Institute for CyberScience to expand computing power for research

University Park, Pa. – The computer that once filled a room is now on your desktop, the memory is on your key chain, but there are still projects that require computing and storage capacity far beyond a room full of desktops. The Penn State Institute for CyberScience (ICS@PSU) aims to advance discovery through computing and information science.

 "We are looking at scalable computing and the transformation of data to knowledge," says Padma Raghavan, Institute director and professor of computer science and engineering. "How do we solve problems that are bigger and more complex?"

The problems targeted by ICS, in collaboration with other Penn State institutes, include infectious diseases, global energy needs, personalized therapies and materials design. Some of these problems may eventually be solved with virtual worlds where viruses can emerge, multiply and evolve; programs for real-time personalized therapies; and tailored nanoscale materials that  can be designed and created. One area of critical importance is in data management and mining, which can aid in model discovery and simulation-based data collection from observatories viewing watersheds, ocean systems or any biologically complex population. The institute targets high-impact, large-scale  research tied to improving life on Earth through the environment, biological sciences and innovation.

"Simply making something bigger does not necessarily work," said Raghavan. "The algorithms are complex, and scaling does not always work in the parallel computing environment from multicore processors to petascale systems." Petascale computing allows speeds of more than a quadrillion floating point operations per second.

Raghavan's expertise is in parallel scientific computing and computational science. Her research includes algorithms for scalable sparse graph and matrix computations including partitioning, mining and fast solvers, energy-aware performance scaling in chip multicore multiprocessor petascale regimes and systems for multiscale modeling, simulation and knowledge extraction.

In the area of complex networking, ICS@psu will involve everything from theory to systems design and processes, networked searching and retrieval and scalable computing, and networked sensors and data fusion. When it comes to models, the possibilities include simulations, continuous models, models of models and real-time systems.

The ICS@psu is the coordinating unit for multidisciplinary computational science research at Penn State and organized under the Office of the senior vice president for Research. The institute is affiliated with the Colleges of Agricultural Sciences, Business, Earth and Mineral Sciences, Engineering, Information Sciences and Technology and the Liberal Arts, the Eberly College of Science, Great Valley School of Graduate and Professional Studies and the University Libraries. 

Also associated with the institute are major University units including the Applied Research Laboratory, Huck Institutes of the Life Sciences, Materials Research Institute, Social Science Research Institute and Penn State Institute of Energy and the Environment. Information Technology Services, which provides the information technology tools and infrastructure necessary to carry out the University's mission, is also a key player in the Institute.

More information about ICS@psu is at http://www.ics.psu.edu on the Internet.