Add gpu tag for ROCm-LLVM and refactor method#228
Conversation
|
What I'm trying to do in #231 is also relevant here |
There was a problem hiding this comment.
I think this PR is in decent shape.
The find_rocm_llvm_dependency function could use a few more comments to make it easier to understand what's going on. I've made some suggestion.
In addition, I agree with Alan's suggestion to store the ROCm version, not the full ROCm-LLVM version+versionsuffix. That'll save us some parsing later on, and it's unlikely we'll be interested in the LLVM version anyway. That will also require an update of the test-eb-hooks.py, see my suggestions.
If you can apply those changes, I think this is good to go. Note that this is blocking progress on #231 , which in turn is blocking progress on our 2026.06 compatibility layer. So I'd like to get this one merged ASAP :)
ad72c73 to
07e4f9f
Compare
Improve search for ROCm-LLVM dep and add tests Fix tuple index access Apply suggestions from code review Co-authored-by: ocaisa <alan.ocais@cecam.org> Co-authored-by: Caspar van Leeuwen <33718780+casparvl@users.noreply.github.com> Get the error in logs Fix numbered indexing to keys Normalise return object for direct dependency of ROCm-LLVM Add more logs Add logs for rompi dry run Simplify check for rocm-compilers based toolchains add logs again, failed the test Add ignore cleanup flag add error log run Take out rompi test until it's available in EESSI
|
bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws for:arch=x86_64/amd/zen2 |
|
New job on instance
|
|
New job on instance
|
|
bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws for:arch=x86_64/amd/zen2 |
|
New job on instance
|
|
New job on instance
|
Easyconfigs which include ROCm-LLVM as a dependency should also be tagged with
gpu. It doesn't need to be dropped to a build dependency as it is redistributable (unlike CUDA).The loop seems to iterate over dependencies twice, it can be simplified to one pass.