The world of supercomputing is pivoting to the novel coronavirus, with some other projects being paused as researchers focus on finding treatments and vaccines as well as studying the spread of the virus and assessing the impact of social-distancing measures.
“This is everybody’s highest priority right now,” said Dan Stanzione, executive director of the Texas Advanced Computing Center. Typical supercomputing research projects related to predicting hurricanes and earthquakes are running more slowly in order to prioritize Covid-19 projects.
Supercomputers can have tens of thousands of processors that work together to perform large calculations. Some calculations on supercomputers can be completed in one day, versus hundreds of years to run on a laptop with only a handful of processors, Mr. Stanzione said.
Approved researchers have free access to supercomputers under an initiative announced earlier this month involving Energy Department national labs, technology companies and academic institutions.
Combined, the approximately 30 supercomputing systems that are part of the Covid-19 High Performance Computing Consortium represent more than 330 petaflops of computing capacity. A petaflop allows for 1,000 trillion, or one quadrillion, operations per second.
Rommie Amaro, a researcher at University of California, San Diego, used a supercomputer based at the Texas Advanced Computing Center to investigate the structure of the novel coronavirus.
Photo:
Texas Advanced Computing Center
The Texas Advanced Computing Center, part of the consortium, is home to two supercomputers, Frontera and Stampede2, as well as other smaller ones, all of which researchers can access remotely. About 100 researchers across the country are using the center’s computers for about 10 different Covid-19 research projects, including those that involve epidemiology and vaccines. They began their work roughly a month ago, Mr. Stanzione said.
Lauren Meyers, a professor of integrative biology at the University of Texas at Austin, is using the center to model the transmission of the virus between people in various regions to get a better understanding of how the disease is spreading. “When cases are reported from different countries, often the data aren’t telling us the whole story,” Ms. Meyers said.
For example, one of Ms. Meyers’s supercomputing models suggested there were probably more than 11,000 cases of Covid-19 in Wuhan, China, center of the pandemic outbreak, by the time officials there imposed lockdown measures in January. At that time, Wuhan had reported about 425 cases, she said. The Chinese Embassy in Washington didn’t respond to a request for comment.
Ms. Meyers is also using supercomputing models to evaluate the impact of various social-distancing measures enacted by state and federal policy makers, she said. The goal is to provide policy makers with information about the consequences of relaxing or strengthening those measures, she said. So far, the models suggest that the U.S. could expect to see “many weeks or even months of various kinds of interventions” to protect people from the virus, Ms. Meyers said.
Researchers at Argonne National Laboratory in Lemont, Ill., are using supercomputers to study the spread of the virus within communities and the evolution of the virus, in addition to other projects such as finding new antiviral drugs and accelerating vaccine development.
Supercomputing partnerships between researchers, government officials and technology companies underscore that there is an urgency to accelerating humanity’s understanding of the virus, said Chirag Dekate, senior research director at research firm
Gartner Inc.
Some of the machines that researchers have access to are orders of magnitude more powerful and faster than computing resources typically used by enterprises, including pharmaceutical companies, Mr. Dekate said.
About 25 researchers from Oak Ridge National Laboratory, the University of Tennessee and other national labs and universities are using Summit, an
International Business Machines Corp.
supercomputer, to find drugs that could potentially treat the pneumonia-like disease caused by the virus. It is likely that more than one drug will be needed to treat the disease because the virus could evolve to become resistant to certain drugs, said Jeremy Smith, director of the University of Tennessee/Oak Ridge National Lab Center for Molecular Biophysics.
The main advantage of the Summit computer is speed; Summit has the computing power of approximately 1 million laptops all working together to solve the same problem, Mr. Smith said. Still, he said it is uncertain how long it will take to find a Covid-19 treatment.
“That’s the infuriating aspect of this. It’s scientific research, and you never know whether you’ll be successful in reaching your end goal,” he said.
Write to Sara Castellanos at sara.castellanos@wsj.com
Copyright ©2020 Dow Jones & Company, Inc. All Rights Reserved. 87990cbe856818d5eddac44c7b1cdeb8
