Faculty Recruiting Support CICS

Shenoy receives grant to build academic cloud at MGHPCC

Professor Prashant Shenoy, along with researchers Chris Hill of MIT, Claudio Rebbi of Boston University, and Gene Cooperman of Northeastern University, recently received a grant from the National Science Foundation's Major Research Instrumentation Program to support a shared computer cluster housed at the Massachusetts Green High-Performance Computing Center (MGHPCC) in Holyoke, MA.

The new cluster, a collection of a few thousand servers, will serve a large number of scientists in a broad variety of disciplines at the four institutions who need high-performance machines that can carry out a high volume of extremely fast calculations and store large amounts of data, Shenoy says.

"Any scientist in the life sciences, genetics, weather modeling, geosciences and engineering who needs high-performance computing capabilities will be able to request time on the new cluster," he adds. "It's really a resource for the broad campus community. Our intent is to have a way for all of our campuses to benefit from and use this cluster, which we think of as an academic cloud."

Computer science researchers from the participating institutions will work closely with scientist users to research new systems-level paradigms to help make high-performance computing more routinely used in day-to-day research workflows. The four principal researchers will also address issues such as effective sharing policies, privacy, security, and how to use energy most efficiently.

Shenoy notes, "This new system will provide support for the many exciting new big-data initiatives across the UMass system such as genomics research and personalized medical treatment studies which can take a huge amount of computational time. There is a pretty strong need among our researchers to invest in this. We are one of the first clusters to go in, so it's an exciting time."

See UMass Amherst Research NEXT article  (https://www.umass.edu/researchnext/big-data) on the NSF MRI grant to support a shared cluster.