diff --git a/llvm/projects/hpvm-tensor-rt/PPoPP_results/knobs.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/knobs.txt new file mode 100644 index 0000000000000000000000000000000000000000..1be644441769e8544901010586bc9842d8b14289 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/knobs.txt @@ -0,0 +1,28 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..f3e1be03b607bcf404a6cb809f1e231d497b19c6 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_120.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_151.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_151.txt new file mode 100644 index 0000000000000000000000000000000000000000..891ef4648247e6a7879f0a8c974b3b9b59193105 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_151.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_152.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_152.txt new file mode 100644 index 0000000000000000000000000000000000000000..0e1dc661467bd46ffab34faf5f85ee1254068b7c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_152.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_153.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_153.txt new file mode 100644 index 0000000000000000000000000000000000000000..211011c1c8c194c7c81ec30abe02b85d03de8f95 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_153.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_154.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_154.txt new file mode 100644 index 0000000000000000000000000000000000000000..4e4718b997d2acdad6386541c328fd778daf9f92 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_154.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_155.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_155.txt new file mode 100644 index 0000000000000000000000000000000000000000..394d24cf90f4313068db4ecf05cdc388d4178799 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_155.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_156.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_156.txt new file mode 100644 index 0000000000000000000000000000000000000000..03318bf3aeeadd855ab379bfcb0db83e3adb17df --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_156.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_157.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_157.txt new file mode 100644 index 0000000000000000000000000000000000000000..492feb3022d61db674a49dde67e8f1c98f5c95dc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_157.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_158.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_158.txt new file mode 100644 index 0000000000000000000000000000000000000000..55b44395936071c1b5c03e233e3d2b1622aab333 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_158.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_159.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_159.txt new file mode 100644 index 0000000000000000000000000000000000000000..c1cd63239796ff96ce9113f5b9d0afb6a565ab5b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_159.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_160.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_160.txt new file mode 100644 index 0000000000000000000000000000000000000000..6ee9d4cd79646b783be1434b4dcfbfedf40cb4dd --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_160.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_161.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_161.txt new file mode 100644 index 0000000000000000000000000000000000000000..3b49fae064a89d59a97f8584aa73d7c48173c14b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_161.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_162.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_162.txt new file mode 100644 index 0000000000000000000000000000000000000000..e1a046ef01a51fd9d03c0f4447a47569b9f65caa --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_162.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_163.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_163.txt new file mode 100644 index 0000000000000000000000000000000000000000..ad8ef2563416b83c6a2661c5d7ec30f90cb37926 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_163.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_164.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_164.txt new file mode 100644 index 0000000000000000000000000000000000000000..3461775f35e069cfe4f597e2a0ec685e4a257c0d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_164.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_165.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_165.txt new file mode 100644 index 0000000000000000000000000000000000000000..a78c775332f9a0b6160456e92466fca7b2294a5a --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_165.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_166.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_166.txt new file mode 100644 index 0000000000000000000000000000000000000000..8c700396a9e51f8addd4db4ff402ff57424c7a50 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_166.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_167.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_167.txt new file mode 100644 index 0000000000000000000000000000000000000000..ed26859f07eb0d4c10b16b1d8a1e1df46c2700b4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_167.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_168.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_168.txt new file mode 100644 index 0000000000000000000000000000000000000000..b9ea435a191c3e97278a0de963a378bf0b41a587 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_perf_fp16_168.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_261.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_261.txt new file mode 100644 index 0000000000000000000000000000000000000000..4ceca9d3b4337e38f2505ded41e53696bce2a3d7 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_261.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_262.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_262.txt new file mode 100644 index 0000000000000000000000000000000000000000..594b3c6cb4546672a49d7d15cfaef6e7cdedd2cc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_262.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_263.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_263.txt new file mode 100644 index 0000000000000000000000000000000000000000..daac5adf34f1af981d7170b5b9405b7bc3acda54 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_263.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_264.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_264.txt new file mode 100644 index 0000000000000000000000000000000000000000..8e9d08eb04228615230676bcf297745e1309040b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_264.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_265.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_265.txt new file mode 100644 index 0000000000000000000000000000000000000000..a9a49265efa868fba12d672e22553fac1245d9fa --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_265.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_266.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_266.txt new file mode 100644 index 0000000000000000000000000000000000000000..17d596384117b149418ce2f2e7588b660aad2328 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_266.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_267.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_267.txt new file mode 100644 index 0000000000000000000000000000000000000000..2a49166de84cb5d607169c5703c2acd1173b2d56 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_267.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_268.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_268.txt new file mode 100644 index 0000000000000000000000000000000000000000..7ffc46b3fa02f3f4cd25aeb7fe2bfece3a6d13f0 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_268.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_269.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_269.txt new file mode 100644 index 0000000000000000000000000000000000000000..dc9228cefaba6b9fa024a854fbd24b1228b144c8 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp16_samp_fp16_269.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp32_perf_fp32_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp32_perf_fp32_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..46c04e6d4c91ff5a1d4f0c98028c6c76a1710a81 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet2_cifar10/alexnet2_cifar10_fp32_perf_fp32_120.txt @@ -0,0 +1,72 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..a9c2b75d063b189adac655c77df93f5dc29b5aa7 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_120.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_151.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_151.txt new file mode 100644 index 0000000000000000000000000000000000000000..0a3f9b79c853297a54edee22f7e34296a3025c2b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_151.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_152.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_152.txt new file mode 100644 index 0000000000000000000000000000000000000000..0e370a50e6ec6468276d597bed6382a367d34133 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_152.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_153.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_153.txt new file mode 100644 index 0000000000000000000000000000000000000000..2f4fe3fc94397ef75b2b4543fcd72d62a251c539 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_153.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_154.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_154.txt new file mode 100644 index 0000000000000000000000000000000000000000..765c4ab9e3829e9793ce7f48b07f997cdbec7f75 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_154.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_155.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_155.txt new file mode 100644 index 0000000000000000000000000000000000000000..495b8e72ce969009a8a1098cffb1b025ff011728 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_155.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_156.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_156.txt new file mode 100644 index 0000000000000000000000000000000000000000..f41af92f3cfc1a6067663065f53738c15b0bf3f0 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_156.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_157.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_157.txt new file mode 100644 index 0000000000000000000000000000000000000000..46cb2c584851b98944162e561e68c378332f13db --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_157.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_158.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_158.txt new file mode 100644 index 0000000000000000000000000000000000000000..ef02807f8f589787ebf2adc839e9410c16414b53 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_158.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_159.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_159.txt new file mode 100644 index 0000000000000000000000000000000000000000..a1dfc26c51f2c81cdbe8af9f3aecc47b1280e1e1 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_159.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_160.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_160.txt new file mode 100644 index 0000000000000000000000000000000000000000..71ac47720756a629f3e14d55cc444a87230c5ba9 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_160.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_161.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_161.txt new file mode 100644 index 0000000000000000000000000000000000000000..71f39ee42f050dc07c92f82ad2155db1c0cc3fdb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_161.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_162.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_162.txt new file mode 100644 index 0000000000000000000000000000000000000000..d5b585eb367a1de62f05f02e78da08c0809625d3 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_162.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_163.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_163.txt new file mode 100644 index 0000000000000000000000000000000000000000..51a8fab32a0d8ff66be1c3027ba4a4cd98dae942 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_163.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_164.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_164.txt new file mode 100644 index 0000000000000000000000000000000000000000..29db7c0b54da6aa7d0431c32bb46542478491427 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_164.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_165.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_165.txt new file mode 100644 index 0000000000000000000000000000000000000000..3792531eb63ede8765add69a80cfe64f2ad54ff6 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_165.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_166.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_166.txt new file mode 100644 index 0000000000000000000000000000000000000000..4e7dee2df5d296223710824ae91ab705e879fc0b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_166.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_167.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_167.txt new file mode 100644 index 0000000000000000000000000000000000000000..8f48476be1a78e1f69dc095ba7f190f736b9f4c9 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_167.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_168.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_168.txt new file mode 100644 index 0000000000000000000000000000000000000000..c2573a11507e7b4f9218acc2d95ac5d9a87cf47b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_perf_fp16_168.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_42.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_42.txt new file mode 100644 index 0000000000000000000000000000000000000000..e8c305acc3d9e02c2cedb41cee945de9aecf6004 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_42.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_44.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_44.txt new file mode 100644 index 0000000000000000000000000000000000000000..3dfc825f8558ace8c058e5a70880f9a4b487213d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_44.