跳到主要內容

臺灣博碩士論文加值系統

(44.192.20.240) 您好!臺灣時間:2024/02/26 02:03
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:黃國展
研究生(外文):Huang, Kuo-Chan
論文名稱:分散式記憶體平行電腦上適應性資料平行計算之研究
論文名稱(外文):Adaptive Data-Parallel Computations on Distributed-Memory Multicomputers
指導教授:王豐堅
指導教授(外文):Feng-Jian Wang
學位類別:博士
校院名稱:國立交通大學
系所名稱:資訊工程學系
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:1998
畢業學年度:86
語文別:中文
論文頁數:103
中文關鍵詞:資料平行計算分散式記憶體平行電腦動態性平行系統適應性計算軟體設計模式物件導向技術
外文關鍵詞:data-parallel computationdistributed-memory multicomputerdynamic parallel systemadaptive computationsoftware design patternobject-oriented technology
相關次數:
  • 被引用被引用:1
  • 點閱點閱:176
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
本論文描述在分散式記憶體平行電腦上關於資料平行計算之適應性排程
的研究成果。此研究著重在平行系統的動態行為及其對執行效能的影響上
。我們發展了一個適應性資料平行計算模型(ADPCM),用來有效地描述程
式及系統的動態行為。奠基於ADPCM,我們探討了適應性資料平行計算中
兩個重要的課題:適應性處理器配置(APA)及資料分割與發送(DPD) 。本
論文採用解析與實驗並重的方式來探討這兩個問題。實驗結果顯示適應性
計算的技術能夠有效地增進動態性平行系統的執行效能。根據這些實驗,
我們發展出一個確定性效能模擬方法(DPSM),以量化的方式來有效地分析
及評估不同的處理器配置及資料分割與發送的方法。藉由DPSM的輔助,我
們發展出一組高效率的演算法及經驗法則,可用來解決ADPCM模式之平行
系統上的適應性處理器配置及資料分割與發送的問題。此外,適應性平行
程式的發展遠比一般平行程式來得複雜,為了幫助程式發展者有效地開發
適應性平行程式,本論文也提出了兩個軟體設計模式(software design
pattern),用來輔助平行程式發展中工作管理(task management)及資料
分割與發送(data partition and distribution)的問題。這兩個模式採
用物件導向的技術,可降低軟體複雜度並促進設計層次的軟體再利用(
software reuse)。 總而言之,本論文中的解析性與實驗性研究促進
了對動態性平行系統行為模式的了解。由於奠基於實際計算環境中的實際
平行系統,所發展出來的方法可有效地改進寬鬆性同步系統(loosely-
synchronous systems) 的執行效能,並可被擴充應用至其他種類的平行
系統上。
This thesis presents adaptively scheduling data-parallel
computation onto distributed-memory multicomputers. An Adaptive
Data-Parallel Computation Model (ADPCM) developed provides
abstract representation of program and system behavior. Based on
ADPCM, two important issues in adaptive data-parallel
computation, the Adaptive Processor Allocation (ADP) and Data
Partition and Distribution (DPD) problems, are addressed. A
combination of both analytical and experimental techniques is
applied to the problems. Experimental results indicate that
adaptive computation can effectively improve the runtime
performance of dynamic parallel systems. A deterministic
performance simulation method (DPSM) developed provides
qualitative insight as well as efficient quantitative evaluation
of different APA and DPD methods. Based on DPSM, a set of
efficient algorithms and heuristics have been developed to solve
the APA and DPD problems for ADPCM based systems. To effectively
develop adaptive programs, this thesis presents two object-
oriented software design patterns addressing the issues of task
management and data partition and distribution in parallel
programming. In summary, the analytical and experimental
results in the thesis contribute towards the understanding of
fundamental behaviors of dynamic parallel systems. Based on real
applications running on real computing environments,the
developed approaches can improve runtime performance of loosely-
synchronous systems, and can be easily extended for other kinds
of parallel systems.
Cover
Abstract (Chinese)
Abstract (English)
Acknowledgment
Table of Contents
List of Figures
List of Tables
Chapter 1 Introduction
Chapter 2 Preliminaries and Previous Work
2.1 A Framework for Task Scheduling in Parallel and Distributed Environments
2.2 The Task Allocation Problem
2.3 Load Divisibility
2.4 Support of Parallel Programming
Chapter 3 Adaptive Date-Parallel Computation Model,ADPCM
3.1 Background
3.2 The Spatial Scheme
3.3 The Temporal Scheme
3.4 Performance of ADPCM Based Systems
Chapter 4 An Experimental Study of Adaptive Processor Allocation in ADPCM
4.1 Adaptive Processor Allocation Subject to Internal Workload Variation
4.2 An Example Dynamic Parallel System
4.3 An Experimental Study
Chapter 5 A Systematic Approach to the APA Problem
5.1 A Deterministic Performance Simulation Method(DPSM)
5.2 Evaluation of DPSM
5.3 Formulation of the APA Problem
5.4 Approaches to the Formulated APA Problem
5.5 Discussion
Chapter 6 Data Partition and Distribution of Matrix Computation
6.1 Extemal Environmental Variation
6.2 Basic Linear Algebra Subprograms for Matrix Computation
6.3 Issues of Heterogeneous and Time-Sharing Parallel Computing Environments
6.4 Parallel Matrix Multiplication Based on ADPCM
6.5 LU Factorization
6.6 Discussion
Chapter 7 Design Patterns Supporting ADPCM
7.1 The Need of Support for Parallel Programming
7.2 Software Design Pattem
7.3 The Computing Space Pattem
7.4 The Group Communication Pattem
7.5 Discussion
Chapter 8 Conclusions and Future Work
8.1 Contributions of the Thesis
8.2 Directions for Future Research
References
Appendix A Definitions of Key Terms
Appendix B Data for the Evaluation of DPSM
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top