Skip to content

Replace Nvidia new site instructions with a tier2 site config#1542

Merged
climbfuji merged 1 commit intoJCSDA:release/1.9.0from
fmahebert:nvhpc_as_tier2_site
Mar 5, 2025
Merged

Replace Nvidia new site instructions with a tier2 site config#1542
climbfuji merged 1 commit intoJCSDA:release/1.9.0from
fmahebert:nvhpc_as_tier2_site

Conversation

@fmahebert
Copy link
Copy Markdown
Contributor

Summary

Various tweaks to the Nvidia instructions ... these may or may not be improvements, so let's discuss.

  1. Provided the nvhpc environment as a site config. This is less flexible to new systems, but easier to set up for the Azure env we're focussed on now. I'm not sure if this is the right move.
  2. Moved from nvhpc 24.9 to 25.1
  3. Removed various customizations from the template's package list. Seems to work without them.
  4. Did NOT change the fact that many low-level tools are installed from the OS package manager. I intended to use gcc as fallback compiler for various packages, but found concretization issues I didn't immediately know how to resolve, so figured I'd put in this PR first.

Testing

Locally on Azure VM, I built this site config then parts of JEDI on top of it.

Applications affected

none

Systems affected

none

Dependencies

Issue(s) addressed

Checklist

  • This PR addresses one issue/problem/enhancement, or has a very good reason for not doing so.
  • These changes have been tested on the affected systems and applications.
  • All dependency PRs/issues have been resolved and this PR can be merged.

@fmahebert fmahebert requested review from climbfuji and stiggy87 March 4, 2025 23:43
@fmahebert
Copy link
Copy Markdown
Contributor Author

@l90lpa does this ping work? For some reason I can't request you as reviewer here.

- spec: [email protected]
prefix: /usr
buildable: False
# TODO: spack-stack won't accept the system version, must build own
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Because it wants a newer version?

Copy link
Copy Markdown
Contributor Author

@fmahebert fmahebert Mar 5, 2025

Choose a reason for hiding this comment

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

System (Ubuntu 24.04) has 4.3, but spack-stack wants either 4.2 or 4.4.

Azure Ubuntu 24.04 with Nvidia compilers
----------------------------------------

This site config describes a minimal spack-stack install suitable for an Ubuntu 24.04 VM with Nvidia HPC SDK 25.1. For instructions on setting up the VM before building spack-stack, consult the ``README.md`` in the ``sites/tier2/ubuntu2404-nvhpc``.
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Suggested change
This site config describes a minimal spack-stack install suitable for an Ubuntu 24.04 VM with Nvidia HPC SDK 25.1. For instructions on setting up the VM before building spack-stack, consult the ``README.md`` in the ``sites/tier2/ubuntu2404-nvhpc``.
This site config describes a minimal spack-stack install suitable for a Ubuntu 24.04 VM with Nvidia HPC SDK 25.1. For instructions on setting up the VM before building spack-stack, consult the ``README.md`` in the ``sites/tier2/ubuntu2404-nvhpc``.

I am never sure which is correct ...

@fmahebert fmahebert changed the base branch from develop to release/1.9.0 March 5, 2025 00:17
@l90lpa
Copy link
Copy Markdown

l90lpa commented Mar 5, 2025

@l90lpa does this ping work? For some reason I can't request you as reviewer here.

Hi @fmahebert, yes the ping worked! Thanks for looping me in, I'll try to test this tomorrow.

@fmahebert fmahebert force-pushed the nvhpc_as_tier2_site branch from 1829777 to 715109b Compare March 5, 2025 05:50
Copy link
Copy Markdown
Collaborator

@climbfuji climbfuji left a comment

Choose a reason for hiding this comment

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

Looks good now, thanks very much for fixing the branch issue.

@climbfuji climbfuji merged commit d118fcf into JCSDA:release/1.9.0 Mar 5, 2025
9 checks passed
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.

4 participants