Fix MPI issue with vmec crashes in 1st evaluation #446
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Outdated suggestions cannot be applied.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.
When vmec fails to converge on the first function evaluation of an optimization, MPI proc 0 exits with an uncaught exception while other procs do not, causing the code to hang due to the MPI communication getting out of sync. This can result is wasted time on the cluster, as jobs don't exit cleanly and keep occupying nodes even though they are not actually working on an optimization. This PR fixes this issue, and the new test here provides coverage for this issue.