U.S. lags in supercomputing, experts say

Department of Energy

The United States has significantly fallen behind Japan in supercomputer technology and must continue an interagency approach to development to close the gap, according to testimony before the House Science Committee July 16.

"We're not at a point of crisis, but we are at a pivotal point when we need to make critical decisions," said House Science Committee Chairman Sherwood Boehlert (R-N.Y.). "When we hear that the U.S. may be losing its lead in supercomputing, that Japan now has the fastest supercomputer, that the U.S. may be returning to a time when our top scientists didn't have access to the best machines, that our government may have too fragmented a supercomputing policy — well, those issues are a red flag that should capture the attention of all of us."

After leading the world's supercomputing efforts for years, the United States has found itself running a distant second since last April's start-up of Japan's Earth Simulator supercomputer, according to committee reports.

The Earth Simulator, used by the Japanese Marine Science and Technology Center to make predictions about Earth's climate and crust, operates at a sustained speed of 35 teraflops. Each teraflop is 1 trillion calculations per second.

The top U.S. supercomputer, the ASCI Q, located at Los Alamos National Laboratory, operates at roughly half that speed.

"Maintaining U.S. leadership requires a coordinated, concerted effort by the federal government," Boehlert said. "Let me stress that: a coordinated, concerted effort by the federal government."

Raymond L. Orbach, director of the Energy Department's Office of Science, emphasized to the committee that his office is rededicating itself to the pursuit of advanced supercomputing capability.

"We are working to develop the next generation of advanced scientific computational capability, a capability that supports economic competitiveness and America's scientific enterprise," Orbach said. "Advanced scientific computing is indispensable to DOE's missions."

In the FY 2003 budget request for high-end computing by agencies participating in the National Information Technology Research and Development program, DOE's Office of Science requested $137.8 million. The largest request came from the National Science Foundation, which requested $283.5 million.

However, NSF was the subject of some very pointed questions regarding the organization's commitment to high-end computing. The foundation is supporting the development of the Extensible Terascale Facility, a nationwide grid of machines that can be used for high-performance computing, and some observers have raised questions regarding NSF's commitment to supercomputer centers as opposed to its focus on grid computing.

Peter A. Freeman, assistant director of computer and information science and engineering at NSF, testified before the committee that the foundation is maintaining its stance toward supercomputing.

"NSF remains absolutely committed to providing young researchers the most advanced computing equipment available, and to sponsoring research that will help create future generations of computational infrastructure, including supercomputers," he said.

The committee urged a strong push toward interagency coordination in the efforts to improve supercomputing in the United States. The National Coordination Office is managing a High End Computing Revitalization Task Force, which wants to develop a road map for interagency research and high-end computing core technologies.

Aside from DOE, other agencies with important supercomputer applications include: the National Security Agency, which uses high-end computing for cryptanalysis; the National Nuclear Security Administration, which develops machines used to design and model nuclear weapons; the National Institutes of Health, which analyzes biological processes; the National Atmospheric and Oceanic Administration, which uses simulations for weather forecasting and climate-change modeling; and NASA, which has a variety of applications.