Google Git
Sign in
opensecura/3p/openxla/iree/HEAD
  1. 79183cb [Bazel] Move HAL materializeConstant to break IR/HALDialect circular dependency (#24203) by RattataKing · 4 hours ago main
  2. 4bd3742 [LinalgExt] Rewriter for Torch::HigherOrderFlexAttentionOp -> LinalgExt::OnlineAttentionOp (#23292) by Keshav Vinayak Jha · 5 hours ago
  3. 702603f [Codegen][LinalgExt] Port attention rewrites to OnlineAttentionOp (#24123) by Keshav Vinayak Jha · 5 hours ago
  4. f3bc127 [DispatchCreation] Check all operands of a consumer before fusing (#24207) by Keshav Vinayak Jha · 5 hours ago
  5. 00dc51c Bump LLVM at llvm/llvm-project@93dce0bf4332 (#24185) by Han-Chung Wang · 24 hours ago latest-snapshot
  6. 43918cd [Codegen] Replace DMA fallback warning with LDBG (#24192) by Zhewen Yu · 25 hours ago
  7. 7011cb8 [LinalgExt] Add TilingInterface support for TopkV2Op (#24129) by Bangtian Liu · 28 hours ago
  8. 8b85022 [Codegen] Enable DMA by default for F16/BF16 Gemm on gfx950 (#24117) by Zhewen Yu · 30 hours ago
  9. f629f21 [LLVMGPU] Legalize ND vectors to 1D vectors (#24028) by Erick Ochoa Lopez · 30 hours ago
  10. 7c9ec25 [LinalgExt] Add scalar implementation for TopkV2Op (#24146) by Bangtian Liu · 2 days ago
  11. ea23432 samples: Refactor input handling with data-driven approach (#23955) by Han-Kuan Chen · 2 days ago
  12. 8306ce0 [Codegen][NFC] Add full pipeline smoke tests for each backend. (#24176) by Han-Chung Wang · 2 days ago
  13. a1ca1cb [GPU] Allow swizzle promotion for transpose linalg.generic producers (preshuffling prs 1/3) (#24086) by Muzammiluddin Syed · 2 days ago
  14. aeabe9c Integrate llvm/torch-mlir@66bac5f (#24168) by Ian Wood · 2 days ago
  15. 5bdbc9c Bump LLVM to llvm-project@2c8c2bdf6aa8 (#24170) by Han-Chung Wang · 2 days ago
  16. 06a5450 [DT] Disable sinking propagation if broadcasting from encoded tensors. (#24171) by Han-Chung Wang · 2 days ago
  17. 32b074e [LinalgExt] Add IndexingMapOpInterface to ArgCompareOp (#24173) by Bangtian Liu · 2 days ago
  18. 7e54116 Revert "[GPUHeuristics] Improve large GEMM intrinsic selection on CDNA4 (#24115)" (#24172) by Vivian Zhang · 2 days ago
  19. 6f099e9 [tools] Avoid noisy failures in IREE bash completion. (#24111) by Ian Wood · 2 days ago
  20. 49c2bcd Fix LinalgFusionOpInterface adaptor for linalg ops (#24159) by Ian Wood · 2 days ago
  21. 50af64b [CI] Fix BYO LLVM build (#24152) by Zhewen Yu · 3 days ago
  22. 3bbf1da [HAL] Renumber export ordinals after pruning (#24162) by Jorn Tuyls · 3 days ago
  23. 63b9cce Integrates/llvm@91339fd (#24153) by Zhewen Yu · 4 days ago
  24. 909afc9 [LLVMGPU] Remove C promotion changes which can now be solved with MapLoad (#23735) by Abhishek Varma · 5 days ago
  25. a89bb74 [LLVMGPU] Add CombineSourceLayoutTransformation to TileAndFuse pipeline (#24158) by Abhishek Varma · 5 days ago
  26. c29cd05 [CI] Fix gfx1100 pipeline test to not require rocm backend plugin (#24151) by Zhewen Yu · 5 days ago
  27. 47d41ab [GPUHeuristics] Improve large GEMM intrinsic selection on CDNA4 (#24115) by Vivian Zhang · 5 days ago
  28. 89b536e [GPU][Codegen] Add skinny gemm vdmfma for gfx950 (#24102) by Eric Feng · 5 days ago
  29. a887ad5 [SPIRV][nfc] Adapt pipeline tests on module-scope. (#24143) by Han-Chung Wang · 5 days ago
  30. bbf54e4 [Codegen][PCF] Add foldForallIntoPCFLoop for split-k loop handling (#23452) by Quinn Dawkins · 5 days ago
  31. ce2de46 [runtime] Restore "any processor" fallback for unspecified Windows … (#24154) by Andrew Woloszyn · 5 days ago
  32. d53f2b6 [Common] Update ComplexChain definition in CombineSourceLayoutTransform (#23860) by Abhishek Varma · 5 days ago
