Refactoring all of the shared cuda helper methods to the shared common file

[Chris Dryden]

Trying to come up with some examples of beginner friendly issues that would be helpful to the development effort:

Lots of the CUDA kernels are using shared methods such as the float4 indexing or the linear multiplication helpers, identify all of the shared methods in the common file and move them there to make sure that all of the CUDA kernels are using the same shared function in the final training code and to make it easier for development in the future.