fix build scripts (#74388) (#74429)

Author: XieYunshen

paddle/scripts/paddle_build.sh

@@ -285,7 +285,7 @@ function cmake_base() {
         -DWITH_ONNXRUNTIME=${WITH_ONNXRUNTIME:-OFF}
         -DWITH_NVCC_LAZY=${WITH_NVCC_LAZY:-ON}
         -DWITH_CUDNN_FRONTEND=${WITH_CUDNN_FRONTEND:-OFF}
-        -DFA_BUILD_WITH_CACHE=${FA_BUILD_WITH_CACHE:ON}
+        -DWITH_FA_BUILD_WITH_CACHE=${WITH_FA_BUILD_WITH_CACHE:ON}
     ========================================
 EOF
     # Disable UNITTEST_USE_VIRTUALENV in docker because
@@ -337,7 +337,7 @@ EOF
         -DWITH_UNITY_BUILD=${WITH_UNITY_BUILD:-OFF}  \
         -DWITH_ONNXRUNTIME=${WITH_ONNXRUNTIME:-OFF}  \
         -DWITH_NVCC_LAZY=${WITH_NVCC_LAZY:-ON} \
-        -DFA_BUILD_WITH_CACHE=${FA_BUILD_WITH_CACHE:ON} \
+        -DWITH_FA_BUILD_WITH_CACHE=${WITH_FA_BUILD_WITH_CACHE:ON} \
         -DWITH_CUDNN_FRONTEND=${WITH_CUDNN_FRONTEND:-OFF};build_error=$?
 
