i’d like to know about the possibilities and challenges of cuda+directx/opengl. i’ve no hardware at hand to try, but from the Guide 1.0, cuda for OpenGL is for all buffer objects; cuda for direcx is only for 9.0 and only for vertex buffers. so here are questions:
1, can i say using opengl is better supported in cuda than using directx?
2, what are the possibilities and challenges to use shaders (vs, gs, ps) in cuda?
3, or on the contrary, since cuda 1.1 is in .dll form, can we include it in opengl/directx programs? in that way, we can use 3d graphics shaders for renderings and cuda for pre-computations.
Thanks for pointing out my ignorance!
I am in no way an expert in this area, but as far as I know OpenGl is indeed better supported (as it is available on all platforms where CUDA runs, so I don’t expect this to change very quickly). About including the dll in programs, as far as I read from NVIDIA people, they don’t allow redistributing the dll, since it is included in the NVIDIA drivers, so people using the latest drivers already have the dll, and as such, there is no need to include it in your distribution.
- I’m not sure I follow this question. Shaders (vs, gs, ps) are not used in CUDA at all. CUDA is a stand-alone toolkit.
Thanks, i see, appreciate your comment very much.
So can we call cuda code/dll/lib within a opengl/directx program? if hard or not, what are the impeding challenges? thanks!