In the Netherlands, for coastal and inland water applications, wave modelling with SWAN has become a main ingredient. However, computational times are relatively high. Therefore we investigated the parallel efficiency of the current MPI and OpenMP versions of SWAN. The MPI version is not that efficient as the OpenMP version within a single node. Therefore, in this paper we propose a hybrid version of SWAN. It combines the efficiency of the current OpenMP version on shared memory with the capability of the current MPI version to distribute memory over nodes. We describe the numerical algorithm. With numerical experiments for important applications we show the potential of this hybrid version, We further optimize the approach and illustrate the behavior for larger number of nodes. Parallel I/O will be subject of future research.
Additional Metadata
Keywords distributed memory, shared memory, hybrid, MPI, OpenMP, wave modelling
MSC Parallel computation (msc 65Y05), Algorithms for specific classes of architectures (msc 65Y10), Complexity and performance of numerical algorithms (msc 65Y20), Analysis of algorithms and problem complexity (msc 68Q25), Analysis of algorithms (msc 68W40)
THEME Null option (theme 11)
Publisher CWI
Series Multiscale Dynamics [MD]
Citation
Genseberger, M, & Donners, J. (2016). A hybrid version of swan for fast and efficient practical wave modelling. Multiscale Dynamics [MD]. CWI.