# Objective
In the Render World, there are a number of collections that are derived
from Main World entities and are used to drive rendering. The most
notable are:
- `VisibleEntities`, which is generated in the `check_visibility` system
and contains visible entities for a view.
- `ExtractedInstances`, which maps entity ids to asset ids.
In the old model, these collections were trivially kept in sync -- any
extracted phase item could look itself up because the render entity id
was guaranteed to always match the corresponding main world id.
After #15320, this became much more complicated, and was leading to a
number of subtle bugs in the Render World. The main rendering systems,
i.e. `queue_material_meshes` and `queue_material2d_meshes`, follow a
similar pattern:
```rust
for visible_entity in visible_entities.iter::<With<Mesh2d>>() {
let Some(mesh_instance) = render_mesh_instances.get_mut(visible_entity) else {
continue;
};
// Look some more stuff up and specialize the pipeline...
let bin_key = Opaque2dBinKey {
pipeline: pipeline_id,
draw_function: draw_opaque_2d,
asset_id: mesh_instance.mesh_asset_id.into(),
material_bind_group_id: material_2d.get_bind_group_id().0,
};
opaque_phase.add(
bin_key,
*visible_entity,
BinnedRenderPhaseType::mesh(mesh_instance.automatic_batching),
);
}
```
In this case, `visible_entities` and `render_mesh_instances` are both
collections that are created and keyed by Main World entity ids, and so
this lookup happens to work by coincidence. However, there is a major
unintentional bug here: namely, because `visible_entities` is a
collection of Main World ids, the phase item being queued is created
with a Main World id rather than its correct Render World id.
This happens to not break mesh rendering because the render commands
used for drawing meshes do not access the `ItemQuery` parameter, but
demonstrates the confusion that is now possible: our UI phase items are
correctly being queued with Render World ids while our meshes aren't.
Additionally, this makes it very easy and error prone to use the wrong
entity id to look up things like assets. For example, if instead we
ignored visibility checks and queued our meshes via a query, we'd have
to be extra careful to use `&MainEntity` instead of the natural
`Entity`.
## Solution
Make all collections that are derived from Main World data use
`MainEntity` as their key, to ensure type safety and avoid accidentally
looking up data with the wrong entity id:
```rust
pub type MainEntityHashMap<V> = hashbrown::HashMap<MainEntity, V, EntityHash>;
```
Additionally, we make all `PhaseItem` be able to provide both their Main
and Render World ids, to allow render phase implementors maximum
flexibility as to what id should be used to look up data.
You can think of this like tracking at the type level whether something
in the Render World should use it's "primary key", i.e. entity id, or
needs to use a foreign key, i.e. `MainEntity`.
## Testing
##### TODO:
This will require extensive testing to make sure things didn't break!
Additionally, some extraction logic has become more complicated and
needs to be checked for regressions.
## Migration Guide
With the advent of the retained render world, collections that contain
references to `Entity` that are extracted into the render world have
been changed to contain `MainEntity` in order to prevent errors where a
render world entity id is used to look up an item by accident. Custom
rendering code may need to be changed to query for `&MainEntity` in
order to look up the correct item from such a collection. Additionally,
users who implement their own extraction logic for collections of main
world entity should strongly consider extracting into a different
collection that uses `MainEntity` as a key.
Additionally, render phases now require specifying both the `Entity` and
`MainEntity` for a given `PhaseItem`. Custom render phases should ensure
`MainEntity` is available when queuing a phase item.
# Objective
- Another step towards #15716
- Remove trait implementations that are dependent on `Handle<T>` being a
`Component`
## Solution
- Remove unused `ExtractComponent` trait implementation for `Handle<T>`
- Remove unused `ExtractInstance` trait implementation for `AssetId`
- Although the `ExtractInstance` trait wasn't used, the `AssetId`s were
being stored inside of `ExtractedInstances` which has an
`ExtractInstance` trait bound on its contents.
I've upgraded the `RenderMaterialInstances` type alias to be its own
resource, identical to `ExtractedInstances<AssetId<M>>` to get around
that with minimal breakage.
## Testing
Tested `many_cubes`, rendering did not explode
# Objective
- Alpha blending can easily fail in many situations and requires sorting
on the cpu
## Solution
- Implement order independent transparency (OIT) as an alternative to
alpha blending
- The implementation uses 2 passes
- The first pass records all the fragments colors and position to a
buffer that is the size of N layers * the render target resolution.
- The second pass sorts the fragments, blends them and draws them to the
screen. It also currently does manual depth testing because early-z
fails in too many cases in the first pass.
## Testing
- We've been using this implementation at foresight in production for
many months now and we haven't had any issues related to OIT.
---
## Showcase
## Future work
- Add an example showing how to use OIT for a custom material
- Next step would be to implement a per-pixel linked list to reduce
memory use
- I'd also like to investigate using a BinnedRenderPhase instead of a
SortedRenderPhase. If it works, it would make the transparent pass
significantly faster.
---------
Co-authored-by: Kristoffer Søholm <k.soeholm@gmail.com>
Co-authored-by: JMS55 <47158642+JMS55@users.noreply.github.com>
Co-authored-by: Charlotte McElwain <charlotte.c.mcelwain@gmail.com>
# Objective
A big step in the migration to required components: meshes and
materials!
## Solution
As per the [selected
proposal](https://hackmd.io/@bevy/required_components/%2Fj9-PnF-2QKK0on1KQ29UWQ):
- Deprecate `MaterialMesh2dBundle`, `MaterialMeshBundle`, and
`PbrBundle`.
- Add `Mesh2d` and `Mesh3d` components, which wrap a `Handle<Mesh>`.
- Add `MeshMaterial2d<M: Material2d>` and `MeshMaterial3d<M: Material>`,
which wrap a `Handle<M>`.
- Meshes *without* a mesh material should be rendered with a default
material. The existence of a material is determined by
`HasMaterial2d`/`HasMaterial3d`, which is required by
`MeshMaterial2d`/`MeshMaterial3d`. This gets around problems with the
generics.
Previously:
```rust
commands.spawn(MaterialMesh2dBundle {
mesh: meshes.add(Circle::new(100.0)).into(),
material: materials.add(Color::srgb(7.5, 0.0, 7.5)),
transform: Transform::from_translation(Vec3::new(-200., 0., 0.)),
..default()
});
```
Now:
```rust
commands.spawn((
Mesh2d(meshes.add(Circle::new(100.0))),
MeshMaterial2d(materials.add(Color::srgb(7.5, 0.0, 7.5))),
Transform::from_translation(Vec3::new(-200., 0., 0.)),
));
```
If the mesh material is missing, previously nothing was rendered. Now,
it renders a white default `ColorMaterial` in 2D and a
`StandardMaterial` in 3D (this can be overridden). Below, only every
other entity has a material:
Why white? This is still open for discussion, but I think white makes
sense for a *default* material, while *invalid* asset handles pointing
to nothing should have something like a pink material to indicate that
something is broken (I don't handle that in this PR yet). This is kind
of a mix of Godot and Unity: Godot just renders a white material for
non-existent materials, while Unity renders nothing when no materials
exist, but renders pink for invalid materials. I can also change the
default material to pink if that is preferable though.
## Testing
I ran some 2D and 3D examples to test if anything changed visually. I
have not tested all examples or features yet however. If anyone wants to
test more extensively, it would be appreciated!
## Implementation Notes
- The relationship between `bevy_render` and `bevy_pbr` is weird here.
`bevy_render` needs `Mesh3d` for its own systems, but `bevy_pbr` has all
of the material logic, and `bevy_render` doesn't depend on it. I feel
like the two crates should be refactored in some way, but I think that's
out of scope for this PR.
- I didn't migrate meshlets to required components yet. That can
probably be done in a follow-up, as this is already a huge PR.
- It is becoming increasingly clear to me that we really, *really* want
to disallow raw asset handles as components. They caused me a *ton* of
headache here already, and it took me a long time to find every place
that queried for them or inserted them directly on entities, since there
were no compiler errors for it. If we don't remove the `Component`
derive, I expect raw asset handles to be a *huge* footgun for users as
we transition to wrapper components, especially as handles as components
have been the norm so far. I personally consider this to be a blocker
for 0.15: we need to migrate to wrapper components for asset handles
everywhere, and remove the `Component` derive. Also see
https://github.com/bevyengine/bevy/issues/14124.
---
## Migration Guide
Asset handles for meshes and mesh materials must now be wrapped in the
`Mesh2d` and `MeshMaterial2d` or `Mesh3d` and `MeshMaterial3d`
components for 2D and 3D respectively. Raw handles as components no
longer render meshes.
Additionally, `MaterialMesh2dBundle`, `MaterialMeshBundle`, and
`PbrBundle` have been deprecated. Instead, use the mesh and material
components directly.
