bv2epr: A Tool for Polynomially Translating Quantifier-Free Bit-Vector Formulas into EPR

Abstract

Abstract. Bit-precise reasoning is essential in many applications of Satisfiability Modulo Theories (SMT). In recent years, efficient approaches for solving fixed-size bit-vector formulas have been developed. Most of these approaches rely on bit-blasting. In [1], we argued that bit-blasting is not polynomial in general, and then showed that solving quantifier-free bit-vector formulas (QF BV) is NExpTime-complete. In this paper, we present a tool based on a new polynomial translation from QF BV into Effectively Propositional Logic (EPR). This allows us to solve QF BV problems using EPR solvers and avoids the exponential growth that comes with bit-blasting. Additionally, our tool allows us to easily generate new challenging benchmarks for EPR solvers.

EPR-based k-induction with Counterexample Guided Abstraction Refinement

“…It can be shown (following [25]) that F is satisfiable if and only if (12) together with the domain axiom for P f :…”

“…Experimental results reported in these papers regarding EPR-based BMC versus SAT and SMT-based approaches showed the promising potential of EPR-based BMC on industrial hardware model checking and equivalence checking problem instances with memories, which were the driving industrial examples for developing the EPR-based word-level BMC. The EPR fragment is NEXPTIME complete, and recent research has also shown that it can be used to encode bit-vectors exponentially more succinctly [12,13].…”

Instantiations, Zippers and EPR Interpolation

“…EPR offers a degree of succinctness over propositional logic that can be a significant advantage. EPR is often also hidden in plain sight [11]: bit-vector logic turns out to be much more succinct than propositional logic and can be translated polynomially back and forth into EPR. Current applications in verification uses EPR as a language for describing assertions and systems.…”