Foreword from the Directors
Michael Levine and Ralph Roskies
PSC co-scientific directors.
This has been a year of achievement for the Pittsburgh Supercomputing Center. As we have done many times over the past 24 years, we are making two new supercomputing systems available to the national research community. Blacklight (see p. 4) is the world’s largest shared-memory computing system. Anton (p. 12) is the world’s most effective system for simulation of proteins and other biomolecules. Furthermore, PSC’s network group, led by Wendy Huntoon, received two major grants from the National Science Foundation (p. 14) and helped to lead a Pennsylvania coalition in securing $100 million in federal stimulus funds to create a statewide broadband infrastructure.
Blacklight’s unprecedented technology will make possible advances in many fields. One of them is epidemiological modeling. Working through the University of Pittsburgh MIDAS Center of Excellence, Shawn Brown and colleagues (p. 20) showed the ability of agent-based modeling to ask and answer questions about the spread of infectious disease in a timely fashion during last year’s H1N1 epidemic. With Blacklight, it now becomes feasible for the first time to do ABM modeling on a national or international scale.
Shared memory also helped with a major step forward in the innovative work (p. 24) of Tom Mitchell and colleagues at Carnegie Mellon. Their studies in machine learning made national news in 2009, and Indra Rustandi’s algorithmic breakthrough — facilitated by PSC consulting, software and hardware — to enable integrated datasets has significantly improved the accuracy of their predictive model.
Atrial fibrillation doesn’t have the almost immediate fatal consequences of ventricular fibrillation, but it’s a serious heart condition that increases risk for stroke and heart failure. With a new approach to low-energy defibrillation (p. 28), Elizabeth Cherry and colleagues have shown the feasibility of treating atrial fibrillation much more effectively than is currently possible.
Actin is one of the most important proteins in the body, involved in metabolic pathways that when disrupted can lead to cancer. The computational work of Greg Voth and Jim Pfaendtner (p. 32), in close collaboration with their laboratory partners, shows how slight structural changes in actin can have profound implications, insights that give new understanding to experimental findings.
Terpenes are fascinating hydrocarbon compounds that have provided biomedicine with one of the most effective cancer-fighting drugs so far developed, taxol. Dean Tantillo is doing quantum chemistry studies at PSC (p. 36) that could lead to much faster, cheaper ways to produce other useful terpene-derived compounds.
Understanding the universe, how it got to be the way it is, is an ongoing project of physics, and supercomputing made a major contribution with the work of University of Washington physicists in modeling dwarf galaxies (p. 40). With improvements in their software and the availability of new powerful TeraGrid systems, they’ve solved one of the major problems in the reigning cold dark matter model of cosmological evolution. In astrophysics too, Blacklight’s shared memory will make possible new kinds of data-analysis, as this work has demonstrated (p. 43).
We’re pleased that PSC resources could enable these advances, and also that we can be a vital resource for research and education in Pennsylvania (p. 6). Along with new scientific insights, this publication also highlights (pp. 9-11) our important work to help educate the next generation of scientists and science-literate citizens.
None of this would have been possible without PSC’s staff, an unparalleled assemblage of talent and experience in high-performance computing. We are also grateful for support from the National Science Foundation, the U.S. Department of Energy, the National Institutes of Health, the Commonwealth of Pennsylvania and many others.