def dumpValidatedConfigs(configs, result_dir, layer_desc_file,
- output_file_name):
-
- os.chdir("../build_tuner/")
-
- layer_desc = loadLayerDesc(layer_desc_file)
- print (layer_desc)
-
- f = open(result_dir + "/" + output_file_name, "w+")
-
- for config in configs:
- index = 0
- for swing in config:
- swing_str = ""
- if swing == 8 or swing == 9:
- layer_size = len(layer_desc[index])
- for i in range(layer_size):
- swing_str += str(swing)
- if i < layer_size - 1:
- swing_str += " "
- elif swing == -9:
- swing_str += "8"
- else:
- swing_str += str(swing)
-
- if index < len(config) - 1:
- swing_str += ","
- f.write(swing_str)
-
- index += 1
- f.write("\n")
- f.close()
-
-
-
-def dumpRankings(validated_ranks, result_dir, rank_file):
-
- os.chdir("../build_tuner/")
- f = open(result_dir + "/" + rank_file, "w+")
- for rank in validated_ranks:
- f.write(rank + "\n")
-
- f.close()
-
-
-
-
-def replaceFP32Configs(loss_confs1, loss_confs2):
-
- for swing_conf in loss_confs1:
- for i in range(0, len(swing_conf)):
- if swing_conf[i] == 9:
- swing_conf[i] = 8
- if i == len(swing_conf) - 1:
- swing_conf[i] = 7
-
- for swing_conf in loss_confs2:
- for i in range(0, len(swing_conf)):
- if swing_conf[i] == 9:
- swing_conf[i] = 8
- if i == len(swing_conf) - 1:
- swing_conf[i] = 7
-
-
- return loss_confs1, loss_confs2
-
-
-
-def replaceGPUConfigs(Bench, loss_confs1, loss_confs2):
-
- skip_layer_str = Bench.skip_layer_str
- layer_ids = skip_layer_str.split("_")
- skip_layers = []
- for layer_id in layer_ids:
- skip_layers.append(int(layer_id))
-
-
- for swing_conf in loss_confs1:
- for i in range(0, len(swing_conf)):
- if i in skip_layers and swing_conf[i] < 8:
- swing_conf[i] = 8
-
- for swing_conf in loss_confs2:
- for i in range(0, len(swing_conf)):
- if i in skip_layers and swing_conf[i] < 8:
- swing_conf[i] = 8
-
-
- return loss_confs1, loss_confs2
-
-
-
-
-def runBenchValidation(Bench):
-
- #Bench = bench_tuner_data[bench_name]
-
- loss_confs, conf_ranks = computeBenchSwings(Bench)
- loss1_confs = loss_confs[0]
- loss2_confs = loss_confs[1]
- conf_ranks1 = conf_ranks[0]
- conf_ranks2 = conf_ranks[1]
-
- #loss1_confs, loss2_confs = replaceFP32Configs(loss1_confs, loss2_confs)
-
-
- validation_dir_1 = "../build_tuner/" + Bench.result_dir_1 + "/validation_runs/"
- if not os.path.exists(validation_dir_1):
- os.mkdir(validation_dir_1)
-
- validation_dir_2 = "../build_tuner/" + Bench.result_dir_2 + "/validation_runs/"
- if not os.path.exists(validation_dir_2):
- os.mkdir(validation_dir_2)
-
-
- ind = 0
- validated_confs1 = []
- validated_ranks1 = []
- failed_confs1 = []
- confidences1 = []
- for layer_swings in loss1_confs:
- print ("len(layer_Swings) = ", len(layer_swings), "\n")
- confidence = testValidationRun(Bench, validation_dir_1,
- layer_swings, 1.0, conf_ranks1[ind])
- if confidence >= 95:
- validated_confs1.append(layer_swings)
- confidences1.append(confidence)
- validated_ranks1.append(conf_ranks1[ind])
- else:
- failed_confs1.append(layer_swings)
- ind += 1
-
-
- ind = 0
- validated_confs2 = []
- validated_ranks2 = []
- failed_confs2 = []
- confidences2 = []
- for layer_swings in loss2_confs:
- confidence = testValidationRun(Bench, validation_dir_2, layer_swings, 2.0, conf_ranks2[ind])
- if confidence >= 92:
- validated_confs2.append(layer_swings)
- confidences2.append(confidence)
- validated_ranks2.append(conf_ranks2[ind])
- else:
- failed_confs2.append(layer_swings)
- ind += 1
-
- dumpValidatedConfigs(validated_confs1, Bench.result_dir_1,
- Bench.layer_file, "validated_confs.txt")
- dumpValidatedConfigs(validated_confs2, Bench.result_dir_2,
- Bench.layer_file, "validated_confs.txt")
