Implemeting heat equation benchmark.

- the timer is now constructed as zero-initialized and stopped
- reset() clears the timer, but does not start it again
- getTime() no longer starts the timer if it is in a stopped state

Refactoring the Dirichlet and the Neumann boundary conditions.

Adding simple heat equation benchmark for CUDA.

Moved tnlObject_impl.h back to tnlObject.cpp since function definitions
in header files cause duplicates when building libtnl. The constructor
of tnlObject is __cuda_callable__ so it's empty definition was moved
directly to tnlObject.h to make it visible to nvcc.

Problem with marking tnlObject's virtual destructor as __cuda_callable__
is yet to be solved...