skip to main content
10.1145/2482759.2482762acmotherconferencesArticle/Chapter ViewAbstractPublication PagesinaConference Proceedingsconference-collections
research-article

Randomizing task placement does not randomize traffic (enough)

Published: 23 January 2013 Publication History

Abstract

Dragonflies are one of the most promising topologies for the Exascale effort for their scalability and cost. Dragonflies achieve very high throughput under uniform traffic, but have a pathological behavior under other regular traffic patterns, some of them very common in HPC applications. A recent study showed that randomization of task placement can make pathological, regular (multi-dimensional stencil) traffic patterns behave similar to uniform traffic.
In this work we provide a theoretical model that is able to predict the expected performance of a generic dragonfly network under uniform traffic and characterize performance-optimal dragonflies. We then analyze whether this model can be extended to other patterns by means of benchmarking the performance of multiple such patterns under both contiguous and randomized task placement. We conclude that, although in comparison with contiguous task placement, randomization does lead to a significant improvement in performance for pathological communication patterns, this performance is not on par with that of uniform traffic, but rather half of it.

References

[1]
A. Bhatele, N. Jain, W. D. Gropp, and L. V. Kale. Avoiding hot-spots on two-level direct networks. In Proceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis, SC '11, pages 76:1--76:11, New York, NY, USA, 2011. ACM.
[2]
W. Dally and B. Towles. Principles and Practices of Interconnection Networks. Morgan Kaufmann Publishers Inc., San Francisco, CA, USA, 2003.
[3]
M. García, E. Vallejo, R. Beivide, M. Odriozola, C. Camarero, M. Valero, G. Rodríguez, J. Labarta, and C. Minkenberg. On-the-fly adaptive routing in high-radix hierarchical networks. In The 41st International Conference on Parallel Processing (ICPP), 09 2012.
[4]
J. Kim, W. J. Dally, S. Scott, and D. Abts. Technology-driven, highly-scalable dragonfly topology. In Proceedings of the 35th Annual International Symposium on Computer Architecture, ISCA '08, pages 77--88, Washington, DC, USA, 2008. IEEE Computer Society.
[5]
L. G. Valiant. A scheme for fast parallel communication. SIAM J. Comput., 11(2):350--361, 1982.

Cited By

View all
  • (2018)Brief AnnouncementProceedings of the 30th on Symposium on Parallelism in Algorithms and Architectures10.1145/3210377.3210665(91-93)Online publication date: 11-Jul-2018
  • (2016)An evaluation of network architectures for next generation supercomputersProceedings of the 7th International Workshop on Performance Modeling, Benchmarking and Simulation of High Performance Computing Systems10.5555/3019057.3019059(11-21)Online publication date: 13-Nov-2016
  • (2016)An Evaluation of Network Architectures for Next Generation Supercomputers2016 7th International Workshop on Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems (PMBS)10.1109/PMBS.2016.007(11-21)Online publication date: Nov-2016
  • Show More Cited By

Recommendations

Comments

Information & Contributors

Information

Published In

cover image ACM Other conferences
IMA-OCMC '13: Proceedings of the 2013 Interconnection Network Architecture: On-Chip, Multi-Chip
January 2013
29 pages
ISBN:9781450317849
DOI:10.1145/2482759
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]

Publisher

Association for Computing Machinery

New York, NY, United States

Publication History

Published: 23 January 2013

Permissions

Request permissions for this article.

Check for updates

Author Tags

  1. dragonfly networks
  2. high performance computing
  3. network throughput
  4. random task placement
  5. uniform traffic

Qualifiers

  • Research-article

Conference

INA-OCMC '13
INA-OCMC '13: On-Chip, Multi-Chip
January 23, 2013
Berlin, Germany

Acceptance Rates

IMA-OCMC '13 Paper Acceptance Rate 7 of 17 submissions, 41%;
Overall Acceptance Rate 12 of 27 submissions, 44%

Contributors

Other Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

  • Downloads (Last 12 months)3
  • Downloads (Last 6 weeks)0
Reflects downloads up to 06 Nov 2024

Other Metrics

Citations

Cited By

View all
  • (2018)Brief AnnouncementProceedings of the 30th on Symposium on Parallelism in Algorithms and Architectures10.1145/3210377.3210665(91-93)Online publication date: 11-Jul-2018
  • (2016)An evaluation of network architectures for next generation supercomputersProceedings of the 7th International Workshop on Performance Modeling, Benchmarking and Simulation of High Performance Computing Systems10.5555/3019057.3019059(11-21)Online publication date: 13-Nov-2016
  • (2016)An Evaluation of Network Architectures for Next Generation Supercomputers2016 7th International Workshop on Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems (PMBS)10.1109/PMBS.2016.007(11-21)Online publication date: Nov-2016
  • (2015)Overcoming far-end congestion in large-scale networks2015 IEEE 21st International Symposium on High Performance Computer Architecture (HPCA)10.1109/HPCA.2015.7056051(415-427)Online publication date: Feb-2015
  • (2015)Comparing Global Link Arrangements for Dragonfly NetworksProceedings of the 2015 IEEE International Conference on Cluster Computing10.1109/CLUSTER.2015.57(361-370)Online publication date: 8-Sep-2015
  • (2014)Efficient task placement and routing of nearest neighbor exchanges in dragonfly networksProceedings of the 23rd international symposium on High-performance parallel and distributed computing10.1145/2600212.2600225(129-140)Online publication date: 23-Jun-2014
  • (2014)Performance implications of remote-only load balancing under adversarial traffic in DragonfliesProceedings of the 8th International Workshop on Interconnection Network Architecture: On-Chip, Multi-Chip10.1145/2556857.2556860(1-4)Online publication date: 22-Jan-2014

View Options

Get Access

Login options

View options

PDF

View or Download as a PDF file.

PDF

eReader

View online with eReader.

eReader

Media

Figures

Other

Tables

Share

Share

Share this Publication link

Share on social media