     if [ "$build_error" != 0 ];then
@@ -4304,13 +4304,13 @@ function run_setup(){
     echo "if you use setup.py to compile,please export envs as following in /paddle ..."
     cat << EOF
     ========================================
-    export CMAKE_BUILD_TYPE=${CMAKE_BUILD_TYPE:-Release} WITH_GPU=${WITH_GPU:-OFF} WITH_SHARED_PHI=${WITH_SHARED_PHI:-OFF} WITH_TENSORRT=${WITH_TENSORRT:-ON} WITH_OPENVINO=${WITH_OPENVINO:-OFF} WITH_ROCM=${WITH_ROCM:-OFF} WITH_CINN=${WITH_CINN:-OFF} WITH_DISTRIBUTE=${distributed_flag} WITH_MKL=${WITH_MKL:-ON} WITH_AVX=${WITH_AVX:-OFF} CUDA_ARCH_NAME=${CUDA_ARCH_NAME:-All} NEW_RELEASE_PYPI=${NEW_RELEASE_PYPI:-OFF} NEW_RELEASE_ALL=${NEW_RELEASE_ALL:-OFF} NEW_RELEASE_JIT=${NEW_RELEASE_JIT:-OFF} WITH_PYTHON=${WITH_PYTHON:-ON} CUDNN_ROOT=/usr/ WITH_TESTING=${WITH_TESTING:-ON} WITH_COVERAGE=${WITH_COVERAGE:-OFF} WITH_INCREMENTAL_COVERAGE=${WITH_INCREMENTAL_COVERAGE:-OFF} CMAKE_MODULE_PATH=/opt/rocm/hip/cmake CMAKE_EXPORT_COMPILE_COMMANDS=ON WITH_INFERENCE_API_TEST=${WITH_INFERENCE_API_TEST:-ON} INFERENCE_DEMO_INSTALL_DIR=${INFERENCE_DEMO_INSTALL_DIR} PY_VERSION=${PY_VERSION:-3.8} CMAKE_INSTALL_PREFIX=${INSTALL_PREFIX:-/paddle/build} WITH_GLOO=${gloo_flag} WITH_XPU=${WITH_XPU:-OFF} WITH_IPU=${WITH_IPU:-OFF} XPU_SDK_ROOT=${XPU_SDK_ROOT:-""} WITH_XPU_BKCL=${WITH_XPU_BKCL:-OFF} WITH_XPU_XHPC=${WITH_XPU_XHPC:-OFF} WITH_XPU_XFT=${WITH_XPU_XFT:-OFF} WITH_XPU_XRE5=${WITH_XPU_XRE5:-OFF} WITH_XPU_FFT=${WITH_XPU_FFT:-OFF} WITH_ARM=${WITH_ARM:-OFF} WITH_STRIP=${WITH_STRIP:-ON} ON_INFER=${ON_INFER:-OFF} CUDA_ARCH_BIN=${CUDA_ARCH_BIN} WITH_RECORD_BUILDTIME=${WITH_RECORD_BUILDTIME:-OFF} WITH_UNITY_BUILD=${WITH_UNITY_BUILD:-OFF} WITH_ONNXRUNTIME=${WITH_ONNXRUNTIME:-OFF} WITH_CUDNN_FRONTEND=${WITH_CUDNN_FRONTEND:-OFF} WITH_CPP_TEST=${WITH_CPP_TEST:-OFF} FA_BUILD_WITH_CACHE=${FA_BUILD_WITH_CACHE:-ON}
+    export CMAKE_BUILD_TYPE=${CMAKE_BUILD_TYPE:-Release} WITH_GPU=${WITH_GPU:-OFF} WITH_SHARED_PHI=${WITH_SHARED_PHI:-OFF} WITH_TENSORRT=${WITH_TENSORRT:-ON} WITH_OPENVINO=${WITH_OPENVINO:-OFF} WITH_ROCM=${WITH_ROCM:-OFF} WITH_CINN=${WITH_CINN:-OFF} WITH_DISTRIBUTE=${distributed_flag} WITH_MKL=${WITH_MKL:-ON} WITH_AVX=${WITH_AVX:-OFF} CUDA_ARCH_NAME=${CUDA_ARCH_NAME:-All} NEW_RELEASE_PYPI=${NEW_RELEASE_PYPI:-OFF} NEW_RELEASE_ALL=${NEW_RELEASE_ALL:-OFF} NEW_RELEASE_JIT=${NEW_RELEASE_JIT:-OFF} WITH_PYTHON=${WITH_PYTHON:-ON} CUDNN_ROOT=/usr/ WITH_TESTING=${WITH_TESTING:-ON} WITH_COVERAGE=${WITH_COVERAGE:-OFF} WITH_INCREMENTAL_COVERAGE=${WITH_INCREMENTAL_COVERAGE:-OFF} CMAKE_MODULE_PATH=/opt/rocm/hip/cmake CMAKE_EXPORT_COMPILE_COMMANDS=ON WITH_INFERENCE_API_TEST=${WITH_INFERENCE_API_TEST:-ON} INFERENCE_DEMO_INSTALL_DIR=${INFERENCE_DEMO_INSTALL_DIR} PY_VERSION=${PY_VERSION:-3.8} CMAKE_INSTALL_PREFIX=${INSTALL_PREFIX:-/paddle/build} WITH_GLOO=${gloo_flag} WITH_XPU=${WITH_XPU:-OFF} WITH_IPU=${WITH_IPU:-OFF} XPU_SDK_ROOT=${XPU_SDK_ROOT:-""} WITH_XPU_BKCL=${WITH_XPU_BKCL:-OFF} WITH_XPU_XHPC=${WITH_XPU_XHPC:-OFF} WITH_XPU_XFT=${WITH_XPU_XFT:-OFF} WITH_XPU_XRE5=${WITH_XPU_XRE5:-OFF} WITH_XPU_FFT=${WITH_XPU_FFT:-OFF} WITH_ARM=${WITH_ARM:-OFF} WITH_STRIP=${WITH_STRIP:-ON} ON_INFER=${ON_INFER:-OFF} CUDA_ARCH_BIN=${CUDA_ARCH_BIN} WITH_RECORD_BUILDTIME=${WITH_RECORD_BUILDTIME:-OFF} WITH_UNITY_BUILD=${WITH_UNITY_BUILD:-OFF} WITH_ONNXRUNTIME=${WITH_ONNXRUNTIME:-OFF} WITH_CUDNN_FRONTEND=${WITH_CUDNN_FRONTEND:-OFF} WITH_CPP_TEST=${WITH_CPP_TEST:-OFF} WITH_FA_BUILD_WITH_CACHE=${WITH_FA_BUILD_WITH_CACHE:-ON}
     ========================================
 EOF
     echo "if you use cmake to compile,please Configuring cmake in /paddle/build ..."
     cat <<EOF
     ========================================
