Highlight Photo
High Performance Lattice Boltzmann Method for Shallow Water Flows
Fig. 1: LBM performance on (a) 16 node CPU (b) on a GPU machine
LBM performance results on (a) a 16 node shared memory system using OpenMP. Parallel scalability with increasing number of threads. (b) on a single Geforce 8800 GT graphics card using Jacket MATLAB GPU engine vs single core of a 3.33 GHz Quad Core CPU. Parallel speedup scalability with increasing problem size.
Credits: Tubbs and Tsai
