A simple mean clock skew estimation algorithm for clock distribution networks in presence of random process variations and nonuniform substrate temperature

This paper presents a simple algorithm to estimate mean clock skew of a well balanced H-tree clock distribution network. The clock skew is a result of random process variations and nonuniform die temperature. The available literature separately discusses how to estimate clock skew in the presence of random process variations and nonuniform die temperature. This work combines the effects of both phenomena into a single equation using some simplifying assumptions. The derived equation can recursively apply to a well-balanced H-tree clock distribution network to estimate the mean clock skew. The proposed method estimates the mean clock skew with a high degree of accuracy. The delay distributions are assumed normal in the calculation. The computational effort needed for the proposed algorithm is O(N).