So its performance that should have been there in the first place and some people are like Finewine baby.![]()
I tend to agree tbh but get that AMD maybe didn't have the resources to do anything about it previously and instead focused on mantel/vulkan to try and move things forwards.
Don't know how true it is but I once read that AMD basically stuck rigidly to the OGL spec, so basically just made sure that the drivers could run it and left it at that.
Where as nVidia technically went out of spec and did a load of driver tweaks to do things like making it more multithreaded and h/w accelerated behind the scenes to get more performance.
Which is why the difference is so large in something like Minecraft which is already very limited by raw cpu core speed, it's already core speed limited and the driver overhead kills it.
In the thread over at Guru3d, where this driver leak/info is coming from, one of the posts seems to think AMD are basically writing their own wrapper and passing it through DX11 to take advantage of their recent performance work with that API.