exdeath on 12 April 2017
| mine said: Much more. As a common x86 CPU isn't able to drive the GPU as efficient. Even DX12 which features multi-threading for this doesn't help PCs: To come close to understand what this new instruction set means, compare those vector instructions like SSE. Those gave huge performance gains. Moving DX12 driver stuff into silicon gives even more gains – as fewer instructions on the CPU side means less use of precious CPU cache. And as the GPU is even earlier ready to compute, this means that the GPU has more time to render and the CPU has more time for the game logic. Lets hope that Microsoft didn't patent it a way no other can use it too. That stuff is perfect for the next version of the Nintendo Switch to use less power for the same graphics... BTW: Microsoft has a big advantage. They could sell the same console power for less money then Sony. From Scorpio on for the next 25 years... . |
You know how I know you aren't a programmer or electrical engineer?
