[Windows] misc fixes to enable Windows build

Summary:
## Context

Implement some small miscellaneous fixes to enable installing ExecuTorch on Windows. Currently, the optimized kernels cannot be built successfully but the installation works in custom ops are not built.

## Installation

```powershell
cd executorch

# Need to disable building custom kernels; there are a bunch of build errors
$env:EXECUTORCH_BUILD_KERNELS_CUSTOM_AOT=0
# For some reason need to explicitly enable tensor extension otherwise I see a linker error
$env:CMAKE_ARGS="-DEXECUTORCH_BUILD_EXTENSION_TENSOR=ON"

python install_executorch.py

# Unset if you want
Remove-Item Env:\EXECUTORCH_BUILD_KERNELS_CUSTOM_AOT
Remove-Item Env:\CMAKE_ARGS
```

## Testing

I am able to follow the [Getting Started tutorial](https://pytorch.org/executorch/stable/getting-started-setup.html#run-your-program) successfully.

## Outstanding Build Issues

Currently, with CMake I can successfully build with the following settings:

```
del -Recurse -Force cmake-out; `
  cmake . `
  -DCMAKE_INSTALL_PREFIX=cmake-out `
  -DPYTHON_EXECUTABLE=C:\\Users\\ssjia\\AppData\\Local\\miniconda3\\python.exe `
  -DCMAKE_PREFIX_PATH=C:\\Users\\ssjia\\AppData\\Local\\miniconda3\\Lib\\site-packages `
  -DCMAKE_BUILD_TYPE=Release `
  -DEXECUTORCH_BUILD_EXTENSION_TENSOR=ON `
  -DEXECUTORCH_BUILD_PYBIND=ON `
  -DEXECUTORCH_BUILD_XNNPACK=ON `
  -DEXECUTORCH_BUILD_KERNELS_QUANTIZED_AOT=ON `
  -DEXECUTORCH_BUILD_KERNELS_CUSTOM_AOT=ON `
  -DEXECUTORCH_BUILD_KERNELS_CUSTOM=OFF `
  -T ClangCL `
  -Bcmake-out; `
  cmake --build cmake-out -j64 --target install
```

If I switch `EXECUTORCH_BUILD_KERNELS_CUSTOM` to `ON`, then the build fails. The primary offenders appear to be the `gelu`. The implementation includes a header from ATen which is causing the majority of the errors.

Author: SS-JIA

backends/xnnpack/CMakeLists.txt

@@ -73,34 +73,42 @@ foreach(fbs_file ${_xnnpack_schema__srcs})
   )
 endforeach()
 
+if(WIN32)
+  set(MV_COMMAND powershell -Command "Move-Item -Path ${_xnnpack_flatbuffer__outputs} -Destination ${_xnnpack_schema__outputs}")
+else()
+  set(MV_COMMAND mv ${_xnnpack_flatbuffer__outputs} ${_xnnpack_schema__outputs})
+endif()
+
 # Generate the headers from the .fbs files.
 add_custom_command(
   OUTPUT ${_xnnpack_schema__outputs}
   COMMAND
     ${FLATC_EXECUTABLE} --cpp --cpp-std c++11 --scoped-enums -o
     "${_xnnpack_schema__include_dir}/executorch/backends/xnnpack/serialization"
     ${_xnnpack_schema__srcs}
-  COMMAND mv ${_xnnpack_flatbuffer__outputs} ${_xnnpack_schema__outputs}
+  COMMAND ${MV_COMMAND}
   WORKING_DIRECTORY ${EXECUTORCH_ROOT}
   COMMENT "Generating xnnpack_schema headers"
   VERBATIM
 )
 
+unset(MV_COMMAND)
+
 add_library(xnnpack_schema INTERFACE ${_xnnpack_schema__outputs})
 set_target_properties(xnnpack_schema PROPERTIES LINKER_LANGUAGE CXX)
 target_include_directories(
   xnnpack_schema INTERFACE ${_xnnpack_schema__include_dir}
                            ${EXECUTORCH_ROOT}/third-party/flatbuffers/include
 )
 
