| 基于机群的并行匹配算法 |
| |
| 引用本文： | 张春玲,邱振戈.基于机群的并行匹配算法[J].测绘科学,2006,31(6):127-128,136. |
| |
| 作者姓名： | 张春玲 邱振戈 |
| |
| 作者单位： | 河南省测绘局,郑州,450052;中国测绘科学研究院重点实验室,北京,100039 |
| |
| 摘 要： | 随着卫星遥感技术的发展,需要快速地将卫星遥感图像数据转化为用户需要的信息,并行图像处理技术是解决“快速”的重要途径。并行程序的性能与计算机体系结构密切相关,不但取决于CPU,还与系统架构、指令结构、存储部件的存取速度等因素有关。一般意义上,提高并行程序的性能采用粗粒度并行,指令级优化(ILP)和存储优化等技术。作为尝试,本文讨论了在工业标准化机群上采用软件式共享存储系统做的并行影像匹配方法,以影像匹配算法为例子,讨论了如何在粗粒度并行、指令级优化(ILP)和存储优化三个方面提高图像处理的计算速度。
|
| 关 键 词： | 影像匹配 粗粒度并行 指令级优化(ILP) 和存储优化 机群 |
| 文章编号： | 1009-2307（2006）06-0127-03 |
| 收稿时间： | 2006-04-20 |
| 修稿时间： | 2006-04-20 |
A parallel image matching algorithm running on cluster |
| |
| ZHANG Chun-ling,QIU Zhen-ge.A parallel image matching algorithm running on cluster[J].Science of Surveying and Mapping,2006,31(6):127-128,136. |
| |
| Authors: | ZHANG Chun-ling QIU Zhen-ge |
| |
| Abstract: | It is the key issue in remote sensing field to transform massive data into information in short time.Parallel image processing on high performance computing is one of the key technologies to solve this problem quickly.The performance of parallel program is closely related to computer architecture,besides CPU,including system framework,instruction structure and access speed of storage unit.Generally there are several ways to improve the performance of parallel program,such as coarse grain parallelism,instruction level and memory optimizing.In this paper,a parallel image matching algorithm on Cluster is studied.Based on this algorithm,the image processing speed can be improved with the help of coarse grain parallelism,instruction level and memory optimizing. |
| |
| Keywords: | image matching coarse grain parallelism instruction level optimizing memory optimizing cluster |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
