2025.02.27 Meeting Notes

Agenda

• Individual/group updates
• Next Parthenon dev/user(?) meeting
• Review non-WIP PRs

Attendees: BB, PM, LR, AR,

Individual/group updates

LR

• looking at coalesced comms again
• PG should send AthenaPK inputs to LR

PG

• first paper on INCITE sims submitted
• Particle load balancing?
  • not implemented
  • Having issues for sims with particles that carry many (~170) field components that results in quite an uneven memory usage across blocks

Next Parthenon dev/user(?) meeting

• PM is leading the effort/org
• Tentative date: 16-20 Jun
• Scope/duration
  • now more downstream users compared to 2022
  • new collab with UMich (potentially many [new] user)
  • Current codes (downstream codes and application areas much more diverse) (roughly day 1)
    • Riot
    • Asail
    • Artemis
    • Jaybenne
    • Phoebus
    • SingularityEOS
    • Flash
    • AthenaPK
    • (AthenaK)
  • Poster session for all codes
  • Larger slot for new downstream codes to present
  • Dedicated session for the new cool stuff in Parthenon since the last workshop (roughly day 2)
    • Non cc-fields
    • Multigrid
    • Forrest OctTree
    • Curvilinear Coords
    • Coalesced comms
    • Outputs: OpenPMD/Ascent
  • Session on future topics/developments
    • Tensor networks
    • Reduced order methods
    • ...
• Hackathon component (roughly day 3/4/5)
  • more structured (with clear targets)
  • more unstructured (focusing on directions/strategic decision) <- probably this

Status of Ascent

• should work as is but has not been used in ~1.5 years
• PG is optimistic that some of the issues (performance, compatibility with the lib stack) have been resolved by now

Tentative next meeting: 13 Mar