Commits

Branch selector

 main

User selector

All users

All time

Commit History

Commits on Mar 6, 2026

• Integrate llvm/llvm-project@a631c3f4077cf84c90f9dfa98a9bdbc1e568f055 (#23670)

  Show description for d8b6df3
  RattataKing

  authored
  d8b6df3

  Copy full SHA for d8b6df3

• [NFC][Codegen] Move distribution pattern infrastructure to VectorExt (#23669)

  Show description for 7de2750
  

  Groverkss

  and

  claude

  authored
  7de2750

  Copy full SHA for 7de2750

• [GlobalOpt] Support Img2Col Transformation for Conv2D Including Quantized Types (#23278)

  phoebesv

  authored
  80d70d6

  Copy full SHA for 80d70d6

Commits on Mar 5, 2026

• [Tokenizer] Add tiktoken format loader for OpenAI BPE vocabularies. (#23663)

  Show description for 205b17f
  

  benvanik

  and

  claude

  authored
  205b17f

  Copy full SHA for 205b17f

• [PCF] Add tensor.collapse_shape fusion into pcf.generic/loop (#23491)

  Show description for c114a89
  

  qedawkins

  and

  claude

  authored
  c114a89

  Copy full SHA for c114a89

• [CPU] Propagate the reduction tile sizes to producers because of fusion. (#23660)

  Show description for a02e85f
  hanhanW

  authored
  a02e85f

  Copy full SHA for a02e85f

• Integrate llvm/llvm-project@03773c3b06b2a9a023f8bab892f237fb3b980439 (#23651)

  Show description for 098465b
  RattataKing

  authored
  098465b

  Copy full SHA for 098465b

• [VectorDistribution] Relax layout size constraint (#23625)

  Show description for 8df2db9
  sommerlukas

  authored
  8df2db9

  Copy full SHA for 8df2db9

• Rewriting the HAL CTS to support bazel and scale better. (#23644)

  Show description for 04784a9
  

  benvanik

  and

  claude

  authored
  04784a9

  Copy full SHA for 04784a9

• [Async] Gracefully degrade io_uring slab registration on RLIMIT_MEMLOCK (#23654)

  Show description for 92bbb28
  

  benvanik

  and

  claude

  authored
  92bbb28

  Copy full SHA for 92bbb28

• [DispatchCreation] Fix fusion of scalar reduction with consumer (#23659)

  Show description for 927c775
  IanWood1

  authored
  927c775

  Copy full SHA for 927c775

• Add samples/hal/hello: pure HAL buffer fill, copy, and readback. (#23645)

  Show description for a385a68
  

  benvanik

  and

  claude

  authored
  a385a68

  Copy full SHA for a385a68

Commits on Mar 4, 2026

• [GPUHeuristics] Prefer larger MMA intrinsics for very large compute-bound GEMMs (#23641)

  Show description for e002a09
  lialan

  authored
  e002a09

  Copy full SHA for e002a09

• Bump stablehlo to d0484f6 (#23649)

  Show description for a657d73
  RattataKing

  authored
  a657d73

  Copy full SHA for a657d73

• [test] Enable e2e attention tests for arm64 targets. (#23635)

  Show description for 2df49e5
  hanhanW

  authored
  2df49e5

  Copy full SHA for 2df49e5

• [Codegen] Remove attention transpose intrinsic hacks (#23633)

  Show description for 1ea0b66
  Groverkss

  authored
  1ea0b66

  Copy full SHA for 1ea0b66

• Integrate llvm/llvm-project@386a3afa553fbf2854ee9d002e83bd456942ad39 (#23639)

  Show description for 46b54e8
  RattataKing

  authored
  46b54e8

  Copy full SHA for 46b54e8

• [docs] Update python versions listed on the website (#23647)

  Show description for ce92d23
  kuhar

  authored
  ce92d23

  Copy full SHA for ce92d23

• [GPU] Make GPUVectorAlloc allocate shared memory based on layout analysis (#23631)

  Show description for 2284319
  Groverkss

  authored
  2284319

  Copy full SHA for 2284319

• [Codegen] Fix layout analysis fixup crashes (#23630)

  Show description for c7072e9
  Groverkss

  authored
  c7072e9

  Copy full SHA for c7072e9

• [GPUHeuristics] Add min-based tile distribution for imbalanced M/N problems (#23619)

  Show description for f63db8d
  

  yzhang93

  and

  claude

  authored
  f63db8d

  Copy full SHA for f63db8d

• A couple of fixes picked up in the fusilli tests using sanitizers. (#23617)

  Show description for a1ef7f9
  AWoloszyn

  authored
  a1ef7f9

  Copy full SHA for a1ef7f9

• [Python] Stop building cp313 wheels on MacOS and Windows (#23640)

  Show description for 776210b
  kuhar

  authored
  776210b

  Copy full SHA for 776210b

• [Torch] Add flag to enable shape refinement (#23632)

  Show description for 2e4ef0f
  IanWood1

  authored
  2e4ef0f

  Copy full SHA for 2e4ef0f

Commits on Mar 3, 2026

• [CMake][Python] Auto-enable Stable ABI (abi3) for CPython 3.12+ with GIL (#23629)

  Show description for ae93b81
  

  kuhar

  and

  claude

  authored
  ae93b81

  Copy full SHA for ae93b81

• [Python][CI] Fix release validation installing wrong package versions (#23634)

  Show description for 39805c6
  

  kuhar

  and

  claude

  authored
  39805c6

  Copy full SHA for 39805c6

• [VMVX] Enable scf.forall distribution for VMVX pipelines. (#23615)

  Show description for 5fcbdbc
  hanhanW

  authored
  5fcbdbc

  Copy full SHA for 5fcbdbc

• Initial plumbing for inner_tiled on CPU with data-tiled MMA attribute. (#23494)

  Show description for 771eac5
  bjacob

  authored
  771eac5

  Copy full SHA for 771eac5

• [Python][CMake] Set py_limited_api on CMakeExtension for abi3 wheels (#23628)

  Show description for f2a2972
  

  kuhar

  and

  claude

  authored
  f2a2972

  Copy full SHA for f2a2972

• [CMake][Python] Bump minimum CMake version to 3.26 (#23607)

  Show description for 84f18ff
  

  kuhar

  and

  claude

  authored
  84f18ff

  Copy full SHA for 84f18ff

• [VectorDistribute] Refactor VectorLayoutAnalysis into 2-phase forward/backward design (#23611)

  Show description for 4e8dde4
  

  Groverkss

  and

  claude

  authored
  4e8dde4

  Copy full SHA for 4e8dde4

• [VectorDistribute] Consider all compute ops for thread tile size (#23394)

  Show description for 34f93d2
  sommerlukas

  authored
  34f93d2

  Copy full SHA for 34f93d2

Commits on Mar 2, 2026

• Integrate llvm/llvm-project@783eeb2ccfcac136f2191bda705ecd4a1a553718 (#23616)

  Show description for 9c4aa35
  RattataKing

  authored
  9c4aa35

  Copy full SHA for 9c4aa35

• [SPIRV] Enable small float support in SPIR-V pipeline. (#23391)

  Show description for e95e5d3
  hanhanW

  authored
  e95e5d3

  Copy full SHA for e95e5d3

• [CPU] Drop scf.for distribution option from CPU pipeline. (#23614)

  Show description for 06fa699
  hanhanW

  authored
  06fa699

  Copy full SHA for 06fa699