Merge CUDAAccel into FullMonteSW
Merge the CUDA accelerator into FullMonteSW as a submodule and create a Kernel (a FullMonteSW Kernel) to interface with the accelerator (akin to TetraMCKernel). When doing so ensure that the compilation of the CUDA accelerator can be turned on/off in the CMake file so the developer does not HAVE to build the CUDA stuff (use a CMake Boolean var).