I don't think it's just wild fantasy. Koduri does have experience and knowledge from Hybrid and typical CrossFire being in AMD. As shown above, in another post, DirectX 12 supports the tech from 2015, so people working with GPUs probably work with that idea for years. For reasons unknown to us, they didn't moved forward to make it available to us. But that's in my opinion decisions made from marketing and financial divitions, not technical problems. If we also consider that much of DirectX 12 is Mantle, maybe multi-adapter was part of Mantle, or it was suppose to be a feature for Mantle 2, 3 or something, if there where more versions of Mantle. Maybe a replacement for the original CrossFire back when CrossFire and SLi where still important. But then Nvidia decided to slowly kill SLi and AMD probably had no problem to follow Nvidia in that decision.
DirectX 12 allows developers more access to the underlying hardware aspects of GPUs, so it is possible to, for example, explicitly do task 1 on GPU1, after that's done, copy the result to GPU2 and use that result to do task 2. So yes, you could say all that was made possible since Mantle. A lot of that control wasn't there in DirectX 11 and before, SLI/CF just merged multiple GPUs into one logical GPU. The problem, to make it work, and especially to make it work well (significantly better than origical CF and SLI), it becomes a lot of work for developers. Just too much work for too little gain.
I don't think it's just wild fantasy. Koduri does have experience and knowledge from Hybrid and typical CrossFire being in AMD. As shown above, in another post, DirectX 12 supports the tech from 2015, so people working with GPUs probably work with that idea for years. For reasons unknown to us, they didn't moved forward to make it available to us. But that's in my opinion decisions made from marketing and financial divisions, not technical problems. If we also consider that much of DirectX 12 is Mantle, maybe multi-adapter was part of Mantle
This I think hits the nail in the head but you fail to draw the right conclusion. If its not a technical problem, its an economical one, and there is absolutely zero reason right now to make it viable all of a sudden. DX12's mGPU bit is up to devs and it still is even if Intel offloads parts of it to its IGP. It won't happen because devs will need to implement it, your perfect proof of that is the current demise of SLI and in somewhat lesser degree, Crossfire. And the few percent you can win on an IGP in absolute performance is a complete waste of time, especially because discrete GPUs get stronger much faster than IGPs. By the time you're done implementing it you're already looking at a cheap discrete GPU that can dwarf its performance for 30 bucks. Those product stacks jump up in perf with 15-30% on a yearly or bi yearly basis. IGP's are still stuck struggling with medium 1080p graphics, hopefully at something above 15 fps.
This I think hits the nail in the head but you fail to draw the right conclusion. If its not a technical problem, its an economical one, and there is absolutely zero reason right now to make it viable all of a sudden. DX12's mGPU bit is up to devs and it still is even if Intel offloads parts of it to its IGP. It won't happen because devs will need to implement it, your perfect proof of that is the current demise of SLI and in somewhat lesser degree, Crossfire. And the few percent you can win on an IGP in absolute performance is a complete waste of time, especially because discrete GPUs get stronger much faster than IGPs. By the time you're done implementing it you're already looking at a cheap discrete GPU that can dwarf its performance for 30 bucks. Those product stacks jump up in perf with 15-30% on a yearly or bi yearly basis. IGP's are still stuck struggling with medium 1080p graphics, hopefully at something above 15 fps.
The economical problems I have in mind is more like "If we keep (for example) SLI, we will have to support it in drivers and we will end up selling two GTX 2070s instead of one 2080 Ti". Well, Intel doesn't have this dilemma. This could end up as their competing option against SLI and Crossfire and Nvidia and AMD killing those two, doesn't mean than Intel will follow. At least not yet, not as long as it is - logically- (far) behind. So, what starts as a combination of an iGPU and a discrete GPU could evolve into combining two discrete GPUs. And the fact that someone would be able to combine two Intel GPUs from different market segments, a low end and a mid range for example, or even different generation discrete GPUs - that was the whole idea in the first place anyway - could offer Intel an advantage.
The economical problems I have in mind is more like "If we keep (for example) SLI, we will have to support it in drivers and we will end up selling two GTX 2070s instead of one 2080 Ti". Well, Intel doesn't have this dilemma. This could end up as their competing option against SLI and Crossfire and Nvidia and AMD killing those two, doesn't mean than Intel will follow. At least not yet, not as long as it is - logically- (far) behind. So, what starts as a combination of an iGPU and a discrete GPU could evolve into combining two discrete GPUs. And the fact that someone would be able to combine two Intel GPUs from different market segments, a low end and a mid range for example, or even different generation discrete GPUs - that was the whole idea in the first place anyway - could offer Intel an advantage.
We've seen that before too, when Raja told us we could put two RX480's in Crossfire to match Nvidia's performance at some point. Didn't quite work out. But you have point regardless, people do already have the IGP in many cases. This hinges entirely on Intel's ability to keep broad support intact, I'm not holding my breath.
