-
Notifications
You must be signed in to change notification settings - Fork 859
WeeklyTelcon_20180227
Geoffrey Paulsen edited this page Jan 15, 2019
·
1 revision
- Dialup Info: (Do not post to public mailing list or public wiki)
- Josh Hursey
- Edgar Gabriel
- Ralph
- Brian
- Jeff Squyres
- Mohan
- Josh Ladd
Review All Open Blockers
Review v2.x Milestones v2.1.3
- v2.1.3rc2 is available - https://www.open-mpi.org/software/ompi/v2.1/
Review v3.0.x Milestones v3.0.1
- v3.0.1rc3 is available - https://www.open-mpi.org/software/ompi/v3.0/
- Ship on Friday, unless a blocker appears.
- Will start the v3.0.2 immediate at that point.
- PMI detection configure: https://github.com/open-mpi/ompi/pull/4861
- If rc3 goes well, then we will push this to v3.0.2.
Review v3.1.x Milestones v3.1.0
- v3.1.0rc2 is available - https://www.open-mpi.org/software/ompi/v3.1/
Review Master Master Pull Requests
-
PR 4854: ORTE update to PMIx v3 PR for review
- Good to merge after discussing the functionality being added.
-
Issue 4795 OB1 out-of-order issue on master
- PR 4852 Did not address the issue.
- Needs more work
-
PR 2025 New activity with a reproducer.
- Multiple threads calling
MPI_Comm_dup
can cause a hang - nextcid needs to progress, but it is blocking... May be worse than that - since it requires re-tooling an older algorithm design.
- We need to discuss this more in the face-to-face.
- Multiple threads calling
- Ralph to present on PMIx-based debugger support (eventual replacement for MPIR)
- CID allocation algorithm per PR 2025
- Update your MTT INI files to reflect new URLs (see email)
It should just be updating lines like:
ompi_snapshot_url = https://www.open-mpi.org/nightly/master
to read:
ompi_snapshot_url = https://download.open-mpi.org/nightly/open-mpi/master
follows the pattern:
https://download.open-mpi.org/nightly/open-mpi/<release branch>/<tarball name>
- Do this before the Face-to-Face meeting (March 18th)
Review Master MTT testing
- Mellanox, Sandia, Intel
- LANL, Houston, IBM, Fujitsu
- Amazon,
- Cisco, ORNL, UTK, NVIDIA