Example jobs

Overview

These examples show different job types on SeaWulf. Each explains the resource choices and when to use that configuration.

1 · Serial job (hello world)

#!/bin/bash
#SBATCH -p short-40core-shared
#SBATCH -N 1
#SBATCH -n 1
#SBATCH --mem 5G
#SBATCH -t 00:05:00
#SBATCH -o hello.out

echo "Hello from SeaWulf!"

One task on a shared partition with modest memory. Suits quick tests or small scripts — a serial program needs minimal resources.

2 · Python job

#!/bin/bash
#SBATCH -p short-40core
#SBATCH -N 1
#SBATCH -n 40
#SBATCH -t 00:10:00
#SBATCH -o python.out

module load anaconda
conda activate my-environment
python script.py

Python scripts are typically single-threaded unless using multi-threaded libraries (NumPy, PyTorch, Dask). If your script is single-threaded, run on one core in a shared partition instead — requesting all 40 cores wastes resources.

3 · MPI job

#!/bin/bash
#SBATCH -p short-40core
#SBATCH -N 2
#SBATCH -n 80
#SBATCH -t 00:10:00
#SBATCH -o mpi.out

module load openmpi
mpirun ./my_mpi_program

Two nodes × 40 tasks fully utilize all cores across both nodes — each MPI rank handles part of the workload.

4 · OpenMP job

#!/bin/bash
#SBATCH -p short-40core
#SBATCH -N 1
#SBATCH -n 1
#SBATCH -c 40
#SBATCH -t 00:10:00
#SBATCH -o openmp.out

export OMP_NUM_THREADS=40
./my_openmp_program

OpenMP uses threads, not separate processes: one task (-n 1) with 40 cores (-c 40), and OMP_NUM_THREADS=40 so the program uses them all.

5 · GPU job

#!/bin/bash
#SBATCH -p a100
#SBATCH -N 1
#SBATCH -n 1
#SBATCH --mem 50G
#SBATCH --gres=gpu:1
#SBATCH -t 00:30:00
#SBATCH -o gpu.out

module load cuda120/toolkit/12.0
./my_gpu_program

Request a GPU with --gres=gpu:1. One CPU task suffices to launch, though you can request more threads if the program also uses CPU cores.

6 · Array job

#!/bin/bash
#SBATCH --job-name=parameter_sweep
#SBATCH --array=1-100
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=40
#SBATCH --time=02:00:00
#SBATCH -p short-40core
#SBATCH -o array_%A_%a.out

module load python/3.9

python simulation.py --param-set $SLURM_ARRAY_TASK_ID

--array=1-100 creates 100 jobs, each receiving a unique $SLURM_ARRAY_TASK_ID — ideal for parameter sweeps, Monte Carlo simulations, or processing many input files. %A is the array job ID, %a the task ID.

Quick reference