With increasing demand in compute performance of HPC systems, accelerators are getting the main focus for application development. Many of the Top500 HPC systems now include accelerators, with the top 3 systems alone having accelerators of three different vendors. This diversity requires application developers to choose portable frameworks to support all at the same time, as developing applications via each native API is time consuming. One of the available frameworks is OpenMP with its offloading capability and availability for C, C++ and Fortran. With OpenMP offloading gaining more traction recently, performance analysis becomes important as well.
With this webinar, Jan André Reuter presented his first results in adding support for OpenMP offloading to the instrumentation and measurement infrastructure Score-P using the OpenMP Tools Interface. He demonstrated how he used both host side callbacks and the device tracing interface to build a measurement adapter capable of analyzing OpenMP applications effectively. He showed the current support landscape between different compilers and present first results for profiles and event traces based on the SPEC HPC 2021 618.tealeaf_s benchmark running on the LUMI HPC cluster at CSC in Finland.
Reference Material
About the Presenter
Jan André Reuter received a Bachelor degree in Scientific Programming,followed by a Master's in Applied Mathematics and Informatics at the FH Aachen in 2018 and 2020, respectively. Since 2015, he has worked at the Research Center Jülich, initially as a part of the vocational training for a mathematical technical software developer (MATSE) and since 2018 as a researcher at the Institute of Neuroscience and Medicine (INM-1). Here, he has participated in the Human Brain Project, developing applications and frameworks tailored towards HPC systems with a main focus on accelerators. In December 2022, he joined the Parallel Performance group at the Jülich Supercomputing Centre to improve support for OpenMP and accelerators in the measurement infrastructure Score-P. His research interests include compilers and their runtimes, lower-level languages like LLVM IR, accelerator frameworks like CUDA and the OpenMP programming model.