Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
FUTURE COMPUTING ENVIRONMENTS FOR MICROSIMULATION MODELING 186 Such an assumption is implicitly biased downward by assuming that advances will occur within the structure of current hardware architectures. Genuine innovations or breakthroughs in computer architecture will render any rate of technical progress based on classic von Neumann architecture too conservative. In particular, any substantial breakthrough in the use of parallel processing for general-purpose computing tasks will allow additional processor power to be added to a system at an incremental cost related to the cost of the processing chip and interface rather than the entire system, so that substantial gains in performance could then be achieved at more modest increases in price. For the purposes of this discussion, advances in computing technology are partitioned into three major hardware categories: (1) processors, primary memory and secondary memory; (2) more general architectural considerations; and (3) software. Other categories, such as advances in computer architecture or in cooperative computer networking (itself a new form of architecture), would have a major impact on the productivity of both the general computing environment and the desktop environment; however, neither the amount nor the timing of any return from research and development in these areas can be predicted accurately. While the 10â30 percent technical dividend per year is a part of the conventional wisdom of the industry, more precise measurement of the actual dividend is more difficult, and information that would yield such a measurement is more difficult to obtain than might be thought. There are several types of sources that yield information regarding the rate of technical progress and therefore the expected characteristics of future computing environments: informed industry observers, announcement and entry information for specific products, economic studies of past and current industry performance, and information and studies from manufacturers and other industry groups. Informed Industry Observers Informed industry observers provide probably the softest information available, but their judgment may well be based on much technical experience and on information that is not yet publicly known with any degree of certainty. Industry observers vary a great deal in their perceptiveness, and their statements should be received in the context of their track records. Recent statements bearing on the medium-term future of computing include the following: â¢ President Bush's science advisor, D.Allan Bromley, has proposed a $2 billion federal program for strengthening the computing industry that could lead to increases in speed of 100â1,000 times in supercomputers in 10 years. Part of the rationale for such an investment is that today's supercomputers are