Commits

Branch selector

 master

User selector

bertmaher

Datepicker

All time

Commit History

Commits on Nov 19, 2020

• [te] Fix pow (#48213)

  Show description for 6da26fe
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  6da26fe

  Copy full SHA for 6da26fe

Commits on Nov 18, 2020

• [tensorexpr] Switch cpp tests to pure gtest (#48160)

  Show description for 07657b6
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  07657b6

  Copy full SHA for 07657b6

• [te][benchmark] Add more optimized versions of gemm (#48159)

  Show description for 464d23e
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  464d23e

  Copy full SHA for 464d23e

• [te] Make BUILD_TENSOREXPR_BENCHMARK a real CMake option (#48158)

  Show description for 8a996dd
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  8a996dd

  Copy full SHA for 8a996dd

Commits on Nov 17, 2020

• [torch][te] aten::type_as is unary, not binary (#48085)

  Show description for 736deef
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  736deef

  Copy full SHA for 736deef

• [pytorch][te] Handle negative axis in chunk (#48084)

  Show description for bbee0ec
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  bbee0ec

  Copy full SHA for bbee0ec

• [pytorch][te] Do not merge Tensor[] variant of aten::where into fusion group (#48063)

  Show description for af37f8f
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  af37f8f

  Copy full SHA for af37f8f

Commits on Nov 15, 2020

• [pytorch][te][easy] Remove KernelScope from fusion pass tests (#47952)

  Show description for 5adf840
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  5adf840

  Copy full SHA for 5adf840

Commits on Nov 13, 2020

• [pytorch][te] Don't start TE fusion groups with an unknown-typed result (#47884)

  Show description for 6b8d20c
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  6b8d20c

  Copy full SHA for 6b8d20c

Commits on Nov 9, 2020

• [tensorexpr] Fix registration of intrinsics on llvm-fb (#47540)

  Show description for 77c49e6
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  77c49e6

  Copy full SHA for 77c49e6

Commits on Oct 27, 2020

• [pytorch] Improve/fix heuristics for using mkldnn vs native conv (#46675)

  Show description for 2397c8d
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  2397c8d

  Copy full SHA for 2397c8d

Commits on Oct 24, 2020

• [pytorch][tensorexpr] Promote integer arguments to sin/cos/tan to float (#46776)

  Show description for c4892c8
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  c4892c8

  Copy full SHA for c4892c8

Commits on Oct 10, 2020

• [pytorch][te] Add compilation time benchmark (#46124)

  Show description for b7261de
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  b7261de

  Copy full SHA for b7261de

Commits on Oct 6, 2020

• [te][llvm] Enable fused multiply-add (fma) in code generation (#45906)

  Show description for 624084e
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  624084e

  Copy full SHA for 624084e

• [te] Tiled (m=32 x n=32) gemm benchmark (#45905)

  Show description for f2e5694
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  f2e5694

  Copy full SHA for f2e5694

• [te] Add a benchmark harness (#45875)

  Show description for 50f8957
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  50f8957

  Copy full SHA for 50f8957

• [te] Add a 2D convolution example test (#45514)

  Show description for 5ff3162
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  5ff3162

  Copy full SHA for 5ff3162

Commits on Oct 2, 2020

• [te] Get llvm codegen to compile with llvm9 and llvm-fb (#45726)

  Show description for e3d2def
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  e3d2def

  Copy full SHA for e3d2def

Commits on Sep 28, 2020

• Add env variable to bypass CUDACachingAllocator for debugging (#45294)

  Show description for 03342af
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  03342af

  Copy full SHA for 03342af

Commits on Sep 24, 2020

• Failing test demonstrating problems with mixed output shapes (#44455)

  Show description for 2d00ebd
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  2d00ebd

  Copy full SHA for 2d00ebd

Commits on Sep 19, 2020

• [pytorch][tensorexpr] Make gtest-style macros in tests match actual gtest signatures (#44861)

  Show description for 0714c00
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  0714c00

  Copy full SHA for 0714c00

Commits on Sep 17, 2020

• [te] Disable flaky test CudaSharedMemReduce_1 (#44862)

  Show description for a40ef25
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  a40ef25

  Copy full SHA for a40ef25

Commits on Sep 13, 2020

• Fix bug simplifying if-then-else when it can be removed (#44462)

  Show description for 6d4a605
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  6d4a605

  Copy full SHA for 6d4a605

Commits on Sep 9, 2020

• Fix frac in CUDA fuser (#44152)

  Show description for 6ec8fab
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  6ec8fab

  Copy full SHA for 6ec8fab

• Prevent the TE fuser from getting datatypes it can't handle (#44160)

  Show description for 350130a
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  350130a

  Copy full SHA for 350130a

• [te] Fix casting of unsigned char, and abs(int) (#44157)

  Show description for 960c088
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  960c088

  Copy full SHA for 960c088

• Dump optimized graph when logging in already-optimized PE (#44315)

  Show description for 8acce55
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  8acce55

  Copy full SHA for 8acce55

Commits on Sep 4, 2020

• [te] Disable reductions by default (#44122)

  Show description for 98ad5ff
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  98ad5ff

  Copy full SHA for 98ad5ff

Commits on Sep 3, 2020

• Test TE fuser unary ops and fix sigmoid(half) (#44094)

  Show description for 55ff9aa
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  55ff9aa

  Copy full SHA for 55ff9aa

• Respect canFuseOn{CPU,GPU} in TE fuser (#43967)

  Show description for 33d51a9
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  33d51a9

  Copy full SHA for 33d51a9

Commits on Sep 1, 2020

• Fix NaN propagation in TE fuser's min/max implementation (#43609)

  Show description for c14a361
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  c14a361

  Copy full SHA for c14a361

Commits on Aug 30, 2020

• Remove unnamed namespace in headers (#43689)

  Show description for 1830e4f
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  1830e4f

  Copy full SHA for 1830e4f

Commits on Aug 27, 2020

• Fix NaN propagation in fuser's min/max implementation (#43590)

  Show description for 0bf27d6
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  0bf27d6

  Copy full SHA for 0bf27d6

Commits on Aug 18, 2020

• [tensorexpr] Fix promotion of booleans (#43097)

  Show description for 6c99d56
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  6c99d56

  Copy full SHA for 6c99d56

Commits on Aug 13, 2020

• Add executor and fuser options to the fastrnn test fixture (#42946)

  Show description for eb47940
  

  bertmaher

  authored and

  facebook-github-bot

  committed
  eb47940

  Copy full SHA for eb47940

Pagination

Previous

Next