The Core i7-6700K and Core i3-6100 delivered similar performance using the GTX 980 Ti and the crazy quality settings with around 53fps. Despite DX12 removing the CPU overhead, the FX-8350 was still unable to match the Core i3-6100 though it did gain around 9% performance under DX12.
Reducing the quality settings to medium we find some intriguing results indeed. Here the Core i7-6700K races ahead of the Core i3-6100 delivering 20% more performance when using DX12 and almost 40% more performance when using DX11. Using DX12 didn’t help with the 6700K and this is why we didn’t see much of a difference in our GPU tests between the two APIs. The Core i3-6100 enjoyed a 16% performance bump when moving from DX11 to DX12 though despite the advantages of DX12 it was still considerably slower than the 6700K. Interestingly, the FX-8350 was still unable to match the Core i3 even with the help of DX12. In fact, when using DX12 the FX-8350 was still slower than the Core i3-6100 using DX11. Very disappointing results for AMD.
Moving to 1440p with the ‘crazy’ quality and 2xMSAA settings we see similar performance trends to what was seen at 1080p.
Again at 1440p on medium quality we see similar performance trends to 1080p and amazingly, even at this resolution the FX-8350 trails the Core i3-6100.
The R9 Fury X results are extremely interesting as they highlight AMD’s driver overhead issue under DX11. While not an issue for the efficient Skylake processors we see that the FX-8350 gets hammered in the DX11 mode while this bottleneck is completely removed under DX12.
Reducing the visual quality preset to medium with the R9 Fury X we find an apparent bottleneck of just over 60fps using DX11. This is AMD’s driver overhead rearing its ugly head once more. Notice that while the Skylake processors can gain a little over 30fps compared to the crazy DX11 quality settings, the FX-8350 becomes just 3fps faster. Again, enabling DX12 does help but even so the FX-8350 using DX12 is still slower than the Skylake Core i3-6100 using DX11.
