| 摘要 |
A computer increases coverage in simulation of a design of a circuit by processing goals for coverage differently depending on whether or not the goals are on input signals of the circuit. Specifically, goals on input signals are used to automatically formulate constraints ("directly-derived constraints") on values of input signals in test vectors. Goals on non-input signals (e.g. internal/output signals) are used with correlations to automatically formulate more additional constraints ("correlation-derived constraints"), by use of goals on non-input signals. The correlations indicate which non-input signals are associated with which input signals. The correlations are received from, for example, a human designer of the circuit. Depending on the embodiment, one or more of the automatically derived constraints are used with human-supplied constraints, to generate test vectors e.g. using a constraints solver, such as a satisfiability (SAT) engine. The test vectors are supplied to a simulator for functional verification.
|
