Adaptive Data Model for Efficient Constraint Handling in AMS IC Design

Further automation of analog and mixed-signal integrated circuit design requires the consideration of a growing number of design constraints. The processing of these constraints needs specific design information and depends on their target parameters and the type of mathematical requirement. Both aspects allow the creation of numerous different constraint types with many demands on the available design data. This paper presents a data model for AMS IC designs that is based on a static model for common design data. In order to store all information necessary for utilized constraint types, this model dynamically adapts and extends its internal structure. Thereby, our data model does not limit the set of supported constraint types and allows uniform access to design and constraint data for constraint-driven algorithms. We preserve the graph nature of our data model by using a graph database for its implementation. Thus, constraint engineering becomes much faster compared to conventional static data models.