Serpent

Here are some simple instructions to get your first test case running.

First connect to cluster and make sure you are in your home directory (optional, you should be there):

cd ~

Create case directory (here named BWR, could be anything) under your home directory and change to that directory:

mkdir BWR
cd BWR

Copy job file template and BWR example to that newly created directory:

cp /shared/tutorials/serpent/template.sh ~/BWR/BWR_2D.sh
cp /shared/tutorials/serpent/BWR_2D.sss ~/BWR

Modify your job file BWR_2D.sh next (example below). Follow instructions inside job script. Script changes to submit directory when executed. When ready hit Ctrl+X and choose to save and overwrite.

nano BWR_2D.sh

Then send job to queue. If you get errors just check std.err and std.out files. Correct things which lead to errors and requeue.

sbatch BWR_2D.sh

Example of the runscript needed for Slurm.

#!/bin/bash -l
###
### BWR_2D example case
###
 
## name of your job
#SBATCH -J BWR_2D
 
## system error message output file
#SBATCH -e BWR_2D.std.err%j
 
## system message output file
#SBATCH -o BWR_2D.std.out%j
 
## send mail after job is finished
#SBATCH --mail-type=end
#SBATCH --mail-user=<username>@lut.fi
 
## a per-process (soft) memory limit
## limit is specified in MB
## example: 1 GB is 1000
#SBATCH --mem-per-cpu=900
 
## how long a job takes, wallclock time d-hh:mm:ss
#SBATCH -t 01:00:00
 
## the number of processes (tasks)
## number of nodes
#SBATCH -N 1
 
## number of tasks (amount of MPI processes)
#SBATCH --ntasks-per-node=2
 
## maximum number of tasks per socket (this should equal to 1)
#SBATCH --ntasks-per-socket=1
 
## threads per process (task)
## (amount of OpenMP threads per MPI process)
#SBATCH --cpus-per-task=8
 
## name of queue
#SBATCH -p phase2
 
## load modules
module load serpent/openmpi-4.0/2.1.30
 
## change directory
cd $SLURM_SUBMIT_DIR
 
## run my MPI executable
srun sss2 -omp $SLURM_CPUS_PER_TASK BWR_2D.sss