A lot of what makes Nanite possible isn't available 10 years ago and that's how compute is integrated into the 3D pipeline. There were tricks then but now it's supported by design. As you can see in the breakdown a whole lot of the pipeline is done in compute.
Parts of this pipeline have been used since 2015. You can see most of the techniques explained in here https://www.advances.realtimerendering.com/s2015/aaltonenhaa...
. What nanite does is to continue the development of this by combining it with Visibility Buffers (decouples rasterization from materials) and adds incredibly fancy LOD system and rasterization for small triangles.
The architecture may be possible in the sense that feature-wise you have the pieces (or most of them). But resources like memory, bandwidth and gpu speed, would fall way below the thresholds and tradeoffs needed to make this useful and not just "possible".
Re:A Macro View of Nanite
by HellDunkel on Monday May 31, 2021 @06:36PM
(#27348837)
GPU memory has been a limiting factor but it also took photogrammetry to capture models of sufficient detail to justify the effort.
As far as I can tell this architecture was possible at least starting from the introduction of DX11 hardware more than 10 years ago, and it seems to be the most straightforward way of rendering a realistic 3D scene.