OpenMP GPU Programming with AMD
The High-Performance Computing Center Stuttgart (HLRS), in cooperation with AMD and the Max Planck Computing and Data Facility, will organize an online training course, Introduction to OpenMP Offloading with AMD GPUs on HLRS's Hunter and Herder systems. The course will be held on 27–28 October 2026.
This training will introduce participants to AMD Instinct GPUs and Accelerated Processing Units (APUs), as well as OpenMP offloading techniques for accelerating computation on heterogeneous high-performance computing (HPC) systems. It will cover unified shared memory, explicit data management, asynchronous offloading, and tools for performance analysis and code optimization.
Designed for developers with basic OpenMP experience, the training combines lectures with practical exercises. Participants will learn to port applications from CPU-only systems to platforms with GPU accelerators and APUs, with hands-on exercises focused on real-world porting and optimization.