Google Git
Sign in
opensecura / 3p / openxla / iree / bb542eee65fa0a498963df1f2ee2f205a3dd8bd0 / tests / e2e / matmul
  1. bb542ee [LLVMGPU] Add Virtual MFMA layout that maximizes load through adjusted K-width (#18930) by Stanley Winata · 5 months ago
  2. 1fc6e5b Add CDNA3 MFMA BF16 intrinsics. (#18892) by Benoit Jacob · 5 months ago
  3. 8ce8bed Simplifications in e2e matmul tests (#18889) by Benoit Jacob · 5 months ago
  4. 225baf2 Add e2e tests for F8E5M2FNUZ and F8E4M3FNUZ data-tiled MFMA on CDNA3 (#18888) by Benoit Jacob · 5 months ago
  5. 0c6a151 Warn when --iree-llvmcpu-target-cpu defaults to "generic". (#18682) by Benoit Jacob · 6 months ago
  6. eb15493 e2e matmul test improvements (#18725) by Benoit Jacob · 6 months ago
  7. 23b63cd [GPU][DT] Add e2e matmul tests for GPU data tiling (#18627) by Max191 · 6 months ago
  8. 84ac47b [LLVMGPU] Switch LLVMGPUVectorDistribute to use iree_gpu.lowering_config (#18651) by Kunwar Grover · 6 months ago
  9. 7050033 [Codegen][GPU] Add support for WMMA_I32_16x16x16_I8 (#18372) by Quinn Dawkins · 7 months ago
  10. 7a7bfe1 [Flow] Move first part of Flow transforms to new pipeline (#18290) by Ian Wood · 7 months ago
  11. c44d29b [compiler] Make cuda/hip/vulkan target cl options consistent (#17710) by Lei Zhang · 7 months ago
  12. 95d5562 [vulkan] Update default RDNA GPU subgroup size to 32 (#18207) by Nithin Meganathan · 8 months ago
  13. e9e24f8 [GPU] Follow the official naming convention for WMMA attributes. (#18147) by Han-Chung Wang · 8 months ago
  14. 31bfc93 [GPU] Fix e2e matmul generator to extract the input element type. (#18140) by Han-Chung Wang · 8 months ago
  15. ca24b96 [GPU] Updates mfma/wmma attribute names in the matmul test generator. (#18134) by Han-Chung Wang · 8 months ago
  16. ae00c4f Revert "Revert "[LLVMGPU][ROCm] Add MFMA_F32_16x16x4_F32 instruction"… (#17921) by Prashant Kumar · 8 months ago
  17. 6a82eb5 Add F8_16x16x32_F32 support for MFMA (#17792) by Stanley Winata · 9 months ago
  18. 2ed3f92 Add nop pass to different backend. by Alan Li · 9 months ago
  19. 02c2000 Revert "[LLVMGPU][ROCm] Add MFMA_F32_16x16x4_F32 instruction" (#17894) by Scott Todd · 9 months ago
  20. d65c6d4 [LLVMGPU][ROCm] Add MFMA_F32_16x16x4_F32 instruction (#17847) by Prashant Kumar · 9 months ago
  21. dcba7c5 [LLVMGPU][ROCm] Plumb through i8, i8 -> i32 MFMA intrinsics (#17764) by Jakub Kuderski · 9 months ago
  22. 695e193 [Codegen] Change CompilationInfoAttr to take a lowering config interface (#17752) by Quinn Dawkins · 9 months ago
  23. 71c07fa [CPU] Signal errors if there are large vectors. (#17620) by Han-Chung Wang · 10 months ago
  24. f7ca45d [ArmSME][test] Enable TransposeMatmulPass and peeling for e2e matmuls (#17452) by Benjamin Maxwell · 11 months ago
  25. 6a9b175 Add test_amd_w7900 CI job including ROCM matmul tests (#17298) by erman-gurses · 11 months ago
  26. 098b0c4 Add test_amd_mi250 CI job including ROCm matmul tests. (#17293) by erman-gurses · 11 months ago
  27. c1cfbfc Add WMMA to matmul test suite for LLVMGPUVectorDistribute (#17285) by Kunwar Grover · 11 months ago
  28. 7b144fc Cleanup: Drop import yaml in generate script (#17253) by Nancy Yuen · 11 months ago
  29. 313c4d7 [LLVMGPU] Remove redundant fields from mma_schedule (#17195) by Kunwar Grover · 11 months ago
  30. 3dde925 [VectorDistribution] Add distribution pattern for vector::MultiDimReductionOp (#17076) by Kunwar Grover · 11 months ago
  31. e8f4948 [DT] Teach encoding about padding. (#17077) by Han-Chung Wang · 12 months ago
  32. 954cb36 Move Codegen pass pipelines to nest on `FunctionOpInterface`. (#16665) by MaheshRavishankar · 12 months ago
  33. 3cce7fc Add the conv2d test generator (for NCHW-FCHW layout). (#16849) by Prashant Kumar · 12 months ago
  34. 2c88e49 [LLVMGPU] Wmma layout for LLVMGPU vector distribute pipeline (#16928) by Stanley Winata · 1 year ago
