created element_operations and matrix_reductions files for farther refactoring

refactored transpose_gpu, made it non template function.

module reorganization: added folder with pure device functions, cuda_shared.hpp renamed to internal_shared.hpp

added gpu exp, log, magnitude, based on NPP.
updated setTo with new NPP functions.
minor fix in tests and comments.

added comments in gpu.hpp

avoid extra gpu memory allocation in BP and CSBP

  fixed some bugs in init_data_cost on first level (added non-reduction version for first level)
  optimized compute_data_cost like init_data_cost (used reduction scheme)
  avoid temp matrix

some refactoring of StereoBeliefPropagation.