On the performance of multicomputer interconnection networks

Mohamed Ould-Khaoua, Samia Loucif, Fethi A. Rabhi

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Several researchers have analysed the performance of k-ary n-cubes taking into account channel bandwidth constraints imposed by implementation technology, namely the constant wiring density and pin-out constraints for VLSI and multiple-chip technology respectively. For instance, Dally [IEEE Trans. Comput. 39(6) (1990) 775], Abraham [Issues in the architecture of direct interconnection networks schemes for multiprocessors, Ph.D. thesis, University of Illinois at Urbana-Champaign, 1992], and Agrawal [IEEE Trans. Parallel Distributed Syst. 2(4) (1991) 398] have shown that low-dimensional k-ary n-cubes (known as tori) outperform their high-dimensional counterparts (known as hypercubes) under the constant wiring density constraint. However, Abraham and Agrawal have arrived at an opposite conclusion when they considered the constant pin-out constraint. Most of these analyses have assumed deterministic routing, where a message always uses the same network path between a given pair of nodes. More recent multicomputers have incorporated adaptive routing to improve performance. This paper re-examines the relative performance merits of the torus and hypercube in the context of adaptive routing. Our analysis reveals that the torus manages to exploit its wider channels under light traffic. As traffic increases, however, the hypercube can provide better performance than the torus. Our conclusion under the constant wiring density constraint is different from that of the works mentioned above because adaptive routing enables the hypercube to exploit its richer connectivity to reduce message blocking.

Original languageEnglish
Pages (from-to)563-574
Number of pages12
JournalJournal of Systems Architecture
Volume50
Issue number9
DOIs
Publication statusPublished - Sep 2004

Fingerprint

Electric wiring
Bandwidth

Keywords

  • Channel bandwidth constraints
  • Interconnection networks
  • Multicomputers
  • Performance comparison

ASJC Scopus subject areas

  • Hardware and Architecture
  • Software

Cite this

On the performance of multicomputer interconnection networks. / Ould-Khaoua, Mohamed; Loucif, Samia; Rabhi, Fethi A.

In: Journal of Systems Architecture, Vol. 50, No. 9, 09.2004, p. 563-574.

Research output: Contribution to journalArticle

Ould-Khaoua, Mohamed ; Loucif, Samia ; Rabhi, Fethi A. / On the performance of multicomputer interconnection networks. In: Journal of Systems Architecture. 2004 ; Vol. 50, No. 9. pp. 563-574.
@article{7b4f4bb5b046470f9692f129ed53096b,
title = "On the performance of multicomputer interconnection networks",
abstract = "Several researchers have analysed the performance of k-ary n-cubes taking into account channel bandwidth constraints imposed by implementation technology, namely the constant wiring density and pin-out constraints for VLSI and multiple-chip technology respectively. For instance, Dally [IEEE Trans. Comput. 39(6) (1990) 775], Abraham [Issues in the architecture of direct interconnection networks schemes for multiprocessors, Ph.D. thesis, University of Illinois at Urbana-Champaign, 1992], and Agrawal [IEEE Trans. Parallel Distributed Syst. 2(4) (1991) 398] have shown that low-dimensional k-ary n-cubes (known as tori) outperform their high-dimensional counterparts (known as hypercubes) under the constant wiring density constraint. However, Abraham and Agrawal have arrived at an opposite conclusion when they considered the constant pin-out constraint. Most of these analyses have assumed deterministic routing, where a message always uses the same network path between a given pair of nodes. More recent multicomputers have incorporated adaptive routing to improve performance. This paper re-examines the relative performance merits of the torus and hypercube in the context of adaptive routing. Our analysis reveals that the torus manages to exploit its wider channels under light traffic. As traffic increases, however, the hypercube can provide better performance than the torus. Our conclusion under the constant wiring density constraint is different from that of the works mentioned above because adaptive routing enables the hypercube to exploit its richer connectivity to reduce message blocking.",
keywords = "Channel bandwidth constraints, Interconnection networks, Multicomputers, Performance comparison",
author = "Mohamed Ould-Khaoua and Samia Loucif and Rabhi, {Fethi A.}",
year = "2004",
month = "9",
doi = "10.1016/j.sysarc.2003.09.008",
language = "English",
volume = "50",
pages = "563--574",
journal = "Journal of Systems Architecture",
issn = "1383-7621",
publisher = "Elsevier",
number = "9",

