White House Expands Coronavirus-Focused Supercomputing Consortium

The White House accepted three new members into the COVID-19 High Performance Computing Consortium this week—less than a month after producing the new partnership to accelerate research around the novel coronavirus and ultimately help halt it.

The National Center for Atmospheric Research’s Wyoming Supercomputing Center, chipmaker AMD and graphics processing units-maker NVIDIA join a variety of other national labs, agencies, companies and academic institutions that have volunteered a range of supercomputing systems and resources for interested individuals to possibly leverage for research efforts that could help combat the COVID-19 pandemic.

U.S. Chief Technology Officer Michael Kratsios announced the new members’ inclusion via tweet Monday, and in it, also added that the consortium “has already matched 15 research proposals with compute power and resources on some of the world’s most powerful machines,” since its launch on March 22.

At this point, the consortium provides access to 30 supercomputers with more than 400 petaflops of compute performance.

NVIDIA released its own announcement Monday regarding the expertise its newly established  “task force of computer scientists” will apply to support the consortium, but its Vice President of Solutions Architecture and Engineering Marc Hamilton also told Nextgov in an email Tuesday that the tech company has already been contributing to several consortium projects. 

“More than 41,000 NVIDIA GPUs deliver about 75% of the combined capability of the 30 consortium supercomputers,” Hamilton noted.

Though this marks the Trump administration’s first dispatch of new consortium members, the cadre is continuously accepting more. Hamilton added that while the initial announcement regarding the consortium’s partners named supercomputing labs and cloud service providers that could have major high-performance computing resources available immediately, the White House Office of Science and Technology Policy invited NVIDIA to join the group in this “second wave.”

“We’ve worked with researchers at Oak Ridge National Laboratory and other consortium supercomputer centers for more than 10 years on accelerated computing using NVIDIA GPUs,” Hamilton said. “So it was a natural fit for us to join.”

On a call with reporters following the consortium’s initial launch in late March, Energy’s Undersecretary for Science Paul Dabbar explained how the new strategic partnership works. To start, researchers anywhere can submit relevant coronavirus-related research proposals to an online portal created and run by the consortium. Those submissions are subsequently reviewed and potentially matched with computing resources and support from the public, private or academic partners involved whose focus areas and materials make the most sense to support the studies.

Dabbar added that the sophisticated systems that the consortium is making available for public use “can process massive amounts of calculations related to bioinformatics, epidemiology, molecular modeling and healthcare system response—helping scientists develop answers to complex scientific questions about COVID-19 in hours or days, rather than weeks or months.”

It’s early to tell, but those assets appear to be paving the way for some results. Since its launch several weeks ago, as Kratsios briefly mentioned, 15 research proposals have been matched with partners to date. So far, 11 of those “active projects” are now highlighted on the consortium’s site. They include several studies that incorporate artificial intelligence and machine learning to advance new discoveries around drugs and therapeutics to fight the virus, one that uses a NASA-run supercomputer to define risk groups for severe pulmonary disease associated with COVID-19, and more. 

In reference to some of those consortium-driven ongoing efforts, NVIDIA’s Hamilton highlighted that on ORNL’s Summit supercomputer, eight COVID-19 research projects “have already consumed nearly 1.5 million GPU-hours of compute.” He said no other supercomputer in the world could have delivered that amount of compute in such a short time—and the power of GPUs helped make it possible.

“The world needs effective treatments and a vaccine 100X faster,” Hamilton said. “So, there’s nothing more important we could be working on.”