Localizing features of ESL models for design understanding

The increasing variety of functionality in embedded systems leads to more and more complex designs. Even abstraction techniques as applied in ESL design solve this problem only to a certain extend. In this paper we present an approach to improve design understanding of ESL models. Our approach localizes features by comparing simulations on ESL models. Code coverage techniques are used to highlight source code implementing a certain feature. This significantly helps the different team members since for major design tasks like e.g. refinement, partitioning or optimization it is required to know where a certain functionality has been implemented. We demonstrate the advantages of our approach for a complex image processing system.