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_46.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_46.txt new file mode 100644 index 0000000000000000000000000000000000000000..587a4bf63ff3e93ffa29ba2b34fd8ecda3453630 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_red_samp_fp16_46.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_261.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_261.txt new file mode 100644 index 0000000000000000000000000000000000000000..7cbbe3bf580489cc033d2b9caec9453143ca51fc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_261.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_262.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_262.txt new file mode 100644 index 0000000000000000000000000000000000000000..b8ce558ccafbc02f7f2875371c2c8861165c9188 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_262.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_263.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_263.txt new file mode 100644 index 0000000000000000000000000000000000000000..dec9f13da9526d4483267d444aacf594f1c70259 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_263.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_264.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_264.txt new file mode 100644 index 0000000000000000000000000000000000000000..2445a2271cd99145bd0e0a5e10364288a28f8987 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_264.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_265.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_265.txt new file mode 100644 index 0000000000000000000000000000000000000000..62d8f0cee8674e21d9597806cf296c9540c428e2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_265.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_266.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_266.txt new file mode 100644 index 0000000000000000000000000000000000000000..ecf5ecb4c08a80289cc115aad2e27074f9481b45 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_266.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_267.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_267.txt new file mode 100644 index 0000000000000000000000000000000000000000..5ad85fb3de959b2b816dd24d066a50b85e831895 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_267.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_268.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_268.txt new file mode 100644 index 0000000000000000000000000000000000000000..c0755261b91bf23e68451ead7a878f72a21f1839 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_268.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_269.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_269.txt new file mode 100644 index 0000000000000000000000000000000000000000..7786d5550a860795cc7a9af0571536c8f002fb0b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp16_samp_fp16_269.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp32_perf_fp32_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp32_perf_fp32_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..f00d5f1ed2fbbe2687b34d1056b0e8f710c876c6 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_fp32_perf_fp32_120.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_121.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_121.txt new file mode 100644 index 0000000000000000000000000000000000000000..507ccfb18484d8336db7f85b3ff3f53e610508dc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_121.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_122.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_122.txt new file mode 100644 index 0000000000000000000000000000000000000000..a88dd7e64b895743978a7dc30a73a0dbc7aea168 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_122.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_123.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_123.txt new file mode 100644 index 0000000000000000000000000000000000000000..f625708f35b3d10af1ac943760493f463cac08c1 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_123.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_124.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_124.txt new file mode 100644 index 0000000000000000000000000000000000000000..f4a26c1fd504dbd5da22ca2e15282635494674e4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_124.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_125.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_125.txt new file mode 100644 index 0000000000000000000000000000000000000000..7804c699e58b893ba877de2f4fa22972df010d24 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_125.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_126.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_126.txt new file mode 100644 index 0000000000000000000000000000000000000000..ea6400d03aa3ad71cac15db58e32481da2483845 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_126.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_127.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_127.txt new file mode 100644 index 0000000000000000000000000000000000000000..11dacf885883227c5be90ef5b5ad89f123c6482c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_127.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_128.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_128.txt new file mode 100644 index 0000000000000000000000000000000000000000..cf76f32f4ceb561231d6ed75a16ba374bb568cdc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_128.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_129.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_129.txt new file mode 100644 index 0000000000000000000000000000000000000000..a6f28a7c602e85be94e5f27feb7a652c82c483c8 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_129.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_130.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_130.txt new file mode 100644 index 0000000000000000000000000000000000000000..636e1870792259d4903ac8708a3c06e07c5ba4c0 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_130.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_131.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_131.txt new file mode 100644 index 0000000000000000000000000000000000000000..50c3eb9ec42d079ea23479abe45ca801abe3360d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_131.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_132.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_132.txt new file mode 100644 index 0000000000000000000000000000000000000000..d2c262d04f052a8a164878434bea1ef52da3b321 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_132.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_133.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_133.txt new file mode 100644 index 0000000000000000000000000000000000000000..e8443470d82519ee6382e895f49dddd6ad947205 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_133.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_134.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_134.txt new file mode 100644 index 0000000000000000000000000000000000000000..f156a272bc0eca63a6594e27aa836509f288799c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_134.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_135.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_135.txt new file mode 100644 index 0000000000000000000000000000000000000000..d93aac33b8164425eb51d4dc814e2e054f163ae5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_135.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_136.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_136.txt new file mode 100644 index 0000000000000000000000000000000000000000..0b546e90369714ce255fdcfdf59eaaa4ad1cfbf9 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_136.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_137.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_137.txt new file mode 100644 index 0000000000000000000000000000000000000000..18601a156bdc8c142870fd2d9555bd3ee59032bc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_137.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_138.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_138.txt new file mode 100644 index 0000000000000000000000000000000000000000..18e2ee76b7ca35b0a962cc80809c02903390cddb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_perf_138.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_41.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_41.txt new file mode 100644 index 0000000000000000000000000000000000000000..070e97bed17bb07898d33e45d7ea90d12f52461f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_41.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_43.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_43.txt new file mode 100644 index 0000000000000000000000000000000000000000..adee4a649a5230b37c713106248be75c1d031a52 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_43.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_45.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_45.txt new file mode 100644 index 0000000000000000000000000000000000000000..151d408eed7b2ed7f739d5916eac202a3cf2fec2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_red_samp_45.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_231.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_231.txt new file mode 100644 index 0000000000000000000000000000000000000000..b1900b41c2d7c78e6b425d7b3e2f52f8c3f08606 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_231.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_232.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_232.txt new file mode 100644 index 0000000000000000000000000000000000000000..acea102c80b68b7a0c693135d1a2bd9dc6dbb418 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_232.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_233.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_233.txt new file mode 100644 index 0000000000000000000000000000000000000000..47aa85aa79ce2df08664ec1ae85b31013416baf5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_233.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_234.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_234.txt new file mode 100644 index 0000000000000000000000000000000000000000..19f49e9d486a58accb45ef9903d341e8b9585c09 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_234.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_235.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_235.txt new file mode 100644 index 0000000000000000000000000000000000000000..f5857d3264f277810e2ca7bd7bde7cfd7f1d46cd --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_235.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_236.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_236.txt new file mode 100644 index 0000000000000000000000000000000000000000..b38950da77e2c0ae11b36cc570e8a3c64b7c100b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_236.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_237.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_237.txt new file mode 100644 index 0000000000000000000000000000000000000000..3c5e6c169a08069d304782a66cf2b965c55e2931 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_237.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_238.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_238.txt new file mode 100644 index 0000000000000000000000000000000000000000..dbc124748794072b323ece1525a773ec28ea633e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_238.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_239.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_239.txt new file mode 100644 index 0000000000000000000000000000000000000000..cc13932a92b884c0e4a74dda6d2bd25edbadd7bb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp16/alexnet_canny_samp_239.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..f00d5f1ed2fbbe2687b34d1056b0e8f710c876c6 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_120.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_121.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_121.txt new file mode 100644 index 0000000000000000000000000000000000000000..507ccfb18484d8336db7f85b3ff3f53e610508dc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_121.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_122.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_122.txt new file mode 100644 index 0000000000000000000000000000000000000000..a88dd7e64b895743978a7dc30a73a0dbc7aea168 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_122.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_123.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_123.txt new file mode 100644 index 0000000000000000000000000000000000000000..f625708f35b3d10af1ac943760493f463cac08c1 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_123.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_124.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_124.txt new file mode 100644 index 0000000000000000000000000000000000000000..f4a26c1fd504dbd5da22ca2e15282635494674e4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_124.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_125.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_125.txt new file mode 100644 index 0000000000000000000000000000000000000000..7804c699e58b893ba877de2f4fa22972df010d24 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_125.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_126.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_126.txt new file mode 100644 index 0000000000000000000000000000000000000000..ea6400d03aa3ad71cac15db58e32481da2483845 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_126.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_127.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_127.txt new file mode 100644 index 0000000000000000000000000000000000000000..11dacf885883227c5be90ef5b5ad89f123c6482c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_127.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_128.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_128.txt new file mode 100644 index 0000000000000000000000000000000000000000..cf76f32f4ceb561231d6ed75a16ba374bb568cdc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_128.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_129.