Skip to content

v0.12.1

Choose a tag to compare

View all tags
@faisalkindi faisalkindi released this 08 Sep 13:49
· 3 commits to main since this release
v0.12.1
6ee6b11

Added

  • A model-resolution dial on the OptiScaler route. OptiScaler's [DlssNr] WorkingScale decides what fraction of the frame the DLSS 5 model works at — the frame itself is never reduced, only the model's own work is done small and enlarged, and its cost falls with the square of this. That makes it the single biggest performance lever available on that route.

    Choose 100%, 75% (about half the cost) or 50% (about a quarter) on the Setup page; it is written into OptiScaler.ini at install.

    This is the same idea as the Neural Upstream option added in v0.11.20, from the other direction: Upstream moves the model before the upscaler on the ReShade route, while this shrinks what it processes on the OptiScaler route. Both attack the fact that the neural pass runs at output resolution and saturates whatever device it lands on.

Assets 3
Loading