GPU加速的自适应积分法研究
详细信息 查看官网全文
- 英文论文题名：Research on Acceleration of the Adaptive Integral Method by Using Graphics Processing Units
- 论文作者：陈金鑫 ; 杨武
- 英文论文作者：CHEN Jin-xin ; YANG Wu ; College of computer science and engineering ; Southeast University
- 年：2016
- 作者机构：东南大学计算机科学与工程学院;
- 论文关键词：自适应积分法(AIM) ; 图形处理单元(GPU) ; 计算设备统一架构(CUDA)
- 英文论文关键词：adaptive integral method(AIM) ; graphics processing units(GPU) ; CUDA
- 会议召开时间：2016-08-17
- 会议录名称：2016年全国军事微波、太赫兹、电磁兼容技术学术会议论文集
- 语种：中文
- 分类号：O441;O411
- 学会代码：WBXB
- 会议名称：2016年全国军事微波、太赫兹、电磁兼容技术学术会议
- 会议地点：中国陕西西安
- 主办单位：中国电子学会微波分会、天线与微波技术重点实验室
- 学会名称：《微波学报》编辑部
- 页数：4
- 文件大小：673k
- 原文格式：D
- 会议级别：全国
摘要
自适应积分法(AIM)是针对矩量法的一种有效改进。但是当未知量数目比较大时,其计算仍较耗时。为了加快AIM的求解,本文引入了CUDA技术来实现在GPU上并行加速自适应积分法。实验结果表明,与传统的CPU实现AIM方法相比,利用GPU加速的AIM方法在保证正确性的同时大大减少了求解计算时间。
The adaptive integral method(AIM) is an effective improvement of the method of moments(MOM). It is time consuming when the number of unknowns is very large. In this paper, we accelerating the AIM by using the GPU parallelization. The experimental results show that the GPU implement of AIM decreased computation time significantly and obtained the same results compared with the CPU implement.
The adaptive integral method(AIM) is an effective improvement of the method of moments(MOM). It is time consuming when the number of unknowns is very large. In this paper, we accelerating the AIM by using the GPU parallelization. The experimental results show that the GPU implement of AIM decreased computation time significantly and obtained the same results compared with the CPU implement.
引文
[1]E.Bleszynski,M.Bleszynski,T.Jaroszewicz.AIM:Adaptive intergral method for solving large-scale electromagnetic scattering and radiation problems.[J].Radio Science,31(5),1225-1251
[2]Stephen D.Gedney and Raj Mittra.The use of the FFTfor the efficient solution of the problem of electromagnetic scattering by a body of revolution[J].Microwave and Optical Technology Letters,1990,15(3),313-322
[3]NVIDIA,CUDA C Programming Guide[Z].USA:NVIDIA Corporation 2013
[4]陈岩,张光辉等,异构平台中并行矩量法的加速技术[J].微波学报,2014,51-54
[5]张溯,电磁散射问题的自适应积分法及其预条件技术的研究[D].南京:东南大学,2009
[6]王晓峰,三维目标电磁散射的自适应积分方法研究[D].成都:电子科技大学,2006
[7]Roverto D.Graglia.On the numerical integration of the linear shape functions times the 3-D Green’s function or its gradient on a plane triangle[J].IEEE Transactions on Antennas and Propagation,1993,41(10),1448-1455
[8]Walton C.Gibson,The Method of Moments in Electromagnetics[M].US:Taylor&Francis Group,2008
[9]王兴,宽带散射和辐射问题的自适应积分法研究[D].西安:西安电子科技大学,2011
[10]NVIDIA Corporation Technical Staff.NVIDIA CUDAprogramming guide version 4.2[Z].USA:NVIDIACorporation,2012
[2]Stephen D.Gedney and Raj Mittra.The use of the FFTfor the efficient solution of the problem of electromagnetic scattering by a body of revolution[J].Microwave and Optical Technology Letters,1990,15(3),313-322
[3]NVIDIA,CUDA C Programming Guide[Z].USA:NVIDIA Corporation 2013
[4]陈岩,张光辉等,异构平台中并行矩量法的加速技术[J].微波学报,2014,51-54
[5]张溯,电磁散射问题的自适应积分法及其预条件技术的研究[D].南京:东南大学,2009
[6]王晓峰,三维目标电磁散射的自适应积分方法研究[D].成都:电子科技大学,2006
[7]Roverto D.Graglia.On the numerical integration of the linear shape functions times the 3-D Green’s function or its gradient on a plane triangle[J].IEEE Transactions on Antennas and Propagation,1993,41(10),1448-1455
[8]Walton C.Gibson,The Method of Moments in Electromagnetics[M].US:Taylor&Francis Group,2008
[9]王兴,宽带散射和辐射问题的自适应积分法研究[D].西安:西安电子科技大学,2011
[10]NVIDIA Corporation Technical Staff.NVIDIA CUDAprogramming guide version 4.2[Z].USA:NVIDIACorporation,2012