Testability-Driven Fault Sampling for Deterministic Test Coverage Estimation of Large Designs

The continuously increasing complexity and size of the modern high performance design often introduces the testability barriers which either restrict the desirable test quality, or require extended test generation time. To detect such testability barriers and accurately estimate the test coverage quickly is important to minimize the risk of test quality issue or any later stage design change. The paper proposes a novel technique to efficiently estimate the deterministic test coverage for large designs. The test coverage estimation with less than 0.5% error can be achieved with more than 30X run time reduction compared to the test generation run time of the entire fault population.