Skip to content

test: fix tests not working with high nranks and change GA to test that #150

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

Open
wants to merge 2 commits into
base: main
Choose a base branch
from

Conversation

mrava87
Copy link
Contributor

@mrava87 mrava87 commented Jul 1, 2025

This PR fixes a number issues that arose when trying to increase the number of ranks in the test suite to 9 or 16.

As a consquence, I have also decided to modify the GA build to run tests with 2,4,9 ranks instead of 2,3,4

@mrava87 mrava87 requested a review from rohanbabbar04 July 1, 2025 08:39
@mrava87
Copy link
Contributor Author

mrava87 commented Jul 1, 2025

@rohanbabbar04 no idea why the openmpi runs with N=9 stall, but those with intelmpi and mpich don't... any idea?

note that without the proposed changes some of the tests failed for N=9 for me locally... most are tests I added with the mask and add some clear if/else statement on the number of ranks and did not stop if a different one was passed, and another one was for the solvers due to some stochasticity in the random (if I run it independently, instead as part of the entire test suite, they were passing for N=9)

PS: in my local conda env, I used install_conda and got mpi4py installed with MPICH (which seems to be the default if not specifying anything... so I guess I can try openmpi and debug locally..

@mrava87
Copy link
Contributor Author

mrava87 commented Jul 2, 2025

This works locally, so not really sure what is happening in the CI...

image

@rohanbabbar04
Copy link
Collaborator

Hmm, let me give it a look

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants