Add MI300X Runners to AMD workflow. #149
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ngc92
reviewed
Jan 22, 2025
| @@ -13,10 +13,18 @@ on: | |||
|
|
|||
| jobs: | |||
| run: | |||
| runs-on: [amdgpu-mi250-x86-64] | |||
| runs-on: ${{ matrix.runs-on }} | |||
| strategy: | |||
There was a problem hiding this comment.
I don't think this is the way we want to handle multiple runners; afaics, this well always run the given code sample on both types of AMD cards. Instead, the card should be selectable by an input argument that is given to the job.
Both runners should be registered in src/discord-cluster-manager/consts.py, and
https://github.com/gpu-mode/discord-cluster-manager/blob/main/src/discord-cluster-manager/cogs/github_cog.py#L35-L40
needs to be updated
| python3 .github/workflows/runner.py | ||
| cat result.json # Debug: show output | ||
|
|
||
| - name: Upload training artifacts | ||
| uses: actions/upload-artifact@v4 | ||
| if: always() | ||
| with: | ||
| name: run-result | ||
| name: run-result-${{ matrix.name }} |
There was a problem hiding this comment.
this name needs to be unchanged so the bot knows how to read the results.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
This commit adds mi300 nodes to the current amd workflow matrix.
What dependencies do we expect to be installed on the runners?
Right now, it is a fresh ubuntu + rocm + python image on the runner.
From the current setup, it looks like we expect dependencies like
torchandpytorch-triton-rocmto be requested by the user.Let me know if that is not the case.
I tested a couple times on the AMD workflow here (https://github.com/gpu-mode/discord-cluster-manager/actions/runs/12902691574)
Checklist
Before submitting this PR, ensure the following steps have been completed:
/verifyrunson your own server./verifyruns.runs may take a little longer. The Modal run is typically quick.)
For more information on running a cluster bot on your own server, see
README.md.