Extending the educational scope of a particle-based simulation framework through parallelization

Particle systems have been incorporated into a wide variety of applications in both academia and industry. They can be employed to investigate complex natural phenomena, illustrate scientific concepts and generate special effects for entertainment purposes. Recently, we implemented an educational simulation framework based on particle systems that can be used to perform interactive virtual experiments involving complex physical laws. The positive feedback received from a pilot deployment of this framework motivated us to look for strategies to increase its scope. However, more complex and engaging simulations require the use of a larger number of geometric primitives (particles), which results in higher computational costs. To mitigate these costs, we resorted to the implementation of parallel techniques through the use of the Message Passing Interface (MPI) standard. In this paper, we describe these techniques and discuss the performance gains resulting from their application to the simulation algorithm that forms the core of our framework. These results were obtained through practical test cases which are also described in detail in this paper.