Debugging Functional Coverage Models: Get the Most of Out of Your Cover Crosses

Applying hole analysis on each cover cross independently can lead to misleading results and is sometimes prohibitive due to the sheer number of crosses. Additionally, we introduce a metric, hole effect, that is proportional to the coverage gains that would result upon resolving the highlighted hole. We evaluate our approach on a real processor’s data processing unit to validate its applicability and usefulness for debugging complex functional coverage models.