-
Notifications
You must be signed in to change notification settings - Fork 900
comm_split_type HW_GUIDED fix MPI_UNDEFINED handling #10702
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
@dalcinl can you take this PR for a spin and let me know if it addresses the issue you mentioned [here]? Thanks! |
@jjhursey Here you have the logs, I believe everything is OK regarding this PR. However, I got PRTE error when trying to run with 5 processes (IIRC, GitHub Action workers have just 2 virtual cores), with oversubscription turned on via
This new issue is most likely unrelated to this PR, but anyway you may be the right person to notify. This is probably a regression after the recent PRTE updates. |
@jjhursey Forgot to mention: PRs related to the new comm split types (this one and the previous ones) should be also merged into branch v5.0.x, right? Maybe you should label them appropriately? Or are they already merged? |
For the runtime issue, I added a note to Issue #10698 on which the team is currently looking at these type of issues. As far as |
FYI: I created a PR for v5.0.x that cherry-picks all of the changes so far. Once this PR is merged then I'll work with the v5 RMs to get it into the v5.0 release cycle. |
Should be fixed in openpmix/prrte#1463 |
Gentle request to move this PR forward. |
@bosilca this is ready to review when you have a chance. Thanks |
@bosilca Can you review? I'd like to get this in so that we can properly test the UNGUIDED PR #10739. @hppritcha I added you as a reviewer in case George doesn't have time. Thanks |
Signed-off-by: Joshua Hursey <[email protected]>
Signed-off-by: Joshua Hursey <[email protected]>
4b6a069
to
6ddcc58
Compare
No description provided.