A parallel algorithm for mapping a special class of task graphs onto linear array multiprocessors

Sibabrata Ray, Hong Jiang, Jitender S. Deogun

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

Abstract

In parallel and distributed computing, mapping a task graph to a specific multiprocessor topology is an important problem. In this paper we investigate a special case of this problem for linear task graph and linear array multiprocessors. A polynomial time sequential algorithm and a parallel algorithm are developed and analyzed. It has been shown that the sequential algorithm developed has a time complexity at least as aood as the existing n log rz algorithm while the parallel algorithm presents the very first attempt at parallelizing the sequential solution. Further, we have proved that the partitioning problem is NP-complete for tree task graphs. Applications of the algorithm are briefly discussed.

Original languageEnglish (US)
Title of host publication1994 ACM Symposium on Applied Computing, SAC 1994
PublisherAssociation for Computing Machinery
Pages473-477
Number of pages5
ISBN (Electronic)0897916476
DOIs
StatePublished - Apr 6 1994
Event1994 ACM Symposium on Applied Computing, SAC 1994 - Phoenix, United States
Duration: Mar 6 1994Mar 8 1994

Publication series

NameProceedings of the ACM Symposium on Applied Computing
VolumePart F129433

Other

Other1994 ACM Symposium on Applied Computing, SAC 1994
CountryUnited States
CityPhoenix
Period3/6/943/8/94

Fingerprint

Parallel algorithms
Distributed computer systems
Parallel processing systems
Computational complexity
Topology
Polynomials

Keywords

  • Computational complexity
  • Load balancing
  • Multiprocessors
  • Parallel algorithms
  • Task graph partitioning

ASJC Scopus subject areas

  • Software

Cite this

Ray, S., Jiang, H., & Deogun, J. S. (1994). A parallel algorithm for mapping a special class of task graphs onto linear array multiprocessors. In 1994 ACM Symposium on Applied Computing, SAC 1994 (pp. 473-477). (Proceedings of the ACM Symposium on Applied Computing; Vol. Part F129433). Association for Computing Machinery. https://doi.org/10.1145/326619.326815

A parallel algorithm for mapping a special class of task graphs onto linear array multiprocessors. / Ray, Sibabrata; Jiang, Hong; Deogun, Jitender S.

1994 ACM Symposium on Applied Computing, SAC 1994. Association for Computing Machinery, 1994. p. 473-477 (Proceedings of the ACM Symposium on Applied Computing; Vol. Part F129433).

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

Ray, S, Jiang, H & Deogun, JS 1994, A parallel algorithm for mapping a special class of task graphs onto linear array multiprocessors. in 1994 ACM Symposium on Applied Computing, SAC 1994. Proceedings of the ACM Symposium on Applied Computing, vol. Part F129433, Association for Computing Machinery, pp. 473-477, 1994 ACM Symposium on Applied Computing, SAC 1994, Phoenix, United States, 3/6/94. https://doi.org/10.1145/326619.326815
Ray S, Jiang H, Deogun JS. A parallel algorithm for mapping a special class of task graphs onto linear array multiprocessors. In 1994 ACM Symposium on Applied Computing, SAC 1994. Association for Computing Machinery. 1994. p. 473-477. (Proceedings of the ACM Symposium on Applied Computing). https://doi.org/10.1145/326619.326815
Ray, Sibabrata ; Jiang, Hong ; Deogun, Jitender S. / A parallel algorithm for mapping a special class of task graphs onto linear array multiprocessors. 1994 ACM Symposium on Applied Computing, SAC 1994. Association for Computing Machinery, 1994. pp. 473-477 (Proceedings of the ACM Symposium on Applied Computing).
@inproceedings{de51cce1d9b9415f80ad81d531849a5c,
title = "A parallel algorithm for mapping a special class of task graphs onto linear array multiprocessors",
abstract = "In parallel and distributed computing, mapping a task graph to a specific multiprocessor topology is an important problem. In this paper we investigate a special case of this problem for linear task graph and linear array multiprocessors. A polynomial time sequential algorithm and a parallel algorithm are developed and analyzed. It has been shown that the sequential algorithm developed has a time complexity at least as aood as the existing n log rz algorithm while the parallel algorithm presents the very first attempt at parallelizing the sequential solution. Further, we have proved that the partitioning problem is NP-complete for tree task graphs. Applications of the algorithm are briefly discussed.",
keywords = "Computational complexity, Load balancing, Multiprocessors, Parallel algorithms, Task graph partitioning",
author = "Sibabrata Ray and Hong Jiang and Deogun, {Jitender S.}",
year = "1994",
month = "4",
day = "6",
doi = "10.1145/326619.326815",
language = "English (US)",
series = "Proceedings of the ACM Symposium on Applied Computing",
publisher = "Association for Computing Machinery",
pages = "473--477",
booktitle = "1994 ACM Symposium on Applied Computing, SAC 1994",

}

TY - GEN

T1 - A parallel algorithm for mapping a special class of task graphs onto linear array multiprocessors

AU - Ray, Sibabrata

AU - Jiang, Hong

AU - Deogun, Jitender S.

PY - 1994/4/6

Y1 - 1994/4/6

N2 - In parallel and distributed computing, mapping a task graph to a specific multiprocessor topology is an important problem. In this paper we investigate a special case of this problem for linear task graph and linear array multiprocessors. A polynomial time sequential algorithm and a parallel algorithm are developed and analyzed. It has been shown that the sequential algorithm developed has a time complexity at least as aood as the existing n log rz algorithm while the parallel algorithm presents the very first attempt at parallelizing the sequential solution. Further, we have proved that the partitioning problem is NP-complete for tree task graphs. Applications of the algorithm are briefly discussed.

AB - In parallel and distributed computing, mapping a task graph to a specific multiprocessor topology is an important problem. In this paper we investigate a special case of this problem for linear task graph and linear array multiprocessors. A polynomial time sequential algorithm and a parallel algorithm are developed and analyzed. It has been shown that the sequential algorithm developed has a time complexity at least as aood as the existing n log rz algorithm while the parallel algorithm presents the very first attempt at parallelizing the sequential solution. Further, we have proved that the partitioning problem is NP-complete for tree task graphs. Applications of the algorithm are briefly discussed.

KW - Computational complexity

KW - Load balancing

KW - Multiprocessors

KW - Parallel algorithms

KW - Task graph partitioning

UR - http://www.scopus.com/inward/record.url?scp=85039451589&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85039451589&partnerID=8YFLogxK

U2 - 10.1145/326619.326815

DO - 10.1145/326619.326815

M3 - Conference contribution

AN - SCOPUS:85039451589

T3 - Proceedings of the ACM Symposium on Applied Computing

SP - 473

EP - 477

BT - 1994 ACM Symposium on Applied Computing, SAC 1994

PB - Association for Computing Machinery

ER -