-
- dumpValidatedConfigs(failed_confs1, Bench.result_dir_1,
- Bench.layer_file, "failed_confs.txt")
- dumpValidatedConfigs(failed_confs2, Bench.result_dir_2,
- Bench.layer_file, "failed_confs.txt")
-
- dumpRankings(validated_ranks1, Bench.result_dir_1, "validated_ranks.txt")
- dumpRankings(validated_ranks2, Bench.result_dir_2, "validated_ranks.txt")
-
- dumpConfigConfidence(validated_confs1, confidences1,
- Bench.result_dir_1, Bench.layer_file)
-
- dumpConfigConfidence(validated_confs2, confidences2,
- Bench.result_dir_2, Bench.layer_file)
-
-
- print (validated_confs1)
- print (validated_confs2)
-
-
-
-def readPromiseResults(loss1_file, loss2_file):
-
- loss_confs = []
- loss1_confs = []
- f1 = open(loss1_file)
- for x in f1:
- print (x)
- swing_toks = x.split(",")
- swing_list = []
- for swing_str in swing_toks:
- swing_val = int(swing_str.split(" ")[0])
- swing_list.append(swing_val)
- loss1_confs.append(swing_list)
-
- loss_confs.append(loss1_confs)
-
- loss2_confs = []
- f2 = open(loss1_file)
- for x in f2:
- swing_toks = x.split(",")
- swing_list = []
- for swing_str in swing_toks:
- swing_val = int(swing_str.split(" ")[0])
- swing_list.append(swing_val)
- loss2_confs.append(swing_list)
-
- loss_confs.append(loss2_confs)
-
- return loss_confs
-
-
-
-
-
-
-def readPromiseResults2(loss1_file, loss2_file, layer_file):
-
- layer_desc = loadLayerDesc(layer_file)
-
- loss_confs = []
- loss1_confs = []
- f1 = open(loss1_file)
- for x in f1:
- print (x)
- swing_toks = x.split(",")
- swing_list = []
-
- it = 0
- for swing_str in swing_toks:
- swing_val = int(swing_str.split(" ")[0])
- if "conv" in layer_desc[it] or "dense" in layer_desc[it]:
- swing_list.append(swing_val)
-
- it += 1
-
- loss1_confs.append(swing_list)
-
- loss_confs.append(loss1_confs)
-
- loss2_confs = []
- f2 = open(loss1_file)
- for x in f2:
- swing_toks = x.split(",")
- swing_list = []
-
- it = 0
- for swing_str in swing_toks:
- swing_val = int(swing_str.split(" ")[0])
- if "conv" in layer_desc[it] or "dense" in layer_desc[it]:
- swing_list.append(swing_val)
-
- it += 1
-
- loss2_confs.append(swing_list)
-
- loss_confs.append(loss2_confs)
-
- return loss_confs
-
-
-
-
-
-def readPromiseResults3(result_dir):
-
- loss_confs = []
- # NOTE: Second parameter is ignored
- config_arr = loadConfigData(result_dir, 100)
-
- for config in config_arr:
- loss_confs.append(config.flags)
-
- return loss_confs
-
-
-
-
-
-
-
-
-def runPromiseBenchValidation(Bench):
-
-
- dir_prefix = "../build_tuner/"
- #Bench = bench_tuner_data[bench_name]
- #loss_confs = readPromiseResults(dir_prefix + Bench.loss1_result_file, dir_prefix + Bench.loss2_result_file)
- loss_confs = readPromiseResults2(dir_prefix + Bench.loss1_result_file, \
- dir_prefix + Bench.loss2_result_file, Bench.layer_file)
-
- loss1_confs = loss_confs[0]
- loss2_confs = loss_confs[1]
-
- ind = 0
- validated_confs1 = []
- failed_confs1 = []
- for layer_swings in loss1_confs:
- confidence = testPromiseRun(Bench, layer_swings, 1.0)
- if confidence >= 95:
- validated_confs1.append(layer_swings)
- else:
- failed_confs1.append(layer_swings)
- ind += 1
-
-
- ind = 0
- validated_confs2 = []
- failed_confs2 = []
- for layer_swings in loss2_confs:
- confidence = testPromiseRun(Bench, layer_swings, 2.0)
- if confidence >= 95:
- validated_confs2.append(layer_swings)
- else:
- failed_confs2.append(layer_swings)
- ind += 1
-
-
- dumpValidatedConfigs(validated_confs1, Bench.result_dir_1,
- Bench.layer_file, "promise_validated_confs.txt")
- dumpValidatedConfigs(validated_confs2, Bench.result_dir_2,
- Bench.layer_file, "promise_validated_confs.txt")
-