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_129.txt new file mode 100644 index 0000000000000000000000000000000000000000..a6f28a7c602e85be94e5f27feb7a652c82c483c8 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_129.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_130.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_130.txt new file mode 100644 index 0000000000000000000000000000000000000000..636e1870792259d4903ac8708a3c06e07c5ba4c0 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_130.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_131.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_131.txt new file mode 100644 index 0000000000000000000000000000000000000000..50c3eb9ec42d079ea23479abe45ca801abe3360d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_131.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_132.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_132.txt new file mode 100644 index 0000000000000000000000000000000000000000..d2c262d04f052a8a164878434bea1ef52da3b321 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_132.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_133.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_133.txt new file mode 100644 index 0000000000000000000000000000000000000000..e8443470d82519ee6382e895f49dddd6ad947205 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_133.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_134.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_134.txt new file mode 100644 index 0000000000000000000000000000000000000000..f156a272bc0eca63a6594e27aa836509f288799c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_134.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_135.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_135.txt new file mode 100644 index 0000000000000000000000000000000000000000..d93aac33b8164425eb51d4dc814e2e054f163ae5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_135.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_136.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_136.txt new file mode 100644 index 0000000000000000000000000000000000000000..0b546e90369714ce255fdcfdf59eaaa4ad1cfbf9 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_136.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_137.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_137.txt new file mode 100644 index 0000000000000000000000000000000000000000..18601a156bdc8c142870fd2d9555bd3ee59032bc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_137.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_138.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_138.txt new file mode 100644 index 0000000000000000000000000000000000000000..18e2ee76b7ca35b0a962cc80809c02903390cddb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_perf_138.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_41.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_41.txt new file mode 100644 index 0000000000000000000000000000000000000000..070e97bed17bb07898d33e45d7ea90d12f52461f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_41.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_43.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_43.txt new file mode 100644 index 0000000000000000000000000000000000000000..adee4a649a5230b37c713106248be75c1d031a52 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_43.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_45.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_45.txt new file mode 100644 index 0000000000000000000000000000000000000000..151d408eed7b2ed7f739d5916eac202a3cf2fec2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_red_samp_45.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_231.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_231.txt new file mode 100644 index 0000000000000000000000000000000000000000..b1900b41c2d7c78e6b425d7b3e2f52f8c3f08606 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_231.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_232.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_232.txt new file mode 100644 index 0000000000000000000000000000000000000000..acea102c80b68b7a0c693135d1a2bd9dc6dbb418 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_232.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_233.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_233.txt new file mode 100644 index 0000000000000000000000000000000000000000..47aa85aa79ce2df08664ec1ae85b31013416baf5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_233.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_234.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_234.txt new file mode 100644 index 0000000000000000000000000000000000000000..19f49e9d486a58accb45ef9903d341e8b9585c09 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_234.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_235.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_235.txt new file mode 100644 index 0000000000000000000000000000000000000000..f5857d3264f277810e2ca7bd7bde7cfd7f1d46cd --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_235.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_236.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_236.txt new file mode 100644 index 0000000000000000000000000000000000000000..b38950da77e2c0ae11b36cc570e8a3c64b7c100b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_236.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_237.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_237.txt new file mode 100644 index 0000000000000000000000000000000000000000..3c5e6c169a08069d304782a66cf2b965c55e2931 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_237.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_238.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_238.txt new file mode 100644 index 0000000000000000000000000000000000000000..dbc124748794072b323ece1525a773ec28ea633e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_238.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_239.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_239.txt new file mode 100644 index 0000000000000000000000000000000000000000..cc13932a92b884c0e4a74dda6d2bd25edbadd7bb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_canny_fp32/alexnet_canny_samp_239.txt @@ -0,0 +1,108 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..ff28fccb9e6a9464df2a3ecf2e0f2f450508f710 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_120.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_151.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_151.txt new file mode 100644 index 0000000000000000000000000000000000000000..bae151beca45f4cfd85d2909d09707078928d782 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_151.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_152.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_152.txt new file mode 100644 index 0000000000000000000000000000000000000000..a6d85b7e69a0a8d6017a20fae3eae2a872189fd8 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_152.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_153.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_153.txt new file mode 100644 index 0000000000000000000000000000000000000000..efc436221c5a779f64b4c606abce1c3c410a66a7 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_153.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_154.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_154.txt new file mode 100644 index 0000000000000000000000000000000000000000..00f06e6d161b2f4b4fc21a7bc9faaceb09188007 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_154.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_155.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_155.txt new file mode 100644 index 0000000000000000000000000000000000000000..5fd4c898cb169b014a03f0fde202e0ca6199b0a7 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_155.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_156.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_156.txt new file mode 100644 index 0000000000000000000000000000000000000000..bbc1088bcd8bf1c07de1df981e03cd6095171473 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_156.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_157.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_157.txt new file mode 100644 index 0000000000000000000000000000000000000000..cec1b029906bdced778e41c051734df7e8b978dc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_157.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_158.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_158.txt new file mode 100644 index 0000000000000000000000000000000000000000..65dd44383241454da0cd35e1684a2d6dadb29c8d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_158.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_159.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_159.txt new file mode 100644 index 0000000000000000000000000000000000000000..bcad05006187854202d93e2e3daf47381f0c18e8 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_159.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_160.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_160.txt new file mode 100644 index 0000000000000000000000000000000000000000..fa9616a04c7532529b9ee6e83a8e38dab29bbee2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_160.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_161.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_161.txt new file mode 100644 index 0000000000000000000000000000000000000000..ee0b75a4e8e8db75946bf0daa98bd8967e737ea2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_161.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_162.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_162.txt new file mode 100644 index 0000000000000000000000000000000000000000..37bce90b936e778afa83970ba465cf3ba11f8d1d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_162.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_163.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_163.txt new file mode 100644 index 0000000000000000000000000000000000000000..33ad0446b56d9654e14069157364a799cadc1984 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_163.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_164.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_164.txt new file mode 100644 index 0000000000000000000000000000000000000000..5f6fe09a6522c5d1824453f0e4ed3725d902328a --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_164.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_165.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_165.txt new file mode 100644 index 0000000000000000000000000000000000000000..2de11a32b2e083f701dee93a6685a4c9102b174b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_165.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_166.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_166.txt new file mode 100644 index 0000000000000000000000000000000000000000..a3e6094c9516db9c96c9563ba1c9cb760f249040 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_166.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_167.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_167.txt new file mode 100644 index 0000000000000000000000000000000000000000..635233830164a2bdbced47023dde8785da3fe82d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_167.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_168.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_168.txt new file mode 100644 index 0000000000000000000000000000000000000000..4939acc2b9f1b971a3aa299826af2d177df97fe1 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_perf_fp16_168.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_261.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_261.txt new file mode 100644 index 0000000000000000000000000000000000000000..d6b5d9b3b2d94898e42aa36ac620cf34d0b76c38 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_261.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_262.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_262.txt new file mode 100644 index 0000000000000000000000000000000000000000..114fd5b4019f70bc14f1e14d97de5daa68f1113e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_262.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_263.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_263.txt new file mode 100644 index 0000000000000000000000000000000000000000..38f5c465aeb55bedf12d25cb499b585b13788821 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_263.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_264.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_264.txt new file mode 100644 index 0000000000000000000000000000000000000000..dce8d97e08c0aa466e8cb19d451ab46e8ee4ae0c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_264.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_265.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_265.txt new file mode 100644 index 0000000000000000000000000000000000000000..b552df392282519064dc22347d342308a906c3a4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_265.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_266.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_266.txt new file mode 100644 index 0000000000000000000000000000000000000000..22b32b349ffcf83de0782941f16da54fff293c4b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_266.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_267.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_267.txt new file mode 100644 index 0000000000000000000000000000000000000000..078360ac09302ce001128a1ea1df997e000ad785 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_267.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_268.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_268.txt new file mode 100644 index 0000000000000000000000000000000000000000..7ef17faf3c5c17e5bc7738f82cd5a261a17345ce --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_268.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_269.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_269.txt new file mode 100644 index 0000000000000000000000000000000000000000..d149a94dd81196c2a6f9006661408e2a48125585 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp16_samp_fp16_269.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp32_perf_fp32_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp32_perf_fp32_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..25cd9de0d053a2797e8942f0ccc84be5ff9766c6 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_cifar10/alexnet_cifar10_fp32_perf_fp32_120.txt @@ -0,0 +1,63 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..391282ad11099343f49a83189f68acbb8bd1942d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_120.