Skip to content

ASC Q4 2022 Meeting

Jump to bottom
Thomas Naughton edited this page Oct 27, 2022 · 5 revisions

PMIx Standard Administrative Steering Committee (ASC) 4Q 2022 Meeting

Quick Links

  • Governance Document [latest]

Agenda (Finalized Oct 13, 2022)

This meeting has a floating agenda with specific synchronization points to keep us on track. Rough time estimates are provided per agenda item, and the co-chairs plan to cover the topics in the order seen below. However, since some agenda items will take longer/shorter than anticipated, an exact start/end timing is not guaranteed, and some items may float to the second day. If you are not able to attend the full meeting and are presenting, then please let the co-chairs know, and we can plan accordingly.

Day 1: Oct 25 (10 am - 1 pm US Central Daylight Time)

Start End Topic
10:00 am 10:05 am Gathering (--)
10:05 am 10:10 am Roll Call (We will start roll call promptly at this time)
10:10 am 11:30 am Discussion of agenda items
11:30 am 11:45 am Break
11:45 am 1:00 pm Discussion of agenda items

Day 2: Oct 27 (10 am - 1 pm US Central Daylight Time)

Start End Topic
10:00 am 10:05 am Gathering (--)
10:05 am 11:30 am Discussion of agenda items
11:30 am 11:50 am Voting and Break
11:50 am 12:30 am Administrative and Working Group agenda items
12:30 am 12:45 pm Technical and Use Case Presentation(s)
12:45 am 1:00 pm Closing discussion and wrap up

Agenda Items

  • Governance PRs up for a Second Vote:
    • None
  • Governance PRs up for a Reading and First Vote:
    • None
  • PMIx Standard PRs up for a Reading (Provisional):
    • None
  • PMIx Standard PRs up for a Reading (Errata):
    • None
  • PMIx Standard PRs up for a Second Vote:
    • None
  • PMIx Standard PRs up for a Reading and First Vote:
  • Plenary discussion items
    • None
  • Revision Exception Votes
    • None
  • Presentation of the v5.0 Standard Release Candidate for discussion (Ken/Dave)

Administrative and Working Group Agenda Items

  • Review quarterly meetings dates and plans
4Q 2022 - Virtual
  - Oct. 25 & 27 (10 am - 1 pm US Central)
---
1Q 2023 - Virtual
  - Jan. 24 & 26 (10 am - 1 pm US Central)
2Q 2023 - Virtual
  - TBD
3Q 2023 - Virtual
  - TBD
4Q 2023 - Virtual
  - TBD
  • Elect new officers
  • ASC Membership
    • Vote on new ASC Members
    • Call for new ASC Members
  • Release Planning
  • Working Group Updates (~ 10-15 minutes each)
    • Client Separation / Implementation Agnostic Document
    • Tools & Dynamic Workflows
    • Open Call for New Working Groups
  • Technical and Use Case presentations
    • TBD
  • Additional discussion items

Meeting Notes:

Attendance

Person Institution Day 1 Day 2
Josh Hursey IBM x x
Aurelien Bouteiller UTK x x
Howard Pritchard LANL x x
Michael Karo Altair x x
Norbert Eicker JSC x x
Thomas Naughton ORNL x x
Ralph Castain Nanook x x
Isaias Compres TUM x x
Dominik Huber TUM x
Ken Raffenetti ANL x x
Kathryn Mohror LLNL x
David Solt IBM x x
Brice Goglin Inria x x

