WeeklyTelcon_20160209

Open MPI Weekly Telcon

---

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

Attendees

• Jeff Squyres
• Geoff Paulsen
• Brad Benton
• Edgar Gabriel
• Howard Pritchard
• Joshua Ladd
• Nathan Hjelm
• Nysal Jan
• ralph
• Ryan Grant
• Sylvain Jeaugey
• Todd Kordenbrock
• Yohann Burette

---

Agenda

Review 1.10

• Milestones: https://github.com/open-mpi/ompi-release/milestones/v1.10.3 - Targeting April, unless there is a need.
  • Nathan will look at 0 byte send issue.
  • dev list of SLURM issues already fixed in 1.10.2
  • verbs usNIC not build by default - wait for review by Howard.
  • Fortran 08 - Jeff will take a look at today.
  • SLES 12 - was a race condition fork/exec before SIGCHILD detection. Fixed.
    • Long running jobs (Linpack) still having SIGCHILD issues.

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
  • Issue 1215 https://github.com/open-mpi/ompi/pull/1335: grpcomm errors
    • Ralph is unable to replicate. Didn't see on Trinity and elsewhere at scale. Found where the problem is, but trying to figure out why solution isn't working. Ralph-and-Jeff-are-iterating phase.
  • https://github.com/open-mpi/ompi/issues/1252: bad perf caused by openib
    • Only fails if openib finds valid procs. As soon as you ibv_cq_poll on 2nd socket. Still like 3ms openib intra-node.
    • Specific Mellanox MOFED 3.0 Verbs.
    • Mellanox has seen far socket on 100ish ns, not 7ms!
• PR 927 - need a Ralph review
  • (the X / test fail was due to github being down -- it's a false failure)
• Issue 1299 - Nathan Hang osc pt2pt.
• Milestones: https://github.com/open-mpi/ompi-release/milestones/v2.0.0
• Mellanox would like new entrypoints in hcoll into 2.0
• Issue with Addprocs on big-endian machines. Now want minimum change to get 2.0 out.
  • Easiest solution - for 32bit and BIG-ENDIAN - don't turn on dynamic add-procs.

Review Master / new RFCs:

• RFC to set the add_procs_cutoff to 32. PR1340 *
• --host vs. --hostfile behavior PR1344
  • how many procs to run
  • Jeff would like consistent with how over subscription works, but no -np runs 1 proc.
  • two issues... how many slots, and how many processes.
  • change behavior so that if user doesn't specify -np but DOES specify --host we'll get 1 slot (and one process).
  • keep hostfile behavior same as today.

MTT status:

Status Updates:

• LANL
• Houston
• HLRS
• IBM

---

Status Update Rotation

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

---

Back to 2016 WeeklyTelcon-2016