Google Git
Sign in
opensecura / 3p / openxla / iree / HEAD
  1. e1f3811 [Codegen][Common] Allow generic conv ops to decompose to lower dim ops (#23294) by Abhishek Varma · 32 hours ago main
  2. e4ab789 Moves threading/synchronization to iree/base/threading/ for public use. (#23337) by Ben Vanik · 2 days ago
  3. 0d71d25 [LinalgExt] Add OuterReduction tiling strategy for ArgCompareOp (#23102) by Bangtian Liu · 3 days ago latest-snapshot
  4. 41cae6d [DispatchCreation] Support split reduction on weight backward CNHW layout (#23343) by Vivian Zhang · 3 days ago
  5. 8659263 [CI] Update iree-test-suite to run convolution tests (#23312) by Erick Ochoa Lopez · 3 days ago
  6. fce95cd [LLVMCPU] Allow generic conv ops too in KernelDispatch (#23295) by Abhishek Varma · 3 days ago
  7. 18fe092 [TensorExt] Propagate and fold iree_tensor_ext.bitcast (#23182) by Zhewen Yu · 3 days ago
  8. c0885ee [StableHLO] Rework zero-extent canonicalizer (#23227) by Lukas Sommer · 3 days ago
  9. 3824de7 iree-bazel-* improvements for handling multiple targets + options. (#23330) by Ben Vanik · 3 days ago
  10. c6ac63e [CI][Torch models] Update golden dispatch counts (#23336) by Eric Feng · 4 days ago
  11. ab71aab Integrate llvm/llvm-project@db6b186e (#23328) by Ian Wood · 4 days ago
  12. 4f92211 Switching iree_time_now() to CLOCK_MONOTONIC + cleanup. (#23325) by Ben Vanik · 4 days ago
  13. 8c014e5 [Dispatch Creation] Restrict broadcasting consumer fusion (#23232) by Ian Wood · 4 days ago
  14. 79166ed [Integrate] Prioritize IREE's PackOp/UnPackOp bufferization implementation. (#23326) by Han-Chung Wang · 4 days ago
  15. 24af126 [plugins][Torch] Add a flag to externalize transients during torch-to-IREE conversion (#23224) by zjgarvey · 4 days ago
  16. ea72d00 [GPU] Don't swap expand with slice in the same block (#23267) by Max191 · 4 days ago
  17. 34d2446 [Stream] Verify `flow.executable` before lowering (#23262) by Lukas Sommer · 4 days ago
  18. 8f8b29c [CI] Try reenabling onnx ops tests for rdna4 (#23324) by Jakub Kuderski · 4 days ago
  19. 602aa40 [Codegen][GPU] Add subgroup and lane execution scopes (#23097) by Quinn Dawkins · 4 days ago
  20. 96f3974 [GPU] Support some collapse_shape ops in GPUReduceBankConflicts (#23301) by Max191 · 4 days ago
  21. d88f560 [Stream] Restrict import of dynamically shaped tensors (#23214) by Lukas Sommer · 4 days ago
  22. a7328f7 [Input] Do not attempt to convert LLVM dialect functions (#23299) by Lukas Sommer · 4 days ago
  23. 709ab4f [LinalgExt] Add pattern to canonicalize identity map_gather into a linalg.copy (#23240) by Abhishek Varma · 4 days ago
  24. e968799 [GPU] MmaSchedule configuration crashes when lacking PerfTflops (#23303) by Rob Suderman · 5 days ago
  25. ea01b8a [AMDGPU][LDS] Support linearized DMA for small innermost dimensions (#23056) by Alan Li · 5 days ago
  26. fc06a5d [GPU] Add workgroupMemoryBankCount parameter to TargetWgpAttr (#23273) by Muzammiluddin Syed · 5 days ago
  27. 315cac2 [compiler][NFC] Follow camelCase naming convention. (#23316) by Han-Chung Wang · 5 days ago
  28. 31f794a Integrate llvm/llvm-project@3446ff1 (#23302) by Ian Wood · 5 days ago
  29. 36ac5f4 Revert "[CPU] Support dynamic attention by tiling K1 when needed." (#23313) by Han-Chung Wang · 5 days ago
  30. d9aec69 [CI] Update golden time for gfx1201 (#23310) by Erick Ochoa Lopez · 5 days ago
  31. 761bd9c [CPU] Support dynamic attention by tiling K1 when needed. (#23304) by Han-Chung Wang · 5 days ago
  32. e3dfd29 [CI] disable failing cts tests (#23300) by Erick Ochoa Lopez · 5 days ago
  33. 81b508d [GlobalOptimization] Fix output_shape handling in SinkTransposeThroughExpandShape (#23308) by Quinn Dawkins · 5 days ago
