Skip to content
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.

Sign up for GitHub

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

Jump to bottom

MPI3 Shared Memory Issues #257

Open
gzagaris opened this issue Jun 13, 2020 · 1 comment
Open

MPI3 Shared Memory Issues #257

gzagaris opened this issue Jun 13, 2020 · 1 comment
Assignees
@gzagaris
Labels
bug Something isn't working memory mpi Related to MPI communication

Comments

@gzagaris
Copy link
Member

gzagaris commented Jun 13, 2020 •
edited
Loading

This is a follow up from the issue in #220.

There was a subsequent segfault related to consecutively creating/destroying shared memory MPI windows.

The segfault is inside MPI. @adammoody has been looking at this and the suggested workaround at the present time is:

  • Use mvapich2-2.3, mvapich2-2.2 has some known bugs and is not recommended.
  • set the following enviroment variable: MV2_SHMEM_COLL_NUM_COMM=1024

The corresponding ticket that has been filed internally to the MVAPICH support team is:

  • Ticket #83 "Bad baseptr address returned from MPI_Win_shared_query"

We should document this in the RELEASE-NOTES for the next Axom release.
@gzagaris gzagaris added bug Something isn't working mpi Related to MPI communication labels Jun 13, 2020
@gzagaris gzagaris self-assigned this Jun 13, 2020
@gzagaris
Copy link
Member Author

gzagaris commented Sep 3, 2020

Just following up on this -- @adammoody got a patch from the MVAPICH team that fixes this issue.

The MVAPICH team is looking to finalize the code and this fix should become available in an upcoming MVAPICH release. Not sure on an actual ETA and schedule for the MVAPICH release.

gzagaris added a commit that referenced this issue Sep 23, 2020
@gzagaris
DOC: document known issues with MPI3 shared memroy
60ce66b 
Add known issues encounted when using MPI3 shared memory to
the "Known Bugs" section of the Release Notes.

This refers to issue #257.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@gzagaris gzagaris

Labels
bug Something isn't working memory mpi Related to MPI communication
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@gzagaris @kennyweiss