Abstract
A modern Petascale System consists of millions of different components, which consume a huge amount of energy. The power rating of each component depends on the type of the current instructions, executed on cores, memory controllers, network units and other various components. There are a lot of influences and complicated dependencies between the software, environment and the energy consumption. The objective of this work is to identify and understand the energy consumption of processors and memory in the consideration of kernel operations. Another important goal is to develop the methodology by which the developers and users could estimate the energy consumption of the different algorithms on different systems with minimal effort and satisfying accuracy.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
We do not run any calculations on the weak graphics card and the fans run at a constant speed. We assume that the power consumption of these components is constant. This was also confirmed with a series of extra tests.
- 2.
We vary the number of active cores, frequency and hierarchy levels of memory, on that the processor operates.
- 3.
The best approximation for the same kernel operation Add on Sandy Bridge Intel i5-2500 (6 M Cache, up to 3.7 GHz, 4 cores) is by ρ = 2. 7.
- 4.
If we compare the obtained values to the data from the article [6] we see that the energy consumption of Sandy Bridge (E5-2687W) and of the previous generation Westmere (Intel XEON X5670) are within the same order of magnitude.
References
Enhanced Intel SpeedStep Technology for the Intel Pentium M Processor - White Paper. ftp://download.intel.com/design/network/papers/30117401.pdf Intel Corparation (2004)
David Kanter, Intel’s Haswell CPU Microarchitecture http://www.realworldtech.com/haswell-cpu (November 2012)
S. Williams, A. Waterman and D. Patterson. Roofline: An Insightful Visual Performance Model for Multicore Architectures Communications of the ACM, Vol. 52, No. 4. (April 2009), pp. 65–76, doi:10.1145/1498765.1498785
Robert Chappell, Bret Toll, Ronal Singhal, Intel Next Generation Micro Architecture Codename Haswell: New Processor Innovations. Presented at IDF (2012)
Markus Wittmann, Georg Hager, Thomas Zeiser, Gerhard Wellein. An analysis of energy-optimized lattice-Boltzmann CFD simulations from the chip to the highly parallel level arXiv:1304.7664 (April 2013)
Daniel Molka, Daniel Hackenberg, Robert Schöne and Matthias S. Müller, Characterizing the Energy Consumption of Data Transfers and Arithmetic Operations on x86-64 Processors, Proceeding of the first International Green Computing Conference, (August 2010)
Acknowledgements
This work has been supported by the CRESTA project that has received funding from the European Community’s Seventh Framework Programme (ICT-2011.9.13) under Grant Agreement no. 287703 and by the ExaSolvers project that has received funding from the German Research Foundation (DFG) as part of the Priority Programme “Software for Exascale Computing–SPPEXA”.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2013 Springer International Publishing Switzerland
About this paper
Cite this paper
Khabi, D., Küster, U. (2013). Power Consumption of Kernel Operations. In: Resch, M., Bez, W., Focht, E., Kobayashi, H., Kovalenko, Y. (eds) Sustained Simulation Performance 2013. Springer, Cham. https://doi.org/10.1007/978-3-319-01439-5_3
Download citation
DOI: https://doi.org/10.1007/978-3-319-01439-5_3
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-01438-8
Online ISBN: 978-3-319-01439-5
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)