Minitrack on computer design: a new grand challenge

Grand challenge computational applications have been identified in a number of scientific and engineering disciplines, and significant attention has been placed on them in part due to the national High-Performance Computing end Communications (HPCC) Initiative. Computational tools are extensively used in designing computers, yet this has not been recognised as a grand challenge comparable in significance to modeling proteins, turbulence and the environment, to name three accepted grand challenges. The increasing recognition of the cost of the time required to bring a new design to market will place new demands on modeling capability. A number of opportunities for improved analysis and design are outlined. Computer manufacturing is a major industrial consumer of computation for design. As competition increases, the design cycle will continue to shorten, requiring significant increases in computer power to support the design process. Scenarios of this sort in other industries prompted the establishment of the HPCC Initiative and the formation of Grand Challenge Application Groups to tackle prototypical problems in science and engineering. The objective of these groups is two-fold. There is the need to develop algorithms and related software suitable for large-scale parallel computation. In addition, there is the need to understand types of computation that can provide useful insight. The author addresses both of these aspects.<>