After months of planning, Brown is set to unveil a new multimillion-dollar supercomputer today.
The result of a partnership with IBM to boost the University's research capacity and visibility, the high-performance computing cluster will allow researchers from all disciplines to conduct powerful, complicated data analyses.
Researchers and faculty have been seeking an advanced computational system for more than three years, said Professor of Applied Mathematics Jan Hesthaven, who, along with Vice President for Research Clyde Briant and Vice President for Computing and Information Services Michael Pickett, led the initiative to bring the system to the University.
The system, which will be the most powerful supercomputer in Rhode Island and is capable of performing 14 trillion operations per second, will be open to educational and research institutions across the state, Hesthaven said.
"You can't have a research institution that would like to be at the top of its class without having this type of facility," Hesthaven said. "Think of it as a necessity rather than a luxury."
The supercomputing network, located in the Center for Computation and Visualization at 180 George St., will provide opportunities to collaborate with other research centers, including the University of Rhode Island and the Marine Biology Laboratory in Woods Hole, Mass., he said.
"At this time, computers are like libraries and laboratories," Hesthaven said. "It's not about physics or engineering or science, but about research."
Today's official ribbon-cutting ceremony, featuring speeches by Gov. Donald Carcieri '65, Mayor David Cicilline '83 and IBM's Vice President of Technology, Nick Bowen, marks the conclusion of the years-long effort to streamline computational research at the University and bring first-rate equipment to the state, Briant said.
"This is a big step forward for us at Brown," Briant said.
‘A vision'
Hesthaven, who will direct the computational center, said he first broached the concept of assembling a computing cluster at Brown three years ago, when he saw a dearth of effective computing power. A centrally located, high-performance cluster would allow researchers to conduct their data analysis on-site, rather than analyzing their data remotely at other institutions, Hesthaven said.
"This was a plan for basically what Brown should have, given what we think we should be," he said. "This became a vision."
Hesthaven said he advocated technological advancements that would inspire statewide collaboration, ultimately securing support for the high-powered cluster from Briant and other administrators.
As part of the Plan for Academic Enrichment, Pickett — who built a similar system at Duke University before he came to Brown in 2007 — drafted an information technology strategic plan last January, which offered recommendations for advancing technology at the University.
Pickett said CIS partnered with the Center for Computation and Visualization to address the logistics behind building a high-powered computing system to promote more effective research collaboration.
"Researchers can't share resources if they have to take all the responsibility themselves," he said, emphasizing the University's need to adapt to modern technological and research models.
Despite widespread support for the network, questions remained over how the University would obtain the necessary machinery to construct the supercomputer. But the problem was solved last May, when Brown signed a memorandum of understanding with IBM, which offered affordable equipment and free consulting to jumpstart the creation of the supercomputer, Briant said.
Over the summer, the University began updating the existing data center to prepare for the cluster, which occupies an entire room in the center. Since early this fall, the cluster has been in a trial phase, operating at 90 percent capacity with researchers volunteering to test the new system before its official launch.
So far, the supercomputer has generated widespread interest from researchers across many disciplines — Hesthaven said there are about 75 users, including undergraduate and graduate students.
"This is all about the institution," he said. "It's not just a toy for the select few."
Branching out
Though many similar supercomputers are reserved for only more advanced research, Hesthaven said the facility will also be used by middle- and high-school students across the state.
"This is an opportunity for high-school students to get their feet wet on equipment," he said. "It's an opportunity to explore the potential for collaborative efforts."
Hesthaven said the University plans to offer courses and seminars to undergraduates that would involve use of the computing cluster.
The opportunity to share resources with other institutions is also more energy-efficient, Hesthaven said, because the network of machines uses an enormous amount of power to prevent the motors from overheating.
The cluster is "more environmentally friendly," he said. "It's a much greener way of doing things." The promise of future collaborative statewide efforts will also help stimulate the knowledge-based economy throughout the state, Briant said, adding that the central cluster will attract new faculty and larger-scale projects. Having a high-performance computational system at Brown may also attract future donors, though the University is not using the supercomputer as a "money-maker," he said.
Because data analysis can now occur at Brown instead of at other institutions with their own networks, Briant said, the University plans to apply for state and national research grants that it previously would not have considered.
Already, a statewide team of researchers led by Professor of Pediatrics James Padbury submitted a grant proposal last month for a clinical and translational sciences award, which would bring in $20 million over a five-year period if the National Institutes of Health awards Brown and its partners the grant next spring. The supercomputer was a key component of the proposal, Padbury said.
"The supercomputer cluster will fund high-dimension analysis," he said, adding that high-performance computing is especially important for biological research. He said collaborative projects that would use the cluster are already in the works, including "genome-wide association studies" and a study on sudden cardiac death.
"Bringing these kinds of tools in-house opens up all kinds of possibilities," said Assistant Professor of Biology Casey Dunn, who conducts research in developmental and evolutionary biology. "My work is absolutely dependent upon high-performance computing."
Dunn said he started using the supercomputer in October to analyze data sets, adding that the undergraduates in his lab have used the cluster.
The computing cluster will also benefit researchers in the humanities, who can use its power to study demographics and analyze elections.
"It's exciting," Dunn said. "It's all very ethereal."
