Abstract
Although the Finite Difference Time- Domain (FDTD) method is over 25 years old, its use for the analysis of crosstalk, propagation delay, and pulse distortion of signals on interconnects in high-speed integrated circuits is relatively recent. With the development of faster computers, the use of FDTD has become more widespread because it has shown itself to be an efficient tool for the full-wave modeling of complicated microstrip circuit interconnects. In this paper, we address the parallelization of the FDTD method on multi-dimensional processor meshes mapped onto a hypercube multiprocessor. The use of parallelization is a natural step for the modeling of circuits because of the tremendous computational requirements to simulate even the simplest interconnect geometries. Experimental results on the Intel iPSC/860 hypercube are given which demonstrate a linear speedup in computation time.
Get full access to this article
View all access options for this article.
