# Objective
- Progress towards #19887.
## Solution
- For cases that don't need to conditionally add systems, we can just
replace FromWorld impls with systems and then add those systems to
`RenderStartup`.
## Testing
- I ran the `lightmaps`, `reflection_probes`, `deferred_rendering`,
`volumetric_fog`, and `wireframe` examples.
A few versions ago, wgpu made it possible to set shader entry point to
`None`, which will select the correct entry point in file where only a
single entrypoint is specified. This makes it possible to implement
`Default` for pipeline descriptors. This PR does so and attempts to
`..default()` everything possible.
# Objective
- PrepassPipelineInternal used to exist to optimize compile time and
binary size when PrepassPipeline was generic over the material.
- After #19667, PrepassPipeline is no longer generic!
## Solution
- Flatten all the fields of `PrepassPipelineInternal` into
`PrepassPipeline`.
Previously, the specialize/queue systems were added per-material and the
plugin prepass/shadow enable flags controlled whether we added those
systems. Now, we make this a property of the material instance and check
for it when specializing. Fixes
https://github.com/bevyengine/bevy/issues/19850.
# Objective
- MaterialProperties uses HashMap for some data that is generally going
to be really small. This is likely using more memory than necessary
## Solution
- Use a SmallVec instead
- I used the size a StandardMaterial would need for all the backing
arrays
## Testing
- Tested the 3d_scene to confirm it still works
## Notes
I'm not sure if it made a measurable difference since I'm not sure how
to measure this. It's a bit hard to create an artificial workflow where
this would be the main bottleneck. This is very in the realm of
microoptimization.
# Objective
Closes#18075
In order to enable a number of patterns for dynamic materials in the
engine, it's necessary to decouple the renderer from the `Material`
trait.
This opens the possibility for:
- Materials that aren't coupled to `AsBindGroup`.
- 2d using the underlying 3d bindless infrastructure.
- Dynamic materials that can change their layout at runtime.
- Materials that aren't even backed by a Rust struct at all.
## Solution
In short, remove all trait bounds from render world material systems and
resources. This means moving a bunch of stuff onto `MaterialProperties`
and engaging in some hacks to make specialization work. Rather than
storing the bind group data in `MaterialBindGroupAllocator`, right now
we're storing it in a closure on `MaterialProperties`. TBD if this has
bad performance characteristics.
## Benchmarks
- `many_cubes`:
`cargo run --example many_cubes --release --features=bevy/trace_tracy --
--vary-material-data-per-instance`:
- @DGriffin91's Caldera
`cargo run --release --features=bevy/trace_tracy -- --random-materials`
- @DGriffin91's Caldera with 20 unique material types (i.e.
`MaterialPlugin<M>`) and random materials per mesh
`cargo run --release --features=bevy/trace_tracy -- --random-materials`
### TODO
- We almost certainly lost some parallelization from removing the type
params that could be gained back from smarter iteration.
- Test all the things that could have broken.
- ~Fix meshlets~
## Showcase
See [the
example](https://github.com/bevyengine/bevy/pull/19667/files#diff-9d768cfe1c3aa81eff365d250d3cbe5a63e8df63e81dd85f64c3c3cd993f6d94)
for a custom material implemented without the use of the `Material`
trait and thus `AsBindGroup`.
---------
Co-authored-by: IceSentry <IceSentry@users.noreply.github.com>
Co-authored-by: IceSentry <c.giguere42@gmail.com>
# Objective
Upgrade to `wgpu` version `25.0`.
Depends on https://github.com/bevyengine/naga_oil/pull/121
## Solution
### Problem
The biggest issue we face upgrading is the following requirement:
> To facilitate this change, there was an additional validation rule put
in place: if there is a binding array in a bind group, you may not use
dynamic offset buffers or uniform buffers in that bind group. This
requirement comes from vulkan rules on UpdateAfterBind descriptors.
This is a major difficulty for us, as there are a number of binding
arrays that are used in the view bind group. Note, this requirement does
not affect merely uniform buffors that use dynamic offset but the use of
*any* uniform in a bind group that also has a binding array.
### Attempted fixes
The easiest fix would be to change uniforms to be storage buffers
whenever binding arrays are in use:
```wgsl
#ifdef BINDING_ARRAYS_ARE_USED
@group(0) @binding(0) var<uniform> view: View;
@group(0) @binding(1) var<uniform> lights: types::Lights;
#else
@group(0) @binding(0) var<storage> view: array<View>;
@group(0) @binding(1) var<storage> lights: array<types::Lights>;
#endif
```
This requires passing the view index to the shader so that we know where
to index into the buffer:
```wgsl
struct PushConstants {
view_index: u32,
}
var<push_constant> push_constants: PushConstants;
```
Using push constants is no problem because binding arrays are only
usable on native anyway.
