llnl / llnl/AMG2023

Caliper annotation fixes

Open
#7 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
C
Stars
9
Forks
18
PR merge metrics
No merged PRs in 30d

Description

Work in progress list, still need to review code in more depth

- [ ] setup-FOM goes to above `HYPRE_BoomerAMGGetCumNnzAP`, end-FOM to right after, not print.
- [ ] replace multiple Adiak calls to single Adiak call
- [ ] FOM1 adiak_namevalue after `FOM1 = `
- [ ] FOM1 calculation incorrect, wall_time not updated -- get MPI time before and after preconditioner, then use the difference as the wall_time denominator in FOM1 calculation
- [ ] remove setup-FOM region, leave adiak reporting of FOM1
- [ ] remove calculate-FOM region, leave adiak reporting of solve and final FOM

`HYPRE_PCGSetup(pcg_solver, (HYPRE_Matrix)parcsr_A, (HYPRE_Vector)b, (HYPRE_Vector)x);` is preconditioner

@pearce8 @daboehme

Contributor guide

No contributing guide indexed for this repository

Research direction

Start at the HYPRE_PCGSetup call and inspect the surrounding setup-FOM and calculate-FOM regions, including the existing Adiak calls. Review the listed annotation placement, reporting, and FOM1 timing requirements against the current code. Done means each checklist item is addressed and the setup, solve, and final FOM values are reported as specified.

Written by the indexing model from the issue text.

Assessment

Tech stack
c
Domain
hpc, performance
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.