Abstract
*† We present two multilevel parallel models based on MPI/MPI (MPI denoting Message Passing Interface) and MPI/OpenMP (OpenMP denoting Open Multi-Processing) for highorder CFD methods and compare their performances. These models are implemented within the spectral/hp element framework to take advantage of the hierarchical structures arising from deterministic and stochastic CFD computations. For MPI/MPI, we employ MPI process groups to decompose the computations into different levels. For MPI/OpenMP, we take a Single-Program-Multiple-Data (SPMD) style approach to OpenMP shared memory parallelism that significantly reduces the OpenMP synchronizations. These models demonstrate a good scalability with respect to the problem size and a good speedup for fixed problem sizes. With identical configurations, the MPI/MPI parallel model is observed to be generally more efficient. The advantage of these multilevel approaches lies in that they reduce the number of processes participating in each communication and the latency overhead, and thus enable the applications to scale to a large number of processors more efficiently. The models have been applied to the direct simulation of turbulent flows past a circular cylinder at Reynolds number Re=10,000.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: Journal of Aerospace Computing, Information, and Communication
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
