Open MPI Weekly Telcon

---

• Dialup Info: (Do not post to public mailing list or public wiki)

Attendees

• Geoff Paulsen
• Jeff Squyres
• Howard Pritchard
• Josh Hursey
• Arm Patinyasakdikul
• Joshua Ladd
• Nathan Hjelm
• Nysal
• Ralph
• Ryan Grant
• Sylvain Jeaugey
• Todd Kordenbrock

---

Agenda

Review 1.10

• Milestones: https://github.com/open-mpi/ompi-release/milestones/v1.10.3
• No news / Good news.

Review 2.0.x

• Wiki: https://github.com/open-mpi/ompi/wiki/Releasev20
• Blocker Issues: https://github.com/open-mpi/ompi/issues?utf8=%E2%9C%93&q=is%3Aopen+milestone%3Av2.0.0+label%3Ablocker *
• Milestones: https://github.com/open-mpi/ompi-release/milestones/v2.0.0
  • 2.0.0 mostly ready to go.
  • RDMACM - broken with MT build. It never worked, and never was tested.
    • Not a regression so community is okay with shipping like this.
    • Will ask Chelcio their opinion.
  • In terms of functionality, we are good to go for 2.0.0.
    • yes.
  • PR 1248 & PR 1249 (needed)
    • Merge both of these to v2.x
    • Allow to run both in MTT tonight.
    • Tomorrow Create a Release Candidate 4 after these merges.
  • 1251 - Fix RDMACM Hang / Locking bug.
    • Nathan will pull out Chelcio fix into seperate PR, and run through tests tonight.
    • Mallanox will temporarily remove RDMACM issue with checkins.
  • Goal of releasing 2.0.0 next Tuesday.
  • After than wait a few days, and start merging in 2.0.1 changes a few at a time.
  • Then need a conversation on when to merge repos.

Review Master MTT testing (https://mtt.open-mpi.org/)

• Has improved. 233 failures, currently on Cisco.
  • Cisco - Many Cisco failures are local cluster issues. Art is working on cleaning up.
• Jeff put in a patch into MTT to allow thread hangs to be marked as hangs.
• nVidia failures are all PMIx failures.
  • Giles found a race condition in PMIx 2.0.
• v2.x failures on Comm_spawn_loop.
• overall not too bad.

MTT Dev status:

New Items:

• Face to face coming up
• Need to discuss ways to take payments.
• WebSite transitions
  • Website itself
  • Nightly tarballs
  • Archives of mailing lists entries.
  • Have mbox archives of all of the lists also. But as soon as we move stuff, where do NEW posts get archived?
• Travis was hung over the weekend. Not sure why.
• ibm jenkins was off over the weekend, should be fixed now.

Status Updates:

• Cisco
  • Have Arm, got usNIC BTL thread multiple in master
  • lots of minor bug fixing and 2.x items.
  • been more focused on libfabric stuff.
• NVIDIA
  • Watching MTT
  • when have 2 cpus and IB card on node, might want to use IB card to do transfers between GPUs.

---

Status Update Rotation

1. Mellanox, Sandia, Intel
2. LANL, Houston, IBM
3. Cisco, ORNL, UTK, NVIDIA

---

Back to 2016 WeeklyTelcon-2016