Estimating parallel execution time of loops with loop-carried dependences

T. Nakanishi, K. Joe, C. D. Polychronopoulos, K. Araki, A. Fukuda

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

3 Citations (Scopus)

Abstract

We propose a scheme to estimate exact minimum parallel execution time of the single loop with loop-carried dependences in medium and fine grain parallel execution. The minimum parallel execution time of a loop is given by the critical path length of the dependence graph which represents the code obtained from the fully unrolled loop. However, unrolling loops with a large number of iterations requires too much computation time and large storage space to be practical. The scheme proposed provides the minimum parallel execution time without unrolling the loop at all by reducing the problem into an integer linear programming problem and employing the simplex method and a branch-and-bound algorithm to solve it. We also show an experimental implementation of the proposed scheme with Livermore Benchmark Kernels to demonstrate that the computational complexity of our scheme is independent of the number of iterations of the given loop.

Original languageEnglish
Title of host publicationSoftware
EditorsK. Pingali
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages61-69
Number of pages9
ISBN (Electronic)081867623X
DOIs
Publication statusPublished - Jan 1 1996
Event25th International Conference on Parallel Processing, ICPP 1996 - Ithaca, United States
Duration: Aug 12 1996Aug 16 1996

Publication series

NameProceedings of the International Conference on Parallel Processing
Volume3
ISSN (Print)0190-3918

Other

Other25th International Conference on Parallel Processing, ICPP 1996
CountryUnited States
CityIthaca
Period8/12/968/16/96

Fingerprint

Linear programming
Execution Time
Computational complexity
Iteration
Critical Path
Simplex Method
Integer Linear Programming
Branch and Bound Algorithm
Path Length
Computational Complexity
Benchmark
kernel
Graph in graph theory
Estimate
Demonstrate

All Science Journal Classification (ASJC) codes

  • Software
  • Mathematics(all)
  • Hardware and Architecture

Cite this

Nakanishi, T., Joe, K., Polychronopoulos, C. D., Araki, K., & Fukuda, A. (1996). Estimating parallel execution time of loops with loop-carried dependences. In K. Pingali (Ed.), Software (pp. 61-69). [538560] (Proceedings of the International Conference on Parallel Processing; Vol. 3). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICPP.1996.538560

Estimating parallel execution time of loops with loop-carried dependences. / Nakanishi, T.; Joe, K.; Polychronopoulos, C. D.; Araki, K.; Fukuda, A.

Software. ed. / K. Pingali. Institute of Electrical and Electronics Engineers Inc., 1996. p. 61-69 538560 (Proceedings of the International Conference on Parallel Processing; Vol. 3).

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

Nakanishi, T, Joe, K, Polychronopoulos, CD, Araki, K & Fukuda, A 1996, Estimating parallel execution time of loops with loop-carried dependences. in K Pingali (ed.), Software., 538560, Proceedings of the International Conference on Parallel Processing, vol. 3, Institute of Electrical and Electronics Engineers Inc., pp. 61-69, 25th International Conference on Parallel Processing, ICPP 1996, Ithaca, United States, 8/12/96. https://doi.org/10.1109/ICPP.1996.538560
Nakanishi T, Joe K, Polychronopoulos CD, Araki K, Fukuda A. Estimating parallel execution time of loops with loop-carried dependences. In Pingali K, editor, Software. Institute of Electrical and Electronics Engineers Inc. 1996. p. 61-69. 538560. (Proceedings of the International Conference on Parallel Processing). https://doi.org/10.1109/ICPP.1996.538560
Nakanishi, T. ; Joe, K. ; Polychronopoulos, C. D. ; Araki, K. ; Fukuda, A. / Estimating parallel execution time of loops with loop-carried dependences. Software. editor / K. Pingali. Institute of Electrical and Electronics Engineers Inc., 1996. pp. 61-69 (Proceedings of the International Conference on Parallel Processing).
@inproceedings{1696cebd706940ac963d520baa19fb6a,
title = "Estimating parallel execution time of loops with loop-carried dependences",
abstract = "We propose a scheme to estimate exact minimum parallel execution time of the single loop with loop-carried dependences in medium and fine grain parallel execution. The minimum parallel execution time of a loop is given by the critical path length of the dependence graph which represents the code obtained from the fully unrolled loop. However, unrolling loops with a large number of iterations requires too much computation time and large storage space to be practical. The scheme proposed provides the minimum parallel execution time without unrolling the loop at all by reducing the problem into an integer linear programming problem and employing the simplex method and a branch-and-bound algorithm to solve it. We also show an experimental implementation of the proposed scheme with Livermore Benchmark Kernels to demonstrate that the computational complexity of our scheme is independent of the number of iterations of the given loop.",
author = "T. Nakanishi and K. Joe and Polychronopoulos, {C. D.} and K. Araki and A. Fukuda",
year = "1996",
month = "1",
day = "1",
doi = "10.1109/ICPP.1996.538560",
language = "English",
series = "Proceedings of the International Conference on Parallel Processing",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "61--69",
editor = "K. Pingali",
booktitle = "Software",
address = "United States",

}

TY - GEN

T1 - Estimating parallel execution time of loops with loop-carried dependences

AU - Nakanishi, T.

AU - Joe, K.

AU - Polychronopoulos, C. D.

AU - Araki, K.

AU - Fukuda, A.

PY - 1996/1/1

Y1 - 1996/1/1

N2 - We propose a scheme to estimate exact minimum parallel execution time of the single loop with loop-carried dependences in medium and fine grain parallel execution. The minimum parallel execution time of a loop is given by the critical path length of the dependence graph which represents the code obtained from the fully unrolled loop. However, unrolling loops with a large number of iterations requires too much computation time and large storage space to be practical. The scheme proposed provides the minimum parallel execution time without unrolling the loop at all by reducing the problem into an integer linear programming problem and employing the simplex method and a branch-and-bound algorithm to solve it. We also show an experimental implementation of the proposed scheme with Livermore Benchmark Kernels to demonstrate that the computational complexity of our scheme is independent of the number of iterations of the given loop.

AB - We propose a scheme to estimate exact minimum parallel execution time of the single loop with loop-carried dependences in medium and fine grain parallel execution. The minimum parallel execution time of a loop is given by the critical path length of the dependence graph which represents the code obtained from the fully unrolled loop. However, unrolling loops with a large number of iterations requires too much computation time and large storage space to be practical. The scheme proposed provides the minimum parallel execution time without unrolling the loop at all by reducing the problem into an integer linear programming problem and employing the simplex method and a branch-and-bound algorithm to solve it. We also show an experimental implementation of the proposed scheme with Livermore Benchmark Kernels to demonstrate that the computational complexity of our scheme is independent of the number of iterations of the given loop.

UR - http://www.scopus.com/inward/record.url?scp=33646740224&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33646740224&partnerID=8YFLogxK

U2 - 10.1109/ICPP.1996.538560

DO - 10.1109/ICPP.1996.538560

M3 - Conference contribution

AN - SCOPUS:33646740224

T3 - Proceedings of the International Conference on Parallel Processing

SP - 61

EP - 69

BT - Software

A2 - Pingali, K.

PB - Institute of Electrical and Electronics Engineers Inc.

ER -