include documentation about numa_api #1028
Open
Conversation
This file contains hidden or 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
t-ramz
reviewed
Dec 12, 2025
| core affinity for PID 1132009: 41 42 43 44 45 46 47 | ||
| Suggested GPU for PID 1132009: 7 | ||
| In contrast, using ``srun --gpus-per-task=8 --gpu-bind=closest torchrun --nproc_per_node=8 --nnodes=1 --rdzv-id=$SLURM_JOBID --rdzv-backend=c10d --rdzv-endpoint=$MASTER_ADDR:3440`` results in the following output, i.e., here ``torchrun`` launches each task on the same core resulting in subpar performance. Note, the GPU suggestions printed by ``numa_api`` are based on the NUMA regions associated with the cores in :ref:`frontier-nodes`. |
Contributor
There was a problem hiding this comment.
This example is incredibly helpful in describing the dilemma.
t-ramz
reviewed
Dec 12, 2025
t-ramz
reviewed
Dec 12, 2025
t-ramz
reviewed
Dec 12, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
Distributed PyTorch launchers such as torchrun often lack the flexibility to map tasks to GPUs as per the NUMA domains of Frontier. Users have often reported subpar performance when using these launchers. Consequently, it is useful for the user to know if their distributed PyTorch program is making use of the node resources in the most optimal manner possible.
numa_apileverages the numactl library to identify which cores a process is bound to, as well as associated GPU for optimal binding. This can be useful because users typically set the GPU IDs manually in frameworks like PyTorch and TensorFlow.