However, this greatly complicates the ability to access `view` in
shaders. For example:
```wgsl
#ifdef BINDING_ARRAYS_ARE_USED
mesh_view_bindings::view.view_from_world[0].z
#else
mesh_view_bindings::view[mesh_view_bindings::view_index].view_from_world[0].z
#endif
```
Using this approach would work but would have the effect of polluting
our shaders with ifdef spam basically *everywhere*.
Why not use a function? Unfortunately, the following is not valid wgsl
as it returns a binding directly from a function in the uniform path.
```wgsl
fn get_view() -> View {
#if BINDING_ARRAYS_ARE_USED
let view_index = push_constants.view_index;
let view = views[view_index];
#endif
return view;
}
```
This also poses problems for things like lights where we want to return
a ptr to the light data. Returning ptrs from wgsl functions isn't
allowed even if both bindings were buffers.
The next attempt was to simply use indexed buffers everywhere, in both
the binding array and non binding array path. This would be viable if
push constants were available everywhere to pass the view index, but
unfortunately they are not available on webgpu. This means either
passing the view index in a storage buffer (not ideal for such a small
amount of state) or using push constants sometimes and uniform buffers
only on webgpu. However, this kind of conditional layout infects
absolutely everything.
Even if we were to accept just using storage buffer for the view index,
there's also the additional problem that some dynamic offsets aren't
actually per-view but per-use of a setting on a camera, which would
require passing that uniform data on *every* camera regardless of
whether that rendering feature is being used, which is also gross.
As such, although it's gross, the simplest solution just to bump binding
arrays into `@group(1)` and all other bindings up one bind group. This
should still bring us under the device limit of 4 for most users.
### Next steps / looking towards the future
I'd like to avoid needing split our view bind group into multiple parts.
In the future, if `wgpu` were to add `@builtin(draw_index)`, we could
build a list of draw state in gpu processing and avoid the need for any
kind of state change at all (see
https://github.com/gfx-rs/wgpu/issues/6823). This would also provide
significantly more flexibility to handle things like offsets into other
arrays that may not be per-view.
### Testing
Tested a number of examples, there are probably more that are still
broken.
---------
Co-authored-by: François Mockers <mockersf@gmail.com>
Co-authored-by: Elabajaba <Elabajaba@users.noreply.github.com>
# Objective
- compute_matrix doesn't compute anything, it just puts an Affine3A into
a Mat4. the name is inaccurate
## Solution
- rename it to conform with to_isometry (which, ironically, does compute
a decomposition which is rather expensive)
## Testing
- Its a rename. If it compiles, its good to go
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
# Objective
- Related to #19024
## Solution
- Use the new `load_shader_library` macro for the shader libraries and
`embedded_asset`/`load_embedded_asset` for the "shader binaries" in
`bevy_pbr` (excluding meshlets).
## Testing
- `atmosphere` example still works
- `fog` example still works
- `decal` example still works
P.S. I don't think this needs a migration guide. Technically users could
be using the `pub` weak handles, but there's no actual good use for
them, so omitting it seems fine. Alternatively, we could mix this in
with the migration guide notes for #19137.
# Objective
Fixes#19150
## Solution
Normally the `validate_cached_entity` in
86cc02dca2/crates/bevy_pbr/src/prepass/mod.rs (L1109-L1126)
marks unchanged entites as clean, which makes them remain in the phase.
If a material is changed to an `alpha_mode` that isn't supposed to be
added to the prepass pipeline, the specialization system just
`continue`s and doesn't indicate to the cache that the entity is not
clean anymore.
I made these invalid entities get removed from the pipeline cache so
that they are correctly not marked clean and then removed from the
phase.
## Testing
Tested with the example code from the issue.
# Objective
Fixes a part of #14274.
Bevy has an incredibly inconsistent naming convention for its system
sets, both internally and across the ecosystem.
<img alt="System sets in Bevy"
src="https://github.com/user-attachments/assets/d16e2027-793f-4ba4-9cc9-e780b14a5a1b"
width="450" />
*Names of public system set types in Bevy*
Most Bevy types use a naming of `FooSystem` or just `Foo`, but there are
also a few `FooSystems` and `FooSet` types. In ecosystem crates on the
other hand, `FooSet` is perhaps the most commonly used name in general.
Conventions being so wildly inconsistent can make it harder for users to
pick names for their own types, to search for system sets on docs.rs, or
to even discern which types *are* system sets.
