Automatic performance analysis of OpenMP codes on a scalable shared memory system using periscope

Shajulin Benedict; Michael Gerndt

doi:10.1007/978-3-642-28145-7_44

Automatic performance analysis of OpenMP codes on a scalable shared memory system using periscope

Shajulin Benedict, Michael Gerndt

Informatics 10 - Chair of Computer Architecture and Parallel Systems
Informatics 10 - Associate Professorship of Architecture of Parallel and Distributed Systems

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

8 Scopus citations

Abstract

OpenMP is a successful interface for programming parallel applications on shared memory systems. It is widely applied on small scale shared memory systems such as multicore processors, but also in hybrid programming on large supercomputers. This paper presents performance properties for OpenMP and their automatic detection by Periscope. We evaluate Periscope's OpenMP analysis strategy in the context of the Altix 4700 supercomputer at Leibniz Computing Center (LRZ) in Garching. On this unique machine OpenMP scales up to 500 cores, one partition of in total 19 partitions. We present results for the NAS parallel benchmarks and for a large hybrid scientific application.

Original language	English
Title of host publication	Applied Parallel and Scientific Computing - 10th International Conference, PARA 2010, Revised Selected Papers
Pages	452-462
Number of pages	11
Edition	PART 2
DOIs	https://doi.org/10.1007/978-3-642-28145-7_44
State	Published - 2012
Event	10th International Conference on Applied Parallel and Scientific Computing, PARA 2010 - Reykjavik, Iceland Duration: 6 Jun 2010 → 9 Jun 2010

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 2
Volume	7134 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	10th International Conference on Applied Parallel and Scientific Computing, PARA 2010
Country/Territory	Iceland
City	Reykjavik
Period	6/06/10 → 9/06/10

Keywords

Memory accesses analysis
OpenMP
Performance analysis
Supercomputers

Access to Document

10.1007/978-3-642-28145-7_44

Cite this

Benedict, S., & Gerndt, M. (2012). Automatic performance analysis of OpenMP codes on a scalable shared memory system using periscope. In Applied Parallel and Scientific Computing - 10th International Conference, PARA 2010, Revised Selected Papers (PART 2 ed., pp. 452-462). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7134 LNCS, No. PART 2). https://doi.org/10.1007/978-3-642-28145-7_44

Benedict, Shajulin ; Gerndt, Michael. / Automatic performance analysis of OpenMP codes on a scalable shared memory system using periscope. Applied Parallel and Scientific Computing - 10th International Conference, PARA 2010, Revised Selected Papers. PART 2. ed. 2012. pp. 452-462 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).

@inproceedings{f25d21eaeaab47d789d766492d9abf9b,

title = "Automatic performance analysis of OpenMP codes on a scalable shared memory system using periscope",

abstract = "OpenMP is a successful interface for programming parallel applications on shared memory systems. It is widely applied on small scale shared memory systems such as multicore processors, but also in hybrid programming on large supercomputers. This paper presents performance properties for OpenMP and their automatic detection by Periscope. We evaluate Periscope's OpenMP analysis strategy in the context of the Altix 4700 supercomputer at Leibniz Computing Center (LRZ) in Garching. On this unique machine OpenMP scales up to 500 cores, one partition of in total 19 partitions. We present results for the NAS parallel benchmarks and for a large hybrid scientific application.",

keywords = "Memory accesses analysis, OpenMP, Performance analysis, Supercomputers",

author = "Shajulin Benedict and Michael Gerndt",

note = "Funding Information: This work is partially funded by BMBF under the ISAR project, grant 01IH08005A and the SILC project, grant 01IH08006E.; 10th International Conference on Applied Parallel and Scientific Computing, PARA 2010 ; Conference date: 06-06-2010 Through 09-06-2010",

year = "2012",

doi = "10.1007/978-3-642-28145-7_44",

language = "English",

isbn = "9783642281440",

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

number = "PART 2",

pages = "452--462",

booktitle = "Applied Parallel and Scientific Computing - 10th International Conference, PARA 2010, Revised Selected Papers",

edition = "PART 2",

}

Benedict, S & Gerndt, M 2012, Automatic performance analysis of OpenMP codes on a scalable shared memory system using periscope. in Applied Parallel and Scientific Computing - 10th International Conference, PARA 2010, Revised Selected Papers. PART 2 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 7134 LNCS, pp. 452-462, 10th International Conference on Applied Parallel and Scientific Computing, PARA 2010, Reykjavik, Iceland, 6/06/10. https://doi.org/10.1007/978-3-642-28145-7_44

Automatic performance analysis of OpenMP codes on a scalable shared memory system using periscope. / Benedict, Shajulin ; Gerndt, Michael.
Applied Parallel and Scientific Computing - 10th International Conference, PARA 2010, Revised Selected Papers. PART 2. ed. 2012. p. 452-462 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7134 LNCS, No. PART 2).

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

TY - GEN

T1 - Automatic performance analysis of OpenMP codes on a scalable shared memory system using periscope

AU - Benedict, Shajulin

AU - Gerndt, Michael

N1 - Funding Information: This work is partially funded by BMBF under the ISAR project, grant 01IH08005A and the SILC project, grant 01IH08006E.

PY - 2012

Y1 - 2012

N2 - OpenMP is a successful interface for programming parallel applications on shared memory systems. It is widely applied on small scale shared memory systems such as multicore processors, but also in hybrid programming on large supercomputers. This paper presents performance properties for OpenMP and their automatic detection by Periscope. We evaluate Periscope's OpenMP analysis strategy in the context of the Altix 4700 supercomputer at Leibniz Computing Center (LRZ) in Garching. On this unique machine OpenMP scales up to 500 cores, one partition of in total 19 partitions. We present results for the NAS parallel benchmarks and for a large hybrid scientific application.

AB - OpenMP is a successful interface for programming parallel applications on shared memory systems. It is widely applied on small scale shared memory systems such as multicore processors, but also in hybrid programming on large supercomputers. This paper presents performance properties for OpenMP and their automatic detection by Periscope. We evaluate Periscope's OpenMP analysis strategy in the context of the Altix 4700 supercomputer at Leibniz Computing Center (LRZ) in Garching. On this unique machine OpenMP scales up to 500 cores, one partition of in total 19 partitions. We present results for the NAS parallel benchmarks and for a large hybrid scientific application.

KW - Memory accesses analysis

KW - OpenMP

KW - Performance analysis

KW - Supercomputers

UR - http://www.scopus.com/inward/record.url?scp=84857469209&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-28145-7_44

DO - 10.1007/978-3-642-28145-7_44

M3 - Conference contribution

AN - SCOPUS:84857469209

SN - 9783642281440

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

SP - 452

EP - 462

BT - Applied Parallel and Scientific Computing - 10th International Conference, PARA 2010, Revised Selected Papers

T2 - 10th International Conference on Applied Parallel and Scientific Computing, PARA 2010

Y2 - 6 June 2010 through 9 June 2010

ER -

Benedict S , Gerndt M. Automatic performance analysis of OpenMP codes on a scalable shared memory system using periscope. In Applied Parallel and Scientific Computing - 10th International Conference, PARA 2010, Revised Selected Papers. PART 2 ed. 2012. p. 452-462. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). doi: 10.1007/978-3-642-28145-7_44

Automatic performance analysis of OpenMP codes on a scalable shared memory system using periscope

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this