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_151.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_151.txt new file mode 100644 index 0000000000000000000000000000000000000000..7b36fde13801276c18018adc16ef4890b4eb4890 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_151.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_152.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_152.txt new file mode 100644 index 0000000000000000000000000000000000000000..dc32d18e93ee5ec80757389dfd4bf1a7e96cef0d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_152.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_153.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_153.txt new file mode 100644 index 0000000000000000000000000000000000000000..5838b4148e0e28bbabd3298535c42729010bc73a --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_153.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_154.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_154.txt new file mode 100644 index 0000000000000000000000000000000000000000..e285249a3496789ec0030d202e4705dde7cf6cbf --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_154.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_155.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_155.txt new file mode 100644 index 0000000000000000000000000000000000000000..5e943f57f9907e6da03e34e63553c2f844ecbccc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_155.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_156.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_156.txt new file mode 100644 index 0000000000000000000000000000000000000000..05d5be47e55158369efbdde8f14412a2b4ba1c5d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_156.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_157.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_157.txt new file mode 100644 index 0000000000000000000000000000000000000000..64684f24133f6f23242ceb1029b6c6dacf73755e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_157.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_158.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_158.txt new file mode 100644 index 0000000000000000000000000000000000000000..905a3f5e14a3c5ec89e252b13b933e1ffdb4e013 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_158.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_159.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_159.txt new file mode 100644 index 0000000000000000000000000000000000000000..32c740c4ce5a6ff9384826b23606df002e1f646d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_159.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_160.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_160.txt new file mode 100644 index 0000000000000000000000000000000000000000..87094f1c2cac516168b0d16205249a4ae613d980 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_160.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_161.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_161.txt new file mode 100644 index 0000000000000000000000000000000000000000..3a56a4bf675a3bf9948d0ef29aca0f7002725cb1 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_161.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_162.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_162.txt new file mode 100644 index 0000000000000000000000000000000000000000..da47c26710f5b8303f9535b705bf97daad8ab2f5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_162.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_163.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_163.txt new file mode 100644 index 0000000000000000000000000000000000000000..5ceaa1add158ee879b8884d085de3430cd1efcc2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_163.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_164.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_164.txt new file mode 100644 index 0000000000000000000000000000000000000000..d5d4542dee5358e1de3d286cf1ad212cffeeced7 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_164.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_165.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_165.txt new file mode 100644 index 0000000000000000000000000000000000000000..adeba7a2fea31bc1224d6dab7d5d76fa79bf1a96 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_165.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_166.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_166.txt new file mode 100644 index 0000000000000000000000000000000000000000..90a5a9bb27c1bebca6314875fa69514039e4080f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_166.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_167.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_167.txt new file mode 100644 index 0000000000000000000000000000000000000000..c3e83a0ac7288f9df2261d3bfc2eed5e5877e3ed --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_167.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_168.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_168.txt new file mode 100644 index 0000000000000000000000000000000000000000..87a8b5b753cca9b81ea1f28369712055c2902022 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_perf_fp16_168.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_261.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_261.txt new file mode 100644 index 0000000000000000000000000000000000000000..93838593b0e86eb9947bc0286ffada8c37d65bad --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_261.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_262.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_262.txt new file mode 100644 index 0000000000000000000000000000000000000000..14eb5aebdc866d3346d6865edc4dc056735a0042 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_262.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_263.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_263.txt new file mode 100644 index 0000000000000000000000000000000000000000..4d7d8bd0660e808ab0e16759a768bd1294db0947 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_263.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_264.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_264.txt new file mode 100644 index 0000000000000000000000000000000000000000..ccb144aca9ef38452537d71bd0dd800618d4fe51 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_264.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_265.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_265.txt new file mode 100644 index 0000000000000000000000000000000000000000..8af78c59c5e7b00d25afbb5d78ef5b95709e043c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_265.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_266.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_266.txt new file mode 100644 index 0000000000000000000000000000000000000000..d0e6e9a17ef4b3dfb59b8f9974ae1d8f215aae00 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_266.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_267.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_267.txt new file mode 100644 index 0000000000000000000000000000000000000000..78ebfbbbb1ea01fb72513f32b0a08423a0258e98 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_267.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_268.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_268.txt new file mode 100644 index 0000000000000000000000000000000000000000..0c2c340b509f603965d439f8617f109f2825e2e1 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_268.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_269.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_269.txt new file mode 100644 index 0000000000000000000000000000000000000000..16cdb9e385561ba7058ccc6ff9c2a33d56864f28 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp16_samp_fp16_269.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp32_perf_fp32_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp32_perf_fp32_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..649a1f51be1545a963b1c6fb717178a837fc701e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/alexnet_imagenet/alexnet_imagenet_fp32_perf_fp32_120.txt @@ -0,0 +1,81 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..ab5123cc67a86fc4a620d0a8bcca343e923258dc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_120.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_151.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_151.txt new file mode 100644 index 0000000000000000000000000000000000000000..f61512071680cf0864c0c16a8e7b21a8a6aee39c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_151.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_152.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_152.txt new file mode 100644 index 0000000000000000000000000000000000000000..a3c51be3cffa7c274431e82fad78197aa6d1836f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_152.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_153.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_153.txt new file mode 100644 index 0000000000000000000000000000000000000000..2dd6e9a88a82649c5b2ab418e598432bec261bf6 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_153.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_154.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_154.txt new file mode 100644 index 0000000000000000000000000000000000000000..389e60070dd2731939dde771502b55c5a437486f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_154.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_155.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_155.txt new file mode 100644 index 0000000000000000000000000000000000000000..ca03841e4af51b8e30ce93734eda3674e7eca805 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_155.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_156.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_156.txt new file mode 100644 index 0000000000000000000000000000000000000000..1f4c7f682568d43d7e4708052ff22280e1198057 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_156.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_157.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_157.txt new file mode 100644 index 0000000000000000000000000000000000000000..4abd2ff1d6abfa5a23c9909030d16f915d6d44c6 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_157.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_158.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_158.txt new file mode 100644 index 0000000000000000000000000000000000000000..ae604d7bdd9f6e85d7a77e9c39d29259b0bdfdff --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_158.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_159.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_159.txt new file mode 100644 index 0000000000000000000000000000000000000000..955744472efb7e677ce14f28f589aa25087e4766 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_159.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_160.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_160.txt new file mode 100644 index 0000000000000000000000000000000000000000..5f142c7920ec552c8b3f8c1a872604b55b6e63b1 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_160.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_161.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_161.txt new file mode 100644 index 0000000000000000000000000000000000000000..092b6f5d79acc04e4622f26006ed1a8de3d1ccea --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_161.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_162.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_162.txt new file mode 100644 index 0000000000000000000000000000000000000000..3ca4277de1d0258709f4d44faf64d9e59937a90f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_162.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_163.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_163.txt new file mode 100644 index 0000000000000000000000000000000000000000..dbfbf37ebc34a0e18fe7b33d296b0e60811e87d5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_163.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_164.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_164.txt new file mode 100644 index 0000000000000000000000000000000000000000..99b2446d3c9a86427679948b7869c3ce4cd645c7 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_164.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_165.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_165.txt new file mode 100644 index 0000000000000000000000000000000000000000..e1c00548bba6f5bf5cd989f354f6855765119752 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_165.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_166.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_166.txt new file mode 100644 index 0000000000000000000000000000000000000000..d0c78df48e4b19fbd1da745db9860b75fb3d81c5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_166.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_167.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_167.txt new file mode 100644 index 0000000000000000000000000000000000000000..4582f6b78296b5f0f5db7168da8bc20a7c8b57a8 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_167.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_168.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_168.txt new file mode 100644 index 0000000000000000000000000000000000000000..55d8d4862bceea092624fef4be804238cafe2d6e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_perf_fp16_168.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_261.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_261.txt new file mode 100644 index 0000000000000000000000000000000000000000..ceeccbe71a1ee4bc7891d2898389c43ffaaae102 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_261.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_262.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_262.txt new file mode 100644 index 0000000000000000000000000000000000000000..ae7893234b4a854afdf2ab6a93804b472d7bde8d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_262.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_263.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_263.txt new file mode 100644 index 0000000000000000000000000000000000000000..bd428223fa0b68dc3eb8d0ec94c2abd03e912b98 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_263.