Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Oct 30, 2025

• [xla:cpu] Do not register legacy runtime symbols with XLA:CPU custom calls

  Show description for 9acb8a4
  

  ezhulenev

  authored and

  Google-ML-Automation

  committed
  9acb8a4

  Copy full SHA for 9acb8a4

• Migrate multioutput_fusion_test to use PjRt.

  Show description for 94775d5
  

  nvgrw

  authored and

  Google-ML-Automation

  committed
  94775d5

  Copy full SHA for 94775d5

• Add additional way to poision a connection (to allow testing different

  Show description for c939d64
  

  pschuh

  authored and

  Google-ML-Automation

  committed
  c939d64

  Copy full SHA for c939d64

• Add the option to dump before/after autotuned instructions in AutotunerConfig.

  Show description for 8a9fc71
  Google-ML-Automation

  committed
  8a9fc71

  Copy full SHA for 8a9fc71

• Refactor: Move common SymbolicMapTest setup to the fixture.

  Show description for f3fa442
  Google-ML-Automation

  committed
  f3fa442

  Copy full SHA for f3fa442

• Only enable YNNPACK for bf16 and int8 for now.

  Show description for 80ae82a
  Google-ML-Automation

  committed
  80ae82a

  Copy full SHA for 80ae82a

• [XLA][codegen] Migrate triton operations that have shared dialect lowerings are implemented for.

  Show description for a211f37
  

  basioli-k

  authored and

  Google-ML-Automation

  committed
  a211f37

  Copy full SHA for a211f37

• Migrate slice_test to use PjRt.

  Show description for cc56f45
  

  nvgrw

  authored and

  Google-ML-Automation

  committed
  cc56f45

  Copy full SHA for cc56f45

• [XLA][codegen] Emit shlo reshape from the fusion emitter and lower it to triton for the triton backend.

  Show description for 1daeea8
  

  basioli-k

  authored and

  Google-ML-Automation

  committed
  1daeea8

  Copy full SHA for 1daeea8

• Allow empty dimension list in SymbolicMap::ReplaceDimsAndSymbols

  Show description for 6d95931
  Google-ML-Automation

  committed
  6d95931

  Copy full SHA for 6d95931

• Migrate params_test to use PjRt.

  Show description for 33b93bb
  

  nvgrw

  authored and

  Google-ML-Automation

  committed
  33b93bb

  Copy full SHA for 33b93bb

• Clear frontend attributes for get-tuple-elements of GlobalToLocal and LocalToGlobal custom-calls.

  Show description for 52ff404
  

  ZixuanJiang

  authored and

  Google-ML-Automation

  committed
  52ff404

  Copy full SHA for 52ff404

• Improve CUDNN error messages

  Show description for b13ea16
  

  wsmoses

  authored and

  Google-ML-Automation

  committed
  b13ea16

  Copy full SHA for b13ea16

• [XLA:GPU] Use a single intra-host ragged-all-to-all in the decomposition.

  Show description for cc7b445
  

  olegshyshkov

  authored and

  Google-ML-Automation

  committed
  cc7b445

  Copy full SHA for cc7b445

• [xla:ffi] Add example of async custom call in XLA:GPU

  Show description for 1de868e
  

  ezhulenev

  authored and

  Google-ML-Automation

  committed
  1de868e

  Copy full SHA for 1de868e

• Migrate map_test to use PjRt.

  Show description for 6b79dad
  

  nvgrw

  authored and

  Google-ML-Automation

  committed
  6b79dad

  Copy full SHA for 6b79dad

• [Autotuner] Add sharding support using KeyValueStore Interface.

  Show description for df61ba7
  Google-ML-Automation

  committed
  df61ba7

  Copy full SHA for df61ba7

• [XLA][codegen] Emit stablehlo reduce op from the fusion emitter and lower it to triton for the triton backend.

