Google Git
Sign in
chromium/external/github.com/google/XNNPACK/HEAD
  1. 5df1362 Mark `binary_test` with `timeout = "moderate"` to avoid flakiness due to timeouts. by Pedro Gonnet · 2 hours ago upstream/master
  2. b58f508 Add `slice_dim0` helper by Dillon Sharlet · 3 hours ago
  3. d06fc11 Merge pull request #9935 from wangw-1991:fix_bugs by XNNPACK Team · 3 hours ago
  4. ac9bc35 Adds missing GEMMBenchmark overload for qp8_f32_qc8w_gemm_minmax and corrects qp8_f32_qc4w microkernels and associated functions to use xnn_f32_qc4w_minmax_params instead of xnn_f32_minmax_params. by Samuel Fuller · 3 hours ago
  5. 704a6d7 Demote some logging for subgraph rewrites from `xnn_log_info` to `xnn_log_debug`. by Pedro Gonnet · 3 hours ago
  6. c9e961c Initial upload. by Wei Wang · 13 hours ago
  7. 7f56155 Minor simplifications of static_broadcast nodes by Dillon Sharlet · 20 hours ago
  8. 9344bcb Give slinky our assumptions to reduce overhead by Dillon Sharlet · 22 hours ago
  9. 7a55567 Update slinky in XNNPACK by Dillon Sharlet · 22 hours ago
  10. eed1807 Handle identity casts for scalar vectors in simd wrappers by Volodymyr Kysenko · 23 hours ago
  11. f4ca248 Add fp64 reduce support by Dillon Sharlet · 23 hours ago
  12. 211341d Add fp64 support to simd::vec for x86 and ARM64 by Dillon Sharlet · 25 hours ago chromium/7784
  13. 1b30163 Merge pull request #9858 from GregoryComer:bf16-rminmax-scalar by XNNPACK Team · 2 days ago
  14. a902cb3 Merge pull request #9879 from keaganchern:rvv-f32-vmulcaddc by XNNPACK Team · 2 days ago
  15. 88a78ab Add subgraph rewrite for `x/sqrt(y)` to `x*reciprocal_square_root(y)` by Dillon Sharlet · 2 days ago
  16. 7ef819c updating kernel and testing infra by kchern2@berkeley.edu · 2 days ago
  17. c7903c9 Add WASM SIMD128 kernel for dequantize_dot. by Volodymyr Kysenko · 2 days ago
  18. f84ade5 Remove formatting check from gn build by Dillon Sharlet · 2 days ago
  19. fcdddfb Don't rewrite dequantize_dot if there is no optimized kernel. by Volodymyr Kysenko · 2 days ago
  20. d6097c2 Avoid `raw_buffer::cast` in dynamic quantization implementation by Dillon Sharlet · 2 days ago
  21. 37d77e5 [gn] fix: do not cache Git repos by Richard Townsend · 2 days ago
  22. c1bff3f Fix mismatching function signatures. by Volodymyr Kysenko · 2 days ago
  23. 925f0e9 Add WASM SIMD128 int8xint8->int32 dot product kernels. by Volodymyr Kysenko · 2 days ago
  24. 6c008cf Optimize the subgraph glue code for reduce kernels by Dillon Sharlet · 2 days ago
  25. 24332a0 Adjust sharding of reduce kernel tests by Dillon Sharlet · 3 days ago
  26. f82740c f16_vexp and f16_vapproxgelu - remove arm fp16 requirement from scalar microkernel tests by Frank Barchard · 3 days ago
  27. 25b9b12 f16_vgelu - remove arm fp16 requirement from scalar microkernel tests by Frank Barchard · 3 days ago
  28. 6938f34 Fixes to enable Hexagon to pass continuous testing by Dillon Sharlet · 3 days ago
  29. 762ae7c Relax tolerances of `dequantize_dot` test by Dillon Sharlet · 3 days ago
  30. ea6c92e Set WASM parameters for threading and update some of the tests accordingly. by Volodymyr Kysenko · 3 days ago
  31. 24d51bb Set stack sizes for WASM. by Volodymyr Kysenko · 3 days ago
  32. 4b71166 Merge pull request #9903 from ken-unger:rvv-fp16-detect-fix by XNNPACK Team · 3 days ago
  33. e2b2b9d Fix F16_VGELU unittest crash on ARM by Frank Barchard · 3 days ago
