Build Linux CUDA Wheels

Workflow runs for Build Linux CUDA Wheels

• [Wasm] Let the tokenizer bindings load HF and Tekken tokenizers

  requires action with the application:
  Build Linux CUDA Wheels#6171:Pull request#25103openedbyabdelaziz-mahdy
  abdelaziz-mahdy:fix/wasm-tokenizers-exceptions

  Nov 15, 2026Action required
  More actions for workflow run 6171

• [LLM runner] Keep stop() requests made during prefill

  requires action with the application:
  Build Linux CUDA Wheels#6170:Pull request#25101openedbyabdelaziz-mahdy
  abdelaziz-mahdy:fix/llm-stop-during-prefill

  Nov 15, 2026Action required
  More actions for workflow run 6170

• ci: consolidate pinned PyTorch installation

  completed successfully:
  Build Linux CUDA Wheels#6169:Pull request#24605synchronizebyzingo
  zingo:ci-consolidate-pinned-PyTorch-installation

  Nov 15, 202635m 17s
  More actions for workflow run 6169

• Run Windows Python unit tests in parallel

  completed successfully:
  Build Linux CUDA Wheels#6168:Pull request#24895synchronizebyzingo
  zingo:Run-Windows-Python-unit-tests-in-parallel

  Nov 15, 202634m 0s
  More actions for workflow run 6168

• Fix TorchAO source installs and automatic pin updates

  completed successfully:
  Build Linux CUDA Wheels#6167:Pull request#24896synchronizebyzingo
  zingo:Fix-TorchAO-source-installs-and-automatic-pin-updates

  Nov 15, 202634m 24s
  More actions for workflow run 6167

• 2026-11-15 nightly release (b3b565b5b62f5e03d57c9054b5a2570a04f0d937)

  completed successfully:
  Build Linux CUDA Wheels#6166:Commitc4af3e1pushedbypytorchbot
  nightly

  Nov 15, 202645m 51s
  More actions for workflow run 6166

• [executorch][cuda] Deprecate triton_kernel_mode=OFF

  completed successfully:
  Build Linux CUDA Wheels#6165:Pull request#25082synchronizebyGasoonjia
  gh/Gasoonjia/108/head

  Nov 15, 202638m 8s
  More actions for workflow run 6165

• [executorch][cuda] Route 3-input-channel window and kernel-equals-stride convolutions to the implicit-GEMM template

  completed successfully:
  Build Linux CUDA Wheels#6164:Pull request#25087synchronizebyGasoonjia
  gh/Gasoonjia/113/head

  Nov 15, 202635m 15s
  More actions for workflow run 6164

• [executorch][cuda] Run 1-D convolutions on the implicit-GEMM template

  completed successfully:
  Build Linux CUDA Wheels#6163:Pull request#25089synchronizebyGasoonjia
  gh/Gasoonjia/115/head

  Nov 15, 202635m 26s
  More actions for workflow run 6163

• [executorch][cuda] Add planned replacement to the Triton replacement pass

  completed successfully:
  Build Linux CUDA Wheels#6162:Pull request#25083synchronizebyGasoonjia
  gh/Gasoonjia/109/head

  Nov 15, 202635m 31s
  More actions for workflow run 6162

• [executorch][cuda] Add an implicit-GEMM template for dense triton::convolution

  completed successfully:
  Build Linux CUDA Wheels#6161:Pull request#25086synchronizebyGasoonjia
  gh/Gasoonjia/112/head

  Nov 15, 202637m 27s
  More actions for workflow run 6161

• [executorch][cuda] Run transposed convolutions on the implicit-GEMM template

  completed successfully:
  Build Linux CUDA Wheels#6160:Pull request#25091synchronizebyGasoonjia
  gh/Gasoonjia/117/head

  Nov 15, 202637m 6s
  More actions for workflow run 6160

• [executorch][cuda] Test 3-D convolutions on the implicit-GEMM template

  completed successfully:
  Build Linux CUDA Wheels#6159:Pull request#25090synchronizebyGasoonjia
  gh/Gasoonjia/116/head

  Nov 15, 202635m 41s
  More actions for workflow run 6159

• [executorch][cuda] Route 2-D depthwise and few-channel grouped convolutions to a Triton direct-stencil kernel

  completed successfully:
  Build Linux CUDA Wheels#6158:Pull request#25085synchronizebyGasoonjia
  gh/Gasoonjia/111/head

  Nov 15, 202635m 35s
  More actions for workflow run 6158

• [executorch][cuda] Converted convolutions follow the output layout the graph recorded

  completed successfully:
  Build Linux CUDA Wheels#6157:Pull request#25088synchronizebyGasoonjia
  gh/Gasoonjia/114/head

  Nov 15, 202635m 30s
  More actions for workflow run 6157

• [executorch][cuda] Add triton::convolution and the conv routing framework

  completed successfully:
  Build Linux CUDA Wheels#6156:Pull request#25084synchronizebyGasoonjia
  gh/Gasoonjia/110/head

  Nov 15, 202634m 47s
  More actions for workflow run 6156

• WIP: Fix complex div broadcasting in CPU kernels

  requires action with the application:
  Build Linux CUDA Wheels#6155:Pull request#24409synchronizebymshr-h
  mshr-h:fix-complex-div-broadcasting

  Nov 15, 2026Action required
  More actions for workflow run 6155

• WIP: Fix complex div broadcasting in CPU kernels

  requires action with the application:
  Build Linux CUDA Wheels#6154:Pull request#24409synchronizebymshr-h
  mshr-h:fix-complex-div-broadcasting

  Nov 15, 2026Action required
  More actions for workflow run 6154

• [PyBindings] Pass float8 tensors through DLPack

  completed successfully:
  Build Linux CUDA Wheels#6153:Pull request#25098synchronizebyshoumikhin
  shoumikhin:float8-dlpack

  Nov 15, 202634m 35s
  More actions for workflow run 6153

• WIP: Fix complex div broadcasting in CPU kernels

  requires action with the application:
  Build Linux CUDA Wheels#6152:Pull request#24409synchronizebymshr-h
  mshr-h:fix-complex-div-broadcasting

  Nov 15, 2026Action required
  More actions for workflow run 6152

• [pytorch hash update] update the pinned pytorch hash

  completed successfully:
  Build Linux CUDA Wheels#6151:Pull request#10955synchronizebypytorchupdatebot
  update-pytorch-commit-hash/15080763801-537-1

  Nov 15, 202634m 1s
  More actions for workflow run 6151

• [PyBindings] Pass float8 tensors through DLPack

  completed successfully:
  Build Linux CUDA Wheels#6150:Pull request#25098synchronizebyshoumikhin
  shoumikhin:float8-dlpack

  Nov 15, 202634m 17s
  More actions for workflow run 6150

• [PyBindings] Pass float8 tensors through DLPack

  completed successfully:
  Build Linux CUDA Wheels#6149:Pull request#25098openedbyshoumikhin
  shoumikhin:float8-dlpack

  Nov 14, 202635m 11s
  More actions for workflow run 6149

• Respect CPU quotas when sizing quantized GEMM compiler workers

  completed successfully:
  Build Linux CUDA Wheels#6148:Pull request#24889synchronizebyGasoonjia
  fix/muse-cgroup-compile-workers

  Nov 14, 202633m 41s
  More actions for workflow run 6148

• Read the CUDA cache counters once per autotune capture

  completed successfully:
  Build Linux CUDA Wheels#6147:Pull request#25060synchronizebyGasoonjia
  gasoonjia/cuda-graph-timing-cheap-stats

  Nov 14, 202634m 13s
  More actions for workflow run 6147