Shared vs. Snoop: Evaluation of cache structure for single-chip multiprocessors

Toru Kisuki, Masaki Wakabayashi, Junji Yamamoto, Keisuke Inoue, Hideharu Amano

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

3 Citations (Scopus)

Abstract

The shared cache structures and snoop cache structures for single-chip multiprocessors are evaluated and compared using an instruction level simulator. Simulation results show that 1-port large shared cache achieves the best performance if there is no delay penalty for arbitration and accessing the bus. However, if 1-clock delay is assumed for accessing the shared cache, a snoop cache with internal wide bus and invalidate style NewKeio protocol overcomes shared caches.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages793-797
Number of pages5
Volume1300 LNCS
Publication statusPublished - 1997
Event3rd International Conference on Parallel Processing, Euro-Par 1997 - Passau, Germany
Duration: 1997 Aug 261997 Aug 29

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1300 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other3rd International Conference on Parallel Processing, Euro-Par 1997
CountryGermany
CityPassau
Period97/8/2697/8/29

Fingerprint

Chip multiprocessors
Cache
Evaluation
Clocks
Simulators
Arbitration
Penalty
Simulator
Internal
Simulation

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Kisuki, T., Wakabayashi, M., Yamamoto, J., Inoue, K., & Amano, H. (1997). Shared vs. Snoop: Evaluation of cache structure for single-chip multiprocessors. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1300 LNCS, pp. 793-797). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1300 LNCS).

Shared vs. Snoop : Evaluation of cache structure for single-chip multiprocessors. / Kisuki, Toru; Wakabayashi, Masaki; Yamamoto, Junji; Inoue, Keisuke; Amano, Hideharu.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1300 LNCS 1997. p. 793-797 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1300 LNCS).

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

Kisuki, T, Wakabayashi, M, Yamamoto, J, Inoue, K & Amano, H 1997, Shared vs. Snoop: Evaluation of cache structure for single-chip multiprocessors. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 1300 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1300 LNCS, pp. 793-797, 3rd International Conference on Parallel Processing, Euro-Par 1997, Passau, Germany, 97/8/26.
Kisuki T, Wakabayashi M, Yamamoto J, Inoue K, Amano H. Shared vs. Snoop: Evaluation of cache structure for single-chip multiprocessors. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1300 LNCS. 1997. p. 793-797. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Kisuki, Toru ; Wakabayashi, Masaki ; Yamamoto, Junji ; Inoue, Keisuke ; Amano, Hideharu. / Shared vs. Snoop : Evaluation of cache structure for single-chip multiprocessors. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1300 LNCS 1997. pp. 793-797 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{1ab5af5ae80e4b7780da0b53f93d3acd,
title = "Shared vs. Snoop: Evaluation of cache structure for single-chip multiprocessors",
abstract = "The shared cache structures and snoop cache structures for single-chip multiprocessors are evaluated and compared using an instruction level simulator. Simulation results show that 1-port large shared cache achieves the best performance if there is no delay penalty for arbitration and accessing the bus. However, if 1-clock delay is assumed for accessing the shared cache, a snoop cache with internal wide bus and invalidate style NewKeio protocol overcomes shared caches.",
author = "Toru Kisuki and Masaki Wakabayashi and Junji Yamamoto and Keisuke Inoue and Hideharu Amano",
year = "1997",
language = "English",
isbn = "9783540634409",
volume = "1300 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "793--797",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Shared vs. Snoop

T2 - Evaluation of cache structure for single-chip multiprocessors

AU - Kisuki, Toru

AU - Wakabayashi, Masaki

AU - Yamamoto, Junji

AU - Inoue, Keisuke

AU - Amano, Hideharu

PY - 1997

Y1 - 1997

N2 - The shared cache structures and snoop cache structures for single-chip multiprocessors are evaluated and compared using an instruction level simulator. Simulation results show that 1-port large shared cache achieves the best performance if there is no delay penalty for arbitration and accessing the bus. However, if 1-clock delay is assumed for accessing the shared cache, a snoop cache with internal wide bus and invalidate style NewKeio protocol overcomes shared caches.

AB - The shared cache structures and snoop cache structures for single-chip multiprocessors are evaluated and compared using an instruction level simulator. Simulation results show that 1-port large shared cache achieves the best performance if there is no delay penalty for arbitration and accessing the bus. However, if 1-clock delay is assumed for accessing the shared cache, a snoop cache with internal wide bus and invalidate style NewKeio protocol overcomes shared caches.

UR - http://www.scopus.com/inward/record.url?scp=84882659090&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84882659090&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84882659090

SN - 9783540634409

VL - 1300 LNCS

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

SP - 793

EP - 797

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

ER -