* Add quick gelu and fused bias add kernel

* fix annotation

* remove useless code

* add fast gelu option and set it in multi transformer op

* add flag to restrict if use fast gelu approximate

* fix flags conflict

* fix use tanh function instead

* add cudart version limit

* use phi fast tanh func

* fix comment

* remove cudaDeviceContext

* remove more template

* fix rocm compile

* remove alias name CUDADeviceContext

* fix compile

* fix tests

* revert changes

* clean distribution_helper, index_impl, aligned_vector code in fluid

* fix conflicts

Fused elementwise_add, dropout, elementwise_add and layer_norm into one operator, only support Forward. 
No Python API changed.