  35. 0c2552f [CI] Run ArmSME tests under emulator as part of `build_test_all_arm64` (#16331) by Benjamin Maxwell · 1 year, 1 month ago
  36. d7de68a [matmul] Add transpose B matrix coverage for CDNA3 (#16558) by Lei Zhang · 1 year, 1 month ago
  37. 8ec45ca [rocdl] Fix mfma accumulator base vector type (#16549) by Lei Zhang · 1 year, 1 month ago
  38. cf3903c [rocdl] Add e2e matmul test for cdna3 matrix core (#16510) by Lei Zhang · 1 year, 1 month ago
  39. 4237053 NFC: Make e2e matmul test names consistent (#16511) by Lei Zhang · 1 year, 1 month ago
  40. 21566f6 [CPU][ArmSME] Add convert-arith-to-arm-sme to the SME pipeline (#16409) by Benjamin Maxwell · 1 year, 2 months ago
  41. a957fde Add a reference matmul to e2e matmul tests (#16280) by Nirvedh Meshram · 1 year, 2 months ago
  42. d7f0a2e [Codegen] Add a configuration attribute dict to translation info (#16224) by Quinn Dawkins · 1 year, 2 months ago
  43. 7472237 Working around bug with aliasing in-place input->output flows. by Ben Vanik · 1 year, 2 months ago
  44. f2952ab Replacing iree-e2e-matmul-test with one using IR instead of YAML. by Ben Vanik · 1 year, 3 months ago
  45. 6dbc227 [CPU] Retire CPUDoubleTilingPadExpert pipeline. (#15931) by Han-Chung Wang · 1 year, 4 months ago
  46. 0842feb [bf16] Rework vector+bf16 support to avoid invalid conversion (#15911) by Rob Suderman · 1 year, 4 months ago
  47. 605aca9 [CPU][ArmSME] Add (initial) tiling and lowering pipeline for ArmSME (#15794) by Benjamin Maxwell · 1 year, 4 months ago
  48. bcbcd25 Disable all `e2e_matmul_nondt` tests pending compile time fix. (#15821) by Scott Todd · 1 year, 4 months ago
  49. d9c26f1 [CPU] Disable fp16 matmul tests for non-dt test suite. (#15801) by Han-Chung Wang · 1 year, 4 months ago
  50. 9b2dfdb [CPU] Add a matmul test suite for data-tiling codegen. (#15738) by Han-Chung Wang · 1 year, 4 months ago
  51. 2eda767 Migrate tests and benchmarks from `--iree-llvmcpu-enable-microkernels` to `--iree-llvmcpu-enable-ukernels` (#15584) by bjacob · 1 year, 5 months ago
  52. 3a3c1a4 Fix `fp16` feature on arm64: the proper feature name is `fullfp16`, not `fp16`. (#15479) by bjacob · 1 year, 5 months ago
  53. 9cc729f [AArch64][SVE] Add e2e tests for small and large matmuls (#15292) by Benjamin Maxwell · 1 year, 5 months ago
  54. aa5602d Improvements to e2e matmul tests (take 2) (#15259) by bjacob · 1 year, 5 months ago
  55. eb9b8b6 Revert "Improvements to e2e matmul tests" (#15252) by bjacob · 1 year, 6 months ago
  56. 71c22da Improvements to e2e matmul tests (#15243) by bjacob · 1 year, 6 months ago
  57. 1a63564 Refactor IREECodegenAttrs to use typed array parameters (#15032) by Benjamin Maxwell · 1 year, 6 months ago
  58. 699b34c [vulkan] Add e2e coop matrix f16 matmul test (#15058) by Jakub Kuderski · 1 year, 6 months ago
  59. 73b04d3 Enable ConstEval for CPU data tiling path. (#14792) by Han-Chung Wang · 1 year, 7 months ago
  60. 09685ee data-tiling: introduce `upper_bound_tile_size` op to defer padding-size choice to MaterializeEncoding. (#14349) by bjacob · 1 year, 9 months ago
  61. d9674d8 Tag `e2e_matmul_direct_f16_gpu_large_unaligned` as requiring sm80 (#14266) by Geoffrey Martin-Noble · 1 year, 9 months ago
  62. 6200ade NFC - Refactor Matmul and GemmLike strategies in preparation for gene… (#14201) by Nicolas Vasilache · 1 year, 9 months ago
  63. d322009 Correctly tag matmul tests requiring sm80 (#14173) by Geoffrey Martin-Noble · 1 year, 10 months ago
  64. be24f02 Use Black to format Python files (#14161) by Jakub Kuderski · 1 year, 10 months ago
  65. 29e8ed0 Remove redundant `requires-gpu-nvidia` added in #14039 (#14076) by Geoffrey Martin-Noble · 1 year, 10 months ago
  66. 2dddf02 Correctly tag Vulkan Ampere tests as requiring sm80 (#14039) by Geoffrey Martin-Noble · 1 year, 10 months ago
