[C] Normalization Refactor + Adding CUDNN backend (NVIDIA#1315)

* cuDNN normalization integration
* TE Norm refactor
* TE Norm APIs changes.

---------

Signed-off-by: Phuong Nguyen <[email protected]>
Co-authored-by: Tim Moon <[email protected]>

Author: phu0ngng

tests/cpp/operator/CMakeLists.txt

@@ -10,8 +10,7 @@ add_executable(test_operator
                test_cast_transpose_dbias_dgelu.cu
                test_cast_transpose_dgeglu.cu
                test_act.cu
-               test_layernorm.cu
-               test_rmsnorm.cu
+               test_normalization.cu
                test_multi_cast_transpose.cu
                test_multi_padding.cu
                test_causal_softmax.cu