基于异构计算的三维FDTD并行算法及其在电磁仿真中的应用  

A 3D FDTD parallel algorithm based on heterogeneous computing and its application in electromagnetic simulation

作　　者：周兰花[1] 付彬[1,2] 李仁发[1] 刘新忠[1] 黄晶[1] 

机构地区：[1]湖南大学信息科学与工程学院,湖南长沙410082 [2]近地空间电磁环境监测与建模重点实验室(长沙理工大学),湖南长沙410114

出　　处：《计算机工程与科学》2017年第7期1241-1248,共8页Computer Engineering & Science

基　　金：国家863计划(2012AA01A301-01);长沙理工大学近地空间电磁环境监测与建模湖南省普通高校重点实验室开放基金(20150105)

摘　　要：时域有限差分(FDTD)法是求解电磁学中麦克斯韦方程组的重要方法之一,一直以来获得了广泛的使用,但是应用于电大尺寸目标仿真时存在巨大的耗时问题。为解决这一问题,利用图形处理器(GPU)的并行处理特性,结合计算统一设备架构(CUDA),以低通滤波器为算例,实现了时域卷积理想匹配层(CPML)吸收边界的三维FDTD高性能加速计算,目标网格数达5百万。实验在Fermi架构的Quadro 4000和Tesla M2050两款GPU上实测,误差均在10^(-4)范围内,相对于同时期的CPU分别可获得36和55倍以上的加速,结果表明该方法具有精度高、效率高、通用性和实用性强等特点。The finite difference time domain （FDTD） me ving Maxwell＇s equations in electromagnetics, and it is wid thod is one of the important methods for sol ely used. But it is time consuming when ap plied to the simulation of electrically large targets. In order to solve this problem, we take advantage of the parallel processing capacity of the graphics processor unit （GPU） together with the compute unified device architecture （CUDA）. Taking a low pass filter as an example and using five million targeting grids, we realize three-dimensional FDTD high performance speed calculation with time-domain convolution perfectly matched layer （CPML） absorbing boundary. Experiments are carried out on the Quadro 4000 and Tesla M2050 GPUs with the Fermi architecture, whose error rate is within the range of 10^-4 , and can obtain 36 and 55 times faster speed than the CPU of the same period. The results show that the method has the characteristics of high precision, high efficiency, versatility and strong practicability.

关 键 词：计算电磁学 FDTD 异构计算 计算统一设备架构 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]