Firedrake meeting 2022 09 14

Date and time 2022-09-14 15:00UTC (16:00BST 22:00AEST)

Action Items

1. Pick Chair and Minuter (SV 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 (done? maybe we can take this out for now)
5. JB: SIAM CSE23 Minisymposium/minisymposterium

Minutes

Present: David, Scott, Rob, Reuben, Connor, Daiane, Colin, Jack, Koki, Nacime, Me

Apologies:

Firedrake"22"

• 4th-6th of January, down of Exeter, co-organised with Jemma and Nele
• invited speakers are not finalised yet
• 48 hour meeting

JB: Status of PETSc/petsc4py GC MR

Since rebasing this now has more memory leaks and needs fixing before it will pass tests. Perhaps revisit this next week.

CW: assign performance issues worth fixing?

I recall that DH described assign as equivalent to interpolate in many cases and that potentially assign would be pared back to only work for the cases where it actually makes sense. TK recently found a performance issue in assign that would require a relatively invasive change to fix. Is it worth me spending time on a fix if we're about to change things anyway?

• problem is that rhs can be anything
• assign should only weights sum of things in the same space (unless maybe different components in the same vector space)
• @Connor don't bother fix the performance bug

RK: Irksome's docs builds are failing

Merge PRs

Date of next meeting

1600 UTC 2022-09-21