Abstract
Lattice Boltzmann Method (LBM) is a powerful numerical simulation method of the fluid flow. With its data parallel nature, it is a promising candidate for a parallel implementation on a GPU. The LBM, however, is heavily data intensive and memory bound. In particular, moving the data to the adjacent cells in the streaming computation phase incurs a lot of uncoalesced accesses on the GPU which affects the overall performance. Furthermore, the main computation kernels of the LBM use a large number of registers per thread which limits the thread parallelism available at the run time due to the fixed number of registers on the GPU. In this paper, we develop high performance parallelization of the LBM on a GPU by minimizing the overheads associated with the uncoalesced memory accesses while improving the cache locality using the tiling optimization with the data layout change. Furthermore, we aggressively reduce the register uses for the LBM kernels in order to increase the run-time thread parallelism. Experimental results on the Nvidia Tesla K20 GPU show that our approach delivers impressive throughput performance: 1210.63 Million Lattice Updates Per Second (MLUPS).
Highlights
Lattice Boltzmann Method (LBM) is a powerful numerical simulation method of the fluid flow, originating from the lattice gas automata methods [1]
We show the performance improvements of our optimization techniques compared with the previous approach based on the Structure of arrays (SoA) Pull implementation: (i) Figure 16 compares the average performance of the SoA with and without removing the branch divergences explained in Section 4.4 in the kernel code
In order to improve the cache locality and minimize the overheads associated with the uncoalesced accesses in moving the data to the adjacent cells in the streaming phase of the LBM, we used the tiling optimization with the data layout change
Summary
Lattice Boltzmann Method (LBM) is a powerful numerical simulation method of the fluid flow, originating from the lattice gas automata methods [1]. LBM models the fluid flow consisting of particles moving with random motions. Such particles exchange the momentum and the energy through the streaming and the collision processes over the discrete lattice grid in the discrete time steps. The architecture of the GPU has gone through a number of innovative design changes in the last decade. It is integrated with a large number of cores and multiple threads per core, levels of the cache hierarchies, and the large amount (>5 GB) of the on-board memory. The advanced GPU architecture and the flexible programming environments have made possible innovative performance improvements in many application areas
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
