Retain RenderMeshInstance and MeshInputUniform data from frame to…

… frame. (#16385)

This commit moves the front end of the rendering pipeline to a retained
model when GPU preprocessing is in use (i.e. by default, except in
constrained environments). `RenderMeshInstance` and `MeshUniformData`
are stored from frame to frame and are updated only for the entities
that changed state. This was rather tricky and requires some careful
surgery to keep the data valid in the case of removals.

This patch is built on top of Bevy's change detection. Generally, this
worked, except that `ViewVisibility` isn't currently properly tracked.
Therefore, this commit adds proper change tracking for `ViewVisibility`.
Doing this required adding a new system that runs after all
`check_visibility` invocations, as no single `check_visibility`
invocation has enough global information to detect changes.

On the Bistro exterior scene, with all textures forced to opaque, this
patch improves steady-state `extract_meshes_for_gpu_building` from
93.8us to 34.5us and steady-state `collect_meshes_for_gpu_building` from
195.7us to 4.28us. Altogether this constitutes an improvement from 290us
to 38us, which is a 7.46x speedup.

![Screenshot 2024-11-13
143841](https://github.com/user-attachments/assets/40b1aacc-373d-4016-b7fd-b0284bc33de4)

![Screenshot 2024-11-13
143850](https://github.com/user-attachments/assets/53b401c3-7461-43b3-918b-cff89ea780d6)

This patch is only lightly tested and shouldn't land before 0.15 is
released anyway, so I'm releasing it as a draft.

crates/bevy_pbr/src/lib.rs

@@ -384,7 +384,8 @@ impl Plugin for PbrPlugin {
                         .in_set(SimulationLightSystems::AssignLightsToClusters)
                         .after(TransformSystem::TransformPropagate)
                         .after(VisibilitySystems::CheckVisibility)
-                        .after(CameraUpdateSystem),
+                        .after(CameraUpdateSystem)
+                        .ambiguous_with(VisibilitySystems::VisibilityChangeDetect),
                     clear_directional_light_cascades
                         .in_set(SimulationLightSystems::UpdateDirectionalLightCascades)
                         .after(TransformSystem::TransformPropagate)
@@ -398,7 +399,8 @@ impl Plugin for PbrPlugin {
                         // We assume that no entity will be both a directional light and a spot light,
                         // so these systems will run independently of one another.
                         // FIXME: Add an archetype invariant for this https://github.com/bevyengine/bevy/issues/1481.
-                        .ambiguous_with(update_spot_light_frusta),
+                        .ambiguous_with(update_spot_light_frusta)
+                        .ambiguous_with(VisibilitySystems::VisibilityChangeDetect),
                     update_point_light_frusta
                         .in_set(SimulationLightSystems::UpdateLightFrusta)
                         .after(TransformSystem::TransformPropagate)
@@ -419,7 +421,8 @@ impl Plugin for PbrPlugin {
                         // NOTE: This MUST be scheduled AFTER the core renderer visibility check
                         // because that resets entity `ViewVisibility` for the first view
                         // which would override any results from this otherwise
-                        .after(VisibilitySystems::CheckVisibility),
+                        .after(VisibilitySystems::CheckVisibility)
+                        .ambiguous_with(VisibilitySystems::VisibilityChangeDetect),
                 ),
             );
 

crates/bevy_pbr/src/render/gpu_preprocess.rs

@@ -404,8 +404,8 @@ pub fn prepare_preprocess_bind_groups(
     } = batched_instance_buffers.into_inner();
 
     let (Some(current_input_buffer), Some(previous_input_buffer), Some(data_buffer)) = (
-        current_input_buffer_vec.buffer(),
-        previous_input_buffer_vec.buffer(),
+        current_input_buffer_vec.buffer.buffer(),
+        previous_input_buffer_vec.buffer.buffer(),
         data_buffer_vec.buffer(),
     ) else {
         return;
@@ -483,5 +483,4 @@ pub fn write_mesh_culling_data_buffer(
     mut mesh_culling_data_buffer: ResMut<MeshCullingDataBuffer>,
 ) {
     mesh_culling_data_buffer.write_buffer(&render_device, &render_queue);
-    mesh_culling_data_buffer.clear();
 }