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

[Woptim] Update software stack for AMD architecture. Introduce rocm 5.7.1 #81

Merged
merged 16 commits into from
Jan 19, 2024

Conversation

adrienbernede
Copy link
Member

No description provided.

@adrienbernede adrienbernede merged commit 1f9193b into main Jan 19, 2024
@adrienbernede adrienbernede deleted the woptim/rocm-5-7-1 branch January 19, 2024 10:55
@adayton1
Copy link
Member

@adrienbernede, this may have broken Umpire on Tioga (with a develop version of radiuss-spack-configs):

==> Error: concretization failed for the following reasons:

  1. No valid version for 'umpire' compiler 'cce' satisfies '@16.0.0'
    [exe: /dev/shm/tioga30-1640271/spack/bin/spack dev-build --fresh --quiet -d /usr/WS2/umdev/gitlab-runner/umdev/builds/9YsVm9VV/002/gitlab/radiuss/Umpire -u initconfig 'umpire@develop~shared +fortran tests=basic %cce@16.0.0']
    ==> Error: concretization failed for the following reasons:
  2. No valid version for 'umpire' compiler 'cce' satisfies '@16.0.0'

@adrienbernede
Copy link
Member Author

@adayton1 It is true that this change is not transparent. This is why there is a dedicated PR in Umpire: LLNL/Umpire#864
I wasn’t expecting someone to try to update to radiuss-spack-configs@main before the above is merged.

I chose to move from cce@16.0.0 to cce@16.0.1 instead of maintaining both.
What you need to do is to look for cce@16.0.0 in the local CI jobs (.gitlab/jobs/[tioga/corona].yml) and change it for cce@16.0.1.

@adrienbernede
Copy link
Member Author

Now, if you update radiuss-spack-configs to main, please beware of the change in the fmt requirement in the Umpire package. I had to specify the cxx standard for intel compiler when using a recent enough Spack that would trigger the build of fmt@10.1.0+.

@adayton1
Copy link
Member

Thanks, @adrienbernede. I see now that umpire was explicitly overriding the cce@16.0.0 job, so I'll update that. The branch I am working on is the one that starts using the fmt requirement in Umpire. I had to make a patch for fmt to work with another compiler, so there's another update in radiuss-spack-configs that brings in fmt with some changes.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants