Memory conscious scheduling for cluster-based NUMA multiprocessors

Takahiro Koita, Tetsuro Katayama, Keizo Saisho, Akira Fukuda

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

For Non-Uniform Memory Access (NUMA) multiprocessors, memory access overhead is crucial to system performance. Processor scheduling and page placement schemes, dominant factors of memory access overhead, are closely related. In particular, if the processor scheduling scheme is dynamic space-sharing, it should be considered together with the page placement scheme for efficient process execution. Most research in this area, however, has focused exclusively on either the processor scheduling scheme or the page placement scheme alone without considering the interaction between the two. This paper proposes several policies for cluster-based NUMA multiprocessors that are combinations of a processor scheduling scheme and a page placement scheme and investigates the interaction between them. The simulation results show that policies that cooperate to employ the home-cluster concept achieve the best performance. The paper also compares the best of the proposed policies with other existing dynamic processor scheduling policies. Based on our study reported here, the best policy is found to perform better than other existing policies.

Original languageEnglish
Pages (from-to)217-235
Number of pages19
JournalJournal of Supercomputing
Volume16
Issue number3
DOIs
Publication statusPublished - Jan 1 2000
Externally publishedYes

Fingerprint

Multiprocessor
Scheduling
Data storage equipment
Placement
Scheduling Policy
Interaction
System Performance
Sharing
Policy
Simulation

All Science Journal Classification (ASJC) codes

  • Software
  • Theoretical Computer Science
  • Information Systems
  • Hardware and Architecture

Cite this

Memory conscious scheduling for cluster-based NUMA multiprocessors. / Koita, Takahiro; Katayama, Tetsuro; Saisho, Keizo; Fukuda, Akira.

In: Journal of Supercomputing, Vol. 16, No. 3, 01.01.2000, p. 217-235.

Research output: Contribution to journalArticle

Koita, Takahiro ; Katayama, Tetsuro ; Saisho, Keizo ; Fukuda, Akira. / Memory conscious scheduling for cluster-based NUMA multiprocessors. In: Journal of Supercomputing. 2000 ; Vol. 16, No. 3. pp. 217-235.
@article{291a3580468942418d22e9783cce9756,
title = "Memory conscious scheduling for cluster-based NUMA multiprocessors",
abstract = "For Non-Uniform Memory Access (NUMA) multiprocessors, memory access overhead is crucial to system performance. Processor scheduling and page placement schemes, dominant factors of memory access overhead, are closely related. In particular, if the processor scheduling scheme is dynamic space-sharing, it should be considered together with the page placement scheme for efficient process execution. Most research in this area, however, has focused exclusively on either the processor scheduling scheme or the page placement scheme alone without considering the interaction between the two. This paper proposes several policies for cluster-based NUMA multiprocessors that are combinations of a processor scheduling scheme and a page placement scheme and investigates the interaction between them. The simulation results show that policies that cooperate to employ the home-cluster concept achieve the best performance. The paper also compares the best of the proposed policies with other existing dynamic processor scheduling policies. Based on our study reported here, the best policy is found to perform better than other existing policies.",
author = "Takahiro Koita and Tetsuro Katayama and Keizo Saisho and Akira Fukuda",
year = "2000",
month = "1",
day = "1",
doi = "10.1023/A:1008165010732",
language = "English",
volume = "16",
pages = "217--235",
journal = "Journal of Supercomputing",
issn = "0920-8542",
publisher = "Springer Netherlands",
number = "3",

}

TY - JOUR

T1 - Memory conscious scheduling for cluster-based NUMA multiprocessors

AU - Koita, Takahiro

AU - Katayama, Tetsuro

AU - Saisho, Keizo

AU - Fukuda, Akira

PY - 2000/1/1

Y1 - 2000/1/1

N2 - For Non-Uniform Memory Access (NUMA) multiprocessors, memory access overhead is crucial to system performance. Processor scheduling and page placement schemes, dominant factors of memory access overhead, are closely related. In particular, if the processor scheduling scheme is dynamic space-sharing, it should be considered together with the page placement scheme for efficient process execution. Most research in this area, however, has focused exclusively on either the processor scheduling scheme or the page placement scheme alone without considering the interaction between the two. This paper proposes several policies for cluster-based NUMA multiprocessors that are combinations of a processor scheduling scheme and a page placement scheme and investigates the interaction between them. The simulation results show that policies that cooperate to employ the home-cluster concept achieve the best performance. The paper also compares the best of the proposed policies with other existing dynamic processor scheduling policies. Based on our study reported here, the best policy is found to perform better than other existing policies.

AB - For Non-Uniform Memory Access (NUMA) multiprocessors, memory access overhead is crucial to system performance. Processor scheduling and page placement schemes, dominant factors of memory access overhead, are closely related. In particular, if the processor scheduling scheme is dynamic space-sharing, it should be considered together with the page placement scheme for efficient process execution. Most research in this area, however, has focused exclusively on either the processor scheduling scheme or the page placement scheme alone without considering the interaction between the two. This paper proposes several policies for cluster-based NUMA multiprocessors that are combinations of a processor scheduling scheme and a page placement scheme and investigates the interaction between them. The simulation results show that policies that cooperate to employ the home-cluster concept achieve the best performance. The paper also compares the best of the proposed policies with other existing dynamic processor scheduling policies. Based on our study reported here, the best policy is found to perform better than other existing policies.

UR - http://www.scopus.com/inward/record.url?scp=0034227688&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0034227688&partnerID=8YFLogxK

U2 - 10.1023/A:1008165010732

DO - 10.1023/A:1008165010732

M3 - Article

AN - SCOPUS:0034227688

VL - 16

SP - 217

EP - 235

JO - Journal of Supercomputing

JF - Journal of Supercomputing

SN - 0920-8542

IS - 3

ER -