Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Mar 21, 2023

• Remove C++ jit support since it has been replaced with Pjit. Keep CompiledFunction alive as a shim which cannot be instantiated but will work for isinstance checks.

  Show description for f5e12ac
  

  yashk2810

  authored and

  copybara-github

  committed
  f5e12ac

  Copy full SHA for f5e12ac

• Rollback "Use erf(x) = 1-erfc(x) for |x| > 1"

  Show description for 898c28d
  

  tensorflower-gardener

  authored and

  copybara-github

  committed
  898c28d

  Copy full SHA for 898c28d

• [NFC] Cleanup unused headers.

  Show description for 5a962e6
  

  jurahul

  authored and

  copybara-github

  committed
  5a962e6

  Copy full SHA for 5a962e6

• Delete the C++ GetEnableJaxArray() flag.

  Show description for 337d403
  

  pschuh

  authored and

  copybara-github

  committed
  337d403

  Copy full SHA for 337d403

• Compare PR benchmark performance to performance at base.

  Show description for 7cd336b
  

  tensorflower-gardener

  authored and

  copybara-github

  committed
  7cd336b

  Copy full SHA for 7cd336b

• [XLA:Python] Delete C++ ShardedDeviceArray class.

  Show description for c08deae
  

  hawkinsp

  authored and

  copybara-github

  committed
  c08deae

  Copy full SHA for c08deae

Commits on Mar 20, 2023

• [XLA/GPU] Change invocation of CollectivesScheduleLinearizer in SPMD

  Show description for 7949e18
  

  jurahul

  authored and

  copybara-github

  committed
  7949e18

  Copy full SHA for 7949e18

• [xla:runtime] Add support of passing async values to runtime executable

  Show description for 9f46b1c
  

  tensorflower-gardener

  authored and

  copybara-github

  committed
  9f46b1c

  Copy full SHA for 9f46b1c

• Integrate StableHLO at openxla/stablehlo@a5d8529

  Show description for 7119b20
  

  sdasgup3

  authored and

  copybara-github

  committed
  7119b20

  Copy full SHA for 7119b20

• Integrate LLVM at llvm/llvm-project@5b1145bc46e2

  Show description for 354e00c
  

  durin42

  authored and

  copybara-github

  committed
  354e00c

  Copy full SHA for 354e00c

• Run mobilenet_v2 HLO benchmark on CPU in A/B diff script.

  Show description for b8aef1a
  

  tensorflower-gardener

  authored and

  copybara-github

  committed
  b8aef1a

  Copy full SHA for b8aef1a

• [XLA] Improve conditional code motion to replicate broadcast operations that have multiple users inside conditional branches, instead of stopping the code motion immediately after a broadcast.

  Show description for b3553cc
  

  tensorflower-gardener

  authored and

  copybara-github

  committed
  b3553cc

  Copy full SHA for b3553cc

• Fix include guards under mlir_hlo

  Show description for 30a6082
  

  ddunl

  authored and

  copybara-github

  committed
  30a6082

  Copy full SHA for 30a6082

• Reverts 918e98f

  Show description for 235419a
  

  pkanwar23

  authored and

  copybara-github

  committed
  235419a

  Copy full SHA for 235419a

• [xla:gpu] Only instantiate cuda graph if the captured function gets called frequently

  Show description for 4325f06
  

  anlunx

  authored and

  copybara-github

  committed
  4325f06

  Copy full SHA for 4325f06

• Added the "kScheduleStart" priority policy in latency-hiding-scheduler.

  Show description for 3ff99d7
  

  tensorflower-gardener

  authored and

  copybara-github

  committed
  3ff99d7

  Copy full SHA for 3ff99d7

• Update the version of Google Benchmark to include the fix to build the Benchmark package with libpfm enabled.

  Show description for 89a4d7c
  

  shiqing117

  authored and

  copybara-github

  committed
  89a4d7c

  Copy full SHA for 89a4d7c

• Apply patch to avoid compile error on Windows.

