Tweet
When GPU rendering in RTX mode, are ONLY the RT cores used or are both CUDA cores and RT cores used simultaneously? My understanding is that both RT cores and CUDA cores run in parallel in applications like gaming, is the same true for RTX GPU rendering? If they are run in parallel for GPU rendering, do you see the current generation of NVidia GPUs being limited by their current number of RT cores, as in do you foresee large gains in the number of RT cores of the next generation of Nvidia GPUs dramatically increasing the performance of RTX GPU rendering? Does the Ray Tracing portion of the rendering process take the majority of the time or is it other things? For example, it seems even the highest end RTX GPUs of the current generation have trouble maintaining high frame rates while playing games with RT enabled. The requirements of gaming are obviously quite different than that of rendering, I wonder how you see the current state of RTX cards performance wise and do you foresee large gains on the next generation of NVidia GPUs (not necessarily from the exact specification perspective), but the future of RTX? Do you see the RT cores of the current generation GPUs as a limiting factor in rendering speeds?
I know some of the above questions are most likely unanswerable(no one knows what NVidia's next hardware specs will be) or naive (i read too many articles about RT cores but dont really have any real technical understanding), but this is my ill-informed attempt at getting some information to make a hardware purchase.
I know some of the above questions are most likely unanswerable(no one knows what NVidia's next hardware specs will be) or naive (i read too many articles about RT cores but dont really have any real technical understanding), but this is my ill-informed attempt at getting some information to make a hardware purchase.
Comment