Code

Triton#6798

Code

Merged

bhavya01 merged 67 commits intomasterpytorch/xla:masterfrom tritonpytorch/xla:tritonCopy head branch name to clipboardJun 7, 2024

+450-14

Commits

Commits on Oct 4, 2023

• Update infra_triggers.tf

  ManfeiBai

  authored
  6dccf0a

Commits on Mar 20, 2024

• Skeleton trition support

  bhavya01

  committed
  9828123

• Merge branch 'master' into triton

  bhavya01

  committed
  99bf48d

Commits on Mar 21, 2024

• Fix bugs

  bhavya01

  committed
  b89e558

• Fix custom call invocation

  bhavya01

  committed
  64189bd

Commits on Mar 22, 2024

• Refactor to include gpu custom call and create triton dir

  bhavya01

  committed
  0c208ef

• Lint fixes

  bhavya01

  committed
  b553ba7

• python lint fix

  bhavya01

  committed
  c5129e6

Commits on Mar 27, 2024

• Updated base image for CI

  bhavya01

  committed
  48e7127

Commits on Mar 28, 2024

• Update github workflow gcr image

  bhavya01

  committed
  e04fc97

• Merge branch 'master' into custom

  bhavya01

  committed
  37bf127

• Remove xrt build and test file

  bhavya01

  committed
  6061895

• Add temporary test to run triton kernel

  bhavya01

  committed
  f59ddbf

• Fix tests

  bhavya01

  committed
  158aed4

Commits on Mar 29, 2024

• Update payload for xla gpu custom call

  bhavya01

  committed
  87b92c5

• Update gpu runner

  bhavya01

  committed
  847ccc5

Commits on Apr 4, 2024

• Merge branch 'master' into triton

  bhavya01

  authored
  eca6d52

Commits on Apr 12, 2024

• Extract payload from triton kernel programatically

  bhavya01

  committed
  2348ca3

• Merge branch 'master' into triton

  bhavya01

  committed
  110c8c6

• Lint fixes

  bhavya01

  committed
  a226150

• Only build triton files for GPU

  bhavya01

  committed
  4c1f4f5

Commits on Apr 13, 2024

• build pytorch for ampere gpus

  bhavya01

  committed
  431f822

• c++ lint fix

  bhavya01

  committed
  4bade16

• Python lint fix

  bhavya01

  committed
  1c5b47d

• Fix torch cuda arch list

  bhavya01

  committed
  3138a92

• Use a bigger machine for CI build

  bhavya01

  committed
  3f00cfd

• Add triton test to run_tests.sh

  bhavya01

  committed
  e729cfb

Commits on Apr 15, 2024

• Update triton env variable

  bhavya01

  committed
  8e304c0

• Set up a separate CI for triton tests

  bhavya01

  committed
  27bdc3a

• Fix github workflow to add _triton.yml

  bhavya01

  committed
  9a3ef84

• Rebuild torch xla for triton tests

  bhavya01

  committed
  ade444d

Commits on Apr 16, 2024

• Create a separate CI tab for triton tests

  bhavya01

  committed
  cb0bb85

• Separate build and test phase for triton

  bhavya01

  committed
  015b1ad

• Fix flags for docker run container

  bhavya01

  committed
  a18028a

• Update triton.yml to output docker image

  bhavya01

  committed
  993ee92

Commits on May 10, 2024

• Add a python binding to register custom calls and remove jax files

  bhavya01

  committed
  a87b782

• Fix lint

  bhavya01

  committed
  bf05d1b

• Merge main

  bhavya01

  committed
  4582fe8

• Merge master

  bhavya01

  committed
  9680167

• Merge master after updating

  bhavya01

  committed
  a7b94c6

• Update CI to use cuda plugin

  bhavya01

  committed
  e14636a

• Install jaxlib while setting up triton tests

  bhavya01

  committed
  256d819

• Install triton package while running triton tests

  bhavya01

  committed
  c616e64

Commits on May 13, 2024

• Experimental: Build pytorch with cuda

  bhavya01

  committed
  60b8d18

Commits on May 14, 2024

• Revert build pytorch with CUDA

  bhavya01

  committed
  2bde624

• Merge branch 'master' into triton

  bhavya01

  committed
  e6c4e0a

• Remove ansible path for triton CI

  bhavya01

  committed
  14ee545

Commits on May 20, 2024

• Style fixes

  bhavya01

  committed
  25acb26

Commits on May 28, 2024

• [Experimental] test new CI

  bhavya01

  committed
  6b0ac18

• [Experimental] Set XLA_CUDA=0 for cuda arch in ansible

  bhavya01

  committed
  4d97150

Commits on May 29, 2024

• [Experimental] Update CI to build pytorch cuda with ansible

  bhavya01

  committed
  e079049

Commits on May 30, 2024

• Update CI

  bhavya01

  committed
  d9c89b6

• Fix CI workflow file

  bhavya01

  committed
  7a6c809

• Fix CI workflow

  bhavya01

  committed
  6b1954d

• Fix the wheels installed for tests requiring torch cuda

  bhavya01

  committed
  21797a6

Commits on May 31, 2024

• Add compute_capability=8.6 for xla cuda plugin

  bhavya01

  committed
  e6e89d3

• update TORCH_CUDA_ARCH_LIST

  bhavya01

  committed
  ac45fe1

• Experimental build torch and torch_xla cuda wheels

  bhavya01

  committed
  f828fbb

• Merge branch 'master' into triton

  bhavya01

  authored
  ac56c00

• Update build_and_test.yml

  bhavya01

  authored
  c3b8653

• Update dlpack test to only use one device

  bhavya01

  committed
  a1168c6

• Remove compute capability 8.6 from cuda plugin

  bhavya01

  committed
  39551a2

• Remove triton.sh

  bhavya01

  committed
  35e0869

• Default empty torch_cuda_arch_list in ansible config

  bhavya01

  committed
  f95d898

Commits on Jun 5, 2024

• Merge branch 'master' into triton

  bhavya01

  authored
  291104d

Commits on Jun 6, 2024

• Revert CI changes

  bhavya01

  committed
  f5c9b1a

• Revert CI changes pt2

  bhavya01

  committed
  5b23969