【作者】 李永旭；

【导师】 林和平；

【作者基本信息】 东北师范大学 ， 计算机应用技术， 2007， 硕士

【摘要】 随着超级计算需求的扩大，人们把高性能计算更多的应用于数据挖掘应用、图像处理业务、基因测序比对处理等数据处理领域。科学计算的规模迅速膨胀，使得传统的串行计算已经不能满足要求，从而提出了并行计算的概念。并行计算是相对于串行计算来说的，它将进程相对独立的分配于不同的节点上，由各自独立的操作系统调度，享有独立的CPU和内存资源；进程间通过消息传递相互交换信息。但是除了近来逐步完善并走向实用化的数据流计算机之外，50年来计算机系统结构虽取得重大进展，但并没有发生革命性的变化，现已实用化的计算机是以冯·诺依曼机器为原型，在此基础上加入并提高并行处理实现的手段和并行处理的能力。MPI是1994年5月发布的一种消息传递接口，是为消息传递程序提供的标准库。MPI以语言独立的形式来定义这个接口库，并提供了与C和Fortran语言的绑定，是目前高效率的超大规模并行计算最可信赖的平台。目前，Internet连接着大量的个人PC机和各种超级计算机，汇集了大量的计算资源。为了利用Internet上广泛分布的空闲计算资源，目前已经有多个分布式计算项目，比如：SETI@Home，GIMPS等等。这些项目已经证实了这种计算模型的可行性。这些系统的资源高度不稳定、用户不可信任、通信带宽有限、延迟巨大。本文采用目前国内外较为流行的面向对象系统分析技术，确定了由对象层、结构与实例连接层、主题层、属性层和事件层构成的静态架构。最后，给出由静态架构反射出来的用例视图，用于表示系统功能的高层逻辑模型，实现在计算机集群的网络范围的分布式计算，该系统资源稳定、用户可信、通信情况优良。更多还原

【Abstract】 According to the expansion of super computing requirements, highperformance computing has helped people in data processing areas such as: datamining applications, image processing business, Gene sequencing comparing.However, considering the rapid expansion of the scientific computing scale, thetraditional serial calculation has been unable to meet the requirements, thus theconcept of parallel computing has been carried out.Parallel computing is in contrast to the serial calculation, it distributes theprocesses to several independent nodes, which are scheduled by independentoperating system, and have independent CPU and memory resources; theexchanges of information between nodes are through message transmission.There are some important improvements on computer system architecture in therecent 50 years although there was no revolutionary changes except the data flowcomputer which gradual improvement toward practical application. Thecomputers have been put into applications are all based on Von Neumannprototype, and improve the means and capabilities of parallel processing on thisbasis.MPI is a kind of message passing interface which is published on May,1994, it provides the standard library for message passing procedure. MPI definesthe interface library by language independent form, and provides the binding withC and FORTRAN language. It is the most reliable platform for high efficiencyand ultra-large-scale parallel computing.At present, the Internet connects with a lot of PC and varioussuper-computer, assembled a large number of computing resources. In order touse the extensively distributed idle calculation resources on the Internet, there arealready a number of distributed computing projects have been set up, for example:SETI @Home, GIMPS and so on. These projects have confirmed the feasibilityof such calculation model. These resources are highly unstable and can not betrusted by users, bandwidth is limited and full of tremendous delay.This paper uses the object-oriented analysis techniques which is popularinside and outside the country, the static structure is made up of by the objectlayer, the structure and link layer, the theme level, the attribute level and theevents constitute layer. Finally, give out the use case view which is reflected bythe static structure, the high level logic model of system functions and theimplementation of the distributed computing in the computer cluster network.The system has steady resources, credible users and good communications.更多还原