An OpenCL implementation for acceleration of the BOR-FDTD algorithm on GPUs

Abstract

The body of revolution finite-difference time-domain (BOR-FDTD) method is accelerated using graphics processing units (GPUs). The developed BOR-FDTD algorithm can be employed to efficiently characterize an axisymmetric gradientindex (GRIN) lens embedded in a planar multilayered medium when illuminated by an oblique incidence plane wave. The BOR-FDTD method is carried out by using the open computing language (OpenCL) programming that can be executed across heterogeneous platforms consisting of CPUs and GPUs. The accuracy and efficiency of the developed parallel techniques are demonstrated by the presented numerical results.