Skip to content
Snippets Groups Projects

Repository graph

You can move around the graph by using the arrow keys.
Select Git revision
  • hpvm-release-epochs
  • hpvm-release-epochs0
  • main default
  • v2.0
  • v2.0rc
  • v1.0
  • v0.5
7 results
Created with Raphaël 2.2.019Apr18171615141312111098765432130Mar292827262423222119181716151312116543228Feb27262524232019181716151413121110742131Jan30292827252423222120191817161514131098Adding routine for Stastical confidence measurement for install-tunerFixing mem handling bug in Lenet_promise sourceUpdating piped scripts to use 5K imagesAdding promise translator support to dump Quantization rangesAdding auto garbage collection in inference loopAdding Quant Range computation across batches of 100 imagesAdding batched inference to get Quant Ranges -- needs more testingMergingReducing quantization percentiles triedprofile in nvprof/nvvpMerge branch 'approx_hpvm' of gitlab.engr.illinois.edu:llvm/hpvm into approx_hpvm1. Now support previously enexpected case of search value resulting to targetAdding test case for PROMISeupdate keras/onnx/tvm modelsmergingAdding test cases for quantizzationMerge branch 'approx_hpvm' of gitlab.engr.illinois.edu:llvm/hpvm into approx_hpvmUsing new knob reading in hpvm controller and approxhpvm_runtime_utilsReplace throw statements with ERROR macrosIncreasing Calibration set size to 5KReplaced throw with ERROR macro - more to be replacedAdd minor fixes to FUJ and CBE, and update Makefiles for Machsuite benchmarkssmall changes to RT and HPVMUtilsfixed cbe handling of loop induction variablecava fixescava changesMerge branch 'hpvm-fpga-9' of gitlab.engr.illinois.edu:llvm/hpvm into hpvm-fpga-9Merge branch 'fpga-opts-WIP' into hpvm-fpga-9unroll and jam test benchmark, latestworking Unroll & Jam pass (except for Denoise in CAVA) along with some necessary fixes in other filesRemove mistakenly committed executablesAdd new FPGA benchmarks and fix Makefilesfuse works except for checking depsAdding test that varies interpolation ratesAdding macro to disable approximation simulationsWiden criteria for direct allocas to include eligible allocas not in entry blockUsing new interpolation_rate knob for Samplingworking fusion with move-upMergingFix CBE bugs: include address space specifier on pointer casts, account for pointer decay on array bitcasts, and fix switch code gen