Fix bug of layernorm ckProfiler and refine code (#448)

* Fix bug of profiler for layernorm

* 1. Rename layernorm into normalization
2. Decouple softmax from normalization

* clang-format

test/CMakeLists.txt

@@ -6,11 +6,10 @@ include(googletest)
 
 add_custom_target(tests)
 
-
 function(add_test_executable TEST_NAME)
     message("adding test ${TEST_NAME}")
     add_executable(${TEST_NAME} ${ARGN})
-    add_test(NAME ${TEST_NAME} COMMAND $<TARGET_FILE:${TEST_NAME}> )
+    add_test(NAME ${TEST_NAME} COMMAND $<TARGET_FILE:${TEST_NAME}>)
     add_dependencies(tests ${TEST_NAME})
     add_dependencies(check ${TEST_NAME})
     rocm_install(TARGETS ${TEST_NAME} COMPONENT tests)
@@ -23,6 +22,7 @@ function(add_gtest_executable TEST_NAME)
     add_executable(${TEST_NAME} ${ARGN})
     add_dependencies(tests ${TEST_NAME})
     add_dependencies(check ${TEST_NAME})
+
     # suppress gtest warnings
     target_compile_options(${TEST_NAME} PRIVATE -Wno-global-constructors -Wno-undef)
     target_link_libraries(${TEST_NAME} PRIVATE gtest_main)
@@ -30,7 +30,6 @@ function(add_gtest_executable TEST_NAME)
     rocm_install(TARGETS ${TEST_NAME} COMPONENT tests)
 endfunction(add_gtest_executable TEST_NAME)
 
-
 add_subdirectory(magic_number_division)
 add_subdirectory(space_filling_curve)
 add_subdirectory(conv_util)
@@ -51,5 +50,5 @@ add_subdirectory(convnd_bwd_data)
 add_subdirectory(grouped_convnd_fwd)
 add_subdirectory(block_to_ctile_map)
 add_subdirectory(softmax)
-add_subdirectory(layernorm)
+add_subdirectory(normalization)
 add_subdirectory(data_type)

test/layernorm/test_layernorm2d_util.hpp → test/normalization/test_layernorm2d_util.hpp

@@ -9,7 +9,7 @@
 
 #include "ck/ck.hpp"
 #include "ck/utility/number.hpp"
-#include "ck/tensor_operation/gpu/device/device_layernorm_impl.hpp"
+#include "ck/tensor_operation/gpu/device/device_normalization_impl.hpp"
 
 #include "ck/library/utility/check_err.hpp"
 #include "ck/library/utility/host_tensor.hpp"
@@ -65,7 +65,7 @@ class TestLayernorm2d : public ::testing::Test
                                                                          Rank,
                                                                          NumReduceDim>;
 
-    using DeviceInstance = tensor_operation::device::DeviceLayernormImpl<XDataType,
+    using DeviceInstance = tensor_operation::device::DeviceNormalizationImpl<XDataType,
                                                                              GammaDataType,
                                                                              BetaDataType,
                                                                              AccDataType,