Previously:
```rust
commands.spawn(MaterialMesh2dBundle {
mesh: meshes.add(Circle::new(100.0)).into(),
material: materials.add(Color::srgb(7.5, 0.0, 7.5)),
transform: Transform::from_translation(Vec3::new(-200., 0., 0.)),
..default()
});
```
Now:
```rust
commands.spawn((
Mesh2d(meshes.add(Circle::new(100.0))),
MeshMaterial2d(materials.add(Color::srgb(7.5, 0.0, 7.5))),
Transform::from_translation(Vec3::new(-200., 0., 0.)),
));
```
If the mesh material is missing, a white default material is now used.
Previously, nothing was rendered if the material was missing.
The `WithMesh2d` and `WithMesh3d` query filter type aliases have also
been removed. Simply use `With<Mesh2d>` or `With<Mesh3d>`.
---------
Co-authored-by: Tim Blackbird <justthecooldude@gmail.com>
Co-authored-by: Carter Anderson <mcanders1@gmail.com>
* Save 16 bytes per vertex by calculating tangents in the shader at
runtime, rather than storing them in the vertex data.
* Based on https://jcgt.org/published/0009/03/04,
https://www.jeremyong.com/graphics/2023/12/16/surface-gradient-bump-mapping.
* Fixed visbuffer resolve to use the updated algorithm that flips ddy
correctly
* Added some more docs about meshlet material limitations, and some
TODOs about transforming UV coordinates for the future.
For testing add a normal map to the bunnies with StandardMaterial like
below, and then test that on both main and this PR (make sure to
download the correct bunny for each). Results should be mostly
identical.
```rust
normal_map_texture: Some(asset_server.load_with_settings(
"textures/BlueNoise-Normal.png",
|settings: &mut ImageLoaderSettings| settings.is_srgb = false,
)),
```
# Objective
- Fixes#6370
- Closes#6581
## Solution
- Added the following lints to the workspace:
- `std_instead_of_core`
- `std_instead_of_alloc`
- `alloc_instead_of_core`
- Used `cargo +nightly fmt` with [item level use
formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A)
to split all `use` statements into single items.
- Used `cargo clippy --workspace --all-targets --all-features --fix
--allow-dirty` to _attempt_ to resolve the new linting issues, and
intervened where the lint was unable to resolve the issue automatically
(usually due to needing an `extern crate alloc;` statement in a crate
root).
- Manually removed certain uses of `std` where negative feature gating
prevented `--all-features` from finding the offending uses.
- Used `cargo +nightly fmt` with [crate level use
formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A)
to re-merge all `use` statements matching Bevy's previous styling.
- Manually fixed cases where the `fmt` tool could not re-merge `use`
statements due to conditional compilation attributes.
## Testing
- Ran CI locally
## Migration Guide
The MSRV is now 1.81. Please update to this version or higher.
## Notes
- This is a _massive_ change to try and push through, which is why I've
outlined the semi-automatic steps I used to create this PR, in case this
fails and someone else tries again in the future.
- Making this change has no impact on user code, but does mean Bevy
contributors will be warned to use `core` and `alloc` instead of `std`
where possible.
- This lint is a critical first step towards investigating `no_std`
options for Bevy.
---------
Co-authored-by: François Mockers <francois.mockers@vleue.com>
# Objective
The names of numerous rendering components in Bevy are inconsistent and
a bit confusing. Relevant names include:
- `AutoExposureSettings`
- `AutoExposureSettingsUniform`
- `BloomSettings`
- `BloomUniform` (no `Settings`)
- `BloomPrefilterSettings`
- `ChromaticAberration` (no `Settings`)
- `ContrastAdaptiveSharpeningSettings`
- `DepthOfFieldSettings`
- `DepthOfFieldUniform` (no `Settings`)
- `FogSettings`
- `SmaaSettings`, `Fxaa`, `TemporalAntiAliasSettings` (really
inconsistent??)
- `ScreenSpaceAmbientOcclusionSettings`
- `ScreenSpaceReflectionsSettings`
- `VolumetricFogSettings`
Firstly, there's a lot of inconsistency between `Foo`/`FooSettings` and
`FooUniform`/`FooSettingsUniform` and whether names are abbreviated or
not.
Secondly, the `Settings` post-fix seems unnecessary and a bit confusing
semantically, since it makes it seem like the component is mostly just
auxiliary configuration instead of the core *thing* that actually
enables the feature. This will be an even bigger problem once bundles
like `TemporalAntiAliasBundle` are deprecated in favor of required
components, as users will expect a component named `TemporalAntiAlias`
(or similar), not `TemporalAntiAliasSettings`.
## Solution
Drop the `Settings` post-fix from the component names, and change some
names to be more consistent.
- `AutoExposure`
- `AutoExposureUniform`
- `Bloom`
- `BloomUniform`
- `BloomPrefilter`
- `ChromaticAberration`
- `ContrastAdaptiveSharpening`
- `DepthOfField`
- `DepthOfFieldUniform`
- `DistanceFog`
- `Smaa`, `Fxaa`, `TemporalAntiAliasing` (note: we might want to change
to `Taa`, see "Discussion")
- `ScreenSpaceAmbientOcclusion`
- `ScreenSpaceReflections`
- `VolumetricFog`
I kept the old names as deprecated type aliases to make migration a bit
less painful for users. We should remove them after the next release.
(And let me know if I should just... not add them at all)
I also added some very basic docs for a few types where they were
missing, like on `Fxaa` and `DepthOfField`.
## Discussion
- `TemporalAntiAliasing` is still inconsistent with `Smaa` and `Fxaa`.
Consensus [on
Discord](https://discord.com/channels/691052431525675048/743663924229963868/1280601167209955431)
seemed to be that renaming to `Taa` would probably be fine, but I think
it's a bit more controversial, and it would've required renaming a lot
of related types like `TemporalAntiAliasNode`,
`TemporalAntiAliasBundle`, and `TemporalAntiAliasPlugin`, so I think
it's better to leave to a follow-up.
- I think `Fog` should probably have a more specific name like
`DistanceFog` considering it seems to be distinct from `VolumetricFog`.
~~This should probably be done in a follow-up though, so I just removed
the `Settings` post-fix for now.~~ (done)
---
## Migration Guide
Many rendering components have been renamed for improved consistency and
clarity.
- `AutoExposureSettings` → `AutoExposure`
- `BloomSettings` → `Bloom`
- `BloomPrefilterSettings` → `BloomPrefilter`
- `ContrastAdaptiveSharpeningSettings` → `ContrastAdaptiveSharpening`
- `DepthOfFieldSettings` → `DepthOfField`
- `FogSettings` → `DistanceFog`
- `SmaaSettings` → `Smaa`
- `TemporalAntiAliasSettings` → `TemporalAntiAliasing`
- `ScreenSpaceAmbientOcclusionSettings` → `ScreenSpaceAmbientOcclusion`
- `ScreenSpaceReflectionsSettings` → `ScreenSpaceReflections`
- `VolumetricFogSettings` → `VolumetricFog`
---------
Co-authored-by: Carter Anderson <mcanders1@gmail.com>
# Objective
- Fixes#14974
## Solution
- Replace all* instances of `NonZero*` with `NonZero<*>`
## Testing
- CI passed locally.
---
## Notes
Within the `bevy_reflect` implementations for `std` types,
`impl_reflect_value!()` will continue to use the type aliases instead,
as it inappropriately parses the concrete type parameter as a generic
argument. If the `ZeroablePrimitive` trait was stable, or the macro
could be modified to accept a finite list of types, then we could fully
migrate.
# Objective
- Faster meshlet rasterization path for small triangles
- Avoid having to allocate and write out a triangle buffer
- Refactor gpu_scene.rs
## Solution
- Replace the 32bit visbuffer texture with a 64bit visbuffer buffer,
where the left 32 bits encode depth, and the right 32 bits encode the
existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga
doesn't support atomic ops on textures yet.
- Instead of writing out a buffer of packed cluster + triangle IDs (per
triangle) to raster, the culling pass now writes out a buffer of just
cluster IDs (per cluster, so less memory allocated, cheaper to write
out).
- Clusters for software raster are allocated from the left side
- Clusters for hardware raster are allocated in the same buffer, from
the right side
- The buffer size is fixed at MeshletPlugin build time, and should be
set to a reasonable value for your scene (no warning on overflow, and no
good way to determine what value you need outside of renderdoc - I plan
to fix this in a future PR adding a meshlet stats overlay)
- Currently I don't have a heuristic for software vs hardware raster
selection for each cluster. The existing code is just a placeholder. I
need to profile on a release scene and come up with a heuristic,
probably in a future PR.
- The culling shader is getting pretty hard to follow at this point, but
I don't want to spend time improving it as the entire shader/pass is
getting rewritten/replaced in the near future.
- Software raster is a compute workgroup per-cluster. Each workgroup
loads and transforms the <=64 vertices of the cluster, and then
rasterizes the <=64 triangles of the cluster.
