Proving sequential consistency by model checking

Sequential consistency is a multiprocessor memory model of both practical and theoretical importance. Unfortunately, the general problem of verifying that a finitestate protocol implements sequential consistency is undecidable, and in practice, validating that a real-world, finitestate protocol implements sequential consistency is very time-consuming and costly. In this work, we show that for memory protocols that occur in practice, a small amount of manual effort can reduce the problem of verifying sequential consistency into a verification task that can be discharged automatically via model checking. Furthermore, we present experimental results on a substantial, directorybased cache coherence protocol, which demonstrate the practicality of our approach.