(3.236.214.19) 您好!臺灣時間:2021/05/06 22:09
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:許俊琛
研究生(外文):Chun-Chen Hsu
論文名稱:叢集計算機平行輸出入處理器的配置問題
論文名稱(外文):I/O Processor Allocation for Mesh Cluster Computers
指導教授:劉邦鋒
指導教授(外文):Pangfeng Liu
學位類別:碩士
校院名稱:國立臺灣大學
系所名稱:資訊工程學研究所
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:2004
畢業學年度:92
語文別:英文
論文頁數:40
中文關鍵詞:平行輸出入器配置叢集計算
外文關鍵詞:processor allocationparallel I/O
相關次數:
  • 被引用被引用:0
  • 點閱點閱:152
  • 評分評分:系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
As cluster systems become increasingly popular, more and more paralle applications require need not only computing power but also significant I/O performance. However, the I/O subsystem has become the bottleneck of the overall system performance for years due to slower improvement of the second storage devices. In recent years parallel I/O has drawn an increasing attention as a promising approach
to eliminate this bottleneck. To improve I/O efficiency of a cluster system computation tasks must be carefully assigned to processors, so that the communication overheads within the group the processors of
the task, and those I/O traffics that connect processors of the task to I/O system are both optimized. Earlier processor allocation strategies considered the optimization of communication traffic or I/O
traffic only. Since both the communication and I/O traffic can cause network contention, we develop two groups of algorithms -- binary tree based methods and Snake-Hilbert curve based methods, that address the issues of both communication and I/O traffics simultaneously. The experimental results indicate that for tasks that
have different mixture of communication and I/O traffics, our algorithms have very good performance in terms of overall parallel I/O efficiency. We also developed two mathematical evaluating criteria -- ``compactness'' and ``spatial compactness'', to determine the fitness of allocation algorithms in terms of geometrical adjacency of
processors. The theoretical results of these two criteria are also presented in this dissertation.
Contents
1 Introduction 6
2 Communication Model 10
3 Algorithms 13
3.1 Binary-Tree Allocations Group . . . . . . . . . . . . . . . . . . . . . . . . 13
3.1.1 Binary-Tree Allocation . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.1.2 Static Binary-Tree Allocation . . . . . . . . . . . . . . . . . . . . . 16
3.1.3 Dynamic Binary-Tree Allocation . . . . . . . . . . . . . . . . . . . . 16
3.1.4 Non-contiguous Binary-Tree Allocations . . . . . . . . . . . . . . . 17
3.2 Snake-Hilbert Allocations . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2.1 Snake-Hilbert Allocation . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2.2 Mirrored Snake-Hilbert Allocation . . . . . . . . . . . . . . . . . . . 22
4 Compact Allocation 24
4.1 Multiple Buddy Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.2 Spatially Compact Allocation . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.2.1 Spatially Compact Binary Tree Allocation . . . . . . . . . . . . . . 27
4.3 Compactness Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5 Experiments 30
5.1 Experimental Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.1.1 Simulation Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.1.2 Communication Patterns . . . . . . . . . . . . . . . . . . . . . . . . 31
5.2 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
5.2.1 The Impacts of Con‾gurable Parameters . . . . . . . . . . . . . . . 31
5.2.2 Result Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6 Conclusions 39
[1] T. Anderson, D. Culler, and D.Patterson. A case for networks of workstations (now).
In IEEE Micro, pages 54{64, Feb 1995.
[2] http://www.top500.org/lists/2003/11/press-release.php.
[3] http://computing.vt.edu/research computing/terascale/pressrelease.html.
[4] http://www.in‾nibandta.org/home.
[5] Applications Working Group of the Scalable I/O Initiative. Preliminary Survey of on I/O Intensive Applications. http://www.cacr.caltech.edu/SIO/pubs/sio apps.ps.
[6] V.M. Lo, K. Windisch, W. Liu, and B. Nitzberg. Non-contiguous processor allocation algorithms for mesh-connected multicomputers. IEEE Transactions on Parallel and
Distributed Systems, 8(7):712{726, July 1997.
[7] Vitus J. Leung, Esther M. Arkin, Michael A. Bender, Jeanette Johnston, Alok Lal,Joseph S. B. Mitchell, Cynthia Phillips, and Steven S. Seiden. Processor allocation on
cplant: Achieving general processor locality using one-dimensional allocation strategies. In Proceedings of the 4th IEEE International Conference on Cluster Computing,
pages 296-304, 2002.
[8] Jens Mache, Virginia Lo, and Sharad Garg. How to schedule parallel I/O intensive jobs. In Proceedings of the 6th Conference on Parallel and Real-Time Systems, 1999.
[9] Jens Mache, Virginia Lo, Marilynn Livingston, and Sharad Garg. The impact of spatial layout of jobs on parallel I/O performance. In Proceedings of the 6th Workshop on I/O in Parallel and Distributed Systems, 1999.
[10] Y. Cho, M. Winslett, S. Kuo, Y. Chen, J. Lee, and K. Motukuri. Parallel I/O on networks of workstations: Performance improvement by careful placement of i/o
servers. In Proceedings of High Performance Computing on Hewlett-Packard Systems,1998.
[11] Sharad Garg. Parallel I/O architecture of the first asci tflops machine. In Proceedings of Intel Supercomputer Users Group, 1997.
[12] D. Hilbert. ÄUber die stetige Abbildung einer Linie auf ein FlÄachenstÄuck. Mathematische Annalen, 38:59{60, 1891.
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
系統版面圖檔 系統版面圖檔