北京航空航天大学学报 ›› 2016, Vol. 42 ›› Issue (9): 1874-1883.doi: 10.13700/j.bh.1001-5965.2015.0593

• 论文 • 上一篇    下一篇

基于分布式平台的FDTD并行算法

冯圆1,2, 代小霞1, 唐晓斌1, 龚晓燕3   

  1. 1. 中国电子科学研究院 预警机研究所, 北京 100041;
    2. 空军预警学院 雷达技术系, 武汉 430019;
    3. 火箭军指挥学院 指挥系, 武汉 430013
  • 收稿日期:2015-09-10 出版日期:2016-09-20 发布日期:2016-02-18
  • 通讯作者: 唐晓斌,Tel.:010-68893613,E-mail:science_xiaoxiao@163.com E-mail:science_xiaoxiao@163.com
  • 作者简介:冯圆,男,博士研究生,讲师。主要研究方向:计算电磁学。Tel.:010-68893271,E-mail:fy-zbh@163.com;代小霞,女,硕士研究生。主要研究方向:雷达信号处理。Tel.:17090084855,E-mail:daixiaoxia5221@163.com;唐晓斌,女,研究员。主要研究方向:电磁兼容测试。Tel.:010-68893613,E-mail:science_xiaoxiao@163.com;龚晓燕女,硕士,讲师。主要研究方向:战役指挥。Tel.:15926498355,E-mail:2228730463@qq.com
  • 基金资助:
    国家“863”计划(2012AA01A308);国家“973”计划(613206)

FDTD parallel algorithm based on distributed platform

FENG Yuan1,2, DAI Xiaoxia1, TANG Xiaobin1, GONG Xiaoyan3   

  1. 1. Institute of Early Warning Aircraft, China Academy of Electronics and Information Technology, Beijing 100041;
    2. Department of Radar Technology, Academy of Air Force Early Warning, Wuhan 430019;
    3. Department of Command, Rocket Army Command Academy, Wuhan 430012
  • Received:2015-09-10 Online:2016-09-20 Published:2016-02-18
  • Supported by:
    National High-tech Research and Development Program of China (2012AA01A308); National Basic Research Program of China (613206)

摘要: 基于分布式平台开展一种新的时域有限差分(FDTD)并行算法研究,该算法基于VC++、CUDA5.0平台开发,调用Intel MPI 4.1.0库进行测试,在上海交通大学高性能计算中心图形处理单元(GPU)集群、上海超级计算机中心的“魔方”商用超级计算机以及国家超级计算济南中心的“神威蓝光”国产超级计算机等平台开展软件调试。通过对纯CPU、GPU以及CPU和GPU的混合测试,线程调度水平、核心函数处理速度得到明显提升,同时减少了通信执行时间比例,提高了加速比和并行效率,最后以2×2微带阵列为验证模型进行拓扑优化测试,结果证明该算法准确、有效。

关键词: Mur, 消息传递接口, 图形处理单元(GPU), 时域有限差分(FDTD), 分布式平台

Abstract: A new finite difference time domain (FDTD) parallel algorithm is developed based on distributed platform, which is based on VC++, CUDA5.0 development platform, calling Intel MPI 4.1.0 library for testing, developing software debugging on the platforms of high performance computing center graphics processing units (GPU) cluster in Shanghai Jiao Tong University, "Rubik's Cube" commercial super computer at Shanghai Supercomputer Center, and "Divinity Blue" domestic super computer at the National Supercomputing Center in Jinan. By pure CPU, GPU, and CPU and GPU hybrid test, thread scheduling level and kernel function processing speed improve significantly, while the proportion of the execution time of communication reduces, and the acceleration ratio and operation efficiency improve. Finally, the topology optimization of the model is verified by 2×2 micro-strip arrays. The results show that the algorithm is accurate and effective.

Key words: Mur, message passing interface, graphics processing units (GPU), finite difference time domain (FDTD), distributed platform

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发