For the last 10-plus years, our industry has been navigating the road to exascale. For the United States, that road hit a major milestone. Cray will deliver the nation’s first exascale supercomputer to the U.S. Department of Energy’s Argonne National Laboratory in 2021 in partnership with Intel.
Named Aurora, the Argonne system will be comprised of more than 200 Cray® Shasta™ system cabinets, our unique software stack optimized for Intel architectures, the Cray® Slingshot™ interconnect, and Intel innovations in compute processor, memory, and storage technologies. As a subcontractor to Intel, our part of the program contract is valued at more than $100 million ― one of the largest in our company’s history.
Are we proud? You bet we are.
The announcement came last week at an Argonne-hosted event stacked with headliners, including U.S. Department of Energy Secretary Rick Perry, U.S. Senator Richard Durbin, U.S. Congressman Daniel Lipinski, Intel CEO Robert Swan, Argonne Director Paul Kearns, University of Chicago President Robert J. Zimmer, and, of course, our own CEO Pete Ungaro. Pete gave some exceptional commentary on the system architecture and how we got to this point in computing history. (See for yourself. Scroll to minute 37:00 of the announcement event.)
But while the road to exascale may have ended, we know the real journey is just beginning. Exascale technology opens possibilities for discovery that humankind has only ever imagined.
Characteristics of the exascale era
You could say the exascale era isn’t really about speed. It’s about the kind of workloads scientists and engineers need to run. It’s an era of extreme, data-intensive computing where AI and analytics workloads converge with modeling and simulation.
As our partners and customers shared their insight and vision with us over the last decade, we saw that meeting their needs demanded a new, data-centric approach to system design. Not only that, we recognized this new design needed a flexible, future-proof compute and cabinet architecture, a modular and extensible software stack that could evolve as exascale workloads matured. And it needed a completely rethought network fabric so diverse workloads could scale seamlessly. The Shasta system was the result. Read the announcement for more details on the Aurora system itself.
And let’s not forget the most important characteristic of the exascale era… scientific achievement. According to the DOE, exascale computing will be central to the national effort to lead in critical areas such as artificial intelligence, health care, and security.
Argonne Early Science Program
A system is only as great as the science it’s doing. With Argonne at the helm, Aurora is already way ahead. The Argonne Leadership Computing Facility (where Aurora will be housed) launched the Aurora Early Science Program two years ago to help prepare key applications, libraries, and infrastructure for the architecture and scale of an exascale system.
The Aurora ESP projects come from a variety or scientific fields ― from neuroscience to materials science – and represent a combination of simulation, data science, and machine learning approaches. The projects will transform how supercomputers are used for scientific discovery and innovation. Argonne has a full list of the projects as well as additional, separate detail broken out between simulation and data and machine learning projects.
Overall, as we look ahead to the discoveries exascale computing will make possible, we think Pete said it best: “It’s going to be an amazing ride.”