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_264.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_264.txt new file mode 100644 index 0000000000000000000000000000000000000000..f6809874580e3592b06fd43812ddf02a388428b2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_264.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_265.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_265.txt new file mode 100644 index 0000000000000000000000000000000000000000..525e1c1bceab761098398811fef3d6f67c8084d8 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_265.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_266.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_266.txt new file mode 100644 index 0000000000000000000000000000000000000000..a7e4b208e395f44cf0a89ecf6755b04b22a0705e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_266.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_267.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_267.txt new file mode 100644 index 0000000000000000000000000000000000000000..3948a143506d5d0eaeabd66c1d60b04fd449f36b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_267.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_268.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_268.txt new file mode 100644 index 0000000000000000000000000000000000000000..2e3cfc46750f56cafde4c30b1b97d33a7ee5d2a1 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_268.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_269.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_269.txt new file mode 100644 index 0000000000000000000000000000000000000000..1910111015a911e4af5f0173caeee0686320fee5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp16_samp_fp16_269.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp32_perf_fp32_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp32_perf_fp32_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..df787165b74a3bb97015b4f0bb9c7afa56c5c6d8 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/lenet_keras/lenet_keras_fp32_perf_fp32_120.txt @@ -0,0 +1,45 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..5d79cb1b4bd9d9c23a307f85b2d265712c3672cb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_120.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_151.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_151.txt new file mode 100644 index 0000000000000000000000000000000000000000..dc27bd9b0cefcba92a6e611bbd04e7abdcf043e4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_151.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_152.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_152.txt new file mode 100644 index 0000000000000000000000000000000000000000..8964f1786a46b940e964e341614251865b47d4e2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_152.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_153.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_153.txt new file mode 100644 index 0000000000000000000000000000000000000000..39c3c4bfc3c45ad536dd32d24c599bbfa1eca85b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_153.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_154.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_154.txt new file mode 100644 index 0000000000000000000000000000000000000000..af20d986d021bcf29c10ad5a89faede4cf16115b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_154.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_155.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_155.txt new file mode 100644 index 0000000000000000000000000000000000000000..3cd198c77ec69951041ad8d1e114a5a5df5072c8 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_155.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_156.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_156.txt new file mode 100644 index 0000000000000000000000000000000000000000..8019eb1b69a25d2c87a036fe7cac6864eaecedea --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_156.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_157.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_157.txt new file mode 100644 index 0000000000000000000000000000000000000000..d8489f200d082fa1d51b5c5d12c1827951a4c4b3 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_157.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_158.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_158.txt new file mode 100644 index 0000000000000000000000000000000000000000..90ffe2f12a60802fc29a7089ae60cd65af0bf6dc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_158.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_159.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_159.txt new file mode 100644 index 0000000000000000000000000000000000000000..f516e0ae4c0a2d7f01b2ed15147ad8d2d14a6eff --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_159.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_160.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_160.txt new file mode 100644 index 0000000000000000000000000000000000000000..c5475d1abb9ac9341738791863861927e48b615e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_160.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_161.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_161.txt new file mode 100644 index 0000000000000000000000000000000000000000..ec4bf136fd09893c177a6bf1002292e43a45cb36 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_161.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_162.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_162.txt new file mode 100644 index 0000000000000000000000000000000000000000..02527d21f3d48f8563ff990061c3c0e2e95f794a --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_162.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_163.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_163.txt new file mode 100644 index 0000000000000000000000000000000000000000..bfaefe1f93f0c06c121c2d40d7a95c0bec41c7d4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_163.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_164.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_164.txt new file mode 100644 index 0000000000000000000000000000000000000000..82957f7e574b11d69bfd9f4026c65dc9dade14e1 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_164.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_165.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_165.txt new file mode 100644 index 0000000000000000000000000000000000000000..cd027ffd1de79aaa5146212bd23f0020c997acc9 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_165.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_166.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_166.txt new file mode 100644 index 0000000000000000000000000000000000000000..0a1f2f2a19baf9ac339ff93c32462d20bec705ce --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_166.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_167.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_167.txt new file mode 100644 index 0000000000000000000000000000000000000000..1b522abbf0d1936a04e32f08c37c4a49a59702b9 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_167.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_168.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_168.txt new file mode 100644 index 0000000000000000000000000000000000000000..feaab7025af2fa86071c5a370adbdaa04b0154c4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_perf_fp16_168.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_261.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_261.txt new file mode 100644 index 0000000000000000000000000000000000000000..5764c923a5d3282cdcb947fdb7362031fdbc155a --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_261.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_262.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_262.txt new file mode 100644 index 0000000000000000000000000000000000000000..9d6d32c0c29b32303dc7183f7bda8f31d6325deb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_262.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_263.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_263.txt new file mode 100644 index 0000000000000000000000000000000000000000..5bd59d30a06f2727acb186f27e93fe96da2ddf1b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_263.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_264.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_264.txt new file mode 100644 index 0000000000000000000000000000000000000000..6ae91b073a0abf842fa52c2091f88f9416c66b21 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_264.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_265.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_265.txt new file mode 100644 index 0000000000000000000000000000000000000000..c74cb8bdba74b5c918d855de74d436dbb2277f5b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_265.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_266.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_266.txt new file mode 100644 index 0000000000000000000000000000000000000000..6c3baca1485041072a6b52c3b76f0cf7a79130b3 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_266.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_267.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_267.txt new file mode 100644 index 0000000000000000000000000000000000000000..e9b0012ca0bcc9f50fc86cc2e7d51e48ae00e410 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_267.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_268.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_268.txt new file mode 100644 index 0000000000000000000000000000000000000000..805f15081a124b40c4b8269fe8ccd2fa1f5eab6a --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_268.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_269.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_269.txt new file mode 100644 index 0000000000000000000000000000000000000000..f64ae15dfc2b9bbe95895df76765388b57c3bde4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp16_samp_fp16_269.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp32_perf_fp32_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp32_perf_fp32_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..a32a995a8279896b28f6a55ddebb630edc916ba4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/mobilenet_cifar10/mobilenet_cifar10_fp32_perf_fp32_120.txt @@ -0,0 +1,255 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..e5768fe5ea99f84981077939f7b7946785db1f65 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_120.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_151.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_151.txt new file mode 100644 index 0000000000000000000000000000000000000000..786d8dbd8e2975e5dcf200279de9ad4f5ee9d9d2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_151.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_152.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_152.txt new file mode 100644 index 0000000000000000000000000000000000000000..600dc88cbedd2c88a6cb6cfa2f84bda1e54ebe7e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_152.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_153.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_153.txt new file mode 100644 index 0000000000000000000000000000000000000000..a8c00d45766dd61abbccb3447c9843273e812b23 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_153.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_154.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_154.txt new file mode 100644 index 0000000000000000000000000000000000000000..6480cd1c9857412d9edfc4452aa8c9ade5d16561 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_154.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_155.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_155.txt new file mode 100644 index 0000000000000000000000000000000000000000..516dfb5b2e62961dae843a98e0d497c345ee5f04 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_155.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_156.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_156.txt new file mode 100644 index 0000000000000000000000000000000000000000..1f5bb7ca4479bd2929694021aa192e83bae5df94 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_156.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_157.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_157.txt new file mode 100644 index 0000000000000000000000000000000000000000..17b884ece43f34f3a1adc72ca028d28e06277f69 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_157.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_158.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_158.txt new file mode 100644 index 0000000000000000000000000000000000000000..e98ad1dac3c713b97c23585b7ba1b320e4d4bc94 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_158.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_159.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_159.txt new file mode 100644 index 0000000000000000000000000000000000000000..84b449752cbe908ff0271b75f6a0bfe8f2cae235 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_159.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_160.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_160.txt new file mode 100644 index 0000000000000000000000000000000000000000..ba5a39afb52f7a01ac475b342a5d6868d04fb006 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_160.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_161.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_161.txt new file mode 100644 index 0000000000000000000000000000000000000000..84970f1fe409bfade732399b96dcca33f74287d5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_161.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_162.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_162.txt new file mode 100644 index 0000000000000000000000000000000000000000..2f1f3b1e5436717e32c0807e70f2d21972b753a4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_162.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_163.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_163.txt new file mode 100644 index 0000000000000000000000000000000000000000..cd96db7497bcf1ae4672492cd0c9b31424348439 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_163.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_164.