Commits

Branch selector

 main

User selector

anlunx

Datepicker

All time

Commit History

Commits on Mar 10, 2023

• [xla:gpu] Initialize BlasSupport before capturing CUDA graphs

  Show description for 20575d4
  

  anlunx

  authored and

  copybara-github

  committed
  20575d4

  Copy full SHA for 20575d4

Commits on Mar 8, 2023

• [xla:gpu] Serialize runtime convolution autotuning

  Show description for a36b4de
  

  anlunx

  authored and

  copybara-github

  committed
  a36b4de

  Copy full SHA for a36b4de

Commits on Mar 6, 2023

• [xla:gpu] NFC: Remove #if GOOGLE_CUDA in convolution custom call implementation

  Show description for aaebc9c
  

  anlunx

  authored and

  copybara-github

  committed
  aaebc9c

  Copy full SHA for aaebc9c

Commits on Mar 5, 2023

• [xla:gpu] Use BFC allocator in runtime autotuning

  Show description for 9772a71
  

  anlunx

  authored and

  copybara-github

  committed
  9772a71

  Copy full SHA for 9772a71

Commits on Mar 2, 2023

• [xla:gpu] Trigger runtime autotuning on AOT autotuning cache miss

  Show description for 64afd11
  

  anlunx

  authored and

  copybara-github

  committed
  64afd11

  Copy full SHA for 64afd11

• [xla:gpu] Add support for runtime convolution autotuning

  Show description for 91d7e82
  

  anlunx

  authored and

  copybara-github

  committed
  91d7e82

  Copy full SHA for 91d7e82

Commits on Feb 24, 2023

• [xla:gpu] Pass NonAtomicallyUpgradeableRWLock to graph launch custom call

  Show description for f1e3b53
  

  anlunx

  authored and

  copybara-github

  committed
  f1e3b53

  Copy full SHA for f1e3b53

Commits on Feb 23, 2023

• [xla:gpu] GEMM with runtime autotuning should not be captured by a CUDA graph

  Show description for cf2d946
  

  anlunx

  authored and

  copybara-github

  committed
  cf2d946

  Copy full SHA for cf2d946

Commits on Feb 22, 2023

• [xla:gpu] Add support for capturing lmhlo_gpu.gemm operations into cuda graphs

  Show description for 24411c0
  

  anlunx

  authored and

  copybara-github

  committed
  24411c0

  Copy full SHA for 24411c0

Commits on Feb 21, 2023

• [xla:gpu] Serialize autotuning on GPU devices

  Show description for af7f28c
  

  anlunx

  authored and

  copybara-github

  committed
  af7f28c

  Copy full SHA for af7f28c

Commits on Feb 17, 2023

• [xla:gpu] Create a non-atomically upgradeable reader mutex lock that acquires and releases the lock via RAII

  Show description for c798026
  

  anlunx

  authored and

  copybara-github

  committed
  c798026

  Copy full SHA for c798026

• [XLA:GPU] Fix dependency error in GpuCompiler

  Show description for 62ba400
  

  anlunx

  authored and

  copybara-github

  committed
  62ba400

  Copy full SHA for 62ba400

Commits on Feb 15, 2023

• [xla:gpu]NFC: Unify AutotuneCacheKey in gpu autotuners

  Show description for 712e0f9
  

  anlunx

  authored and

  copybara-github

  committed
  712e0f9

  Copy full SHA for 712e0f9

Commits on Feb 10, 2023

• Rollback of [XLA:GPU][NFC] Make PickBestAlgorithmNoCacheCuda suitable for runtime autotuning

  Show description for 0b49787
  

  anlunx

  authored and

  copybara-github

  committed
  0b49787

  Copy full SHA for 0b49787

Commits on Feb 8, 2023

• [XLA:GPU][NFC] Make PickBestAlgorithmNoCacheCuda suitable for runtime autotuning

  Show description for 2998711
  

  anlunx

  authored and

  copybara-github

  committed
  2998711

  Copy full SHA for 2998711

Commits on Feb 6, 2023

• [xla:gpu] Create abstract GpuSerializableAutotuner class for storing shared data/functions between GemmAlgorithmPicker and GpuConvAlgorithmPicker

