One-to-all broadcast algorithm for constant degree 4 Cayley graphs

Mohamed Benmaiza, Abderezak Touzene

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

Broadcasting in interconnection networks is crucial because of its multiple utilization in linear algebra problems, neural networks, optimization problems and other fields alike. Many global communication algorithms have been studied for different topologies of interconnection networks such as hypercubes, meshes, De Bruijn graphs, and Star graphs. In a recent paper, Vadapalli and Srimani proposed a new class of constant degree 4 Cayley graphs along with an optimal routing algorithm. We extended this work by proposing an efficient one-to-all broadcasting algorithm for constant degree 4 Cayley graphs of dimension n presented in this paper. The proposed algorithm is based on the construction of edge-disjoint spanning trees of height 2n and uses pipelining techniques. Communication is assumed to be full-duplex and based on store-and-forward techniques. We also assume a communication model with linear communication times. Based on these assumptions, we show that the performances of the proposed algorithm depend only on the dimension of the graph and are near optimal in terms of propagation delay.

Original languageEnglish
Pages (from-to)249-264
Number of pages16
JournalParallel Computing
Volume25
Issue number3
DOIs
Publication statusPublished - Mar 1999

Fingerprint

Cayley Graph
Broadcast
Communication
Interconnection Networks
Broadcasting
De Bruijn Graph
Star Graph
Pipelining
Network Optimization
Linear algebra
Alike
Routing algorithms
Routing Algorithm
Optimal Algorithm
Hypercube
Spanning tree
Stars
Disjoint
Topology
Mesh

ASJC Scopus subject areas

  • Computer Science Applications
  • Hardware and Architecture
  • Control and Systems Engineering

Cite this

One-to-all broadcast algorithm for constant degree 4 Cayley graphs. / Benmaiza, Mohamed; Touzene, Abderezak.

In: Parallel Computing, Vol. 25, No. 3, 03.1999, p. 249-264.

Research output: Contribution to journalArticle

@article{e1d5daeb4121484a9721a1dc9a61fd74,
title = "One-to-all broadcast algorithm for constant degree 4 Cayley graphs",
abstract = "Broadcasting in interconnection networks is crucial because of its multiple utilization in linear algebra problems, neural networks, optimization problems and other fields alike. Many global communication algorithms have been studied for different topologies of interconnection networks such as hypercubes, meshes, De Bruijn graphs, and Star graphs. In a recent paper, Vadapalli and Srimani proposed a new class of constant degree 4 Cayley graphs along with an optimal routing algorithm. We extended this work by proposing an efficient one-to-all broadcasting algorithm for constant degree 4 Cayley graphs of dimension n presented in this paper. The proposed algorithm is based on the construction of edge-disjoint spanning trees of height 2n and uses pipelining techniques. Communication is assumed to be full-duplex and based on store-and-forward techniques. We also assume a communication model with linear communication times. Based on these assumptions, we show that the performances of the proposed algorithm depend only on the dimension of the graph and are near optimal in terms of propagation delay.",
author = "Mohamed Benmaiza and Abderezak Touzene",
year = "1999",
month = "3",
doi = "10.1016/S0167-8191(98)00109-4",
language = "English",
volume = "25",
pages = "249--264",
journal = "Parallel Computing",
issn = "0167-8191",
publisher = "Elsevier",
number = "3",

}

TY - JOUR

T1 - One-to-all broadcast algorithm for constant degree 4 Cayley graphs

AU - Benmaiza, Mohamed

AU - Touzene, Abderezak

PY - 1999/3

Y1 - 1999/3

N2 - Broadcasting in interconnection networks is crucial because of its multiple utilization in linear algebra problems, neural networks, optimization problems and other fields alike. Many global communication algorithms have been studied for different topologies of interconnection networks such as hypercubes, meshes, De Bruijn graphs, and Star graphs. In a recent paper, Vadapalli and Srimani proposed a new class of constant degree 4 Cayley graphs along with an optimal routing algorithm. We extended this work by proposing an efficient one-to-all broadcasting algorithm for constant degree 4 Cayley graphs of dimension n presented in this paper. The proposed algorithm is based on the construction of edge-disjoint spanning trees of height 2n and uses pipelining techniques. Communication is assumed to be full-duplex and based on store-and-forward techniques. We also assume a communication model with linear communication times. Based on these assumptions, we show that the performances of the proposed algorithm depend only on the dimension of the graph and are near optimal in terms of propagation delay.

AB - Broadcasting in interconnection networks is crucial because of its multiple utilization in linear algebra problems, neural networks, optimization problems and other fields alike. Many global communication algorithms have been studied for different topologies of interconnection networks such as hypercubes, meshes, De Bruijn graphs, and Star graphs. In a recent paper, Vadapalli and Srimani proposed a new class of constant degree 4 Cayley graphs along with an optimal routing algorithm. We extended this work by proposing an efficient one-to-all broadcasting algorithm for constant degree 4 Cayley graphs of dimension n presented in this paper. The proposed algorithm is based on the construction of edge-disjoint spanning trees of height 2n and uses pipelining techniques. Communication is assumed to be full-duplex and based on store-and-forward techniques. We also assume a communication model with linear communication times. Based on these assumptions, we show that the performances of the proposed algorithm depend only on the dimension of the graph and are near optimal in terms of propagation delay.

UR - http://www.scopus.com/inward/record.url?scp=0033099703&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033099703&partnerID=8YFLogxK

U2 - 10.1016/S0167-8191(98)00109-4

DO - 10.1016/S0167-8191(98)00109-4

M3 - Article

VL - 25

SP - 249

EP - 264

JO - Parallel Computing

JF - Parallel Computing

SN - 0167-8191

IS - 3

ER -