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_164.txt new file mode 100644 index 0000000000000000000000000000000000000000..7f3da6f1d981d3f5ad957abc0318ae1ece976cf9 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_164.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_165.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_165.txt new file mode 100644 index 0000000000000000000000000000000000000000..f535f820c943898d9f59a276eab1f036f793a172 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_165.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_166.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_166.txt new file mode 100644 index 0000000000000000000000000000000000000000..4568dca5cc198e32be6a44e7ef13e0a7a0b089e7 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_166.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_167.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_167.txt new file mode 100644 index 0000000000000000000000000000000000000000..3a5e9b50bc4f7e1d36243820a746abfff29e080d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_167.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_168.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_168.txt new file mode 100644 index 0000000000000000000000000000000000000000..80dc433310b114fa54345faac279f4532cf8b4c8 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_perf_fp16_168.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_261.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_261.txt new file mode 100644 index 0000000000000000000000000000000000000000..6caeadddfdd40b8638bb3ef555563ea32760ed07 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_261.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_262.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_262.txt new file mode 100644 index 0000000000000000000000000000000000000000..9194ee12f9f9d142886a9915e8705f649b0f80bd --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_262.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_263.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_263.txt new file mode 100644 index 0000000000000000000000000000000000000000..c5025fd8696135af2d5ddcd46afef3c19fa2cfaa --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_263.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_264.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_264.txt new file mode 100644 index 0000000000000000000000000000000000000000..951b8c465bdacd5d15f980305c97861b76160712 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_264.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_265.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_265.txt new file mode 100644 index 0000000000000000000000000000000000000000..49017013dc5f934f9f238c0a687f8cadf1f64283 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_265.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_266.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_266.txt new file mode 100644 index 0000000000000000000000000000000000000000..595da5c025a8b2d21a8859f7b81759e0b119fa09 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_266.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_267.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_267.txt new file mode 100644 index 0000000000000000000000000000000000000000..df31c8040e2bee945a9f413ea6c2fe1bab857755 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_267.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_268.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_268.txt new file mode 100644 index 0000000000000000000000000000000000000000..0fdf2c3e3f76b6851c39755e368bd4ac9013fbda --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_268.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_269.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_269.txt new file mode 100644 index 0000000000000000000000000000000000000000..b7d2ecb6a5df940372f102775172e8c560ffd7c4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp16_samp_fp16_269.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp32_perf_fp32_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp32_perf_fp32_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..2323ecc3dacbed4c5b302a388e99a12e7836a20b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet18_cifar10/resnet18_cifar10_fp32_perf_fp32_120.txt @@ -0,0 +1,222 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..76d15fea328a534e3c1a7e3ea2c179996dbdea76 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_120.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_151.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_151.txt new file mode 100644 index 0000000000000000000000000000000000000000..17423a25717e5f1ef4224d5c5c1fdf2d49e757fd --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_151.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_152.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_152.txt new file mode 100644 index 0000000000000000000000000000000000000000..65b63cbff6ee7eb4de969593d130343c438ab6d6 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_152.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_153.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_153.txt new file mode 100644 index 0000000000000000000000000000000000000000..5b0c6da1e9835f283fdb03412acded28d054a5e6 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_153.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_154.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_154.txt new file mode 100644 index 0000000000000000000000000000000000000000..067c920d312c59f9744c6b44995cba3dc9da1698 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_154.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_155.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_155.txt new file mode 100644 index 0000000000000000000000000000000000000000..37a7257c20638a86157576ba9927b5aaa0c3ffbe --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_155.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_156.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_156.txt new file mode 100644 index 0000000000000000000000000000000000000000..f9f71a75a2e78e34be922ae773bfa9c791c2a3c4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_156.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_157.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_157.txt new file mode 100644 index 0000000000000000000000000000000000000000..d83b30ed36748e31285c98b911c2d0f051f16b0b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_157.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_158.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_158.txt new file mode 100644 index 0000000000000000000000000000000000000000..e2b008c23901b088b1280c581a89c8234995fc71 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_158.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_159.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_159.txt new file mode 100644 index 0000000000000000000000000000000000000000..71d4f29c81dff8609f20a5f860737044d1f35efa --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_159.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_160.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_160.txt new file mode 100644 index 0000000000000000000000000000000000000000..bd610058084cfde81ae8bd5522f782c71fc495aa --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_160.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_161.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_161.txt new file mode 100644 index 0000000000000000000000000000000000000000..393a1d43f9c935e9e0c166595c3a969961908f2f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_161.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_162.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_162.txt new file mode 100644 index 0000000000000000000000000000000000000000..fb4626bdf9e07cf25b1ca67575ea1754dc11292e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_162.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_163.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_163.txt new file mode 100644 index 0000000000000000000000000000000000000000..aeaff59555bf4dd4488057ebe107d423e452a728 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_163.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_164.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_164.txt new file mode 100644 index 0000000000000000000000000000000000000000..2d2d83329218d03a220a795739867f212389b0ce --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_164.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_165.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_165.txt new file mode 100644 index 0000000000000000000000000000000000000000..1b536588b3a9bb1d3c390b99f8694ba3f35a9114 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_165.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_166.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_166.txt new file mode 100644 index 0000000000000000000000000000000000000000..36738891fb7bdaf914c3df108c203d821697398e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_166.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_167.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_167.txt new file mode 100644 index 0000000000000000000000000000000000000000..023a48c70f216caab996a5f7754998e9daee720f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_167.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_168.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_168.txt new file mode 100644 index 0000000000000000000000000000000000000000..22ccbe9e1149bf3c0fbf39844e932b17acfcb7cb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_perf_fp16_168.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_261.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_261.txt new file mode 100644 index 0000000000000000000000000000000000000000..45af9a24777fe9eaecf431e5a0d510e295491e31 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_261.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_262.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_262.txt new file mode 100644 index 0000000000000000000000000000000000000000..afd86204a851f474212cc8a01710567f273029fc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_262.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_263.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_263.txt new file mode 100644 index 0000000000000000000000000000000000000000..49071a42cb0be1bbfc0c55e8c370b45addc02c69 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_263.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_264.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_264.txt new file mode 100644 index 0000000000000000000000000000000000000000..ed370c72a35a37cf5558de4eb8e00f30a5d396d0 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_264.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_265.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_265.txt new file mode 100644 index 0000000000000000000000000000000000000000..50304795f436db23f1b47f41b544d664fa71eac9 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_265.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_266.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_266.txt new file mode 100644 index 0000000000000000000000000000000000000000..2fec11a0373a923cedd956091368a97e10916b03 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_266.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_267.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_267.txt new file mode 100644 index 0000000000000000000000000000000000000000..93685543e165975790b21ed3b63feb5f731b76e3 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_267.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_268.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_268.txt new file mode 100644 index 0000000000000000000000000000000000000000..15ea4be9c331557db770acf1fa4e5439f4e7357c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_268.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_269.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_269.txt new file mode 100644 index 0000000000000000000000000000000000000000..ffe9c7f43dbf4070056d9c9b3dc7defa01c4ba82 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp16_samp_fp16_269.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp32_perf_fp32_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp32_perf_fp32_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..8d9d8fee56451e2961b8d7b53aed70aef75d0acb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/resnet50_imagenet/resnet50_imagenet_fp32_perf_fp32_120.txt @@ -0,0 +1,687 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..35b105d97a68bcbfbff43da403a925200b5f5382 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_120.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_151.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_151.txt new file mode 100644 index 0000000000000000000000000000000000000000..8280ad73a88b4ca9a6b56c4118210a280283cd5e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_151.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_152.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_152.txt new file mode 100644 index 0000000000000000000000000000000000000000..2bb5c473d4584e3193f3b3390053700b27506565 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_152.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_153.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_153.txt new file mode 100644 index 0000000000000000000000000000000000000000..2fc0ebfbe5c2e0fe96637b22efcecbcc79d1ac13 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_153.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_154.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_154.txt new file mode 100644 index 0000000000000000000000000000000000000000..d1e7a3a32ff94dcbad5f9812f10e034afca71201 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_154.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_155.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_155.txt new file mode 100644 index 0000000000000000000000000000000000000000..6e417397af8fb9593a07eca1c6bb12065668dc51 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_155.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_156.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_156.txt new file mode 100644 index 0000000000000000000000000000000000000000..a7a3c079d38e5f5f638c5c65b5f92b01dce1533d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_156.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_157.