Improving the efficiency and quality of simulation-based behavioral model verification using dynamic Bayesian criteria

In order to improve the effectiveness of simulation-based behavioral verification, it is important to determine when to stop the current test strategy and to switch to an expectantly more rewarding test strategy. The location of a stopping point is dependent on the statistical model one chooses to describe the coverage behavior during verification. In this paper, we present dynamic Bayesian (DB) and confidence-based dynamic Bayesian (CDB) stopping rules for behavioral VHDL model verification. The statistical assumptions of the proposed stopping rules are based on experimental evaluation of probability distribution functions and correlation functions. Fourteen behavioral VHDL models were experimented with to determine the high efficiency of the proposed stopping rules over the existing ones. Results show that the DB and the CDB stopping rules outperform all the existing stopping rules with an average improvement of at least 69% in coverage per testing patterns used.