In our HPC Tips and Tricks sessions, we focus on several advanced topics in HPC.
Best practices: files and I/O
- Have you ever wondered why /scratch slows down sometimes?
- In this session we explore filesystem fundamentals and the workings of a distributed parallel filesystem like /scratch. Using real-life examples, you will learn how to optimize your jobs and data workflows to bypass common storage bottlenecks.
- The session is geared towards all existing users, and highly recommended for anyone running I/O-intensive workloads. While focused on CalcUA clusters, concepts apply broadly across HPC environments.
- Next session: October 7, 13.00-15.00 h (CMI G.005)