  34. fbc4499 [Encoding] Add verifier for encoding_dims on (Un)SetEncodingOp (#23245) by Jorn Tuyls · 5 days ago
  35. 839085a Implement missing stablehlo.fft operations (#22829) by pstarkcdpr · 6 days ago
  36. 789859e [Codegen] Use safer hoisting in OptimizeTensorInsertExtractSlices (#23280) by Max191 · 6 days ago
  37. af093a8 [AMDGPU][LDS] Adding 1k, 2k, 4k, 8k static shape e2e tests for coalesced gather DMA op (#22884) by Alan Li · 6 days ago
  38. 3d3d912 [RISCV] Separate bare-metal and Linux build scripts. (#21800) by Han-Kuan Chen · 6 days ago
  39. 9e167dd [Codegen] Preserve DPS when vectorizing iree_vector_ext.to_layout (#23285) by Max191 · 6 days ago
  40. 00184c7 Integrate llvm/llvm-project@648cb36 (#23288) by Ian Wood · 6 days ago
  41. 7aa7be5 [Util]Fix memory corruption in LiftCFGToSCF when processing empty regions (#23131) by kimm240 · 6 days ago
  42. a413305 [e2e] Increase test timeout for gfx1250 (#23286) by Jakub Kuderski · 7 days ago
  43. bb04a48 Set CMAKE_CXX_EXTENSIONS to OFF to align with LLVM (#23284) by Bangtian Liu · 7 days ago
  44. 7edade7 [ROCm][gfx1250] Add e2e matmul tests for gfx1250 (#23282) by Jakub Kuderski · 7 days ago
  45. 5ee0652 [Codegen][IGEMM] Support Conv with no input channel dimension (#23271) by Vivian Zhang · 7 days ago
  46. 1ce2fa2 [LLVMCPU] Fix crash in limitVectorTileSizes with dynamic operand shapes. (#23281) by Han-Chung Wang · 7 days ago
  47. d4216bb New distribution tile heuristic for CPU data-tiled matmuls, take two. (#23272) by Benoit Jacob · 7 days ago
  48. de381bd [Bazel][ConstEval] Add missing tool dependency to LITs (#23241) by Artem Gindinson · 7 days ago
  49. 73bdcff [Encoding] Drop experimental i1 packing flag (#23186) by Lukas Sommer · 7 days ago
  50. ac97724 [Util] Verify tied operands for util.func (#23173) by Lukas Sommer · 7 days ago
  51. 1d89835 [NFC] Make status test macros take ownership of iree_status_t. (#23276) by Ben Vanik · 8 days ago
  52. 1a912be [CPU][NFC] Fix incorrect mmt4d dimension names in comments. (#23234) by Han-Chung Wang · 10 days ago
  53. 56acf7e Integrate LLVM@7a10fc8d542 (#23264) by Erick Ochoa Lopez · 10 days ago
  54. bc80992 [Encoding] Fix encoding dims propagation in SinkUnsetEncodingOp (#23265) by Jorn Tuyls · 10 days ago
  55. e9b69aa [Codegen] Add gpu.subgroup_size to dispatch bounds handling (#23233) by Krzysztof Drewniak · 10 days ago
  56. 31c3b34 [CI] Build clang-tidy from source and use in presubmit checks (#23258) by Jakub Kuderski · 10 days ago
  57. f816205 [Flow][NFC] Use region verifier for `flow.executable.export` (#23263) by Lukas Sommer · 10 days ago
  58. ccae7f0 [StreamToHal] Use rewriter to create block (#23249) by Lukas Sommer · 10 days ago
  59. 689176c [LLVMGPU] Promote C when DPS inits come from compute ops (#23254) by Max191 · 10 days ago
  60. 818f45f Integrate LLVM@5c35af8f1e6ebc7c32 (#23252) by Erick Ochoa Lopez · 11 days ago
  61. caa708f [GPU] Add divisibility comparision to buffer optimization (#23248) by Nirvedh Meshram · 11 days ago
  62. 8e0aa2b Revert "New distribution tile heuristic for CPU data-tiled matmuls." (#23255) by Erick Ochoa Lopez · 11 days ago
  63. 65a48b6 [LLVMGPU] Add pass to lower vector loads to amdgpu.transpose_load (#23081) by Max191 · 11 days ago
  64. 71595be Integrate LLVM@2e53764f2da742ba3 (#23250) by Erick Ochoa Lopez · 11 days ago
  65. f7ed024 New distribution tile heuristic for CPU data-tiled matmuls. (#23197) by Benoit Jacob · 11 days ago
  66. 63d1f64 [Codegen/Common] Skip generating padding scf.forall loops when padding is effectively a no-op (#23035) by Pooja Hemashekar · 11 days ago