- Two variants are implemented: Scanline for clusters with any larger
triangles (still smaller than hardware is good at), and brute-force for
very very tiny triangles
- Once the shader determines that a pixel should be filled in, it does
an atomicMax() on the visbuffer to store the results, copying how Nanite
works
- On devices with a low max workgroups per dispatch limit, an extra
compute pass is inserted before software raster to convert from a 1d to
2d dispatch (I don't think 3d would ever be necessary).
- I haven't implemented the top-left rule or subpixel precision yet, I'm
leaving that for a future PR since I get usable results without it for
now
- Resources used:
https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters
6-8 of
https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index
- Hardware raster now spawns 64*3 vertex invocations per meshlet,
instead of the actual meshlet vertex count. Extra invocations just
early-exit.
- While this is slower than the existing system, hardware draws should
be rare now that software raster is usable, and it saves a ton of memory
using the unified cluster ID buffer. This would be fixed if wgpu had
support for mesh shaders.
- Instead of writing to a color+depth attachment, the hardware raster
pass also does the same atomic visbuffer writes that software raster
uses.
- We have to bind a dummy render target anyways, as wgpu doesn't
currently support render passes without any attachments
- Material IDs are no longer written out during the main rasterization
passes.
- If we had async compute queues, we could overlap the software and
hardware raster passes.
- New material and depth resolve passes run at the end of the visbuffer
node, and write out view depth and material ID depth textures
### Misc changes
- Fixed cluster culling importing, but never actually using the previous
view uniforms when doing occlusion culling
- Fixed incorrectly adding the LOD error twice when building the meshlet
mesh
- Splitup gpu_scene module into meshlet_mesh_manager, instance_manager,
and resource_manager
- resource_manager is still too complex and inefficient (extract and
prepare are way too expensive). I plan on improving this in a future PR,
but for now ResourceManager is mostly a 1:1 port of the leftover
MeshletGpuScene bits.
- Material draw passes have been renamed to the more accurate material
shade pass, as well as some other misc renaming (in the future, these
will be compute shaders even, and not actual draw calls)
---
## Migration Guide
- TBD (ask me at the end of the release for meshlet changes as a whole)
---------
Co-authored-by: vero <email@atlasdostal.com>
# Objective
Adding more features to `AsBindGroup` proc macro means making the trait
arguments uglier. Downstream implementors of the trait without the proc
macro might want to do different things than our default arguments.
## Solution
Make `AsBindGroup` take an associated `Param` type.
## Migration Guide
`AsBindGroup` now allows the user to specify a `SystemParam` to be used
for creating bind groups.
# Objective
currently if we use an image with the wrong sampler type in a material,
wgpu panics with an invalid texture format. turn this into a warning and
fail more gracefully.
## Solution
the expected sampler type is specified in the AsBindGroup derive, so we
can just check the image sampler is what it should be.
i am not totally sure about the mapping of image sampler type to
#[sampler(type)], i assumed:
```
"filtering" => [ TextureSampleType::Float { filterable: true } ],
"non_filtering" => [
TextureSampleType::Float { filterable: false },
TextureSampleType::Sint,
TextureSampleType::Uint,
],
"comparison" => [ TextureSampleType::Depth ],
```
# Objective
- It's possible to have errors in a draw command, but these errors are
ignored
## Solution
- Return a result with the error
## Changelog
Renamed `RenderCommandResult::Failure` to `RenderCommandResult::Skip`
Added a `reason` string parameter to `RenderCommandResult::Failure`
## Migration Guide
If you were using `RenderCommandResult::Failure` to just ignore an error
and retry later, use `RenderCommandResult::Skip` instead.
This wasn't intentional, but this PR should also help with
https://github.com/bevyengine/bevy/issues/12660 since we can turn a few
unwraps into error messages now.
---------
Co-authored-by: Charlotte McElwain <charlotte.c.mcelwain@gmail.com>
Switches `Msaa` from being a globally configured resource to a per
camera view component.
Closes#7194
# Objective
Allow individual views to describe their own MSAA settings. For example,
when rendering to different windows or to different parts of the same
view.
## Solution
Make `Msaa` a component that is required on all camera bundles.
## Testing
Ran a variety of examples to ensure that nothing broke.
TODO:
- [ ] Make sure android still works per previous comment in
`extract_windows`.
---
## Migration Guide
`Msaa` is no longer configured as a global resource, and should be
specified on each spawned camera if a non-default setting is desired.
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
Co-authored-by: François Mockers <francois.mockers@vleue.com>
This commit uses the [`offset-allocator`] crate to combine vertex and
index arrays from different meshes into single buffers. Since the
primary source of `wgpu` overhead is from validation and synchronization
when switching buffers, this significantly improves Bevy's rendering
performance on many scenes.
This patch is a more flexible version of #13218, which also used slabs.
Unlike #13218, which used slabs of a fixed size, this commit implements
slabs that start small and can grow. In addition to reducing memory
usage, supporting slab growth reduces the number of vertex and index
buffer switches that need to happen during rendering, leading to
improved performance. To prevent pathological fragmentation behavior,
slabs are capped to a maximum size, and mesh arrays that are too large
get their own dedicated slabs.
As an additional improvement over #13218, this commit allows the
application to customize all allocator heuristics. The
`MeshAllocatorSettings` resource contains values that adjust the minimum
and maximum slab sizes, the cutoff point at which meshes get their own
dedicated slabs, and the rate at which slabs grow. Hopefully-sensible
defaults have been chosen for each value.
Unfortunately, WebGL 2 doesn't support the *base vertex* feature, which
is necessary to pack vertex arrays from different meshes into the same
buffer. `wgpu` represents this restriction as the downlevel flag
`BASE_VERTEX`. This patch detects that bit and ensures that all vertex
buffers get dedicated slabs on that platform. Even on WebGL 2, though,
we can combine all *index* arrays into single buffers to reduce buffer
changes, and we do so.
The following measurements are on Bistro:
Overall frame time improves from 8.74 ms to 5.53 ms (1.58x speedup):
Render system time improves from 6.57 ms to 3.54 ms (1.86x speedup):
Opaque pass time improves from 4.64 ms to 2.33 ms (1.99x speedup):
## Migration Guide
### Changed
* Vertex and index buffers for meshes may now be packed alongside other
buffers, for performance.
* `GpuMesh` has been renamed to `RenderMesh`, to reflect the fact that
it no longer directly stores handles to GPU objects.
* Because meshes no longer have their own vertex and index buffers, the
responsibility for the buffers has moved from `GpuMesh` (now called
`RenderMesh`) to the `MeshAllocator` resource. To access the vertex data
for a mesh, use `MeshAllocator::mesh_vertex_slice`. To access the index
data for a mesh, use `MeshAllocator::mesh_index_slice`.
[`offset-allocator`]: https://github.com/pcwalton/offset-allocator
As reported in #14004, many third-party plugins, such as Hanabi, enqueue
entities that don't have meshes into render phases. However, the
introduction of indirect mode added a dependency on mesh-specific data,
breaking this workflow. This is because GPU preprocessing requires that
the render phases manage indirect draw parameters, which don't apply to
objects that aren't meshes. The existing code skips over binned entities
that don't have indirect draw parameters, which causes the rendering to
be skipped for such objects.
To support this workflow, this commit adds a new field,
`non_mesh_items`, to `BinnedRenderPhase`. This field contains a simple
list of (bin key, entity) pairs. After drawing batchable and unbatchable
objects, the non-mesh items are drawn one after another. Bevy itself
doesn't enqueue any items into this list; it exists solely for the
application and/or plugins to use.
Additionally, this commit switches the asset ID in the standard bin keys
to be an untyped asset ID rather than that of a mesh. This allows more
flexibility, allowing bins to be keyed off any type of asset.
This patch adds a new example, `custom_phase_item`, which simultaneously
serves to demonstrate how to use this new feature and to act as a
regression test so this doesn't break again.
Fixes#14004.
## Changelog
### Added
* `BinnedRenderPhase` now contains a `non_mesh_items` field for plugins
to add custom items to.
# Objective
- Fixes#13811 (probably, I lost my test code...)
## Solution
- Turns out that Queue and PrepareAssets are _not_ ordered. We should
probably either rethink our system sets (again), or improve the
documentation here. For reference, I've included the current ordering
below.
- The `prepare_meshlet_meshes_X` systems need to run after
`prepare_assets::<PreparedMaterial<M>>`, and have also been moved to
QueueMeshes.
```rust
schedule.configure_sets(
(
ExtractCommands,
ManageViews,
Queue,
PhaseSort,
Prepare,
Render,
Cleanup,
)
.chain(),
);
schedule.configure_sets((ExtractCommands, PrepareAssets, Prepare).chain());
schedule.configure_sets(QueueMeshes.in_set(Queue).after(prepare_assets::<GpuMesh>));
schedule.configure_sets(
(PrepareResources, PrepareResourcesFlush, PrepareBindGroups)
.chain()
.in_set(Prepare),
);
```
## Testing
- Ambiguity checker to make sure I don't have ambiguous system ordering
This is a revamped equivalent to #9902, though it shares none of the
code. It handles all special cases that I've tested correctly.
