Supercomputing, big data and artificial intelligence are crucial tools in the fight against the coronavirus pandemic. Around the world, researchers, corporations and governments are urgently devoting their computing resources to this global crisis. This column collects the biggest news about how advanced technologies are helping us fight back against COVID-19.
In a briefing last week, the Department of Energy went into more detail regarding the new COVID-19 High-Performance Computing Consortium. In the briefing, the DOE referenced several projects that are already underway, highlighting how the division of labor for fighting COVID-19 is being spread across major supercomputing facilities. The DOE also launched a portal for submitting proposals to the consortium. To read more, visit the HPCwire article here.
The MVS-10P cluster at the Joint Supercomputer Center of the Russian Academy of Sciences has joined the fight against the coronavirus. The 771-peak teraflop supercomputer is being used by an international research team representing a dozen institutions and a half-dozen countries to study the telltale spike protein of the coronavirus. Specifically, the researchers are looking to design aptamers (molecules that bind to specific target molecules) that are more attractive to COVID-19 than the human ACE2 protein. To read more about this work, visit the HPCwire article here.
Aiming to identify what makes individuals more or less susceptible to severe symptoms of COVID-19, Bionano Genomics has launched the largest study yet on genomic variations and active substances that affect resistance or sensitivity to the pandemic. A number of companies are supporting the study through the Tech Against Covid initiative, including Rescale, which is providing its high-performance computing cloud platform, and AWS. To read more, visit the press release here.
Lawrence Livermore National Laboratory (LLNL) highlighted its COVID-19 research, outlining an interdisciplinary team of researchers working to find antibodies. LLNL’s work began with the production of a 3D protein structure of COVID-19. That structure was then leveraged by another team at LLNL, which used a pair of HPC clusters to apply AI to antibody analysis simulations. Yet another group examined 26 million molecules on the Quartz cluster (delivering 3.25 peak petaflops) to see how they interacted with key proteins in COVID-19. To read more, visit the HPCwire article here.
The National Energy Research Scientific Computing Center (NERSC) and the Energy Sciences Network (ESnet) both operate under Lawrence Berkeley National Laboratory, which has spun down much of its activity to ensure social distancing – but NERSC and ESnet, both providing essential services, continue to support the nation’s scientists. NERSC, a member of the COVID-19 High Performance Computing Consortium, is working to prioritize COVID-19 research. To read more, visit the HPCwire article here.
PRACE awards its first COVID-19 project under fast-tracked proposal process
Last week, PRACE announced that it was fast-tracking COVID-19 proposals – and it’s already awarded its first fast-tracked project. The project, titled COVID-HP, is led by Jean-Philip Piquemal, a professor of theoretical chemistry at the Sorbonne. The research team will use 20,000,000 core hours on the Joliot-Curie Rome supercomputer (a 6.6-peak petaflop system hosted by GENCI) to research inhibitors targeting COVID-19’s key proteins and certain nucleic acid sequences in its genome.
The Barcelona Supercomputing Center (BSC) is examining the development of COVID-19, studying how it evolved through different epidemics in order to understand how it passes from species to species. BSC is crunching the data – which includes genomic sequences of virus mutations and animal species – with its MareNostrum 4 supercomputer, which delivers 13.7 peak petaflops. BSC is also using MareNostrum 4 to examine possible treatments for COVID-19 by testing molecular docking. To read more, visit the press release here.
Immunotherapy company ImmunityBio is pitting cloud supercomputing against the coronavirus, running molecular dynamics simulations of the spike protein on a cloud HPC platform through Microsoft Azure. The simulations produced a detailed model of the spike protein’s solution structure, affording researchers a clearer pathway to producing a drug of treatment to debilitate COVID-19. The Azure cluster used to conduct the simulations contained over 1,250 Nvidia V100 GPUs. To read more, visit the press release here.
Following in the footsteps of crowdsourced computing projects like Folding@home and Rosetta@home, IBM has announced that its own crowdsourced computing platform, the World Community Grid, will be contributing to COVID-19 research. The project, called OpenPandemics – COVID-19, is designed and led by Scripps Research and promises to deliver “hundreds of millions of calculations” for COVID-19-related simulations. To read more, visit the press release here.
IU has announced that Jetstream, its cloud-based, on-demand computing resource, is fast-tracking COVID-19-related projects within the Extreme Science and Engineering Discovery Environment (XSEDE). Jetstream is operating through the COVID-19 HPC Consortium. “Jetstream is proud to be part of the COVID-19 HPC Consortium, putting the power of our collective resources into the hands of scientists who can use them to respond to this unprecedented crisis,” said David Hancock, director of advanced cyberinfrastructure at IU and PI of Jetstream. To read more, visit the press release here.
Do you know about COVID-19 research that should be featured on this list? If so, send us an email at [email protected]. We look forward to hearing from you.