To reign in the inconsistency a bit and help unify the ecosystem, it
would be good to establish a common recommended naming convention for
system sets in Bevy itself, similar to how plugins are commonly suffixed
with `Plugin` (ex: `TimePlugin`). By adopting a consistent naming
convention in first-party Bevy, we can softly nudge ecosystem crates to
follow suit (for types where it makes sense to do so).
Choosing a naming convention is also relevant now, as the [`bevy_cli`
recently adopted
lints](https://github.com/TheBevyFlock/bevy_cli/pull/345) to enforce
naming for plugins and system sets, and the recommended naming used for
system sets is still a bit open.
## Which Name To Use?
Now the contentious part: what naming convention should we actually
adopt?
This was discussed on the Bevy Discord at the end of last year, starting
[here](<https://discord.com/channels/691052431525675048/692572690833473578/1310659954683936789>).
`FooSet` and `FooSystems` were the clear favorites, with `FooSet` very
narrowly winning an unofficial poll. However, it seems to me like the
consensus was broadly moving towards `FooSystems` at the end and after
the poll, with Cart
([source](https://discord.com/channels/691052431525675048/692572690833473578/1311140204974706708))
and later Alice
([source](https://discord.com/channels/691052431525675048/692572690833473578/1311092530732859533))
and also me being in favor of it.
Let's do a quick pros and cons list! Of course these are just what I
thought of, so take it with a grain of salt.
`FooSet`:
- Pro: Nice and short!
- Pro: Used by many ecosystem crates.
- Pro: The `Set` suffix comes directly from the trait name `SystemSet`.
- Pro: Pairs nicely with existing APIs like `in_set` and
`configure_sets`.
- Con: `Set` by itself doesn't actually indicate that it's related to
systems *at all*, apart from the implemented trait. A set of what?
- Con: Is `FooSet` a set of `Foo`s or a system set related to `Foo`? Ex:
`ContactSet`, `MeshSet`, `EnemySet`...
`FooSystems`:
- Pro: Very clearly indicates that the type represents a collection of
systems. The actual core concept, system(s), is in the name.
- Pro: Parallels nicely with `FooPlugins` for plugin groups.
- Pro: Low risk of conflicts with other names or misunderstandings about
what the type is.
- Pro: In most cases, reads *very* nicely and clearly. Ex:
`PhysicsSystems` and `AnimationSystems` as opposed to `PhysicsSet` and
`AnimationSet`.
- Pro: Easy to search for on docs.rs.
- Con: Usually results in longer names.
- Con: Not yet as widely used.
Really the big problem with `FooSet` is that it doesn't actually
describe what it is. It describes what *kind of thing* it is (a set of
something), but not *what it is a set of*, unless you know the type or
check its docs or implemented traits. `FooSystems` on the other hand is
much more self-descriptive in this regard, at the cost of being a bit
longer to type.
Ultimately, in some ways it comes down to preference and how you think
of system sets. Personally, I was originally in favor of `FooSet`, but
have been increasingly on the side of `FooSystems`, especially after
seeing what the new names would actually look like in Avian and now
Bevy. I prefer it because it usually reads better, is much more clearly
related to groups of systems than `FooSet`, and overall *feels* more
correct and natural to me in the long term.
For these reasons, and because Alice and Cart also seemed to share a
preference for it when it was previously being discussed, I propose that
we adopt a `FooSystems` naming convention where applicable.
## Solution
Rename Bevy's system set types to use a consistent `FooSet` naming where
applicable.
- `AccessibilitySystem` → `AccessibilitySystems`
- `GizmoRenderSystem` → `GizmoRenderSystems`
- `PickSet` → `PickingSystems`
- `RunFixedMainLoopSystem` → `RunFixedMainLoopSystems`
- `TransformSystem` → `TransformSystems`
- `RemoteSet` → `RemoteSystems`
- `RenderSet` → `RenderSystems`
- `SpriteSystem` → `SpriteSystems`
- `StateTransitionSteps` → `StateTransitionSystems`
- `RenderUiSystem` → `RenderUiSystems`
- `UiSystem` → `UiSystems`
- `Animation` → `AnimationSystems`
- `AssetEvents` → `AssetEventSystems`
- `TrackAssets` → `AssetTrackingSystems`
- `UpdateGizmoMeshes` → `GizmoMeshSystems`
- `InputSystem` → `InputSystems`
- `InputFocusSet` → `InputFocusSystems`
- `ExtractMaterialsSet` → `MaterialExtractionSystems`
- `ExtractMeshesSet` → `MeshExtractionSystems`
- `RumbleSystem` → `RumbleSystems`
- `CameraUpdateSystem` → `CameraUpdateSystems`
- `ExtractAssetsSet` → `AssetExtractionSystems`
- `Update2dText` → `Text2dUpdateSystems`
- `TimeSystem` → `TimeSystems`
- `AudioPlaySet` → `AudioPlaybackSystems`
- `SendEvents` → `EventSenderSystems`
- `EventUpdates` → `EventUpdateSystems`
A lot of the names got slightly longer, but they are also a lot more
consistent, and in my opinion the majority of them read much better. For
a few of the names I took the liberty of rewording things a bit;
definitely open to any further naming improvements.
