CUDA Visual Studio 2010 Express build error

Question

I am trying to get started with CUDA programming on Windows using Visual Studio 2010 Express on a 64 bit Windows 7. It took me a while setting up the environment, and I just wrote my first program, helloWorld.cu :)

Currently I am working with the following program:

#include <stdio.h>


__global__ void add(int a, int b, int *c){
    *c = a + b; 
}

int main(void){
    int c;
    int *dev_c;
    HANDLE_ERROR( cudaMalloc( (void**)&dev_c, sizeof(int) ) );

    add<<<1,1>>>(2, 7, dev_c);

    HANDLE_ERROR( cudaMemcpy( &c, dev_c, sizeof(int), cudaMemcpyDeviceToHost ) );

    printf("2 + 7 = %d\n", c);
    cudaFree( dev_c );

    return 0;
}

And here is the output of build with Windows7.1SDK platform toolset:

1>  C:\Users\User\documents\visual studio 2010\Projects\CudaTest2\CudaTest2>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\bin\nvcc.exe" -gencode=arch=compute_10,code=\"sm_10,compute_10\" --use-local-env --cl-version  -ccbin "C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\bin\x86_amd64"  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\include"  -G   --keep-dir "x64\Debug" -maxrregcount=0  --machine 64 --compile  -g    -Xcompiler "/EHsc /W1 /nologo /O2 /Zi  /MD  " -o "x64\Debug\hello.cu.obj" "C:\Users\User\documents\visual studio 2010\Projects\CudaTest2\CudaTest2\hello.cu" 
1>nvcc : fatal error : '-ccbin': expected a number
1>C:\Program Files (x86)\MSBuild\Microsoft.Cpp\v4.0\BuildCustomizations\CUDA 5.0.targets(592,9): error MSB3721: The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\bin\nvcc.exe" -gencode=arch=compute_10,code=\"sm_10,compute_10\" --use-local-env --cl-version  -ccbin "C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\bin\x86_amd64"  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\include"  -G   --keep-dir "x64\Debug" -maxrregcount=0  --machine 64 --compile  -g    -Xcompiler "/EHsc /W1 /nologo /O2 /Zi  /MD  " -o "x64\Debug\hello.cu.obj" "C:\Users\User\documents\visual studio 2010\Projects\CudaTest2\CudaTest2\hello.cu"" exited with code -1.

However, with v100 platform toolset I am getting this error:

C:\Users\User\documents\visual studio 2010\Projects\CudaTest2\CudaTest2>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\bin\nvcc.exe" -gencode=arch=compute_10,code=\"sm_10,compute_10\" --use-local-env --cl-version 2010 -ccbin "C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\bin\x86_amd64"  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\include"  -G   --keep-dir "x64\Debug" -maxrregcount=0  --machine 64 --compile  -g    -Xcompiler "/EHsc /W1 /nologo /O2 /Zi  /MD  " -o "x64\Debug\hello.cu.obj" "C:\Users\User\documents\visual studio 2010\Projects\CudaTest2\CudaTest2\hello.cu" 
1>C:\Program Files (x86)\Microsoft Visual Studio 9.0\VC\include\math.h(455): error : dllexport/dllimport requires external linkage
1>C:/Users/User/documents/visual studio 2010/Projects/CudaTest2/CudaTest2/hello.cu(12): error : identifier "HANDLE_ERROR" is undefined
1>c:\program files\nvidia gpu computing toolkit\cuda\v5.0\include\math_functions.h(2900): error : function "hypotf" was referenced but not defined
1>C:\Program Files (x86)\MSBuild\Microsoft.Cpp\v4.0\BuildCustomizations\CUDA 5.0.targets(592,9): error MSB3721: The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\bin\nvcc.exe" -gencode=arch=compute_10,code=\"sm_10,compute_10\" --use-local-env --cl-version 2010 -ccbin "C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\bin\x86_amd64"  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\include"  -G   --keep-dir "x64\Debug" -maxrregcount=0  --machine 64 --compile  -g    -Xcompiler "/EHsc /W1 /nologo /O2 /Zi  /MD  " -o "x64\Debug\hello.cu.obj" "C:\Users\User\documents\visual studio 2010\Projects\CudaTest2\CudaTest2\hello.cu"" exited with code 2.
========== Build: 0 succeeded, 1 failed, 0 up-to-date, 0 skipped ==========

Could anyone point me in right direction? I have been reading almost all posts about related issues, installed and uninstalled tools couple of times and still can't get a proper build.

Answer 1

This is because visual studio doesn't have the privilege to access the Temp folder.

navigate to C:\Users\userName\AppData\Local\Temp and change the access security rules of the current user (in this case it is userName) to full control

Answer 2

The answer is two-fold. The HANDLE_ERROR build error you're getting using the v100 platform is related to the HANDLE_ERROR function being defined in the additional headers for the book "Cuda By Example" - I recognize the example, so I assume that's the code you're trying to compile. You haven't included this filem because it's not listed in the book code listings ... a small oversight from the author(s), imho.

You can download the code examples, and the additional header "book.h" you need from the book page:https://developer.nvidia.com/content/cuda-example-introduction-general-purpose-gpu-programming-0

For reference, this is the HANDLE_ERROR code from book.h:

static void HandleError( cudaError_t err,
                     const char *file,
                     int line ) {
if (err != cudaSuccess) {
    printf( "%s in %s at line %d\n", cudaGetErrorString( err ),
            file, line );
    exit( EXIT_FAILURE );
  }
}
#define HANDLE_ERROR( err ) (HandleError( err, __FILE__, __LINE__ ))

Answer 3

The following worked for me. Edit the file *C:\Program Files (x86)\MSBuild\Microsoft.Cpp\v4.0\BuildCustomizations\CUDA 5.0.prop*s. Search for PlatformToolset and add an entry for Windows7.1SDK. Set the value to 2010.

<CudaClVersion Condition="'$(PlatformToolset)' == 'Windows7.1SDK'">2010</CudaClVersion>