International Journal of Computer Theory and Engineering
The article describes the implementation of the new synchronization method for parallel discrete event simulation and presents the experimental results obtained from parallel environment. In the proposed method the time steps are introduced and the synchronization occurs only at the end of a time step to achieve reduction of messages, exchanged between local processes of simulation. The method was implemented in C++, using the MPI libraty. The studies have shown that it is possible, under certain conditions and using appropriate hardware architecture, the significant acceleration of distributed simulations.