An algorithm for parallel discrete event simulation using common memory

Abstract

Most work on parallel discrete event simulation has been based on a distributed model of computation in which processes can only communicate through message passing. Here we study parallel discrete event simulation under a common memory model of computation. An algorithm for parallel discrete event simulation is developed based on the assumption that every process has direct access to the state of any other process. The objective is to avoid the high overhead associated with null messages and request messages in distributed algorithms. This algorithm is then compared to distributed synchronization algorithms.