Parallel programming of clusters is an effective way to improve computing performance and Intel hardware is the most used on the list of Top 500 supercomputers in the world. When implementing parallel computational algorithms that use several threads or processes, new types of errors related to simultaneous use of the shared resources emerge. This significantly complicates implementation and debugging of parallel programs. This presentation considers questions related to the correctness analysis of MPI programs and presents a special automated analysis tool developed at Intel, the Intel Message Checker, which is the first commercial offering of its kind. The new Intel® Message Checker is an MPI correctness tool that improves programmer productivity and increases confidence in the correctness of MPI programs. It features a graphical user interface to a comprehensive analysis engine that automatically detects many MPI errors such as race conditions, deadlocks and potential deadlocks. This helps programmers to find issues earlier in the debugging process and in less time.