Abstract
In this paper, we propose an improved model for predicting HPL (High performance Linpack) performance. In order to accurately predict the maximal LINPACK performance we first divide the performance model into two parts: computational cost and message passing overhead. In the message passing overhead, we adopt Xu and Hwang’s broadcast model instead of the point-to-point message passing model. HPL performance prediction is a multi-variables problem. In this proposed model we improved the existing model by introducing a weighting function to account for many effects such that the proposed model could more accurately predict the maximal LINPACK performance R max . This improvement in prediction accuracy has been verified on a variety of architectures, including IA64 and IA32 CPUs in a Myrinet-based environment, as well as in Quadrics, Gigabits Ethernet and other network environments. Our improved model can help cluster users in estimating the maximal HPL performance of their systems.
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
Sterling, T., Becker, D., Savarese, D., et al.: BEOWULF: A Parallel Workstation for Scientific Computation. In: Proc. Of the 1995 International Conf. On Parallel Processing (1995)
Sterling, T., Savarese, D., Becker, D., et al.: Communication Overhead for Space Science Applications on the Beowulf Parallel Workstation. In: Proc. of 4th IEEE Symposium on High Performance Distributed Computing (1995)
Reschke, C., Sterling, T., Ridge, D.: A Design Study of Alternative Network Topologies for the Beowulf Parallel Workstation. In: Proceedings of the 5th IEEE Symposium on High Performance and Distributed Computing (1996)
Ridge, D., Becker, D., Merkey, P.: Beowulf: Harnessing the Power of Parallelism in a Pile-of-PCs. In: Proceedings of IEEE Aerospace (1997)
Pfister, G.F.: In Search of Clusters. Prentice-Hall, Englewood Cliffs (1998)
Top 500 List, http://www.top500.org
HPL Web site, http://www.netlib.org/benchmark/hpl/
Wang, P., et al.: LINPACK Performance on a Geographically Distributed Linux Cluster. In: 18th International Parallel and Distributed Processing Symposium (IPDPS’04), Santa Fe, New Mexico (2004)
Hockney, R.W.: The Communication Challenge for MPP: Intel Paragon and Meiko CS-2. Parallel Computing 20, 389–398 (1994)
Xu, Z., Hwang, K.: Modeling Communication Overhead: MPI and MPL Performance on the IBM SP2. IEEE Parallel & Distributed Technology 4(1), 9–23 (1996)
NCHC Formosa PC Cluster Home Page, http://formosa.nchc.org.tw
NCHC Triton Cluster Home Page, http://www/english/pcCluster.php
Zhang, W., Chen, M., Fan, J.: HPL Performance Prevision to Intending System Improvement. In: Second International Symposium on Parallel and Distributed Processing and Applications (2004)
Boden, N.J., et al.: Myrinet: A Giga-bit-per-second Local-area Network. In: IEEE micro (1995)
Burns, G., Daoud, R., Vaigl, J.: LAM: An Open Cluster Environment for MPI. In: Proceedings of Supercomputing Symposium’94, pp. 379–386 (1994)
Petrini, F., et al.: Performance Evaluation of the Quadrics Interconnection Network. Cluster Computing (2003)
GOTO library, http://www.cs.utexas.edu/users/kgoto
Chou, C.-Y., et al.: Modeling Message-Passing Overhead on NCHC Formosa PC Cluster. In: Chung, Y.-C., Moreira, J.E. (eds.) GPC 2006. LNCS, vol. 3947, pp. 299–307. Springer, Heidelberg (2006)
Zhang, W., Fan, J., Chen, M.: Efficient Determination of Block Size NB for Parallel Linpack Test. In: The 16th IASTED International Conference on Parallel and Distributed Computing and Systems (2004)
Gunawan, T., Cai, W.: Performance Analysis of a Myrinet-Based Cluster. Cluster Computing 6, 229–313 (2003)
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 2007 Springer Berlin Heidelberg
About this paper
Cite this paper
Chou, CY., Chang, HY., Wang, ST., Huang, KC., Shen, CY. (2007). An Improved Model for Predicting HPL Performance. In: Cérin, C., Li, KC. (eds) Advances in Grid and Pervasive Computing. GPC 2007. Lecture Notes in Computer Science, vol 4459. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-72360-8_14
Download citation
DOI: https://doi.org/10.1007/978-3-540-72360-8_14
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-72359-2
Online ISBN: 978-3-540-72360-8
eBook Packages: Computer ScienceComputer Science (R0)