Title: Performance comparison of different parallel lattice Boltzmann implementations on multi-core multi-socket systems

Authors: S. Donath, K. Iglberger, G. Wellein, T. Zeiser, A. Nitsure, U. Rude

Addresses: System Simulation – Computer Science 10 (LSS), University of Erlangen-Nuremberg, Germany. ' System Simulation – Computer Science 10 (LSS), University of Erlangen-Nuremberg, Germany. ' Regional Computing Center of Erlangen (RRZE), University of Erlangen-Nuremberg, Germany. ' Regional Computing Center of Erlangen (RRZE), University of Erlangen-Nuremberg, Germany. ' Regional Computing Center of Erlangen (RRZE), University of Erlangen-Nuremberg, Germany. ' System Simulation – Computer Science 10 (LSS), University of Erlangen-Nuremberg, Germany

Abstract: In this report, we discuss the performance behaviour of different parallel lattice Boltzmann implementations. In previous works, we already proposed a fast serial implementation and a cache oblivious spatial and temporal blocking algorithm for the lattice Boltzmann method (LBM) in three spatial dimensions. The cache oblivious update scheme has originally been proposed by Frigo et al. The main idea is to provide maximum performance results for stencil-based methods by dividing the space-time domain in an optimal way, independently of any external parameters, such as cache size. In view of the increasing gap between processor speed and memory performance, this approach offers a promising path to increase cache utilisation. We present results for the shared memory parallelisation of the cache oblivious implementation based on task queueing in comparison to the iterative standard implementation, thereby focusing on the special issues for multi-core and multi-socket systems.

Keywords: parallel lattice Boltzmann; cache optimisation; cache oblivious; multi-core multi-socket systems; task queueing; shared-memory parallelisation.

DOI: 10.1504/IJCSE.2008.021107

International Journal of Computational Science and Engineering, 2008 Vol.4 No.1, pp.3 - 11

Available online: 04 Nov 2008 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article