Fix merge conflicts

.vscode/settings.json

@@ -13,6 +13,7 @@
         "editor.codeActionsOnSave": {
             "source.organizeImports": true
         },
+        "editor.defaultFormatter": "ms-python.black-formatter"
     },
     // Enable Python linting and Pylance type checking
     "python.analysis.typeCheckingMode": "basic",

cmake/deps.txt

@@ -54,4 +54,4 @@ tensorboard;https://github.com/tensorflow/tensorboard/archive/373eb09e4c5d2b3cc2
 cutlass;https://github.com/NVIDIA/cutlass/archive/refs/tags/v3.1.0.zip;757f90a795034a89d4f48a79d1f009f7a04c8dee
 utf8_range;https://github.com/protocolbuffers/utf8_range/archive/72c943dea2b9240cd09efde15191e144bc7c7d38.zip;9925739c9debc0efa2adcb194d371a35b6a03156
 extensions;https://github.com/microsoft/onnxruntime-extensions/archive/94142d8391c9791ec71c38336436319a2d4ac7a0.zip;4365ac5140338b4cb75a39944a4be276e3829b3c
-composable_kernel;https://github.com/ROCmSoftwarePlatform/composable_kernel/archive/a4f72a314a85732ed67d5aa8d1088d207a7e0e61.zip;f57357ab6d300e207a632d034ebc8aa036a090d9
+composable_kernel;https://github.com/ROCmSoftwarePlatform/composable_kernel/archive/5356c4a943a35e74d7cdc69486afcb8703b9a59a.zip;522382c2af437e09124287e5879ab64af5b2e299

cmake/onnxruntime.cmake

@@ -282,11 +282,7 @@ endif()
 
 # Assemble the Apple static framework (iOS and macOS)
 if(onnxruntime_BUILD_APPLE_FRAMEWORK)
-  if(${CMAKE_SYSTEM_NAME} STREQUAL "iOS")
-    set(STATIC_FRAMEWORK_OUTPUT_DIR ${CMAKE_CURRENT_BINARY_DIR}/${CMAKE_BUILD_TYPE}-${CMAKE_OSX_SYSROOT})
-  else() # macOS
-    set(STATIC_FRAMEWORK_OUTPUT_DIR ${CMAKE_CURRENT_BINARY_DIR})
-  endif()
+  set(STATIC_FRAMEWORK_OUTPUT_DIR ${CMAKE_CURRENT_BINARY_DIR}/${CMAKE_BUILD_TYPE}-${CMAKE_OSX_SYSROOT})
 
   # Setup the various directories required. Remove any existing ones so we start with a clean directory.
   set(STATIC_LIB_DIR ${CMAKE_CURRENT_BINARY_DIR}/static_libraries)

cmake/onnxruntime_optimizer.cmake

@@ -86,6 +86,8 @@ if (onnxruntime_ENABLE_TRAINING)
     "${ORTTRAINING_SOURCE_DIR}/core/optimizer/*.cc"
     "${ORTTRAINING_SOURCE_DIR}/core/optimizer/compute_optimizer/*.h"
     "${ORTTRAINING_SOURCE_DIR}/core/optimizer/compute_optimizer/*.cc"
+    "${ORTTRAINING_SOURCE_DIR}/core/optimizer/memory_optimizer/*.h"
+    "${ORTTRAINING_SOURCE_DIR}/core/optimizer/memory_optimizer/*.cc"
   )
 endif()
 

cmake/onnxruntime_providers_cuda.cmake

@@ -172,10 +172,8 @@
       target_link_libraries(${target} PRIVATE cuda)
     endif()
 
-    if (onnxruntime_USE_FLASH_ATTENTION OR onnxruntime_USE_MEMORY_EFFICIENT_ATTENTION)
-      include(cutlass)
-      target_include_directories(${target} PRIVATE ${cutlass_SOURCE_DIR}/include ${cutlass_SOURCE_DIR}/examples)
-    endif()
+    include(cutlass)
+    target_include_directories(${target} PRIVATE ${cutlass_SOURCE_DIR}/include ${cutlass_SOURCE_DIR}/examples)
 
