Loop scheduling for heterogeneity

Michal Cierniak, Wei Li, Mohammed Javeed Zaki

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

22 Citations (Scopus)

Abstract

In this paper, we study the problem of scheduling parallel loops at compile-time for a heterogeneous network of machines. We consider heterogeneity in three aspects of parallel programming: program, processor and network. A heterogeneous program has parallel loops with different amount of work in each iteration; heterogeneous processors have different speeds; and a heterogeneous network has different cost of communication between processors. We propose a simple yet comprehensive model for use in compiling for a network of processors, and develop compiler algorithms for generating optimal and sub-optimal schedules of loops for load balancing, communication optimizations and network contention. Experiments show that a significant improvement of performance is achieved using our techniques.

Original languageEnglish
Title of host publicationIEEE International Symposium on High Performance Distributed Computing, Proceedings
PublisherIEEE
Pages78-85
Number of pages8
Publication statusPublished - 1995
Externally publishedYes
EventProceedings of the 4th IEEE International Symposium on High Performance Distributed Computing - Washington, DC, USA
Duration: 2 Aug 19954 Aug 1995

Other

OtherProceedings of the 4th IEEE International Symposium on High Performance Distributed Computing
CityWashington, DC, USA
Period2/8/954/8/95

Fingerprint

Heterogeneous networks
Scheduling
Parallel programming
Communication
Resource allocation
Program processors
Costs
Experiments

ASJC Scopus subject areas

  • Hardware and Architecture

Cite this

Cierniak, M., Li, W., & Zaki, M. J. (1995). Loop scheduling for heterogeneity. In IEEE International Symposium on High Performance Distributed Computing, Proceedings (pp. 78-85). IEEE.

Loop scheduling for heterogeneity. / Cierniak, Michal; Li, Wei; Zaki, Mohammed Javeed.

IEEE International Symposium on High Performance Distributed Computing, Proceedings. IEEE, 1995. p. 78-85.

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

Cierniak, M, Li, W & Zaki, MJ 1995, Loop scheduling for heterogeneity. in IEEE International Symposium on High Performance Distributed Computing, Proceedings. IEEE, pp. 78-85, Proceedings of the 4th IEEE International Symposium on High Performance Distributed Computing, Washington, DC, USA, 2/8/95.
Cierniak M, Li W, Zaki MJ. Loop scheduling for heterogeneity. In IEEE International Symposium on High Performance Distributed Computing, Proceedings. IEEE. 1995. p. 78-85
Cierniak, Michal ; Li, Wei ; Zaki, Mohammed Javeed. / Loop scheduling for heterogeneity. IEEE International Symposium on High Performance Distributed Computing, Proceedings. IEEE, 1995. pp. 78-85
@inproceedings{e6220111d729471887aaf5b2442acd03,
title = "Loop scheduling for heterogeneity",
abstract = "In this paper, we study the problem of scheduling parallel loops at compile-time for a heterogeneous network of machines. We consider heterogeneity in three aspects of parallel programming: program, processor and network. A heterogeneous program has parallel loops with different amount of work in each iteration; heterogeneous processors have different speeds; and a heterogeneous network has different cost of communication between processors. We propose a simple yet comprehensive model for use in compiling for a network of processors, and develop compiler algorithms for generating optimal and sub-optimal schedules of loops for load balancing, communication optimizations and network contention. Experiments show that a significant improvement of performance is achieved using our techniques.",
author = "Michal Cierniak and Wei Li and Zaki, {Mohammed Javeed}",
year = "1995",
language = "English",
pages = "78--85",
booktitle = "IEEE International Symposium on High Performance Distributed Computing, Proceedings",
publisher = "IEEE",

}

TY - GEN

T1 - Loop scheduling for heterogeneity

AU - Cierniak, Michal

AU - Li, Wei

AU - Zaki, Mohammed Javeed

PY - 1995

Y1 - 1995

N2 - In this paper, we study the problem of scheduling parallel loops at compile-time for a heterogeneous network of machines. We consider heterogeneity in three aspects of parallel programming: program, processor and network. A heterogeneous program has parallel loops with different amount of work in each iteration; heterogeneous processors have different speeds; and a heterogeneous network has different cost of communication between processors. We propose a simple yet comprehensive model for use in compiling for a network of processors, and develop compiler algorithms for generating optimal and sub-optimal schedules of loops for load balancing, communication optimizations and network contention. Experiments show that a significant improvement of performance is achieved using our techniques.

AB - In this paper, we study the problem of scheduling parallel loops at compile-time for a heterogeneous network of machines. We consider heterogeneity in three aspects of parallel programming: program, processor and network. A heterogeneous program has parallel loops with different amount of work in each iteration; heterogeneous processors have different speeds; and a heterogeneous network has different cost of communication between processors. We propose a simple yet comprehensive model for use in compiling for a network of processors, and develop compiler algorithms for generating optimal and sub-optimal schedules of loops for load balancing, communication optimizations and network contention. Experiments show that a significant improvement of performance is achieved using our techniques.

UR - http://www.scopus.com/inward/record.url?scp=0029546721&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0029546721&partnerID=8YFLogxK

M3 - Conference contribution

SP - 78

EP - 85

BT - IEEE International Symposium on High Performance Distributed Computing, Proceedings

PB - IEEE

ER -