DocumentCode
655082
Title
A Generic Methodology to Derive Empirical Power Consumption Prediction Models for Multi-Core Processors
Author
Basmadjian, Robert ; Rainer, Sebastian ; de Meer, Hermann
Author_Institution
Univ. of Passau, Passau, Germany
fYear
2013
fDate
Sept. 30 2013-Oct. 2 2013
Firstpage
167
Lastpage
174
Abstract
With the emergence of multi-core processors and their P-states, deriving power consumption models expressed in terms of easily extractable parameters becomes a necessity. In this paper, we introduce a generic methodology to devise power consumption estimation models for multi-core processors. The proposed approach takes into account three metrics: The number of active cores (i.e. executing instructions), frequency and utilization rate of the processor. Consequently, the derived models are expressed in terms of the above-mentioned parameters which can be extracted by any monitoring system. In order to assess the correctness as well as the accuracy of the proposed methodology, an experimental analysis is performed on Intel quad- and hexa-core processors. The results confirm the exactitude of the proposed methodology.
Keywords
microprocessor chips; multiprocessing systems; power aware computing; Intel quad-core processors; P-states; active cores; empirical power consumption prediction models; experimental analysis; generic methodology; hexa-core processors; multicore processors; processor utilization rate; Analytical models; Computational modeling; Equations; Mathematical model; Multicore processing; Power demand; Predictive models; Power consumption; empirical models; methodology; multi-core processors;
fLanguage
English
Publisher
ieee
Conference_Titel
Cloud and Green Computing (CGC), 2013 Third International Conference on
Conference_Location
Karlsruhe
Type
conf
DOI
10.1109/CGC.2013.32
Filename
6686024
Link To Document