Microsoft Research has released CHESS, an automated tool for finding errors -- data-races, deadlocks, hangs, and data-corruption induced access violations -- in multithreaded software by systematically exploring thread schedules. Once CHESS locates an error, it provides a fully repeatable execution of the program leading to the error, thus greatly aiding the debugging process. In addition, CHESS provides a novel notion of test coverage suitable for multithreaded programs. CHESS can use existing concurrent test cases and is therefore easy to deploy.
In their paper CHESS: A Systematic Testing Tool for Concurrent, researchers Madanlal Musuvathi, Shaz Qadeer, and Thomas Ball describe how CHESS uses model checking techniques to systematically generate all interleaving of a given scenario. CHESS scales to large concurrent programs and has found numerous previously unknown bugs in systems that had been stress tested for many months prior to being tested by CHESS. For instance, CHESS discovered a bug in PLINQ (short for "Parallel LINQ," which is part of the Parallel FX extensions for .NET, that was due to an incorrect use of LiteEvents, a concurrency primitive implemented in the library.
According to the researchers, CHESS has been integrated into the test frameworks of many code bases inside Microsoft and is being used by testers on a daily basis.
You can download a version of CHESS for Win32 here. A .NET version will be available in the near future.
