Parallelization of the FDTD algorithm over MPICH-G2

In this paper, a parallel implementation of the FDTD algorithm based on parallel sockets communication is proposed. The optimized MPI programs run on a grid-enabled MPI (MPICH-G2), which is a complete implementation of the MPI-1 standard using Globus toolkit services to support efficient and transparent execution in heterogeneous grid environments. The numerical results show that more efficient communication can be achieved, and this kind of grid parallel computing environment proved to be low-cost, more practical and more efficient.