cuUnmapRegister has Launch Failure error. Cuda OpenGL interop. with driver API.

Hi, all

I wrote a simple program that use driver api and opengl.

In the code below, the CU_SAFE_CALL(cuGLUnmapBufferObject(vbo)) in function `runCuda()’ has error and the drawing fails.

Currently the kernel function takes just the vertex array and change its value a little. That’s the only thing. But this causes problem.
If I keep the value of vertices the same, then the above call shows CUDA_SUCCESS.

The kernel function and calling program are as follows:

It turned out to be running okey in m32 version. Thank you for viewing.