Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Use ACCESS-NRI fork of generic tracers with accessom_coupler #26

Closed
wants to merge 3 commits into from
Closed

Use ACCESS-NRI fork of generic tracers with accessom_coupler #26

wants to merge 3 commits into from

Conversation

dougiesquire
Copy link

See mom-ocean#388 for context and details.

This PR was originally made to mom-ocean/MOM5 but it was decided that we would make the changes here first while some details about the build system and testing are still being resolved. Comments from @aidanheerdegen's review of the original PR have been incorporated here.

Note, this PR includes the ACCESS-NRI fork of generic_tracers as a git submodule.

Note also, this PR removes the need for the ACCESS-OM-BGC build type (and hence the CSIRO_BGC pp def). The ACCESS-OM build type now always builds with the USE_OCEAN_BGC pp def and generic tracer WOMBAT can be configured at runtime. I’ve verified that always building with USE_OCEAN_BGC doesn’t impact performance when generic tracers are not configured. To do this, I ran two ACCESS-OM2 experiments on the NCI Gadi supercomputer:

  • Exp 1 uses this configuration modified to run for 1 year at a time and to use a MOM5 executable built from 59e0aad (the most recent commit on main at the time of opening this PR).
  • Exp 2 uses the same configuration, but the MOM5 executable used includes the changes in this PR.

I ran three years of each. The experiments were run at the same time. The chksum outputs in the logs are identical between the experiments and the walltimes reported by PBS for each year are given in the table below.

  Exp 1 Exp 2
year 1 959 s 908 s
year 2 1031 s 1002 s
year 3 1148 s 1028 s
average 1046 s 979 s

Closes mom-ocean#388

@dougiesquire
Copy link
Author

Build CI is failing. I haven't looked in any detail, but one thing sprang to mind - does the build CI need me to make this from a branch in this repo (rather than a fork)? @CodeGat

@CodeGat
Copy link

CodeGat commented Aug 7, 2024

Hi @dougiesquire - I reran the CI to see if the error was reproducible, and yeah I think you're on the money. Seems like it's looking for the ref on ACCESS-NRI/MOM5 rather than your fork, since that is what is defined in the SPD.

@dougiesquire
Copy link
Author

Thanks @CodeGat . I'll also need to add submodules = True to the SPD. Does the build CI just use the latest version on main of ACCESS-NRI/spack-packages?

@CodeGat
Copy link

CodeGat commented Aug 7, 2024 •
edited
Loading

No, it uses the latest tagged version - in which case you'll need to push a tag after you've merged the PR (2024.08.08). I'll have to do a workaround after that because this issue is still plaguing me :D ACCESS-NRI/build-ci#170

@dougiesquire
Copy link
Author

Closing this and will reopen an equivalent PR from a branch in this repo

@dougiesquire dougiesquire mentioned this pull request Aug 7, 2024
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@aidanheerdegen aidanheerdegen Awaiting requested review from aidanheerdegen

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

Allow generic tracers with accessom coupler
2 participants
@dougiesquire @CodeGat