It is funny when Fottemberg from Semiaccurate forum in, If I remember correctly, April last year wrote that there will be storm over forums, about DirectX12, and Nvidia GPUs. Looks like he had very good source of this information. It was also him who said that DirectX 12.1 feature level was added just because Nvidia specifically asked for it to Microsoft. What he implied is that it was more "political" decision, not "technical".
I do not believe he is wrong on this topic, also.
End off-top.
continuing with the "slightly OT discussion, NVidia fans have been recently clamoring about Just Cause 3 adding CR (Conservative Rasterization), however.. what they seem to ignore is that this was added with Intel's work, who supports Teir 3 of Conservative Rasterization, where as NV only supports Teir 1, in effect NV is like DX 12.05 where as Intel is the most complete DX part to date (DX12.1). This could actually hamper NV more than it benefits where as Intel iGPUs have the potential to get significant gains. Intel gains MUCH more than NV vs NV vs AMD.. (1/2 pixel coverage vs intels 1/256 pixelx , NV also lacks Post-snap degenerate triangles and has no Inner input coverage) for greater details see
https://msdn.microsoft.com/en-us/library/windows/desktop/dn903791(v=vs.85).aspx
from MSDN: (primary benefits
bolded)
NV-->
Tier 1 enforces a maximum 1/2 pixel uncertainty region and does not support post-snap degenerates. This is good for tiled rendering, a
texture atlas, light map generation and sub-pixel shadow maps.
Intel ->
Tier 2 reduces the maximum uncertainty region to 1/256 and requires post-snap degenerates not be culled. This tier is helpful for
CPU-based algorithm acceleration (such as
voxelization).
• Tier 3 maintains a maximum 1/256 uncertainty region and adds support for inner input coverage. Inner input coverage adds the new value SV_InnerCoverage to High Level Shading Language (HLSL). This is a 32-bit scalar integer that can be specified on input to a pixel shader, and represents the underestimated Conservative Rasterization information (that is, whether a pixel is guaranteed-to-be-fully covered). This tier is helpful for
occlusion culling
In effect NV only gains minimal perf from tex altas (think lookup), light maps and shadow maps (notice NV as of late PUSHING shadows a lot in last GW titles..!) where as they do NOT get the more substantial performance benefits from occlusion culling or efficient use of voxels ..
Edit video:
https://youtu.be/zL0oSY_YmDY?t=137