If adding low latency RAM is relatively expensive, I could see where AMD and Nvidia would have to avoid using it in a core consumer/prosumer architecture because most of their products need to be scaleable to various price points. WHile it could make sense in there $600 offering it just wouldnt be feasible at $200 or $100, for example.
Also, would the driver maintenance for them increase dramatically to optimize for such different architectures simultaneously?
Also, would the driver maintenance for them increase dramatically to optimize for such different architectures simultaneously?