An Efficient Implementation of Parallelization in the Domain Decomposition of TELEMAC

In this study, we present an MPI-parallelized implementation of the domain decomposer for TELEMAC. The decomposer, called PARTEL, splits the computational domain into several partitions, which forms the basis of parallel simulations. In current TELEMAC releases (version 8), the only serial mode has been implemented in PARTEL, which gives room for speeding up the code with a parallel implementation together with a specific enhancement for the internal data structures used by the code. In this work, we fully parallelized the domain decomposition using MPI and utilized different implementations of Hash table for representing several data structures in PARTEL. This approach allows us to decompose a huge computational domain consisting of a hundred million elements into ten thousand partitions on an ordinary workstation. The benchmark also revealed that the speed-up by a factor of 4 is readily obtained compared to the original PARTEL before the enhancement.     «

In this study, we present an MPI-parallelized implementation of the domain decomposer for TELEMAC. The decomposer, called PARTEL, splits the computational domain into several partitions, which forms the basis of parallel simulations. In current TELEMAC releases (version 8), the only serial mode has been implemented in PARTEL, which gives room for speeding up the code with a parallel implementation together with a specific enhancement for the internal data structures used by the code. In this wor...     »