The problem of wave scattering by rough surfaces has been studied extensively by scientists and engineers because of its wide applications in science and technology. In this letter, the Physical Optics method, which is a high frequency technique, is presented for analyzing the scattering of rough surface. In addition, the compute unified device architecture of NVIDIA takes advantage of the Graphics Processing Units for parallel computing, and greatly improves the speed of computation. As there is a large number of data to deal with, a parallelization concept is presented which is based on the utilization of GPU to further improve the computational efficiency. In the end, the simulation time of CPU-based Physical Optics method and GPU-based Physical Optics method are compared, and it can be found that good acceleration effect has been achieved.
Radio wave propagation prediction is very important for the design of the mobile communication network. The raytracing algorithm is a commonly used computational method for site-specific prediction of the radio channel characteristics of wireless communication systems. However, it does not consider the diffuse scattering. Therefore, an indoor diffuse scattering model which based on diffuse scattering theory and FDTD is established. The diffuse scattering of indoor walls and ceiling and floor is calculated at a series of discrete time instance in this method. In recent years, the compute unified device architecture (CUDA) of NVIDIA takes advantage of the GPU for parallel computing, and greatly improve the speed of computation. Because there is a large number of data to deal with, in order to reduce the computation time, a GPU-based diffuse scattering model for indoor radio prediction is introduced in this paper, which fully utilizes the parallel processing capabilities of CUDA to further improve the computational efficiency. It can be found that good acceleration effect has been achieved.