This article is available in PDF format as it originally appeared in Projects in Scientific Computing, 2001.

As this publication goes to press, the PSC computer room at Westinghouse Energy Center in Monroeville, Pa. is a busy scene, the focus of activity for a team of engineers from PSC and Compaq Computer Corporation. Systematically, cable-by-cable, box-by-box, test-by-test, they are bringing into being an unprecedented computing system. When fully installed (Terascale Supercomputing System Installed at PSC, Oct. 1, 2001), it will be the most powerful system in the world available for public research.

With 3,000 Compaq Alpha EV 68 microprocessors, housed in 750 four-processor AlphaServer systems, the Terascale Computing System will provide six teraflops (six trillion calculations per second) of computational capability to U.S. engineers and scientists nationwide. While several other terascale systems are available for classified research, the PSC system will be the most capable to date provided as an open resource for scientists attacking a broad range of problems.

As it becomes a productive research tool, first with a "friendly user" testing period and then, by early 2002, integrated into the NSF's Partnerships for Advanced Computational Infrastructure (PACI) program, the TCS will be used to advance knowledge in many fields. These include earthquake modeling, storm-scale weather forecasting, global climate change, and protein genomics modeling that is integral to the development of new drug therapies.

Developed and implemented at PSC, in close collaboration with Compaq, pursuant to a three-year $45 million award to PSC from the NSF-PACI program, the TCS employs the very fast EV68 chip. This latest evolution of Alpha microchip technology, widely used in both commercial and scientific computing for 10 years, has peak floating-point capability of two gigaflops (two billion calculations per second) and boasts exceptional data-transfer rates.

Along with six teraflops of processing power, the TCS features 3.0 terabytes of memory, high-bandwidth, low-latency interconnections and remarkable capabilities for large-scale data handling, including the ability to write the entire memory to disk in under 40 seconds. This extremely short system-write time, developed through PSC software engineering, is critical to efficient checkpointing, needed to preserve research data in the event of component failure.

A prototype 256-processor TCS became operational in October 2000, well ahead of schedule. Since April, it's been an allocated component of the PACI program, demonstrating reliability of the TCS concept. Several projects featured in Projects in Scientific Computing (In Search of Planetary Pancakes, Touchy Proteins, Electronic Nirvana and Flutter in the Sky) represent significant scientific results made possible by this early-version TCS, which is just the beginning of fruitful relationships between the TCS and science.