The overall technique consists of double-buffering the joint matrix and
morph weights buffers, as most of the previous attempts to solve this
problem did. The process is generally straightforward. Note that, to
avoid regressing the ability of mesh extraction, skin extraction, and
morph target extraction to run in parallel, I had to add a new system to
rendering, `set_mesh_motion_vector_flags`. The comment there explains
the details; it generally runs very quickly.
I've tested this with modified versions of the `animated_fox`,
`morph_targets`, and `many_foxes` examples that add TAA, and the patch
works. To avoid bloating those examples, I didn't add switches for TAA
to them.
Addresses points (1) and (2) of #8423.
## Changelog
### Fixed
* Motion vectors, and therefore TAA, are now supported for meshes with
skins and/or morph targets.
This commit makes us stop using the render world ECS for
`BinnedRenderPhase` and `SortedRenderPhase` and instead use resources
with `EntityHashMap`s inside. There are three reasons to do this:
1. We can use `clear()` to clear out the render phase collections
instead of recreating the components from scratch, allowing us to reuse
allocations.
2. This is a prerequisite for retained bins, because components can't be
retained from frame to frame in the render world, but resources can.
3. We want to move away from storing anything in components in the
render world ECS, and this is a step in that direction.
This patch results in a small performance benefit, due to point (1)
above.
## Changelog
### Changed
* The `BinnedRenderPhase` and `SortedRenderPhase` render world
components have been replaced with `ViewBinnedRenderPhases` and
`ViewSortedRenderPhases` resources.
## Migration Guide
* The `BinnedRenderPhase` and `SortedRenderPhase` render world
components have been replaced with `ViewBinnedRenderPhases` and
`ViewSortedRenderPhases` resources. Instead of querying for the
components, look the camera entity up in the
`ViewBinnedRenderPhases`/`ViewSortedRenderPhases` tables.
Implement visibility ranges, also known as hierarchical levels of detail
(HLODs).
This commit introduces a new component, `VisibilityRange`, which allows
developers to specify camera distances in which meshes are to be shown
and hidden. Hiding meshes happens early in the rendering pipeline, so
this feature can be used for level of detail optimization. Additionally,
this feature is properly evaluated per-view, so different views can show
different levels of detail.
This feature differs from proper mesh LODs, which can be implemented
later. Engines generally implement true mesh LODs later in the pipeline;
they're typically more efficient than HLODs with GPU-driven rendering.
However, mesh LODs are more limited than HLODs, because they require the
lower levels of detail to be meshes with the same vertex layout and
shader (and perhaps the same material) as the original mesh. Games often
want to use objects other than meshes to replace distant models, such as
*octahedral imposters* or *billboard imposters*.
The reason why the feature is called *hierarchical level of detail* is
that HLODs can replace multiple meshes with a single mesh when the
camera is far away. This can be useful for reducing drawcall count. Note
that `VisibilityRange` doesn't automatically propagate down to children;
it must be placed on every mesh.
Crossfading between different levels of detail is supported, using the
standard 4x4 ordered dithering pattern from [1]. The shader code to
compute the dithering patterns should be well-optimized. The dithering
code is only active when visibility ranges are in use for the mesh in
question, so that we don't lose early Z.
Cascaded shadow maps show the HLOD level of the view they're associated
with. Point light and spot light shadow maps, which have no CSMs,
display all HLOD levels that are visible in any view. To support this
efficiently and avoid doing visibility checks multiple times, we
precalculate all visible HLOD levels for each entity with a
`VisibilityRange` during the `check_visibility_range` system.
A new example, `visibility_range`, has been added to the tree, as well
as a new low-poly version of the flight helmet model to go with it. It
demonstrates use of the visibility range feature to provide levels of
detail.
[1]: https://en.wikipedia.org/wiki/Ordered_dithering#Threshold_map
[^1]: Unreal doesn't have a feature that exactly corresponds to
visibility ranges, but Unreal's HLOD system serves roughly the same
purpose.
## Changelog
### Added
* A new `VisibilityRange` component is available to conditionally enable
entity visibility at camera distances, with optional crossfade support.
This can be used to implement different levels of detail (LODs).
## Screenshots
High-poly model:
Low-poly model up close:
Crossfading between the two:
---------
Co-authored-by: Carter Anderson <mcanders1@gmail.com>
This commit implements opt-in GPU frustum culling, built on top of the
infrastructure in https://github.com/bevyengine/bevy/pull/12773. To
enable it on a camera, add the `GpuCulling` component to it. To
additionally disable CPU frustum culling, add the `NoCpuCulling`
component. Note that adding `GpuCulling` without `NoCpuCulling`
*currently* does nothing useful. The reason why `GpuCulling` doesn't
automatically imply `NoCpuCulling` is that I intend to follow this patch
up with GPU two-phase occlusion culling, and CPU frustum culling plus
GPU occlusion culling seems like a very commonly-desired mode.
Adding the `GpuCulling` component to a view puts that view into
*indirect mode*. This mode makes all drawcalls indirect, relying on the
mesh preprocessing shader to allocate instances dynamically. In indirect
mode, the `PreprocessWorkItem` `output_index` points not to a
`MeshUniform` instance slot but instead to a set of `wgpu`
`IndirectParameters`, from which it allocates an instance slot
dynamically if frustum culling succeeds. Batch building has been updated
to allocate and track indirect parameter slots, and the AABBs are now
supplied to the GPU as `MeshCullingData`.
A small amount of code relating to the frustum culling has been borrowed
from meshlets and moved into `maths.wgsl`. Note that standard Bevy
frustum culling uses AABBs, while meshlets use bounding spheres; this
means that not as much code can be shared as one might think.
This patch doesn't provide any way to perform GPU culling on shadow
maps, to avoid making this patch bigger than it already is. That can be
a followup.
## Changelog
### Added
* Frustum culling can now optionally be done on the GPU. To enable it,
add the `GpuCulling` component to a camera.
* To disable CPU frustum culling, add `NoCpuCulling` to a camera. Note
that `GpuCulling` doesn't automatically imply `NoCpuCulling`.
# Objective
- Fixes#12976
## Solution
This one is a doozy.
- Run `cargo +beta clippy --workspace --all-targets --all-features` and
fix all issues
- This includes:
- Moving inner attributes to be outer attributes, when the item in
question has both inner and outer attributes
- Use `ptr::from_ref` in more scenarios
- Extend the valid idents list used by `clippy:doc_markdown` with more
names
- Use `Clone::clone_from` when possible
- Remove redundant `ron` import
- Add backticks to **so many** identifiers and items
- I'm sorry whoever has to review this
---
## Changelog
- Added links to more identifiers in documentation.
[Alpha to coverage] (A2C) replaces alpha blending with a
hardware-specific multisample coverage mask when multisample
antialiasing is in use. It's a simple form of [order-independent
transparency] that relies on MSAA. ["Anti-aliased Alpha Test: The
Esoteric Alpha To Coverage"] is a good summary of the motivation for and
best practices relating to A2C.
This commit implements alpha to coverage support as a new variant for
`AlphaMode`. You can supply `AlphaMode::AlphaToCoverage` as the
`alpha_mode` field in `StandardMaterial` to use it. When in use, the
standard material shader automatically applies the texture filtering
method from ["Anti-aliased Alpha Test: The Esoteric Alpha To Coverage"].
Objects with alpha-to-coverage materials are binned in the opaque pass,
as they're fully order-independent.
The `transparency_3d` example has been updated to feature an object with
alpha to coverage. Happily, the example was already using MSAA.
This is part of #2223, as far as I can tell.
[Alpha to coverage]: https://en.wikipedia.org/wiki/Alpha_to_coverage
[order-independent transparency]:
https://en.wikipedia.org/wiki/Order-independent_transparency
["Anti-aliased Alpha Test: The Esoteric Alpha To Coverage"]:
https://bgolus.medium.com/anti-aliased-alpha-test-the-esoteric-alpha-to-coverage-8b177335ae4f
---
## Changelog
### Added
* The `AlphaMode` enum now supports `AlphaToCoverage`, to provide
limited order-independent transparency when multisample antialiasing is
in use.
This commit splits `VisibleEntities::entities` into four separate lists:
one for lights, one for 2D meshes, one for 3D meshes, and one for UI
elements. This allows `queue_material_meshes` and similar methods to
avoid examining entities that are obviously irrelevant. In particular,
this separation helps scenes with many skinned meshes, as the individual
bones are considered visible entities but have no rendered appearance.
Internally, `VisibleEntities::entities` is a `HashMap` from the `TypeId`
representing a `QueryFilter` to the appropriate `Entity` list. I had to
do this because `VisibleEntities` is located within an upstream crate
from the crates that provide lights (`bevy_pbr`) and 2D meshes
(`bevy_sprite`). As an added benefit, this setup allows apps to provide
their own types of renderable components, by simply adding a specialized
`check_visibility` to the schedule.