}

TY - JOUR

T1 - On the performance of multicomputer interconnection networks

AU - Ould-Khaoua, Mohamed

AU - Loucif, Samia

AU - Rabhi, Fethi A.

PY - 2004/9

Y1 - 2004/9

N2 - Several researchers have analysed the performance of k-ary n-cubes taking into account channel bandwidth constraints imposed by implementation technology, namely the constant wiring density and pin-out constraints for VLSI and multiple-chip technology respectively. For instance, Dally [IEEE Trans. Comput. 39(6) (1990) 775], Abraham [Issues in the architecture of direct interconnection networks schemes for multiprocessors, Ph.D. thesis, University of Illinois at Urbana-Champaign, 1992], and Agrawal [IEEE Trans. Parallel Distributed Syst. 2(4) (1991) 398] have shown that low-dimensional k-ary n-cubes (known as tori) outperform their high-dimensional counterparts (known as hypercubes) under the constant wiring density constraint. However, Abraham and Agrawal have arrived at an opposite conclusion when they considered the constant pin-out constraint. Most of these analyses have assumed deterministic routing, where a message always uses the same network path between a given pair of nodes. More recent multicomputers have incorporated adaptive routing to improve performance. This paper re-examines the relative performance merits of the torus and hypercube in the context of adaptive routing. Our analysis reveals that the torus manages to exploit its wider channels under light traffic. As traffic increases, however, the hypercube can provide better performance than the torus. Our conclusion under the constant wiring density constraint is different from that of the works mentioned above because adaptive routing enables the hypercube to exploit its richer connectivity to reduce message blocking.

AB - Several researchers have analysed the performance of k-ary n-cubes taking into account channel bandwidth constraints imposed by implementation technology, namely the constant wiring density and pin-out constraints for VLSI and multiple-chip technology respectively. For instance, Dally [IEEE Trans. Comput. 39(6) (1990) 775], Abraham [Issues in the architecture of direct interconnection networks schemes for multiprocessors, Ph.D. thesis, University of Illinois at Urbana-Champaign, 1992], and Agrawal [IEEE Trans. Parallel Distributed Syst. 2(4) (1991) 398] have shown that low-dimensional k-ary n-cubes (known as tori) outperform their high-dimensional counterparts (known as hypercubes) under the constant wiring density constraint. However, Abraham and Agrawal have arrived at an opposite conclusion when they considered the constant pin-out constraint. Most of these analyses have assumed deterministic routing, where a message always uses the same network path between a given pair of nodes. More recent multicomputers have incorporated adaptive routing to improve performance. This paper re-examines the relative performance merits of the torus and hypercube in the context of adaptive routing. Our analysis reveals that the torus manages to exploit its wider channels under light traffic. As traffic increases, however, the hypercube can provide better performance than the torus. Our conclusion under the constant wiring density constraint is different from that of the works mentioned above because adaptive routing enables the hypercube to exploit its richer connectivity to reduce message blocking.

KW - Channel bandwidth constraints

KW - Interconnection networks

KW - Multicomputers

KW - Performance comparison

UR - http://www.scopus.com/inward/record.url?scp=3242717613&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=3242717613&partnerID=8YFLogxK

U2 - 10.1016/j.sysarc.2003.09.008

DO - 10.1016/j.sysarc.2003.09.008

M3 - Article

VL - 50

SP - 563

EP - 574

JO - Journal of Systems Architecture

JF - Journal of Systems Architecture

SN - 1383-7621

IS - 9

ER -