As the twentieth century draws to a close, it is becoming clear that our Nation's prosperity and security in future decades will rely heavily on its productivity and competitiveness in an increasingly aggressive international market place. Crucial to leadership in this rapidly changing global economy are its capabilities in developing, manufacturing, and applying high-performance computing and information management technologies. Yet, although the demand for R&Dinvestment to meet this challenge has never been greater, a combination of tightening constraints on available resources and the accelerating pace of technical innovation has created a climate of exceptionally hard choices for researchers and policy makers alike. Complicating the task of setting research directions is that major objectives of the future may depend on much earlier work in creating the necessary enabling technologies and developing essential capabilities in design and production. The extended lead times from research to delivered systems that result make critical early and active planning to (1) establish long-term national technical goals, (2) identify enabling technologies for which near-term R&Dis critical, (3) determine research agendas to meet these needs, and (4) prepare for support of initiatives in selected directions. It is no longer feasible to engage in reactionary or incremental strategies to R&Dplanning if the U.S. is to sustain dominance in future high technology. Only long-term planning cycles begun now will ensure a strong U.S. position in strategic technologies decades hence.
This report captures the findings of an historical meeting of key technical leaders in high-performance computing convened to set the trajectory of technology investment and innovation over the next two decades to a target of sustainable PetaFLOPS performance capability. No goal envisioned will be more challenging, demand greater coordination and collaboration among all sectors of the high-performance computing community, or more strongly promote ultimate U.S. leadership in all facets of computing into the next century.