I know this is an odd combination, but I am getting a crash in the CUDA runtime while debugging a program in VS2008, on a Windows Vista system logged in through RDP (Remote Desktop).
If I run the program it works as expected (not using any GPU’s because RDP doesn’t support it), but while running in the debugger, it happens the first time I make any CUDA runtime library call such as cudaGetDeviceCount() that shouldn’t even need the existence of a GPU. The crash is a memory exception happening inside crtIsValidHeapPointer, but there are several layers above that call in the call stack.
Has anyone experienced this? It’s not a critical issue, just a little annoying.