  Show description for 36a6095
  

  akuegel

  authored and

  copybara-github

  committed
  36a6095

  Copy full SHA for 36a6095

Commits on Mar 19, 2023

• Relax dealloc/alloc coalescing restriction.

  Show description for 5711c56
  

  jreiffers

  authored and

  copybara-github

  committed
  5711c56

  Copy full SHA for 5711c56

Commits on Mar 18, 2023

• [XLA:CPU Next] Filter-out scf.forall-specific attrs when lowering to scf.for.

  Show description for 97173dd
  

  pifon2a

  authored and

  copybara-github

  committed
  97173dd

  Copy full SHA for 97173dd

• [xla:runtime] Add support for f8 data types to XLA runtime and custom calls

  Show description for 918e98f
  

  ezhulenev

  authored and

  copybara-github

  committed
  918e98f

  Copy full SHA for 918e98f

• [XLA:CPU] Propagate HloXlaRuntimePipelineOptions for fusion outlining

  Show description for 666a7f2
  

  frgossen

  authored and

  copybara-github

  committed
  666a7f2

  Copy full SHA for 666a7f2

• Integrate LLVM at llvm/llvm-project@a7cf2892e964

  Show description for 9e1f49c
  

  slackito

  authored and

  copybara-github

  committed
  9e1f49c

  Copy full SHA for 9e1f49c

Commits on Mar 17, 2023

• Roll cl/517070827 forward with fix for ROCm.

  Show description for 16e953a
  

  chsigg

  authored and

  copybara-github

  committed
  16e953a

  Copy full SHA for 16e953a

• Remove extraneous | from check_contents.yml

  Show description for 6061922
  

  ddunl

  authored and

  copybara-github

  committed
  6061922

  Copy full SHA for 6061922

• [XLA GPU] Modifying deduplicator pass to support marking duplicate fusions without changing the HLO through a flag. The pass is invoked at the end of the pipeline to capture such information.

  Show description for 926153c
  

  Moerafaat

  authored and

  copybara-github

  committed
  926153c

  Copy full SHA for 926153c

• Replace usage of the tsl::Status(tsl::error::Code, ...) constructor.

  Show description for d6d9544
  

  jblespiau

  authored and

  copybara-github

  committed
  d6d9544

  Copy full SHA for d6d9544

• Integrate LLVM at llvm/llvm-project@e495eabd3268

  Show description for 529b039
  

  slackito

  authored and

  copybara-github

  committed
  529b039

  Copy full SHA for 529b039

• [XLA:CPU] Add xla_cpu_enable_mlir_fusion_outlining flag

  Show description for 5f59984
  

  frgossen

  authored and

  copybara-github

  committed
  5f59984

  Copy full SHA for 5f59984

• [PJRT C API] Implement PjRtCApiBuffer::logical_on_device_shape() reusing PJRT_Buffer_OnDeviceTrimmedShape_Args.

  Show description for a593d60
  

  Jieying Luo

  authored and

  copybara-github

  committed
  a593d60

  Copy full SHA for a593d60

• Update dlpack types.

  Show description for 29b660c
  

  pschuh

  authored and

  copybara-github

  committed
  29b660c

  Copy full SHA for 29b660c

• Remove dispatch.result_handlers since they are not used.

  Show description for a1469a6
  

  yashk2810

  authored and

  copybara-github

  committed
  a1469a6

  Copy full SHA for a1469a6

• [XLA:CPU Next] Add a pattern to fold tensor.cast into scf.forall.

  Show description for a2fef71
  

  pifon2a

  authored and

  copybara-github

  committed
  a2fef71

  Copy full SHA for a2fef71

• Introduce call to benchmark_deps() to recursively fetch dependencies required

  Show description for 7f8cfc9
  

  tensorflower-gardener

  authored and

  copybara-github

  committed
  7f8cfc9

  Copy full SHA for 7f8cfc9

• PR openxla#1965: [Code style]clean header guards from TensorFlow

  Show description for 5d8f791
  

  GHGmc2

  authored and

  copybara-github

  committed
  5d8f791

  Copy full SHA for 5d8f791

Pagination

Previous

Next