In contrast to serial processing, typically used in personal computers as well as in many larger machines, parallel processing allows a task to be broken down into many subtasks that are performed simultaneously on multiple processors sometimes numbering in the hundreds or even the thousands. Writing programs for these powerful systems is a major challenge for software and computer engineers.

In light of this year's modest performance gains, which contrast with last year's spectacular improvements, the size of the cash awards was reduced.

Three entries were awarded cash prizes:

-- Panayotis Skordos of the Massachusetts Institute of Technology was recognized with a $500 award in the price/performance category for his modeling of air flow in flue pipes. He achieved 3.6 billion floating-point operations per second (gigaflops) per million dollars on a cluster of 20 Hewlett-Packard workstations.

-- Masahiro Yoshida, Masahiro Fukuda, and Takashi Nakamura of the National Aerospace Laboratory in Japan, Atushi Nakamura of Yamagata University, and Shinji Hioki of Hiroshima University submitted a quantum chromodynamics simulation run on 128 processors of the Numerical Wind Tunnel. They reported a computation rate of 179 gigaflops, for which they received $500 in the performance category.

-- Junichiro Makino and Makoto Taiji of the University of Tokyo presented a simulation of over 100,000 stars at 112 gigaflops on a machine with 288 computation engines having a peak performance of 176 gigaflops. Their entry earned a $250 award for special-purpose machines.

No prize was awarded in the compiler speedup category.

The judges were Alan Karp of Hewlett-Packard Laboratories, Michael Heath of the University of Illinois, and Al Geist of Oak Ridge National Laboratory. "Although the performance gains were not as impressive as last year's, one trend did continue: The highest performance came from the big iron," said Karp, who also serves as Gordon Bell Prize administrator. "In fact, only one submission used a cluster of workstations. This entry won an award for price/performance, but it lagged by a factor of over 1,000 the best raw performance reported this year."

The prizes are sponsored by Gordon Bell, a former National Science Foundation division director, now an independent consultant. They are intended to spur the transition of parallel processing from computer science research to useful applications. Entries are coordinated by Computer magazine, a publication of the IEEE Computer Society. The January issue of Computer will include a special feature describing the work of the Gordon Bell Prize winners.