Long range research investments for a semiconductor systems future

Revolutionary technological progress since 1970 in the electronic and semiconductor industries has drawn heavily on early sustained programs of basis research and graduate training in the solid-state sciences. The potential for further progress through higher density, lower cost-per-function ICs is widely held to hinge upon advances in systems architecture and software competence, both for commercial applications and to enable the complex design tasks for 10⁶active elements or more per chip. Understandably, we might expect correspondingly to have a substantial research and graduate training base in systems architecture, design, and software. But a systems research investment comparable to that in solid-state science is not readily discernible-certainly not in the U.S. where the solid-state foundation was laid. Indeed, basic research has been suffering from inflation and decreased emphasis. University research equipment is generally obsolescent, and few universities have had the equipment and faculty with which to launch frontier research on systems issues. Will the several new university-industry initiatives in systems and microelectronics help respond to this need, or will they influence universities toward short-term, rather than long-range, perspectives? Can industrial research, always under short-term pressures, take a longer view after recent years of inflationary and high-interest rate discounting of the research dollar? Are multi-company joint research ventures an answer? This address will assess these questions and consider steps to stimulate systems research and graduate training as well as to mitigate a general disinvestment in long range research.