Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Jul 30, 2025

• Add unit test with non-flat call graph on shardy xla pass.

  Show description for 4717b2c
  tensorflower-gardener

  committed
  4717b2c

  Copy full SHA for 4717b2c

• PR #29332: Fix the xla/service/gpu:gpu_aot_compilation_test

  Show description for 97f2232
  

  serach24

  authored and

  tensorflower-gardener

  committed
  97f2232

  Copy full SHA for 97f2232

• PR #29342: Fix the oneAPI Presubmit failure

  Show description for 59b3fb5
  

  mraunak

  authored and

  tensorflower-gardener

  committed
  59b3fb5

  Copy full SHA for 59b3fb5

• Add a missing symbol to unbreak Windows.

  Show description for 0adf105
  

  belitskiy

  authored and

  tensorflower-gardener

  committed
  0adf105

  Copy full SHA for 0adf105

• #sdy if we need to use tuples, tuplify entry computation layout even if no layout has been set.

  Show description for 8d778f7
  

  bartchr808

  authored and

  tensorflower-gardener

  committed
  8d778f7

  Copy full SHA for 8d778f7

• [XLA:CPU] Add logging for LLVM compilation time.

  Show description for 03f861c
  

  WillFroom

  authored and

  tensorflower-gardener

  committed
  03f861c

  Copy full SHA for 03f861c

• Remove PJRT v70 checks since the 3 week forward compatibility window has passed.

  Show description for df656de
  

  bartchr808

  authored and

  tensorflower-gardener

  committed
  df656de

  Copy full SHA for df656de

• [XLA:CPU] Run inline pass multiple times.

  Show description for 1688645
  

  WillFroom

  authored and

  tensorflower-gardener

  committed
  1688645

  Copy full SHA for 1688645

• [XLA:GPU] Do not call ComposePermutations for non-permutations (NFC).

  Show description for f9d95bb
  

  akuegel

  authored and

  tensorflower-gardener

  committed
  f9d95bb

  Copy full SHA for f9d95bb

• Add a variadic scatter test with repeated operand.

  Show description for d5e4be7
  

  akuegel

  authored and

  tensorflower-gardener

  committed
  d5e4be7

  Copy full SHA for d5e4be7

• [XLA:GPU/TMA] Restrict TMA for use-cases where the offsets don't meet the constraints. The original issue manifested in the form of CUDA_ERROR_ILLEGAL_INSTRUCTION on microbenchmarks that used TMA.

  Show description for 8605de8
  

  Moerafaat

  authored and

  tensorflower-gardener

  committed
  8605de8

  Copy full SHA for 8605de8

• [XLA:GPU] Enable Triton TMA tuning in new autotuner. Added support in both backends that support Triton. Also refactored a common tma utility and updated the usages.

  Show description for 0622b2b
  

  Moerafaat

  authored and

  tensorflower-gardener

  committed
  0622b2b

  Copy full SHA for 0622b2b

• compat: Update forward compatibility horizon to 2025-07-30

  Show description for 171f812
  tensorflower-gardener

  committed
  171f812

  Copy full SHA for 171f812

• Update GraphDef version to 2304.

  Show description for 74dd9be
  tensorflower-gardener

  committed
  74dd9be

  Copy full SHA for 74dd9be

• [XLA] Support DimensionVector return type in math_util, permutation_util, util.

  Show description for f727033
  tensorflower-gardener

  committed
  f727033

  Copy full SHA for f727033

• #sdy set `PropagationOptions::enableAutoPartitioning` in shardy_xla_pass and cleanup legacy auto partitioner code, as it's no longer needed.

  Show description for 02473fd
  

  tomnatan30

  authored and

  tensorflower-gardener

  committed
  02473fd

  Copy full SHA for 02473fd

• [XLA:GPU] Code style improvements for matmul_indexing_utils

  Show description for 4a25beb
  

  loislo

  authored and

  tensorflower-gardener

  committed
  4a25beb

  Copy full SHA for 4a25beb

• [XLA:GPU] Use MapVector instead of DenseMap for storing constraints.

