2004 | OriginalPaper | Buchkapitel
Variable Reuse for Efficient Image Computation
verfasst von : Zijiang Yang, Rajeev Alur
Erschienen in: Formal Methods in Computer-Aided Design
Verlag: Springer Berlin Heidelberg
Enthalten in: Professional Book Archive
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Image computation, that is, computing the set of states reachable from a given set in one step, is a crucial component in typical tools for BDD-based symbolic reachability analysis. It has been shown that the size of the intermediate BDDs during image computation can be dramatically reduced via conjunctive partitioning of the transition relation and ordering the conjuncts for facilitating early quantification. In this paper, we propose to enhance the effectiveness of these techniques by reusing the quantified variables. Given an ordered set of conjuncts, if the last conjunct that uses a variable u appears before the first conjunct that uses another variable v, then v can be renamed to u, assuming u will be quantified immediately after its last use. In general, multiple variables can share the same identifier so the BDD nodes that are inactive but not garbage collected may be activated. We give a polynomial-time algorithm for generating the optimum number of variables that are required for image computation and show how to modify the image computation accounting for variable reuse. The savings for image computation are demonstrated on ISCAS’89 and Texas’97 benchmark models.