Analysis and optimization of a demographic simulator for parallel environments

In the past years, the advent of multi-core machines has led to the need for adapting current simulation solutions to modern hardware architectures. In this poster, we present a solution to exploit multicore shared-memory capacities in Yades, a parallel tool for running socio-demography dynamic simulations. We propose to abandon the single-threaded programming approach addresses in Yades by using ROOT-Sim, a library which allows to apply discrete event simulation to parallel environments profiting share-memory capabilities. As a result of this new approach, our results show the improvement in Yades' performance and scalability.