Parallel discrete event simulation for DEVS cellular models using a GPU

Abstract

The discrete event systems specification (DEVS) simulation has been studied to analyze complex homogeneous systems which is represented by the cellular models. In the simulation of large-scale DEVS cellular model, it requires a high-performance simulation technique such as parallel simulation. This paper presents a parallel discrete event simulation (PDES) environment using a graphics processing unit (GPU) to simulate DEVS cellular models. In the proposed PDES environment, the DEVS cellular models and their corresponding simulators are mapped to the GPU. Due to the features of the GPU, two parts of the traditional DEVS simulation algorithm are revised. The first is about managing the event list, and the second is about routing output events. To verify the efficiency of the proposed environment, we designed and experimented the case study for a fire-spreading phenomenon, and analyzed the experimental results.