  67. 0e03852 Test with ASAN in bytecode modules (#14005) by bjacob · 1 year, 10 months ago
  68. e23561d Improvements to target CPU features variants in e2e tests (#13915) by bjacob · 1 year, 10 months ago
  69. 1038648 [TransformStrategies] Add support for aligned and partially aligned matmul (#13541) by Quinn Dawkins · 1 year, 10 months ago
  70. bf8588e Microkernels: add arm64 bitcode. Test everywhere. (#13846) by bjacob · 1 year, 10 months ago
  71. 1423d09 Enable AVX2+FMA in e2e matmul + ukernels test; support comma-separated CPU features. (#13837) by bjacob · 1 year, 10 months ago
  72. f1545bb move CUDA check tests out of e2e/matmul (#13805) by bjacob · 1 year, 10 months ago
  73. 61551c0 [LLVMGPU] Turn on TD strategy for unaligned matmul (#13492) by Thomas · 1 year, 10 months ago
  74. 3e2d243 limit ukernel bitcode tests to x86-64 (#13710) by bjacob · 1 year, 11 months ago
  75. 29647b3 CPU ukernels as bitcode (x86-only for now) (#13460) by MaheshRavishankar · 1 year, 11 months ago
  76. a8daf0e e2e matmul test improvements (#13657) by bjacob · 1 year, 11 months ago
  77. 475af42 [spirv][vulkan] Add f16 e2e matmul tests (#13327) by Jakub Kuderski · 2 years ago
  78. dc1684d [spirv][vulkan] Run e2 i8 matmul tests in CI (#13312) by Jakub Kuderski · 2 years ago
  79. 0b92d79 Enable passing tests on CPU. (#13147) by Han-Chung Wang · 2 years ago
  80. 7758e48 Add a `noriscv` test label, skip in RISC-V emulator tests (#12854) by bjacob · 2 years ago
  81. ab0f86a Improved fine-grained Instruction Pipelining for F32 Tensor Cores (mma.sync) (#12761) by Manish Gupta · 2 years ago
  82. d1e7ef8 Only exclude e2e_matmul_mmt4d_i8_small on RISC-V. (#12764) by Han-Chung Wang · 2 years ago
  83. 8144cba Disable f32 mma.sync test until fine-grained schedule is debugged (#12751) by Manish Gupta · 2 years ago
  84. e6addfa Integrate llvm-project and bump dependencies. (#12653) by Manish Gupta · 2 years ago
  85. 8f5ced9 CMake IREE_ARCH variable, a canonicalized CMAKE_SYSTEM_PROCESSOR (#12687) by bjacob · 2 years ago
  86. 8626fdb Rename all Bazel BUILD files to BUILD.bazel (#12663) by Geoffrey Martin-Noble · 2 years, 1 month ago
  87. e2151d3 [LLVMGPU][NFC] Break up mma.sync into its own codegen pipeline (#12582) by Thomas · 2 years, 1 month ago
  88. 380bde7 Renaming `--iree-llvm-` CPU flags to `--iree-llvmcpu-`. by Ben Vanik · 2 years, 1 month ago
  89. 9691c91 Functional support mma.sync.1688.f32.tf32 for F32 datatype (#12054) by Manish Gupta · 2 years, 2 months ago
  90. 2231682 Adds Native Tensor Core (F16) Support [mma.sync.16816.f16.f16 and ldmatrix] (#11817) by Manish Gupta · 2 years, 2 months ago
  91. 3dd670f Switch e2e/matmul tests on vmvx+ukernel to data-tiling (#11522) by bjacob · 2 years, 3 months ago
  92. 52c2e35 Relands "Switch e2e tests and benchmarks to `--iree-flow-enable-data-tiling` on `llvm-cpu`" (#11443) by Han-Chung Wang · 2 years, 4 months ago
  93. b3fa021 Revert "Switch e2e tests and benchmarks to `--iree-flow-enable-data-tiling` on `llvm-cpu`" (#11435) by Jerry Wu · 2 years, 4 months ago
  94. c781d6a Re-enabling cuda's split-k test (#11431) by Murali Vijayaraghavan · 2 years, 4 months ago
  95. 044017f Switch e2e tests and benchmarks to `--iree-flow-enable-data-tiling` on `llvm-cpu` (#11411) by bjacob · 2 years, 4 months ago
  96. 323305c Integrate llvm-project at 5cfc22cafe3f and bump dependencies (#11275) by Kojo Acquah · 2 years, 4 months ago
  97. 242ffbb [NFC] Remove trailing whitespaces. (#11107) by Han-Chung Wang · 2 years, 5 months ago
  98. 0a6cdf0 Add support for GEMM e2e Test For CUDA backend on F16 input (#10842) by Manish Gupta · 2 years, 5 months ago
  99. 8f39d27 Integrate llvm-project at b9898e7ed1ce and bump dependencies (#10740) by Thomas · 2 years, 6 months ago
  100. 2500a0a [spirv] NFC: Rename existing pipelines to be consistent (#10735) by Lei Zhang · 2 years, 6 months ago