There are still also cases where the `FooSystems` naming doesn't really
make sense, and those I left alone. This primarily includes system sets
like `Interned<dyn SystemSet>`, `EnterSchedules<S>`, `ExitSchedules<S>`,
or `TransitionSchedules<S>`, where the type has some special purpose and
semantics.
## Todo
- [x] Should I keep all the old names as deprecated type aliases? I can
do this, but to avoid wasting work I'd prefer to first reach consensus
on whether these renames are even desired.
- [x] Migration guide
- [x] Release notes
Fixes#18809Fixes#18823
Meshes despawned in `Last` can still be in visisible entities if they
were visible as of `PostUpdate`. Sanity check that the mesh actually
exists before we specialize. We still want to unconditionally assume
that the entity is in `EntitySpecializationTicks` as its absence from
that cache would likely suggest another bug.
# Objective
The goal of `bevy_platform_support` is to provide a set of platform
agnostic APIs, alongside platform-specific functionality. This is a high
traffic crate (providing things like HashMap and Instant). Especially in
light of https://github.com/bevyengine/bevy/discussions/18799, it
deserves a friendlier / shorter name.
Given that it hasn't had a full release yet, getting this change in
before Bevy 0.16 makes sense.
## Solution
- Rename `bevy_platform_support` to `bevy_platform`.
Currently, `RenderMaterialInstances` and `RenderMeshMaterialIds` are
very similar render-world resources: the former maps main world meshes
to typed material asset IDs, and the latter maps main world meshes to
untyped material asset IDs. This is needlessly-complex and wasteful, so
this patch unifies the two in favor of a single untyped
`RenderMaterialInstances` resource.
This patch also fixes a subtle issue that could cause mesh materials to
be incorrect if a `MeshMaterial3d<A>` was removed and replaced with a
`MeshMaterial3d<B>` material in the same frame. The problematic pattern
looks like:
1. `extract_mesh_materials<B>` runs and, seeing the
`Changed<MeshMaterial3d<B>>` condition, adds an entry mapping the mesh
to the new material to the untyped `RenderMeshMaterialIds`.
2. `extract_mesh_materials<A>` runs and, seeing that the entity is
present in `RemovedComponents<MeshMaterial3d<A>>`, removes the entry
from `RenderMeshMaterialIds`.
3. The material slot is now empty, and the mesh will show up as whatever
material happens to be in slot 0 in the material data slab.
This commit fixes the issue by splitting out `extract_mesh_materials`
into *three* phases: *extraction*, *early sweeping*, and *late
sweeping*, which run in that order:
1. The *extraction* system, which runs for each material, updates
`RenderMaterialInstances` records whenever `MeshMaterial3d` components
change, and updates a change tick so that the following system will know
not to remove it.
2. The *early sweeping* system, which runs for each material, processes
entities present in `RemovedComponents<MeshMaterial3d>` and removes each
such entity's record from `RenderMeshInstances` only if the extraction
system didn't update it this frame. This system runs after *all*
extraction systems have completed, fixing the race condition.
3. The *late sweeping* system, which runs only once regardless of the
number of materials in the scene, processes entities present in
`RemovedComponents<ViewVisibility>` and, as in the early sweeping phase,
removes each such entity's record from `RenderMeshInstances` only if the
extraction system didn't update it this frame. At the end, the late
sweeping system updates the change tick.
Because this pattern happens relatively frequently, I think this PR
should land for 0.16.
## Objective
Fix motion blur not working on skinned meshes.
## Solution
`set_mesh_motion_vector_flags` can set
`RenderMeshInstanceFlags::HAS_PREVIOUS_SKIN` after specialization has
already cached the material. This can lead to
`MeshPipelineKey::HAS_PREVIOUS_SKIN` never getting set, disabling motion
blur.
The fix is to make sure `set_mesh_motion_vector_flags` happens before
specialization.
Note that the bug is fixed in a different way by #18074, which includes
other fixes but is a much larger change.
## Testing
Open the `animated_mesh` example and add these components to the
`Camera3d` entity:
```rust
MotionBlur {
shutter_angle: 5.0,
samples: 2,
#[cfg(all(feature = "webgl2", target_arch = "wasm32", not(feature = "webgpu")))]
_webgl2_padding: Default::default(),
},
#[cfg(all(feature = "webgl2", target_arch = "wasm32", not(feature = "webgpu")))]
Msaa::Off,
```
Tested on `animated_mesh`, `many_foxes`, `custom_skinned_mesh`,
Win10/Nvidia with Vulkan, WebGL/Chrome, WebGPU/Chrome. Note that testing
`many_foxes` WebGL requires #18715.
