I am setting up a windows 7 workstation to work with cuda/opencl applications; the main GPU that should be used for computations is the Quadro 4000.
In order to avoid the 5 seconds limit imposed to the kernel if the GPU is used also to drive the display, I installed an old Geforce 7900GTX to be used only for the display.
Now if I install the drivers for the geforce, the kernel run on the quadro crashes after 5 seconds also if on the quadro there is no display attached; otherwise, if I install the drivers for the quadro, the cuda/opencl software say that it can not find any device with cuda/opencl capabilities.
Does someone has some hints about how such an hardware configuration should be set up from the software/drivers point of view? What am I doing wrong? Is it possible to use such configuration?
Thanks in advance,