This title appears in the Scientific Report :
2014
Please use the identifier:
http://dx.doi.org/10.3233/978-1-61499-381-0-113 in citations.
Please use the identifier: http://hdl.handle.net/2128/5970 in citations.
Characterizing Performance of Applications on Blue Gene/Q
Characterizing Performance of Applications on Blue Gene/Q
Recently the latest generation of Blue Gene machines became available. In this paper we introduce general metrics to characterize the performance of applications and apply it to a diverse set of applications running on Blue Gene/Q. The applications range from regular, floating-point bound to irregul...
Saved in:
Personal Name(s): | Baumeister, Paul F. (Corresponding Author) |
---|---|
Boettiger, Hans / Hater, Thorsten / Knobloch, Michael / Maurer, Thilo / Nobile, Andrea / Pleiter, Dirk / Vandenbergen, Nicolas | |
Contributing Institute: |
Jülich Supercomputing Center; JSC |
Published in: |
Parallel Computing: Accelerating Computational Science and Engineering (CSE) |
Imprint: |
IOS Press
2014
|
Physical Description: |
113-122 |
ISBN: |
978-1-61499-380-3 |
DOI: |
10.3233/978-1-61499-381-0-113 |
Conference: | International Conference on Parallel Computing, Munich (Germany), 2013-09-10 - 2013-09-13 |
Document Type: |
Contribution to a book Contribution to a conference proceedings |
Research Program: |
Computational Science and Mathematical Methods Supercomputer Facility |
Series Title: |
Advances in Parallel Computing
25 |
Link: |
OpenAccess |
Publikationsportal JuSER |
Please use the identifier: http://hdl.handle.net/2128/5970 in citations.
Recently the latest generation of Blue Gene machines became available. In this paper we introduce general metrics to characterize the performance of applications and apply it to a diverse set of applications running on Blue Gene/Q. The applications range from regular, floating-point bound to irregular event-simulator like types. We argue that the proposed metrics are suitable to characterize the performance for a larger set of computational science applications running on today's massively-parallel systems. They therefore do not only allow to assess usability of the Blue Gene/Q architecture for the considered (types of) applications. They also provide more general information on application requirements and valuable input for evaluating the usability of various architectural features, i.e. information, which is needed for future co-design efforts aiming for exascale performance. |