Effectiveness of neural networks for power modeling for Cloud and HPC: It's worth it!

Power consumption of servers and applications are of utmost importance as computers are becoming ubiquitous, from smart phones to IoT and full-fledged computers. To optimize their power consumption, knowledge is necessary during execution at different levels: for the Operating System to take decisions of scheduling, for users to choose between different applications. Several models exist to evaluate the power consumption of computers without relying on actual wattmeters: Indeed, these hardware are costly but also usually have limits on their pooling frequency (usually a one-second frequency is observed) except for dedicated professional hardware. The models link applications behavior with their power consumption, but up to now there is a 5% wall: Most models cannot reduce their error under this threshold and are usually linked to a particular hardware configuration. This article demonstrates how to break the 5% wall of power models. It shows that by using neural networks it is possible to create models with 1% to 2% error. It also quantifies the reachable precision obtainable with other classical methods such as analytical models.

[1]  Alejandro Rico,et al.  Experiences with mobile processors for energy efficient HPC , 2013, 2013 Design, Automation & Test in Europe Conference & Exhibition (DATE).

[2]  Weisong Shi,et al.  Fine-grained power management using process-level profiling , 2012, Sustain. Comput. Informatics Syst..

[3]  Sameer Alawnah,et al.  Modeling of smartphones’ power using neural networks , 2017, EURASIP J. Embed. Syst..

[4]  Thierry Monteil,et al.  Mixed integer linear programming for quality of service optimization in Clouds , 2017, Future Gener. Comput. Syst..

[5]  Haifeng Wang,et al.  Predicting power consumption of GPUs with fuzzy wavelet neural networks , 2015, Parallel Comput..

[6]  Bernd Mohr,et al.  Modeling CPU Energy Consumption of HPC Applications on the IBM POWER7 , 2014, 2014 22nd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing.

[7]  Shufeng Tan,et al.  Reducing data dimensionality through optimizing neural network inputs , 1995 .

[8]  Eduard Ayguadé,et al.  Decomposable and responsive power models for multicore processors using performance counters , 2010, ICS '10.

[9]  D. van der Spoel,et al.  GROMACS: A message-passing parallel molecular dynamics implementation , 1995 .

[10]  Bernhard Schölkopf,et al.  A tutorial on support vector regression , 2004, Stat. Comput..

[11]  Kevin Leyton-Brown,et al.  Algorithm runtime prediction: Methods & evaluation , 2012, Artif. Intell..

[12]  Qi Zhao,et al.  iMeter: An integrated VM power model based on performance profiling , 2013, Future Gener. Comput. Syst..

[13]  Somnath Mazumdar,et al.  Power efficient server consolidation for Cloud data center , 2017, Future Gener. Comput. Syst..

[14]  Dong Li,et al.  PowerPack: Energy Profiling and Analysis of High-Performance Systems and Applications , 2010, IEEE Transactions on Parallel and Distributed Systems.

[15]  Leandro Fontoura Cupertino,et al.  Modeling the power consumption of computing systems and applications through machine learning techniques , 2015 .

[16]  Tajana Simunic,et al.  P4: Phase-based power/performance prediction of heterogeneous systems via neural networks , 2017, 2017 IEEE/ACM International Conference on Computer-Aided Design (ICCAD).

[17]  Sudhakar Yalamanchili,et al.  Power Modeling for GPU Architectures Using McPAT , 2014, TODE.

[18]  Ching-Hsien Hsu,et al.  Experimental and quantitative analysis of server power model for cloud data centers , 2016, Future Gener. Comput. Syst..

[19]  R. Tibshirani.  Regression Shrinkage and Selection via the Lasso , 1996 .

[20]  Ananta Tiwari,et al.  Modeling Power and Energy Usage of HPC Kernels , 2012, 2012 IEEE 26th International Parallel and Distributed Processing Symposium Workshops & PhD Forum.

[21]  Mahadev Satyanarayanan,et al.  PowerScope: a tool for profiling the energy usage of mobile applications , 1999, Proceedings WMCSA'99. Second IEEE Workshop on Mobile Computing Systems and Applications.

[22]  I K Fodor,et al.  A Survey of Dimension Reduction Techniques , 2002 .

[23]  Andreas Gerstlauer,et al.  Accurate phase-level cross-platform power and performance estimation , 2016, 2016 53nd ACM/EDAC/IEEE Design Automation Conference (DAC).

[24]  James L. Peterson,et al.  Design and validation of a performance and power simulator for PowerPC systems , 2003, IBM J. Res. Dev..

[25]  Pascal Bouvry,et al.  Performance Evaluation and Energy Efficiency of High-Density HPC Platforms Based on Intel, AMD and ARM Processors , 2013, EE-LSDS.

[26]  Rizos Sakellariou,et al.  A Survey of Power and Energy Predictive Models in HPC Systems and Applications , 2017, ACM Comput. Surv..

[27]  Jan Weglarz,et al.  Practical power consumption estimation for real life HPC applications , 2013, Future Gener. Comput. Syst..

[28]  Carlo Ghezzi,et al.  A quality driven extension to the QVT-relations transformation language , 2011, Computer Science - Research and Development.

[29]  Jack Dongarra,et al.  HPC Challenge: Design, History, and Implementation Highlights , 2017 .

[30]  J. Węglarz,et al.  Runtime power usage estimation of HPC servers for various classes of real-life applications , 2014, Future Gener. Comput. Syst..