Skip to content

freedreno: Supports 16-bit float type variables and operations

Hyunjun Ko requested to merge zzoon/mesa:mediump-freedreno into master

This MR is following MR !1043 (closed) and !1044 (closed). As the type of variables and operations are lowered, also needs to handle them on freedreno. This includes a debug option to enable precision lowering at GLSL IR.

Edited by Emma Anholt

Merge request reports