Running Jobs on Deep Blizzard

Deep Blizzard uses the SLURM workload manager to schedule jobs on compute nodes. To run a job, you must create a SLURM submit script that defines the resources your job requires (CPUs, memory, GPUs, walltime, etc.).
Slurm User Documentation: https://slurm.schedmd.com/quickstart.html

Quick Start (Recommended Method)

Research Computing maintains a set of tested example job scripts that reflect the current cluster configuration.
To begin:

cp /mnt/it_research/examples/deepblizzard/v2/* . ls

Submit a simple test job:

sbatch 00_hello_world_cpu.slurm

When the job completes, view the output:

cat slurm-[jobid].out

These example scripts are the authoritative reference for job submission on Deep Blizzard.

Choosing the Correct Partition and Account

Deep Blizzard uses different SLURM partitions and accounts depending on your research group.

cd $SLURM_SUBMIT_DIR echo "Hello World! I ran on compute node $(/bin/hostname -s)"

The final section appends the job's completion time to the output file:

echo "" echo "======================================================" echo "End Time : $(date)" echo "======================================================"

To run this first job. Login to the login node, create a file called basic-submit.slurm and past the above script into the file.

Then submit the script using the slurm scheduler:

$ sbatch basic-submit.slurm Submitted batch job 18513

Once your job completes you should see the following(or similar since the actual compute node could be different) when you cat your output file:

$ cat slurm-18513.out ====================================================== Start Time : Mon Sep 29 09:48:26 EDT 2025 Submit Dir : /home/jlkarau Job ID/Name : 18513 / basic_slurm_job Num Tasks : total [1 nodes @ 1 CPUs/node] ====================================================== Hello World! I ran on compute node compute-0-7 ====================================================== End Time : Mon Sep 29 09:48:26 EDT 2025 ======================================================

Additional sample batch files

There are additional sample batch files on the system that you can copy into your home directory:

$ cp /mnt/it_research/examples/bioinformatics/* . $ cp /mnt/it_research/examples/mri/* . $ sbatch bio_cpu1_test Submitted batch job 18514 cat slurm-18514.out

Additional scripts you copied to your home directory:

$ ls bio* bio_gpu1_test bio_gpu4_test bio_jobs_test bio_largemem_test $ ls mri_* mri_gpu1_test mri_gpu4_test mri_jobs_test mri_largemem_test

Since you have copied these to your home directory you can edit and modify them as needed or use them as the start of your own batch file.