Using CUDA with two displays

Accelerated Computing CUDA CUDA Programming and Performance
1

I have an application, in which I need to send CUDA results to two fullscreen displays. Note that I want the two display shows the different thing, not just two same copies. Then I set my NVIDIA display card to horzontal spam mode, set resolution of each monitor to be 1024x768. And in my OpenGL code, I create a window with size (2048x768).

In my CUDA program, I write results to a 2048x768 PBO buffer and display this buffer to my big window. I expect to see the whole window is spanned to two monitors. But it turned out that the content of my output was squeezed into one monitor and the other monitor has nothing on it.

Then I change the window size to 1024x768, hope horizontal span can just span it evenly to two displays, but the same thing happened again.

Can expert tell me why is this and how to make it right?

2

At work with CUDA, I turn off the second monitor as CUDA not always with it works correctly. It is necessary to wait for the fresh version of the driver of the device.
Activation " Single display performance mode " in options of the driver not always helps.