Set local_devices from CUDA_VISIBLE_DEVICES in initialize_jax_for_gpu #2945
+9
−0
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Outdated suggestions cannot be applied.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
When running in Slurm, jax.init_distributed.initialize only attaches to one GPU unless local_device_ids is passed.
This PR uses the standard CUDA_VISIBLE_DEVICES environment variable to set local_device_ids,
falling back to the default behavior if CUDA_VISIBLE_DEVICES is None or cannot be parsed.
We have been using this internally at AMD for some time.
This change will not affect TPUs in any way.
FIXES: #865
Tests
Run a job in Slurm on a machine with multiple GPUs with and without CUDA_VISIBLE_DEVICES set
Checklist
Before submitting this PR, please make sure (put X in square brackets):
gemini-reviewlabel.