DRW: Split ViewProjectionMatrix in order to increase precision

This also removes the need to compute the persmat and saves some memory
from the `ViewInfos` struct. This is needed to allow multiview support.

Initial testing found no major performance regression during vertex
heavy workload.

Test file: {F13610017}

Results:
| Platform | Master | Split Matrix|
| Linux + Mesa + AMD W6600 | 48 fps | 47 fps |
| Macbook Pro M1 | 50 fps | 51 fps |
| Linux + NVidia 1080Ti | 51 fps | 52 fps |
| Linux + Radeon Vega 64 | 25.6 fps | 26.7 fps |

Increased precision when far from origin:
{F13610024}

{F13610025}

Reviewed By: jbakker

Differential Revision: https://developer.blender.org/D16125

1 files changed, 1 insertions, 1 deletions

diff --git a/source/blender/draw/engines/overlay/shaders/overlay_outline_prepass_gpencil_frag.glsl b/source/blender/draw/engines/overlay/shaders/overlay_outline_prepass_gpencil_frag.glsl
index 92be9ec3bcb..a575d42231e 100644
--- a/source/blender/draw/engines/overlay/shaders/overlay_outline_prepass_gpencil_frag.glsl
+++ b/source/blender/draw/engines/overlay/shaders/overlay_outline_prepass_gpencil_frag.glsl
@@ -25,7 +25,7 @@ void main()
     bool is_persp = drw_view.winmat[3][3] == 0.0;
     vec2 uvs = vec2(gl_FragCoord.xy) * drw_view.viewport_size_inverse;
     vec3 pos_ndc = vec3(uvs, gl_FragCoord.z) * 2.0 - 1.0;
-    vec4 pos_world = drw_view.persinv * vec4(pos_ndc, 1.0);
+    vec4 pos_world = drw_view.viewinv * (drw_view.wininv * vec4(pos_ndc, 1.0));
     vec3 pos = pos_world.xyz / pos_world.w;
 
     vec3 ray_ori = pos;