Code Optimization for Solving Large 3D EMP Problems

Abstract

This paper describes techniques for implementing and optimizing large (50×50×50) three-dimensional finite-difference EMP codes. A number of procedures for speeding execution and minimizing memory requirements are presented and evaluated. As an example, we illustrate numerical results obtained when a B52 is modeled within a 50×43×59 mesh and illuminated by a nominal EMP simulation.