T L  C S C a Sat S P

In not much more than a decade the practice of implementing programs for testing satisfiability of propositional formulas has progressed from being practically a curiosity to an active and competitive endeavor. Impressive performance improvements have made SAT solvers useful in some industrial applications, most notably in formal verification of hardware and software systems as the engine for bounded model checking (BMC) tools. Here we describe the main design and implementation features of the family of SAT solvers which represent the state of the art for solving of industrial instances, and includes the solvers Chaff, BerkMin and siege. We attempt to give a self-contained presentation in a uniform style, and in doing so hope to encourage readers to ask interesting questions, as well to tackle some problems we mention. It should also be possible to produce a creditable solver with little more at hand than this paper and a penchant for highly efficient