2021.07.22 Meeting Notes

Agenda

• Individual/group updates
• Paper
• Review non-WIP PR

Individual/group updates

LANL/External CS

Jonas Lippuner

• Sparse interface PR
• Ran into issue with DtoH and HtoD CUDA profiler output making CI fail
• Need look into regression test suite now failing because of missing new gold (probably)
• Added some text to paper

Tanner Broaddus

• Will look further into communication patterns and load balancing
• Looking at Parthenon code and comparing back to Athena++ paper
• Question: All calculations on AMR octree are local, correct? Yes (tree construction and rank assignment are all done locally and identically on each rank)

Galen Shipman

• Heads up: Prototyping a benchmarking tool suite that should include Parthenon at some point, also for downstream codes (sending out some slides)

LANL Physics

Question: Are LANL downstream codes doing automated integration tests? No, only manual at this point

Jonah Miller

• Participated in astrophysics conferences, not much time on Parthenon
• Jonah and Ben have Monte Carlo transport running on Phoebus on top of Parthenon, pushing particles
• Working on opacity library for use on GPUs (separate public library: https://github.com/lanl/singularity-opac)
• Added text to paper. Are we adding code snippets?

AthenaPK

Philipp Grete

• Tracking down 0's (see Matrix channel), should be hopefully fixed now
• Reviewed PRs
• Working on adding AMR performance test to regression suite
• Looks like we're getting new CI machine soon, current one is very dated. With new CI machine add Intel and Clang compilers
• More interest in using AthenaPK for actual science applications and pleased with performance, results in more user support requirements and potential postdoc opportunities

Agenda

Paper