  Show description for cb6d6c7
  

  basioli-k

  authored and

  Google-ML-Automation

  committed
  cb6d6c7

  Copy full SHA for cb6d6c7

• Migrate reshape_test to use PjRt.

  Show description for 352bf05
  

  nvgrw

  authored and

  Google-ML-Automation

  committed
  352bf05

  Copy full SHA for 352bf05

• [XTile] Modify Stable HLO check on iota to restrict it to the 1D case.

  Show description for 958ba77
  

  WillFroom

  authored and

  Google-ML-Automation

  committed
  958ba77

  Copy full SHA for 958ba77

• Integrate LLVM at llvm/llvm-project@4c46ae394841

  Show description for 124d7c0
  Google-ML-Automation

  committed
  124d7c0

  Copy full SHA for 124d7c0

• Increase the maximum HLO op chain length for profiling from 8192 to 16384.

  Show description for 92766a9
  

  chsigg

  authored and

  Google-ML-Automation

  committed
  92766a9

  Copy full SHA for 92766a9

• Update Bazel version to 7.7.0.

  Show description for ec30f5d
  

  meteorcloudy

  authored and

  Google-ML-Automation

  committed
  ec30f5d

  Copy full SHA for ec30f5d

• Reuse tuple elements field from existing `HloSharding`

  Show description for 5985bae
  

  KanishAnand

  authored and

  Google-ML-Automation

  committed
  5985bae

  Copy full SHA for 5985bae

• PR #33278: Bump keras from 3.11.3 to 3.12.0 in /xla/backends/cpu/benchmarks/e2e/gemma2/keras

  Show description for df68859
  

  dependabot[bot]

  authored and

  Google-ML-Automation

  committed
  df68859

  Copy full SHA for df68859

• Add helper functions for creating and inspecting symbolic dimensions and symbols

  Show description for 05e62cf
  Google-ML-Automation

  committed
  05e62cf

  Copy full SHA for 05e62cf

• [XLA:GPU] CustomCallThunk: enable use of lambdas with captures

  Show description for 7761ba0
  

  Marcin Radomski

  authored and

  Google-ML-Automation

  committed
  7761ba0

  Copy full SHA for 7761ba0

• [XLA:CPU] Compare host cpu features when loading AOT result to the compilation machine features

  Show description for aca11aa
  

  basioli-k

  authored and

  Google-ML-Automation

  committed
  aca11aa

  Copy full SHA for aca11aa

• Add hashing support for SymbolicMap

  Show description for 6d47a09
  Google-ML-Automation

  committed
  6d47a09

  Copy full SHA for 6d47a09

• [XLA:Collective] Remove unnecessary const for a function argument

  Show description for 0679383
  Google-ML-Automation

  committed
  0679383

  Copy full SHA for 0679383

• Add serialization for ffi::Attribute

  Show description for dd65fcc
  

  beckerhe

  authored and

  Google-ML-Automation

  committed
  dd65fcc

  Copy full SHA for dd65fcc

• [XLA:GPU] Add multi-host ragged-all-to-all decomposer pattern for combine ra2a.

  Show description for a375c2b
  

  olegshyshkov

  authored and

  Google-ML-Automation

  committed
  a375c2b

  Copy full SHA for a375c2b

• [XLA][codegen] Emit ReshapeToScalar as a tensor.extract op, and lower to triton specific impl.

  Show description for 09d112d
  

  basioli-k

  authored and

  Google-ML-Automation

  committed
  09d112d

  Copy full SHA for 09d112d

• Add helper functions for creating constant SymbolicExpr

  Show description for 4b388ab
  Google-ML-Automation

  committed
  4b388ab

  Copy full SHA for 4b388ab

• [XLA:GPU] Introduce `CollectiveOpsE2ETestBase` for common functionality.

  Show description for 91cc4f9
  

  olegshyshkov

  authored and

  Google-ML-Automation

  committed
  91cc4f9

  Copy full SHA for 91cc4f9

Pagination

Previous

Next