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_157.txt new file mode 100644 index 0000000000000000000000000000000000000000..df8e948520da6185160850f198075fed6ed2d31d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_157.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_158.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_158.txt new file mode 100644 index 0000000000000000000000000000000000000000..48144ad26dc7dc140a5edfa180c479cc59d88769 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_158.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_159.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_159.txt new file mode 100644 index 0000000000000000000000000000000000000000..a7a8b5d9a3ec2fabec459d839011db147bfc6740 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_159.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_160.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_160.txt new file mode 100644 index 0000000000000000000000000000000000000000..00bd498faf8a57e71a8921395f24aeabc40f2c3d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_160.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_161.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_161.txt new file mode 100644 index 0000000000000000000000000000000000000000..feb7f4fb58c604f0fefaf03df971cae5992546a6 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_161.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_162.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_162.txt new file mode 100644 index 0000000000000000000000000000000000000000..9b5c953995fcee3f5a87294d58c0d8bfbd03b7b5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_162.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_163.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_163.txt new file mode 100644 index 0000000000000000000000000000000000000000..d440002ea739909721ab71f6a2637e3e4e5e43ae --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_163.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_164.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_164.txt new file mode 100644 index 0000000000000000000000000000000000000000..97bdfb6ee438be0723f2e206b035d0718c2e3c8a --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_164.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_165.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_165.txt new file mode 100644 index 0000000000000000000000000000000000000000..9fa8b59e7c6a0f489ea8f6e979c0624c852a1fa5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_165.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_166.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_166.txt new file mode 100644 index 0000000000000000000000000000000000000000..c02af1e8401a6cec2ec4354272dc9abd0e36d20f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_166.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_167.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_167.txt new file mode 100644 index 0000000000000000000000000000000000000000..ac8c32b7b4bd1794f0611a7602470a84e6e113dd --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_167.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_168.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_168.txt new file mode 100644 index 0000000000000000000000000000000000000000..33036bcf77a68012a02a3338ee2c58a69038a499 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_perf_fp16_168.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_261.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_261.txt new file mode 100644 index 0000000000000000000000000000000000000000..93c47b97c816b1c1dc19a5400efa40dc506985d3 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_261.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_262.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_262.txt new file mode 100644 index 0000000000000000000000000000000000000000..3068a697954c6bf4d9f86b4a3d24ec31b7d4a496 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_262.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_263.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_263.txt new file mode 100644 index 0000000000000000000000000000000000000000..2b5bf1c1e37cf8db751dc4bb033bde819d2bb3ba --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_263.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_264.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_264.txt new file mode 100644 index 0000000000000000000000000000000000000000..6c046346913f021ca6bd916ed293565c7fabbf7a --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_264.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_265.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_265.txt new file mode 100644 index 0000000000000000000000000000000000000000..2befd2504c3dc98626f0cf144d9f914b26a00f52 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_265.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_266.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_266.txt new file mode 100644 index 0000000000000000000000000000000000000000..9aa16c022611cdd686be61a166338c885af0a1ed --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_266.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_267.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_267.txt new file mode 100644 index 0000000000000000000000000000000000000000..d819d236d8e802c02e1de724a58a17b19d4d88a7 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_267.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_268.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_268.txt new file mode 100644 index 0000000000000000000000000000000000000000..f83aa5a4709dc52995ef15de36e536773af41e2c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_268.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_269.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_269.txt new file mode 100644 index 0000000000000000000000000000000000000000..c6c446c551e592d1da97339ecb863280e123ac11 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp16_samp_fp16_269.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp32_perf_fp32_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp32_perf_fp32_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..abb7299681b6fb721682e76b110e429fa58b0bff --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar10/vgg16_cifar10_fp32_perf_fp32_120.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..ecd5f1f273a176bed586d43a30ce54ac3128fc96 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_120.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_151.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_151.txt new file mode 100644 index 0000000000000000000000000000000000000000..c526f423205cedfa53089441044ee9bbc63d55b2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_151.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_152.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_152.txt new file mode 100644 index 0000000000000000000000000000000000000000..22b1e98bab229ab157903d6bf9fb60615a218d73 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_152.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_153.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_153.txt new file mode 100644 index 0000000000000000000000000000000000000000..b3b55cf48c5aa2edccc42166c8a667fa9ba0525d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_153.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_154.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_154.txt new file mode 100644 index 0000000000000000000000000000000000000000..92ec828264303390e919dffb5b8202e7056a7d2e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_154.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_155.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_155.txt new file mode 100644 index 0000000000000000000000000000000000000000..189a292990df4a90fcbde996d521074e0caf9aa9 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_155.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_156.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_156.txt new file mode 100644 index 0000000000000000000000000000000000000000..04a89e57adad7d4a3efd4d25589a57176222c512 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_156.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_157.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_157.txt new file mode 100644 index 0000000000000000000000000000000000000000..3355365dc018744ad1a13d28b4bcbb9ea9aa3544 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_157.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_158.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_158.txt new file mode 100644 index 0000000000000000000000000000000000000000..d519f756fb24c124a78a441beb64ef6b500bda78 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_158.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_159.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_159.txt new file mode 100644 index 0000000000000000000000000000000000000000..84702c96ffc953fd2e1350a3818c1d1dc2ca411c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_159.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_160.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_160.txt new file mode 100644 index 0000000000000000000000000000000000000000..7cca3d0807f2f07b371ed2d7ecc09c46cb98814c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_160.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_161.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_161.txt new file mode 100644 index 0000000000000000000000000000000000000000..74007f27b84264ab048cdf0559788f01861c4b3d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_161.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_162.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_162.txt new file mode 100644 index 0000000000000000000000000000000000000000..f7efa4236329b869a83bada18c7d6e0886466beb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_162.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_163.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_163.txt new file mode 100644 index 0000000000000000000000000000000000000000..18b0d55b59930fba7d49e229cf26daf519c17c74 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_163.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_164.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_164.txt new file mode 100644 index 0000000000000000000000000000000000000000..9f999df508b0097dc35d12b781d61a2db0267ccf --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_164.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_165.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_165.txt new file mode 100644 index 0000000000000000000000000000000000000000..4adb029c932d180ff157c72449ef5e176dd6318d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_165.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_166.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_166.txt new file mode 100644 index 0000000000000000000000000000000000000000..743386a5264e4118f7902564f26afdd60c6cb4fd --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_166.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_167.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_167.txt new file mode 100644 index 0000000000000000000000000000000000000000..0c0ce45f0e9ca21798834a2dcc5d89480c3aa4b8 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_167.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_168.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_168.txt new file mode 100644 index 0000000000000000000000000000000000000000..b99d2b30cba81cfa2ce0a9d416db8d0e021c7150 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_perf_fp16_168.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_261.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_261.txt new file mode 100644 index 0000000000000000000000000000000000000000..2c67cf74bc5de24304a0e61e2b1a29240f169f77 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_261.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_262.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_262.txt new file mode 100644 index 0000000000000000000000000000000000000000..f14d74e574c91ea99311d1704ce7509fd59bd68e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_262.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_263.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_263.txt new file mode 100644 index 0000000000000000000000000000000000000000..247eb7b675a38e9081891c54cacb6c161ed4dbdb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_263.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_264.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_264.txt new file mode 100644 index 0000000000000000000000000000000000000000..878f8d87162aa551eb44d6e2c79c6dccccafefea --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_264.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_265.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_265.txt new file mode 100644 index 0000000000000000000000000000000000000000..36826ba6b9c86fa7510621e4962a02433d1be746 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_265.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_266.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_266.txt new file mode 100644 index 0000000000000000000000000000000000000000..45da23840f0db91dba22812086508d2375272ab5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_266.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_267.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_267.txt new file mode 100644 index 0000000000000000000000000000000000000000..1244da47a9f5f51489887a3bb1c97f9274c9ab3f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_267.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_268.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_268.txt new file mode 100644 index 0000000000000000000000000000000000000000..83fb5487c921a1a851ea3249cbb1ea04376f2f92 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_268.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_269.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_269.txt new file mode 100644 index 0000000000000000000000000000000000000000..b1d756884be10fdeb15b96febf92375c5ffe95aa --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp16_samp_fp16_269.