I thought Ray Tracing was some kind of technology developed by Nvidia, like the DLSS
Thanks for the answers!
Not trying to look down at you, but this is really impressive for NVIDIA in my opinion. If they've actually convinced some people that they invented ray tracing, that's fucking wild. Not your fault at all, since that's clearly what their marketing is going for. But damn, that's really upsetting for me personally.
Ray tracing is ancient. It was used to generate some of the first computer generated images back in the day. It's a really simple mathematical technique, much like the words "differentiation", "integration" or "calculus". NVIDIA's hardware includes a special chip that performs a single math operation used in ray tracing,
very fast. Normally, ray tracing is slow because that single operation takes a while. The operation is ray -> triangle intersection. Does this light ray hit this triangle? needs to be asked millions of times per second. NVIDIA implemented a super cool chip for doing intersection very fast. AMD has also implemented a chip to do this, but it works differently internally. AMD was slower to make the chip and by every account, there's will likely perform slower. That is, they won't be able to perform as many ray-triangle intersections as quickly. But all the API functions will look the same.
Microsoft and the Khronos consortium (the committee who makes OpenGL and Vulkan) work with AMD and NVIDIA to determine what hardware they're adding and a way to present a common interface for that hardware. They're both committees that both companies work with so that developers have a standardized set of functions to use. Then you don't have to relearn everything to change which card you develop for. NVIDIA presented a proposal to both Khronos and Microsoft for their RTX hardware, with some functions to use it back in 2017 or so. From those proposed functions, Microsoft/Khronos talked to AMD and asked what their hardware is going to look like, so they can choose functions that work for both. It's all set up this way due to a lot of history. With AMD talked to, Microsoft developed and announced DXR. Now both companies will implement the same features, but differentiation will occur in performance.
It doesn't make sense to disable the features on AMD since the code should look practically identical. The only differences in code will be for performance rather than functionality. Now on a console game, this would imply a removal of the feature. Since you don't want any low-performance features in a console game. But this is PC, and I don't think they are resisting adding the option because it would tank frame-rates on AMD or something. That doesn't make much sense to me, and as a developer myself I really don't see how there would be any distinction between the code for the different hardware at least in terms of functionality. That's only because Microsoft has explicitly standardized the functions you need to call and how they work.
It all leads back to the pressure that NVIDIA is putting on this game. Like in my work, we are hearing about the stress from this game constantly. They put
so much money into this game that it needs to return big time for NVIDIA. But hey, this will be trivial to test when the game comes out. Should be able to run it on AMD cards anyway with some creative binary patching/hooking and see how much bullshit they're spouting.