Skip to content

Fix the libnbc MPILAllreduce ring algorithm. #9393

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

Merged
merged 1 commit into from
Oct 7, 2021

Conversation

bosilca
Copy link
Member

@bosilca bosilca commented Sep 20, 2021

The ring version did not support MPI_IN_PLACE and had no protection against it,
so instead of nicely bailing out was delivering wrong results. This patch
addresses this issue, simplifies the computation of the segment sizes, and
minimize the temporary memory size.

Fixes #9385.

Signed-off-by: George Bosilca bosilca@icl.utk.edu

@bosilca bosilca added this to the master milestone Sep 20, 2021
The ring version did not support MPI_IN_PLACE and had no protection against it,
so instead of nicely bailing out was delivering wrong results. This patch
addresses this issue, simplifies the computation of the segment sizes, and
minimize the temporary memory size.

Fixes open-mpi#9385.

Signed-off-by: George Bosilca <bosilca@icl.utk.edu>
@jsquyres jsquyres merged commit 8702ff5 into open-mpi:master Oct 7, 2021
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.

Iallreduce gives an incorrect result for algorithms 1 and 3
3 participants