Originally Posted by looniam
well yeah and that's why AMD went to johan andersson @DICE to help develop mantle. did anyone see anyone from AMD at conferences when presenting mantle?
NO. they saw and heard johan; it was only later that robert hallok started working with other developers to incorporate mantle in their game engines.
*sigh* its always the same in vulkan/dx12 discussions; it always degenerates into another mantle history lesson . .
so yeah on topic -
great. crytek going to have a bunch of free to play games with vulkan and dx12 . . . yay. well more implementation would hopefully lead to better performance.
their beta implementation is 6 months and several patches
old. one can always check the steam discussions for the improvements since it was released as some post their benchmarks each patch/driver such as:
You think mantle was the first name. They had some hefty "mantle" options already in micro code for the die. You don't just all of a sudden have hardware do an unintended function. Many of the base mantle functions are done because the gpu can, which means they were already planning on something along those lines.
So AMD had ideas, that weren't in an API but they were already developing them. And that's what you would do if a new API is to be pushed. You develop hardware that can do what the API needs first. Then I'd build a driver base, all of this taking a lot of time. Then you start hashing out an API. Without hardware to test, what good is a theoretical API?