Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit History

Commits on May 20, 2026

• [VM] Add VMASM roundtrip tooling (#24508)

  Show description for 2098bbe
  benvanik

  authored
  2098bbe

  Copy full SHA for 2098bbe

Commits on May 19, 2026

• [Codegen][CPU] Add x86 AVX-512 VNNI vpdpbusd MMA intrinsic. (#24494)

  Show description for e8ac558
  

  bjacob

  and

  claude

  authored
  e8ac558

  Copy full SHA for e8ac558

Commits on May 15, 2026

• [Vulkan] HAL Rewrite (#24459)

  Show description for b46def4
  benvanik

  authored
  b46def4

  Copy full SHA for b46def4

• [Codegen][CPU] Add e2e matmul tests for the inner_tiled lowering path. (#24492)

  Show description for cff4b04
  bjacob

  authored
  cff4b04

  Copy full SHA for cff4b04

• [LinalgExt] Initialize top-k output values in test

  benvanik

  committed
  3546cbf

  Copy full SHA for 3546cbf

• [PkgCI] Refresh gfx942 torch AB goldens

  benvanik

  committed
  caa5ee1

  Copy full SHA for caa5ee1

• [Vulkan] Use current BDA compiler flag in e2e tests

  benvanik

  committed
  c510104

  Copy full SHA for c510104

• [Vulkan] Harden BDA executable serialization

  Show description for 4385f43
  benvanik

  committed
  4385f43

  Copy full SHA for 4385f43

• [DMA] Support DMA in convolution lowering (#23907)

  Show description for 901adad
  lialan

  authored
  901adad

  Copy full SHA for 901adad

• [VectorDistribute] Distribute `vector.scan` across subgroups (#24475)

  Show description for 74278ee
  sommerlukas

  authored
  74278ee

  Copy full SHA for 74278ee

Commits on May 14, 2026

• [Codegen][GPU] Relax 32-bit reduction/arg_compare ceiling for VectorDistribute (#24432)

  Show description for 65df46f
  bangtianliu

  authored
  65df46f

  Copy full SHA for 65df46f

Commits on May 13, 2026

• [Bazel] Add missing loads and glob support for bazel 8+ (#24434)

  Show description for 9d211b4
  cerisier

  authored
  9d211b4

  Copy full SHA for 9d211b4

Commits on May 12, 2026

• [CI] Update vulkan jobs to include arch in names and drop matmul benchmarking. (#24420)

  Show description for 00593ea
  hanhanW

  authored
  00593ea

  Copy full SHA for 00593ea

Commits on May 11, 2026

• [Codegen][GPU] Avoid generic TileAndFuse promotion for convolutions. (#24435)

  Show description for b8aec7a
  hanhanW

  authored
  b8aec7a

  Copy full SHA for b8aec7a

• [GPU][Codegen] Choose VDMFMA intrinsics for 16bit input skinny gemms (#24103)

  Show description for 67daf81
  efric

  authored
  67daf81

  Copy full SHA for 67daf81

• [Codegen] Fix RAW in_bounds false for R&W (#24418)

  Show description for cd8e78a
  amd-eochoalo

  authored
  cd8e78a

  Copy full SHA for cd8e78a

Commits on May 9, 2026

• Revert "[Codegen][CPU] Add e2e matmul tests for the inner_tiled lowering path." (#24421)

  Show description for 919e956
  bjacob

  authored
  919e956

  Copy full SHA for 919e956

Commits on May 8, 2026

• [tools] Account both atol and rtol in floating point comparison. (#24394)

  Show description for f6e1872
  hanhanW

  authored
  f6e1872

  Copy full SHA for f6e1872

• [Codegen][CPU] Add e2e matmul tests for the inner_tiled lowering path. (#24405)

  Show description for fbe60d8
  bjacob

  authored
  fbe60d8

  Copy full SHA for fbe60d8

• [INTEGRATION] Bump llvm to 180409dc6757 (#24409)

  Show description for 2bf97ad
  lialan

  authored
  2bf97ad

  Copy full SHA for 2bf97ad

• Revert "[LinalgExt] Fix attention NaN for fully-masked rows" (#24413)

  Show description for 83b0606
  hanhanW

  authored
  83b0606

  Copy full SHA for 83b0606

• [Codegen] Move remaining pipelines to `iree_codegen` attrs (#24398)

  Show description for c0f5d4b
  kuhar

  authored
  c0f5d4b

  Copy full SHA for c0f5d4b

Commits on May 6, 2026

• [INTEGRATION] Bump llvm to 3ed76d05a78d (#24376)

  Show description for 4030534
  lialan

  authored
  4030534

  Copy full SHA for 4030534

• [LinalgExt] Fix attention NaN for fully-masked rows (#24178)

  Show description for 5fbfe29
  keshavvinayak01

  authored
  5fbfe29

  Copy full SHA for 5fbfe29

• [InputConversion] Lower AtenArgmax/AtenArgmin to iree_linalg_ext.arg_compare (#24291)

  Show description for 9092be0
  bangtianliu

  authored
  9092be0

  Copy full SHA for 9092be0

• [test][nfc] Add regression tests about strided vector.gather back. (#24370)

  Show description for a1b8b72
  hanhanW

  authored
  a1b8b72

  Copy full SHA for a1b8b72

Commits on May 5, 2026

• [DispatchCreation][LinalgExt] Add OnlineAttentionOp support in dispatch formation and reshape fusion (#24068)

  Show description for 4f6cd98
  keshavvinayak01

  authored
  4f6cd98

  Copy full SHA for 4f6cd98

Commits on May 2, 2026

• [DispatchCreation] Tighten scatter-skip predicate in CollapseDimensions (#24334)

  Show description for 0935008
  

  yzhang93

  and

  claude

  authored
  0935008

  Copy full SHA for 0935008

Commits on May 1, 2026

• [Codegen][DMA] Fix unaligned swizzle offset computation in gather-to-lds lowering (#24241)

  Show description for f4d1908
  Yu-Zhewen

  authored
  f4d1908

  Copy full SHA for f4d1908

Commits on Apr 30, 2026

• [CPU] Add ContiguousMemrefGather1DToConditionalLoads vector lowering. (#24327)

  Show description for 967b794
  hanhanW

  authored
  967b794

  Copy full SHA for 967b794

• [CI] Update iree-test-suite ref (#24304)

  Show description for 9f7a14e
  amd-eochoalo

  authored
  9f7a14e

  Copy full SHA for 9f7a14e

Commits on Apr 29, 2026

• [Test] Update ONNX Vulkan O0 xfails

  benvanik

  committed
  7b5817b

  Copy full SHA for 7b5817b

Commits on Apr 24, 2026

• [CI][TorchModels] Relax mi325 golden for flaky llama_8b_fp16 decode_seq128 benchmark (#24249)

  Show description for 89e348c
  hanhanW

  authored
  89e348c

  Copy full SHA for 89e348c

Commits on Apr 23, 2026

• [CPU] Drop benchmarking requirement from cpu torch op tests. (#24194)

  Show description for b2df6f5
  hanhanW

  authored
  b2df6f5

  Copy full SHA for b2df6f5

• [CI][TorchModels] Relax mi325 goldens for two flaky llama_8b_fp16 benchmarks (#24234)

  Show description for ce41a40
  hanhanW

  authored
  ce41a40

  Copy full SHA for ce41a40