Summary:
Adding windows CUDA smoke tests on PRs (master should run the full suite).
Next step:
- Automate data update so we get a new smoke test list without manual effort
Pull Request resolved: https://github.com/pytorch/pytorch/pull/59686
Test Plan: https://github.com/pytorch/pytorch/actions/runs/958296267 The sharded smoke tests take long still because of dependencies installation
Reviewed By: walterddr
Differential Revision: D29243533
Pulled By: janeyx99
fbshipit-source-id: dde7ba127fa15c95bda0e833cc5311598fb85e2b
Summary:
Enables sharding for Windows on CI. To make that possible, we currently remove the smoke tests tested in shard 1 which don't seem all that important as they are
1. tested on nightlies
2. seems to be tested anyway by running the test suite
Pull Request resolved: https://github.com/pytorch/pytorch/pull/59970
Reviewed By: seemethere
Differential Revision: D29268484
Pulled By: janeyx99
fbshipit-source-id: 7f90d73037cfeb2c267b28714550316eb471b4dd
Summary:
This is branch off of https://github.com/pytorch/pytorch/issues/59970 to only shard on linux so far (we're running in issues with windows gflags).
This would enable sharding of tests on a few Linux jobs on GHA, allowing tts to be essentially halved.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/60124
Reviewed By: zou3519
Differential Revision: D29204211
Pulled By: janeyx99
fbshipit-source-id: 1cc31d1eccd564d96e2aef14c0acae96a3f0fcd0
Summary:
This removes our cancel_redundant_workflows job in favor of GitHub's built in [`concurrency`](https://docs.github.com/en/actions/reference/workflow-syntax-for-github-actions#concurrency) keyword which limits runs of a particularly named group. Since the group names have to be unique per job per PR, it should end up looking something like `filename-job_name-{pr number | sha (for non-PR workflows)}`. There's also a script to check workflows and ensure that it is being properly gated so people don't forget to add the key in the future.
`ruamel.YAML` also didn't like some of the spacing so that is changed but it also makes it more consistent so �
This also has a minor change of renaming the workflow templates from `.in` to `.j2` which is the standard Jinja2 extension that the VSCode extension automatically picks up for syntax highlighting / errors.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/59019
Test Plan: pushed a commit `reset` and then immediately another commit `test`: the jobs from `reset` are cancelled: https://github.com/pytorch/pytorch/actions/runs/880099510
Reviewed By: samestep
Differential Revision: D28722419
Pulled By: driazati
fbshipit-source-id: c547a161877a0583be9d7edb29244b086b6bcad1
Summary:
Followup to https://github.com/pytorch/pytorch/issues/58491:
- use f-string to remove the literal `generated` string from the generator script, so Phabricator no longer thinks it is a generated file
- remove the special logic for `test_runner_type` and instead explicitly specify for every workflow
Pull Request resolved: https://github.com/pytorch/pytorch/pull/58498
Test Plan:
```
make generate-gha-workflows
```
Also, check that Phabricator doesn't classify `.github/scripts/generate_ci_workflows.py` as "Generated changes" in this diff.
Reviewed By: seemethere
Differential Revision: D28516291
Pulled By: samestep
fbshipit-source-id: 8736eaad5d28082490be0a9b2e271c9493c2ba9d
Summary:
This PR simplifies `.github/scripts/generate_ci_workflows.py` by using the same strategy as https://github.com/pytorch/pytorch/issues/54344, representing workflows as plain data to avoid duplicating the definition of the `generate_workflow_file` function. This will make the script easier to maintain if/when that function is modified and/or more workflow types are added.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/58491
Test Plan:
The Lint job in CI; specifically:
```
make generate-gha-workflows
mypy --config mypy-strict.ini
```
Reviewed By: malfet, seemethere
Differential Revision: D28511918
Pulled By: samestep
fbshipit-source-id: aaf415a954d938a29aee7c9367c9bc2b9f44bb01