  34. 6b30acd Add RMS norm benchmark by Dillon Sharlet · 3 days ago
  35. b5bdf59 fix comment by Ken Unger · 3 days ago
  36. 225c8e8 fix RISCV_HWPROBE_EXT_ZVFH when not defined by Ken Unger · 3 days ago
  37. d0a0412 Actually use WASM SIMD128 FP64 dot kernels by including them in kernel.inc by Volodymyr Kysenko · 3 days ago
  38. 202cf6e added additional branching for vmulcaddc to prevent crashing by kchern2@berkeley.edu · 3 days ago
  39. 5872ba3 Fix for MSVC build - remove avx2 from fma3 header. by Frank Barchard · 3 days ago
  40. 3f281b3 Remove capture of variables by reference across threads. by Marie White · 4 days ago
  41. 5ab29b0 Add more m=1 kernels by Dillon Sharlet · 4 days ago
  42. aee6b1b Add FP16 detect for RVV by Frank Barchard · 4 days ago
  43. 5ab74a4 Print ccache stats for Windows builds. by Alexander Shaposhnikov · 4 days ago
  44. 79f8208 Use `fuse_and_slice_leading_dims` + `for_each_element` for `compute_qd8_params` by Dillon Sharlet · 4 days ago
  45. 48365ce Optimize `fuse_and_slice_leading_dims` by Dillon Sharlet · 4 days ago
  46. e238480 Add m = 1 kernel for tile_k = 16 dot kernels by Dillon Sharlet · 4 days ago
  47. 80bc38c Add fp64 dot kernels for wasm simd. by Volodymyr Kysenko · 4 days ago
  48. b5e60c2 Merge pull request #9693 from ken-unger:unary-rvv by XNNPACK Team · 4 days ago
  49. 9db1fb7 Add benchmark of `fuse_and_slice_leading_dims` by Dillon Sharlet · 4 days ago
  50. 0dc6aba Replace `broadcast_or_load` with specialization for `offset` in `dequantize_dot` by Dillon Sharlet · 4 days ago
  51. 802acce Add WASM SIMD128 FP32 dot product kernels. by Volodymyr Kysenko · 4 days ago
  52. 76f79df Add a `dequantize_dot` kernel by Dillon Sharlet · 4 days ago
  53. e4ae519 Make wasm simd128 arch name consistent with others. by Volodymyr Kysenko · 4 days ago
  54. ad25fa0 Add variable-length RVV f32-vmulcaddc-minmax microkernel by kchern2@berkeley.edu · 4 days ago
  55. 6293fad Fix test:buffer_test asan failure by Frank Barchard · 4 days ago
  56. b095cc6 Remove F32 ReLU GEMM/IGEMM microkernels by Frank Barchard · 4 days ago
  57. db167df Minor cleanups by Dillon Sharlet · 4 days ago
  58. 5ca7819 Add scalar parameters for more unary kernels by Dillon Sharlet · 4 days ago
  59. c60479f Merge branch 'google:master' into unary-rvv by Ken Unger · 4 days ago
  60. de3fca8 adding f32-vmulcaddc rvv microkernel by kchern2@berkeley.edu · 5 days ago
  61. 9263081 Fix neonsme2 warning script tools/generate-gemm-test.py by Frank Barchard · 7 days ago
  62. 5407a22 Run generators to update RVV microkernels by Frank Barchard · 7 days ago
  63. 8c540e4 Refactor `schedule_dot` and `run_dot` to not capture parameters of A and B by Dillon Sharlet · 7 days ago
  64. 4453b06 Refactor Dockerfile.standard and add riscv64 container build by Alexander Shaposhnikov · 7 days ago
  65. 6d11164 GEMM test generator fix for linear and relu - set output to input if not set by Frank Barchard · 7 days ago
  66. 3af7056 Remove whitespace from rvv qd8 gemm by Frank Barchard · 7 days ago
  67. 353d065 Strengthen fusion of constant arithmetic by Dillon Sharlet · 7 days ago
  68. 8bfc09d Add benchmark support for f32-bf16-vcvt by Gregory Comer · 7 days ago
  69. ffba580 Add transpose and interleave kernels for wasm simd128. by Volodymyr Kysenko · 7 days ago
