GPU+COW parallel computing system to accelerate the FDTD method

摘要

Since Finite-Difference Time-Domain (FDTD) method was proposed by Yee in 1966, its hardware acceleration problem arouses much concern. In this paper, how to use GPU+COW parallel computing system to accelerate the FDTD method was introduced briefly. The acceleration scheme and the simulation results were presented, and then the feasibility and efficiency of the proposed scheme were proved.