A basic model for profiling power consumption in HPC subsystems

The emergence of the term ‘Green Computing’ itself is sufficient to describe its need. Regarding computers, energy estimation is the first step towards optimization. This paper describes a simple model which generates profiling results for scientific applications. It is designed to analyze a software process in order to identify its power consuming components (e.g. Instruction Count). The study specifically aims at the CPU and the RAM as two main subsystems, since the Computing Machine is a highly complex association and every part of it contributes to the energy consumption. The formulation process of this model as well as its application are presented and discussed. The results obtained out of these experimentations are also presented.

[1]  Aie.  World Energy Outlook 2015 , 2015 .

[2]  J. Steele.  ACPI thermal sensing and control in the PC , 1998, Wescon/98. Conference Proceedings (Cat. No.98CH36265).

[3]  Ioana Banicescu,et al.  A Utility Based Power-Aware Autonomic Approach for Running Scientific Applications , 2012, 2012 IEEE 26th International Parallel and Distributed Processing Symposium Workshops & PhD Forum.

[4]  Deva Bodas,et al.  Simple Power-Aware Scheduler to Limit Power Consumption by HPC System within a Budget , 2014, 2014 Energy Efficient Supercomputing Workshop.

[5]  Zhili Sun,et al.  A Survey of Power-Saving Techniques on Data Centers and Content Delivery Networks , 2013, IEEE Communications Surveys & Tutorials.

[6]  Shrirang M. Yardi,et al.  CAMP: A technique to estimate per-structure power at run-time using a few simple parameters , 2009, 2009 IEEE 15th International Symposium on High Performance Computer Architecture.

[7]  Mitsuhisa Sato,et al.  Runtime DVFS control with instrumented Code in power-scalable cluster system , 2008, 2008 IEEE International Conference on Cluster Computing.

[8]  Michael Voss,et al.  Dynamic instrumentation of large-scale MPI and OpenMP applications , 2003, Proceedings International Parallel and Distributed Processing Symposium.

[9]  Anand Raghunathan,et al.  Power monitors: a framework for system-level power estimation using heterogeneous power models , 2005, 18th International Conference on VLSI Design held jointly with 4th International Conference on Embedded Systems Design.

[10]  Sujit Dey,et al.  High-level macro-modeling and estimation techniques for switching activity and power consumption , 2003, IEEE Trans. Very Large Scale Integr. Syst..

[11]  Bo Li,et al.  Harnessing renewable energy in cloud datacenters: opportunities and challenges , 2014, IEEE Network.

[12]  Stephen W. Poole,et al.  Power measurement for high performance computing: State of the art , 2011, 2011 International Green Computing Conference and Workshops.

[13]  Nikil D. Dutt,et al.  A Multi-Granularity Power Modeling Methodology for Embedded Processors , 2011, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[14]  Vijay P. Bhatkar.  PARAM parallel supercomputer: architecture, programming environment, and applications , 1994, Proceedings of 8th International Parallel Processing Symposium.

[15]  Vincent M. Weaver.  Self-monitoring overhead of the Linux perf_ event performance counter interface , 2015, 2015 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS).

[16]  Lizy Kurian John,et al.  Complete System Power Estimation: A Trickle-Down Approach Based on Performance Events , 2007, 2007 IEEE International Symposium on Performance Analysis of Systems & Software.

[17]  Lizy Kurian John,et al.  Complete System Power Estimation Using Processor Performance Events , 2012, IEEE Transactions on Computers.