Practical Implementation of a CPML Absorbing Boundary for GPU Accelerated FDTD Technique

摘要

The use of graphical processing units (GPU) has been recently documented for the implementation of the FDTD technique; however, little has been reported about the necessary additions to three dimensional FDTD codes to make the technique more useful for fast antenna analysis and design. This paper details the addition of a convolutional perfectly matched layer absorbing boundary (CPML) to a three dimensional GPU accelerated FDTD code.