1. Create an example script named slurm_example.sh:

#!/bin/bash  				
#SBATCH --mem 10gb			# the amount of memory you request for your job
#SBATCH --cpus-per-task=4               # Number of CPU cores per task
#SBATCH --output=slurm-%j.log		# file name where the job console output will be stored
#SBATCH --mail-type=BEGIN,END		# if specified you will receive email when your job starts, ends	
#SBATCH --mail-user=<your_mail>		# your email address

for i in {1..100..1}				your script
do
    echo "This is my first script with slurm :)"
done

2. Submit the job:

sbatch slurm_example.sh

3. Check the job status:

squeue 

4. Cancel your pending or running job:

scancel 3564

For more detailed information see the https://slurm.schedmd.com/documentation.html.

It is recommended to use one script per task. If you happen to run the same script with different parameters, have a separate script for each run and name them accordingly.

Suppose you have a script that performs alignment named: align.sh. And you want to run the alignment in slow and fast modes. Instead of changing the modes inside the script, you can have two scripts

align_slow.sh align_fast.sh

In this way, you can also separate the log files: align_slow.log align_fast.log

This will help you to trace your jobs back easily!

In this way, you will

TBA (please contribute)