| 全球气象资料客观分析系统的分布式并行化 |
| |
| 作者姓名： | ZHAO Jun SONG Junqiang LI Zhenjun |
| |
| 作者单位： | [1]CollegeofComputerScience,NationalUniversityofDefenseTechnology,Changsha410073 [2]MeteorologicalCenterofAirforce,Beijing100843 |
| |
| 摘 要： | 对已有的率行算法进行并行化，是一项很困难的工作。通过对全球气象资料客观分析系统率行算法的研究，提出了在MPP高性能计算机上的一种静态分配数据的分布式并行算法。该算法通过间隔选取分析盒子和模式格点纬圈行，将数据分配给不同的处理机实现分布式并行。该并行算法负载平衡好，并行效率高，而且并行化代价较低，具有良好的可扩展性。最后，给出了并行算法的性能测试结果。
|
| 关 键 词： | 全球气象资料 静态分配数据 分布式并行算法 并行效率 客观分析系统 |
Distributed Parallelization of a Global Atmospheric Data Objective Analysis System |
| |
| ZHAO Jun,SONG Junqiang,LI Zhenjun.Distributed Parallelization of a Global Atmospheric Data Objective Analysis System[J].Advances in Atmospheric Sciences,2003,20(1):159-163. |
| |
| Authors: | Jun Zhao Junqiarig Song Zhenjun Li |
| |
| Institution: | College of Computer Science, National University of Defense Technology, Changsha 410073,College of Computer Science, National University of Defense Technology, Changsha 410073,Meteorological Center of Airforce, Beijing 100843 |
| |
| Abstract: | It is difficult to parallelize a subsistent sequential algorithm. Through analyzing the sequentialalgorithm of a Global Atmospheric Data Objective Analysis System, this article puts forward a distributedparallel algorithm that statically distributes data on a massively parallel processing (MPP) computer.The algorithm realizes distributed parallelization by extracting the analysis boxes and model grid pointlatitude rows with leaped steps, and by distributing the data to different processors. The parallel algorithmachieves good load balancing, high parallel efficiency, and low parallel cost. Performance experiments ona MPP computer are also presented. |
| |
| Keywords: | distributed parallelization analysis box data distribution objective analysis |
| 本文献已被 CNKI 维普 万方数据 SpringerLink 等数据库收录！ |
| 点击此处可从《大气科学进展》浏览原始摘要信息 |
| 点击此处可从《大气科学进展》下载免费的PDF全文 |