Day 1: Oct. 25, 2022

  • PMIx Standard PRs up for a Reading and First Vote:
    • Publish/Lookup Chapter (Dave ~30 min) https://github.com/pmix/pmix-standard/pull/398
      • The name of the function is Publish2, which is not great, but do we have a better idea that still link to the original?
      • New retrieval rules: the retrieval is done on the same range it was published (no inheritance, hierarchy)
        • Ralph says that the new rule is too strict: the publisher should decide what range should see, the requestor then can restrict if it sees an info published from a specific range
        • Discussion about changing the semantics a bit to limit publish to own namespaces or to global (avoiding problem scenario of publish to other namespace directly).
        • Also, consideration for avoiding the lookup range argument to be generic from lookup side (not specific to origin of data).
        • Ralph raised idea of possibly using “qualified value” recently added in OpenPMIx might be relevant/useful.  Value of key and qualifiers indicate namespace and publisher.
        • Synopsis:
            1. lookup process specify range data published to, but not indicate where publisher is located
            • restricts range only
            • not restrict who published it
            1. no specify ranges at all on lookup, figure out how to return all matches w/ info on where published to
            1. do not allow publishing to range namespace that is outside of your namespace
            • problematic b/c limit to only use range global and possibly result in conflicts (e.g., force encoding "namespace" in the key to avoid conflict, but duplicates problem trying to solve here)
        • Reminder - original task was to rectify issues identified in the Lookup Example in the spec.  In addition to updates to interface, need to update example.
        • Plan to do straw-poll in break to get feedback on how to proceed
  • Presentation of the v5.0 Standard Release Candidate for discussion (Ken/Dave)
    • Timeline is Q1 2023 for rc approval vote
    • V5 branch will soon split off from the master branch
    • Scope: use cases; implementation agnostic; ABI; cleanup and minor additions
    • Full list of changes available in the slices (todo: add slides to the full meeting notes)
    • New milestones for v5.1 and v6 have been created on github
  • Administrative and Working Group Agenda Items
    • Review quarterly meetings dates and plans
    • 4Q 2022 - Virtual - Oct. 25 & 27 (10 am - 1 pm US Central)
    • 1Q 2023 - Virtual - Jan. 24 & 26 (10 am - 1 pm US Central)
    • 2Q 2023 - Virtual - TBD
    • 3Q 2023 - Virtual - TBD
    • 4Q 2023 - Virtual - TBD
  • Elect new officers
  • ASC Membership
    • Vote on new ASC Members
    • Call for new ASC Members
  • Release Planning
  • Working Group Updates (~ 10-15 minutes each)
    • Client Separation / Implementation Agnostic Document
      • Working on event notification chapter (⅔ of the way)
      • Only minor clarifications found, no major API changes
      • connect/disconnect clarifications
    • Tools & Dynamic Workflows
      • Prototyping with Open PMIx for malleable/dynamic processes
      • Allocation request already existing (yeah), need for a cancellation directive (this requires some careful consideration due to timing in distributed systems)
      • Allocation from the RM side, new nodes available, new jobs need space => scheduler offers more nodes, or takes back nodes, still WIP (scheduler driven malleability), prototyping will start soon
      • We need to advertise again to the former Dynamic Workflows group, they have not joined en masse after the WG merger
    • Logistical concerns with WG
      • Where is the info for the WG meetings? WG meeting info has been added to the connection info for the ASG/monthly meetings
      • This info will also appear in the PMIx calendar
    • Open Call for New Working Groups

Day 2: Oct. 27, 2022

  • Voting Results
    • Aurelien Bouteiller and Norbert Eicker confirmed to officer positions unanimous
  • Continued discussion on Publish-Lookup
    • https://github.com/pmix/pmix-standard/pull/398
    • Results from the straw poll about options expressed on day 1: https://github.com/pmix/pmix-standard/pull/398#issuecomment-1293631854
    • Ralph presented a diagram about filtering rules for published access data
    • We are coalescing towards
    • lookup will have a new key “allow_multiple_matches”, when that key is not set, and multiple matches are possible, an error is returned
    • When the key is set, and multiple matches are possible, the array of matches are produced, if only one key, the value is produced
    • It is now possible to publish multiple times to the same range, since multiple matches are now possible, the requestor can pick
    • The requestor filter range is not matching the publisher, it is a hierarchical filter, e.g., requestor calling for GLOBAL range will get a value published by any publisher, even if the value was published on the range NAMESPACE (given that the namespace filter for the requestor would match)
    • Publisher would have a key to ‘overwrite’ its own published value with new access rules
    • Long discussion about what to do for publishing multiple times the same key on multiple ranges and/or same range, our current idea is that this is possible and will be returned as multiple matches in the lookup
    • Name could be ‘publish_data, lookup_data, unpublish_data’ rather than publish2 etc.
    • We will discuss this again on Nov. 10th Monthly meeting
  • Technical and Use Case presentations
    • None
  • Additional discussion items
    • None

Home

Working Groups

Events

Helpful How To...

ASC Quarterly Meetings

ASC Monthly Meetings

Clone this wiki locally