On efficiency of supercomputers in CFD simulations
The present work is reporting the experience on different aspects of using supercomputers for CFD applications. Several problems that appear with new supercomputers built of multi-core nodes are considered. In particular, it is a problem of increased influence of irregular memory access, problems with efficient use of such a big numbers of processors that are available in present time. A hybrid parallelization using MPI and OpenMP technologies is also considered to improve efficiency of computations. Results of performance tests where the hybrid approach outperformed the MPI-only parallelization are shown. Finally, a performance comparison of Marenostrum and MVS-50000 supercomputers is presented, illustrative numerical results on the DNS of turbulent flows are also included.
Key wordssupercomputers parallel algorithms multi-core architecture hybrid parallelization
Unable to display preview. Download preview PDF.
- 1.I. Abalakin, A. Dervieux, and T. Kozubskaya. High Accuracy Finite Volume Method for Solving Nonlinear Aeroacoustics Problems on Unstructured Meshes. Chinese Journal of Aeroanautics, pages 97–104, 2006.Google Scholar
- 3.M.V. Iakobovski, S.N. Boldyrev, and S.A. Sukov. Big Unstructured Mesh Processing on Multiprocessor Computer Systems. Parallel Computational Fluid Dynamics: Advanced numerical methods software and applications. Proc. of the Parallel CFD 2003 Conference Moscow, Russia (May 13-15, 2003), Elsevier, Amsterdam, pages 73–79, 2004.Google Scholar
- 4.T. Kozubskaya, I. Abalakin, A. Gorobets, and Alexey Mironov. Simulation of Acoustic Fields in Resonator-Type Problems Using Unstructured Meshes. AIAA Paper 2519-2006, 2006.Google Scholar
- 5.F. X. Trias, A. Gorobets, M. Soria, and A. Oliva. DNS of Turbulent Natural Convection Flows on the MareNostrum Supercomputer. In Parallel Computational Fluid Dynamics, Antalya, Turkey, May 2007. Elsevier.Google Scholar