Skip to content

Add --mpi=pmix to COMMAND srun in Ray Slurm template #439

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
athitten wants to merge 2 commits into from
Closed

Add --mpi=pmix to COMMAND srun in Ray Slurm template #439

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
786bfc5
Add --mpi=pmix to COMMAND srun in ray template
athitten Feb 27, 2026
b38e09c
Update test artifacts with --mpi=pmix
athitten Mar 2, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion nemo_run/run/ray/templates/ray.sub.j2
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -454,7 +454,7 @@ COMMAND="${COMMAND:-{{ command | default('', true) }}}"
COMMAND_WORKDIR={{ command_workdir | default('$CONTAINER_CWD') }}

if [[ -n "$COMMAND" ]]; then
srun {% if heterogeneous %}--het-group=0 {% endif %}--no-container-mount-home --gpus=0 --overlap --container-name=ray-head --container-workdir=$COMMAND_WORKDIR --nodes=1 --ntasks=1 -w "$head_node" -o $LOG_DIR/{{ ray_log_prefix }}job.log bash -c "$COMMAND"
srun {% if heterogeneous %}--het-group=0 {% endif %}--no-container-mount-home --mpi=pmix --gpus=0 --overlap --container-name=ray-head --container-workdir=$COMMAND_WORKDIR --nodes=1 --ntasks=1 -w "$head_node" -o $LOG_DIR/{{ ray_log_prefix }}job.log bash -c "$COMMAND"
else
echo "[INFO]: Ray Cluster is idled, run this on the slurm head node to get a shell to the head node:"
cat <<EOF >$CLUSTER_DIR/scripts/${SLURM_JOB_ID}-attach.sh
Expand Down
4 changes: 2 additions & 2 deletions test/core/execution/artifacts/expected_ray_cluster.sub
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -423,7 +423,7 @@ COMMAND="${COMMAND:-python train.py}"
COMMAND_WORKDIR=/workspace

if [[ -n "$COMMAND" ]]; then
srun --no-container-mount-home --gpus=0 --overlap --container-name=ray-head --container-workdir=$COMMAND_WORKDIR --nodes=1 --ntasks=1 -w "$head_node" -o $LOG_DIR/ray-job.log bash -c "$COMMAND"
srun --no-container-mount-home --mpi=pmix --gpus=0 --overlap --container-name=ray-head --container-workdir=$COMMAND_WORKDIR --nodes=1 --ntasks=1 -w "$head_node" -o $LOG_DIR/ray-job.log bash -c "$COMMAND"
else
echo "[INFO]: Ray Cluster is idled, run this on the slurm head node to get a shell to the head node:"
cat <<EOF >$CLUSTER_DIR/scripts/${SLURM_JOB_ID}-attach.sh
Expand All @@ -440,4 +440,4 @@ EOF
chmod +x $CLUSTER_DIR/scripts/${SLURM_JOB_ID}-attach.sh
echo " bash $CLUSTER_DIR/scripts/${SLURM_JOB_ID}-attach.sh"
sleep infinity
fi
fi
4 changes: 2 additions & 2 deletions test/core/execution/artifacts/expected_ray_cluster_ssh.sub
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -430,7 +430,7 @@ COMMAND="${COMMAND:-ray job submit --address ray://localhost:10001 --job-id trai
COMMAND_WORKDIR=/workspace/training

if [[ -n "$COMMAND" ]]; then
srun --no-container-mount-home --gpus=0 --overlap --container-name=ray-head --container-workdir=$COMMAND_WORKDIR --nodes=1 --ntasks=1 -w "$head_node" -o $LOG_DIR/ray-job.log bash -c "$COMMAND"
srun --no-container-mount-home --mpi=pmix --gpus=0 --overlap --container-name=ray-head --container-workdir=$COMMAND_WORKDIR --nodes=1 --ntasks=1 -w "$head_node" -o $LOG_DIR/ray-job.log bash -c "$COMMAND"
else
echo "[INFO]: Ray Cluster is idled, run this on the slurm head node to get a shell to the head node:"
cat <<EOF >$CLUSTER_DIR/scripts/${SLURM_JOB_ID}-attach.sh
Expand All @@ -447,4 +447,4 @@ EOF
chmod +x $CLUSTER_DIR/scripts/${SLURM_JOB_ID}-attach.sh
echo " bash $CLUSTER_DIR/scripts/${SLURM_JOB_ID}-attach.sh"
sleep infinity
fi
fi
4 changes: 2 additions & 2 deletions test/core/execution/artifacts/expected_ray_het_cluster.sub
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -463,7 +463,7 @@ COMMAND="${COMMAND:-}"
COMMAND_WORKDIR=None

if [[ -n "$COMMAND" ]]; then
srun --het-group=0 --no-container-mount-home --gpus=0 --overlap --container-name=ray-head --container-workdir=$COMMAND_WORKDIR --nodes=1 --ntasks=1 -w "$head_node" -o $LOG_DIR/ray-job.log bash -c "$COMMAND"
srun --het-group=0 --no-container-mount-home --mpi=pmix --gpus=0 --overlap --container-name=ray-head --container-workdir=$COMMAND_WORKDIR --nodes=1 --ntasks=1 -w "$head_node" -o $LOG_DIR/ray-job.log bash -c "$COMMAND"
else
echo "[INFO]: Ray Cluster is idled, run this on the slurm head node to get a shell to the head node:"
cat <<EOF >$CLUSTER_DIR/scripts/${SLURM_JOB_ID}-attach.sh
Expand All @@ -480,4 +480,4 @@ EOF
chmod +x $CLUSTER_DIR/scripts/${SLURM_JOB_ID}-attach.sh
echo " bash $CLUSTER_DIR/scripts/${SLURM_JOB_ID}-attach.sh"
sleep infinity
fi
fi
Loading