  33. 899bc48 [LinalgExt] Verify input/output indices element types match in topk_v2 (#24147) by Bangtian Liu · 5 days ago
  34. 552122f [DMA] Reject invalid pads early in DMA convertible check (#24132) by Zhewen Yu · 5 days ago
  35. 2287440 Revert "[SPIRV] Fix ub.poison legalization failure for 2d vectors (#23789)" (#24140) by Zhewen Yu · 6 days ago
  36. 273592d [TensorExt] Add LinearizeRaggedDimsOp (#24101) by MaheshRavishankar · 6 days ago
  37. 99ac461 Add data tiling hints for convolutions (#24061) by Jelle Schühmacher · 6 days ago
  38. 62872f4 [Codegen][GPU] Add DistributeArgCompare pattern (#23793) by Bangtian Liu · 6 days ago
  39. 551ef31 Bump torch-mlir and stablehlo (#24141) by Rob Suderman · 6 days ago
  40. 3374771 [VMVX] Switch to module-scope pipeline. (#24133) by Han-Chung Wang · 6 days ago
  41. bc6d56d Integrates/llvm@c463b75 (#24137) by Zhewen Yu · 6 days ago
  42. 9ed4a4e [Codegen] Move codegen pipelines to work on online_attention (#24110) by Kunwar Grover · 6 days ago
  43. 78d4a2c [LinalgExt] Add mask operand to linalg_ext gather/scatter (#24126) by Kunwar Grover · 6 days ago
  44. 1c72ccf [Codegen] Add transfer_{gather/scatter} to vector.{gather/scatter} lowering (#24104) by Nouman Amir · 6 days ago
  45. 1bf33a8 Integrates torch-mlir@762ccbc (#24124) by Abhishek Varma · 7 days ago
  46. 192602f [DispatchCreation] Limit attention broadcast fusion (#24127) by Ian Wood · 7 days ago
  47. bae8f56 [Codegen][GPU] Guard multi-buffering behind pipelining pre-flight check (#24114) by Zhuoran Yin · 7 days ago
  48. 8789483 [NFC] Clean up collapsed AttentionOp builder use (#24128) by Ian Wood · 7 days ago
  49. ad66dc0 [GPU][nfc] Migrate the remaining pipeline test to module-scope test. 3/3 (#24118) by Han-Chung Wang · 7 days ago
  50. f3220d8 [DispatchCreation] Fuse producers with multiple users (#24050) by Ian Wood · 7 days ago
  51. ebb8bdc Integrates/llvm@738ead2 (#24120) by Zhewen Yu · 7 days ago
  52. 68f5d46 [DMA][Swizzle] Enforce access_width >= elementsPerLane in DMA lowering (#24113) by Alan Li · 7 days ago
  53. 041f412 Fix ExpandDestinationForallOp rejecting non-store users (#24073) by Jorn Tuyls · 7 days ago
  54. 4f10faa [LLVMGPU][nfc] Modernize rocdl pipeline tests to module-scope tests. 2/n (#24099) by Han-Chung Wang · 8 days ago
  55. 3ca47ca Add strnlen fallback for newlib in thread.c (#24026) by Han-Kuan Chen · 8 days ago
  56. 1ff9203 [LLVMGPU] Add `includeLLVMLowering` option to pipeline and migrate tests 1/n (#24080) by Han-Chung Wang · 8 days ago
  57. cffd309 [Codegen] Do not fuse parallel ops if any of users is not LinalgOp. (#24066) by Han-Chung Wang · 8 days ago
  58. 15ca40a [Dispatch Creation] Fix crash on tensor.empty (#24112) by Ian Wood · 8 days ago
  59. 39d380d [CI] Enable sccache on windows. (#24096) by Erick Ochoa Lopez · 8 days ago
  60. 78d7b1e [CPU] Fix tile sizes based on number of threads. (#24108) by Benoit Jacob · 8 days ago
  61. 15008e0 [CPU] Stop scaling PackOp distribution tiles by inner pack sizes (#24107) by Benoit Jacob · 8 days ago
  62. 7ead77e [CPU] Factor thread utilization into mmt4d lowering tile selection (#24109) by Benoit Jacob · 8 days ago
  63. f6439d4 [CPU] Block root op chain fusion in non-root anchored tiling. (#24098) by Han-Chung Wang · 8 days ago
  64. b7f1cf2 [LinalgExt] Add topk_v2 op with roundtrip and invalid mlir test (#24054) by Bangtian Liu · 8 days ago
  65. 7ca09a2 [Codegen] Add XOR swizzle for BF16 matmul with DMA (#23932) by Zhewen Yu · 8 days ago
  66. 65f88b5 Bump the github-actions group with 4 updates (#24097) by dependabot[bot] · 8 days ago
  67. 259d9d2 Integrates/llvm@c838322 (#24105) by Zhewen Yu · 8 days ago
