Scientific American
IEEE Spectrum
Programming pearls: perspective on performance
Communications of the ACM
High performance parallel architectures
Proceedings of the 1989 ACM/IEEE conference on Supercomputing
Hi-index | 0.00 |
High-performance computer technology is undergoing a period of unusually rapid change, and this paper attempts to describe the patterns of these changes in a systematic way. Pattern recognition is the basis of technology forecasting, and it is through technology forecasting that we obtain the anticipatory information that allows us to avoid problems and create opportunities. We will first identify the stages in which technological changes occur, and then define “change” as the first derivative of an information function that describes the state of a technology. We then explore the driving forces that cause three generic patterns of technological change: incremental, exponential, and logistic. Some areas in high-performance computer technology that are following these patterns are then identified, and a forecast is developed. Finally, some limits of high-performance computing are discussed.