Firedrake meeting 2022 08 25 - firedrakeproject/firedrake GitHub Wiki

Date and time 2022-08-25 12:00UTC (13:00BST 22:00AEST)

Action Items

  1. Pick Chair and Minuter (JB to pick).
  2. ALL: (ongoing) triage the open issues and confirm if they are indeed still open (and perhaps provide labels)
  3. JB: Look into updating the @parallel test marker (ongoing)
  4. DH: talk to Jemma about Firedrake 2022 (ongoing)
  5. JB: SIAM CSE23 Minisymposium/minisymposterium
  6. DH: Produce an MFE for "adjoint of Interp" issue - RESOLVED

Agenda

Present: CW (minuter), DH, RNH, SV, KS, JB, RK

Apologies:

JB: Rescheduling of this meeting

  • JB: In on Tuesdays and Thursdays
  • DH: In on Wednesday next week
  • No meeting for next week
  • DH: Aim for 4pm so US can attend
  • RK: Monday, Wednesday, Friday all work

JB: CSE23 minisymposium

  • JB: 5 confirmed. Still waiting on Jack Hale. 2 slots for Firedrakers left. These can be given to JB and CW.

DH: EGU

  • DH: Plans are afoot

DH: Firedrake 22

  • DH: Guest speaker from Dedalus ('spectral element Firedrake')

KS: I/O

DH: Firedrake UFL fork Serendipity elements

  • DH: Some elements got merged into the Firedrake UFL fork and not into FEniCS UFL creating an unnecessary conflict.
  • RK: Keep ENF in UFL, make them go to cubes, yank div and curl, get merged to FEniCS then propagate through Firedrake. Action point DH

KS: Archer2

  • OFI (implementation of MPI) network seems working (can achieve ~4.5 [GiB/s])

  • UCX (which is supposed to be better) is still not working

  • JB: This is the fabric layer. Archer 2 supports two (which is unusual for clusters). UCX is supposed to be substantially better.

  • KS: PETSc works fine with UCX, but Firedrake breaks.

  • DH: Try to create a MFE (maybe with mpi4py) Action point KS

Merge PRs

Date of next meeting

Next meeting: Not next week

1600 UTC 2022-09-08