  68. d87d49e [Codegen][PCF] Add convertForallToGenericNest transform (#23332) by Quinn Dawkins · 9 days ago
  69. 58f594f [DMA][Swizzle] Apply inverse source swizzle in DMA lowering (#24094) by Alan Li · 9 days ago
  70. 5b4b421 [TensorExt] Add resolveRange and getDistributionInfo to SparseCastOpInterface (#23989) by MaheshRavishankar · 9 days ago
  71. d26baca [Codegen] Relax identity layout check in ResolveSwizzleHints verification (#24091) by Zhewen Yu · 9 days ago
  72. f299db8 [LinalgExt] Move ArgCompareOp region verification to verifyRegions (#24095) by Bangtian Liu · 9 days ago
  73. 705d3f4 [GPU] Implement CDNA block intrinsics (6/6) (#24020) by Nirvedh Meshram · 9 days ago
  74. 3555b91 Integrates/llvm@9a5bc72 (#24090) by Zhewen Yu · 9 days ago
  75. b631b00 [Preprocessing] Register backward-data-conv preprocessing pipeline (#24055) by Vivian Zhang · 11 days ago
  76. f737a3e Integrate llvm-project to 15e46e25f26c. (#24077) by Muzammiluddin Syed · 12 days ago
  77. f4c5153 [CPU] Improve splitReduction for bounded-dynamic cases. (#24067) by Han-Chung Wang · 12 days ago
  78. bcd9618 [Codegen] Use createWriteOrMaskedWrite utility (#24069) by Lukas Sommer · 12 days ago
  79. bf4615d [LLVMGPU] Enable swizzle absorb/reinsert passes (#24002) by Zhewen Yu · 12 days ago
  80. 3717052 [Codegen] Extend generated E2E attention tests (#23342) by Lukas Sommer · 12 days ago
  81. 5ef8ca1 [VectorDistribute] Masking pipeline tests (#24044) by Lukas Sommer · 12 days ago
  82. 40a49b8 [Codegen] Fix BufferizeDispatchTensorLoadStore to handle nested ops (#24048) by Zhewen Yu · 13 days ago
  83. 20dbcb7 [Codegen] Add pass to reinsert swizzle hints from alloc attributes (#24001) by Zhewen Yu · 13 days ago
  84. 4aaca88 [Codegen] Add pass to absorb swizzle hints into alloc attributes (#24000) by Zhewen Yu · 13 days ago
  85. 3d3cb37 [Codegen] Deprecate the old ExportOp path from PropagateDispatchSizeBounds (#24038) by Han-Chung Wang · 13 days ago
  86. 8b36422 [Codegen] Create the temp DispatchConfigOp with FromSliceOp. (#24037) by Han-Chung Wang · 13 days ago
  87. cdfa628 Integrate llvm-project to 2c0078b22469. (#24062) by Muzammiluddin Syed · 13 days ago
  88. 0ed80b3 [LinalgExt] Fix FuseTransposeWithAttentionOp skipping valid operands (#24059) by Ian Wood · 13 days ago
  89. 37e17e2 Bump iree-org/torch-mlir@7629fd5363ee1ec1a18f94652e237ffba8d54860 (#24063) by Keshav Vinayak Jha · 13 days ago
  90. b518b35 [Preprocessing] Add pass to swap strided insert_slice with contraction (#23979) by Vivian Zhang · 14 days ago
  91. dfddea4 Integrates/llvm@4b00302c99e0. (#24047) by Muzammiluddin Syed · 2 weeks ago
  92. 55c8a43 [SPIRV] Add native shapes for step op (#24046) by Erick Ochoa Lopez · 2 weeks ago
  93. 034d5ef [CI] Updating torch vae test golden time (#24052) by Muzammiluddin Syed · 2 weeks ago
  94. df67eb8 [Preprocessing] Add pass to convert strided insert_slice to linalg.generic (#23990) by Vivian Zhang · 2 weeks ago
  95. 53dc660 [DispatchCreation] Skip collapse for scatter-like generics with tensor.extract (#24034) by Vivian Zhang · 2 weeks ago
  96. 94170e7 [LinalgExt] Implement direct vectorization for im2col op (#23855) by Max191 · 2 weeks ago
  97. 007b1ee [Codegen] Fix transfer_scatter failing tests (#24043) by Kunwar Grover · 2 weeks ago
  98. c51fec1 [Codegen] Support more operations in tile size analysis (#23971) by Lukas Sommer · 2 weeks ago
  99. d539fd4 [Codegen][VectorExt] Add TransferScatterOp unrolling support (#23704) by Keshav Vinayak Jha · 2 weeks ago
  100. 485e401 [Codegen] Masked vectorization of inner_tiled (#23935) by Lukas Sommer · 2 weeks ago