Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / XNNPACK / 20c3b921a15dc47c6344ac3f5c43a21fb30c558d
  1. 20c3b92 Size test for Subgraph API by Marat Dukhan · 4 years, 8 months ago
  2. 64debce Cortex-A55 cpu detect and select 6x8 GEMM and IGEMM microkernels. 9% faster end to end on mobile net v2. 16.5% faster on mobile net v1. 6.4% faster on mobile net v3. by Frank Barchard · 4 years, 8 months ago
  3. fe7acb6 Targets for requantization tests and benchmarks by Marat Dukhan · 4 years, 8 months ago
  4. 91e1999 6x8 GEMM and IGEMM microkernels for Cortex A55. 9% faster end to end: by Frank Barchard · 4 years, 8 months ago
  5. 303ad00 Remove redundant dependency in NEON micro-kernels by Marat Dukhan · 4 years, 8 months ago
  6. 4245f43 Update size_test with newly added operators by Marat Dukhan · 4 years, 8 months ago
  7. 9fd7e25 Use cpuinfo_get_max_cache_size in bench-utils by Marat Dukhan · 4 years, 9 months ago
  8. b6862d1 Cleanup unused SoftMax benchmark presets by Marat Dukhan · 4 years, 9 months ago
  9. f42facc Completely remove NaCl and PNaCl support by Marat Dukhan · 4 years, 9 months ago
  10. 6206b4a Fix XNNPACK build on Linux/Android with Bazel by Marat Dukhan · 4 years, 9 months ago
  11. 8d3c693 Optional SoftMax benchmarks vs DNNL by Marat Dukhan · 4 years, 9 months ago
  12. cc51633 Update cpuinfo dependency by Marat Dukhan · 4 years, 9 months ago
  13. 49a5969 More descriptive entry point names for operators with groups by Marat Dukhan · 4 years, 9 months ago
  14. 892687b Fix typo in comment by Marat Dukhan · 4 years, 9 months ago
  15. f092a4a f32-maxpool microkernel for ARM Neon. by Frank Barchard · 4 years, 9 months ago
  16. 488c334 Update status of PyTorch integration in the readme by Marat Dukhan · 4 years, 9 months ago
  17. 818d1d4 Merge pull request #383 from mattn:update-pthreadpool by XNNPACK Team · 4 years, 9 months ago
  18. 4ea95be Update pthreadpoool for support Windows by Yasuhiro Matsumoto · 4 years, 9 months ago
  19. 5b3185b Merge pull request #381 from mattn:build-windows by XNNPACK Team · 4 years, 9 months ago
  20. 081b77c Adding -fno-asynchronous-unwind-tables for MinGW by Yasuhiro Matsumoto · 4 years, 9 months ago
  21. 21d3bd6 Support Average Pooling and Max Pooling in Subgraph API by Marat Dukhan · 4 years, 9 months ago
  22. dac1a0e Fix macro for exception to disable exception by Yasuhiro Matsumoto · 4 years, 9 months ago
  23. 462be05 Build on Windows by Yasuhiro Matsumoto · 4 years, 9 months ago
  24. 466da75 Support TF SAME padding in Average Pooling operator by Marat Dukhan · 4 years, 9 months ago
  25. bee7825 Support TF SAME padding in Max Pooling operator by Marat Dukhan · 4 years, 9 months ago
  26. 96171aa Refactor Average Pooling operator by Marat Dukhan · 4 years, 9 months ago
  27. ee1f63e Support input_offset in AVGPOOL and PAVGPOOL micro-kernels by Marat Dukhan · 4 years, 9 months ago
  28. f5fec4b Fix MAXPOOL test generator for two-pass subtile and multi-pass test cases by Marat Dukhan · 4 years, 9 months ago
  29. 6ee435a Refactor AVGPOOL & PAVGPOOL micro-kernels & unit tests by Marat Dukhan · 4 years, 9 months ago
  30. 5cb8ff0 Re-generate F32 MAXPOOL micro-kernel tests by Marat Dukhan · 4 years, 9 months ago
  31. 50bb910 Update to upstream version of cpuinfo by Marat Dukhan · 4 years, 9 months ago
  32. 7493bfb Add Qualcomm Krait cpu detect and initial kernel. by Frank Barchard · 4 years, 9 months ago
