A Stochastic Model for Interconnection Complexity based on Rent’s Rule

In the past, Rent’s rule has been successfully applied for a priori estimation of wire length distributions. Deviations to Rent’s rule appear due to the existence of heterogeneity. This can be classified in hierarchical and spatial heterogeneity. Stochastic models for the interconnection complexity, based on Rent’s rule, are introduced. A coarse model shows that the variance follows a power law relationship. A more refined model incorporates the effect of local spatial heterogeneity. Experiments show that this is sufficient to model the variance of the terminal count distribution. Finally, the model is further extended to incorporate global spatial heterogeneity.