     target_include_directories(${target} PRIVATE ${ONNXRUNTIME_ROOT} ${CMAKE_CURRENT_BINARY_DIR}  ${eigen_INCLUDE_DIRS} ${TVM_INCLUDES} PUBLIC ${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES})
     # ${CMAKE_CURRENT_BINARY_DIR} is so that #include "onnxruntime_config.h" inside tensor_shape.h is found

cmake/onnxruntime_unittests.cmake

@@ -783,7 +783,7 @@ if (onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS)
   onnxruntime_add_shared_library_module(onnxruntime_providers_cuda_ut ${onnxruntime_test_providers_cuda_ut_src} $<TARGET_OBJECTS:onnxruntime_providers_cuda_obj>)
   config_cuda_provider_shared_module(onnxruntime_providers_cuda_ut)
   onnxruntime_add_include_to_target(onnxruntime_providers_cuda_ut GTest::gtest GTest::gmock)
-  target_link_libraries(onnxruntime_providers_cuda_ut PRIVATE GTest::gtest GTest::gmock)
+  target_link_libraries(onnxruntime_providers_cuda_ut PRIVATE GTest::gtest GTest::gmock ${ONNXRUNTIME_MLAS_LIBS} onnxruntime_common)
   list(APPEND onnxruntime_test_providers_dependencies onnxruntime_providers_cuda_ut)
 endif()
 

cmake/patches/composable_kernel/Fix_Clang_Build.patch

@@ -1,5 +1,5 @@
 diff --git a/CMakeLists.txt b/CMakeLists.txt
-index b09da41a8..fca2bdf69 100644
+index 04674124c..12e8b8b00 100644
 --- a/CMakeLists.txt
 +++ b/CMakeLists.txt
 @@ -19,7 +19,7 @@ endif()
@@ -48,7 +48,18 @@ index b09da41a8..fca2bdf69 100644
 
  ## tidy
  include(EnableCompilerWarnings)
-@@ -489,11 +466,3 @@ rocm_install(FILES
+@@ -376,7 +353,9 @@ if(BUILD_DEV)
+     add_compile_options(-Werror -Weverything)
+ endif()
+ #add flags to reduce the size of binaries
+-add_compile_options(-Oz -flto=thin)
++# -flto requires ORT to use a linker that support LTO and -flto flag shoud be passed to linker together.
++# add_compile_options(-Oz -flto=thin)
++add_compile_options(-Oz)
+ message("CMAKE_CXX_FLAGS: ${CMAKE_CXX_FLAGS}")
+
+ add_custom_target(check COMMAND ${CMAKE_CTEST_COMMAND} --output-on-failure -C ${CMAKE_CFG_INTDIR})
+@@ -482,11 +461,3 @@ rocm_install(FILES
 
  set(CPACK_RESOURCE_FILE_LICENSE "${CMAKE_CURRENT_SOURCE_DIR}/LICENSE")
  set(CPACK_RPM_PACKAGE_LICENSE "MIT")
@@ -61,7 +72,7 @@ index b09da41a8..fca2bdf69 100644
 -    HEADER_ONLY
 -)
 diff --git a/library/src/tensor_operation_instance/gpu/CMakeLists.txt b/library/src/tensor_operation_instance/gpu/CMakeLists.txt
-index a0478c9f0..1e7782cd4 100644
+index 9cb5d0e9a..141a46f3d 100644
 --- a/library/src/tensor_operation_instance/gpu/CMakeLists.txt
 +++ b/library/src/tensor_operation_instance/gpu/CMakeLists.txt
 @@ -44,8 +44,14 @@ function(add_instance_library INSTANCE_NAME)

cmake/winml.cmake

@@ -451,6 +451,8 @@ onnxruntime_add_static_library(winml_lib_api
   ${winml_lib_api_dir}/impl/TensorKindFrom.h
   ${winml_lib_api_dir}/impl/TensorMemoryBufferReference.h
   ${winml_lib_api_dir}/NumericData.cpp
+  ${winml_lib_api_dir}/HardwareCoreEnumerator.cpp
+  ${winml_lib_api_dir}/HardwareCoreEnumerator.h
   ${winml_lib_api_dir}/ImageFeatureDescriptor.cpp
   ${winml_lib_api_dir}/ImageFeatureDescriptor.h
   ${winml_lib_api_dir}/ImageFeatureValue.cpp