- dumpValidatedConfigs(failed_confs1, Bench.result_dir_1,
- Bench.layer_file, "promise_failed_confs.txt")
- dumpValidatedConfigs(failed_confs2, Bench.result_dir_2,
- Bench.layer_file, "promise_failed_confs.txt")
-
-
-
-
-
-def copyValidatedConf(result_dir, validated_confs):
-
- src_dir = result_dir + "/promise_tuner/high_confidence/"
- dest_dir = result_dir + "/promise_tuner/validated/"
-
- if not os.path.isdir(dest_dir):
- os.mkdir(dest_dir)
-
- for fname in validated_confs:
- shutil.copy(src_dir + fname, dest_dir + fname)
-
-
-
-def copyFailedConf(result_dir, failed_confs):
-
- src_dir = result_dir + "/promise_tuner/high_confidence/"
- dest_dir = result_dir + "/promise_tuner/failed/"
-
- if not os.path.isdir(dest_dir):
- os.mkdir(dest_dir)
-
- for fname in failed_confs:
- shutil.copy(src_dir + fname, dest_dir + fname)
-
-
-
-
-def validateConfigs(Bench, result_dir, configs_arr, acc_thresh):
-
- validated_confs = []
- failed_confs = []
- for conf in configs_arr:
- layer_swings = conf.flags
- confidence = testPromiseRun(Bench, layer_swings, acc_thresh)
- if confidence >= 95:
- validated_confs.append(conf.fname)
- else:
- failed_confs.append(conf.fname)
-
-
- copyValidatedConf(result_dir, validated_confs)
- copyFailedConf(result_dir, failed_confs)
-
-
-
-
-
-
-
-def runPromiseBenchValidation2(Bench):
-
-
- config_arr1 = loadConfigData(Bench.result_dir_1, 100)
- config_arr2 = loadConfigData(Bench.result_dir_2, 100)
- config_arr3 = loadConfigData(Bench.result_dir_3, 100)
-
-
- validateConfigs(Bench, Bench.result_dir_1, config_arr1, 1.0)
- validateConfigs(Bench, Bench.result_dir_2, config_arr2, 2.0)
- validateConfigs(Bench, Bench.result_dir_3, config_arr3, 3.0)
-
-
-
-
-### NOTE: Algo Tuner Validation routines
-
-
-
-
-
-def addAccuracyLoss(dest_file, accuracy_loss):
-
- f = open(dest_file, "r")
- file_str = ""
- ind = 0
- for x in f:
- line_str = x
- if ind == 0:
- line_str = x.replace("\n", "")
- line_str += "\tvalidation_loss=" + str(accuracy_loss) + "\n"
-
- file_str += line_str
- ind += 1
- f.close()
-
-
- f_out = open(dest_file, "w+")
- f_out.write(file_str)
- f_out.close()
-
-
-
-
-def dumpValidConfigs(result_dir, validated_confs, src_dir):
-
- src_dir = result_dir + "/algo_tuner/" + src_dir + "/" # high_confidence/"
- dest_dir = result_dir + "/algo_tuner/validated/"
-
- if not os.path.isdir(dest_dir):
- os.mkdir(dest_dir)
-
- for (fname, accuracy_loss) in validated_confs:
- dest_file = dest_dir + fname
- shutil.copy(src_dir + fname, dest_file)
- addAccuracyLoss(dest_file, accuracy_loss)
-
-
-
-def dumpFailedConfigs(result_dir, failed_confs, src_dir):
-
- src_dir = result_dir + "/algo_tuner/" + src_dir + "/" # high_confidence/"
- dest_dir = result_dir + "/algo_tuner/failed/"
-
- if not os.path.isdir(dest_dir):
- os.mkdir(dest_dir)
-
- for (fname, accuracy_loss) in failed_confs:
- dest_file = dest_dir + fname
- shutil.copy(src_dir + fname, dest_file)
- addAccuracyLoss(dest_file, accuracy_loss)
-
-
-
-
-def readAccuracy(file_name):
-
- file = open(file_name, "r")
- file_str = file.read()
- file.close()
-
- accuracy = 0.0
- try:
- accuracy = float(file_str)
- except:
- print ("ERROR: Reading accuracy from 'final_accuracy' file")
- sys.exit(0)
-
- print ("accuracy = ", accuracy)
- return accuracy
-
-def getBaselineConfig(num_layers):
-
- fp32_swing = 11
- swings = []
-
- for i in range(num_layers):
- swings.append(str(fp32_swing))
-
- return swings
-
-
-
-def readConfidence(target_acc):
-
- f = open("run_accuracies.txt")
- index = 0.0
- unsuccessful = 0.0
- sum_acc = 0.0
- for x in f:
- x = x.strip()
- acc = float(x)
- if acc < target_acc:
- unsuccessful += 1
- index += 1
