Originally Posted by TranquilTempest
The entire point of mantle is that it allows low level access to the hardware. This means you can squeeze more performance out of GCN with code that is optimized for that specific architecture. If you make it abstract enough to run the same code on every kind of GPU, you lose the performance advantage that justifies a switch in the first place.
Understood, but if that is truely the case, then AMD should come out and SAY "Mantle will ONLY work for GCN cards."
They have hinted (as well as the developers) that it COULD be extended to Intel and nVidia, as well as older non-GCN AMD cards.
So which is it? CAN it be extended to Intel, nVidia, and non-GCN cards or not?
This is part of my "AMD and the developers aren't telling the whole truth" and why I have adopted a "wait and see" approach about performance.
If it is only going to be on GCN cards ONLY, then that will be only about 10% of the hardware out there, and from a managers point of view, spending lots of time and energy (ie money) on adopting a new standard that is controlled by just one company doesn't make much sense. If Mantle and really only make a difference on 10% of the hardware out there, and that they really have no plans to extend it, then AMD is lying and has pulled the wool over developers eyes, and this is nothing more than a PR campaign.
If Mantle will only be used on GCN, then it's limited, and then statements by devs who are working with it from Chris Roberts just show that they are wishing for something that can never be.
I would love NVidia and Intel to have Mantle drivers (as the API is designed to be non GPU architecture specific) but if not we would support NVidia or Intel drivers that would allow us to get to the metal (GPU Hardware) efficiently and take advantage of parallelism in CPU cores (for efficient batching of data between the game and the GPU).
Bottom line, things aren't adding up.