Hi Kathryn,

I attached the slides I presented to this group in January of 2017. They are the so-called, “What I did on my XMas vacation” slides that recap work I did (with a lot of help from Ralph) to learn how PMIx might interact with a tool like TotalView. It also contains some thoughts about what the next generation of multi-process debugging interfaces might look like. Given the presentation’s age and the fact that I slapped it together, I think it’s fairly close to what’s needed, but should serve only as a strawman.

Slide 5 has a rough breakdown of the “layers” a tool needs. Refining that a bit, I think the layers /  APIs are:

• Job discovery. Not just for MPI. Allows the tool to find existing jobs.
• Job submission, status, cancelation. Not just for MPI. Allows a tool to submit jobs to a batch/RM system, monitor job status, and cancel jobs.
• Process discovery. Not just for MPI. Allows the tool to find existing or dynamically created processes in a job, and query proctable-like info (pids, nodes, and executables).
• Tool daemon spawning. Not just for MPI. Allows a tool to spawn its daemon processes.
• MPI task labeling. MPI-specific. Any time after process acquisition, the tool can discover a process's or thread's "role" as an MPI task, and "label" it accordingly.
• MPI Message Queue Dumping. MPI-specific. Already have the MPI MQD API, not sure how or why we’d change it.
• MPI Opaque Handles. MPI-specific. Open MPI / TotalView prototype (Jeff / Laust).

Cheers, John D.

From: mpiwg-tools <mpiwg-tools-bounces@lists.mpi-forum.org> On Behalf Of Mohror, Kathryn via mpiwg-tools
Sent: Sunday, March 17, 2019 10:47 AM
To: mpiwg-tools@lists.mpi-forum.org
Cc: Mohror, Kathryn <mohror1@llnl.gov>
Subject: [mpiwg-tools] Meeting March 21, 2019

Hey everybody,

For our call this Thursday (3/21) we’ll come back to debugger topics again. The call is at the usual time in the US (8 am Pacific/ 11 am Eastern / 4 pm CET) but note that the US has moved to DST but I don’t think the EU has done so yet.

Webex Info:

https://apps.fz-juelich.de/scalasca/mpiwg-tools-webex/index.php

Agenda:

In the last debugger call, I said that the plan for this meeting would be to get an overview of OMPD and start talking about analogous interfaces for MPI (e.g. revamp MQD). However, in the meantime, I was convinced that we should focus on process acquisition first. So, the plan will be to talk about process acquisition in this meeting. I’ll go through our notes from 2017(ish) to hopefully find the straw men we drafted back then. Please bring your use cases and any ideas you have on this front.

_________________________________________________________________

Kathryn Mohror, kathryn@llnl.gov, https://people.llnl.gov/kathryn

Data Analysis Group @ Lawrence Livermore National Laboratory, Livermore, CA, USA