LUMI Profiling and Optimization Workshop

An advanced, hands-on workshop for experienced HPC users who want to get more performance out of LUMI. Experts from HPE and AMD go through the tools and techniques for profiling and optimising CPU and GPU workloads. The workshop runs 09:00–17:30 on Wednesday and Thursday, and 09:00–12:30 on Friday.

Topics

• LUMI’s architecture: system design and components
• Compiler optimisation, performance analysis and profiling
• Cray MPI on Slingshot, GPU support and rank reordering
• I/O optimisation: handling and storing data
• HPE’s profiling tools, and AMD’s tools for GPU workloads

Who should attend

Experienced HPC users with a good knowledge of LUMI – for example from the LUMI introductory course – and some active experience on it. You should know the Linux command line and C, Fortran or Python, and have an active LUMI project or an application in progress.

The introductory course Moving your HPC workloads to LUMI runs on 9–10 November at the same place.

All events