Implementation of large-scale cellular automata models on multi-core computers and clusters
Olga L. Bandman · 2013
The purpose of the paper is to develop an approach to allocation a large-scale CA-model onto processing units of a multi-processor system. Two methods are proposed that are based on the principle of maximum similarity between the parallel CA algorithm structure and computer architecture: (1) domain decomposition with implementation on a cluster, if CA model is a sequential CA composition, and (2) allocation of each CA on a processor of a multicore computer with shared memory, if the CA model is a parallel CA composition. Each proposed method is illustrated by an example of parallel implementation of a complex CA-model on multi-core computer or cluster.1