Contradiction analysis for constraint-based random simulation

Constraint-based random simulation is state-of-the-art in verification of multi-million gate industrial designs. This method is based on stimulus generation by constraint solving. The resulting stimuli will particularly cover corner case test scenarios which are usually hard to identify manually by the verification engineer. Consequently, constraint-based random simulation will catch corner case bugs that would remain undetected otherwise. Therefore, the quality of design verification is increased significantly. However, in the process of constraint specification for a specific test scenario, the verification engineer is faced with the problem of over-constraining, i.e. the overall constraint specified for a test scenario has no solution. In this case the root cause of the contradiction has to be identified and resolved. Given the complexity of constraints used to describe test scenarios, this can be a very time-consuming process. In this paper we propose a fully automated contradiction analysis method. Our method determines all "non relevant" constraints and computes all reasons that lead to the over-constraining. Thus, we pinpoint the verification engineer to exactly the sets of constraints that have to be considered to resolve the over-constraining. Experiments have been conducted in a real-life SystemC-based verification environment at AMD Dresden Design Center. They demonstrate a significant reduction of the constraint contradiction debug time.