GPU Accelerated Fluid Flow Computations Using the Latice Boltzmann Method

Abstract

We propose a numerical implementation based on a Graphics Processing Unit (GPU) for the acceleration of the execution time of the Lattice Boltzmann Method. The performance analysis is based on three three-dimensional benchmark applications: Poisseuille flow, lid-driven cavity flow, and flow in an elbow-shaped domain. Three different, recently released GPU cards are considered for parallel implementation. To correctly evaluate the speed-up potential of the GPUs, both single-core and multi-core CPU-based implementations are used. The results indicate that the GTX 680 GPU card leads to the best performance, with a speed-up ranging between 6.7 and 14.35 over the multi-core CPU-based implementation, depending on the application and on the grid density.