Skip to content

WIP: Add support for Athena/Turin#65

Merged
amolod merged 9 commits into
mainfrom
feature/add-turin-support
Apr 13, 2026
Merged

WIP: Add support for Athena/Turin#65
amolod merged 9 commits into
mainfrom
feature/add-turin-support

Conversation

@mathomp4
Copy link
Copy Markdown
Member

@mathomp4 mathomp4 commented Mar 19, 2026

This PR is an attempt to collect the changes needed to run S2S on Athena at NAS.

Current Status

I was able to run a c180 1440x1080 job on Athena with these changes. Some stats:

Run Wall Time Nodes SBU Rate Cores Per Node Approx SBUs
Rome 15:53 18 4.06 120 20
Rome HT 26:23 9 4.06 240 16
Turin 12:53 9 12.20 240 24
Turin HT 13:53 5 12.20 480 14

S2S does seem to not hate using hyperthreading on the Turins.

It does run with hyperthreading on the Romes but was slower. It seems almost too slow, so I'm re-running

Build Updates

The main changes are a new Baselibs build as well as updates to the Make system. One to support MPICH and the other because of a pinstall issue.

Scripting changes

Well, I did try to update parallel_build.csh to work. I think it should work once I get the manual build working but...not sure.

NOTE: At NAS, I now try to have "sensible" defaults for parallel_build.csh. If you submit on pfe, you get rom_ait; on afe, mil_ait; and on athfe, tur_ath.

I also put in a bit of a hack in gcm_setup so that if you run that from Athena (or I guess TOSS5) you get a Turin job out.

What will be needed more is help from @qvis for all the scripting around the build to know about Turin. A simple example is gcm_setup but then we have all the things that ask "am I at NAS".

For Athena, that means you don't look for pfe but for athfe and compute nodes have hostnames like:

($node =~ x[0-9]*c[0-9]*s[0-9]*b[0-9]*n[0-9]*) ) then

@mathomp4 mathomp4 self-assigned this Mar 19, 2026
@mathomp4 mathomp4 added the enhancement New feature or request label Mar 19, 2026
@mathomp4 mathomp4 added the Skip Changelog Skips the Changelog Enforcer label Mar 20, 2026
@mathomp4 mathomp4 marked this pull request as ready for review March 20, 2026 17:58
@mathomp4 mathomp4 requested a review from a team as a code owner March 20, 2026 17:58
@amolod amolod merged commit d7a6829 into main Apr 13, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

enhancement New feature or request Skip Changelog Skips the Changelog Enforcer

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants