Fluent example script for SLURM
Below a simple script for running Fluent v16.2 in parallel job using 4 nodes each 36 cores:
#!/bin/bash
#SBATCH --partition cpus
#SBATCH --job-name=EXAMPLE_JOB
#SBATCH --output EXAMPLE_JOB.output
#SBATCH -N 4
#SBATCH --ntasks-per-node 36
#SBATCH --time=60:00
# N=4*--ntasks-per-node=36=144
TASKS=144
OPTIONS=" -cflush -nmon -mpi=openmpi"
JOURNAL=example_journal_file.jou
# Loading MPI variables
module load hyp/mpi/intel/4.1.3
# Loading Fluent 16.2 variables
module load hyp/commercial/fluent/16.2
# DON'T CHANGE BELOW THIS LINE
export FLUENT_ENABLE_SLURM_SUPPORT=1
scontrol show hostname $SLURM_JOB_NODELIST > nodes_list
fluent 3ddp -g -$OPTIONS -t$TASKS -cnf=nodes_list -i$JOURNAL -ssh > fluent.log
Fluent version | corresponding MPI version |
16.2 | Intel 4.1.3 |
18.2 | Intel 5.1.3.223 |
2023R2 | Intel 2018.3.222; Intel 2021.8.0; OpenMPI 4.0.5 |