Overview

The clusters Puma and Ocelote both have high memory nodes available that offer more memory per CPU than standard nodes. For more information, see our system documentation.

Submission Script

#!/bin/bash
#SBATCH --job-name=High-Memory-Node-Job
#SBATCH --ntasks=1
#SBATCH --nodes=1 
#SBATCH --constraint=hi_mem
#SBATCH --time=00:00:30   
#SBATCH --partition=standard
#SBATCH --account=YOUR_GROUP

echo "Running job on a high memory node!"
echo "Hostname: $(hostname)"
echo "Memory per CPU: $SLURM_MEM_PER_CPU"

Job Submission

[netid@wentletrap ~]$ sbatch High-Memory-Node.slurm 
Submitted batch job 2106313

Output

[netid@wentletrap ~]$ cat slurm-2106313.out 
Running job on a high memory node!
Hostname: r5u19n1.puma.hpc.arizona.edu
Memory per CPU: 32768

---