# Objective
- The prepass pipeline has a generic bound on the specialize function
but 95% of it doesn't need it
## Solution
- Move most of the fields to an internal struct and use a separate
specialize function for those fields
## Testing
- Ran the 3d_scene and it worked like before
---
## Migration Guide
If you were using a field of the `PrepassPipeline`, most of them have
now been move to `PrepassPipeline::internal`.
## Notes
Here's the cargo bloat size comparison (from this tool
https://github.com/bevyengine/bevy/discussions/14864):
```
before:
(
"<bevy_pbr::prepass::PrepassPipeline<M> as bevy_render::render_resource::pipeline_specializer::SpecializedMeshPipeline>::specialize",
25416,
0.05582993,
),
after:
(
"<bevy_pbr::prepass::PrepassPipeline<M> as bevy_render::render_resource::pipeline_specializer::SpecializedMeshPipeline>::specialize",
2496,
0.005490916,
),
(
"bevy_pbr::prepass::PrepassPipelineInternal::specialize",
11444,
0.025175499,
),
```
The size for the specialize function that is generic is now much
smaller, so users won't need to recompile it for every material.
# Objective
For materials that aren't being used or a visible entity doesn't have an
instance of, we were unnecessarily constantly checking whether they
needed specialization, saying yes (because the material had never been
specialized for that entity), and failing to look up the material
instance.
## Solution
If an entity doesn't have an instance of the material, it can't possibly
need specialization, so exit early before spending time doing the check.
Fixes#18388.
# Objective
Allow prepass to run without ATTRIBUTE_NORMAL.
This is needed for custom materials with non-standard vertex attributes.
For example a voxel material with manually packed vertex data.
Fixes#13054.
This PR covers the first part of the **stale** PR #13569 to only focus
on fixing #13054.
## Solution
- Only push normals `vertex_attributes` when the layout contains
`Mesh::ATTRIBUTE_NORMAL`
## Testing
- Did you test these changes? If so, how?
**Tested the fix on my own project with a mesh without normal
attribute.**
- Are there any parts that need more testing?
**I don't think so.**
- How can other people (reviewers) test your changes? Is there anything
specific they need to know?
**Prepass should not be blocked on a mesh without normal attributes
(with or without custom material).**
- If relevant, what platforms did you test these changes on, and are
there any important ones you can't test?
**Probably irrelevant, but Windows/Vulkan.**
Currently, Bevy's implementation of bindless resources is rather
unusual: every binding in an object that implements `AsBindGroup` (most
commonly, a material) becomes its own separate binding array in the
shader. This is inefficient for two reasons:
1. If multiple materials reference the same texture or other resource,
the reference to that resource will be duplicated many times. This
increases `wgpu` validation overhead.
2. It creates many unused binding array slots. This increases `wgpu` and
driver overhead and makes it easier to hit limits on APIs that `wgpu`
currently imposes tight resource limits on, like Metal.
This PR fixes these issues by switching Bevy to use the standard
approach in GPU-driven renderers, in which resources are de-duplicated
and passed as global arrays, one for each type of resource.
Along the way, this patch introduces per-platform resource limits and
bumps them from 16 resources per binding array to 64 resources per bind
group on Metal and 2048 resources per bind group on other platforms.
(Note that the number of resources per *binding array* isn't the same as
the number of resources per *bind group*; as it currently stands, if all
the PBR features are turned on, Bevy could pack as many as 496 resources
into a single slab.) The limits have been increased because `wgpu` now
has universal support for partially-bound binding arrays, which mean
that we no longer need to fill the binding arrays with fallback
resources on Direct3D 12. The `#[bindless(LIMIT)]` declaration when
deriving `AsBindGroup` can now simply be written `#[bindless]` in order
to have Bevy choose a default limit size for the current platform.
Custom limits are still available with the new
`#[bindless(limit(LIMIT))]` syntax: e.g. `#[bindless(limit(8))]`.
The material bind group allocator has been completely rewritten. Now
there are two allocators: one for bindless materials and one for
non-bindless materials. The new non-bindless material allocator simply
maintains a 1:1 mapping from material to bind group. The new bindless
material allocator maintains a list of slabs and allocates materials
into slabs on a first-fit basis. This unfortunately makes its
performance O(number of resources per object * number of slabs), but the
number of slabs is likely to be low, and it's planned to become even
lower in the future with `wgpu` improvements. Resources are
de-duplicated with in a slab and reference counted. So, for instance, if
multiple materials refer to the same texture, that texture will exist
only once in the appropriate binding array.
