Ray Tracing CUDA - RT cards vs. NonRT cards

Accelerated Computing CUDA CUDA Programming and Performance
1

Hello,

I am looking to develop some ray tracing code on a RT card as well as a Non-RT graphics card which will be rendered offline - I would really like to get performance measurements such as FPS between the two. Is CUDA the way to go if one would like to have the ability to execute on either of these kinds of hardware architectures? Is OptiX the way to go?

Thank you to anyone with some advise on how to proceed.

2

OptiX would seem like the natural fit. Have you tried it? If so, what did you find lacking with regard to your use case?

3

@njuffa, thank you for the reply. No I haven’t tried OptiX, maybe it is the way to go. I am currently using MSVS 2017 - do you know of any tutorial or simple example code using this setup with OpitX? I haven’t been able to find too much in this regard.

Thanks again

4

As far as I am aware, OptiX requires Linux, or the Linux subsystem of Windows 10. Don’t know about supported toolchains. There is a dedicated OptiX forum where you could ask:

https://forums.developer.nvidia.com/c/professional-graphics-and-rendering/advanced-graphics/optix

5

@njuffa, thanks again. I have been trying to get OptiX to run on my windows machine using MSVS 2017 with no luck. I will check out the link you gave.