Skip to content

Using the SC cluster with Python and R

This guide shows you how to run your Python and R code on the cluster of Scientific Computing (SC) at Leipzig University. The cluster helps with analyses that are too large or too slow for your own computer. You need no experience with clusters to begin, only an account at the university and some Python or R.

Four steps through this guide

  1. Set up your computer

    Get access to the cluster and connect to it for the first time.

    Time: 60 min

    Set up your computer

  2. Learn the basics

    How work on a cluster goes, how a job script tells the cluster what to run, and the tools for each step: the terminal, uv or renv for your packages, and Git.

    Time: 60 min

    Basics · Slurm · Your project on the cluster

  3. Build your first project

    Build a small research project from an empty folder, test it on your computer and run it on the cluster.

    Time: 45 min

    Code-alongs

  4. Bring your own project

    Prepare a project that already runs on your computer for the cluster.

    Bring your own project

What sc-starter does

sc-starter is a small program that runs on your computer. You type its commands in a terminal opened in the folder of your project. sc-starter then takes care of the steps on the cluster for you. It copies your project there, asks the cluster to run it and brings the results back to the same folder:

Your computer
sc-starter submit    # copy the project to the cluster and start it
sc-starter status    # see how each run is doing
sc-starter pull      # bring the results back

Your scripts stay ordinary Python or R scripts. A short file in the project, the job script, tells the cluster how much time, memory and computing power each run needs. Working with sc-starter explains each step, and the command reference lists every option.

Code-alongs from the social sciences

Seven small research projects, in Python and R. In each one you build the project yourself, from an empty folder: you write the code, test it on your computer, write the job scripts and run it on the cluster. They cover many runs side by side, R packages, GPUs and runs longer than the time limit.

See all code-alongs

Help

  • Something went wrong

    The messages you may see, what they mean and what to do. Troubleshooting

  • A word is unclear

    The technical terms used in this guide, explained. Glossary

  • Your account or the cluster itself

    sc-starter is not run by the SC team. For questions about your account or the cluster, contact SC support.