  33. 15d1f51 Clarify input_padding_* arguments with XNN_FLAG_TENSORFLOW_SAME_PADDING flag by Marat Dukhan · 4 years, 9 months ago
  34. c8230a4 Remove output_min & output_max arguments in PReLU operator by Marat Dukhan · 4 years, 9 months ago
  35. 3ba868c Update psimd dependency by Marat Dukhan · 4 years, 9 months ago
  36. 7353eea Fix typo in initialization of GEMM micro-kernels for WAsm SIMD by Marat Dukhan · 4 years, 9 months ago
  37. 2995427 Optimized Indirect Deconvolution algorithm for 1x1 subconvolutions by Marat Dukhan · 4 years, 9 months ago
  38. 5b5a062 Use generic LD64 GEMM/IGEMM micro-kernels on Cortex-A5/A7 by Marat Dukhan · 4 years, 9 months ago
  39. efc184a Properly set XNN_ENABLE_ASSEMBLY macro in CMake builds by Marat Dukhan · 4 years, 9 months ago
  40. b00004d 4x2c4 GEMM micro-kernels for PSIMD and SSE by Marat Dukhan · 4 years, 9 months ago
  41. 74427b5 Update Raspberry Pi performance by Marat Dukhan · 4 years, 9 months ago
  42. f85fc33 Fix iOS build with Bazel by Marat Dukhan · 4 years, 9 months ago
  43. 52261a8 IGEMM for Cortex-A75 aarch32 pad stack so vector push is aligned. by Frank Barchard · 4 years, 9 months ago
  44. 775d349 Detect Cortex-A53/A55 and use IGEMM optimized for Cortex-A53 on aarch32 by Frank Barchard · 4 years, 9 months ago
  45. 16d7272 ks loop use B.HI instead of B.NE to avoid bugs causing infinite loop. by Frank Barchard · 4 years, 9 months ago
  46. d6f77a6 Fix bug in computing batch size in activations by Marat Dukhan · 4 years, 9 months ago
  47. c87a8fd Cortex A53 IGEMM 32 bit ARM by Frank Barchard · 4 years, 9 months ago
  48. c1a0697 Replace load with mov for ks in xnn_f32_igemm_ukernel_4x8__aarch32_neon_cortex_a75 by Frank Barchard · 4 years, 9 months ago
  49. 9b499d6 Load parameters in order of usage. by Frank Barchard · 4 years, 9 months ago
  50. 3f90adb Document machine learning frameworks using XNNPACK by Marat Dukhan · 4 years, 9 months ago
  51. 1498d1d Support iOS in Bazel builds by Marat Dukhan · 4 years, 9 months ago
  52. 52bd86f Support Clamp, HardSwish, Sigmoid, Softmax in Subgraph API by Marat Dukhan · 4 years, 9 months ago
  53. 8155854 Direct branch to source remainder handler for GEMM/IGEMM. by Frank Barchard · 4 years, 9 months ago
  54. 79ade18 LD64 microkernels branch directly to remainder if less than 2 channels. by Frank Barchard · 4 years, 9 months ago
  55. 7cdeede CPU detect and select Cortex-A75 IGEMM for 32 bit ARM. by Frank Barchard · 4 years, 9 months ago
  56. fd262e1 IGEMM 4x8 LD64 pad stack with 4 bytes to align to 8 bytes by Frank Barchard · 4 years, 9 months ago
  57. 90ce789 Cortex A75 IGEMM 32 bit ARM. by Frank Barchard · 4 years, 9 months ago
  58. 90dff80 Work around MemorySanitizer alarm in AVX512 DWCONV micro-kernels by Marat Dukhan · 4 years, 9 months ago
  59. dc38f07 LD64 IGEMM 32 bit ARM by Frank Barchard · 4 years, 9 months ago
  60. 2fd2ba1 Support PReLU operator in Subgraph API by Marat Dukhan · 4 years, 9 months ago
  61. 54dcb46 Support Add2 and Multiply2 Nodes in Subgraph API by Marat Dukhan · 4 years, 9 months ago
  62. 01a1e98 Fix data pointers for static inputs in Subgraph API by Marat Dukhan · 4 years, 9 months ago
  63. dd2b588 Improve Subgraph API documentation by Marat Dukhan · 4 years, 10 months ago
  64. bdb56f5 FP16 versions of SpMM micro-kernels by Marat Dukhan · 4 years, 10 months ago