This provides a 16.23% end-to-end speedup on `many_foxes` with 10,000
foxes (24.06 ms/frame to 20.70 ms/frame).
## Migration guide
* `check_visibility` and `VisibleEntities` now store the four types of
renderable entities--2D meshes, 3D meshes, lights, and UI
elements--separately. If your custom rendering code examines
`VisibleEntities`, it will now need to specify which type of entity it's
interested in using the `WithMesh2d`, `WithMesh`, `WithLight`, and
`WithNode` types respectively. If your app introduces a new type of
renderable entity, you'll need to add an explicit call to
`check_visibility` to the schedule to accommodate your new component or
components.
## Analysis
`many_foxes`, 10,000 foxes: `main`:
`many_foxes`, 10,000 foxes, this branch:
`queue_material_meshes` (yellow = this branch, red = `main`):
`queue_shadows` (yellow = this branch, red = `main`):
I ported the two existing PCF techniques to the cubemap domain as best I
could. Generally, the technique is to create a 2D orthonormal basis
using Gram-Schmidt normalization, then apply the technique over that
basis. The results look fine, though the shadow bias often needs
adjusting.
For comparison, Unity uses a 4-tap pattern for PCF on point lights of
(1, 1, 1), (-1, -1, 1), (-1, 1, -1), (1, -1, -1). I tried this but
didn't like the look, so I went with the design above, which ports the
2D techniques to the 3D domain. There's surprisingly little material on
point light PCF.
I've gone through every example using point lights and verified that the
shadow maps look fine, adjusting biases as necessary.
Fixes#3628.
---
## Changelog
### Added
* Shadows from point lights now support percentage-closer filtering
(PCF), and as a result look less aliased.
### Changed
* `ShadowFilteringMethod::Castano13` and
`ShadowFilteringMethod::Jimenez14` have been renamed to
`ShadowFilteringMethod::Gaussian` and `ShadowFilteringMethod::Temporal`
respectively.
## Migration Guide
* `ShadowFilteringMethod::Castano13` and
`ShadowFilteringMethod::Jimenez14` have been renamed to
`ShadowFilteringMethod::Gaussian` and `ShadowFilteringMethod::Temporal`
respectively.
Currently, `MeshUniform`s are rather large: 160 bytes. They're also
somewhat expensive to compute, because they involve taking the inverse
of a 3x4 matrix. Finally, if a mesh is present in multiple views, that
mesh will have a separate `MeshUniform` for each and every view, which
is wasteful.
This commit fixes these issues by introducing the concept of a *mesh
input uniform* and adding a *mesh uniform building* compute shader pass.
The `MeshInputUniform` is simply the minimum amount of data needed for
the GPU to compute the full `MeshUniform`. Most of this data is just the
transform and is therefore only 64 bytes. `MeshInputUniform`s are
computed during the *extraction* phase, much like skins are today, in
order to avoid needlessly copying transforms around on CPU. (In fact,
the render app has been changed to only store the translation of each
mesh; it no longer cares about any other part of the transform, which is
stored only on the GPU and the main world.) Before rendering, the
`build_mesh_uniforms` pass runs to expand the `MeshInputUniform`s to the
full `MeshUniform`.
The mesh uniform building pass does the following, all on GPU:
1. Copy the appropriate fields of the `MeshInputUniform` to the
`MeshUniform` slot. If a single mesh is present in multiple views, this
effectively duplicates it into each view.
2. Compute the inverse transpose of the model transform, used for
transforming normals.
3. If applicable, copy the mesh's transform from the previous frame for
TAA. To support this, we double-buffer the `MeshInputUniform`s over two
frames and swap the buffers each frame. The `MeshInputUniform`s for the
current frame contain the index of that mesh's `MeshInputUniform` for
the previous frame.
This commit produces wins in virtually every CPU part of the pipeline:
`extract_meshes`, `queue_material_meshes`,
`batch_and_prepare_render_phase`, and especially
`write_batched_instance_buffer` are all faster. Shrinking the amount of
CPU data that has to be shuffled around speeds up the entire rendering
process.
| Benchmark | This branch | `main` | Speedup |
|------------------------|-------------|---------|---------|
| `many_cubes -nfc` | 17.259 | 24.529 | 42.12% |
| `many_cubes -nfc -vpi` | 302.116 | 312.123 | 3.31% |
| `many_foxes` | 3.227 | 3.515 | 8.92% |
Because mesh uniform building requires compute shader, and WebGL 2 has
no compute shader, the existing CPU mesh uniform building code has been
left as-is. Many types now have both CPU mesh uniform building and GPU
mesh uniform building modes. Developers can opt into the old CPU mesh
uniform building by setting the `use_gpu_uniform_builder` option on
`PbrPlugin` to `false`.
Below are graphs of the CPU portions of `many-cubes
--no-frustum-culling`. Yellow is this branch, red is `main`.
`extract_meshes`:
It's notable that we get a small win even though we're now writing to a
GPU buffer.
`queue_material_meshes`:
There's a bit of a regression here; not sure what's causing it. In any
case it's very outweighed by the other gains.
`batch_and_prepare_render_phase`:
There's a huge win here, enough to make batching basically drop off the
profile.
`write_batched_instance_buffer`:
There's a massive improvement here, as expected. Note that a lot of it
simply comes from the fact that `MeshInputUniform` is `Pod`. (This isn't
a maintainability problem in my view because `MeshInputUniform` is so
simple: just 16 tightly-packed words.)
## Changelog
### Added
* Per-mesh instance data is now generated on GPU with a compute shader
instead of CPU, resulting in rendering performance improvements on
platforms where compute shaders are supported.
## Migration guide
* Custom render phases now need multiple systems beyond just
`batch_and_prepare_render_phase`. Code that was previously creating
custom render phases should now add a `BinnedRenderPhasePlugin` or
`SortedRenderPhasePlugin` as appropriate instead of directly adding
`batch_and_prepare_render_phase`.
# Objective
- Replace `RenderMaterials` / `RenderMaterials2d` / `RenderUiMaterials`
with `RenderAssets` to enable implementing changes to one thing,
`RenderAssets`, that applies to all use cases rather than duplicating
changes everywhere for multiple things that should be one thing.
- Adopts #8149
## Solution
- Make RenderAsset generic over the destination type rather than the
source type as in #8149
- Use `RenderAssets<PreparedMaterial<M>>` etc for render materials
---
## Changelog
- Changed:
- The `RenderAsset` trait is now implemented on the destination type.
Its `SourceAsset` associated type refers to the type of the source
asset.
- `RenderMaterials`, `RenderMaterials2d`, and `RenderUiMaterials` have
been replaced by `RenderAssets<PreparedMaterial<M>>` and similar.
## Migration Guide
- `RenderAsset` is now implemented for the destination type rather that
the source asset type. The source asset type is now the `RenderAsset`
trait's `SourceAsset` associated type.
This commit makes the following optimizations:
## `MeshPipelineKey`/`BaseMeshPipelineKey` split
`MeshPipelineKey` has been split into `BaseMeshPipelineKey`, which lives
in `bevy_render` and `MeshPipelineKey`, which lives in `bevy_pbr`.
Conceptually, `BaseMeshPipelineKey` is a superclass of
`MeshPipelineKey`. For `BaseMeshPipelineKey`, the bits start at the
highest (most significant) bit and grow downward toward the lowest bit;
for `MeshPipelineKey`, the bits start at the lowest bit and grow upward
toward the highest bit. This prevents them from colliding.
The goal of this is to avoid having to reassemble bits of the pipeline
key for every mesh every frame. Instead, we can just use a bitwise or
operation to combine the pieces that make up a `MeshPipelineKey`.
## `specialize_slow`
Previously, all of `specialize()` was marked as `#[inline]`. This
bloated `queue_material_meshes` unnecessarily, as a large chunk of it
ended up being a slow path that was rarely hit. This commit refactors
the function to move the slow path to `specialize_slow()`.
Together, these two changes shave about 5% off `queue_material_meshes`:
## Migration Guide
- The `primitive_topology` field on `GpuMesh` is now an accessor method:
`GpuMesh::primitive_topology()`.
- For performance reasons, `MeshPipelineKey` has been split into
`BaseMeshPipelineKey`, which lives in `bevy_render`, and
`MeshPipelineKey`, which lives in `bevy_pbr`. These two should be
combined with bitwise-or to produce the final `MeshPipelineKey`.
# Objective
This is a necessary precursor to #9122 (this was split from that PR to
reduce the amount of code to review all at once).
Moving `!Send` resource ownership to `App` will make it unambiguously
`!Send`. `SubApp` must be `Send`, so it can't wrap `App`.