To support these new features, this patch adds the concept of a
*bindless descriptor* to the `AsBindGroup` trait. The bindless
descriptor allows the material bind group allocator to probe the layout
of the material, now that an array of `BindGroupLayoutEntry` records is
insufficient to describe the group. The `#[derive(AsBindGroup)]` has
been heavily modified to support the new features. The most important
user-facing change to that macro is that the struct-level `uniform`
attribute, `#[uniform(BINDING_NUMBER, StandardMaterial)]`, now reads
`#[uniform(BINDLESS_INDEX, MATERIAL_UNIFORM_TYPE,
binding_array(BINDING_NUMBER)]`, allowing the material to specify the
binding number for the binding array that holds the uniform data.
To make this patch simpler, I removed support for bindless
`ExtendedMaterial`s, as well as field-level bindless uniform and storage
buffers. I intend to add back support for these as a follow-up. Because
they aren't in any released Bevy version yet, I figured this was OK.
Finally, this patch updates `StandardMaterial` for the new bindless
changes. Generally, code throughout the PBR shaders that looked like
`base_color_texture[slot]` now looks like
`bindless_2d_textures[material_indices[slot].base_color_texture]`.
This patch fixes a system hang that I experienced on the [Caldera test]
when running with `caldera --random-materials --texture-count 100`. The
time per frame is around 19.75 ms, down from 154.2 ms in Bevy 0.14: a
7.8× speedup.
[Caldera test]: https://github.com/DGriffin91/bevy_caldera_scene
Currently, the specialized pipeline cache maps a (view entity, mesh
entity) tuple to the retained pipeline for that entity. This causes two
problems:
1. Using the view entity is incorrect, because the view entity isn't
stable from frame to frame.
2. Switching the view entity to a `RetainedViewEntity`, which is
necessary for correctness, significantly regresses performance of
`specialize_material_meshes` and `specialize_shadows` because of the
loss of the fast `EntityHash`.
This patch fixes both problems by switching to a *two-level* hash table.
The outer level of the table maps each `RetainedViewEntity` to an inner
table, which maps each `MainEntity` to its pipeline ID and change tick.
Because we loop over views first and, within that loop, loop over
entities visible from that view, we hoist the slow lookup of the view
entity out of the inner entity loop.
Additionally, this patch fixes a bug whereby pipeline IDs were leaked
when removing the view. We still have a problem with leaking pipeline
IDs for deleted entities, but that won't be fixed until the specialized
pipeline cache is retained.
This patch improves performance of the [Caldera benchmark] from 7.8×
faster than 0.14 to 9.0× faster than 0.14, when applied on top of the
global binding arrays PR, #17898.
[Caldera benchmark]: https://github.com/DGriffin91/bevy_caldera_scene
Currently, invocations of `batch_and_prepare_binned_render_phase` and
`batch_and_prepare_sorted_render_phase` can't run in parallel because
they write to scene-global GPU buffers. After PR #17698,
`batch_and_prepare_binned_render_phase` started accounting for the
lion's share of the CPU time, causing us to be strongly CPU bound on
scenes like Caldera when occlusion culling was on (because of the
overhead of batching for the Z-prepass). Although I eventually plan to
optimize `batch_and_prepare_binned_render_phase`, we can obtain
significant wins now by parallelizing that system across phases.
This commit splits all GPU buffers that
`batch_and_prepare_binned_render_phase` and
`batch_and_prepare_sorted_render_phase` touches into separate buffers
for each phase so that the scheduler will run those phases in parallel.
At the end of batch preparation, we gather the render phases up into a
single resource with a new *collection* phase. Because we already run
mesh preprocessing separately for each phase in order to make occlusion
culling work, this is actually a cleaner separation. For example, mesh
output indices (the unique ID that identifies each mesh instance on GPU)
are now guaranteed to be sequential starting from 0, which will simplify
the forthcoming work to remove them in favor of the compute dispatch ID.
On Caldera, this brings the frame time down to approximately 9.1 ms with
occlusion culling on.
Currently, we look up each `MeshInputUniform` index in a hash table that
maps the main entity ID to the index every frame. This is inefficient,
cache unfriendly, and unnecessary, as the `MeshInputUniform` index for
an entity remains the same from frame to frame (even if the input
uniform changes). This commit changes the `IndexSet` in the `RenderBin`
to an `IndexMap` that maps the `MainEntity` to `MeshInputUniformIndex`
(a new type that this patch adds for more type safety).
On Caldera with parallel `batch_and_prepare_binned_render_phase`, this
patch improves that function from 3.18 ms to 2.42 ms, a 31% speedup.
