Load balancing in distributed shared memory systems

An Chow Lai, Ce-Kuen Shieh, Yih Tzye Kok, Jyh Chang Ueng, Ling Yang Kung

Research output: Chapter in Book/Report/Conference proceedingConference contribution

20 Citations (Scopus)

Abstract

Despite the fast evolution of Distributed Shared Memory (DSM) systems, the load balance problem has not received enough attentions. This problem arises naturally after multithreading was introduced to DSM systems few years ago. The cognizance of it would bring us a significant improvement in system performance. In this paper, we address it by proposing and experimentally evaluating a load balancing method called Dependence-Driven Load Balancing (DDLB) that is dedicated for multithreaded DSM systems. The most attractive feature of this method is to take thread dependence into account in making decisions for migration. In contrast to existing thread scheduling works which for the most part have relied on the information supplied before execution, we require no a priori information. Typically, DDLB embraces three policies, i.e. transfer policy, location policy and selection policy, and applies affinity scheduling. Finally, from the experimental results, the performance of the system with load balancing is improving evidently.

Original languageEnglish
Title of host publicationIEEE International Performance, Computing & Communications Conference, Proceedings
Editors Anon
PublisherIEEE
Pages152-158
Number of pages7
Publication statusPublished - 1997
EventProceedings of the 1997 IEEE International Performance Computing & Communications Conference - Phoenix, AZ, USA
Duration: 1997 Feb 51997 Feb 7

Other

OtherProceedings of the 1997 IEEE International Performance Computing & Communications Conference
CityPhoenix, AZ, USA
Period97-02-0597-02-07

Fingerprint

Resource allocation
Computer systems
Data storage equipment
Scheduling
Decision making

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

Lai, A. C., Shieh, C-K., Kok, Y. T., Ueng, J. C., & Kung, L. Y. (1997). Load balancing in distributed shared memory systems. In Anon (Ed.), IEEE International Performance, Computing & Communications Conference, Proceedings (pp. 152-158). IEEE.
Lai, An Chow ; Shieh, Ce-Kuen ; Kok, Yih Tzye ; Ueng, Jyh Chang ; Kung, Ling Yang. / Load balancing in distributed shared memory systems. IEEE International Performance, Computing & Communications Conference, Proceedings. editor / Anon. IEEE, 1997. pp. 152-158
@inproceedings{832f2907c8e34746ad8080e4b985766d,
title = "Load balancing in distributed shared memory systems",
abstract = "Despite the fast evolution of Distributed Shared Memory (DSM) systems, the load balance problem has not received enough attentions. This problem arises naturally after multithreading was introduced to DSM systems few years ago. The cognizance of it would bring us a significant improvement in system performance. In this paper, we address it by proposing and experimentally evaluating a load balancing method called Dependence-Driven Load Balancing (DDLB) that is dedicated for multithreaded DSM systems. The most attractive feature of this method is to take thread dependence into account in making decisions for migration. In contrast to existing thread scheduling works which for the most part have relied on the information supplied before execution, we require no a priori information. Typically, DDLB embraces three policies, i.e. transfer policy, location policy and selection policy, and applies affinity scheduling. Finally, from the experimental results, the performance of the system with load balancing is improving evidently.",
author = "Lai, {An Chow} and Ce-Kuen Shieh and Kok, {Yih Tzye} and Ueng, {Jyh Chang} and Kung, {Ling Yang}",
year = "1997",
language = "English",
pages = "152--158",
editor = "Anon",
booktitle = "IEEE International Performance, Computing & Communications Conference, Proceedings",
publisher = "IEEE",

}

Lai, AC, Shieh, C-K, Kok, YT, Ueng, JC & Kung, LY 1997, Load balancing in distributed shared memory systems. in Anon (ed.), IEEE International Performance, Computing & Communications Conference, Proceedings. IEEE, pp. 152-158, Proceedings of the 1997 IEEE International Performance Computing & Communications Conference, Phoenix, AZ, USA, 97-02-05.

Load balancing in distributed shared memory systems. / Lai, An Chow; Shieh, Ce-Kuen; Kok, Yih Tzye; Ueng, Jyh Chang; Kung, Ling Yang.

IEEE International Performance, Computing & Communications Conference, Proceedings. ed. / Anon. IEEE, 1997. p. 152-158.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Load balancing in distributed shared memory systems

AU - Lai, An Chow

AU - Shieh, Ce-Kuen

AU - Kok, Yih Tzye

AU - Ueng, Jyh Chang

AU - Kung, Ling Yang

PY - 1997

Y1 - 1997

N2 - Despite the fast evolution of Distributed Shared Memory (DSM) systems, the load balance problem has not received enough attentions. This problem arises naturally after multithreading was introduced to DSM systems few years ago. The cognizance of it would bring us a significant improvement in system performance. In this paper, we address it by proposing and experimentally evaluating a load balancing method called Dependence-Driven Load Balancing (DDLB) that is dedicated for multithreaded DSM systems. The most attractive feature of this method is to take thread dependence into account in making decisions for migration. In contrast to existing thread scheduling works which for the most part have relied on the information supplied before execution, we require no a priori information. Typically, DDLB embraces three policies, i.e. transfer policy, location policy and selection policy, and applies affinity scheduling. Finally, from the experimental results, the performance of the system with load balancing is improving evidently.

AB - Despite the fast evolution of Distributed Shared Memory (DSM) systems, the load balance problem has not received enough attentions. This problem arises naturally after multithreading was introduced to DSM systems few years ago. The cognizance of it would bring us a significant improvement in system performance. In this paper, we address it by proposing and experimentally evaluating a load balancing method called Dependence-Driven Load Balancing (DDLB) that is dedicated for multithreaded DSM systems. The most attractive feature of this method is to take thread dependence into account in making decisions for migration. In contrast to existing thread scheduling works which for the most part have relied on the information supplied before execution, we require no a priori information. Typically, DDLB embraces three policies, i.e. transfer policy, location policy and selection policy, and applies affinity scheduling. Finally, from the experimental results, the performance of the system with load balancing is improving evidently.

UR - http://www.scopus.com/inward/record.url?scp=0030714241&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0030714241&partnerID=8YFLogxK

M3 - Conference contribution

SP - 152

EP - 158

BT - IEEE International Performance, Computing & Communications Conference, Proceedings

A2 - Anon, null

PB - IEEE

ER -

Lai AC, Shieh C-K, Kok YT, Ueng JC, Kung LY. Load balancing in distributed shared memory systems. In Anon, editor, IEEE International Performance, Computing & Communications Conference, Proceedings. IEEE. 1997. p. 152-158