A comparative study of memory structures for DSM systems on wireless environments

Authors:
Hsiao-Hsi Wang;Kuang-Jui Wang;Chien-Lung Chou;Ssu-Hsuan Lu;Kuan-Ching Li
Affiliations:
Parallel and Distributed Processing Center, Dept. of Computer Science and Information Management, Providence University Shalu, Taichung, Taiwan;Parallel and Distributed Processing Center, Dept. of Computer Science and Information Management, Providence University Shalu, Taichung, Taiwan;Parallel and Distributed Processing Center, Dept. of Computer Science and Information Management, Providence University Shalu, Taichung, Taiwan;Parallel and Distributed Processing Center, Dept. of Computer Science and Information Management, Providence University Shalu, Taichung, Taiwan;Parallel and Distributed Processing Center, Dept. of Computer Science and Information Management, Providence University Shalu, Taichung, Taiwan
Venue:
ICCS'06 Proceedings of the 6th international conference on Computational Science - Volume Part I
Year:
2006

Citing 3
Cited 0

The SPLASH-2 programs: characterization and methodological considerations

ISCA '95 Proceedings of the 22nd annual international symposium on Computer architecture
Reducing System Overheads in Home-based Software DSMs

IPPS '99/SPDP '99 Proceedings of the 13th International Symposium on Parallel Processing and the 10th Symposium on Parallel and Distributed Processing
A Proposal of a Memory Management Architecture for Mobile Computing Environments

DEXA '00 Proceedings of the 11th International Workshop on Database and Expert Systems Applications

Quantified Score

Hi-index	0.00

Visualization

Abstract

In recent years, wireless networking has become more popular than ever. Distributed shared memory (DSM) combines computer hardware resources in order to achieve the efficiency and high performance provided by parallel computing. Unfortunately, the major overhead of DSM software is the communication time, especially in wireless network environments. In this paper, we have implemented three memory structures over JIAJIA DSM system on wireless network, and then analyzed their performance. From experimental results, we could find out the relation between communication time and memory layouts. In addition, we have also discovered relationship between characteristics of application programs and memory structures. Experimental results of five well-known benchmark applications show that a suitable memory layout can effectively reduce communication overhead in wireless network. We have analyzed advantages and disadvantages of these memory structures, to improve future designs of wireless DSM systems.