3 ms·
The article says that MS (like everyone else) expects there to be fewer and fewer "fixed-function" features in a GPU with more shifting over to software-defined
by tgb 9y ago
The article says that MS (like everyone else) expects there to be fewer and fewer "fixed-function" features in a GPU with more shifting over to software-defined shader code. But isn't this just an introduction of more fixed-functionality to the pipeline? So what's the advantage of doing it this way versus writing a ray tracer in the current compute shaders?
- bhouston 9y agoThe main issue with current compute shaders is intersecting with arbitrary geometry in a fast way. I expect that the main thing they will add is raycast() function that propagates the ray to the next surface, similar to how RenderMan shaders or OSL works.
- kevingadd 9y agoAPIs like a ray tracing API can be implemented in driver software (on-CPU) or in device firmware (on-GPU, but programmatic, not in silicon). Technically this also applies for the old fixed-function pipeline, of course, but it's worth considering the difference. A raytracing API also isn't forced into the pipeline for every rasterized polygon like old features - hardware T&L, geometry shaders, etc - were. It's something you use on-demand in a compute or fragment shader.