Friday, Sept. 23, 2011
HUNTINGTON, W.Va. – Taking on some of mankind’s greatest challenges and advancing cutting-edge science, research and learning all require enormous computing power.
Now, researchers, faculty and students at Marshall University have access to a new high-performance computing (HPC) cluster that will enable them to make significant advancements in fields as diverse as bioinformatics, climate research, physics, computational chemistry and engineering.
According to Dr. Jan I. Fox, Marshall’s senior vice president for information technology and chief information officer, all these disciplines rely on the state-of-the-art computing tools and methods provided by the new cluster.
“This new HPC cluster makes possible scholarly innovation and discoveries that were, until recently, possible only at the most prestigious research institutions,” she said. “Along with our connection to Internet2, our students and faculty now have access to computing power, data and information we could only imagine just a few years ago.”
Nicknamed “BigGreen,” Marshall’s new cluster is made up of 23 high-end computer systems housed in the university’s Drinko Library. Once they have a user account on BigGreen, researchers can access the cluster and its resources from anywhere.
For the computer-savvy, it is important to note that BigGreen features 276 central processing unit cores, 552 gigabytes of memory and more than 10 terabytes of storage. Eight NVidia Tesla graphics processing units with 448 cores each provide support for massively parallel computation, pushing BigGreen to roughly six Teraflops—or six trillion floating point operations per second—of theoretical peak computing power. A variety of scientific software packages are installed and available for use on the cluster, including COMSOL Multiphysics, Mathematica and CLC Genomics WorkBench.
For most people, it is enough to understand that BigGreen is powerful enough to allow simulations of black holes and gravitational waves. It can provide data support for sequencing of DNA at unprecedented speeds, and it can make possible the design of complex underground mine ventilation systems. The cluster can also accommodate molecular modeling, disaster simulations and gait analysis in the university’s Visualization Lab.
Fox added that even those disciplines not traditionally associated with bits and bytes, like the humanities, can benefit from the university’s new computing power.
“For example, a 3-D scan of Michelangelo’s statue ‘David’ contains billions of raw data points. Rendering all that data into a 3-D model would be nearly impossible on a desktop computer,” she said. “Using our high-performance computing capabilities, a student or professor could run that same data and produce the model in just a fraction of the time. It will literally change the way we work and do research at Marshall University.”
Dr. John M. Maher, vice president for research, said Marshall’s new computing capabilities will be transformational.
“As our research enterprise continues to grow, it will become increasingly important for us to be able to recruit and retain the brightest students and most-successful faculty researchers,” he said. “To compete, we need to offer top-notch facilities and laboratories. This computing cluster is one of the most powerful available at any institution in the entire region, so it truly levels the field as we work to advance research and research collaborations, win grant funding and enhance students’ classroom experiences.”
BigGreen was made possible in part by a National Science Foundation grant that funds “Cyberinfrastructure for Transformational Scientific Discovery in West Virginia and Arkansas (CI-TRAIN),” a partnership among eight higher education institutions in West Virginia and Arkansas.
For more information about high-performance computing at Marshall or obtaining an account on the university’s HPC cluster, visit http://www.marshall.edu/it/research.
Contact: Ginny Painter, Communications Director, Marshall University Research Corporation, 304.746.1964