Implementation and Experiments of a Distributed SMT Solving Environment

SMT-based Bounded Model Checking (BMC) consists of two primary tasks: (1) encoding a bounded model checking problem into a propositional formula that represents the problem, and (2) using a SMT solver to solve the formula, that is, finding a set of variable assignments that makes the formula true. Solving the formula (namely, SMT solving) involves computation-intensive processes and is thus time-consuming. Furthermore, as the model-checking bound increases, the encoded formulas become larger in size and harder to solve. The target of this paper is to accelerate SMT solving through distributed computation so as to enhance the performance of SMT-based BMC. To this end, we implement a multi-CPUs & multi-cores based distributed computation environment, to which different SMT solvers can be integrated. In this paper, we describe the implementation details and report our preliminary experimental results. Keywords-acceleration, distributed smt solving, mpi, openmp