Solving bigger problems-by decreasing the operation count and increasing the computation bandwidth

摘要

Illustrates the computational complexity of modeling large (in wavelengths) electromagnetic problems and suggests some alternatives for reducing the cost of both the basic model computation itself (by reducing the order of the frequency dependence) as well as the number of frequency samples needed for defining the wideband response (by more efficiently representing the frequency behavior). The massively parallel computers becoming available, whose computation bandwidths of 10/sup 16/ hour and 10/sup 18/ hour are being projected by 1995 and 2000 respectively, and their potential impact on computational electromagnetics (CEM) are also considered.