ABI Knowledge Base

You've loaded an old revision of the document! If you save it, you will create a new version with this data.

====== Using SLURM to run jobs ======

1. Create an example script named slurm_example.sh:

<code>
#!/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

</code>

2. Submit the job:

<code>
sbatch slurm_example.sh
</code>

3. Check the job status:

<code>
squeue 
</code>

4. Cancel your pending or running job:

<code>
scancel 3564
</code>

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

====== Tips & Recommendations ======

===== Use one script per task =====

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

Edit summary