As seen in the Pitt Research October 2021 Newsletter
A multi-disciplinary team of Pitt researchers has landed a $1.2 million National Science Foundation (NSF) award for new computing resources that will greatly boost the Center for Research Computing’s (CRC) capabilities in speed, power, and scope. Chemistry associate professor Geoffrey Hutchison led the proposal along with associate professors Lillian Chong in chemistry, Inanc Senocak in mechanical engineering and materials science, and David Koes in computational and systems biology. Internal Pitt funding added to the grant creates a total of over $1.5 million for new resources.
“This is not a grant to one researcher or one project but an investment in research itself. We don’t know how wide the impact will be over the life of the technology. The multiplier effect could be huge,” said Hutchison.
More than 30 existing research groups across the university stand to immediately benefit from the new resources, groups representing over $18 million in NSF grants in fields including chemistry, computational biology, materials science, psychology, astrophysics, weather forecasting, energy, and sustainability.
For CRC, the grant enables a leap in technology based on state-of-the-art graphics processing units – GPUs, a technology originally developed for computer gaming and now widely adopted for artificial intelligence and machine learning. The new GPUs are twice as fast as the previous generation – and up to 50 times faster than chips used in standard computing.
“GPU technology is driving a revolution of unprecedented discovery in science, engineering, and beyond,” says Hutchison. “Not only are these advanced resources, but we’re going to have a lot of them. Colleagues are surprised when I tell them how many GPUs we are getting.”
The grant also funds advanced resources for distributed computing, boasting a total of 18 terabytes of aggregate memory (a terabyte equals roughly 1,000 gigabytes) and 2,300 central processing unit cores, enabling an immense range of computing being used at Pitt to create better weather models, molecular simulations of disease cells, and new models of financial credit scores.
CRC co-director Kim Wong calls the boost in computing capability “remarkable.” "This opens the ceiling for innovation. Pitt researchers will be able to create more detailed, realistic simulations and models, leading to discoveries that directly impact everyday life. We are excited to be part of a team empowering the unbounded creativity of our research community."
A significant aspect of the team’s success in securing the grant was CRC’s educational outreach with online workshops that make the expertise of its research faculty consultants available to a wide audience. Initially in response to the COVID-related shutdown, CRC began presenting online workshops ranging from basic cluster usage to data science and machine learning. The workshops are regularly attended by students and faculty at Howard University in Washington, DC, and the center is expanding its outreach to Pitt regional campuses, regional research consortiums, and Bowie State University in Maryland, as well as other Historically Black Colleges and Universities. Locally, CRC is working to make research and learning resources available for students in Pittsburgh Public Schools.