- sum_acc += acc
-
- f.close()
-
- confidence = ( (index - unsuccessful) / index) * 100.0
- print ("run_confidence = ", confidence)
- avg_acc = sum_acc / index
-
- return confidence, avg_acc
-
-
-
-def invokeBinary(binary_path, layer_swings, runs, input_size, offset, target_acc): # threshold):
-
- default_skip = 4
- # Write to promise_flags
- fout = open("promise_flags", "w+")
- for swing in layer_swings:
- int_swing = int(swing)
- if int_swing > 0:
- fout.write(str(swing) + "\n")
- fout.close()
-
- run_cmd = "./" + binary_path + " " + str(runs) + " " + str(target_acc) + " " + str(default_skip) + " " + str(input_size) + " " + str(offset)
- # Execute Validation Run
- #p = subprocess.Popen("./" + validation_binary, shell=True)
-
- p = subprocess.Popen(run_cmd, shell=True)
- p.wait()
-
-
-
-
-
-def validateAlgoConfigs(binary_path, result_dir, configs_arr, gold_acc, \
- acc_thresh, runs, src_dir = "high_confidence"):
-
- # NOTE: Use confidence target as 95%
- confidence_target = 95
- # NOTE: 1 run sufficient for software approximations
-
- validated_confs = []
- failed_confs = []
-
- #validation_acc = Bench.validation_accuracy
- target_acc = gold_acc - acc_thresh
-
- for conf in configs_arr:
- layer_swings = conf.flags
- invokeBinary(binary_path, layer_swings, runs, 2000, 8000, target_acc)
- confidence, avg_acc = readConfidence(target_acc)
-
- accuracy_loss = gold_acc - avg_acc
- if confidence >= confidence_target:
- validated_confs.append((conf.fname, accuracy_loss))
- else:
- failed_confs.append((conf.fname, accuracy_loss))
-
-
- dumpValidConfigs(result_dir, validated_confs, src_dir)
- dumpFailedConfigs(result_dir, failed_confs, src_dir)
-
-
-
-
-def runAlgoBenchValidate(Bench):
-
- num_layers = Bench.num_layers
- base_conf = getBaselineConfig(num_layers)
- # Path to binary to run
- binary_path = Bench.promise_binary
- # NOTE: 'target_acc' passed 0.0 since unused for baseline run
- invokeBinary(binary_path, base_conf, 1, 2000, 8000, 0.0)
- gold_acc = readAccuracy("final_accuracy")
-
-
- loss1_dir = Bench.result_dir_1
- loss2_dir = Bench.result_dir_2
- loss3_dir = Bench.result_dir_3
-
- loss1_configs = loadConfigData(loss1_dir, 100)
- loss2_configs = loadConfigData(loss2_dir, 100)
- loss3_configs = loadConfigData(loss3_dir, 100)
-
- runs = 1
- validateAlgoConfigs(binary_path, loss1_dir, loss1_configs, gold_acc, 1.0, runs)
- validateAlgoConfigs(binary_path, loss2_dir, loss2_configs, gold_acc, 2.0, runs)
- validateAlgoConfigs(binary_path, loss3_dir, loss3_configs, gold_acc, 3.0, runs)
-
-
-
-
-
-
-def getStatisticalConfidence(binary_path, layer_swings, \
- gold_accuracy, accuracy_slack, \
- total_runs, abort_after):
-
-
- target_acc = gold_accuracy - accuracy_slack
-
- # Write to promise_flags
- fout = open("promise_flags", "w+")
- for swing in layer_swings:
- int_swing = int(swing)
- if int_swing > 0:
- fout.write(str(swing) + "\n")
- fout.close()
-
- extra_args = str(total_runs) + " " + str(target_acc) + " " + str(abort_after)
- # Execute Validation Run
- p = subprocess.Popen("./" + binary_path + " " + extra_args, shell=True)
- p.wait()
-
- f = open("run_accuracies.txt")
- index = 0.0
- unsuccessful = 0.0
- sum_acc = 0.0
- for x in f:
- x = x.strip()
- acc = float(x)
- if acc < target_acc:
- unsuccessful += 1
- index += 1
- sum_acc += acc
-
- f.close()
-
- confidence = ( (index - unsuccessful) / index) * 100.0
- avg_acc = sum_acc / index
- print ("run_confidence = ", confidence, " avg_acc = ", avg_acc)
-
- return avg_acc, confidence
-
-
-
-
-
-
-
-
-if __name__ == "__main__" :
-
-
- getStatisticalConfidence("lenet_promise", [7, 1, 1, 1], 99.0, 2, 50, 5)
-
--
GitLab