The widespread use of massively parallel computing as a scientific and industrial tool has been impeded by technical problems associated with available hardware and software. U.S. industries do not have the applications and necessary software tools and languages to facilitate use of massively parallel computing environments. To address this need, the Industrial Computing Initiative (ICI), which is part of the $66-million High-Performance Parallel Processing Project, is developing a set of efficient applications that can realistically simulate complex problems and are written in a way that allows them to run on massively parallel hardware from various vendors. The ICI involves more than 40 scientists from LLNL and Los Alamos National Laboratory, six specialists in parallelization from Cray Research, two specialists from Thinking Machines Corporation, and nearly a score of industrial scientists. The delivery of a set of efficient parallel applications, serving as guideposts for subsequent work, can help U.S. industry compete more effectively in the global marketplace. The new applications, together with improved methods and tools, can reduce product cycle times, produce higher-quality products, and speed the development of more efficient industrial processes. At the same time, the new computational advances in massively parallel computing technology can be reintegrated into ongoing programs at the national laboratories to serve the missions defined by the DOE and to address national needs.
Materials scientists at Smith and Nephew Richards, a firm that makes artificial hip joints, decided that a femoral head made of zirconium, which can be oxidized to acquire a surface layer of zirconia, a ceramic, would combine the lower production costs of metal heads with the much greater longevity of ceramic heads. Smith and Nephew Richards machinists, however, encountered difficulties in grinding zirconium heads to the needed size, contour, and smoothness. Because of LLNL's expertise in precision engineering, Smith and Nephew Richards and the Lab signed a research agreement under the National Machine Tool Partnership to help them solve their problem. We soon established that the preferred technique for use with zirconium was not grinding, but single-point turning. We thus had an opportunity to turn decades of experience acquired through weapons work to advances in medical technology. Identifying an appropriate turning machine that the Lab and Smith and Nephew Richards both have, we turned some pieces on it to their criteria. We delivered to Smith and Nephew Richards a turnkey package: a custom workpiece holder for high-precision single-point turning and the software for operating the compensated tool path. During the course of this work, Smith and Nephew Richards and the Laboratory discovered that the Lab's inspection capabilities for measuring the size, contour, and smoothness of the femoral heads were more precise than theirs, and so we undertook and completed a secondary task of helping establish the exact limits of their fabricating and inspection equipment.
(pdf file, 524K)
and LLNL Disclaimers