Assessment of COMA multiprocessors

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

2 Citations (Scopus)

Abstract

In Cache Only Memory Architecture (COMA) for distributed shared memory multiprocessors, the physical location of a datum is completely decoupled from its address by organizing the memory local to each node as a cache for shared address space. As in traditional cache-coherent multiprocessors, the overhead involved in coherence enforcement strongly affects the performance. The overhead seems more pronounced in a COMA machine than the traditional multiprocessors because of its relatively huge size of the local memory acting as cache and the level of memory hierarchy at which coherence needs to be enforced. Using trace driven simulations of the Perfect Club Benchmark Suite, this paper studies the change in the miss ratio and the network traffic with the two coherence policies, update and invalidate. Our study shows that the two policies provide distinct characteristics which reveal different opportunity of improving COMA multiprocessors for different choice of coherence policy.

Original languageEnglish
Title of host publicationIEEE Symposium on Parallel and Distributed Processing - Proceedings
Editors Anon
PublisherIEEE
Pages388-392
Number of pages5
Publication statusPublished - 1995
Externally publishedYes
EventProceedings of the IEEE 9th International Parallel Processing Symposium - Santa Barbara, CA, USA
Duration: 1995 Apr 251995 Apr 28

Other

OtherProceedings of the IEEE 9th International Parallel Processing Symposium
CitySanta Barbara, CA, USA
Period95/4/2595/4/28

Fingerprint

Memory architecture
Data storage equipment

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Lee, K. H. (1995). Assessment of COMA multiprocessors. In Anon (Ed.), IEEE Symposium on Parallel and Distributed Processing - Proceedings (pp. 388-392). IEEE.

Assessment of COMA multiprocessors. / Lee, Kyung Ho.

IEEE Symposium on Parallel and Distributed Processing - Proceedings. ed. / Anon. IEEE, 1995. p. 388-392.

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

Lee, KH 1995, Assessment of COMA multiprocessors. in Anon (ed.), IEEE Symposium on Parallel and Distributed Processing - Proceedings. IEEE, pp. 388-392, Proceedings of the IEEE 9th International Parallel Processing Symposium, Santa Barbara, CA, USA, 95/4/25.
Lee KH. Assessment of COMA multiprocessors. In Anon, editor, IEEE Symposium on Parallel and Distributed Processing - Proceedings. IEEE. 1995. p. 388-392
Lee, Kyung Ho. / Assessment of COMA multiprocessors. IEEE Symposium on Parallel and Distributed Processing - Proceedings. editor / Anon. IEEE, 1995. pp. 388-392
@inproceedings{f6e11a835c064e309e1af7ca66c453a1,
title = "Assessment of COMA multiprocessors",
abstract = "In Cache Only Memory Architecture (COMA) for distributed shared memory multiprocessors, the physical location of a datum is completely decoupled from its address by organizing the memory local to each node as a cache for shared address space. As in traditional cache-coherent multiprocessors, the overhead involved in coherence enforcement strongly affects the performance. The overhead seems more pronounced in a COMA machine than the traditional multiprocessors because of its relatively huge size of the local memory acting as cache and the level of memory hierarchy at which coherence needs to be enforced. Using trace driven simulations of the Perfect Club Benchmark Suite, this paper studies the change in the miss ratio and the network traffic with the two coherence policies, update and invalidate. Our study shows that the two policies provide distinct characteristics which reveal different opportunity of improving COMA multiprocessors for different choice of coherence policy.",
author = "Lee, {Kyung Ho}",
year = "1995",
language = "English",
pages = "388--392",
editor = "Anon",
booktitle = "IEEE Symposium on Parallel and Distributed Processing - Proceedings",
publisher = "IEEE",

}

TY - GEN

T1 - Assessment of COMA multiprocessors

AU - Lee, Kyung Ho

PY - 1995

Y1 - 1995

N2 - In Cache Only Memory Architecture (COMA) for distributed shared memory multiprocessors, the physical location of a datum is completely decoupled from its address by organizing the memory local to each node as a cache for shared address space. As in traditional cache-coherent multiprocessors, the overhead involved in coherence enforcement strongly affects the performance. The overhead seems more pronounced in a COMA machine than the traditional multiprocessors because of its relatively huge size of the local memory acting as cache and the level of memory hierarchy at which coherence needs to be enforced. Using trace driven simulations of the Perfect Club Benchmark Suite, this paper studies the change in the miss ratio and the network traffic with the two coherence policies, update and invalidate. Our study shows that the two policies provide distinct characteristics which reveal different opportunity of improving COMA multiprocessors for different choice of coherence policy.

AB - In Cache Only Memory Architecture (COMA) for distributed shared memory multiprocessors, the physical location of a datum is completely decoupled from its address by organizing the memory local to each node as a cache for shared address space. As in traditional cache-coherent multiprocessors, the overhead involved in coherence enforcement strongly affects the performance. The overhead seems more pronounced in a COMA machine than the traditional multiprocessors because of its relatively huge size of the local memory acting as cache and the level of memory hierarchy at which coherence needs to be enforced. Using trace driven simulations of the Perfect Club Benchmark Suite, this paper studies the change in the miss ratio and the network traffic with the two coherence policies, update and invalidate. Our study shows that the two policies provide distinct characteristics which reveal different opportunity of improving COMA multiprocessors for different choice of coherence policy.

UR - http://www.scopus.com/inward/record.url?scp=0029217708&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0029217708&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0029217708

SP - 388

EP - 392

BT - IEEE Symposium on Parallel and Distributed Processing - Proceedings

A2 - Anon, null

PB - IEEE

ER -