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.

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

Update oneapi modules on Sunspot and Aurora #6944

Merged
merged 8 commits into from
Jan 31, 2025

Conversation

amametjanov
Copy link
Member

@amametjanov amametjanov commented Jan 25, 2025

Update oneapi modules on Sunspot and Aurora:

  • compiler to latest available 2024.07.30.002
  • (p)netcdf libraries to latest versions built with updated compiler
  • update queues on aurora
  • default F-compsets single-node PEs
  • update project paths on Aurora
  • remove -fpe0 from debug flags

[BFB]

@amametjanov amametjanov added Machine Files BFB PR leaves answers BFB Aurora labels Jan 25, 2025
@amametjanov amametjanov self-assigned this Jan 25, 2025
@amametjanov
Copy link
Member Author

amametjanov commented Jan 25, 2025

Testing on Aurora:

@rljacob
Copy link
Member

rljacob commented Jan 27, 2025

How does this relate to the work being done in #6916 ?

@oksanaguba
Copy link
Contributor

Az asked me if we want to "unify" machine blocks in aurora and in my spot-clean branch. i originally said no, but maybe we should. right now there are other issues with spot-clean branch, after they are resolved, we can get config_machines sorted.

@amametjanov
Copy link
Member Author

This PR should be complementary.

@amametjanov amametjanov marked this pull request as ready for review January 27, 2025 23:02
amametjanov added a commit that referenced this pull request Jan 31, 2025
(PR #6944)

Update oneapi modules on Sunspot and Aurora:
- compiler to latest available 2024.07.30.002
- (p)netcdf libraries to latest versions built with updated compiler
- update queues on aurora
- default F-compsets single-node PEs
- update project paths on Aurora

[BFB]
@@ -3240,10 +3240,8 @@
<BATCH_SYSTEM>pbspro</BATCH_SYSTEM>
<SUPPORTED_BY>e3sm</SUPPORTED_BY>
<MAX_TASKS_PER_NODE>208</MAX_TASKS_PER_NODE>
<MAX_TASKS_PER_NODE compiler="oneapi-ifx">208</MAX_TASKS_PER_NODE>
<MAX_TASKS_PER_NODE compiler="oneapi-ifxgpu">104</MAX_TASKS_PER_NODE>
<MAX_TASKS_PER_NODE compiler="oneapi-ifxgpu">96</MAX_TASKS_PER_NODE>
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why is this changed for a specific compiler?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

For now, setting it to 2x the number of MPI tasks per node (48): 96 threads/node for GPU builds.
On my to-do is to look at best combinations of tasks and threads for each compiler build.

@amametjanov amametjanov merged commit caf0f37 into master Jan 31, 2025
6 checks passed
@amametjanov amametjanov deleted the azamat/alcf-machines/update-oneapi-stack branch January 31, 2025 20:11
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Aurora BFB PR leaves answers BFB intel Intel compilers Machine Files Ready to merge
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants