Large Slurm Job

This is an example of a very large Slurm job which is still running on a single node. A single-node Slurm job can use up to 256 cores and 1.5TB of RAM.

Uses the HPC Project group myhpcproject; change to use your real HPC Project name
Submitted to the free highmem_free partition (–partition=highmem_free)
Requests 256 CPU cores (–cpus-per-task=256); this the largest single node job possible
Requests 512GB of RAM (–mem=512G)
Requests up to 10 hours of runtime (–time=10:00:00), for a maximum possible total of 2560 Compute Hours (cpus_per_task * time_in_hours)
Prints the time the job started to the log file
Prints the name of the compute node(s) it will run on
Prints the time the job finished to the log file

#!/bin/bash

#SBATCH --account=myhpcproject
#SBATCH --partition=highmem_free
#SBATCH --cpus-per-task=256
#SBATCH --mem=512G
#SBATCH --time=10:00:00

# Log when we started
echo "Job started at: `date`"

# Show which node(s) we are running on
HOSTNAMES=`scontrol show hostnames $SLURM_JOB_NODELIST`
echo "Job is running on: $HOSTNAMES"

# Add any 'module load' commands here

# Add your custom commands you want to run here

# Log when we finished
echo "Job finished at: `date`"

Jobs submitted to the free highmem_free partition may need to wait for longer until sufficient resources are free. If your HPC Project has funds available, consider applying for Funded access and then submitting to the highmem_paid partition instead, where your job has an opportunity to be scheduled on a much larger set of available compute nodes than are available to free users.

Back to Getting Started