NVIDIA_SDK PROJECT Compilation "cudaEvent_t" is undefined

I Work on Dell Precision M4300 with a Quadro FX 360M. My OS is a Debian. 4.0
I Install CUDA 1.0 and I create a link with gcc-4.1 in /usr/local/cuda/bin.

I try to compile NVIDIA_SDK project, and there is no problem for boxFilter, convolutionSeparable, deviceQuery, dwtHaar1D, dxtc and simpleGL but for asyncAPI I have an error message:

“asyncAPI.cu”, line 89: error: identifier “cudaEvent_t” is undefined
cudaEvent_t start, stop;



CUDA_1.0 will not work with gcc-4.1. You need to install the CUDA_2.0-beta packages. Also note that Debian is not supported at this time.