APPLICATION OF GPU NEIGHBOR SEARCH ALGORITHM IN SPH ALGORITHM OF WIND-SAND FLOW

In order to solve the problem of low computational efficiency caused by the increase in the number of particles in the smoothed particle hydrodynamics (SPH) algorithm of wind-sand flow, GPU parallel computing is applied to the SPH algorithm of wind-sand flow. The reason why SPH algorithm was suitable for parallel computing was analyzed. Based on the neighbor search method of cell-linked list (CLL) method, the parallel computing model of SPH wind-sand flow was established. Through the verification of the calculation model, the slope field was analyzed, and the variation law of sand horizontal velocity along the height and the change of sand spatial distribution were obtained. The computational efficiency of the four wind-sand flow SPH algorithms under different particle numbers was compared. The results show that the proposed algorithm can improve the computational efficiency.