A handful of engineers at a lab in Poughkeepsie, N.Y., have assembled what they expect will become--at least for a while--the world's most powerful computer. The IBM Roadrunner likely will go down in history as the first computer to consistently crank out 1 petaflops--a quadrillion floating-point operations per second.

But the true significance of the IBM supercomputer--and many similar efforts gearing up around the world--might not be the milestone of cracking the petaflops barrier. The bigger impact may lie in what the creators and users of these powerful machines are learning about science and how to harness parallel computing.

Leading-edge supercomputers are making great strides in reducing cost and power consumption. But there is a growing gap between their theoretical performance and the amount of real work they can accomplish. The gap is due to the growing complexity of programming systems with many processors.

Making Roadrunner beep
A group of 15 to 20 engineers working at an IBM lab last week were expected to finish slotting together some 3,240 compute modules and plugging in cables to link them on a two-tier Infiniband network. Each Roadrunner module consists of two AMD dual-core Opteron processors and four PowerXCell 8i CPUs--65-nanometer versions of the Cell processor developed by IBM, Sony and Toshiba for the Play- station. The completed system will consume 4 megawatts and fit into a 6,000-square-foot room.

"We have been manufacturing and populating this system for the last several months, and the last pieces should be in place today; then we will make a run for the roses with the Linpack test," Don Grice, chief engineer of the Roadrunner project at IBM, said in an interview.

The Linpack test is a standard benchmark of theoretical performance used as a measure for entry onto the Top 500 list of the world's most powerful computers.

The Top 500 list is published in June and November each year around the major supercomputing conferences. Typically, April 15 is the cutoff date for the June rankings, but list administrators have granted an extension to the Roadrunner team.

"We have made a major exception for them until May 15 and can provide a few more days beyond that," said Jack Dongarra, a professor at the University of Tennessee who helps manage the Top 500 list. "It's a big deal if they can do it, so we want to give them every possibility."

Grice said, "We are pushing to be on Jack's list and hope we have a few weeks left. I'll be at the June supercomputer conference in Dresden [Germany] with whatever numbers we have."

Once fully tested by IBM, the system will be packed up and shipped to Los Alamos National Laboratory in New Mexico, where it will be used to run classified physics experiments as part of the U.S. nuclear missile program.

Doing real work
Breaking the petaflops barrier is "certainly a milestone. What it really accomplishes is giving scientists an ability to do more science than they could before," said Buddy Bland, project director for a major supercomputer center at Oak Ridge National Laboratory, which expects to install its own petaflops system this fall.

The lab has a Cray XT-4 that can deliver 263 teraflops and can be used by any scientist for unclassified scientific work. Its petaflops system will be a follow-on to the XT-4 that raises the number of AMD Opteron cores from 31,000 to 100,000.

The system has run global climate simulations that fed into the work of Nobel Peace Prize winner Al Gore. It is also being used to help design a fusion reactor being built in France.

"Our mission is to deliver the largest systems possible to the unclassified science community to solve some of the most challenging problems facing the world today," said Bland.

He noted that the teraflops systems of a generation ago could model in one day five years of climate change. His current system can model 40 years of change in a day.

"Getting applications to scale is our biggest challenge," he said, looking ahead to his 100,000-core petaflops system.

Thus, Oak Ridge researchers have been testing new parallel programming languages, including X-10 from IBM, Chapel from Cray and Fortress from Sun. "It turns out you get just as much advancement from better software and algorithms as you do from better hardware," Bland said.

A NASA supercomputer specialist agreed that a deeper problem lurks below the excitement of breaking the petaflops barrier. "Roadrunner is very impressive, but the bottom line is how much real work you can do on the machine," said Bill Thigpen, chief of the engineering branch at NASA's advanced supercomputing division in the Ames Research Center. "One of the challenges is being able to get the available work out of the theoretical performance peak."