Using multi core computers for implementing cellular automata systems

Abstract

A concept of cellular automata system (CA-system) is introduced as a model of comp[lex phenomena in which several interacting species are involved. CA system suggests a common work of several CA where each processes its own cellular array using in its transition rules cell states of others CA of the system. Taking into account that multi core computers with shared memory are nowadays widely used, a temptation to accelerate the computation by allocating each CA of the system onto one of computer cores is quite natural. Hence, it would be helpful to know what speedup can be obtained by such a parallelization. The paper is aimed to get an answer to this question by determining the conditions, when multicore parallel implementation of CA systems is expedient and correct, and develop the parallelization algorithms for typical CA systems. The results are illustrated by simulation experiments. © 2011 Springer-Verlag Berlin Heidelberg.