two GPUs in one computer how to make sure programs run on which

We bought a dell workstation with Quadro FX570, but also installed a Telsa on it. How to make sure that our GPU program will run with Tesla, instead of FX570.
Or is it possible to ask FX570 to run a kernel and ask Tesla to run another kernel?

Thank you,

Run-time API includes calls to determine how many cuda-capable devices you have. There are also calls to query for each device’s information (which includes a name), set the current device, etc.

To use both cards simultaneously, you’ll have to have one CPU thread for each GPU.

Check the programming guide for more details.