Accelerating 3D FDTD algorithm using GPU computing

In order to accelerate 3D finite-difference time-domain(FDTD)algorithm using graphics processing unit(GPU),with the implement of "XY plane parallel,Z direction loop update" and other techniques,a GPU FDTD programming model is proposed which is both efficient and accurate.A project is simulated using the proposed model, and the result shows good agreement with CPU computing while simulation speed is 174 times faster, thus proves the effectiveness and great potential of GPU acceleration in FDTD research.