Policy Decisions: Supercomputing's Federal Road MapPolicy Decisions: Supercomputing's Federal Road Map

The Defense Advanced Research Projects Agency's High Productivity Computing Systems program, begun last year, aims to supplement high-performance computing research and development, according to Darpa's Web site, that's stuck between late '80s technology and "the promise of quantum computing"--an experimental approach that could take years to yield practical results.

The Integrated High-End Computing Program, a cooperation among several federal agencies, aims to consolidate the National Nuclear Security Administration's ASCI program with other Defense Department and National Security Agency R&D programs. An April report to Congress by the National Security Agency, Darpa, and other agencies proposes spending $250 million to $390 million per year to upgrade supercomputing applications in weapons development, nuclear stockpile stewardship, cryptographic analysis, weather forecasting, biology, and 10 other fields.

The White House Office of Science and Technology Policy's High-End Computing Revitalization Task Force held its latest meeting last week in Washington to create a five-year road map, starting in 2005, for federal high-productivity computing investments. A report to help guide the federal 2005 budget is due in August.

The National Science Foundation, which advises the government on policy issues, has assembled a "Future of Supercomputing" committee; members include the University of Tennessee's Jack Dongarra, Microsoft researcher Butler Lampson, and Steven Wallach, an executive at Chiaro Networks who worked at Data General in 1970s. The NSF in February proposed spending more than $1 billion a year to modernize its high-performance computing infrastructure.