Abstract
Collectives are an important component of parallel programs, and have a significant impact on performance and scalability of an application. To obtain best performance, platform specific implementations of various parallel programming frameworks, such as MPI and Charm++, are done. As a result, when systems with new network topologies are built, new topology aware algorithms for collectives are added to these frameworks that also contain the topology oblivious algorithms. In this paper, we propose topology aware algorithms for collectives performed on two-tier direct networks such as IBM PERCS and Dragonfly. We observe that, for large message operations, significant performance gains can be made by taking advantage of large number of links in a two-tier direct network. We evaluate proposed algorithms using an analytical model based on link utilization.
Keywords
- Collectives
- Topology
- Two-tier networks
- PERCS
- Dragonfly
This is a preview of subscription content, access via your institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
MPI: A Message Passing Interface Standard. MPI Forum
Arimilli, B., Arimilli, R., Chung, V., Clark, S., Denzel, W., Drerup, B., Hoefler, T., Joyner, J., Lewis, J., Li, J., Ni, N., Rajamony, R.: The PERCS High-Performance Interconnect. In: 2010 IEEE 18th Annual Symposium on High Performance Interconnects (HOTI), pp. 75–82 (August 2010)
Barnett, M., Gupta, S., Payne, D.G., Shuler, L., Geijn, R., Watts, J.: Interprocessor Collective Communication Library (InterCom). In: Proceedings of the Scalable High Performance Computing Conference, pp. 357–364 (1994)
Faraj, A., Kumar, S., Smith, B., Mamidala, A., Gunnels, J., Heidelberger, P.: Mpi collective communications on the blue gene/p supercomputer: algorithms and optimizations. In: Proceedings of the 23rd International Conference on Supercomputing, ICS 2009, pp. 489–490 (2009)
Jain, N., Sabharwal, Y.: Optimal bucket algorithms for large mpi collectives on torus interconnects. In: Proceedings of the 24th ACM International Conference on Supercomputing, ICS 2010, pp. 27–36 (2010)
Kim, J., Dally, W.J., Scott, S., Abts, D.: Technology-driven, highly-scalable dragonfly topology. SIGARCH Comput. Archit. News 36, 77–88 (2008)
Rabenseifner, R.: A new optimized MPI reduce algorithm (1997)
Thakur, R., Gropp, W.D.: Improving the Performance of Collective Operations in MPICH. In: Dongarra, J., Laforenza, D., Orlando, S. (eds.) EuroPVM/MPI 2003. LNCS, vol. 2840, pp. 257–267. Springer, Heidelberg (2003)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2012 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Jain, N., Lau, J., Kale, L. (2012). Collectives on Two-Tier Direct Networks. In: Träff, J.L., Benkner, S., Dongarra, J.J. (eds) Recent Advances in the Message Passing Interface. EuroMPI 2012. Lecture Notes in Computer Science, vol 7490. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-33518-1_12
Download citation
DOI: https://doi.org/10.1007/978-3-642-33518-1_12
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-33517-4
Online ISBN: 978-3-642-33518-1
eBook Packages: Computer ScienceComputer Science (R0)
