Request CUDA code suggestion to allow a selective monitor output

I am a new to CUDA . Please, suggest me a code sample or function that allows to output an image to a selected monitor from a multi-monitor configuration.

To better describe what I need to accomplish, I have three monitors connected to a GeoForce 630 Card. I am writting a code that create an image. One is dedicated to display the desktop. Now, I want to project the images using the second or third monitor.

This is similar to what we do in a PowePoint presentation, where your run your software in your laptop and the slides are projected in the LDP for your audience.

In my case, I use Mathematica with CUDA to create a series images from computation of complex equations. I need these images projected to the second, third or both monitor simultaneously. I need a CUDA code allowing to display the images to a specified monitor.

I hope my problem statement is not confusing.

You’re probably looking for something along the lines of this?

Not sure opening a window on a particular monitor has anything to do with CUDA, well perhaps specifically to OpenGL or whatever API is creating a graphical window… perhaps something like:

You’d have to be a little more specific to be able to search for what you specifically want to do.

Thank you a lot. But this is not exactly what I am looking after because the suggested post deals with opening different Windows within a monitor area.

In my situation, I need control over each MONITOR separately.

I like to do this from CUDA. That is because my code is executed from a computational software Wolfram Mathematica. The software provides to create complicated plot figures which I want outputted and displayed in the second monitor.

Any help?

Jose E. Calderon

CUDA is a sandbox for GPGPU programming purely for computation. CUDA doesn’t provide much graphical control aside from having abilities to work with DirectX / OpenGL. For absolute control over what pixels are being sent to each monitor, the only option is to deal with drivers (which would have nothing to do with CUDA).

Windows (and probably Linux) requires that things drawn to the screen be done using the normal routine; request window space, draw to window space. That being said, you can achieve a very similar solution by creating a border-less window for each monitor which is automatically full sized. You can manage these windows from an application which simply draws to each one accordingly. If you go down this route, you should look at drawing libraries such as GDI, DirectX or OpenGL.

If there is some better option I am unaware of, I do hope someone replies with it.