A Metamodel Residual-based Stopping Criterion for Adaptive Verification of Integrated Circuits

Adaptive verification is gaining increasing use as solution to overcome the coverage problem of complex integrated circuits' verification. As its name suggests, it places more points in regions of interest by learning the information from previous data, offering great potential to the engineer to understand the behavior of the system under study and to identify interesting regions in the design space. Since it is an iterative task, it should be stopped at the point that is optimum or near-optimum. This paper proposes a stopping criterion for planning adaptive experiments (simulations/measurements) for the identification of extreme regions (minimum/maximum) of the response. It is based on the concept of metamodeling and helps to efficiently characterize the targeted extreme region of complex integrated circuits with the advantage of low experimental effort. The approach was evaluated on synthetic test functions and an analog integrated circuit. The stopping criterion shows substantial improvements in the number of simulations/measurements while maintaining high resolution in the region of interest.