After running build_windows.bat, I got error MSB6006: "cmd.exe" e xited with code 1

[danfouer]

Hi, I run build_windows.bat, then I got: "C:\deepvision\caffe2\build\ALL_BUILD.vcxproj" (default target) (1) -> "C:\deepvision\caffe2\build\caffe2\Caffe2_GPU.vcxproj" (default target) (7) -> (CustomBuild target) -> C:\Program Files (x86)\MSBuild\Microsoft.Cpp\v4.0\V140\Microsoft.CppCommon.targets(171,5): error MSB6006: "cmd.exe" e xited with code 1. [C:\deepvision\caffe2\build\caffe2\Caffe2_GPU.vcxproj]

773 Warning(s)
1 Error(s)

Can anyone help me out with that?

[sebaap]

Did you find any solution to it? I have the same issue. I've been able to compile it with USE_CUDA=OFF but if i set it to ON it fails throwing that error.

If it's useful, I'm running W10 with CUDA 8.0 and VS 2017

[sebaap]

Ok, I found the problem. For anyone having this issue seems like CUDA has problems with VS 2017, because it tries to find the compiler in C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\bin\ but for VS 2017 it's located in C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.10.25017\bin\HostX64\x64\. So I switched to VS2015 and it works.

[rajat95]

I am facing same issue with CUDA 8 and VS2015 on Win 10

[tozimmer]

I have the same issue with CUDA 8/9 and VS2017 on Windows 10 (master branch [b1b4e8d511c7206174c7a83de89e15e2620b856e]), building with USE_CUDA=OFF is working fine, but with USE_CUDA=ON 1 get the above mentioned Error

Microsoft.CppCommon.targets(171,5): error MSB6006: "cmd.exe" exited with code 1. [..\caffe2_gpu.vcxp
roj]

With VS2015 I encouter following errors:

"C:\Users\zimmermann\caffe2\build\ALL_BUILD.vcxproj" (Standardziel) (1) ->
"C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj" (Standardziel) (5) ->
(ClCompile Ziel) ->
  C:\Users\zimmermann\caffe2\caffe2\operators\elementwise_op_schema.cc(71): error C2664: 'caffe2::OpSchema &caffe2::OpSchema::CostInferenceFunction(caffe2::OpSchema::CostInferenceFunctionType &&)': cannot convert argument 1 from 'caffe2
::OpSchema::Cost (__cdecl *)(const caffe2::OperatorDef &,const std::vector<caffe2::TensorShape,std::allocator<_Ty>> &)' to 'caffe2::OpSchema::CostInferenceFunctionType &&' [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\elementwise_op_schema.cc(72): error C2228: left of '.IdenticalTypeAndShapeOfInput' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\elementwise_op_schema.cc(73): error C2228: left of '.FillUsing' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\elementwise_op_schema.cc(78): error C2664: 'caffe2::OpSchema &caffe2::OpSchema::CostInferenceFunction(caffe2::OpSchema::CostInferenceFunctionType &&)': cannot convert argument 1 from 'caffe2
::OpSchema::Cost (__cdecl *)(const caffe2::OperatorDef &,const std::vector<caffe2::TensorShape,std::allocator<_Ty>> &)' to 'caffe2::OpSchema::CostInferenceFunctionType &&' [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\elementwise_op_schema.cc(79): error C2228: left of '.IdenticalTypeAndShapeOfInput' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\elementwise_op_schema.cc(80): error C2228: left of '.FillUsing' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\elementwise_op_schema.cc(85): error C2664: 'caffe2::OpSchema &caffe2::OpSchema::CostInferenceFunction(caffe2::OpSchema::CostInferenceFunctionType &&)': cannot convert argument 1 from 'caffe2
::OpSchema::Cost (__cdecl *)(const caffe2::OperatorDef &,const std::vector<caffe2::TensorShape,std::allocator<_Ty>> &)' to 'caffe2::OpSchema::CostInferenceFunctionType &&' [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\elementwise_op_schema.cc(86): error C2228: left of '.IdenticalTypeAndShapeOfInput' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\elementwise_op_schema.cc(87): error C2228: left of '.FillUsing' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\elementwise_op_schema.cc(92): error C2664: 'caffe2::OpSchema &caffe2::OpSchema::CostInferenceFunction(caffe2::OpSchema::CostInferenceFunctionType &&)': cannot convert argument 1 from 'caffe2
::OpSchema::Cost (__cdecl *)(const caffe2::OperatorDef &,const std::vector<caffe2::TensorShape,std::allocator<_Ty>> &)' to 'caffe2::OpSchema::CostInferenceFunctionType &&' [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\elementwise_op_schema.cc(93): error C2228: left of '.IdenticalTypeAndShapeOfInput' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\elementwise_op_schema.cc(94): error C2228: left of '.FillUsing' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\leaky_relu_op.cc(60): error C2664: 'caffe2::OpSchema &caffe2::OpSchema::CostInferenceFunction(caffe2::OpSchema::CostInferenceFunctionType &&)': cannot convert argument 1 from 'caffe2::OpSche
ma::Cost (__cdecl *)(const caffe2::OperatorDef &,const std::vector<caffe2::TensorShape,std::allocator<_Ty>> &)' to 'caffe2::OpSchema::CostInferenceFunctionType &&' [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\leaky_relu_op.cc(61): error C2228: left of '.IdenticalTypeAndShape' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\leaky_relu_op.cc(62): error C2228: left of '.SetDoc' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\leaky_relu_op.cc(67): error C2228: left of '.Input' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\leaky_relu_op.cc(68): error C2228: left of '.Output' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\relu_op.cc(78): error C2664: 'caffe2::OpSchema &caffe2::OpSchema::CostInferenceFunction(caffe2::OpSchema::CostInferenceFunctionType &&)': cannot convert argument 1 from 'caffe2::OpSchema::Co
st (__cdecl *)(const caffe2::OperatorDef &,const std::vector<caffe2::TensorShape,std::allocator<_Ty>> &)' to 'caffe2::OpSchema::CostInferenceFunctionType &&' [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\relu_op.cc(79): error C2228: left of '.IdenticalTypeAndShape' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\relu_op.cc(80): error C2228: left of '.SetDoc' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\relu_op.cc(85): error C2228: left of '.Input' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]
  C:\Users\zimmermann\caffe2\caffe2\operators\relu_op.cc(86): error C2228: left of '.Output' must have class/struct/union [C:\Users\zimmermann\caffe2\build\caffe2\caffe2.vcxproj]


