Abstract
With energy costs now accounting for nearly 30 % of a datacenter’s operating expenses, energy consumption has become an important issue when designing and executing a parallel algorithm. This paper analyzes the energy consumption of MPI applications following the master–slave paradigm. The analytical model is derived for this paradigm and is validated over a master–slave matrix-multiplication. This analytical model is parameterized through architectural and algorithmic parameters, and it is capable of predicting the energy consumption for a given instance of the problem over a given architecture. We use an external, metered, power distribution unit that allows to easily measure the power consumption of computing nodes without the needing of dedicated hardware.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Alonso P, Dolz MF, Igual FD, Mayo R, Quintana-Ortí ES Saving energy in the lu factorization with partial pivoting on multi-core processors. In: Stotzka et al [20], pp 353–358
Bedard D, Lim MY, Fowler RJ, Porterfield A (2010) Powermon: fine-grained and integrated power monitoring for commodity computer systems. In: Proceedings of the IEEE SoutheastCon 2010 (SoutheastCon), March 2010, pp 479–484
Castillo M, Fernández JC, Mayo R, Quintana-Ortí ES, Roca V Analysis of strategies to save energy for message-passing dense linear algebra kernels. In: Stotzka et al [20], pp 346–352
Chou C-Y, Chang H-Y, Wang S-T, Tcheng S-C (2006) Modeling message-passing overhead on nchc formosa pc cluster. In: Chung YC, Moreira JE (eds) GPC. Lecture notes in computer science, vol 3947. Springer, Berlin, pp 299–307
Davis J, Rivoire S, Goldszmidt M, Ardestani E (2011) Accounting for variability in large-scale cluster power models. In: The second exascale evaluation and research techniques workshop, Held in conjunction with ASPLOS 2011
Enos J, Steffen CP, Fullop J, Showerman MT, Shi G, Esler K, Kindratenko VV, Stone JE, Phillips JC (2010) Quantifying the impact of gpus on performance and energy efficiency in hpc clusters. In: Green computing conference. IEEE Press, New York, pp 317–324
Freeh VW, Lowenthal DK, Pan F, Kappiah N, Springer R, Rountree B, Femal ME (2007) Analyzing the energy-time trade-off in high-performance computing applications. IEEE Trans Parallel Distrib Syst 18(6):835–848
Hähnel M, Döbel B, Völp M, Härtig H (2012) Measuring energy consumption for short code paths using rapl. In: GreenMetrics 2012 workshop, in conjunction with ACM sigmetrics/performance 2012, ACM sigmetrics, London, UK, 11 June 2012. ACM, New York
Hockney RW (1994) The communication challenge for mpp: intel paragon and meiko cs-2. Parallel Comput 20(3):389–398
Keys L, Rivoire S, Davis JD (2010) The search for energy-efficient building blocks for the data center. In: Varbanescu AL, Molnos AM, van Nieuwpoort R (eds) ISCA workshops. Lecture notes in computer science, vol 6161. Springer, Berlin, pp 172–182
Lim MY, Porterfield A, Fowler RJ (2010) Softpower: fine-grain power estimations using performance counters. In: Hariri S, Keahey K (eds) HPDC. ACM, New York, pp 308–311
Lim MY, Freeh VW, Lowenthal DK (2011) Adaptive, transparent cpu scaling algorithms leveraging inter-node mpi communication regions. Parallel Comput 37(10–11):667–683
Martinez DR, Cabaleiro JC, Pena TF, Rivera FF, Blanco V (2010) Performance modeling of mpi applications using model selection techniques. In: Danelutto M, Bourgeais J, Gross T (eds) 18th euromicro conference on parallel, distributed and network–based processing, PDP2010, Pisa, Italy, February 2010. IEEE Computer Society, Los Alamitos, pp 95–102
Martínez DR, Albín JL, Pena TF, Cabaleiro JC, Rivera FF, Blanco V (2011) Using accurate AIC-based performance models to improve the scheduling of parallel applications. J Supercomput 58(3):332–340. doi:10.1007/s11227-011-0589-1
MPICH-a portable implementation of MPI, November 2005
Mullen KM, van Stokkum IHM (2010) nnls: the Lawson–Hanson algorithm for non-negative least squares (nnls), April 2010
Nurmi D, Mandal A, Brevik J, Koelbel C, Wolski R, Kennedy K (2006) Evaluation of a workflow scheduler using integrated performance modelling and batch queue wait time prediction. In: Proceedings of the ACM/IEEE SC2006 conference on high performance networking and computing, Tampa, FL, USA, 11–17 November 2006. ACM, New York, p 119
Rivoire S, Shah MA, Ranganathan P, Kozyrakis C, Meza J (2007) Models and metrics to enable energy-efficiency optimizations. IEEE Comput 40(12):39–48
Schuermans A (2012) Schleifenbauer products bv, March 2012
Stotzka R, Schiffers M, Cotronis Y (eds) (2012) Proceedings of the 20th euromicro international conference on parallel, distributed and network-based processing, PDP 2012, Munich, Germany, 15–17 February 2012. IEEE Press, New York, pp 15–17
Various (2012) R, language and environment for statistical computing and graphics, March 2012
Xu Z, Hwang K (1996) Modeling communication overhead: mpi and mpl performance on the ibm sp2. IEEE Parallel Distrib Technol, Syst Appl 4(1):9–24
Acknowledgements
This work was supported by the Spanish MEC projects TIN2011-24598 and TIN2008-06570-C04-03, the FPU program, ACIISI contract ProID20100222, and COST-ICT-0805 and CAPAP-H3 research networks.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Almeida, F., Blanco, V., Cabrera, A. et al. Modeling energy consumption for master–slave applications. J Supercomput 65, 1137–1149 (2013). https://doi.org/10.1007/s11227-013-0914-y
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11227-013-0914-y