  Show description for 9453299
  

  akuegel

  authored and

  tensorflower-gardener

  committed
  9453299

  Copy full SHA for 9453299

• Change return value of IFRT `MakeDeviceList` from `DeviceListRef` to `absl::StatusOr<DeviceListRef>` so that upcoming CLs can do validation that the devices match requirements.

  Show description for 3dd0750
  tensorflower-gardener

  committed
  3dd0750

  Copy full SHA for 3dd0750

Commits on Jul 29, 2025

• Allow CPU embedding lookup inputs to be a subset of embedding feature configs.

  Show description for b4353a4
  

  WAangzE

  authored and

  tensorflower-gardener

  committed
  b4353a4

  Copy full SHA for b4353a4

• PR #29390: pjrt/ffi: add FFI handler registration API to the FFI PjRt extension

  Show description for 8c48d7f
  

  hugomano

  authored and

  tensorflower-gardener

  committed
  8c48d7f

  Copy full SHA for 8c48d7f

• Integrate LLVM at llvm/llvm-project@22b083539051

  Show description for b5e11e3
  tensorflower-gardener

  committed
  b5e11e3

  Copy full SHA for b5e11e3

• [XLA:GPU] Use XLA_VLOG_LINES instead of VLogModule.

  Show description for 10ae91f
  

  olegshyshkov

  authored and

  tensorflower-gardener

  committed
  10ae91f

  Copy full SHA for 10ae91f

• Fix crash on reentrant call to LRUCache::Clear().

  Show description for 6233b53
  

  hawkinsp

  authored and

  tensorflower-gardener

  committed
  6233b53

  Copy full SHA for 6233b53

• Disable de-Morgan's law clang-tidy check on macro. Users of the macro aren't aware, don't control, and shouldn't care that the macro expands to "!(x)", and it's reasonable for users to pass in "a |…

  Show description for e337221
  

  tkoeppe

  authored and

  tensorflower-gardener

  committed
  e337221

  Copy full SHA for e337221

• [XLA:GPU] move matmul_indexing_utils one dir up

  Show description for a49de24
  

  loislo

  authored and

  tensorflower-gardener

  committed
  a49de24

  Copy full SHA for a49de24

• [XLA:GPU] Refactor `KernelArguments::Create`.

  Show description for fcd37bf
  

  olegshyshkov

  authored and

  tensorflower-gardener

  committed
  fcd37bf

  Copy full SHA for fcd37bf

• [XLA:GPU] Remove MaterializeOp (NFC).

  Show description for bf3ffb1
  

  akuegel

  authored and

  tensorflower-gardener

  committed
  bf3ffb1

  Copy full SHA for bf3ffb1

• Update the `coords` of `PjRtStreamExecutorDeviceDescription`.

  Show description for ea47b73
  tensorflower-gardener

  committed
  ea47b73

  Copy full SHA for ea47b73

• [xla:cpu] Register runtime symbols as weak

  Show description for 20d58c3
  

  ezhulenev

  authored and

  tensorflower-gardener

  committed
  20d58c3

  Copy full SHA for 20d58c3

• [xla:cpu] Fix use-after-free on error handling path

  Show description for f208737
  

  ezhulenev

  authored and

  tensorflower-gardener

  committed
  f208737

  Copy full SHA for f208737

• [XLA:GPU] Remove redundant `needed_operands` argument from KernelArguments::Create.

  Show description for 0d18071
  

  olegshyshkov

  authored and

  tensorflower-gardener

  committed
  0d18071

  Copy full SHA for 0d18071

• PR #29386: Bump github/codeql-action from 3.29.3 to 3.29.4

  Show description for c21aab6
  

  dependabot[bot]

  authored and

  tensorflower-gardener

  committed
  c21aab6

  Copy full SHA for c21aab6

• [XLA:GPU] Remove FuseLoops pass

  Show description for 4c835f7
  

  akuegel

  authored and

  tensorflower-gardener

  committed
  4c835f7

  Copy full SHA for 4c835f7

• [XLA:GPU] Pass operands to BuildKernelThunkForNonFusionOp for RNG and sort kernels.

  Show description for f2d6c4c
  

  olegshyshkov

  authored and

  tensorflower-gardener

  committed
  f2d6c4c

  Copy full SHA for f2d6c4c

Pagination

Previous

Next