  70. 7019283 Add interleave simd wrappers for wasm. by Volodymyr Kysenko · 7 days ago
  71. 3582cbd Implement approxgelu in YNNPACK by Dillon Sharlet · 7 days ago
  72. b8455c9 Add WASM SIMD128 kernels for FP32 sum and sum of squares. by Volodymyr Kysenko · 7 days ago
  73. 43df024 Add transpose and its supporting functions to wasm simd wrappers. by Volodymyr Kysenko · 7 days ago
  74. 9dd4192 Add min, max, minmax reduce kernels for wasm. by Volodymyr Kysenko · 7 days ago
  75. 8c9c00f Add explicit dimensions for transposed A by Dillon Sharlet · 7 days ago
  76. 319c384 Add horizontal_min/horizontal_max implementation to wasm simd wrappers by Volodymyr Kysenko · 7 days ago
  77. fd746bf Add missing min/max overloads for WASM SIMD. by Volodymyr Kysenko · 7 days ago
  78. a2d8b20 Add params to `erf` kernel and use them to implement `gelu`. Mostly AI generated, here is what AI had to say: by Dillon Sharlet · 7 days ago
  79. 9cd282f Fix discrepancy in a_stride_m/a_stride_k for transposed dot kernels by Dillon Sharlet · 7 days ago
  80. dfa0f4b Add support for passing scalar parameters to elementwise kernels. by Volodymyr Kysenko · 8 days ago
  81. e4a3e71 Update bazel build for bf16 minmax kernels by Gregory Comer · 8 days ago
  82. 4f95611 Remove unnecessary convert after quantized dots with float outputs by Dillon Sharlet · 8 days ago
  83. f086008 Add scalar bf16-rminmax microkernels, configs, and tests by Gregory Comer · 8 days ago
  84. 015e3e6 Add `ynn_define_slice_like` and use it to implement `xnn_define_rope` by Dillon Sharlet · 8 days ago
  85. 94a13a8 Merge pull request #9844 from yolanda15:qd8_update_simd by XNNPACK Team · 8 days ago
  86. 3ef1548 Fix warning on some builds by Dillon Sharlet · 9 days ago
  87. cbd072c Merge pull request #9846 from yolanda15:qd8_update_sdot by XNNPACK Team · 9 days ago
  88. 8ebc71c Merge pull request #9812 from ken-unger:f16-softmax2-rvv by XNNPACK Team · 9 days ago
  89. 3b9ec79 Merge pull request #9727 from GregoryComer:f32-bf16-vcvt-scalar by XNNPACK Team · 9 days ago
  90. f3976a3 Update qd8 gemm config to enable c4 microkernels for wasmsdot by Yolanda Chen · 9 days ago
  91. cf89f69 Update qd8 gemm config to enable 4x16c2s2 microkernels for wasm simd by Yolanda Chen · 10 days ago
  92. e0fe33a Merge pull request #9753 from yolanda15:qd8_update_usdot by XNNPACK Team · 10 days ago
  93. 1812bbe Merge pull request #9833 from mohammadmseet-hue:fix/bmm-num-batch-dims-check by XNNPACK Team · 10 days ago chromium/7767 chromium/7768 chromium/7769 chromium/7770 chromium/7771 chromium/7772 chromium/7773 chromium/7774 chromium/7775 chromium/7776 chromium/7777 chromium/7778 chromium/7778_2 chromium/7779 chromium/7780 chromium/7781 chromium/7782 chromium/7783
  94. 0304f05 Fix incorrectly disabled benchmarks and tests by Dillon Sharlet · 10 days ago
  95. d0a338b Fix warnings on some builds by Dillon Sharlet · 10 days ago
  96. 856538e Update build_srcs.bzl for bf16 scalar cvt by Gregory Comer · 10 days ago
  97. e0de998 Add notap tag to manual targets by Dillon Sharlet · 10 days ago
  98. 28e1a13 Fix stack buffer overflow in batch matrix multiply reshape by mohammadmseet-hue · 10 days ago
  99. 3d7a7d6 Fix stack buffer overflow in static_slice and missing bounds checks in copy ops by mohammadmseet-hue · 10 days ago
  100. beea6f7 Remove redundant dim bounds checks by Dillon Sharlet · 10 days ago