SLURM

What is SLURM?

The Simple Linux Utility for Resource Management (SLURM) is a cluster management and job scheduling system. We use SLURM to allocate access to resources (i.e. compute nodes), prevent resource conflicts, and ensure fair and equitable access to University resources. SLURM also provides a means of coordinating workflows, establishing job dependencies, and queueing jobs when the system is fully utilized.

Where can I learn more about SLURM?

The definitive resource for SLURM documentation is published by SchedMD. Full documentation, quick start guides, and tutorials for the version of SLURM we run on Talapas can be found at https://slurm.schedmd.com/archive/slurm-16.05.7/. For user's who are familiar with other job schedulers (i.e. PBS Torque) and simply wish to "translate" their job scripts, SchedMD provides a handy Rosetta Stone of Workload Managers. Lastly, we have several how-to guides on this site covering the job submission process. Just search for the keyword "SLURM".

Â