Introduction to Linux and Cluster Computing

  • Duration: 6 hours
  • Taught by: Carlos Teijeiro Barjas¬†(SURFsara)

Compute clusters are useful when you have to run many calculations or analyses, or when the calculations are too big or would take too long on your own system. Knowing how to effectively use compute clusters will help you do your research by running more calculations in less time. This goal of this course is to have you run your first jobs on a supercomputer, with no previous Linux or supercomputing experience. The examples given are simple enough to be understood fully, yet can be modified easily to be directly applied your research.

The first part of the course focuses on making you ‘feel at home’ with the Linux system, present in many of the current compute clusters and the largest supercomputers in the world. You will learn:

* how find your way around a Linux system
* how to run commands
* how to exchange files between your own desktop computer and a Linux cluster
* manipulate files, and organize your files in directories
* how to edit text files using a basic text editor.

In the second part, we will introduce the concept of a “compute job”, and you will run your first jobs on a cluster. We will discuss:

* what compute clusters are
* how to run simple jobs
* how to run multiple jobs
* best practices

You’ll learn how to work with the Cartesius supercomputer and the national compute cluster Lisa. We explain when you need the supercomputer and what applications feature on the Lisa cluster. This hands-on course includes both an interactive part on the login nodes and a batch part on the worker nodes. You will use capability and capacity cluster computers.

Prerequisite courses or equivalent knowledge: none