Recent Advances in Parallel Virtual Machine and Message Passing Interface
Volume 4192 of the series Lecture Notes in Computer Science pp 40-48
MPI Collective Algorithm Selection and Quadtree Encoding
- Jelena Pješivac–GrbovićAffiliated withInnovative Computing Laboratory, University of Tennessee Computer Science Department
- , Graham E. FaggAffiliated withInnovative Computing Laboratory, University of Tennessee Computer Science Department
- , Thara AngskunAffiliated withInnovative Computing Laboratory, University of Tennessee Computer Science Department
- , George BosilcaAffiliated withInnovative Computing Laboratory, University of Tennessee Computer Science Department
- , Jack J. DongarraAffiliated withInnovative Computing Laboratory, University of Tennessee Computer Science Department
Abstract
Selecting the close-to-optimal collective algorithm based on the parameters of the collective call at run time is an important step in achieving good performance of MPI applications. In this paper, we focus on MPI collective algorithm selection process and explore the applicability of the quadtree encoding method to this problem. We construct quadtrees with different properties from the measured algorithm performance data and analyze the quality and performance of decision functions generated from these trees. The experimental data shows that in some cases, the decision function based on a quadtree structure with a mean depth of 3 can incur as little as a 5% performance penalty on average. The exact, experimentally measured, decision function for all tested collectives could be fully represented using quadtrees with a maximum of 6 levels. These results indicate that quadtrees may be a feasible choice for both processing of the performance data and automatic decision function generation.
- Title
- MPI Collective Algorithm Selection and Quadtree Encoding
- Book Title
- Recent Advances in Parallel Virtual Machine and Message Passing Interface
- Book Subtitle
- 13th European PVM/MPI User’s Group Meeting Bonn, Germany, September 17-20, 2006 Proceedings
- Pages
- pp 40-48
- Copyright
- 2006
- DOI
- 10.1007/11846802_14
- Print ISBN
- 978-3-540-39110-4
- Online ISBN
- 978-3-540-39112-8
- Series Title
- Lecture Notes in Computer Science
- Series Volume
- 4192
- Series ISSN
- 0302-9743
- Publisher
- Springer Berlin Heidelberg
- Copyright Holder
- Springer-Verlag Berlin Heidelberg
- Additional Links
- Topics
- Industry Sectors
- eBook Packages
- Editors
-
- Bernd Mohr (16)
- Jesper Larsson Träff (17)
- Joachim Worringen (18)
- Jack Dongarra (19)
- Editor Affiliations
-
- 16. Forschungszentrum Jülich, ZAM
- 17. NEC Europe Ltd., NEC Laboratories Europe
- 18. Dolphin Interconnect Solutions ASA R&D Germany
- 19. Computer Science Department, University of Tennessee
- Authors
-
- Jelena Pješivac–Grbović (20)
- Graham E. Fagg (20)
- Thara Angskun (20)
- George Bosilca (20)
- Jack J. Dongarra (20)
- Author Affiliations
-
- 20. Innovative Computing Laboratory, University of Tennessee Computer Science Department, 1122 Volunteer Blvd., Knoxville, TN, 37996-3450, USA
Continue reading...
To view the rest of this content please follow the download PDF link above.