SC19 Proceedings

The International Conference for High Performance Computing, Networking, Storage, and Analysis

Achieving Performance on Current and Future Large-Scale Intel-Based Systems


Authors: Richard Gerber (National Energy Research Scientific Computing Center (NERSC), Lawrence Berkeley National Laboratory), Glenn Brook (University of Tennessee), David Martin (Argonne National Laboratory), Estela Suarez (Forschungszentrum Juelich)

Abstract: This BoF, organized by the Intel eXtreme Performance Computing Users Group (IXPUG), will focus on achieving performance at scale on current and future large Intel-based systems. The IXPUG community has in-depth knowledge and experience preparing codes for Xeon Phi and Xeon Scalable family of processors and coprocessors. With a planned 2021 exascale system that includes Intel GPUs, IXPUG’s optimization targets will be expanding to include the full line of Intel processors, including FPGAs. Through invited talks and open discussion, the BoF will provide a forum to share tips and techniques and look to the future.

Long Description: This BoF, organized by the Intel eXtreme Performance Computing Users Group (IXPUG), will focus on achieving performance at scale on current and future large Intel-based systems. The IXPUG community has in-depth knowledge and experience based on its success preparing codes for generations of Xeon Phi and Xeon Scalable family of processors and coprocessors. With the recent announcement of a planned 2021 exascale system that includes Intel GPUs, IXPUG’s optimization targets will be expanding to include the full line of Intel processors, including FPGAs. Through invited talks and open discussion, the BoF will provide a forum for researchers, tool developers, application programmers, HPC center staff and industry experts to share tips and techniques gained through deployments on large-scale systems and look to the future.

IXPUG is a worldwide community of hundreds of application users, software developers and HPC center staff that utilize Intel processors to solve some of the world’s most challenging problems on some of the world’s most powerful supercomputers. This BoF follows productive and well-attended IXPUG BoFs starting from SC14 and continuing through SC18. IXPUG has also hosted BoFs and workshops at ISC and HPC Asia. IXPUG will solicit proposals for 4-5 short talks on best practices and lessons learned from optimizing codes using multi-threaded and SIMD constructs and how those efforts will translate to achieving performance on future systems with GPUs, FPGAs, and more traditional - yet highly multithreaded - Xeon processors. The BoF will also include a brief discussion of the unique capabilities of Intel performance tools and the Intel OneAPI programing framework. A panel with open discussion will complete the session and allow attendees to interact with speakers and share their own experiences. The BoF will educate attendees about how to achieve performance at scale and will strengthen the IXPUG community by increasing knowledge sharing and establishing personal connections.


URL: https://www.ixpug.org/SC19-BOF


Back to Birds of a Feather Archive Listing