GPU Paddle release-1.8
win10, cuda 10.0, cudnn 7.6, python 3.7,Visual Studio 2017
按照官方文档里win编译需要的python库：umpy, protobuf, wheel都安装。

cmake命令：
cmake .. -G "Visual Studio 15 2017 Win64" -DPY_VERSION=3.7 -DPYTHON_INCLUDE_DIR=${PYTHON_INCLUDE_DIRS} -DPYTHON_LIBRARY=${PYTHON_LIBRARY} -DPYTHON_EXECUTABLE=${PYTHON_EXECUTABLE} -DWITH_FLUID_ONLY=ON -DWITH_GPU=ON -DWITH_TESTING=OFF -DCMAKE_BUILD_TYPE=Release -DCUDA_TOOLKIT_ROOT_DIR=${CUDA_TOOLKIT_ROOT_DIR}

报错地方：
CMake Error at cmake/generic.cmake:313 (add_dependencies): The dependency target "python" of target "py_func_op" does not exist.

CMake Error at cmake/generic.cmake:313 (add_dependencies): The dependency target "python" of target "paddle_pybind" does not exist.

以及诸多警告：

CMake Warning at CMakeLists.txt:126 (MESSAGE):
  Disable NCCL when compiling for Windows.  Force WITH_NCCL=OFF.

CMake Warning at CMakeLists.txt:144 (MESSAGE):
  If the environment is multi-card, the WITH_NCCL option needs to be turned
  on, otherwise only a single card can be used.

CMake Warning at cmake/tensorrt.cmake:46 (message):
  TensorRT is NOT found when WITH_DSO is ON.

CMake Warning at CMakeLists.txt:224 (message):
  On inference mode, will take place some specific optimization.  Turn on the
  ON_INFER flag when building inference_lib only.

CMake Warning:
  Manually-specified variables were not used by the project:

把完整的编译过程贴下来：

\Paddle\build>cmake .. -G "Visual Studio 15 2017 Win64" -DPY_VERSION=3.7 -DPYTHON_INCLUDE_DIR=${PYTHON_INCLUDE_DIRS} -DPYTHON_LIBRARY=${PYTHON_LIBRARY} -DPYTHON_EXECUTABLE=${PYTHON_EXECUTABLE} -DWITH_FLUID_ONLY=ON -DWITH_GPU=ON -DWITH_TESTING=OFF -DCMAKE_BUILD_TYPE=Release -DCUDA_TOOLKIT_ROOT_DIR=${CUDA_TOOLKIT_ROOT_DIR}
-- Found Paddle host system: win32, version:
-- Found Paddle host system's CPU: 12 cores
-- Selecting Windows SDK version 10.0.17763.0 to target Windows 10.0.18363.
-- The CXX compiler identification is MSVC 19.16.27041.0
-- The C compiler identification is MSVC 19.16.27041.0
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: D:/AppInstall/Work/Microsoft Visual Studio/2017/Community/VC/Tools/MSVC/14.16.27023/bin/Hostx86/x64/cl.exe - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: D:/AppInstall/Work/Microsoft Visual Studio/2017/Community/VC/Tools/MSVC/14.16.27023/bin/Hostx86/x64/cl.exe - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- CXX compiler: D:/AppInstall/Work/Microsoft Visual Studio/2017/Community/VC/Tools/MSVC/14.16.27023/bin/Hostx86/x64/cl.exe, version: MSVC 19.16.27041.0
-- C compiler: D:/AppInstall/Work/Microsoft Visual Studio/2017/Community/VC/Tools/MSVC/14.16.27023/bin/Hostx86/x64/cl.exe, version: MSVC 19.16.27041.0
-- AR tools: D:/AppInstall/Work/Microsoft Visual Studio/2017/Community/VC/Tools/MSVC/14.16.27023/bin/Hostx86/x64/lib.exe
-- Use static C runtime time, refer to https://docs.microsoft.com/en-us/cpp/c-runtime-library/crt-library-features?view=vs-2019
-- Using parallel compiling (/MP)
-- Found Git: D:/AppInstall/Work/Git/cmd/git.exe (found version "2.27.0.windows.1")
-- Looking for pthread.h
-- Looking for pthread.h - not found
-- Found Threads: TRUE
-- Performing Test MMX_FOUND
-- Performing Test MMX_FOUND - Failed
-- Performing Test SSE2_FOUND
-- Performing Test SSE2_FOUND - Success
-- Performing Test SSE3_FOUND
-- Performing Test SSE3_FOUND - Success
-- Performing Test AVX_FOUND
-- Performing Test AVX_FOUND - Success
-- Performing Test AVX2_FOUND
-- Performing Test AVX2_FOUND - Success
-- Performing Test AVX512F_FOUND
-- Performing Test AVX512F_FOUND - Failed
CMake Warning at CMakeLists.txt:126 (MESSAGE):
  Disable NCCL when compiling for Windows.  Force WITH_NCCL=OFF.

CMake Warning at CMakeLists.txt:144 (MESSAGE):
  If the environment is multi-card, the WITH_NCCL option needs to be turned
  on, otherwise only a single card can be used.

