Abstract: Massive congestion is a very serious concern that can lead to disasters. The development of crowd dynamics models and crowd simulation tools is essential to better represent congestion aspects as well as to evaluate proposed solutions. However, model development normally involves increasing the mathematical complexity, imposing higher computational demands. This has led researchers to investigate solutions that can reduce or minimize the computational demands of such models where among them is the parallel computing approaches. In this study we highlight the application of parallel computing in reducing computational demands for crowd dynamic simulators while simultaneously improving their performances. This study also includes a comprehensive overview of the state-of-the-art parallel computing approaches used in crowd dynamic simulators.