Parallel-Computing Symposium Turns 25

First held in a basement, this year's event expects 600 attendees

7 April 2011

For decades, computing was a game of leapfrog: Processors got faster and able to handle more and more program instructions per millisecond. Meanwhile, programs got longer and more complex, giving processors a lot of instructions to chew through. But the faster the processor, the hotter it runs, decreasing power efficiency and increasing the risk of meltdown. The solution: Let multiple processors—running at different speeds—work on different parts of the program simultaneously. Those processors might be multiple cores on a single CPU chip, multiple CPUs within a computer, or networks of standalone computers known as distributed processing.

"Parallel processing was once something exotic used on supercomputers," says Viktor Prasanna, steering committee cochair and cofounder of the IEEE International Parallel & Distributed Processing Symposium (IPDPS). "Today it's everywhere—on your desktop and on your laptop. It's become mainstream."

This year's IPDPS, to be held from 16 to 20 May in Anchorage, Alaska, expects about 600 participants and 100 papers. The papers cover four main subjects: parallel and distributed algorithms, applications, architecture, and software. The five-day program includes panels, tutorials, keynote speakers, poster sessions for doctoral students, and about 20 workshops.

"We consider IPDPS the premier event in its field," says IEEE Fellow David A. Bader, a steering committee member and a computational science and engineering professor at Georgia Tech. "Its attendance is uniformly high, about twice that of other academic conferences in parallel processing, and very stable. It attracts international researchers from top academic institutions, national labs, and industry, and attendees present and learn of the research results from projects supported by sponsors such as the National Science Foundation and the U.S. Department of Energy.

This year the symposium is celebrating its 25th year. When it was first held in 1987, it was a local affair known as the Orange County Parallel Processing Conference, and met in a basement on the Cal State Fullerton campus. Its sponsor was the IEEE Orange County section. Today it is sponsored by the IEEE Computer Society's Technical Committee on Parallel Processing in cooperation with the Association for Computing Machinery.

Special programs and events to celebrate the conference's 25th year have been planned. The two most significant, according to Sally Jelinek Westrom, production chair and another cofounder of the conference, are a pair of panels, "Looking Back," moderated by IEEE Fellow Yves Robert of the Ecole Normale Supérieure de Lyon, in France, and "What's Ahead," moderated by IEEE Fellow Per Stenström of the Chalmers University of Technology, in Gothenburg, Sweden.

Also planned is a new, retrospective award for papers presented in prior years that have proven especially insightful or to have had great impact, to be nominated by conference attendees. There is an IPDPS history exhibit as well, including photographs and posters from prior years, copies of past conference proceedings, and souvenir T-shirts. If all works out, says Bader—who also chairs the anniversary committee—there will be a DVD incorporating all previous proceedings, a memory book, and videos featuring luminaries in the field.

"It's sort of a family reunion," Bader says. "But an extended family, bringing together people from different areas—algorithms, applications, architectures, and software. It's rare to have such a crosscutting meeting, where maybe an applications expert can influence an architecture designer, or an algorithm designer can learn from a hardware designer how best to exploit that hardware's capabilities."

Over the years, Bader adds, many IPDPS papers have been widely cited, with impact beyond the computing field: "We've covered some of the first work in computational biology, computational finance, and other emerging areas that parallel computing has affected."

Learn More