Virtual ALCF workshop helps participants boost supercomputer performance

outreach
2020 ALCF Computational Performance Workshop

Argonne computational scientist Colleen Bertoni provides an overview of the OpenMP programming model at the virtual ALCF Computational Performance Workshop. (Image: Argonne National Laboratory)

The 2020 ALCF Computational Performance Workshop went virtual this year, providing an interactive training event to help researchers enhance code performance on leadership-class supercomputers.

When trying to harness the power of leadership-class supercomputers to accelerate scientific research, hands-on assistance is sometimes the best approach to getting started — even in the midst of a pandemic.

Earlier this year, the Argonne Leadership Computing Facility (ALCF) held its annual Computational Performance Workshop, albeit with one major difference compared to preceding years: 2020’s workshop was entirely virtual. The ALCF — part of the U.S. Department of Energy’s (DOE) Argonne National Laboratory — is a DOE Office of Science User Facility.

During three days in May, over 130 attendees — more than twice that of the 2019 workshop — interacted with 47 ALCF staff members and speakers to prepare code for the extreme scale and unique architectures that characterize leadership-class computing systems. The annual event is designed to help attendees boost performance and prepare for future ALCF projects.

Our goal is always to connect attendees with the experts who know ALCF systems best. This year, our approach had to change due to the circumstances surrounding the coronavirus pandemic,” said Yasaman Ghadar, an ALCF assistant computational scientist and lead organizer of the event. ​But by leveraging the BlueJeans videoconferencing service and the Slack instant messaging platform, we were able to enable interaction between participants and staff and, with relative seamlessness, educate attendees on the various tools and services that can help accelerate research on our machines. Happily, this resulted in a relatively large number of breakout sessions with guided, hands-on work.”

Some 80 percent of this year’s participants were part of an ongoing project at the ALCF, and over half intended to submit proposals for new research following the workshops.

The yearly workshop, typically held at the ALCF during the first week of May, is designed to help prepare users improve code performance and prepare for major allocation awards, such as the ALCF Data Science Program (ADSP), the ASCR Leadership Computing Challenge (ALCC), and DOE’s Innovative and Novel Computational Impact on Theory and Experiment (INCITE) program.

A sprawling array of topics were covered over the course of the three-day workshop, co-organized by Ray Loy, ALCF Lead for Training, Debuggers, and Math Libraries. The schedule included an OpenMP overview and interactive session led by Colleen Bertoni; data visualization overview and breakout session delivered by Joseph Insley, Silvio Rizzi and Janet Knowles; a presentation from Loy about debugging code on ALCF machines; and a presentation delivered by Misha Salim about the workflow management tool and edge service Balsam. Also popular were presentations about Aurora, the ALCF’s upcoming exascale system.

The wide variety of topics covered reflects the variety of tools and services the ALCF provides, as well as the vibrancy of the research the facility enables,” said Loy.

Additionally, it’s of course always encouraging to see the steadily growing interest in and enthusiasm for exascale,” he continued. ​With a robust community engaged in advancing research computing, we fully expect to deliver science on day one following the arrival of Aurora.”

Videos of presentations and sessions from the ALCF Computational Performance Workshop are available here.

Stay tuned to the ALCF events webpage for details on the facility's virtual workshop this fall - the 2020 Simulation, Data, and Learning Workshop.

 

The Argonne Leadership Computing Facility provides supercomputing capabilities to the scientific and engineering community to advance fundamental discovery and understanding in a broad range of disciplines. Supported by the U.S. Department of Energy’s (DOE’s) Office of Science, Advanced Scientific Computing Research (ASCR) program, the ALCF is one of two DOE Leadership Computing Facilities in the nation dedicated to open science.

Argonne National Laboratory seeks solutions to pressing national problems in science and technology. The nation’s first national laboratory, Argonne conducts leading-edge basic and applied scientific research in virtually every scientific discipline. Argonne researchers work closely with researchers from hundreds of companies, universities, and federal, state and municipal agencies to help them solve their specific problems, advance America’s scientific leadership and prepare the nation for a better future. With employees from more than 60 nations, Argonne is managed by UChicago Argonne, LLC for the U.S. Department of Energy’s Office of Science.

The U.S. Department of Energy’s Office of Science is the single largest supporter of basic research in the physical sciences in the United States and is working to address some of the most pressing challenges of our time. For more information, visit https://​ener​gy​.gov/​s​c​ience.