-    cmake .. -DCMAKE_BUILD_TYPE=${CMAKE_BUILD_TYPE:-Release} -DWITH_GPU=${WITH_GPU:-OFF} -DWITH_SHARED_PHI=${WITH_SHARED_PHI:-OFF} -DWITH_TENSORRT=${WITH_TENSORRT:-ON} -DWITH_OPENVINO=${WITH_OPENVINO:-OFF} -DWITH_ROCM=${WITH_ROCM:-OFF} -DWITH_CINN=${WITH_CINN:-OFF} -DWITH_DISTRIBUTE=${distributed_flag} -DWITH_MKL=${WITH_MKL:-ON} -DWITH_AVX=${WITH_AVX:-OFF} -DCUDA_ARCH_NAME=${CUDA_ARCH_NAME:-All} -DNEW_RELEASE_PYPI=${NEW_RELEASE_PYPI:-OFF} -DNEW_RELEASE_ALL=${NEW_RELEASE_ALL:-OFF} -DNEW_RELEASE_JIT=${NEW_RELEASE_JIT:-OFF} -DWITH_PYTHON=${WITH_PYTHON:-ON} -DCUDNN_ROOT=/usr/ -DWITH_TESTING=${WITH_TESTING:-ON} -DWITH_COVERAGE=${WITH_COVERAGE:-OFF} -DWITH_INCREMENTAL_COVERAGE=${WITH_INCREMENTAL_COVERAGE:-OFF} -DCMAKE_MODULE_PATH=/opt/rocm/hip/cmake -DCMAKE_EXPORT_COMPILE_COMMANDS=ON -DWITH_INFERENCE_API_TEST=${WITH_INFERENCE_API_TEST:-ON} -DINFERENCE_DEMO_INSTALL_DIR=${INFERENCE_DEMO_INSTALL_DIR} -DPY_VERSION=${PY_VERSION:-3.8} -DCMAKE_INSTALL_PREFIX=${INSTALL_PREFIX:-/paddle/build} -DWITH_PSLIB=${pslib_flag} -DWITH_GLOO=${gloo_flag} -DWITH_XPU=${WITH_XPU:-OFF} -DWITH_IPU=${WITH_IPU:-OFF} -DXPU_SDK_ROOT=${XPU_SDK_ROOT:-""} -DWITH_XPU_BKCL=${WITH_XPU_BKCL:-OFF} -DWITH_XPU_XHPC=${WITH_XPU_XHPC:-OFF} -DWITH_XPU_XFT=${WITH_XPU_XFT:-OFF} -DWITH_XPU_XRE5=${WITH_XPU_XRE5:-OFF} -DWITH_XPU_FFT=${WITH_XPU_FFT:-OFF} -DWITH_ARM=${WITH_ARM:-OFF} -DWITH_STRIP=${WITH_STRIP:-ON} -DON_INFER=${ON_INFER:-OFF} -DCUDA_ARCH_BIN=${CUDA_ARCH_BIN} -DWITH_RECORD_BUILDTIME=${WITH_RECORD_BUILDTIME:-OFF} -DWITH_UNITY_BUILD=${WITH_UNITY_BUILD:-OFF} -DWITH_ONNXRUNTIME=${WITH_ONNXRUNTIME:-OFF} -DWITH_CUDNN_FRONTEND=${WITH_CUDNN_FRONTEND:-OFF} -DWITH_CPP_TEST=${WITH_CPP_TEST:-OFF} -DFA_BUILD_WITH_CACHE=${FA_BUILD_WITH_CACHE:-ON}
+    cmake .. -DCMAKE_BUILD_TYPE=${CMAKE_BUILD_TYPE:-Release} -DWITH_GPU=${WITH_GPU:-OFF} -DWITH_SHARED_PHI=${WITH_SHARED_PHI:-OFF} -DWITH_TENSORRT=${WITH_TENSORRT:-ON} -DWITH_OPENVINO=${WITH_OPENVINO:-OFF} -DWITH_ROCM=${WITH_ROCM:-OFF} -DWITH_CINN=${WITH_CINN:-OFF} -DWITH_DISTRIBUTE=${distributed_flag} -DWITH_MKL=${WITH_MKL:-ON} -DWITH_AVX=${WITH_AVX:-OFF} -DCUDA_ARCH_NAME=${CUDA_ARCH_NAME:-All} -DNEW_RELEASE_PYPI=${NEW_RELEASE_PYPI:-OFF} -DNEW_RELEASE_ALL=${NEW_RELEASE_ALL:-OFF} -DNEW_RELEASE_JIT=${NEW_RELEASE_JIT:-OFF} -DWITH_PYTHON=${WITH_PYTHON:-ON} -DCUDNN_ROOT=/usr/ -DWITH_TESTING=${WITH_TESTING:-ON} -DWITH_COVERAGE=${WITH_COVERAGE:-OFF} -DWITH_INCREMENTAL_COVERAGE=${WITH_INCREMENTAL_COVERAGE:-OFF} -DCMAKE_MODULE_PATH=/opt/rocm/hip/cmake -DCMAKE_EXPORT_COMPILE_COMMANDS=ON -DWITH_INFERENCE_API_TEST=${WITH_INFERENCE_API_TEST:-ON} -DINFERENCE_DEMO_INSTALL_DIR=${INFERENCE_DEMO_INSTALL_DIR} -DPY_VERSION=${PY_VERSION:-3.8} -DCMAKE_INSTALL_PREFIX=${INSTALL_PREFIX:-/paddle/build} -DWITH_PSLIB=${pslib_flag} -DWITH_GLOO=${gloo_flag} -DWITH_XPU=${WITH_XPU:-OFF} -DWITH_IPU=${WITH_IPU:-OFF} -DXPU_SDK_ROOT=${XPU_SDK_ROOT:-""} -DWITH_XPU_BKCL=${WITH_XPU_BKCL:-OFF} -DWITH_XPU_XHPC=${WITH_XPU_XHPC:-OFF} -DWITH_XPU_XFT=${WITH_XPU_XFT:-OFF} -DWITH_XPU_XRE5=${WITH_XPU_XRE5:-OFF} -DWITH_XPU_FFT=${WITH_XPU_FFT:-OFF} -DWITH_ARM=${WITH_ARM:-OFF} -DWITH_STRIP=${WITH_STRIP:-ON} -DON_INFER=${ON_INFER:-OFF} -DCUDA_ARCH_BIN=${CUDA_ARCH_BIN} -DWITH_RECORD_BUILDTIME=${WITH_RECORD_BUILDTIME:-OFF} -DWITH_UNITY_BUILD=${WITH_UNITY_BUILD:-OFF} -DWITH_ONNXRUNTIME=${WITH_ONNXRUNTIME:-OFF} -DWITH_CUDNN_FRONTEND=${WITH_CUDNN_FRONTEND:-OFF} -DWITH_CPP_TEST=${WITH_CPP_TEST:-OFF} -DWITH_FA_BUILD_WITH_CACHE=${WITH_FA_BUILD_WITH_CACHE:-ON}
     ========================================
 EOF
     export CMAKE_BUILD_TYPE=${CMAKE_BUILD_TYPE:-Release}
@@ -4357,7 +4357,7 @@ EOF
     export WITH_SHARED_PHI=${WITH_SHARED_PHI:-OFF}
     export WITH_NVCC_LAZY=${WITH_NVCC_LAZY:-ON}
     export WITH_CPP_TEST=${WITH_CPP_TEST:-ON}
-    export FA_BUILD_WITH_CACHE=${FA_BUILD_WITH_CACHE:-ON}
+    export WITH_FA_BUILD_WITH_CACHE=${WITH_FA_BUILD_WITH_CACHE:-ON}
 
