Equivalence checking on ESL utilizing a priori knowledge

Abstract

We propose EASY, an algorithm for functional equivalence checking of ESL descriptions written in a high-level programming language like C++. Given two ESL descriptions, in a PDR-like fashion EASY systematically refines a candidate invariant to characterize the reachable states of a miter of the descriptions until either the invariant becomes inductive or a counterexample has been found. The algorithm is flexible and allows to incorporate a priori knowledge about the design to speed up the verification process. We provide an implementation of EASY on top of a standard model checker and show in two case studies that EASY outperforms other state-of-the-art equivalence checking tools.