Chip level statistical leakage power estimation using generalized extreme value distribution

Authors:
Alireza Khosropour;Hossein Aghababa;Ali Afzali-Kusha;Behjat Forouzandeh
Affiliations:
Nanoelectronics Center of Excellence, School of Electrical and Computer Engineering, University of Tehran, Tehran, Iran;Nanoelectronics Center of Excellence, School of Electrical and Computer Engineering, University of Tehran, Tehran, Iran;Nanoelectronics Center of Excellence, School of Electrical and Computer Engineering, University of Tehran, Tehran, Iran;Nanoelectronics Center of Excellence, School of Electrical and Computer Engineering, University of Tehran, Tehran, Iran
Venue:
PATMOS'11 Proceedings of the 21st international conference on Integrated circuit and system design: power and timing modeling, optimization, and simulation
Year:
2011

Citing 7
Cited 0

Parameter variations and impact on circuits and microarchitecture

Proceedings of the 40th annual Design Automation Conference
Statistical estimation of leakage current considering inter- and intra-die process variation

Proceedings of the 2003 international symposium on Low power electronics and design
Parametric yield estimation considering leakage variability

Proceedings of the 41st annual Design Automation Conference
Full-chip analysis of leakage power under process variations, including spatial correlations

Proceedings of the 42nd annual Design Automation Conference
Projection-based statistical analysis of full-chip leakage power with non-log-normal distributions

Proceedings of the 43rd annual Design Automation Conference
Statistical analysis of full-chip leakage power considering junction tunneling leakage

Proceedings of the 44th annual Design Automation Conference
Efficient additive statistical leakage estimation

IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

Previous works for full-chip leakage power estimation are all based on Wilkinson's approach which approximates sum of lognormal random variables as another lognormal by matching the first and second moments. In this paper we will show that natural logarithm of leakage deviates from normal distribution by scaling transistor sizes, as a result distribution of leakage power cannot be described by lognormal distribution anymore. We will introduce generalized extreme value distribution as the best candidate for full-chip leakage power estimation and we will prove its superiority over lognormal approximation through simulation results in 45nm technology.