Mapping to a GL_TEXTURE_RECTANGLE

Accelerated Computing CUDA CUDA Programming and Performance
5

So changing the kernel to:

__global__ void kernel(float *dev_ouput, int width, int height)
{
    int x = threadIdx.x + blockIdx.x * blockDim.x;
    int y = threadIdx.y + blockIdx.y * blockDim.y;
    int offset = x + y * blockDim.x * gridDim.x;

    if ( x >= width || y >= height ) return;

    float4 pixel = tex2D(texRef, x / (float)width, y / (float)height);

    dev_ouput[offset*4 + 0] = pixel.x;
    dev_ouput[offset*4 + 1] = pixel.y;
    dev_ouput[offset*4 + 2] = pixel.z;
    dev_ouput[offset*4 + 3] = pixel.w;
}

Sadly gives me the same result.

Oh I should say that I am using SDK 5.5, looking at your link the interop stuff has changed yet again, but I am stuck with 5.5.

I saw the textureReference.normalized value earlier in the docs, it is set to the default 0.