Framework for statistical analysis of homogeneous multicore power grid networks

Abstract

In this paper, we propose a framework to analyze the large-scaled multicore power grid network statistically by first building a simplified multicore power supply distribution model. We then apply the Modified Nodal Analysis (MNA) method on a simplified power gird circuit. Under such a framework, most statistical approaches, including Monte Carlo (MC), Importance Sampling, and Stochastic Spectrum Analysis, can be applied to analyze the processinduced variation of homogeneous multicore power grid networks. In the experiment, we focus on the subthreshold leakage current variations, which are modeled as lognormal distribution random variables, by using MC approach as an example to demonstrate the feasibility of such a framework. 1