CUDA with a real time extension to windows How do I integrate CUDA with RTX?

Hi everyone,

I have just installed an NVIDIA graphics card within an industrial PC with the intent of doing some parallel computing. The existing system has a dual processor with one processor running windows XP and the other dedictated to executing control algorithms on a real time kernel (Ardence RTX).

All the documentation I can find on CUDA talks about how to use Microsoft Visual C++ to run code on XP that calls the GPU. Is anyone familiar with using visual c++ to run code on a real time kernel which uses the GPU? i.e. how/can I set up visual c++ to do this?

Any help is greatly appreciated.



Can you just call ordinal program from your rt kernel and track its execution on winxp core?

I am not familiar with these programs but I assume that I’m not the first to integrate a GPU into a real time extension to windows. I would have though that many of the CUDA applications are time critical and need to be deployed in a system where you have the tight regulation over the timing that a real time extension would provide?