  Show description for 28bf67b
  

  anlunx

  authored and

  copybara-github

  committed
  28bf67b

  Copy full SHA for 28bf67b

Commits on Feb 1, 2023

• [xla:gpu] Remove check on XLA_FLAGS when doing deserialization

  Show description for c50c90a
  

  anlunx

  authored and

  copybara-github

  committed
  c50c90a

  Copy full SHA for c50c90a

Commits on Jan 31, 2023

• [xla:gpu] Fix a bug that string constructor can receive nullptr as input

  Show description for 5b3ba47
  

  anlunx

  authored and

  copybara-github

  committed
  5b3ba47

  Copy full SHA for 5b3ba47

• [xla:gpu] Use reader lock in GpuExecutable::ExecuteAsyncOnStreamImpl

  Show description for 30c35e9
  

  anlunx

  authored and

  copybara-github

  committed
  30c35e9

  Copy full SHA for 30c35e9

Commits on Jan 27, 2023

• [XLA:GPU] Reuse GetBestAlgorithm in runtime autotuning

  Show description for 92a0ba5
  

  anlunx

  authored and

  copybara-github

  committed
  92a0ba5

  Copy full SHA for 92a0ba5

Commits on Jan 26, 2023

• [XLA:GPU] Reuse GetBestAlgorithm in runtime autotuning

  Show description for 0b3cca9
  

  anlunx

  authored and

  copybara-github

  committed
  0b3cca9

  Copy full SHA for 0b3cca9

Commits on Jan 24, 2023

• [XLA:GPU] Abstract away CudaComputeCapability in GpuCompiler

  Show description for 3c53e7e
  

  anlunx

  authored and

  copybara-github

  committed
  3c53e7e

  Copy full SHA for 3c53e7e

Commits on Jan 21, 2023

• [XLA:GPU] GPU genrule should be able to run without autotune results

  Show description for bc33aae
  

  anlunx

  authored and

  copybara-github

  committed
  bc33aae

  Copy full SHA for bc33aae

Commits on Jan 20, 2023

• [XLA:GPU] Implement GEMM runtime autotuning

  Show description for 09558ee
  

  anlunx

  authored and

  copybara-github

  committed
  09558ee

  Copy full SHA for 09558ee

Commits on Jan 17, 2023

• [XLA:GPU] Separate autotune results from GpuTargetConfig

  Show description for e0f8d74
  

  anlunx

  authored and

  copybara-github

  committed
  e0f8d74

  Copy full SHA for e0f8d74

Commits on Jan 14, 2023

• [XLA:GPU] Enable AOT autotuning for Convolutions

  Show description for c150d5f
  

  anlunx

  authored and

  copybara-github

  committed
  c150d5f

  Copy full SHA for c150d5f

Commits on Jan 13, 2023

• [XLA:GPU] Enable AOT autotuning for GEMMs

  Show description for bbc753f
  

  anlunx

  authored and

  copybara-github

  committed
  bbc753f

  Copy full SHA for bbc753f

Commits on Jan 12, 2023

• [XLA:GPU] Add a deviceless API for GpuConvAlgorithmPicker

  Show description for 0a53568
  

  anlunx

  authored and

  copybara-github

  committed
  0a53568

  Copy full SHA for 0a53568

Commits on Jan 9, 2023

• [XLA:GPU] NFC:Improve the documentation of GemmAlgorithmPicker APIs

  Show description for 07ba2fe
  

  anlunx

  authored and

  copybara-github

  committed
  07ba2fe

  Copy full SHA for 07ba2fe

Commits on Dec 31, 2022

• [XLA:GPU] Support Gemm autotuning without device

  Show description for dc4ac84
  

  anlunx

  authored and

  copybara-github

  committed
  dc4ac84

  Copy full SHA for dc4ac84

Commits on Dec 21, 2022

• [XLA:GPU] Delay the use of se::AsDnn to SupportsBF16 call site

  Show description for 41726e4
  

  anlunx

  authored and

  copybara-github

  committed
  41726e4

  Copy full SHA for 41726e4

Commits on Dec 19, 2022

• [XLA:GPU] Support convolutions in GPU AOT compilation

  Show description for 6c27865
  

  anlunx

  authored and

  copybara-github

  committed
  6c27865

  Copy full SHA for 6c27865

Commits on Dec 18, 2022

• [XLA:GPU] Update AMDGPUCompiler to match the base class GpuCompiler

  Show description for f2ecf8d
  

  anlunx

  authored and

  copybara-github

  committed
  f2ecf8d

  Copy full SHA for f2ecf8d

Commits on Dec 17, 2022

• [XLA:GPU] Enable HLO passes in xla_compile

  Show description for 959be3a
  

  anlunx

  authored and

  copybara-github

  committed
  959be3a

  Copy full SHA for 959be3a

Commits on Dec 15, 2022

• [XLA:GPU] Add Dnn VersionInfo to GpuTargetConfig

  Show description for dd107c8
  

  anlunx

  authored and

  copybara-github

  committed
  dd107c8

  Copy full SHA for dd107c8

Pagination

Previous

Next