The University of Wisconsin-Milwaukee has been awarded funding from the National Science Foundation to upgrade the university’s aging supercomputing infrastructure. The $400,000 grant will expand campus research capabilities and research-related educational opportunities.
Supercomputers are used in fields that require high-speed computation, such as weather forecasting, physical simulation, molecular modeling and astrophysics. However, UWM’s cluster, which was established in 2009, is no longer powerful enough to accommodate many researchers’ needs.
“Supercomputing looms large in both the volume and funding associated with UWM’s computation-intensive research,” said Robert Beck, UWM associate vice chancellor and chief information officer. “These NSF grant funds will add unique resources not currently available within our aging cluster environment.”
The need for computation in research has exploded over the last 10 years, said Dan Siercks, interim director of research computing. Last year alone, the supercomputing cluster ran more than 600,000 computational jobs, and seven of the top 10 research projects at UWM used the cluster.
Designing a model that works
Philip Chang, associate professor of physics, recruited 11 other researchers across disciplines to write the grant. Coauthors on the grant represented UWM research in astrophysics, atmospheric science, cancer research, genomics, fluid dynamics, biophysics and quantum mechanics.
“We found that no single kind of supercomputer would satisfy all the competing requirements for different computational projects in terms of computing, storage and memory,” Chang said, “so we designed one that does.”
This new investment will support science and engineering throughout the region, not just UWM, Chang said.
“Innovation sparked by the research will directly drive high-end industrial development and the resulting economic rewards,” he said.
For example, Professor Ryo Amano’s research involves modeling the performance of turbine blades for wind energy, and Associate Professor Kevin Renken optimizes heat exchangers used in air separation plants and transport industries. Both require simulation software with hefty computing requirements.
Distributed computing consortium
The upgrade also will allow UWM to participate more fully in the Open Science Grid, a distributed computing consortium that includes 42 universities, including UW-Madison and UWM.
Distributed computing links many partners’ individual supercomputing clusters so that the idle time of any computers across the consortium can be used to process large datasets. “The idea is to keep the computers of all partners in use all the time,” said Siercks, and that increases the high-performance computing resources for partners.
The new supercomputer will have educational benefits as well. In the next five years, between 60 and 100 undergraduates will use the cluster for education and research, while more than 100 graduate students will use it for coursework and research. And it will enhance UWM’s ability to recruit and retain students from underrepresented groups in the STEM disciplines.
Besides Chang, Renken and Amano, the following researchers helped write the grant: Sarah Vigeland, David Kaplan, Abbas Ourmazd, Peter Schwander, Russell Fung and Michael Weinert from physics; Clark Evans from atmospheric science; Mahsa Dabagh from engineering; and Rebecca Klaper from freshwater sciences.
The grant comes from the NSF’s Campus Cyberinfrastructure program, which invests in campus-level networking and cyberinfrastructure improvements for science applications and distributed research projects.