Currently, we *sweep*, or remove entities from bins when those entities
became invisible or changed phases, during `queue_material_meshes` and
similar phases. This, however, is wrong, because `queue_material_meshes`
executes once per material type, not once per phase. This could result
in sweeping bins multiple times per phase, which can corrupt the bins.
This commit fixes the issue by moving sweeping to a separate system that
runs after queuing.
This manifested itself as entities appearing and disappearing seemingly
at random.
Closes#17759.
---------
Co-authored-by: Robert Swain <robert.swain@gmail.com>
# Objective
Things were breaking post-cs.
## Solution
`specialize_mesh_materials` must run after
`collect_meshes_for_gpu_building`. Therefore, its placement in the
`PrepareAssets` set didn't make sense (also more generally). To fix, we
put this class of system in ~`PrepareResources`~ `QueueMeshes`, although
it potentially could use a more descriptive location. We may want to
review the placement of `check_views_need_specialization` which is also
currently in `PrepareAssets`.
This PR makes Bevy keep entities in bins from frame to frame if they
haven't changed. This reduces the time spent in `queue_material_meshes`
and related functions to near zero for static geometry. This patch uses
the same change tick technique that #17567 uses to detect when meshes
have changed in such a way as to require re-binning.
In order to quickly find the relevant bin for an entity when that entity
has changed, we introduce a new type of cache, the *bin key cache*. This
cache stores a mapping from main world entity ID to cached bin key, as
well as the tick of the most recent change to the entity. As we iterate
through the visible entities in `queue_material_meshes`, we check the
cache to see whether the entity needs to be re-binned. If it doesn't,
then we mark it as clean in the `valid_cached_entity_bin_keys` bit set.
If it does, then we insert it into the correct bin, and then mark the
entity as clean. At the end, all entities not marked as clean are
removed from the bins.
This patch has a dramatic effect on the rendering performance of most
benchmarks, as it effectively eliminates `queue_material_meshes` from
the profile. Note, however, that it generally simultaneously regresses
`batch_and_prepare_binned_render_phase` by a bit (not by enough to
outweigh the win, however). I believe that's because, before this patch,
`queue_material_meshes` put the bins in the CPU cache for
`batch_and_prepare_binned_render_phase` to use, while with this patch,
`batch_and_prepare_binned_render_phase` must load the bins into the CPU
cache itself.
On Caldera, this reduces the time spent in `queue_material_meshes` from
5+ ms to 0.2ms-0.3ms. Note that benchmarking on that scene is very noisy
right now because of https://github.com/bevyengine/bevy/issues/17535.
# Objective
- Make use of the new `weak_handle!` macro added in
https://github.com/bevyengine/bevy/pull/17384
## Solution
- Migrate bevy from `Handle::weak_from_u128` to the new `weak_handle!`
macro that takes a random UUID
- Deprecate `Handle::weak_from_u128`, since there are no remaining use
cases that can't also be addressed by constructing the type manually
## Testing
- `cargo run -p ci -- test`
---
## Migration Guide
Replace `Handle::weak_from_u128` with `weak_handle!` and a random UUID.
# Cold Specialization
## Objective
An ongoing part of our quest to retain everything in the render world,
cold-specialization aims to cache pipeline specialization so that
pipeline IDs can be recomputed only when necessary, rather than every
frame. This approach reduces redundant work in stable scenes, while
still accommodating scenarios in which materials, views, or visibility
might change, as well as unlocking future optimization work like
retaining render bins.
## Solution
Queue systems are split into a specialization system and queue system,
the former of which only runs when necessary to compute a new pipeline
id. Pipelines are invalidated using a combination of change detection
and ECS ticks.
### The difficulty with change detection
Detecting “what changed” can be tricky because pipeline specialization
depends not only on the entity’s components (e.g., mesh, material, etc.)
but also on which view (camera) it is rendering in. In other words, the
cache key for a given pipeline id is a view entity/render entity pair.
As such, it's not sufficient simply to react to change detection in
order to specialize -- an entity could currently be out of view or could
be rendered in the future in camera that is currently disabled or hasn't
spawned yet.
### Why ticks?
Ticks allow us to ensure correctness by allowing us to compare the last
time a view or entity was updated compared to the cached pipeline id.
This ensures that even if an entity was out of view or has never been
seen in a given camera before we can still correctly determine whether
it needs to be re-specialized or not.
## Testing
TODO: Tested a bunch of different examples, need to test more.
## Migration Guide
TODO
- `AssetEvents` has been moved into the `PostUpdate` schedule.