## Solution
Refactor `App` and `SubApp` to not have a recursive relationship. Since
`SubApp` no longer wraps `App`, once `!Send` resources are moved out of
`World` and into `App`, `SubApp` will become unambiguously `Send`.
There could be less code duplication between `App` and `SubApp`, but
that would break `App` method chaining.
## Changelog
- `SubApp` no longer wraps `App`.
- `App` fields are no longer publicly accessible.
- `App` can no longer be converted into a `SubApp`.
- Various methods now return references to a `SubApp` instead of an
`App`.
## Migration Guide
- To construct a sub-app, use `SubApp::new()`. `App` can no longer
convert into `SubApp`.
- If you implemented a trait for `App`, you may want to implement it for
`SubApp` as well.
- If you're accessing `app.world` directly, you now have to use
`app.world()` and `app.world_mut()`.
- `App::sub_app` now returns `&SubApp`.
- `App::sub_app_mut` now returns `&mut SubApp`.
- `App::get_sub_app` now returns `Option<&SubApp>.`
- `App::get_sub_app_mut` now returns `Option<&mut SubApp>.`
Today, we sort all entities added to all phases, even the phases that
don't strictly need sorting, such as the opaque and shadow phases. This
results in a performance loss because our `PhaseItem`s are rather large
in memory, so sorting is slow. Additionally, determining the boundaries
of batches is an O(n) process.
This commit makes Bevy instead applicable place phase items into *bins*
keyed by *bin keys*, which have the invariant that everything in the
same bin is potentially batchable. This makes determining batch
boundaries O(1), because everything in the same bin can be batched.
Instead of sorting each entity, we now sort only the bin keys. This
drops the sorting time to near-zero on workloads with few bins like
`many_cubes --no-frustum-culling`. Memory usage is improved too, with
batch boundaries and dynamic indices now implicit instead of explicit.
The improved memory usage results in a significant win even on
unbatchable workloads like `many_cubes --no-frustum-culling
--vary-material-data-per-instance`, presumably due to cache effects.
Not all phases can be binned; some, such as transparent and transmissive
phases, must still be sorted. To handle this, this commit splits
`PhaseItem` into `BinnedPhaseItem` and `SortedPhaseItem`. Most of the
logic that today deals with `PhaseItem`s has been moved to
`SortedPhaseItem`. `BinnedPhaseItem` has the new logic.
Frame time results (in ms/frame) are as follows:
| Benchmark | `binning` | `main` | Speedup |
| ------------------------ | --------- | ------- | ------- |
| `many_cubes -nfc -vpi` | 232.179 | 312.123 | 34.43% |
| `many_cubes -nfc` | 25.874 | 30.117 | 16.40% |
| `many_foxes` | 3.276 | 3.515 | 7.30% |
(`-nfc` is short for `--no-frustum-culling`; `-vpi` is short for
`--vary-per-instance`.)
---
## Changelog
### Changed
* Render phases have been split into binned and sorted phases. Binned
phases, such as the common opaque phase, achieve improved CPU
performance by avoiding the sorting step.
## Migration Guide
- `PhaseItem` has been split into `BinnedPhaseItem` and
`SortedPhaseItem`. If your code has custom `PhaseItem`s, you will need
to migrate them to one of these two types. `SortedPhaseItem` requires
the fewest code changes, but you may want to pick `BinnedPhaseItem` if
your phase doesn't require sorting, as that enables higher performance.
## Tracy graphs
`many-cubes --no-frustum-culling`, `main` branch:
<img width="1064" alt="Screenshot 2024-03-12 180037"
src="https://github.com/bevyengine/bevy/assets/157897/e1180ce8-8e89-46d2-85e3-f59f72109a55">
`many-cubes --no-frustum-culling`, this branch:
<img width="1064" alt="Screenshot 2024-03-12 180011"
src="https://github.com/bevyengine/bevy/assets/157897/0899f036-6075-44c5-a972-44d95895f46c">
You can see that `batch_and_prepare_binned_render_phase` is a much
smaller fraction of the time. Zooming in on that function, with yellow
being this branch and red being `main`, we see:
<img width="1064" alt="Screenshot 2024-03-12 175832"
src="https://github.com/bevyengine/bevy/assets/157897/0dfc8d3f-49f4-496e-8825-a66e64d356d0">
The binning happens in `queue_material_meshes`. Again with yellow being
this branch and red being `main`:
<img width="1064" alt="Screenshot 2024-03-12 175755"
src="https://github.com/bevyengine/bevy/assets/157897/b9b20dc1-11c8-400c-a6cc-1c2e09c1bb96">
We can see that there is a small regression in `queue_material_meshes`
performance, but it's not nearly enough to outweigh the large gains in
`batch_and_prepare_binned_render_phase`.
---------
Co-authored-by: James Liu <contact@jamessliu.com>
# Objective
- Not all materials need shadow, but a queue_shadows system is always
added to the `Render` schedule and executed
## Solution
- Make a setting for shadows, it defaults to true
## Changelog
- Added `shadows_enabled` setting to `MaterialPlugin`
## Migration Guide
- `MaterialPlugin` now has a `shadows_enabled` setting, if you didn't
spawn the plugin using `::default()` or `..default()`, you'll need to
set it. `shadows_enabled: true` is the same behavior as the previous
version, and also the default value.
# Objective
assets that don't load before they get removed are retried forever,
causing buffer churn and slowdown.
## Solution
stop trying to prepare dead assets.
Although we cached hashes of `MeshVertexBufferLayout`, we were paying
the cost of `PartialEq` on `InnerMeshVertexBufferLayout` for every
entity, every frame. This patch changes that logic to place
`MeshVertexBufferLayout`s in `Arc`s so that they can be compared and
hashed by pointer. This results in a 28% speedup in the
`queue_material_meshes` phase of `many_cubes`, with frustum culling
disabled.
Additionally, this patch contains two minor changes:
1. This commit flattens the specialized mesh pipeline cache to one level
of hash tables instead of two. This saves a hash lookup.
2. The example `many_cubes` has been given a `--no-frustum-culling`
flag, to aid in benchmarking.
See the Tracy profile:
<img width="1064" alt="Screenshot 2024-02-29 144406"
src="https://github.com/bevyengine/bevy/assets/157897/18632f1d-1fdd-4ac7-90ed-2d10306b2a1e">
## Migration guide
* Duplicate `MeshVertexBufferLayout`s are now combined into a single
object, `MeshVertexBufferLayoutRef`, which contains an
atomically-reference-counted pointer to the layout. Code that was using
`MeshVertexBufferLayout` may need to be updated to use
`MeshVertexBufferLayoutRef` instead.
# Objective
- As part of the migration process we need to a) see the end effect of
the migration on user ergonomics b) check for serious perf regressions
c) actually migrate the code
- To accomplish this, I'm going to attempt to migrate all of the
remaining user-facing usages of `LegacyColor` in one PR, being careful
to keep a clean commit history.
- Fixes#12056.
## Solution
I've chosen to use the polymorphic `Color` type as our standard
user-facing API.
- [x] Migrate `bevy_gizmos`.
- [x] Take `impl Into<Color>` in all `bevy_gizmos` APIs
- [x] Migrate sprites
- [x] Migrate UI
- [x] Migrate `ColorMaterial`
- [x] Migrate `MaterialMesh2D`
- [x] Migrate fog
- [x] Migrate lights
- [x] Migrate StandardMaterial
- [x] Migrate wireframes
- [x] Migrate clear color
- [x] Migrate text
- [x] Migrate gltf loader
- [x] Register color types for reflection
- [x] Remove `LegacyColor`
- [x] Make sure CI passes
Incidental improvements to ease migration:
- added `Color::srgba_u8`, `Color::srgba_from_array` and friends
- added `set_alpha`, `is_fully_transparent` and `is_fully_opaque` to the
`Alpha` trait
- add and immediately deprecate (lol) `Color::rgb` and friends in favor
of more explicit and consistent `Color::srgb`
- standardized on white and black for most example text colors
- added vector field traits to `LinearRgba`: ~~`Add`, `Sub`,
`AddAssign`, `SubAssign`,~~ `Mul<f32>` and `Div<f32>`. Multiplications
and divisions do not scale alpha. `Add` and `Sub` have been cut from
this PR.
- added `LinearRgba` and `Srgba` `RED/GREEN/BLUE`
- added `LinearRgba_to_f32_array` and `LinearRgba::to_u32`
## Migration Guide
Bevy's color types have changed! Wherever you used a
`bevy::render::Color`, a `bevy::color::Color` is used instead.
These are quite similar! Both are enums storing a color in a specific
color space (or to be more precise, using a specific color model).
However, each of the different color models now has its own type.
TODO...
- `Color::rgba`, `Color::rgb`, `Color::rbga_u8`, `Color::rgb_u8`,
`Color::rgb_from_array` are now `Color::srgba`, `Color::srgb`,
`Color::srgba_u8`, `Color::srgb_u8` and `Color::srgb_from_array`.
