Commits

Branch selector

 compiler_gym

User selector

dskhudia

Datepicker

All time

Commit History

Commits on Oct 14, 2021

• Improve im2col for certain cases (pytorch#715)

  Show description for 0375f13
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  0375f13

Commits on Oct 12, 2021

• Fix intermittant UBSAN error

  Show description for 9698cad
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  9698cad

Commits on Aug 3, 2021

• Small commit to force a PyTorch submodule update PR (pytorch#659)

  Show description for 10ec0d3
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  10ec0d3

Commits on Jul 21, 2021

• Fix OSS ubuntu test failure (pytorch#653)

  Show description for d2440b4
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  d2440b4

Commits on Jun 9, 2021

• Put github actions build status at the top of readme (pytorch#625)

  Show description for 0520ad5
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  0520ad5

Commits on Jun 8, 2021

• fix build with gcc 5.4 (pytorch#622)

  Show description for 8aaefde
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  8aaefde

Commits on Jun 7, 2021

• Unroll outer loop to improve performance (pytorch#617)

  Show description for baeb9a4
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  baeb9a4

• index remapping benchmark (pytorch#615)

  Show description for f433d5f
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  f433d5f

• Vectorize index remapping for pruned embedding tables (pytorch#607)

  Show description for 76cf190
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  76cf190

Commits on May 12, 2021

• add missing copyright headers

  Show description for 4b8aaad
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  4b8aaad

Commits on Mar 25, 2021

• Build fix with certain compilers versions (pytorch#575)

  Show description for 6c051e1
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  6c051e1

Commits on Mar 16, 2021

• Update asmjit for fbcode and oss (pytorch#538)

  Show description for 5d36b7d
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  5d36b7d

Commits on Mar 1, 2021

• Fix build with older gcc versions (pytorch#525)

  Show description for a431ee3
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  a431ee3

Commits on Feb 27, 2021

• open source sparse-dense matrix-matrix multiplication (int8 and fp32) (pytorch#523)

  Show description for 953e671
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  953e671

Commits on Feb 8, 2021

• removing parallel build for OSS (pytorch#509)

  Show description for c5e2273
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  c5e2273

Commits on Oct 16, 2020

• Add missing copyright header (pytorch#443)

  Show description for e1db75c
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  e1db75c

Commits on Oct 2, 2020

• Matrix transpose kernels for int8/uint8 datatypes (pytorch#431)

  Show description for fe91640
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  fe91640

Commits on Sep 14, 2020

• part3 Move embedding dequantization kernels to fbgemm for better sharing between C2/PT (pytorch#426)

  Show description for 1d71039
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  1d71039

Commits on Sep 12, 2020

• Improve SLS performance on BDW by reducing port 5 pressure (pytorch#429)

  Show description for 0725301
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  0725301

Commits on Sep 4, 2020

• part2: Move embedding quantization kernels to fbgemm for better sharing between C2/PT (pytorch#425)

  Show description for 7e401c0
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  7e401c0

Commits on Sep 2, 2020

• To fix a windows failure (pytorch#424)

  Show description for a591925
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  a591925

Commits on Aug 31, 2020

• Move embedding quantization kernels to fbgemm for better sharing between C2/PT (pytorch#419)

  Show description for a9397bf
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  a9397bf

Commits on Aug 29, 2020

• Remove usage of _mm_maskmoveu_si128 (pytorch#421)

  Show description for 20d994f
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  20d994f

Commits on Aug 27, 2020

• Reduce code include in files compiled with avx2 flag (pytorch#416)

  Show description for be51cdd
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  be51cdd

Commits on Aug 11, 2020

• Adding missing copyright headers (pytorch#407)

  Show description for c043abc
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  c043abc

Commits on Aug 5, 2020

• Fix confusing comment

  Show description for 4abc34a
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  4abc34a

Commits on Jul 21, 2020

• Bug fix in depthwise conv (pytorch#395)

  Show description for 139c6f2
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  139c6f2

Commits on May 14, 2020

• Minor improvements in GEMM Kernels (pytorch#368)

  Show description for 46981b8
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  46981b8

Commits on Apr 30, 2020

• asmjit submodule update (pytorch#364)

  Show description for b2bdf84
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  b2bdf84

Commits on Apr 23, 2020

• Zero point addition after rounding in quantization routines (pytorch#362)

  Show description for 91063b0
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  91063b0

Commits on Apr 9, 2020

• Merge pull request pytorch#345 from jspark1105/fixup-T64383689-master

  Show description for 96bf855
  dskhudia

  authored
  96bf855

Commits on Apr 7, 2020

• Match numerics of quantize with vector path and fake quantize (pytorch#340)

  Show description for 284e1c7
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  284e1c7

Commits on Mar 21, 2020

• Quantize should work even if the value after quantization becomes inf (pytorch#317)

  Show description for 3ee0b7d
  

  dskhudia

  authored and

  jspark1105

  committed
  3ee0b7d

• Quantize should work even if the value after quantization becomes inf (pytorch#107)

  Show description for 2a50f19
  

  dskhudia

  authored and

  jspark1105

  committed
  2a50f19

• Make spmmfp32 and spmmi8 benchmark consistent with tests (pytorch#315)

  Show description for 40235b3
  

  dskhudia

  authored and

  jspark1105

  committed
  40235b3

Pagination

PreviousNext