Relaxing the inclusion property in cache only memory architecture

Jinseok Kong, Kyung Ho Lee

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

3 Citations (Scopus)

Abstract

Cache only memory architecture (COMA), even with its additional memory overhead, can incur longer inter/intra-node communication latency than cache-coherent nonuniform memory access (CCNUMA). Some studies on COMA suggest that the inclusion property applied between the processor cache and its local memory is one of the major causes of less-than-desirable performance. The inclusion property creates extra accesses to the slow local memory. We consider the binding time of data address to the local memory to be an important factor related to the long latency in COMA. This paper considers the inclusion property in COMA and introduces a variant of COMA, dubbed Dynamic Memory Architecture (DYMA), where the local memory is utilized as a backing store for blocks discarded from the processor cache. Thus, by delaying the binding time, the long latency due to the inclusion property can be avoided. This paper examines the potential performance of DYMA compared to COMA and CC-NUMA.

Original languageEnglish
Title of host publicationEuro-Par 1996 Parallel Processing - 2nd International Euro-Par Conference, Proceedings
PublisherSpringer Verlag
Pages435-444
Number of pages10
Volume1124
ISBN (Print)3540616276, 9783540616276
Publication statusPublished - 1996
Externally publishedYes
Event2nd International Euro-Par Conference on Parallel Processing, Euro-Par 1996 - Lyon, France
Duration: 1996 Aug 261996 Aug 29

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1124
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other2nd International Euro-Par Conference on Parallel Processing, Euro-Par 1996
CountryFrance
CityLyon
Period96/8/2696/8/29

Fingerprint

Memory architecture
Cache
Inclusion
Data storage equipment
Latency
Architecture
Communication

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Kong, J., & Lee, K. H. (1996). Relaxing the inclusion property in cache only memory architecture. In Euro-Par 1996 Parallel Processing - 2nd International Euro-Par Conference, Proceedings (Vol. 1124, pp. 435-444). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1124). Springer Verlag.

Relaxing the inclusion property in cache only memory architecture. / Kong, Jinseok; Lee, Kyung Ho.

Euro-Par 1996 Parallel Processing - 2nd International Euro-Par Conference, Proceedings. Vol. 1124 Springer Verlag, 1996. p. 435-444 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1124).

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

Kong, J & Lee, KH 1996, Relaxing the inclusion property in cache only memory architecture. in Euro-Par 1996 Parallel Processing - 2nd International Euro-Par Conference, Proceedings. vol. 1124, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1124, Springer Verlag, pp. 435-444, 2nd International Euro-Par Conference on Parallel Processing, Euro-Par 1996, Lyon, France, 96/8/26.
Kong J, Lee KH. Relaxing the inclusion property in cache only memory architecture. In Euro-Par 1996 Parallel Processing - 2nd International Euro-Par Conference, Proceedings. Vol. 1124. Springer Verlag. 1996. p. 435-444. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Kong, Jinseok ; Lee, Kyung Ho. / Relaxing the inclusion property in cache only memory architecture. Euro-Par 1996 Parallel Processing - 2nd International Euro-Par Conference, Proceedings. Vol. 1124 Springer Verlag, 1996. pp. 435-444 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{df865c62fc01442f85698bb89820022c,
title = "Relaxing the inclusion property in cache only memory architecture",
abstract = "Cache only memory architecture (COMA), even with its additional memory overhead, can incur longer inter/intra-node communication latency than cache-coherent nonuniform memory access (CCNUMA). Some studies on COMA suggest that the inclusion property applied between the processor cache and its local memory is one of the major causes of less-than-desirable performance. The inclusion property creates extra accesses to the slow local memory. We consider the binding time of data address to the local memory to be an important factor related to the long latency in COMA. This paper considers the inclusion property in COMA and introduces a variant of COMA, dubbed Dynamic Memory Architecture (DYMA), where the local memory is utilized as a backing store for blocks discarded from the processor cache. Thus, by delaying the binding time, the long latency due to the inclusion property can be avoided. This paper examines the potential performance of DYMA compared to COMA and CC-NUMA.",
author = "Jinseok Kong and Lee, {Kyung Ho}",
year = "1996",
language = "English",
isbn = "3540616276",
volume = "1124",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "435--444",
booktitle = "Euro-Par 1996 Parallel Processing - 2nd International Euro-Par Conference, Proceedings",

}

TY - GEN

T1 - Relaxing the inclusion property in cache only memory architecture

AU - Kong, Jinseok

AU - Lee, Kyung Ho

PY - 1996

Y1 - 1996

N2 - Cache only memory architecture (COMA), even with its additional memory overhead, can incur longer inter/intra-node communication latency than cache-coherent nonuniform memory access (CCNUMA). Some studies on COMA suggest that the inclusion property applied between the processor cache and its local memory is one of the major causes of less-than-desirable performance. The inclusion property creates extra accesses to the slow local memory. We consider the binding time of data address to the local memory to be an important factor related to the long latency in COMA. This paper considers the inclusion property in COMA and introduces a variant of COMA, dubbed Dynamic Memory Architecture (DYMA), where the local memory is utilized as a backing store for blocks discarded from the processor cache. Thus, by delaying the binding time, the long latency due to the inclusion property can be avoided. This paper examines the potential performance of DYMA compared to COMA and CC-NUMA.

AB - Cache only memory architecture (COMA), even with its additional memory overhead, can incur longer inter/intra-node communication latency than cache-coherent nonuniform memory access (CCNUMA). Some studies on COMA suggest that the inclusion property applied between the processor cache and its local memory is one of the major causes of less-than-desirable performance. The inclusion property creates extra accesses to the slow local memory. We consider the binding time of data address to the local memory to be an important factor related to the long latency in COMA. This paper considers the inclusion property in COMA and introduces a variant of COMA, dubbed Dynamic Memory Architecture (DYMA), where the local memory is utilized as a backing store for blocks discarded from the processor cache. Thus, by delaying the binding time, the long latency due to the inclusion property can be avoided. This paper examines the potential performance of DYMA compared to COMA and CC-NUMA.

UR - http://www.scopus.com/inward/record.url?scp=84947904656&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84947904656&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84947904656

SN - 3540616276

SN - 9783540616276

VL - 1124

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 435

EP - 444

BT - Euro-Par 1996 Parallel Processing - 2nd International Euro-Par Conference, Proceedings

PB - Springer Verlag

ER -