An ALCF, NERSC, OLCF joint workshop held January 27-29 at ORNL focused on architecture and performance portability as a critical objective for the application readiness programs at the three facilities. National Laboratories’ attendees at the meeting were (in alphabetical order): Katie Antypas, NERSC; Yuri Alexeev, ALCF; Valentine Anantharaj, OLCF; Ashley Barker, OLCF; Anouar Benali, ALCF; David Bernholdt, ORNL; Mark Berrill, OLCF; Swen Boehm, OLCF; Michael Brim, OLCF; Susan Coghlan, ALCF; Matthew Cordery, NERSC; Ed D’Azevedo, OLCF; Markus Eisenbach, OLCF; Hal Finkel, ALCF; Graham Fletcher, ALCF; Fernanda Foertter, OLCF; Chris Fuson, OLCF; Scott Futral, LLNL; Richard Gerber, NERSC; Benjamin Hernandez, OLCF; Oscar Hernandez, OLCF; Judy Hill, OLCF; Antti-Pekka Hynninen, OLCF; Wayne Joubert, OLCF; Supada Laosooksathit, OLCF; Dmitry Liakh, OLCF; Ray Loy, ALCF; Mike Matheson, ORNL; Bronson Messer, OLCF; Vitali Morozov, ALCF; Rob Neely, LLNL; Matthew Norman, OLCF; James Osborn, ALCF; Suzanne Parete-Koon, OLCF; Adrian Pope, ALCF; Bill Renaud, OLCF; Jon Rood, NERSC; Ramanan Sankaran, OLCF; Adam Simpson, OLCF; Tjerk Straatsma, OLCF; Arnold Tharrington, OLCF; Veronica Vergara Larrea, OLCF; Ying Wai Li, OLCF; Tim Williams, ALCF; Woo-Sun Yang, NERSC.
HPC vendors share details on next-generation systems
Team members of the Oak Ridge Leadership Computing Facility (OLCF), a US Department of Energy (DOE) Office of Science User Facility, hosted staff from Argonne Leadership Computing Facility (ALCF) and National Energy Research Scientific Computing Center (NERSC) for a meeting with high-performance computing (HPC) vendors late last month to discuss strategies for application readiness and performance portability in preparation for the next generation of supercomputers.
More than 50 people attended the event at DOE’s Oak Ridge National Laboratory (ORNL), including representatives from Lawrence Livermore National Laboratory and Sandia National Laboratories.
The three-day meeting, a continuation of discussions that took place at NERSC in September 2014, focused on enabling performance portability across computing architectures and using current software tools to prepare for the next supercomputers. Vendors shared details on hardware and software development tools for the next-generation systems during the meeting’s first two days. The third day featured an open discussion on how to prepare for upcoming application readiness activities.
The discussions will help inform efforts at each institution to train and assist users with code preparation for systems being deployed in the next 2–4 years. DOE’s Office of Advanced Scientific Computing Research (ASCR)—which manages the OLCF, ALCF, and NERSC—considers architecture portability and performance portability to be critical components of application readiness to better accommodate users who run their applications at multiple computing facilities.
“Addressing the performance portability that our users expect for the applications that they will run on the next generation systems at the ALCF, NERSC, and OLCF requires a detailed understanding of the roadmap for the hardware and especially the software development environments,” said Tjerk Straatsma, ORNL Scientific Computing group leader. “This workshop provided some of the details necessary to start the process of successfully preparing our applications to be ready for the future systems of the ASCR facilities.”
Application readiness programs at ASCR facilities include the OLCF’s Center for Accelerated Application Readiness, NERSC’s Exascale Science Applications Program, and the ALCF’s Early Science Program. The next planned application readiness workshops will cover team training and best practices.—Jonathan Hines
Oak Ridge National Laboratory is supported by the US Department of Energy’s Office of Science. The single largest supporter of basic research in the physical sciences in the United States, the Office of Science is working to address some of the most pressing challenges of our time. For more information, please visit science.energy.gov.
