Power grid analysis and verification considering temperature variations

Process variations significantly impact leakage power, a pivotal parameter in designing a power grid. Because of the strong relationship between temperature and leakage power, the variations also impose statistical behavior on the operating temperature. In addition, the metal resistivity of a power grid increases with temperature. Therefore, ignoring the interdependency between leakage and temperature can introduce large errors in the power grid design. In this paper, a power grid analysis is proposed considering a statistical thermal profile across the grid. In addition, an efficient verification method is provided that ensures the robustness of the power grid in the presence of variations. A reliable estimation of the statistical thermal profile at the architecture level makes it possible to address any voltage violation early in the design process.