Scheduling Overview¶
The BIH HPC uses the Slurm scheduling system for resource allocation.
This section of the manual attempts to give an overview of what scheduling is and how to use the Slurm scheduler.
For more detailed information, you will have to refer to the Slurm website and the Slurm man pages (e.g., by entering man sbatch
or man srun
on the HPC terminal's command line).
For a quick introduction and hands-on examples, please see the manual sections
- Overview, starting with Slurm Quickstart, and
- HPC Tutorial, starting with Episode 0.
Also, make sure that you are aware of our How-To: Debug Software and How-To: Debug Software on HPC Systems guides in the case that something goes wrong.
Annotated Contents¶
- Background on Scheduling -- some background on scheduling and the terminology used
- Quickstart -- explains the most important Slurm commands, with examples
- Cheat Sheet -- for quick reference
- Job Scripts -- how to setup job scripts with Slurm
- Memory Allocation -- memory allocation ( one of the most important concepts that is most often found confusing)
- Introduction to Slurm Commands
srun
-- running parallel jobs nowsbatch
-- submission of batch jobsscancel
-- stop/kill jobssinfo
-- display information about the Slurm clustersqueue
-- information about pending and running jbosscontrol
-- detailed information (and control)sacct
-- access Slurm accounting information (pending, running, and past jobs)- Format Strings in Slurm -- format strings allow to display extended information about Slurm scheduler objects
- Slurm and Snakemake -- how to use Snakemake with Slurm
- X11 Forwarding -- X11 forwarding in Slurm (simple; short)
- Rosetta Stone -- lookup table for SGE <-> Slurm
A Word on "Elsewhere"¶
Many other facilities run Slurm clusters and make their documentation available on the internet. We list some that we found useful below. However, be aware that Slurm is a highly configurable and extensible system. Other sites may have different configurations and plugins enabled than we have (or might even have written custom plugins that are not available at BIH). In any case, it's always useful to look "über den Tellerrand".
- Quick Start User Guide - the official guide from the Slurm creators.
- Slurm
man
Pages - web versions of Unix manual (man
) pages. - TU Dresden Slurm Compendium - nice documentation from the installation in Dresden. Note that their installation is highly customized, in particular, their partition selection is automated (but is not for us).
- Slurm at CECI - CECI is a HPC consortium from Belgium.
- Slurm at the Arctic University of Norway
- Slurm at Technical University of Denmark - if you want to get an insight in how this looks to administrator.