 
     if [ "$SYSTEM" == "Linux" ];then

python/env_dict.py.in

@@ -115,5 +115,12 @@ env_dict={
     'TENSORRT_LIBRARY_DIR':'@TENSORRT_LIBRARY_DIR@',
     'WITH_FLAGCX':'@WITH_FLAGCX@',
     'FLAGCX_LIB':'@FLAGCX_LIB@',
-    'COMPILED_CUDA_ARCHS':'@COMPILED_CUDA_ARCHS@'
+    'COMPILED_CUDA_ARCHS':'@COMPILED_CUDA_ARCHS@',
+    'WITH_DISTRIBUTE':'@WITH_DISTRIBUTE@',
+    'WITH_NVSHMEM':'@WITH_NVSHMEM@',
+    'NVSHMEM_BOOTSTRAP_UID_LIB':'@NVSHMEM_BOOTSTRAP_UID_LIB@',
+    'NVSHMEM_BOOTSTRAP_PMI_LIB':'@NVSHMEM_BOOTSTRAP_PMI_LIB@',
+    'NVSHMEM_BOOTSTRAP_PMI2_LIB':'@NVSHMEM_BOOTSTRAP_PMI2_LIB@',
+    'NVSHMEM_TRANSPORT_IBRC_LIB':'@NVSHMEM_TRANSPORT_IBRC_LIB@',
+    'NVSHMEM_TRANSPORT_IBGDA_LIB':'@NVSHMEM_TRANSPORT_IBGDA_LIB@'
 }