首页 | 本学科首页   官方微博 | 高级检索  
     

基于视窗操作系统网络并行计算环境的建立与应用
引用本文:化春键,赵升吨,申光宪. 基于视窗操作系统网络并行计算环境的建立与应用[J]. 西安交通大学学报, 2004, 38(7): 758-761
作者姓名:化春键  赵升吨  申光宪
作者单位:1. 西安交通大学机械工程学院,710049,西安
2. 燕山大学机械工程学院,066004,秦皇岛
基金项目:国家自然科学基金资助项目(50375117),陕西省自然科学基金资助项目(2003E204).
摘    要:针对单台微型计算机进行数值模拟耗时长的问题,在微软公司开发的32位视窗操作系统下建立了网络并行计算环境.把2台微型计算机连成一个星形结构的以太网,以并行虚拟机作为网络并行计算平台、Vi sualC 作为开发工具,通过建立帐号及主目录、配置网络、设置系统环境变量和测试等步骤,完成了并行虚拟机的配置与调试.采用C语言自行研制了并行凝缩算法的源程序,并进一步以主从进程模式对金属平板轧制过程的弹塑性力学行为进行了数值模拟.研究结果表明,所建立的网络并行计算环境运行可靠,并行加速比可达1 785,并行效率达到了89%,与单台微型计算机相比,不仅提高了运行速度,且大大缩短了计算时间.

关 键 词:并行计算环境  网络  数值模拟  弹塑性力学
文章编号:0253-987X(2004)07-0758-04
修稿时间:2003-12-24

Establishment and Application of a Network Parallel Computation Environment Based on Windows Network Operating System
Hua Chunjian,Zhao Shengdun,Shen Guangxian. Establishment and Application of a Network Parallel Computation Environment Based on Windows Network Operating System[J]. Journal of Xi'an Jiaotong University, 2004, 38(7): 758-761
Authors:Hua Chunjian  Zhao Shengdun  Shen Guangxian
Affiliation:Hua Chunjian~1,Zhao Shengdun~1,Shen Guangxian~2
Abstract:To shorten CPU time-consuming in numerical simulation, a network parallel computation environment was established, which is based on the win32 network operating system developed by Microsoft. Two microcomputers were connected into an ethernet with one 100 M fast ethernet switch. Parallel virtual machine is adopted as parallel platform and Visual C~( ) was applied as development tool in the environment, the configuration of parallel virtual machine was accomplished by establishing account number and main catalogue, collocating network, setting system environment variables, and testing.Moreover, a source program for exclusive parallel reducing computing method is developed with C language.With this program the elastic-plastic mechanical behavior of a strip rolling process was simulated numerically by means of master-slave model. The results indicate that the parallel computation can be realized dependably in the established network parallel computation environment, the ratio of parallel speedup can reach 1.785 and the efficiency of parallel computing can reach 89%. Compared with the CPU time-consuming of single microcomputer, and thus the computing speed is raised and the CPU time-consuming is reduced greatly.
Keywords:parallel computing platform  network  numerical computation  elastic-plastic  
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号