Skip to content

Enable BFloat16 and TensorFloat32 conversions for cooperative matrices #2635

Enable BFloat16 and TensorFloat32 conversions for cooperative matrices

Enable BFloat16 and TensorFloat32 conversions for cooperative matrices #2635

Triggered via pull request November 14, 2023 16:03
Status Failure
Total duration 32m 26s
Artifacts 1

check-code-style.yml

on: pull_request
clang-format & clang-tidy
42s
clang-format & clang-tidy
Fit to window
Zoom out
Zoom in

Annotations

1 error and 2 warnings
clang-format & clang-tidy
Process completed with exit code 1.
clang-format & clang-tidy
The following actions uses node12 which is deprecated and will be forced to run on node16: actions/upload-artifact@v2. For more info: https://github.blog/changelog/2023-06-13-github-actions-all-actions-will-run-on-node16-instead-of-node12-by-default/
clang-format & clang-tidy
The `set-output` command is deprecated and will be disabled soon. Please upgrade to using Environment Files. For more information see: https://github.blog/changelog/2022-10-11-github-actions-deprecating-save-state-and-set-output-commands/

Artifacts

Produced during runtime
Name Size
clang-format.patch Expired
2.31 KB