So a lot of people are claiming that Dx12 20% increase is exclusive to PC but I'm not so sure about, granted Spencer was vague in his statements as he always is(great way to get people talking about the XB1 Phil), but I highly doubt that the improvement will be that small for PC. Here's why:




So I ran the current Star Swarm test on my PC and jumped from 20fps(Dx11) to 50fps(Mantle), now like Daious mentioned in the other thread the test is based on best case scenario but I find it hard to believe that in the worse case scenario that PC performance would drop that much. If we use the 980 from the last pic as an example, a 20% increase would only give us 25fps, that can be achieved simply by overclocking the gpu with no need for DX12/Mantle for Amd.
In the video below Wardell talks about DX12 for PC and Xb1 and his statements coincide with what Spencer says about devs having to implement DX 12 meaning there wont be much of a difference on older engines. He uses BF4 as an example, when people claimed Mantle only gave a 20% increase, it was because it was an existing engine. About 17 minutes in is where he talks about it, but this is about PC performance so I'm going to jump to the part where he actually talks about PC.
Don't really have much else to say except that I doubt PC performance will drop to 20% and 50% in the worse case scenario, it would basically mean AMD and MS spent tons of money on API's that achieve the same thing you could achieve by uping your gpu clock.
http://www.anandtech.com/show/8962/the-directx-12-performance-preview-amd-nvidia-star-swarm/5
Log in to comment