"C:\Users\zimmermann\caffe2\build\ALL_BUILD.vcxproj" (Standardziel) (1) ->
"C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj" (Standardziel) (7) ->
(CustomBuild Ziel) ->
  C:/Users/zimmermann/caffe2\caffe2/utils/GpuBitonicSort.cuh(42): error : expected an identifier [C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj]
  C:/Users/zimmermann/caffe2\caffe2/utils/GpuBitonicSort.cuh(44): error : expected an identifier [C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj]
  C:/Users/zimmermann/caffe2\caffe2/utils/GpuBitonicSort.cuh(42): error : expected an identifier [C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj]
  C:/Users/zimmermann/caffe2\caffe2/utils/GpuBitonicSort.cuh(44): error : expected an identifier [C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj]
  C:/Users/zimmermann/caffe2\caffe2/utils/GpuBitonicSort.cuh(42): error : expected an identifier [C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj]
  C:/Users/zimmermann/caffe2\caffe2/utils/GpuBitonicSort.cuh(44): error : expected an identifier [C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj]
  C:/Users/zimmermann/caffe2\caffe2/utils/GpuBitonicSort.cuh(42): error : expected an identifier [C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj]
  C:/Users/zimmermann/caffe2\caffe2/utils/GpuBitonicSort.cuh(44): error : expected an identifier [C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj]
  C:/Users/zimmermann/caffe2\caffe2/utils/GpuBitonicSort.cuh(42): error : expected an identifier [C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj]
  C:/Users/zimmermann/caffe2\caffe2/utils/GpuBitonicSort.cuh(44): error : expected an identifier [C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj]
  C:/Users/zimmermann/caffe2\caffe2/utils/GpuBitonicSort.cuh(42): error : expected an identifier [C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj]
  C:/Users/zimmermann/caffe2\caffe2/utils/GpuBitonicSort.cuh(44): error : expected an identifier [C:\Users\zimmermann\caffe2\build\caffe2\caffe2_gpu.vcxproj]

Any way to fix this issue?

[andrewwasielewski]

Also having the same problem with CUDA 9 and VS2017 on Windows 10 - any suggestions for resolution?

[DavideCatto]

Me too, any suggestions?

[Yangqing]

I am 95% towards dealing with cuda + MSVC. Right now MSVC 2017 and CUDA9 do not work for any framework unfortunately. See https://devtalk.nvidia.com/default/topic/1027299/cuda-setup-and-installation/cuda-9-failed-to-support-the-latest-visual-studio-2017-version-15-5/ . For CUDA9 and VS2015 I am able to build it successfully, stay tuned for a few days.

I'll loop back when the windows cuda (and shared library mode) works.

[Yangqing]

For now, if you sync beyond https://github.com/caffe2/caffe2/commit/4372b8d04b203c85f68c144df82dae560887ef13 , at least you will be able to overcome the MSB6006 issue.

[notobject]

Also having the same problem with CUDA 8.0 and VS2017 on Windows 10 #

[Yangqing]

Update: PR #1793 is out to make windows build a bit better.

[hanna-kim]

I still have an same error that says

I have installed CUDA 9.0, python 3.5, Visual Studio Community 2015. (originally there was only vs2013 but now I have both version 2013 and 2015. Should I uninstall vs 2013?)

can anyone help me please? I struggled merely 3 more days to install caffe...