A Beginner's guide to using Slurm on Oscar

This workshop is designed to be run on Oscar, Brown's computing cluster. The workshop is aimed at people getting started with Slurm on Oscar. We assume you have some knowledge of the linux command line.

The workshop presentation slides can be viewed at

google slide deck

The exercises can be found at

github.com/brown-ccv/beginner-slurm

You can attend this workshop in person. To find upcoming CCV events visit:

events.brown.edu/ccv

You are also welcome follow the slides and do the exercises on your own. If you find any problems with the exercises, please create an issue on this repository or email [email protected].

Important concept: Login nodes vs. compute nodes

Login nodes are for editing files, compiling, moving files, changing permission, submitting jobs. Compute nodes are where you should do your intensive computing. To use the compute nodes you are going to use Slurm. Do not run Matlab on the login nodes!

Exercises

To copy the exercise materials:

1. Login into Oscar
2. Clone this repostiory into your home directory
3. cd into the beginner-slurm directory

git clone https://github.com/brown-ccv/beginner-slurm
cd beginner-slurm

Example Slurm batch scripts for future reference

Example batch scripts can be found in your home directory on Oscar. To list all the example scripts do:

ls ~/batch_scripts

You can use these scripts as a reference. You can copy these scripts and edit to fit your needs, for example to copy the serial.sh

cp ~/batch_scripts/serial.sh .

Checkout other workshops offered by the CCV HPC team.

To sign up for live events, go to the CCV Events Calendar