Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Oct 2, 2024

• [XLA:Python] Fix more bugs in the weakref_lru_cache implementation.

  Show description for 82269c5
  

  hawkinsp

  authored and

  tensorflower-gardener

  committed
  82269c5

  Copy full SHA for 82269c5

• Remove unneeded stream_executor/platform/port.h inclusion.

  Show description for 5e01e43
  

  klucke

  authored and

  tensorflower-gardener

  committed
  5e01e43

  Copy full SHA for 5e01e43

• [XLA:SPMD] Fix ReshapeSharding for dimensions of size 1 and >1 partitions.

  Show description for de753e9
  

  ZixuanJiang

  authored and

  tensorflower-gardener

  committed
  de753e9

  Copy full SHA for de753e9

• This CL correctly computes the tensor dim to mesh axis mapping for mixed mesh strategies when computing resharding costs involving such a strategy.

  Show description for 6a9a0a7
  tensorflower-gardener

  committed
  6a9a0a7

  Copy full SHA for 6a9a0a7

• [HLO Componentization] Create hlo/builder sub-component (Phase II).

  Show description for c08344f
  

  sdasgup3

  authored and

  tensorflower-gardener

  committed
  c08344f

  Copy full SHA for c08344f

• Correct some spell check mistakes

  Show description for 5c3f3c5
  tensorflower-gardener

  committed
  5c3f3c5

  Copy full SHA for 5c3f3c5

• Enable expand dims in the xnnpack delegate

  Show description for 3e84c99
  

  alankelly

  authored and

  tensorflower-gardener

  committed
  3e84c99

  Copy full SHA for 3e84c99

• Integrate LLVM at llvm/llvm-project@6292f117c39b

  Show description for 98e1fd6
  tensorflower-gardener

  committed
  98e1fd6

  Copy full SHA for 98e1fd6

• Use tsl DsoLoader functions instead of stream_executor aliases.

  Show description for c30560b
  

  klucke

  authored and

  tensorflower-gardener

  committed
  c30560b

  Copy full SHA for c30560b

• [StableHLO] Pin StableHLOv0.19.0 for older PJRT plugins.

  Show description for 67e9c59
  

  GleasonK

  authored and

  tensorflower-gardener

  committed
  67e9c59

  Copy full SHA for 67e9c59

• Delete xla_client.execute_with_python_values.

  Show description for 10aa1eb
  

  hawkinsp

  authored and

  tensorflower-gardener

  committed
  10aa1eb

  Copy full SHA for 10aa1eb

• [XLA:GPU] Introduce the `FusionBlockLevelRewriter` pass.

  Show description for 119dda6
  

  bchetioui

  authored and

  tensorflower-gardener

  committed
  119dda6

  Copy full SHA for 119dda6

• [XLA:GPU] Unify DimVar, RangeVar and RTVar.

  Show description for 03e8ff5
  

  pifon2a

  authored and

  tensorflower-gardener

  committed
  03e8ff5

  Copy full SHA for 03e8ff5

• [XLA:GPU] Clean up symbol visibility in Triton Fusion emitter

  Show description for bed0a51
  

  dimitar-asenov

  authored and

  tensorflower-gardener

  committed
  bed0a51

  Copy full SHA for bed0a51

• Adds num_features to SparsecoreLayoutProto.

  Show description for 7b44278
  tensorflower-gardener

  committed
  7b44278

  Copy full SHA for 7b44278

• Integrate LLVM at llvm/llvm-project@bf25ecb6caff

  Show description for 23a4c01
  tensorflower-gardener

  committed
  23a4c01

  Copy full SHA for 23a4c01

• [XLA:GPU][Emitters] Remove name from comparison operator for DimVar/RangeVar/RTVar.

  Show description for 1ffde05
  

  pifon2a

  authored and

  tensorflower-gardener

  committed
  1ffde05

  Copy full SHA for 1ffde05

• Integrate Triton up to [6fa4f504](https://github.com/openai/triton/commits/6fa4f504d61e48f7cea454ce0c1f6169907d5aaa)

  Show description for b259d92
  

  vwbaker

  authored and

  tensorflower-gardener

  committed
  b259d92

  Copy full SHA for b259d92

• [XLA:GPU] Generalize the Reduce-Precision in MHLO to also work on Tensors and use it in a Triton emitter.

  Show description for 50160d4
  

  dimitar-asenov

  authored and

  tensorflower-gardener

  committed
  50160d4

  Copy full SHA for 50160d4

• compat: Update forward compatibility horizon to 2024-10-02

  Show description for 07206ac
  tensorflower-gardener

  committed
  07206ac

  Copy full SHA for 07206ac

• Update GraphDef version to 2003.

  Show description for 316529b
  tensorflower-gardener

  committed
  316529b

  Copy full SHA for 316529b

• Automated Code Change

  Show description for b41bada
  tensorflower-gardener

  committed
  b41bada

  Copy full SHA for b41bada

• Automated Code Change

  Show description for 621ee72
  tensorflower-gardener

  committed
  621ee72

  Copy full SHA for 621ee72

• Automated Code Change

  Show description for 10fd5a6
  tensorflower-gardener

  committed
  10fd5a6

  Copy full SHA for 10fd5a6

• Automated Code Change

  Show description for b463ef1
  tensorflower-gardener

  committed
  b463ef1

  Copy full SHA for b463ef1

• Fix an IFRT Proxy version compatibility bug introduced by `ExecuteOptions::fill_status`

  Show description for b1c2ce6
  

  junwhanahn

  authored and

  tensorflower-gardener

  committed
  b1c2ce6

  Copy full SHA for b1c2ce6

• Add FP8 support to reverse_test

  Show description for 5f0a3c4
  tensorflower-gardener

  committed
  5f0a3c4

  Copy full SHA for 5f0a3c4

Commits on Oct 1, 2024

• PR #17828: Fix pylint C0301: Line too long in xla_client_test.py

  Show description for 617fbcc
  

  apivovarov

  authored and

  tensorflower-gardener

  committed
  617fbcc

  Copy full SHA for 617fbcc

• Use tsl DsoLoader function instead of stream_executor aliases.

  Show description for 60dcc40
  

  klucke

  authored and

  tensorflower-gardener

  committed
  60dcc40

  Copy full SHA for 60dcc40

• Stop using stream_executor aliases to tsl DsoLoader functions.

  Show description for a0a0407
  

  klucke

  authored and

  tensorflower-gardener

  committed
  a0a0407

  Copy full SHA for a0a0407

• Remove unneeded dso_loader.h include from ROCM files.

  Show description for b537321
  

  klucke

  authored and

  tensorflower-gardener

  committed
  b537321

  Copy full SHA for b537321

• [HLO Componentization] Create hlo/parser sub-component (Phase I).

  Show description for cef82f5
  

  sdasgup3

  authored and

  tensorflower-gardener

  committed
  cef82f5

  Copy full SHA for cef82f5

• Use TF_PREDICT_TRUE in gpu_device_functions.h rather than the SE aliases.

  Show description for 4e7b9d3
  

  klucke

  authored and

  tensorflower-gardener

  committed
  4e7b9d3

  Copy full SHA for 4e7b9d3

• Stop using unneeded stream_executor/platform/port.h header file & dependency.

  Show description for 1b8c33a
  

  klucke

  authored and

  tensorflower-gardener

  committed
  1b8c33a

  Copy full SHA for 1b8c33a

• Stop using SE_PREDICT_TRUE in favor of the ABSL variant.

  Show description for 21af557
  

  klucke

  authored and

  tensorflower-gardener

  committed
  21af557

  Copy full SHA for 21af557

Pagination

Previous

Next