Programming finite-difference time-domain for graphics processor units using compute unified device architecture

Recently graphic processing units (GPU's) have become the hardware platforms to perform high performance scientific computing them. The unavailability of high level languages to program graphics cards had prevented the widespread use of GPUs. Relatively recently Compute Unified Device Architecture (CUDA) development environment has been introduced by NVIDIA and made GPU programming much easier. This contribution presents an implementation of finite-difference time-domain (FDTD) method using CUDA. A thread-to-cell mapping algorithm is presented and performance of this algorithm is provided.