  65. d16d0f0 Work around msan failured in AVX micro-kernels by Marat Dukhan · 4 years, 10 months ago
  66. 022c659 Support multithreading in Subgraph API by Marat Dukhan · 4 years, 10 months ago
  67. 484bfc4 Kryo, Cortex-A77 and Exynos-M5 use IGEMM 4x8 with no prefetch. by Frank Barchard · 4 years, 10 months ago
  68. fbbb40e Add Cortex-A77 and Exynos-M5 to init. by Frank Barchard · 4 years, 10 months ago
  69. 7278a95 Fix description in README by Marat Dukhan · 4 years, 10 months ago
  70. 8170d37 Merge pull request #329 from xta0:master by XNNPACK Team · 4 years, 10 months ago
  71. 4d402b4 Fix another typo bug in runtime initialization by Marat Dukhan · 4 years, 10 months ago
  72. ca27b40 Add pipelined to gemm tests for aarch32. by Frank Barchard · 4 years, 10 months ago
  73. 93d61a7 Fix typo bug in XNNPACK runtime construction by Marat Dukhan · 4 years, 10 months ago
  74. a7b22c1 Fix wrong MR specifications in IGEMM benchmark by Marat Dukhan · 4 years, 10 months ago
  75. 74208c8 Clean up whitespaces by Tao Xu · 4 years, 10 months ago
  76. 8426814 Fix iOS build by Tao Xu · 4 years, 10 months ago
  77. ccfdbd1 Fix internal dependencies of XNNPACK by Marat Dukhan · 4 years, 10 months ago
  78. 6866099 Harden DWCONV micro-kernels with extra asserts by Marat Dukhan · 4 years, 10 months ago
  79. 98ca635 CMake build for Subgraph API by Marat Dukhan · 4 years, 10 months ago
  80. 1d75a54 Subgraph API by Marat Dukhan · 4 years, 10 months ago
  81. 03bc407 Reformat tabulation in xnnpack.h header by Marat Dukhan · 4 years, 10 months ago
  82. 9c0db96 F32 Softmax operator benchmark by Marat Dukhan · 4 years, 10 months ago
  83. 49adc2d Document Softmax support in README by Marat Dukhan · 4 years, 10 months ago
  84. fd8e689 Rename SoftArgMax operator to SoftMax by Marat Dukhan · 4 years, 10 months ago
  85. 1edc454 SoftArgMax operator by Marat Dukhan · 4 years, 10 months ago
  86. e104aa3 Remove extraneous argument from mobilenetv3 igemm benchmark. by Frank Barchard · 4 years, 10 months ago
  87. 8137e4c NEON/NEONFMA RAddStoreExpMinusMax micro-kernels by Marat Dukhan · 4 years, 10 months ago
  88. b39689d SSE2/PSIMD RAddStoreExpMinusMax micro-kernels by Marat Dukhan · 4 years, 10 months ago
  89. f46f675 Scalar RAddStoreExpMinusMax micro-kernels by Marat Dukhan · 4 years, 10 months ago
  90. d6ebf0c 6x8 a53 use X8 for GPR shadow register. Eliminate GPR push/pop by Frank Barchard · 4 years, 10 months ago
  91. 534375d A53 GEMM / IGEMM kernel prefetches adjust by 1 by Frank Barchard · 4 years, 10 months ago
  92. c03b2bd 4x12 A53 GEMM and IGEMM use X8 for temp GPR by Frank Barchard · 4 years, 10 months ago
  93. 640b35d Add link to the two-pass softmax paper in README by Marat Dukhan · 4 years, 10 months ago
  94. 324f2bb 4x8 A53 GEMM use X4 for temp GPR Saves a push/pop of X19. by Frank Barchard · 4 years, 10 months ago
  95. 7693acf 4x8 Cortex-A53 GEMM / IGEMM use 1 GPR instead of 2. by Frank Barchard · 4 years, 10 months ago
  96. f884a7b 6X8 Cortex-A53 GEMM use 1 GPR instead of 2. by Frank Barchard · 4 years, 10 months ago
  97. 54afb13 Use 2 GPR registers instead of 4 for GPR loads. by Frank Barchard · 4 years, 10 months ago
  98. cbb35d0 4x12 IGEMM use prefetch on A and B by Frank Barchard · 4 years, 10 months ago
  99. 3216758 6X8 Cortex-A53 IGEMM use 1 GPR instead of 2. by Frank Barchard · 4 years, 10 months ago
  100. 4cd8907 4x12 A53 kernel use prefetches on A by Frank Barchard · 4 years, 10 months ago