NHR PerfLab Seminar: OpenMP Offloading

Symbolic picture for the article. The link opens the image in a large view.

Speaker: Dr. Christian Terboven (RWTH Aachen IT Center)

Title: OpenMP Offloading

Date and time: Tuesday, April 5, 2:00 p.m. – 3:00 p.m. CEST

Abstract:

OpenMP is a popular, portable, widely supported, and easy-to-use shared-memory model. Since version 4.0 released in 2013, OpenMP offers support for heterogeneous parallel programming, including offloading to GPUs. Contrary to other models, the OpenMP approach reuses existing Host-oriented functionality to express parallelism for accelerators.

This seminar talk will examine the OpenMP Offloading model and discuss it’s relation to Host-oriented parallel programming. It will briefly present the additions to that model made with OpenMP 4.5 and 5.x and comment on future developments. It will connect to the previous NHR Perflab seminar on OpenMP by showing that asynchronous kernel offloads actually are OpenMP tasks.

Short Bio:
Dr. Christian Terboven is a senior scientist and leads the HPC group at RWTH Aachen University. His research interests center around Parallel Programming and related Software Engineering aspects. Dr. Terboven has been involved in the Analysis, Tuning and Parallelization of several large-scale simulation codes for various architectures. He is responsible for several research projects in the area of programming models and approaches to improve the productivity and efficiency of modern HPC systems.