- `Color::set_a` and `Color::a` is now `Color::set_alpha` and
`Color::alpha`. These are part of the `Alpha` trait in `bevy_color`.
- `Color::is_fully_transparent` is now part of the `Alpha` trait in
`bevy_color`
- `Color::r`, `Color::set_r`, `Color::with_r` and the equivalents for
`g`, `b` `h`, `s` and `l` have been removed due to causing silent
relatively expensive conversions. Convert your `Color` into the desired
color space, perform your operations there, and then convert it back
into a polymorphic `Color` enum.
- `Color::hex` is now `Srgba::hex`. Call `.into` or construct a
`Color::Srgba` variant manually to convert it.
- `WireframeMaterial`, `ExtractedUiNode`, `ExtractedDirectionalLight`,
`ExtractedPointLight`, `ExtractedSpotLight` and `ExtractedSprite` now
store a `LinearRgba`, rather than a polymorphic `Color`
- `Color::rgb_linear` and `Color::rgba_linear` are now
`Color::linear_rgb` and `Color::linear_rgba`
- The various CSS color constants are no longer stored directly on
`Color`. Instead, they're defined in the `Srgba` color space, and
accessed via `bevy::color::palettes::css`. Call `.into()` on them to
convert them into a `Color` for quick debugging use, and consider using
the much prettier `tailwind` palette for prototyping.
- The `LIME_GREEN` color has been renamed to `LIMEGREEN` to comply with
the standard naming.
- Vector field arithmetic operations on `Color` (add, subtract, multiply
and divide by a f32) have been removed. Instead, convert your colors
into `LinearRgba` space, and perform your operations explicitly there.
This is particularly relevant when working with emissive or HDR colors,
whose color channel values are routinely outside of the ordinary 0 to 1
range.
- `Color::as_linear_rgba_f32` has been removed. Call
`LinearRgba::to_f32_array` instead, converting if needed.
- `Color::as_linear_rgba_u32` has been removed. Call
`LinearRgba::to_u32` instead, converting if needed.
- Several other color conversion methods to transform LCH or HSL colors
into float arrays or `Vec` types have been removed. Please reimplement
these externally or open a PR to re-add them if you found them
particularly useful.
- Various methods on `Color` such as `rgb` or `hsl` to convert the color
into a specific color space have been removed. Convert into
`LinearRgba`, then to the color space of your choice.
- Various implicitly-converting color value methods on `Color` such as
`r`, `g`, `b` or `h` have been removed. Please convert it into the color
space of your choice, then check these properties.
- `Color` no longer implements `AsBindGroup`. Store a `LinearRgba`
internally instead to avoid conversion costs.
---------
Co-authored-by: Alice Cecile <alice.i.cecil@gmail.com>
Co-authored-by: Afonso Lage <lage.afonso@gmail.com>
Co-authored-by: Rob Parrett <robparrett@gmail.com>
Co-authored-by: Zachary Harrold <zac@harrold.com.au>
# Objective
- followup to https://github.com/bevyengine/bevy/pull/11671
- I forgot to change the alpha masked phases.
## Solution
- Change the sorting for alpha mask phases to sort by pipeline+mesh
instead of distance, for much better batching for alpha masked
materials.
I also fixed some docs that I missed in the previous PR.
---
## Changelog
- Alpha masked materials are now sorted by pipeline and mesh.
# Objective
- Add the new `-Zcheck-cfg` checks to catch more warnings
- Fixes#12091
## Solution
- Create a new `cfg-check` to the CI that runs `cargo check -Zcheck-cfg
--workspace` using cargo nightly (and fails if there are warnings)
- Fix all warnings generated by the new check
---
## Changelog
- Remove all redundant imports
- Fix cfg wasm32 targets
- Add 3 dead code exceptions (should StandardColor be unused?)
- Convert ios_simulator to a feature (I'm not sure if this is the right
way to do it, but the check complained before)
## Migration Guide
No breaking changes
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
# Objective
The migration process for `bevy_color` (#12013) will be fairly involved:
there will be hundreds of affected files, and a large number of APIs.
## Solution
To allow us to proceed granularly, we're going to keep both
`bevy_color::Color` (new) and `bevy_render::Color` (old) around until
the migration is complete.
However, simply doing this directly is confusing! They're both called
`Color`, making it very hard to tell when a portion of the code has been
ported.
As discussed in #12056, by renaming the old `Color` type, we can make it
easier to gradually migrate over, one API at a time.
## Migration Guide
THIS MIGRATION GUIDE INTENTIONALLY LEFT BLANK.
This change should not be shipped to end users: delete this section in
the final migration guide!
---------
Co-authored-by: Alice Cecile <alice.i.cecil@gmail.com>
# Objective
- Some properties of public types are private but sometimes it's useful
to be able to set those
## Solution
- Make more stuff pub
---
## Changelog
- `MaterialBindGroupId` internal id is now pub and added a new()
constructor
- `ExtractedPointLight` and `ExtractedDirectionalLight` properties are
now all pub
---------
Co-authored-by: James Liu <contact@jamessliu.com>
This fixes a `FIXME` in `extract_meshes` and results in a performance
improvement.
As a result of this change, meshes in the render world might not be
attached to entities anymore. Therefore, the `entity` parameter to
`RenderCommand::render()` is now wrapped in an `Option`. Most
applications that use the render app's ECS can simply unwrap the
`Option`.
Note that for now sprites, gizmos, and UI elements still use the render
world as usual.
## Migration guide
* For efficiency reasons, some meshes in the render world may not have
corresponding `Entity` IDs anymore. As a result, the `entity` parameter
to `RenderCommand::render()` is now wrapped in an `Option`. Custom
rendering code may need to be updated to handle the case in which no
`Entity` exists for an object that is to be rendered.
# Objective
Bevy could benefit from *irradiance volumes*, also known as *voxel
global illumination* or simply as light probes (though this term is not
preferred, as multiple techniques can be called light probes).
Irradiance volumes are a form of baked global illumination; they work by
sampling the light at the centers of each voxel within a cuboid. At
runtime, the voxels surrounding the fragment center are sampled and
interpolated to produce indirect diffuse illumination.
## Solution
This is divided into two sections. The first is copied and pasted from
the irradiance volume module documentation and describes the technique.
The second part consists of notes on the implementation.
### Overview
An *irradiance volume* is a cuboid voxel region consisting of
regularly-spaced precomputed samples of diffuse indirect light. They're
ideal if you have a dynamic object such as a character that can move
about
static non-moving geometry such as a level in a game, and you want that
dynamic object to be affected by the light bouncing off that static
geometry.
To use irradiance volumes, you need to precompute, or *bake*, the
indirect
light in your scene. Bevy doesn't currently come with a way to do this.
Fortunately, [Blender] provides a [baking tool] as part of the Eevee
renderer, and its irradiance volumes are compatible with those used by
Bevy.
The [`bevy-baked-gi`] project provides a tool, `export-blender-gi`, that
can
extract the baked irradiance volumes from the Blender `.blend` file and
package them up into a `.ktx2` texture for use by the engine. See the
documentation in the `bevy-baked-gi` project for more details as to this
workflow.
Like all light probes in Bevy, irradiance volumes are 1×1×1 cubes that
can
be arbitrarily scaled, rotated, and positioned in a scene with the
[`bevy_transform::components::Transform`] component. The 3D voxel grid
will
be stretched to fill the interior of the cube, and the illumination from
the
irradiance volume will apply to all fragments within that bounding
region.
Bevy's irradiance volumes are based on Valve's [*ambient cubes*] as used
in
*Half-Life 2* ([Mitchell 2006], slide 27). These encode a single color
of
light from the six 3D cardinal directions and blend the sides together
according to the surface normal.
The primary reason for choosing ambient cubes is to match Blender, so
that
its Eevee renderer can be used for baking. However, they also have some
advantages over the common second-order spherical harmonics approach:
ambient cubes don't suffer from ringing artifacts, they are smaller (6
colors for ambient cubes as opposed to 9 for spherical harmonics), and
evaluation is faster. A smaller basis allows for a denser grid of voxels
with the same storage requirements.
If you wish to use a tool other than `export-blender-gi` to produce the
irradiance volumes, you'll need to pack the irradiance volumes in the
following format. The irradiance volume of resolution *(Rx, Ry, Rz)* is
expected to be a 3D texture of dimensions *(Rx, 2Ry, 3Rz)*. The
unnormalized
texture coordinate *(s, t, p)* of the voxel at coordinate *(x, y, z)*
with
side *S* ∈ *{-X, +X, -Y, +Y, -Z, +Z}* is as follows:
```text
s = x
t = y + ⎰ 0 if S ∈ {-X, -Y, -Z}
⎱ Ry if S ∈ {+X, +Y, +Z}
⎧ 0 if S ∈ {-X, +X}
p = z + ⎨ Rz if S ∈ {-Y, +Y}
⎩ 2Rz if S ∈ {-Z, +Z}
```
Visually, in a left-handed coordinate system with Y up, viewed from the
right, the 3D texture looks like a stacked series of voxel grids, one
for
each cube side, in this order:
| **+X** | **+Y** | **+Z** |
| ------ | ------ | ------ |
| **-X** | **-Y** | **-Z** |
A terminology note: Other engines may refer to irradiance volumes as
*voxel
global illumination*, *VXGI*, or simply as *light probes*. Sometimes
*light
probe* refers to what Bevy calls a reflection probe. In Bevy, *light
probe*
is a generic term that encompasses all cuboid bounding regions that
capture
indirect illumination, whether based on voxels or not.
