As preparations begin for unveiling their new facility, the Massachusetts Green High-Performance Computing Center (MGHPCC) announced the first winners of their seed grants, $600,000 in funding that range from individual grants valued at $45K to $130K.
These project winners will be among the first to use the center, and the recent recipients included several BU faculty, including the Department of Electrical & Computer Engineering’s Ayse Coskun and Martin Herbordt.
The center won’t open until later this year, but the green facility is already hailed by IEEE Spectrum as “one of the world’s most ambitious university computer centers.”
The research building, located in Holyoke, Mass., will have some of the fastest supercomputers on hand and house much of the research computing resources of five area schools – Boston University, Harvard University, MIT, Northeastern University, and the University of Massachusetts.
The MGHPCC will support Coskun, Herbordt, and Northeastern’s Gunar Schirner in their work on measuring and improving the energy efficiency of large-scale computing.
“We’re really looking forward to this opportunity,” said Herbordt. “If we had to conduct these experiments in our own labs, it would be a real challenge. Instead, we’ll have a chance to conduct them at an industrial-scale high-performance computing center.”
Specifically, Coskun, Herbordt, and Schirner aim to create systematic, cheaper methods for developing software that will also use less energy.
Other BU professors who received MGHPCC grants include Lorena Barba (ME) and Jonathan Appavoo (CS). Like Coskun’s and Herbordt’s research, their projects will use funding from MGHPCC, the BU Provost’s Office, and the Rafik B. Hariri Institute for Computing and Computation Science & Engineering.
Barba will work with Cris Cecka (Harvard) and Hans Johnston (UMass Amherst) to develop “exascale” software platforms, which have the potential to make computing speeds 1,000 times faster than what are currently used on high-performance computers. As part of this work, they also hope to develop algorithms that would be capable of solving diverse, complex problems in areas such as biophysics, acoustics, and fluid dynamics.
Appavoo will work with Patricia Ellen Grant (Harvard and Children’s Hospital, Boston) to use high-performance computing to improve medical imaging analysis by making it more automated. This would make the process less costly and easier for radiological clinicians to use. Patients may see results faster, too.
-Rachel Harrington (firstname.lastname@example.org)