Transistor-Level Gate Modeling for Nano CMOS Circuit Verification Considering Statistical Process Variations

Equationor table-based gate-level models (GLMs) have been applied in static timing analysis (STA) for decades. In order to evaluate the impact of statistical process variabilities, Monte Carlo (MC) simulations are utilized with GLMs for statistical static timing analysis (SSTA), which requires a massive amount of CPU time. Driven by the challenges associated with CMOS technology scaling to 45nm and below, intensive efforts have been contributed to optimize GLMs for higher accuracy at the expense of enhanced complexity. In order to maintain both accuracy and efficiency at 45nm node and below, in this paper we present a gate model built from a simplified transistor model. Considering the increasing statistical process variabilities, the model is embedded in our new statistical simulation engine, which can do both implicit non-MC statistical as well as deterministic simulations. Results of timing, noise and power grid analysis are presented using a 45nm PTMLP technology.