Developing Scalable Applications with Vampir, VampirServer and VampirTrace

Matthias S. Mueller; Andreas Knuepfer; Matthias Jurenz; Matthias Lieber; Holger Brunst; Hartmut Mix; Wolfgang E. Nagel

Developing Scalable Applications with Vampir, VampirServer and VampirTrace

Research output: Contribution to book/Conference proceedings/Anthology/Report › Conference contribution › Contributed › peer-review

Contributors

Matthias S. Mueller - , Center for Information Services and High Performance Computing (ZIH) (Author)
Andreas Knuepfer - , Center for Information Services and High Performance Computing (ZIH) (Author)
Matthias Jurenz - , Center of Information Services and High Performance Computing (ZIH) (Author)
Matthias Lieber - , Center for Information Services and High Performance Computing (ZIH) (Author)
Holger Brunst - , Center for Information Services and High Performance Computing (ZIH) (Author)
Hartmut Mix - , Center for Information Services and High Performance Computing (ZIH) (Author)
Wolfgang E. Nagel - , Center for Information Services and High Performance Computing (ZIH), Chair of Computer Architecture (Author)

Abstract

This paper presents some scalability studies of the performance analysis tools Vampir and VampirTrace. The usability is analyzed with data collected from real applications, i.e. the thirteen applications contained in the SPEC MPI 1.0 benchmark suite. The analysis covers all phases of performance analysis: instrumenting the application, collecting the performance data, and finally viewing and analyzing the data. The aspects examined include instrumenting effort, monitoring overhead, trace file sizes, load time and response time during analysis.

Details

Original language	English
Title of host publication	Parallel Computing: Architectures, Algorithms and Applications
Editors	C Bischof, C Bucker, P Gibbon, G Joubert, T Lippert, B Mohr, F Peters
Publisher	IOS Press, Amsterdam [u. a.]
Pages	637-644
Number of pages	8
ISBN (print)	978-1-58603-796-3
Publication status	Published - 2008
Peer-reviewed	Yes

Publication series

Series	Advances in Parallel Computing
Volume	15
ISSN	0927-5452

Conference

Title	International Parallel Computing Conference 2007
Abbreviated title	ParCo 2007
Duration	4 - 7 September 2007
Website	https://www.parco.org/ParCo2007/parco2007.html
Degree of recognition	International event
City	Juelich
Country	Germany

External IDs

WOS	000273549500076
Scopus	70349863847
ORCID	/0000-0001-6520-4563/work/142236625
ORCID	/0000-0003-3137-0648/work/142238842