A novel SAT all-solutions solver for efficient preimage computation

In this paper, we present a novel all-solutions preimage SAT solver, SOLALL, with the following features: (1) a new success-driven learning algorithm employing smaller cut sets; (2) a marked CNF database non-trivially combining success/conflict-driven learning; (3) quantified-jump-back dynamically quantifying primary input variables from the preimage; (4) improved free BDD built on the fly, saving memory and avoiding inclusion of PI variables; finally, (5) a practical method of storing all solutions into a canonical OBDD format. Experimental results demonstrated the efficiency of the proposed approach for very large sequential circuits.