Two improved nearest neighbor search algorithms for SPH and their parallelization

Abstract

Efficient search algorithm is very important for Smoothed Particle Hydrodynamics (SPH). In this Chapter, the entire search process is divided into pre-search process and main searching process. On this basis, two improved search algorithms based on sliding windows and background grid are proposed. Application was made for three-dimensional shear cavity flow. Results show that the sequential program can greatly reduce the computational cost while maintaining the same accuracy. Furthermore, the improved search algorithms have been implemented on a cluster, the experimental results demonstrate that the parallel program has a higher speedup. © 2012 Springer Science+Business Media B.V.