* cublas emitter

* clang format fixes

* Initial comment incorporation from Chris

* Chris's If-else change comment incorporation

* incorporating Bob's comments phase 1

*  Remove unnecessary headers in cublas emitter hpp & cpp (as per Bob's comments)

* clang format on previous commit

* incorporate fenglei's refactoring comment

* incorporating comments

* Incorporate Chris's final comment

* All comments resolved

* Resolve Geoff's comments

* Change cache_primitive to register_primitive