-set(xnnpack_third_party pthreadpool cpuinfo)
+set(xnnpack_third_party pthreadpool extension_threadpool cpuinfo)
 
 include(cmake/Dependencies.cmake)
 
 list(TRANSFORM _xnnpack_backend__srcs PREPEND "${EXECUTORCH_ROOT}/")
 add_library(xnnpack_backend STATIC ${_xnnpack_backend__srcs})
 target_link_libraries(
-  xnnpack_backend PRIVATE ${xnnpack_third_party} executorch_core xnnpack_schema
+  xnnpack_backend PUBLIC ${xnnpack_third_party} executorch_core xnnpack_schema
 )
 
 target_include_directories(

extension/llm/custom_ops/CMakeLists.txt

@@ -49,7 +49,7 @@ if(CMAKE_SYSTEM_PROCESSOR MATCHES "^(aarch64|arm64|armv7)$")
   list(APPEND _custom_ops__srcs
        "extension/llm/custom_ops/spinquant/third-party/FFHT/fht_neon.c"
   )
-elseif(CMAKE_SYSTEM_PROCESSOR STREQUAL "x86_64")
+elseif(CMAKE_SYSTEM_PROCESSOR MATCHES "^(x86_64|AMD64)")
   list(APPEND _custom_ops__srcs
        "extension/llm/custom_ops/spinquant/third-party/FFHT/fht_avx.c"
   )

install_executorch.bat

@@ -7,14 +7,8 @@ rem This batch file provides a basic functionality similar to the bash script.
 
 cd /d "%~dp0"
 
-rem Find the names of the python tools to use (replace with your actual python installation)
-if "%PYTHON_EXECUTABLE%"=="" (
-  if "%CONDA_DEFAULT_ENV%"=="" OR "%CONDA_DEFAULT_ENV%"=="base" OR NOT EXIST "python" (
-    set PYTHON_EXECUTABLE=python3
-  ) else (
-    set PYTHON_EXECUTABLE=python
-  )
-)
+rem Under windows, it's always python
+set PYTHON_EXECUTABLE=python
 
 "%PYTHON_EXECUTABLE%" install_executorch.py %*
 

setup.py

@@ -345,22 +345,30 @@ def inplace_dir(self, installer: "InstallerBuildExt") -> Path:
 class BuiltExtension(_BaseExtension):
     """An extension that installs a python extension that was built by cmake."""
 
-    def __init__(self, src: str, modpath: str):
+    def __init__(self, src: str, modpath: str, src_dir: Optional[str] = None):
         """Initializes a BuiltExtension.
 
         Args:
-            src: The path to the file to install (typically a shared library),
-                relative to the cmake-out directory. May be an fnmatch-style
-                glob that matches exactly one file. If the path ends in `.so`,
-                this class will also look for similarly-named `.dylib` files.
+            src_dir: The directory of the file to install, relative to the cmake-out
+                directory. A placeholder %BUILD_TYPE% will be replaced with the build
+                type for multi-config generators (like Visual Studio) where the build
+                output is in a subdirectory named after the build type. For single-
+                config generators (like Makefile Generators or Ninja), this placeholder
+                will be removed.
+            src_name: The name of the file to install. If the path ends in `.so`,
             modpath: The dotted path of the python module that maps to the
                 extension.
         """
         assert (
             "/" not in modpath
         ), f"modpath must be a dotted python module path: saw '{modpath}'"
+        full_src = src
+        if src_dir is None and platform.system() == "Windows":
+            src_dir = "%BUILD_TYPE%/"
+        if src_dir is not None:
+            full_src = os.path.join(src_dir, src)
         # This is a real extension, so use the modpath as the name.
-        super().__init__(src=f"%CMAKE_CACHE_DIR%/{src}", dst=modpath, name=modpath)
+        super().__init__(src=f"%CMAKE_CACHE_DIR%/{full_src}", dst=modpath, name=modpath)
 
     def src_path(self, installer: "InstallerBuildExt") -> Path:
         """Returns the path to the source file, resolving globs.