Note that, if binding arrays aren't supported (e.g. on WebGPU or WebGL
2),
then only the closest irradiance volume to the view will be taken into
account during rendering.
[*ambient cubes*]:
https://advances.realtimerendering.com/s2006/Mitchell-ShadingInValvesSourceEngine.pdf
[Mitchell 2006]:
https://advances.realtimerendering.com/s2006/Mitchell-ShadingInValvesSourceEngine.pdf
[Blender]: http://blender.org/
[baking tool]:
https://docs.blender.org/manual/en/latest/render/eevee/render_settings/indirect_lighting.html
[`bevy-baked-gi`]: https://github.com/pcwalton/bevy-baked-gi
### Implementation notes
This patch generalizes light probes so as to reuse as much code as
possible between irradiance volumes and the existing reflection probes.
This approach was chosen because both techniques share numerous
similarities:
1. Both irradiance volumes and reflection probes are cuboid bounding
regions.
2. Both are responsible for providing baked indirect light.
3. Both techniques involve presenting a variable number of textures to
the shader from which indirect light is sampled. (In the current
implementation, this uses binding arrays.)
4. Both irradiance volumes and reflection probes require gathering and
sorting probes by distance on CPU.
5. Both techniques require the GPU to search through a list of bounding
regions.
6. Both will eventually want to have falloff so that we can smoothly
blend as objects enter and exit the probes' influence ranges. (This is
not implemented yet to keep this patch relatively small and reviewable.)
To do this, we generalize most of the methods in the reflection probes
patch #11366 to be generic over a trait, `LightProbeComponent`. This
trait is implemented by both `EnvironmentMapLight` (for reflection
probes) and `IrradianceVolume` (for irradiance volumes). Using a trait
will allow us to add more types of light probes in the future. In
particular, I highly suspect we will want real-time reflection planes
for mirrors in the future, which can be easily slotted into this
framework.
## Changelog
> This section is optional. If this was a trivial fix, or has no
externally-visible impact, you can delete this section.
### Added
* A new `IrradianceVolume` asset type is available for baked voxelized
light probes. You can bake the global illumination using Blender or
another tool of your choice and use it in Bevy to apply indirect
illumination to dynamic objects.
# Objective
Bevy does ridiculous amount of drawcalls, and our batching isn't very
effective because we sort by distance and only batch if we get multiple
of the same object in a row. This can give us slightly better GPU
performance when not using the depth prepass (due to less overdraw), but
ends up being massively CPU bottlenecked due to doing thousands of
unnecessary drawcalls.
## Solution
Change the sort functions to sort by pipeline key then by mesh id for
large performance gains in more realistic scenes than our stress tests.
Pipelines changed:
- Opaque3d
- Opaque3dDeferred
- Opaque3dPrepass
---
## Changelog
- Opaque3d drawing order is now sorted by pipeline and mesh, rather than
by distance. This trades off a bit of GPU time in exchange for massively
better batching in scenes that aren't only drawing huge amounts of a
single object.
# Objective
TypeUuid is deprecated, remove it.
## Migration Guide
Convert any uses of `#[derive(TypeUuid)]` with `#[derive(TypePath]` for
more complex uses see the relevant
[documentation](https://docs.rs/bevy/latest/bevy/prelude/trait.TypePath.html)
for more information.
---------
Co-authored-by: ebola <dev@axiomatic>
# Objective
> Can anyone explain to me the reasoning of renaming all the types named
Query to Data. I'm talking about this PR
https://github.com/bevyengine/bevy/pull/10779 It doesn't make sense to
me that a bunch of types that are used to run queries aren't named Query
anymore. Like ViewQuery on the ViewNode is the type of the Query. I
don't really understand the point of the rename, it just seems like it
hides the fact that a query will run based on those types.
[@IceSentry](https://discord.com/channels/691052431525675048/692572690833473578/1184946251431694387)
## Solution
Revert several renames in #10779.
## Changelog
- `ViewNode::ViewData` is now `ViewNode::ViewQuery` again.
## Migration Guide
- This PR amends the migration guide in
https://github.com/bevyengine/bevy/pull/10779
---------
Co-authored-by: atlas dostal <rodol@rivalrebels.com>
This pull request re-submits #10057, which was backed out for breaking
macOS, iOS, and Android. I've tested this version on macOS and Android
and on the iOS simulator.
# Objective
This pull request implements *reflection probes*, which generalize
environment maps to allow for multiple environment maps in the same
scene, each of which has an axis-aligned bounding box. This is a
standard feature of physically-based renderers and was inspired by [the
corresponding feature in Blender's Eevee renderer].
## Solution
This is a minimal implementation of reflection probes that allows
artists to define cuboid bounding regions associated with environment
maps. For every view, on every frame, a system builds up a list of the
nearest 4 reflection probes that are within the view's frustum and
supplies that list to the shader. The PBR fragment shader searches
through the list, finds the first containing reflection probe, and uses
it for indirect lighting, falling back to the view's environment map if
none is found. Both forward and deferred renderers are fully supported.
A reflection probe is an entity with a pair of components, *LightProbe*
and *EnvironmentMapLight* (as well as the standard *SpatialBundle*, to
position it in the world). The *LightProbe* component (along with the
*Transform*) defines the bounding region, while the
*EnvironmentMapLight* component specifies the associated diffuse and
specular cubemaps.
A frequent question is "why two components instead of just one?" The
advantages of this setup are:
1. It's readily extensible to other types of light probes, in particular
*irradiance volumes* (also known as ambient cubes or voxel global
illumination), which use the same approach of bounding cuboids. With a
single component that applies to both reflection probes and irradiance
volumes, we can share the logic that implements falloff and blending
between multiple light probes between both of those features.
2. It reduces duplication between the existing *EnvironmentMapLight* and
these new reflection probes. Systems can treat environment maps attached
to cameras the same way they treat environment maps applied to
reflection probes if they wish.
Internally, we gather up all environment maps in the scene and place
them in a cubemap array. At present, this means that all environment
maps must have the same size, mipmap count, and texture format. A
warning is emitted if this restriction is violated. We could potentially
relax this in the future as part of the automatic mipmap generation
work, which could easily do texture format conversion as part of its
preprocessing.
An easy way to generate reflection probe cubemaps is to bake them in
Blender and use the `export-blender-gi` tool that's part of the
[`bevy-baked-gi`] project. This tool takes a `.blend` file containing
baked cubemaps as input and exports cubemap images, pre-filtered with an
embedded fork of the [glTF IBL Sampler], alongside a corresponding
`.scn.ron` file that the scene spawner can use to recreate the
reflection probes.
Note that this is intentionally a minimal implementation, to aid
reviewability. Known issues are:
* Reflection probes are basically unsupported on WebGL 2, because WebGL
2 has no cubemap arrays. (Strictly speaking, you can have precisely one
reflection probe in the scene if you have no other cubemaps anywhere,
but this isn't very useful.)
* Reflection probes have no falloff, so reflections will abruptly change
when objects move from one bounding region to another.
* As mentioned before, all cubemaps in the world of a given type
(diffuse or specular) must have the same size, format, and mipmap count.
Future work includes:
* Blending between multiple reflection probes.
* A falloff/fade-out region so that reflected objects disappear
gradually instead of vanishing all at once.
* Irradiance volumes for voxel-based global illumination. This should
reuse much of the reflection probe logic, as they're both GI techniques
based on cuboid bounding regions.
* Support for WebGL 2, by breaking batches when reflection probes are
used.
These issues notwithstanding, I think it's best to land this with
roughly the current set of functionality, because this patch is useful
as is and adding everything above would make the pull request
significantly larger and harder to review.
---
## Changelog
### Added
* A new *LightProbe* component is available that specifies a bounding
region that an *EnvironmentMapLight* applies to. The combination of a
*LightProbe* and an *EnvironmentMapLight* offers *reflection probe*
functionality similar to that available in other engines.
[the corresponding feature in Blender's Eevee renderer]:
https://docs.blender.org/manual/en/latest/render/eevee/light_probes/reflection_cubemaps.html
[`bevy-baked-gi`]: https://github.com/pcwalton/bevy-baked-gi
[glTF IBL Sampler]: https://github.com/KhronosGroup/glTF-IBL-Sampler
