Nvidia Corp. is helping the U.S. Department of Energy’s Argonne National Laboratory build another incredibly powerful supercomputer, it announced today.

The Polaris supercomputer (pictured) is said to be the DOE’s largest graphics processing unit-based machine so far. It’ll be hosted at the Argonne Leadership Computing Facility, where it will aid research and discovery with extreme scale for user’s algorithms and science, the DOE said. An exascale computer is one that can calculate at least 10¹⁸ floating point operations per second.

Hewlett-Packard Enterprise Co. is building Polaris, which will be supercharged with 2,240 Nvidia A100 Tensor Core GPUs squeezed into 560 total nodes when it’s deployed in early 2022. That will enable it to achieve almost 1.4 exaflops of theoretical artificial intelligence performance, Nvidia said, and up to 44 petaflops, or quadrillion FLOPS of peak double-precision performance.

Ian Buck, Nvidia’s vice president and general manager of accelerated computing, said the new era of exascale AI that Polaris will usher in will lead to new scientific breakthroughs that bring incredible benefits to society. “Nvidia’s GPU-accelerated computing platform provides pioneers like the ALCF breakthrough performance for next-generation supercomputers such as Polaris that let researchers push the boundaries of scientific exploration,” he said.

Polaris has been earmarked for accelerating transformative scientific exploration into things such as new cancer treatments, clean energy sources and propelling particle collision research to discover new approaches to physics, the ALCF said. It will also enable ALCF’s researchers to update their scientific workloads for Aurora, a planned exascale supercomputer system that will go online next year.

“Polaris is a powerful platform that will allow our users to enter the era of exascale AI,” said ALCF Director Michael E. Papka. “Harnessing the huge number of Nvidia A100 GPUs will have an immediate impact on our data-intensive and AI HPC workloads, allowing Polaris to tackle some of the world’s most complex scientific problems.”

ALCF said it will also make Polaris available to researchers from academia, government agencies and industry through its peer-reviewed allocation and application programs. The idea is to assist the scientific community in addressing some of the “grand challenges” in science and engineering.

Nvidia enjoys a growing presence in the supercomputing market. In July it announced it had participated in the building of Cambridge-1, the U.K.’s most powerful supercomputer that 80 of its DGX A100 computing modules. The company’s hardware also powers Perlmutter, said to be the world’s fastest supercomputer built specifically for AI workloads, as well as another machine hosted by the Argonne National Laboratory, called Selene.

Image: Nvidia