Fix builds for ONNX Runtime

Author: saudet

onnxruntime/cppbuild.sh

@@ -16,9 +16,9 @@ export CUDNN_HOME="/usr/local/cuda"
 export MAKEFLAGS="-j $MAKEJ"
 export PYTHON_BIN_PATH=$(which python3)
 
-export GPU_FLAGS=
+export CUDAFLAGS=
 if [[ "$EXTENSION" == *gpu ]]; then
-    GPU_FLAGS="--use_cuda"
+    CUDAFLAGS="--use_cuda"
 fi
 
 ONNXRUNTIME=1.17.3
@@ -75,18 +75,24 @@ sedinplace '/WX/d' cmake/CMakeLists.txt
 
 # allow cross compilation for linux-arm64
 sedinplace 's/if args.arm or args.arm64:/if False:/g' tools/ci_build/build.py
+sedinplace 's/is_linux() and platform.machine() == "x86_64"/False/g' tools/ci_build/build.py
+sedinplace 's/onnxruntime_target_platform STREQUAL "aarch64"/FALSE/g' cmake/CMakeLists.txt
+sedinplace 's/if (NOT APPLE)/if (FALSE)/g' cmake/onnxruntime_mlas.cmake
+sedinplace 's/!defined(__APPLE__)/0/g' onnxruntime/core/mlas/inc/mlas.h
+sedinplace 's/defined(__aarch64__) && defined(__linux__)/0/g' `find . -name *.cpp -o -name *.cc -o -name *.h`
+sedinplace 's/MLAS_CPUIDINFO::GetCPUIDInfo().HasArmNeon_I8MM()/false/g' onnxruntime/core/mlas/lib/platform.cpp
 
 # work around toolchain issues on Mac and Windows
 patch -p1 < ../../../onnxruntime.patch
 #patch -p1 < ../../../onnxruntime-windows.patch # https://github.com/microsoft/onnxruntime/pull/7883
 sedinplace '/--Werror/d' cmake/CMakeLists.txt
 sedinplace "s/return 'ON'/return 'OFF'/g" tools/ci_build/build.py
 sedinplace "s/Visual Studio 1. 20../Ninja/g" tools/ci_build/build.py
-sedinplace 's/Darwin|iOS/iOS/g' cmake/onnxruntime_providers.cmake
-sedinplace 's/-fvisibility=hidden//g' cmake/CMakeLists.txt cmake/adjust_global_compile_flags.cmake cmake/onnxruntime_providers.cmake
-sedinplace 's:/Yucuda_pch.h /FIcuda_pch.h::g' cmake/onnxruntime_providers.cmake
-sedinplace 's/${PROJECT_SOURCE_DIR}\/external\/cub//g' cmake/onnxruntime_providers.cmake
-sedinplace 's/ONNXRUNTIME_PROVIDERS_SHARED)/ONNXRUNTIME_PROVIDERS_SHARED onnxruntime_providers_shared)/g' cmake/onnxruntime_providers.cmake
+sedinplace 's/Darwin|iOS/iOS/g' cmake/onnxruntime_providers_cpu.cmake cmake/onnxruntime_providers.cmake
+sedinplace 's/-fvisibility=hidden//g' cmake/CMakeLists.txt cmake/adjust_global_compile_flags.cmake cmake/onnxruntime_providers_cpu.cmake cmake/onnxruntime_providers.cmake
+sedinplace 's:/Yucuda_pch.h /FIcuda_pch.h::g' cmake/onnxruntime_providers_cuda.cmake cmake/onnxruntime_providers.cmake
+sedinplace 's/${PROJECT_SOURCE_DIR}\/external\/cub//g' cmake/onnxruntime_providers_cuda.cmake cmake/onnxruntime_providers.cmake
+sedinplace 's/ONNXRUNTIME_PROVIDERS_SHARED)/ONNXRUNTIME_PROVIDERS_SHARED onnxruntime_providers_shared)/g' cmake/onnxruntime_providers_cpu.cmake cmake/onnxruntime_providers.cmake
 sedinplace 's/DNNL_TAG v.*)/DNNL_TAG v3.4.1)/g' cmake/external/dnnl.cmake
 sedinplace 's/DNNL_SHARED_LIB libdnnl.1.dylib/DNNL_SHARED_LIB libdnnl.2.dylib/g' cmake/external/dnnl.cmake
 sedinplace 's/DNNL_SHARED_LIB libdnnl.so.1/DNNL_SHARED_LIB libdnnl.so.2/g' cmake/external/dnnl.cmake
@@ -160,7 +166,7 @@ sedinplace 's/UTFChars(javaNameStrings/UTFChars((jstring)javaNameStrings/g' java
 sedinplace 's/initializers = allocarray/initializers = (const OrtValue**)allocarray/g' java/src/main/native/ai_onnxruntime_OrtSession_SessionOptions.cpp
 
 which ctest3 &> /dev/null && CTEST="ctest3" || CTEST="ctest"
-"$PYTHON_BIN_PATH" tools/ci_build/build.py --build_dir ../build --config Release --cmake_path "$CMAKE" --ctest_path "$CTEST" --build_shared_lib $ARCH_FLAGS $DNNL_FLAGS $OPENMP_FLAGS $GPU_FLAGS
+"$PYTHON_BIN_PATH" tools/ci_build/build.py --build_dir ../build --config Release --cmake_path "$CMAKE" --ctest_path "$CTEST" --build_shared_lib $ARCH_FLAGS $DNNL_FLAGS $OPENMP_FLAGS $CUDAFLAGS
 
 # install headers and libraries in standard directories
 cp -r include/* ../include