Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Dec 20, 2025

• Don't use `EnumerateIndices` unnecessariily

  Show description for fc1d7b1
  

  dsharletg

  authored and

  xnnpack-bot

  committed
  fc1d7b1

  Copy full SHA for fc1d7b1

Commits on Dec 19, 2025

• Change simd::vec to be recursively defined

  Show description for 82d56fb
  

  dsharletg

  authored and

  xnnpack-bot

  committed
  82d56fb

  Copy full SHA for 82d56fb

• Fixed full byte requirement in 2-bit Fully Connected test for qs8 version.

  Show description for f52fbff
  

  Aelphy

  authored and

  xnnpack-bot

  committed
  f52fbff

  Copy full SHA for f52fbff

• Don't try to test int2 fully connected with YNNPACK

  Show description for ecdf584
  

  dsharletg

  authored and

  xnnpack-bot

  committed
  ecdf584

  Copy full SHA for ecdf584

Commits on Dec 18, 2025

• Added qs8_qc2w FullyConnected.

  Show description for 91bc5c2
  

  Aelphy

  authored and

  xnnpack-bot

  committed
  91bc5c2

  Copy full SHA for 91bc5c2

• Support tile_k > 1 in transpose `stencil_copy` rewrite.

  Show description for e31a359
  

  mariecwhite

  authored and

  xnnpack-bot

  committed
  e31a359

  Copy full SHA for e31a359

• Fix `multi_vec`'s partial `load`/`store`

  Show description for 183297d
  

  dsharletg

  authored and

  xnnpack-bot

  committed
  183297d

  Copy full SHA for 183297d

• Rewrite transpose_a(stencil_copy(x)) to stencil_copy(transpose_a(x))

  Show description for ff23b55
  xnnpack-bot

  committed
  ff23b55

  Copy full SHA for ff23b55

• Minor fixes to stencil_copy

  Show description for 0723f38
  xnnpack-bot

  committed
  0723f38

  Copy full SHA for 0723f38

• Enable transpose_a kernels for num_k_dims > 1

  Show description for 944d44d
  

  dsharletg

  authored and

  xnnpack-bot

  committed
  944d44d

  Copy full SHA for 944d44d

Commits on Dec 17, 2025

• Tweak K1 = 1 reduction case

  Show description for 51dd4c8
  

  dsharletg

  authored and

  xnnpack-bot

  committed
  51dd4c8

  Copy full SHA for 51dd4c8

• Merge pull request #9316 from salmanmkc:upgrade-github-actions-node24

  Show description for 996234f
  xnnpack-bot

  committed
  996234f

  Copy full SHA for 996234f

• Pass down the original kernel and bias buffers when creating a convolution operation.

  Show description for e317c78
  

  qukhan

  authored and

  xnnpack-bot

  committed
  e317c78

  Copy full SHA for e317c78

• Improved k1=1 reductions. Most of the cases have significant speed-up except scalar kernels, sum bf16_f32 for f16c and sse2, and x8_s32 for avx2.

  Show description for 103c631
  

  Aelphy

  authored and

  xnnpack-bot

  committed
  103c631

  Copy full SHA for 103c631

• Allowed not to pass channel wise zero point to accommodate asymmetric (zero point is assumed 0) 2-bit case.

  Show description for adf70ec
  

  Aelphy

  authored and

  xnnpack-bot

  committed
  adf70ec

  Copy full SHA for adf70ec

• Enable weight caching for quantized depthwise convolutions.

  Show description for 2c1a512
  

  qukhan

  authored and

  xnnpack-bot

  committed
  2c1a512

  Copy full SHA for 2c1a512

• Rewrite SIMD tests to make a new test for each architecture

  Show description for 4574c4d
  

  dsharletg

  authored and

  xnnpack-bot

  committed
  4574c4d

  Copy full SHA for 4574c4d

Commits on Dec 16, 2025

• Fixed n=0 case that left uninitialized vectors.

  Show description for 30cd310
  

  Aelphy

  authored and

  xnnpack-bot

  committed
  30cd310

  Copy full SHA for 30cd310

• Upgrade GitHub Actions for Node 24 compatibility

  salmanmkc

  committed
  7926389

  Copy full SHA for 7926389

• Pass down the original kernel and bias buffers when creating a fully connected operation.

  Show description for 9e77b69
  

  qukhan

  authored and

  xnnpack-bot

  committed
  9e77b69

  Copy full SHA for 9e77b69

• Add stencil_copy helper function

  Show description for 5dab025
  xnnpack-bot

  committed
  5dab025

  Copy full SHA for 5dab025

• Attepmted to fix 2-bit fully-connected operator flakyness.

  Show description for d9ad8b4
  

  Aelphy

  authored and

  xnnpack-bot

  committed
  d9ad8b4

  Copy full SHA for d9ad8b4

• Create transpose_a helper

  Show description for 7b059b1
  xnnpack-bot

  committed
  7b059b1

  Copy full SHA for 7b059b1

• Don't allow computing ops in-place if the output of the op is an external output.

  Show description for becb9f5
  

  dsharletg

  authored and

  xnnpack-bot

  committed
  becb9f5

  Copy full SHA for becb9f5

Commits on Dec 15, 2025

• Hexagon V81 portable simd

  Show description for 5d89a13
  

  fbarchard

  authored and

  xnnpack-bot

  committed
  5d89a13

  Copy full SHA for 5d89a13

• Added scalar kernels to qc2w gemm-config.

  Show description for f51a395
  

  Aelphy

  authored and

  xnnpack-bot

  committed
  f51a395

  Copy full SHA for f51a395

• Fix assert in dot.cc

  Show description for 7b275c9
  

  vksnk

  authored and

  xnnpack-bot

  committed
  7b275c9

  Copy full SHA for 7b275c9

• Tighten up reduce kernels and architectures

  Show description for a423689
  

  dsharletg

  authored and

  xnnpack-bot

  committed
  a423689

  Copy full SHA for a423689

• Add `fma`, `concat`, and more `convert` functions to SIMD headers

  Show description for 021f28e
  

  dsharletg

  authored and

  xnnpack-bot

  committed
  021f28e

  Copy full SHA for 021f28e

• Add a function to get fingerprints by index for testing.

  Show description for 3fc92c7
  

  qukhan

  authored and

  xnnpack-bot

  committed
  3fc92c7

  Copy full SHA for 3fc92c7

Commits on Dec 12, 2025

• Added simd extract tests.

  Show description for 7746844
  

  Aelphy

  authored and

  xnnpack-bot

  committed
  7746844

  Copy full SHA for 7746844

• Fix uninitialized variable build error for params to rsum for 2 bit

  Show description for ddc8e1a
  

  fbarchard

  authored and

  xnnpack-bot

  committed
  ddc8e1a

  Copy full SHA for ddc8e1a

• Merge pull request #9270 from JonathanC-ARM:jclohess_fix_missing_debug_logging

  Show description for 3a7b710
  xnnpack-bot

  committed
  3a7b710

  Copy full SHA for 3a7b710

• Use /clang: to pass copts to clang-cl

  Show description for e436865
  

  dsharletg

  authored and

  xnnpack-bot

  committed
  e436865

  Copy full SHA for e436865

• Minor code cleanup

  Show description for 4e2113d
  

  dsharletg

  authored and

  xnnpack-bot

  committed
  4e2113d

  Copy full SHA for 4e2113d

Pagination

Previous

Next