PHILADELPHIA The University of Pennsylvania has installed an IBM supercomputer to help solve some of humanity most perplexing problems, such as research on gene therapy, AIDS accessory proteins and atmospheric ozone depletion. The RS/6000 SP supercomputer is twice as powerful as the IBM Deep Blue supercomputer that defeated world chess champion Garry Kasparov.

"This system represents a significant computational resource," said Michael Klein, the Hepburn Professor of Physical Science and director of the School of Arts and Sciences' Center for Molecular Modeling. "The supercomputers that are currently available at the national supercomputing centers are only about an order of magnitude larger than the University new machine, but they must be shared among many groups throughout the country. Having a machine of our own of this size means that we will now more quickly be able to tackle even more complicated problems than before."

Rod Adkins, general manager of IBM RS/6000 brand, said, "The University of Pennsylvania has long been at the forefront of history-making milestones in computing. That began more than 50 years ago, with its development of ENIAC, the world first programmable computer. Now, the university is deploying the computing prowess of the RS/6000 SP to help achieve breakthroughs that potentially could impact on the treatment of diseases and help solve environmental problems."

The 32-way RS/6000 SP at the University of Pennsylvania's Laboratory for Research on the Structure of Matter consists of 16 dual processor units, featuring IBM POWER3 computer chip. The POWER3 chip can perform up to two billion operations per second and is the successor to the POWER2 chip in the Deep Blue supercomputer.

The system can store more than a half terabyte of data, which is necessary to run large-scale simulations of molecules found in protein structures as well as new materials. That is enough disk capacity to store all the information contained in half a million unabridged dictionaries.

The supercomputer was purchased by the University of Pennsylvania, in part, with a Shared University Research (SUR) grant from IBM.