高速信息網(wǎng)絡(luò)的飛速發(fā)展以及網(wǎng)絡(luò)并行計(jì)算支撐軟件如PVM、MPI、Express、Linda,P4等的出現(xiàn),使得網(wǎng)絡(luò)并行計(jì)算的投資少、見(jiàn)效快、靈活性強(qiáng)、性能價(jià)格比高等諸多優(yōu)點(diǎn)顯得更加突出,受到國(guó)內(nèi)外越來(lái)越多的大學(xué)和科研單位的青睞,已成為并行計(jì)算和分布式計(jì)算技術(shù)的重要發(fā)展方向。本研究結(jié)合具體的數(shù)值試驗(yàn),主要完成了以下工作:（1）探討了PVM on Win32的網(wǎng)絡(luò)并行數(shù)值計(jì)算平臺(tái)的構(gòu)架,并行環(huán)境變量的設(shè)置,PVM的啟動(dòng)及虛擬機(jī)的配置等;（2）較詳細(xì)地討論了基于PVM on Win32的網(wǎng)絡(luò)并行數(shù)值計(jì)算程序設(shè)計(jì)的方法及注意事項(xiàng),包括PVM基本編程模式的研究、任務(wù)的分解、通信的設(shè)計(jì)及PVM并行數(shù)值算法的數(shù)值穩(wěn)定性等;（3）重點(diǎn)研究了影響PVM并行程序性能的幾個(gè)重要因素,包括負(fù)載平衡、通信開(kāi)銷、網(wǎng)絡(luò)性能、任務(wù)粒度、處理機(jī)個(gè)數(shù)以及處理機(jī)內(nèi)存容量問(wèn)題等,并提出了相應(yīng)的策略,以最大限度地提高PVM并行程序的性能;（4）針對(duì)基于PVM on Win32的由桌面PC機(jī)聯(lián)網(wǎng)而成的網(wǎng)絡(luò)并行計(jì)算環(huán)境中,處理機(jī)的運(yùn)算速度較快而處理機(jī)間的通信相對(duì)較慢的實(shí)際情況,將求解線性方程組的Gauss-Jordan消去法與Ga... 

【文章來(lái)源】：貴州師范大學(xué)貴州省

【文章頁(yè)數(shù)】：45 頁(yè)

【學(xué)位級(jí)別】：碩士

【文章目錄】：
Abstract in Chinese
Abstract
1 Introduction
    1.1 The Prospect of Network Parallel Computing
    1.2 PVM and Its Characteristics
    1.3 The Current Research Status
    1.4 Main Work in This Study
    1.5 The Significance of This Study
2 Building PVM on Wi1132 Network Parallel Environments
    2.1 Installing and Setting PVM on Wi1132 System
    2.2 Starting PVM and Configuring Virtual Machine
3 Parallel Programming Under PVM on Win32 Network Parallel Environments
    3.1 Decomposing Task
    3.2 Designing Communication
    3.3 Choosing Programming Models
    3.4 Writing, Compiling and Executing PVM Parallel Programs
    3.5 Debugging PVM Parallel Programs
    3.6 Optimizing PVM Parallel Numerical Programs
4 Factors Affecting the Performance of PVM Parallel Programs
    4.1 Load-Balancing
        4.1.1 The Importance of Load-Balancing
        4.1.2 Methods for Load-Balancing
    4.2 Communication Costs
        4.2.1 Communication Costs Are the Key Factor Affecting the Performance of PVM Parallel Programs
        4.2.2 Techniques of Decreasing Communication Costs
    4.3 Network Performance
        4.3.1 Network Performance Affects the Performance of PVM Parallel Programs
        4.3.2 How to Improve Network Performance
    4.4 Task Granularity and the Number of Processors
        4.4.1 Task Granularity and the Number of Processors Affect Speedup
        4.4.2 A Strategy for Optimizing the Number of Processors
    4.5 Memory Problems
5 Network Parallel Computing Practices
    5.1 A Parallel Algorithm for Solving Dense Linear Equations
        5.1.1 Basic Idea
        5.1.2 Numerical Experiment
        5.1.3 Analysis and Enlightenment of the Experimental Results
    5.2 A Piecewise Gauss-Seidel Parallel Iterative Algorithm for Solving Linear Equations
        5.2.1 Basic Idea
        5.2.2 Numerical Experiment
        5.2.3 Analysis and Enlightenment of the Experimental Results
    5.3 A Parallel Subspace Iterative Algorithm for Eigenvalue Problems and Its Implementation on LAN
        5.3.1 Serial Subspace Iterative Algorithm
        5.3.2 Basic Idea of Parallel Algorithm
        5.3.3 Numerical Experiment
        5.3.4 Analysis and Enlightenment of the Experimental Results
6 Conclusion
Bibliography
Appendix: Published Papers of the Author
Acknowledgements
原創(chuàng)性聲明
關(guān)于學(xué)位論文使用授權(quán)的聲明


【參考文獻(xiàn)】：
期刊論文
[1]微機(jī)網(wǎng)絡(luò)環(huán)境下提高PVM并行程序性能的策略[J]. 尚月強(qiáng).  計(jì)算機(jī)工程與設(shè)計(jì). 2007(13)
[2]局域網(wǎng)上求解三角形方程組的一種并行算法[J]. 尚月強(qiáng).  計(jì)算機(jī)工程與應(yīng)用. 2007(19)

碩士論文
[1]基于主機(jī)負(fù)載預(yù)測(cè)的機(jī)群動(dòng)態(tài)任務(wù)調(diào)度策略研究[D]. 陳榮征.廣東工業(yè)大學(xué) 2008



本文編號(hào)：3641436