Do you think there's actual future in pathtracing tech? We are unironically hitting physical limits of chips. Making them smaller is no longer the answer because at these scales electrons can quantum-tunnel, creating leaks and compute errors. Pathtracing is very expensive and even software optimizations for volumetrics cannot scale infinitely. Are we just going to run 2000W GPUs? I don't think so. So what's the point of pathtracing and other "next big thing" in the RT, other than curiosity, and maybe RT audio?
>>732275636no because ray tracing will be inefficient compared to whatever small dedicated local models (like DLSS style) that inevitably come out to enable AI enhanced realistic visuals over 3D models in 3D engines.It will be used more for videogame logic than for lighting anymore, like how the last Doom used it for hitboxes / collisions
>>732275636Think about what a single 'ray trace' takes in instructions. It's sustainable at a given limit but is it efficient? Hell no. That is why good software + hardware accel is king.
>>732275636We're hitting the limit of consumer computers.From power draw needed, cooling and size.Everyone likes having their air cooled GPUs that fit in a relatively small case the can put out of the way. Nobody wants a computer with a massive GPU that needs to be water cooled, costs a lot to run and is too big for current size cases.At the end of the day, they're going to have to optimise.Either the developers themselves are going to have to learn to make UE5 run by stripping it apart or UE5 is going to have to be optimised.The only alternative is game streaming but that doesn't scale to the gaming consumer base and even if they could scale, the average person doesn't have good internet and nobody would prefer streaming with the artifacts that come with it over just running it themselves.Pathtracing was just the scam until they realised AI was an easier scam.
there's no future for rat racing
