How to run a supercompter
Scheduled time: Sunday, 16:00 - Room V1 - Duration 30 Min.
A brief overview of the infrastructure used to run a supercomputer with over a thousand nodes in a heterogeneous cluster at NHR@ZIB. This talk explores the core stack in production: from Warewulf, for node provisioning, Slurm, for distributed jobs, and the tiered storage systems.
Desired previous knowledge: Basic knowledge of *nix systems and Networking.
Further information: Short paper
