Skip to content

Conversation

@TroyGarden
Copy link
Contributor

@TroyGarden TroyGarden commented Sep 9, 2025

Summary:

context

  • the proposal order might vary on different rank in github action
  • skip this test
  • error log: P1939004930
  • order comparison
image
  • ref: D76303748

Differential Revision: D82031515

Summary:

# context
* previous test failed due to `invalid device pointer` while an embedding module (EBC/EC) is created on `cuda:0`, and then the multiprocess forks the test into two threads to enable multi-GPU env.
* the issue comes from the test case breakdown, where the same tenor is freed twice (likely)
* here we modify the test case so that the embedding modules are created after the fork.

Reviewed By: iamzainhuda

Differential Revision: D82009423
Summary:
# context
* the proposal order might vary on different rank in [github action](https://github.com/pytorch/torchrec/actions/runs/17589918113/job/49967902728?fbclid=IwY2xjawMtaPxleHRuA2FlbQIxMQABHqSimhEWaKNU_G5SH_R2nioBbN7otIqoM1RDmlcjIZwB37M0cNWPEiOERRv1_aem_bhDL2MD4niEtgGMZrytbJQ)
* skip this test
* error log: P1939004930

Differential Revision: D82031515
@meta-cla meta-cla bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Sep 9, 2025
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D82031515

TroyGarden added a commit to TroyGarden/torchrec that referenced this pull request Sep 9, 2025
Summary:

# context
* the proposal order might vary on different rank in [github action](https://github.com/pytorch/torchrec/actions/runs/17589918113/job/49967902728?fbclid=IwY2xjawMtaPxleHRuA2FlbQIxMQABHqSimhEWaKNU_G5SH_R2nioBbN7otIqoM1RDmlcjIZwB37M0cNWPEiOERRv1_aem_bhDL2MD4niEtgGMZrytbJQ)
* skip this test
* error log: P1939004930
* order [comparison](https://www.internalfb.com/intern/diffing/?paste_number=1939241286)
{F1981848607, width=300} 
* ref: D76303748

Reviewed By: aporialiao

Differential Revision: D82031515
@TroyGarden TroyGarden deleted the export-D82031515 branch September 19, 2025 22:02
@TroyGarden TroyGarden changed the title why hashes are different skip consistent hash test in OSS Oct 8, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. fb-exported

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants