Universal Callable ad Runnable Wrappers and thread safety test

[PavelSosin-320]

To run Java cuda applications as regular ( not main ) class Collable and Runable wrappers are the prerequisit. Running Collable and Runnable classes requirs thread safety test. Both are almost done. I will assign these issues to myself

[jcuda]

I'm not sure what this is aiming at, but if the goal is to show that it's easy to crash a CUDA/JCuda application without synchronization: I know that, and I wrote a somewhat extensive answer at jcuda/jcuda#25 (comment)

In any case: Please do the same test in C++, so that we have a comparison to plain CUDA.

[jcuda]

If there are any specific results (issues) or tasks that can be derived from that, just let me know.