LUMI Profiling and Optimization Workshop, May 11-13 in Stockholm

This advanced three-day on-site workshop offers an in-depth exploration of performance profiling and optimisation on the LUMI supercomputer. Designed for experienced HPC users, the course combines theoretical insights with hands-on exercises to enhance efficiency on both CPU and GPU partitions. Topics include deep dives into LUMI architecture, compiler optimisation and performance analysis tools, Cray MPI environment tuning, I/O optimisation strategies, and specialised profiling tools for AMD GPUs. Participants are expected to have solid HPC knowledge, familiarity with Linux and programming in C/Fortran/Python, and an active LUMI project. Attendance is free of charge.