Experiments with Single Core, Multi-core, and GPU Based Computation of Cellular Automata

Cellular automata are a well-known modeling formalism exploited in a wide range of application areas. In many of those, the complexity of models hampers a thorough analysis of the system under study. Therefore, efficient simulation algorithms are required. We present here a comparison of seven different simulation algorithms for cellular automata: the classical ldquofullrdquo simulator, the classical ldquodiscrete eventrdquo simulator, a threaded (multicore) variant of each of these, an adaptable threaded variant, and a GPU based algorithm with and without readback of calculated states. The comparison is done based on the M&S framework JAMES II by using a set of well-known models.