Immediately delete cuda-windows GitHub Artifacts - #22863
Merged
Merged
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/22863
Note: Links to docs will display an error until the docs builds have been completed. ⏳ No Failures, 2 PendingAs of commit 94742c0 with merge base 0aa375a ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
Bot
added
the
CLA Signed
jathu
force-pushed
the
jathu/ci-reduce--cuda-windows
branch
from
September 16, 2026 00:52
c34ca98 to
94742c0
Compare
jathu
marked this pull request as ready for review
September 16, 2026 00:53
Gasoonjia
approved these changes
Sep 16, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
In #22848 and #22854 we could export and test on the same machine. We can't do that here because the models are exported on Linux and tested on Windows, so we still have to upload them for the Windows job to download. This PR just deletes them once the Windows job is done.
GitHub charges for artifact storage by the hour. Each run uploads 18.4 GB, and right now that sits around for the full 2 days of retention even though the Windows job is done with it within an hour. In August this workflow ran about 70 times a day, so at that rate this brings us from roughly $650/month down to $12. It also takes about 500 GB off our live artifact storage.
If a test fails we leave the artifacts alone so you can still re-run the failed jobs or debug. They expire with the normal retention (currently 2 days).
Test plan
CI. The
delete-model-cuda-windows-artifactsjob should log six deletions and the run should end with zero artifacts.