By Neal Singer

Thursday, October 13, 2016

Red Storm architect honored

Sandia emeritus director Bill Camp wins Seymour Cray Computer Engineering Award

Insights into what happens in explosions
Bill Camp

When Bill Camp learned by phone last month that he had won the IEEE’s top-of-the-line Seymour Cray Computer Engineering Award, the Sandia emeritus director said that compared with the giants who had won the award in previous years, he felt “like the kid who had snuck into the movie theatre.”  However, no one else does.

“This is the top career award in high-performance computing architecture. I’m delighted Bill is receiving this recognition, which is richly deserved,” says Rob Leland, Science and Technology vice president (1000).

Bruce Hendrickson, director of the Center for Computing Research (1400), adds, “This is one of the IEEE's most prestigious awards, and is largely based upon Bill's leadership in the development of Red Storm.”

Red Storm supercomputer was 'crowning achievement

The Red Storm supercomputer — specifically cited in the IEEE announcement as a crowning achievement of Bill’s “visionary leadership” — saved the bacon of the then-foundering Cray Inc., according to its current president and chief executive officer Pete Ungaro.

“Without Red Storm, I wouldn’t be here in front of you today,” Ungaro told a Sandia audience in 2012. The machine went on to serve as intellectual stud to successive generations of the best-selling Cray XT supercomputer line, he says.

“Red Storm is arguably the most successful general-purpose supercomputer in high-end computing,” says Bill, who achieved support for the vision first proposed by himself and Sandia technical adept Jim Tomkins (retired), who provided much of the detailed design work. “Red Storm is over, but its influence is not.”

But there’s much more.

In 1998, Bill led development of Sandia’s Massively Parallel Computing Research Lab, which pioneered using many processors working in concert to solve large problems in science and engineering. In its first five years, the lab won the inaugural Gordon Bell Prize, several international awards, and eight R&D100 Awards — all for pioneering highly scalable applications, algorithms, and methods as well as scalable systems software and hardware.  The researchers working in the lab received more than 30 patents in the technique then becoming known as massively parallel-processing. Their achievment included standing up Sandia’s Paragon, the first massively parallel supercomputer to lead the world’s bi-annual Top 500 list in computational speed.

Led development of first teraflop computer

Later, as director of NNSA/DOE’s Accelerated Strategic Computing Initiative, Bill led development of the world’s first teraflop computer, ASCI Red, which led the Top 500 list for an unprecedented three and a half years.

Bill’s work frequently demonstrated computation to be a major pillar of science and engineering. He  co-founded Sandia’s programs in cognitive science and biotechnology. He and several Sandia teams  provided computational expertise to Celera Genomics in its successful sequencing of the human genome, and they developed the family of light-weight kernel operating systemsSUNMOS, Cougar, Puma, Catamount, and Kitten — that foreshadowed operating systems in the high-end computing industry. They also created the world’s first Linux cluster-based supercomputer, and produced the first tera-scale cluster-based supercomputing environment (CPlant).

Bill has published more than 60 journal articles in physics, materials, engineering, and computing  and made keynote presentations at international conferences. He is a Fellow of the American Physical Society, a member of IEEE, and was co-founder and second chairman of the International Union of Pure and Applied Physics’ International Commission on Computational Physics. He has served on the editorial boards of several journals and is on the editorial board of Concurrency, a software journal. He holds a doctorate from Cornell in theoretical and computational physics.

Currently a consultant at Sandia and other labs, Bill’s recent projects include research on quantum annealing and Ising spin-glass systems as a type of quantum computer, optical interconnects for router-less supercomputing, technical approaches for neuromorphic computing, and intelligent stacked memories for processor-in-memory architectures.

Bill also consults on post-exascale supercomputing ideas beyond CMOS architectures. Exascale requires a degree of computational speed that DOE laboratories are only beginning to attempt.

Asked how he keeps up with so many projects, the retiree says, “I think as well as ever. Just more slowly.”

It’s just not obvious.

About the prize . . . and Seymour Cray

The Seymour Cray prize, widely regarded in the computer engineering community as the IEEE’s top award, consists of a crystal memento, an illuminated certificate, and a $10,000 honorarium, which will be presented to Bill at the SC2016 convention on in Salt Lake City, Utah, on Nov. 15. On Nov. 16 he is scheduled to present an invited talk there.

Seymour Cray enjoyed near-mythical stature for his brilliant innovations that kept his company’s supercomputers the fastest in the world for more than a decade. He died in a car wreck in 1997. The memorial award recognizes design, engineering, and intellectual leadership in creating innovative and successful HPC systems.