Anatomy and Empirical Evaluation of Modern SAT Solvers

The Boolean Satisfiability (SAT) decision problem can be deservedly declared a success story of computer science. Although SAT was the first problem to be proved NP-complete, the last decade and a half have seen dramatic improvements in the performance of SAT solvers on many practical problem instances. These performance improvements enabled a wide range of real-world applications, several of which have key industrial significance. This article surveys the organization of modern conflict-driven clause learning (CDCL) SAT solvers, focusing on the principal techniques that have contributed to this impressive performance. The article also empirically evaluates these techniques on a comprehensive suite of problem instances taken from a range of representative applications, allowing for a better understanding of their relative contribution.