  67. f09cea1 Integrate LLVM@78481a2444b1d4 (#23243) by Erick Ochoa Lopez · 11 days ago
  68. 8b62d22 [Codegen] Apply clang-tidy fixes to KernelConfig. NFC. (#23244) by Jakub Kuderski · 11 days ago
  69. 4cc4576 [CI] Add clang-tidy workflow (#23237) by Jakub Kuderski · 11 days ago
  70. 142aa59 Carry encoding in the preferred storage type of a hoistable type (#22221) by Jorn Tuyls · 11 days ago
  71. 5aa6453 Reapply "LLVM Integrate@6cc18a8e4338 (#23226)" (#23236) by Erick Ochoa Lopez · 12 days ago
  72. 5853971 Revert "LLVM Integrate@6cc18a8e4338" (#23235) by Erick Ochoa Lopez · 12 days ago
  73. ce1244f [ReductionVectorDistribute] Avoid adding lowering configs on failure (#23228) by Rahul Kayaith · 12 days ago
  74. 44a5bea LLVM Integrate@6cc18a8e4338 (#23226) by Erick Ochoa Lopez · 12 days ago
  75. 4e5d3da [CI][Torch] Enable split reduction and O3 for llama_8b_fp16 gfx942 config (#23231) by Bangtian Liu · 12 days ago
  76. 59d0a10 [Util] Support loop IVs in divisibility analysis (#22729) by Max191 · 12 days ago
  77. c815c2f [LinalgExt] Support and use arg_compare with explicit-index mode in split reduction (#23218) by Bangtian Liu · 12 days ago
  78. b7e2382 [Encoding] Propagate (Un)SetEncodingOp with dynamic encoding dims (#23125) by Jorn Tuyls · 12 days ago
  79. f1e63bd [DispatchCreation] Fold extract_slice of broadcast during split reduction tiling (#23012) by Bangtian Liu · 12 days ago
  80. e9b7f96 [Dispatch Creation] Add FoldReshapesIntoTensorBarriers to pass pipeline (#23222) by Ian Wood · 12 days ago
  81. 56d25a6 [DT] Remap linalg.index ops during encoding materialization. (#23159) by Han-Chung Wang · 12 days ago
  82. 44f2a68 [TensorExt] Improves `BitCastOfTensorCastStaticInfo` to handle constant dynamic dims (#23183) by Zhewen Yu · 12 days ago
  83. 6709ef9 [CI] Add MI355 e2e tests (#23090) by Jorn Tuyls · 12 days ago
  84. 9a021ae [runtime][bindings][python] Allow >= on sympy versions (#23223) by Quinn Dawkins · 13 days ago
  85. 3da7a63 [Flow] Always choose attention as the best op for dispatch annotation (#19696) by Kunwar Grover · 13 days ago
  86. b555852 [Codegen] Fix dynamic dim issue in getCopyTileSizes (#23121) by Ian Wood · 13 days ago
  87. 60b6fb9 [LinalgExt] MSVC Bug fix - useExp / useExp2 in AggregatedOpInterfaceImpl (#23219) by Keshav Vinayak Jha · 13 days ago
  88. bb00d01 Integrate LLVM@783fbdc54e (#23217) by Erick Ochoa Lopez · 13 days ago
  89. d54c845 [CPU] Enable E2E MX (scaled matmul) tests for CPU backends. (#23202) by Han-Chung Wang · 13 days ago
  90. 71fd3c7 [Codegen][GPU] Fix lane offset handling in coalesced DMA lowering (#23110) by Jorn Tuyls · 13 days ago
  91. 9753465 [LinalgExt] Added toggle for using useExp2 for onlineAttention Decomposition (#23211) by Keshav Vinayak Jha · 13 days ago
  92. 566455b [CI] Fix flag name for reverse iteration (#23215) by Jakub Kuderski · 13 days ago
  93. 3c3f9b8 Fixes AsyncUpdateOp elision for tied operations like copy. (#23208) by Ben Vanik · 13 days ago
  94. a0a9a50 [docs] Add policy for AI tool use (#23188) by Jakub Kuderski · 13 days ago
  95. 4197fe3 Integrate llvm-project@ad947503831a [ours a60d6603fbf8] (#23130) by Krzysztof Drewniak · 13 days ago
  96. c6ead2e Add initial clang-tidy configuration (#23203) by Jakub Kuderski · 13 days ago
  97. 744b303 Apply naming convention fixes. NFC. (#23209) by Jakub Kuderski · 13 days ago
  98. 7a600d7 Integrate torch-mlir@ac33bab4 (#23138) by Krzysztof Drewniak · 13 days ago
  99. dc11f63 [DispatchCreation] Include TensorExt ops in compute regions for barrier insertion (#23181) by Zhewen Yu · 13 days ago
  100. 7d91727 Adding overflow-safe allocation helpers to iree/base. (#23155) by Ben Vanik · 13 days ago