Analyses for the translation of OpenMP codes into SPMD style with array privatization

Zhenying Liu; Barbara Chapman; Yi Wen; Lei Huang; Tien Hsiung Weng; Oscar Hernandez

doi:10.1007/3-540-45009-2_3

Analyses for the translation of OpenMP codes into SPMD style with array privatization

Zhenying Liu, Barbara Chapman, Yi Wen, Lei Huang, Tien Hsiung Weng, Oscar Hernandez

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

15 Scopus citations

Abstract

A so-called SPMD style OpenMP program can achieve scalability on ccNUMA systems by means of array privatization, and earlier research has shown good performance under this approach. Since it is hard to write SPMD OpenMP code, we showed a strategy for the automatic translation of many OpenMP constructs into SPMD style in our previous work. In this paper, we first explain how to interprocedurally detect whether the OpenMP program consistently schedules the parallel loops. If the parallel loops are consistently scheduled, we may carry out array privatization according to OpenMP semantics. We give two examples of code patterns that can be handled despite the fact that they are not consistent, and where the strategy used to translate them differs from the straightforward approach that can otherwise be applied.

Original language	English
Title of host publication	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Editors	Michael J. Voss
Publisher	Springer Verlag
Pages	26-41
Number of pages	16
ISBN (Print)	9783540404354
DOIs	https://doi.org/10.1007/3-540-45009-2_3
State	Published - 2003
Externally published	Yes

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	2716
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Access to Document

10.1007/3-540-45009-2_3

Cite this

Liu, Z., Chapman, B., Wen, Y., Huang, L., Weng, T. H., & Hernandez, O. (2003). Analyses for the translation of OpenMP codes into SPMD style with array privatization. In M. J. Voss (Ed.), Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (pp. 26-41). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2716). Springer Verlag. https://doi.org/10.1007/3-540-45009-2_3

Liu, Zhenying ; Chapman, Barbara ; Wen, Yi et al. / Analyses for the translation of OpenMP codes into SPMD style with array privatization. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). editor / Michael J. Voss. Springer Verlag, 2003. pp. 26-41 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inbook{3e07fc7ad7d24fcd82c9be10d5d67d89,

title = "Analyses for the translation of OpenMP codes into SPMD style with array privatization",

abstract = "A so-called SPMD style OpenMP program can achieve scalability on ccNUMA systems by means of array privatization, and earlier research has shown good performance under this approach. Since it is hard to write SPMD OpenMP code, we showed a strategy for the automatic translation of many OpenMP constructs into SPMD style in our previous work. In this paper, we first explain how to interprocedurally detect whether the OpenMP program consistently schedules the parallel loops. If the parallel loops are consistently scheduled, we may carry out array privatization according to OpenMP semantics. We give two examples of code patterns that can be handled despite the fact that they are not consistent, and where the strategy used to translate them differs from the straightforward approach that can otherwise be applied.",

author = "Zhenying Liu and Barbara Chapman and Yi Wen and Lei Huang and Weng, {Tien Hsiung} and Oscar Hernandez",

year = "2003",

doi = "10.1007/3-540-45009-2_3",

language = "English",

isbn = "9783540404354",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "26--41",

editor = "Voss, {Michael J.}",

booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

Liu, Z, Chapman, B, Wen, Y, Huang, L, Weng, TH & Hernandez, O 2003, Analyses for the translation of OpenMP codes into SPMD style with array privatization. in MJ Voss (ed.), Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 2716, Springer Verlag, pp. 26-41. https://doi.org/10.1007/3-540-45009-2_3

Analyses for the translation of OpenMP codes into SPMD style with array privatization. / Liu, Zhenying; Chapman, Barbara; Wen, Yi et al.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). ed. / Michael J. Voss. Springer Verlag, 2003. p. 26-41 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2716).

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

TY - CHAP

T1 - Analyses for the translation of OpenMP codes into SPMD style with array privatization

AU - Liu, Zhenying

AU - Chapman, Barbara

AU - Wen, Yi

AU - Huang, Lei

AU - Weng, Tien Hsiung

AU - Hernandez, Oscar

PY - 2003

Y1 - 2003

N2 - A so-called SPMD style OpenMP program can achieve scalability on ccNUMA systems by means of array privatization, and earlier research has shown good performance under this approach. Since it is hard to write SPMD OpenMP code, we showed a strategy for the automatic translation of many OpenMP constructs into SPMD style in our previous work. In this paper, we first explain how to interprocedurally detect whether the OpenMP program consistently schedules the parallel loops. If the parallel loops are consistently scheduled, we may carry out array privatization according to OpenMP semantics. We give two examples of code patterns that can be handled despite the fact that they are not consistent, and where the strategy used to translate them differs from the straightforward approach that can otherwise be applied.

AB - A so-called SPMD style OpenMP program can achieve scalability on ccNUMA systems by means of array privatization, and earlier research has shown good performance under this approach. Since it is hard to write SPMD OpenMP code, we showed a strategy for the automatic translation of many OpenMP constructs into SPMD style in our previous work. In this paper, we first explain how to interprocedurally detect whether the OpenMP program consistently schedules the parallel loops. If the parallel loops are consistently scheduled, we may carry out array privatization according to OpenMP semantics. We give two examples of code patterns that can be handled despite the fact that they are not consistent, and where the strategy used to translate them differs from the straightforward approach that can otherwise be applied.

UR - http://www.scopus.com/inward/record.url?scp=35248874348&partnerID=8YFLogxK

U2 - 10.1007/3-540-45009-2_3

DO - 10.1007/3-540-45009-2_3

M3 - Chapter

AN - SCOPUS:35248874348

SN - 9783540404354

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 26

EP - 41

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

A2 - Voss, Michael J.

PB - Springer Verlag

ER -

Liu Z, Chapman B, Wen Y, Huang L, Weng TH, Hernandez O. Analyses for the translation of OpenMP codes into SPMD style with array privatization. In Voss MJ, editor, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Springer Verlag. 2003. p. 26-41. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/3-540-45009-2_3

Analyses for the translation of OpenMP codes into SPMD style with array privatization

Abstract

Publication series

Access to Document

Other files and links

Fingerprint

Cite this