-- Current cuDNN header is D:/AppInstall/Work/Nivida GPU Computing/include/cudnn.h. Current cuDNN version is v7.6.
-- Disable XBYAK in Windows and MacOS
-- Looking for C++ include shlwapi.h
-- Looking for C++ include shlwapi.h - found
-- BOOST_VERSION: 1.41.0, BOOST_URL: http://paddlepaddledeps.bj.bcebos.com/boost_1_41_0.tar.gz
-- warp-ctc library: C:/Users/xx/Paddle/build/third_party/install/warpctc/bin/warpctc.dll
-- Download dependence[cudaerror] from http://paddlepaddledeps.bj.bcebos.com/cudaErrorMessage.tar.gz
-- MKLML_VER: mklml_win_2019.0.1.20181227, MKLML_URL: https://paddlepaddledeps.bj.bcebos.com/mklml_win_2019.0.1.20181227.zip
-- Found cblas and lapack in MKLML (include: C:/Users/xx/Paddle/build/third_party/install/mklml/include, library: C:/Users/xx/Paddle/build/third_party/install/mklml/lib/mklml.lib)
-- Set C:/Users/xx/Paddle/build/third_party/install/mkldnn/llib to runtime path
-- MKLDNN library: C:/Users/xx/Paddle/build/third_party/install/mkldnn/bin/mkldnn.lib
-- Protobuf protoc executable: C:/Users/xx/Paddle/build/third_party/install/protobuf/bin/protoc.exe
-- Protobuf-lite library: C:/Users/xx/Paddle/build/third_party/install/protobuf/lib/libprotobuf-lite.lib
-- Protobuf library: C:/Users/xx/Paddle/build/third_party/install/protobuf/lib/libprotobuf.lib
-- Protoc library: C:/Users/xx/Paddle/build/third_party/install/protobuf/lib/libprotoc.lib
-- Protobuf version: 3.1.0
-- Found PythonInterp: C:/Users/xx/Paddle/build/${PYTHON_EXECUTABLE} (Required is at least version "3.7")
-- Found PythonLibs: C:/Users/xx/Paddle/build/${PYTHON_LIBRARY} (Required is at least version "3.7")
-- Looking for UINT64_MAX
-- Looking for UINT64_MAX - found
-- Looking for sys/types.h
-- Looking for sys/types.h - found
-- Looking for stdint.h
-- Looking for stdint.h - found
-- Looking for stddef.h
-- Looking for stddef.h - found
-- Check size of pthread_spinlock_t
-- Check size of pthread_spinlock_t - failed
-- Check size of pthread_barrier_t
-- Check size of pthread_barrier_t - failed
-- CUDA detected: 10.0
-- WARNING: This is just a warning for publishing release.
      You are building GPU version without supporting different architectures.
      So the wheel package may fail on other GPU architectures.
      You can add -DCUDA_ARCH_NAME=All in cmake command
      to get a full wheel package to resolve this warning.
      While, this version will still work on local GPU architecture.
-- Automatic GPU detection failed. Building for all known architectures.
-- Added CUDA NVCC flags for: sm_30 sm_35 sm_50 sm_52 sm_60 sm_61 sm_70 sm_75
CMake Warning at cmake/tensorrt.cmake:46 (message):
  TensorRT is NOT found when WITH_DSO is ON.
Call Stack (most recent call first):
  CMakeLists.txt:183 (include)

-- Paddle version is 0.0.0
-- Found CUDA: ${CUDA_TOOLKIT_ROOT_DIR} (found version "10.0")
-- Cannot find CUPTI, GPU Profiling is incorrect.
-- Enable Intel OpenMP with C:/Users/xx/Paddle/build/third_party/install/mklml/lib/libiomp5md.lib
CMake Warning at CMakeLists.txt:224 (message):
  On inference mode, will take place some specific optimization.  Turn on the
  ON_INFER flag when building inference_lib only.

-- commit: 659da23f17
-- branch: release/1.8
-- WARNING: This is just a warning for publishing release.
      You are building AVX version without NOAVX core.
      So the wheel package may fail on NOAVX machine.
      You can add -DFLUID_CORE_NAME=/path/to/your/core_noavx.* in cmake command
      to get a full wheel package to resolve this warning.
      While, this version will still work on local machine.
-- Configuring done
CMake Error at cmake/generic.cmake:313 (add_dependencies):
  The dependency target "python" of target "py_func_op" does not exist.
Call Stack (most recent call first):
  paddle/fluid/operators/CMakeLists.txt:123 (cc_library)

CMake Error at cmake/generic.cmake:313 (add_dependencies):
  The dependency target "python" of target "paddle_pybind" does not exist.
Call Stack (most recent call first):
  paddle/fluid/pybind/CMakeLists.txt:114 (cc_library)

-- Generating done
CMake Warning:
  Manually-specified variables were not used by the project:

    WITH_FLUID_ONLY

CMake Generate step failed.  Build files cannot be regenerated correctly.