The US government, industry and academia have pooled efforts to fight the COVID-19 High Performance Computing Consortium can be used for work like projecting the disease’s spread and modeling possible medicines.by offering researchers access to some of the world’s most powerful supercomputers from IBM, along with cloud-computing resources from Amazon, Microsoft and Google. The
“By pooling the supercomputing capacity under a consortium of partners … we can offer extraordinary supercomputing power to scientists, medical researchers and government agencies as they respond to and mitigate this global emergency,” said Dario Gil, director of IBM Research, in a statement. IBM is coordinating the effort.
Supercomputers, mammoth machines that can take up entire floors of buildings and consume as much power as a town, are used for tasks like simulating nuclear weapons explosions, global climate change effects and the physics of the cosmos. They also can be good at medical research like drug discovery — a key ability given the fast spread of the new coronavirus and the COVID-19 pandemic it’s caused.
IBM built several of the biggest machines involved in the effort, like the Lassen supercomputer at Lawrence Livermore National Laboratory, with 34,848 IBM Power9 processor cores assisted by 3,168 Nvidia graphics chips.
But a wide variety of machines are involved, including horsepower from the three biggest cloud-computing services: Amazon Web Services, Microsoft Azure and Google Cloud. Those services offer enormous quantities of computer power, but typically it’s spread more widely across different machines that don’t have the superfast internal data connections of supercomputers.
The IBM-built Summit machine, the world’s fastest supercomputer today, already has been used to screen 8,000 chemical compounds on a search for COVID-19 medicine that could thwart its infectious power. So far, researchers in that effort at Oak Ridge National Laboratory and the University of Tennessee recommended 77 drug compounds for experimental testing.