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp32_perf_fp32_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp32_perf_fp32_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..ef02bae6041afdd39151c18c7f8c5dd9634b2602 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_cifar100/vgg16_cifar100_fp32_perf_fp32_120.txt @@ -0,0 +1,150 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..589d3a4ac05f4b6c1ab01fac9f4d2a21357859d9 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_120.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_151.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_151.txt new file mode 100644 index 0000000000000000000000000000000000000000..2971ceeea6c8dd895551ec1ad644e950a2c5bf0d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_151.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_152.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_152.txt new file mode 100644 index 0000000000000000000000000000000000000000..06209b398390083a7e26f5fdb4aeeda3cb209f85 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_152.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_153.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_153.txt new file mode 100644 index 0000000000000000000000000000000000000000..1643c576ca002cc4b8cc09c4a06b97c91f35457c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_153.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_154.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_154.txt new file mode 100644 index 0000000000000000000000000000000000000000..8cbd143a8a8e2064f740d51b93696e84268031dd --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_154.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_155.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_155.txt new file mode 100644 index 0000000000000000000000000000000000000000..918a2cfce4ae2cdbe2512f329d6bc7d247e48b7c --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_155.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_156.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_156.txt new file mode 100644 index 0000000000000000000000000000000000000000..fba1d66b52595c5c4daed3631f4937073a8d1713 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_156.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_157.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_157.txt new file mode 100644 index 0000000000000000000000000000000000000000..b282e13d3160a6dbc9ff539af043a14042c6591d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_157.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_158.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_158.txt new file mode 100644 index 0000000000000000000000000000000000000000..1d3d7298488d10c4b14b5adf33424f91d23eddad --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_158.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_159.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_159.txt new file mode 100644 index 0000000000000000000000000000000000000000..6008d5f448fea1da8a972d0261a0f86c061f4297 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_159.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_160.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_160.txt new file mode 100644 index 0000000000000000000000000000000000000000..cf13d7359db25c20ce331a4114a8945cdad1872a --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_160.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_161.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_161.txt new file mode 100644 index 0000000000000000000000000000000000000000..233279f0829fbf4e776cb877d053dc49184c9971 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_161.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_162.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_162.txt new file mode 100644 index 0000000000000000000000000000000000000000..48c3d43c77cb5a825c84d99ee0e66c50ded6cf45 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_162.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_163.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_163.txt new file mode 100644 index 0000000000000000000000000000000000000000..28c204df0fc5274e49e7cb1ffd7818470b709b15 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_163.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_164.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_164.txt new file mode 100644 index 0000000000000000000000000000000000000000..ba32130b28fe4be08803034683f3371ea182602e --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_164.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_165.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_165.txt new file mode 100644 index 0000000000000000000000000000000000000000..c59e479d329481849d8d3dac85ceab6dacb695fd --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_165.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_166.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_166.txt new file mode 100644 index 0000000000000000000000000000000000000000..c64f4f3e52e6fa04b95b4612aa5c3bcb4a4c3ed2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_166.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_167.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_167.txt new file mode 100644 index 0000000000000000000000000000000000000000..54673f6021e799da059454b88b7fa00f4037aa37 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_167.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_168.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_168.txt new file mode 100644 index 0000000000000000000000000000000000000000..89db826f20bbc72ed4380b8ba9dbbe7525a4e4a8 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_perf_fp16_168.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_261.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_261.txt new file mode 100644 index 0000000000000000000000000000000000000000..e4548f7e02c2297a343cf2900af977e1c3ae1ccb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_261.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_262.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_262.txt new file mode 100644 index 0000000000000000000000000000000000000000..10e7503edfe46bb07fa6346a48480c1fdd62264b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_262.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_263.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_263.txt new file mode 100644 index 0000000000000000000000000000000000000000..b7b8bf2f6395e09572921abd7b0d7af36b1fd02b --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_263.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_264.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_264.txt new file mode 100644 index 0000000000000000000000000000000000000000..eba44020106d7110831698fe6797c3e08759b3c4 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_264.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_265.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_265.txt new file mode 100644 index 0000000000000000000000000000000000000000..db65e012b6c594f9919d33e8de8ba1edbe30485f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_265.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_266.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_266.txt new file mode 100644 index 0000000000000000000000000000000000000000..2e8956e67f1a0a2f3eb8cb56b93c442a477407e5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_266.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_267.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_267.txt new file mode 100644 index 0000000000000000000000000000000000000000..5f30e7718826a5a44d04258d0a461470bf361dfc --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_267.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_268.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_268.txt new file mode 100644 index 0000000000000000000000000000000000000000..a1c035492a86751cb107c73d6ff19d309b9d549f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_268.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_269.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_269.txt new file mode 100644 index 0000000000000000000000000000000000000000..ed7cda546e596b4f5d9a42cda6899c1973f370b5 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp16_samp_fp16_269.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp32_perf_fp32_120.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp32_perf_fp32_120.txt new file mode 100644 index 0000000000000000000000000000000000000000..e5708b3dbd6d733bf6b9af75952d72bb0bb9f974 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/profiling_results/vgg16_imagenet/vgg16_imagenet_fp32_perf_fp32_120.txt @@ -0,0 +1,159 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet/alexnet_valid_soc.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet/alexnet_valid_soc.txt new file mode 100644 index 0000000000000000000000000000000000000000..1b7aeb981c745717c52c841f99672cfbd532f7cb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet/alexnet_valid_soc.txt @@ -0,0 +1,231 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet2/alexnet2_valid_soc.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet2/alexnet2_valid_soc.txt new file mode 100644 index 0000000000000000000000000000000000000000..a888b5ee5a50d140f60d6579a3f6bdb6aa5ddfbd --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/alexnet2/alexnet2_valid_soc.txt @@ -0,0 +1,188 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/resnet18/resnet18_valid_soc.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/resnet18/resnet18_valid_soc.txt new file mode 100644 index 0000000000000000000000000000000000000000..942789c1c4defd1139e75209ffbcb073a2b39b30 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/resnet18/resnet18_valid_soc.txt @@ -0,0 +1,1576 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar10/vgg16_cifar10_valid_soc.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar10/vgg16_cifar10_valid_soc.txt new file mode 100644 index 0000000000000000000000000000000000000000..789f4e21cf4a778535d1df0f9f7be22c1415d672 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar10/vgg16_cifar10_valid_soc.txt @@ -0,0 +1,1027 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar100/vgg16_cifar100_valid_soc.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar100/vgg16_cifar100_valid_soc.txt new file mode 100644 index 0000000000000000000000000000000000000000..ef6509b99bee287bf0e3dfbaa035d51f9e3cb0ea --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/runtime_experiments/vgg16_cifar100/vgg16_cifar100_valid_soc.txt @@ -0,0 +1,210 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/alexnet2_cifar10/alexnet2_cifar10.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/alexnet2_cifar10/alexnet2_cifar10.txt new file mode 100644 index 0000000000000000000000000000000000000000..6ec4a06d3dbd2e088d6db287d23dd3bd5aad7ddb --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/alexnet2_cifar10/alexnet2_cifar10.txt @@ -0,0 +1,419 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/alexnet_cifar10/alexnet_cifar10.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/alexnet_cifar10/alexnet_cifar10.txt new file mode 100644 index 0000000000000000000000000000000000000000..a9ccba6eb63f620c0e3b6f95fd7c50892018f00f --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/alexnet_cifar10/alexnet_cifar10.txt @@ -0,0 +1,511 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/alexnet_imagenet/alexnet_imagenet.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/alexnet_imagenet/alexnet_imagenet.txt new file mode 100644 index 0000000000000000000000000000000000000000..b0e42a5aaa5d7b5a06b6422a5c33a0047b6eff8d --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/alexnet_imagenet/alexnet_imagenet.txt @@ -0,0 +1,229 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/lenet_keras/lenet_keras.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/lenet_keras/lenet_keras.txt new file mode 100644 index 0000000000000000000000000000000000000000..b4e51dff426f4d3c5cb7b9572e6aa5940212acbd --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/lenet_keras/lenet_keras.txt @@ -0,0 +1,409 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/mobilenet_cifar10/mobilenet_cifar10.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/mobilenet_cifar10/mobilenet_cifar10.txt new file mode 100644 index 0000000000000000000000000000000000000000..b4d8bd893c8d9395fce6a3484d75f543f1e72da2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/mobilenet_cifar10/mobilenet_cifar10.txt @@ -0,0 +1,3220 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/resnet18_cifar10/resnet18_cifar10.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/resnet18_cifar10/resnet18_cifar10.txt new file mode 100644 index 0000000000000000000000000000000000000000..654cffbf632686dca6310a93ecf56b6521e32039 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/resnet18_cifar10/resnet18_cifar10.txt @@ -0,0 +1,2296 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/resnet50_imagenet/resnet50_imagenet.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/resnet50_imagenet/resnet50_imagenet.txt new file mode 100644 index 0000000000000000000000000000000000000000..094eed413b520f9dd661797b96735438861d1c08 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/resnet50_imagenet/resnet50_imagenet.txt @@ -0,0 +1,1057 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/vgg16_cifar10/vgg16_cifar10.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/vgg16_cifar10/vgg16_cifar10.txt new file mode 100644 index 0000000000000000000000000000000000000000..2b325a9fe2d122e74cdd2b80e2768e68591313bf --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/vgg16_cifar10/vgg16_cifar10.txt @@ -0,0 +1,913 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/vgg16_cifar100/vgg16_cifar100.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/vgg16_cifar100/vgg16_cifar100.txt new file mode 100644 index 0000000000000000000000000000000000000000..2c29bedd096aec2c7f66afbe729353e372fac403 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/vgg16_cifar100/vgg16_cifar100.txt @@ -0,0 +1,970 @@ +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/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/vgg16_imagenet/vgg16_imagenet.txt b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/vgg16_imagenet/vgg16_imagenet.txt new file mode 100644 index 0000000000000000000000000000000000000000..108a101c810f4ebe488e6f2029be4d970d7869a2 --- /dev/null +++ b/llvm/projects/hpvm-tensor-rt/PPoPP_results/soc_sim_results/dev_time/vgg16_imagenet/vgg16_imagenet.txt @@ -0,0 +1,561 @@ +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 +-----