IBM claims fastest supercomputer title--for now

Sure, 36.01 trillion calculations per second is pretty fast. But some say SGI could eventually beat that.

Stephen Shankland principal writer
Stephen Shankland has been a reporter at CNET since 1998 and writes about processors, digital photography, AI, quantum computing, computer science, materials science, supercomputers, drones, browsers, 3D printing, USB, and new computing technology in general. He has a soft spot in his heart for standards groups and I/O interfaces. His first big scoop was about radioactive cat poop.
Expertise processors, semiconductors, web browsers, quantum computing, supercomputers, AI, 3D printing, drones, computer science, physics, programming, materials science, USB, UWB, Android, digital photography, science Credentials
  • I've been covering the technology industry for 24 years and was a science writer for five years before that. I've got deep expertise in microprocessors, digital photography, computer hardware and software, internet standards, web technology, and other dee
Stephen Shankland
3 min read
IBM claimed first place Wednesday in a supercomputer speed competition with an expanded version of its Blue Gene/L, which beat out a Japanese machine from NEC. But a new Silicon Graphics machine has a chance at giving Big Blue a run for the title.

Researchers at the Blue Gene/L production facility in Rochester, Minn., clocked the machine at a sustained performance of 36.01 trillion calculations per second, or 36.01 teraflops. By comparison, since 2002, NEC's Earth Simulator has held the top spot on a ranking of the world's 500 fastest supercomputers while running at a speed of 35.9 teraflops.

"The U.S. supercomputing industry is still alive and well," Dave Turek, vice president of Deep Computing at IBM, said during a conference call Wednesday. IBM is building the Blue Gene/L for nuclear weapons lab Lawrence Livermore National Laboratory but has begun selling the system to other customers.

IBM's achievement was expected. When two Blue Gene/L machines made it into the Top500 list, Turek said there was a "good chance" Blue Gene/L would top the forthcoming November list, adding that "the trajectory of it is already clear."

There's a new contender on the horizon, though, a NASA system called Columbia that SGI is building with 10,240 Intel Itanium processors. Intel President Paul Otellini said in September the system will be finished in 2004 and will have a speed of 60 teraflops.

Another contender for a high rank is a system at the Virginia Polytechnic Institute, which hit the No. 3 spot in 2003 using 1,100 Apple G5 computers, each with dual 2GHz processors. An upgrade with the same number of slimmer Apple Xserve computers is now complete, said Virginia Tech spokeswoman Lynn Nystrom.

"The goal will be to remain in the top 10 (of the Top500) and the fastest at a university," Nystrom said. Although Apple sells Xserves only with 2GHz processors today, the Virginia Tech upgrade uses 2.3GHz chips, she said.

The flagship Blue Gene/L machine at Livermore isn't going to make its initial 2004 deadline, though. "The total delivery to Livermore will come roughly in May of 2005. One of the challenges to face any company when going down the path of building scalable systems is that you really do penetrate certain unknown areas," Turek said.

Although the 64-rack Blue Gene/L that IBM is building for the Livermore lab is the largest, the company also has sold a six-rack system to the Astron Dutch radio telescope project, a one-rack system to Argonne National Laboratory and a four-rack system to Japan's National Institute of Advanced Industrial Science and Technology.

Don Dossa, a computational physicist at the Livermore lab, said in an earlier interview that Blue Gene/L is well adapted for some calculations but not for others.

For example, it's appropriate for simulating materials in a relatively unchanging environment, such as a solid that's cracking under pressure. But when it comes to the nuclear weapons simulations at the heart of the lab's mission, the 512MB of memory in each Blue Gene/L computing node isn't enough to deal with the broad range of radiation, temperature and pressure conditions, Dossa said.

For those nuclear weapons simulations, IBM is building a system called ASCI Purple that uses a smaller number of more powerful computers that rely on IBM's Power5 processor.

Blue Gene/L, when finished, will have about 65,000 dual-core processors.