基于GPU并行射线跟踪的电波预测算法研究
|
摘要
针对利用射线跟踪算法进行小区电波预测的计算量非常大,运行时间较长,在实际工程中难以应用等的问题,本文在研究已有并行射线跟踪算法,即主从模式和对等模式并行射线跟踪的基础上,采用图形处理器(GPU)作为计算平台,结合GPU高并行处理能力,运用CUDA的编程框架,并在NVIDIA公司出厂的GEFORCE 9800GT上,实现了对城市小区的简单模型环境的电波预测。实验结果表明,基于GPU并行射线跟踪算法适用于城市小区的电波预测研究,该算法能大大节省通信时间,缩短计算时间,更方便更实用。
Ray-tracing algorithm for predicting radio-wave propagation in urban microcell needs a great deal of computation time. In order to more widely using this algorithm, we need to solve these problems. In this paper, based on previous research of parallel ray-tracing, namely Master/Slaves Model (MSM) and Peer to Peer Model (PPM), present a new method which combines with the capability for parallel processing of GPU, and uses CUDA programming framework. In GEFORCE 9800GT produced by NVIDIA, we conduct some experiments to forecast radio-wave propagation in simple city environment. The results show that this method in the paper which could save communication-time and computation-time is more convenient than that uses several computes.
Ray-tracing algorithm for predicting radio-wave propagation in urban microcell needs a great deal of computation time. In order to more widely using this algorithm, we need to solve these problems. In this paper, based on previous research of parallel ray-tracing, namely Master/Slaves Model (MSM) and Peer to Peer Model (PPM), present a new method which combines with the capability for parallel processing of GPU, and uses CUDA programming framework. In GEFORCE 9800GT produced by NVIDIA, we conduct some experiments to forecast radio-wave propagation in simple city environment. The results show that this method in the paper which could save communication-time and computation-time is more convenient than that uses several computes.
引文
[1]赵毅,卓永宁等.一种室内射线跟踪的二维模型实现[J].微计算机信息.2010,12-1:166-168
[2]刘海涛,黎滨洪,谢勇.并行射线跟踪算法及其在城市电波预测的应用[J].电波科学学报.2004,10
[3]刘海涛,黎滨洪,戚冬生.基于主从模式的并行射线跟踪算法的研究[C].2003全国微波毫米波会议论文集
[4]杨俊华,符红光,郭惠.基于GPU快速光线跟踪算法的设计与实现[J].计算机应用.2007,8
[5]钱悦.图形处理器CUDA编程模型的应用研究[J].计算机与应用工程.2008,12
[6]Alvaro Valcarce,Guillaume De La Roche,Jie Zhang,“A GPU approach to FDTD for Radio Coverage Prediction”IEEE ICCS2008
[7]陶煜波.基于图形硬件的快速电磁计算方法与系统[D].浙江大学博士学位论文.2009,1
[8]陆建勇,曹雪虹,焦良葆.基于GPU交互式光线跟踪算法的设计与实现[J].南京工程学院学报(自然科学版).2009,9
[9]吴连贵,易瑜,李肯立.基于CUDA的地震数据相干并行算法(计算机应用).2009,3
[2]刘海涛,黎滨洪,谢勇.并行射线跟踪算法及其在城市电波预测的应用[J].电波科学学报.2004,10
[3]刘海涛,黎滨洪,戚冬生.基于主从模式的并行射线跟踪算法的研究[C].2003全国微波毫米波会议论文集
[4]杨俊华,符红光,郭惠.基于GPU快速光线跟踪算法的设计与实现[J].计算机应用.2007,8
[5]钱悦.图形处理器CUDA编程模型的应用研究[J].计算机与应用工程.2008,12
[6]Alvaro Valcarce,Guillaume De La Roche,Jie Zhang,“A GPU approach to FDTD for Radio Coverage Prediction”IEEE ICCS2008
[7]陶煜波.基于图形硬件的快速电磁计算方法与系统[D].浙江大学博士学位论文.2009,1
[8]陆建勇,曹雪虹,焦良葆.基于GPU交互式光线跟踪算法的设计与实现[J].南京工程学院学报(自然科学版).2009,9
[9]吴连贵,易瑜,李肯立.基于CUDA的地震数据相干并行算法(计算机应用).2009,3
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心