I have got two video cards, one CUDA ready and one not. In Windows 7 64-bits, I have the older card as my primary display and the CUDA card as secondary display by extending the screen to it. However, I can’t run a kernel longer than a few seconds. Taking advice from the release notes, I tried not to extend the screen and disconnected the display on the secondary device. However, Cuda runtime wasn’t able to detect the secondary CUDA device at all.
Anybody knows how to do it in Windows Vista / 7?
Thanks,