BIOLOGY

University’s fourth and fastest supercomputer attracts new researchers to campus: Experiments at UC Riverside that once might have required 256 days of computer time on a personal computer can now be performed in just one day thanks to a new supercomputer the campus has acquired. Named “Biocluster,” the supercomputer – UCR’s fourth – was purchased for $280,000 by the Bioinformatics Facility of the Institute for Integrative Genome Biology (IIGB). The new computing resource, a Linux blade cluster system, is now the most powerful computer at UCR. It significantly reduces the computing time of large-scale genome analysis, drug discovery and simulation studies for many research projects. “This new system is a valuable resource for many of our scientists,” said Thomas Girke, the director of the Bioinformatics Facility and an assistant professor of bioinformatics in the Department of Botany and Plant Sciences. “It changes not only the nature of science at UCR but its quality as well, making UCR more competitive for large federal grants. Computational research is being done faster and more efficiently than ever before on our campus.” Girke oversees Biocluster’s use, and is responsible for setting up installation services, designing the cluster and the software tools available, as well as communicating with different labs on campus. Acquired and custom configured from Verari Systems, a computer company located in San Diego, Biocluster already is being used by many campus researchers, allowing computationally-intensive research projects to be conducted on campus that previously had to be done elsewhere. Thomas Girke, the director of UC Riverside's Bioinformatics Facility, stands next to Biocluster, a new supercomputer the campus has acquired. Photo credit: UCR Strategic Communications. “Before Biocluster became available, I could run my simulations only on national supercomputers supported by the National Science Foundation’s Supercomputer Cluster program,” said Chia-en Chang, an assistant professor of chemistry who came to UCR from UC San Diego in January 2008. Chang works on molecular interactions and computer-aided drug design, a computationally demanding task involving highly complex simulations. “Biocluster enables us to have much more computer power and much better flexibility to do our research in studying protein-protein interactions and protein dynamics,” she said. “Although I can do parallel computing using a national supercomputer, I would need to share computer power with researchers nationwide. This means having to wait in a queue for running our jobs. With Biocluster, we don’t have to wait for days before we can run our simulations.” The supercomputer also has made possible the hiring of new faculty at UCR. “The university just hired three faculty members in computational biology,” Girke said. “Hiring them would have been very difficult – if not impossible – without a state-of-the-art infrastructure for high-performance computing.” Currently, research programs in all bioscience, biomedical, engineering and statistics departments are using Biocluster for a moderate user fee, including researchers not affiliated with UCR. Besides Biocluster, UCR has supercomputers in the Department of Physics and Astronomy; the Department of Computer Science and Engineering; and an older cluster in IIGB. The new supercomputer was delivered to campus in early April. Located temporarily in the Computing and Communications data center, it will be moved, in December, into a server room in the new Genomics Building. Biocluster details: The new supercomputer from Verari Systems has 256 CPU cores (64x Xeon Quad Harpertown, 12 MB cache), 640 GB of total RAM and the latest Infiniband interconnect. Its high-memory node has a total of 64 GB of usable memory available for a single process. If needed, the memory on this node can be expanded to a total of 192 GB.