Skip to content

nvk: update limits for Kepler/Maxwell, drop a stale todo

Emma Anholt requested to merge anholt/mesa:nvk-k-limits into nvk/main

Just a bit of looking into why Kepler might be in rough shape.

Looks like K/M are probably also 256 UBO/TBO alignment instead of 64 -- are we doing something special where the lower alignment might apply? @gfxstrand since I saw you were doing some maxwell stuff.

Merge request reports