Abstract
HeteroMPI is an extension of MPI designed for high performance computing on heterogeneous networks of computers. The recent new feature of HeteroMPI is the optimized version of collective communications. The optimization is based on a novel performance communication model of switch-based computational clusters. In particular, the model reflects significant non-deterministic and non-linear escalations of the execution time of many-to-one collective operations for medium-sized messages. The paper outlines this communication model and describes how HeteroMPI uses this model to optimize one-to-many (scatter-like) and many-to-one (gather-like) communications. We also demonstrate that HeteroMPI collective communications outperform their native counterparts for various MPI implementations and cluster platforms.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Dongarra, J., Huss-Lederman, S., Otto, S., Snir, M., Walker, D.: MPI: The Complete Reference. The MIT Press, Cambridge (1996)
Lastovetsky, A., Reddy, R.: HeteroMPI: Towards a message-passing library for heterogeneous networks of computers. J. of Parallel and Distr. Comp. 66, 197–220 (2006)
Lastovetsky, A., O’Flynn, M.: A Performance Model of Many-to-One Collective Communications for Parallel Computing. In: Proc. of IPDPS 2007, Long Beach, CA (2007)
Vadhiyar, S.S., Fagg, G.E., Dongarra, J.: Automatically tuned collective communications. In: Proc. of Supercomputing 1999, Portland, OR (1999)
Thakur, R., Rabenseifner, R., Gropp, W.: Optimization of Collective Communication Operations in MPICH. Int. J. of High Perf. Comp. App. 19, 49–66 (2005)
Pjesivac-Grbovic, J., Angskun, T., Bosilca, G., Fagg, G.E., Gabriel, E., Dongarra, J.J.: Performance Analysis of MPI Collective Operations. In: Proc. of IPDPS 2005, Denver, CO (2005)
Kielmann, T., Hofman, R.F.H., Bal, H., Plaat, A., Bhoedjang, R.A.F.: MagPIe: MPI’s collective communication operations for clustered wide area systems. In: Proc. of PPoPP 1999, pp. 131–140. ACM Press, New York (1999)
Iannello, G.: Efficient algorithms for the reduce-scatter operation in LogGP. IEEE Transactions on Parallel and Distr. Systems 8(9), 970–982 (1997)
Benson, G.D., Chu, C-W., Huang, Q., Caglar, S.G.: A comparison of MPICH allgather algorithms on switched networks. In: Dongarra, J.J., Laforenza, D., Orlando, S. (eds.) Recent Advances in Parallel Virtual Machine and Message Passing Interface. LNCS, vol. 2840, pp. 335–343. Springer, Heidelberg (2003)
Lastovetsky, A., Mkwawa, I., O’Flynn, M.: An Accurate Communication Model of a Heterogeneous Cluster Based on a Switch-Enabled Ethernet Network. In: Proc. of ICPADS 2006, Minneapolis, MN, pp. 15–20 (2006)
Culler, D., Karp, R., Patterson, R., Sahay, A., Schauser, K.E., Santos, R.S.E., von Eicken, T.: LogP: Towards a realistic model of parallel computation. In: Proc. of the 4th ACM SIGPLAN, ACM Press, New York (1993)
LAM/MPI User’s Guide, http://www.lam-mpi.org/
Open MPI Publications, http://www.open-mpi.org/
MPICH/MPICH-2 User’s Guide, http://www-unix.mcs.anl.gov/mpi/
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 2007 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Lastovetsky, A., O’Flynn, M., Rychkov, V. (2007). Optimization of Collective Communications in HeteroMPI. In: Cappello, F., Herault, T., Dongarra, J. (eds) Recent Advances in Parallel Virtual Machine and Message Passing Interface. EuroPVM/MPI 2007. Lecture Notes in Computer Science, vol 4757. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-75416-9_23
Download citation
DOI: https://doi.org/10.1007/978-3-540-75416-9_23
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-75415-2
Online ISBN: 978-3-540-75416-9
eBook Packages: Computer ScienceComputer Science (R0)