Transformation method to reduce loop overhead in HPF compiler

Mi Soon Koo, Sung Soon Park, Hyun Gyoo Yook, Myong Soon Park

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Among data parallel languages, which were suggested to solve programming problems of distributed memory machines, HPF is regarded as a standard data parallel language. And most HPF compilers are source-to-source translators because they can be implemented easily. By the way, they currently generate significant amount of ineffective codes. In particular, FORALL construct is converted into several DO loops, so loop overhead, which exacerbates program performance, of these codes is very increased. Therefore HPF source-to-source compilers require some techniques to produce optimized DO loop from FORALL construct. In this paper, we propose transformation method to reduce loop overhead for FORALL construct. For this, we define and use relation distance vector which can represent both data dependence information and flow information. Then we evaluate and analyze execution time for the codes converted by our method and by PARADIGM method in aspect of array size and dependence distance.

Original languageEnglish
Title of host publicationProceedings of the Conference on High Performance Computing on the Information Superhighway, HPC Asia'97
Editors Anon
Place of PublicationLos Alamitos, CA, United States
PublisherIEEE
Pages306-311
Number of pages6
Publication statusPublished - 1997 Jan 1
EventProceedings of the 1997 2nd High Performance Computing on the Information Superhighway, HPC Asia'97 - Seoul, South Korea
Duration: 1997 Apr 281997 May 2

Other

OtherProceedings of the 1997 2nd High Performance Computing on the Information Superhighway, HPC Asia'97
CitySeoul, South Korea
Period97/4/2897/5/2

Fingerprint

Computer programming
Data storage equipment

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Koo, M. S., Park, S. S., Yook, H. G., & Park, M. S. (1997). Transformation method to reduce loop overhead in HPF compiler. In Anon (Ed.), Proceedings of the Conference on High Performance Computing on the Information Superhighway, HPC Asia'97 (pp. 306-311). Los Alamitos, CA, United States: IEEE.

Transformation method to reduce loop overhead in HPF compiler. / Koo, Mi Soon; Park, Sung Soon; Yook, Hyun Gyoo; Park, Myong Soon.

Proceedings of the Conference on High Performance Computing on the Information Superhighway, HPC Asia'97. ed. / Anon. Los Alamitos, CA, United States : IEEE, 1997. p. 306-311.

Research output: Chapter in Book/Report/Conference proceedingChapter

Koo, MS, Park, SS, Yook, HG & Park, MS 1997, Transformation method to reduce loop overhead in HPF compiler. in Anon (ed.), Proceedings of the Conference on High Performance Computing on the Information Superhighway, HPC Asia'97. IEEE, Los Alamitos, CA, United States, pp. 306-311, Proceedings of the 1997 2nd High Performance Computing on the Information Superhighway, HPC Asia'97, Seoul, South Korea, 97/4/28.
Koo MS, Park SS, Yook HG, Park MS. Transformation method to reduce loop overhead in HPF compiler. In Anon, editor, Proceedings of the Conference on High Performance Computing on the Information Superhighway, HPC Asia'97. Los Alamitos, CA, United States: IEEE. 1997. p. 306-311
Koo, Mi Soon ; Park, Sung Soon ; Yook, Hyun Gyoo ; Park, Myong Soon. / Transformation method to reduce loop overhead in HPF compiler. Proceedings of the Conference on High Performance Computing on the Information Superhighway, HPC Asia'97. editor / Anon. Los Alamitos, CA, United States : IEEE, 1997. pp. 306-311
@inbook{d024badbb42f49a8b08338bef8f170b3,
title = "Transformation method to reduce loop overhead in HPF compiler",
abstract = "Among data parallel languages, which were suggested to solve programming problems of distributed memory machines, HPF is regarded as a standard data parallel language. And most HPF compilers are source-to-source translators because they can be implemented easily. By the way, they currently generate significant amount of ineffective codes. In particular, FORALL construct is converted into several DO loops, so loop overhead, which exacerbates program performance, of these codes is very increased. Therefore HPF source-to-source compilers require some techniques to produce optimized DO loop from FORALL construct. In this paper, we propose transformation method to reduce loop overhead for FORALL construct. For this, we define and use relation distance vector which can represent both data dependence information and flow information. Then we evaluate and analyze execution time for the codes converted by our method and by PARADIGM method in aspect of array size and dependence distance.",
author = "Koo, {Mi Soon} and Park, {Sung Soon} and Yook, {Hyun Gyoo} and Park, {Myong Soon}",
year = "1997",
month = "1",
day = "1",
language = "English",
pages = "306--311",
editor = "Anon",
booktitle = "Proceedings of the Conference on High Performance Computing on the Information Superhighway, HPC Asia'97",
publisher = "IEEE",

}

TY - CHAP

T1 - Transformation method to reduce loop overhead in HPF compiler

AU - Koo, Mi Soon

AU - Park, Sung Soon

AU - Yook, Hyun Gyoo

AU - Park, Myong Soon

PY - 1997/1/1

Y1 - 1997/1/1

N2 - Among data parallel languages, which were suggested to solve programming problems of distributed memory machines, HPF is regarded as a standard data parallel language. And most HPF compilers are source-to-source translators because they can be implemented easily. By the way, they currently generate significant amount of ineffective codes. In particular, FORALL construct is converted into several DO loops, so loop overhead, which exacerbates program performance, of these codes is very increased. Therefore HPF source-to-source compilers require some techniques to produce optimized DO loop from FORALL construct. In this paper, we propose transformation method to reduce loop overhead for FORALL construct. For this, we define and use relation distance vector which can represent both data dependence information and flow information. Then we evaluate and analyze execution time for the codes converted by our method and by PARADIGM method in aspect of array size and dependence distance.

AB - Among data parallel languages, which were suggested to solve programming problems of distributed memory machines, HPF is regarded as a standard data parallel language. And most HPF compilers are source-to-source translators because they can be implemented easily. By the way, they currently generate significant amount of ineffective codes. In particular, FORALL construct is converted into several DO loops, so loop overhead, which exacerbates program performance, of these codes is very increased. Therefore HPF source-to-source compilers require some techniques to produce optimized DO loop from FORALL construct. In this paper, we propose transformation method to reduce loop overhead for FORALL construct. For this, we define and use relation distance vector which can represent both data dependence information and flow information. Then we evaluate and analyze execution time for the codes converted by our method and by PARADIGM method in aspect of array size and dependence distance.

UR - http://www.scopus.com/inward/record.url?scp=0030644545&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0030644545&partnerID=8YFLogxK

M3 - Chapter

SP - 306

EP - 311

BT - Proceedings of the Conference on High Performance Computing on the Information Superhighway, HPC Asia'97

A2 - Anon, null

PB - IEEE

CY - Los Alamitos, CA, United States

ER -