---------
Co-authored-by: Patrick Walton <pcwalton@mimiga.net>
This patch fixes a bug whereby we're re-extracting every mesh every
frame. It's a regression from PR #17413. The code in question has
actually been in the tree with this bug for quite a while; it's that
just the code didn't actually run unless the renderer considered the
previous view transforms necessary. Occlusion culling expanded the set
of circumstances under which Bevy computes the previous view transforms,
causing this bug to appear more often.
This patch fixes the issue by checking to see if the previous transform
of a mesh actually differs from the current transform before copying the
current transform to the previous transform.
*Occlusion culling* allows the GPU to skip the vertex and fragment
shading overhead for objects that can be quickly proved to be invisible
because they're behind other geometry. A depth prepass already
eliminates most fragment shading overhead for occluded objects, but the
vertex shading overhead, as well as the cost of testing and rejecting
fragments against the Z-buffer, is presently unavoidable for standard
meshes. We currently perform occlusion culling only for meshlets. But
other meshes, such as skinned meshes, can benefit from occlusion culling
too in order to avoid the transform and skinning overhead for unseen
meshes.
This commit adapts the same [*two-phase occlusion culling*] technique
that meshlets use to Bevy's standard 3D mesh pipeline when the new
`OcclusionCulling` component, as well as the `DepthPrepass` component,
are present on the camera. It has these steps:
1. *Early depth prepass*: We use the hierarchical Z-buffer from the
previous frame to cull meshes for the initial depth prepass, effectively
rendering only the meshes that were visible in the last frame.
2. *Early depth downsample*: We downsample the depth buffer to create
another hierarchical Z-buffer, this time with the current view
transform.
3. *Late depth prepass*: We use the new hierarchical Z-buffer to test
all meshes that weren't rendered in the early depth prepass. Any meshes
that pass this check are rendered.
4. *Late depth downsample*: Again, we downsample the depth buffer to
create a hierarchical Z-buffer in preparation for the early depth
prepass of the next frame. This step is done after all the rendering, in
order to account for custom phase items that might write to the depth
buffer.
Note that this patch has no effect on the per-mesh CPU overhead for
occluded objects, which remains high for a GPU-driven renderer due to
the lack of `cold-specialization` and retained bins. If
`cold-specialization` and retained bins weren't on the horizon, then a
more traditional approach like potentially visible sets (PVS) or low-res
CPU rendering would probably be more efficient than the GPU-driven
approach that this patch implements for most scenes. However, at this
point the amount of effort required to implement a PVS baking tool or a
low-res CPU renderer would probably be greater than landing
`cold-specialization` and retained bins, and the GPU driven approach is
the more modern one anyway. It does mean that the performance
improvements from occlusion culling as implemented in this patch *today*
are likely to be limited, because of the high CPU overhead for occluded
meshes.
Note also that this patch currently doesn't implement occlusion culling
for 2D objects or shadow maps. Those can be addressed in a follow-up.
Additionally, note that the techniques in this patch require compute
shaders, which excludes support for WebGL 2.
This PR is marked experimental because of known precision issues with
the downsampling approach when applied to non-power-of-two framebuffer
sizes (i.e. most of them). These precision issues can, in rare cases,
cause objects to be judged occluded that in fact are not. (I've never
seen this in practice, but I know it's possible; it tends to be likelier
to happen with small meshes.) As a follow-up to this patch, we desire to
switch to the [SPD-based hi-Z buffer shader from the Granite engine],
which doesn't suffer from these problems, at which point we should be
able to graduate this feature from experimental status. I opted not to
include that rewrite in this patch for two reasons: (1) @JMS55 is
planning on doing the rewrite to coincide with the new availability of
image atomic operations in Naga; (2) to reduce the scope of this patch.
A new example, `occlusion_culling`, has been added. It demonstrates
objects becoming quickly occluded and disoccluded by dynamic geometry
and shows the number of objects that are actually being rendered. Also,
a new `--occlusion-culling` switch has been added to `scene_viewer`, in
order to make it easy to test this patch with large scenes like Bistro.
[*two-phase occlusion culling*]:
https://medium.com/@mil_kru/two-pass-occlusion-culling-4100edcad501
[Aaltonen SIGGRAPH 2015]:
https://www.advances.realtimerendering.com/s2015/aaltonenhaar_siggraph2015_combined_final_footer_220dpi.pdf
[Some literature]:
https://gist.github.com/reduz/c5769d0e705d8ab7ac187d63be0099b5?permalink_comment_id=5040452#gistcomment-5040452
[SPD-based hi-Z buffer shader from the Granite engine]:
https://github.com/Themaister/Granite/blob/master/assets/shaders/post/hiz.comp
## Migration guide
* When enqueuing a custom mesh pipeline, work item buffers are now
created with
`bevy::render::batching::gpu_preprocessing::get_or_create_work_item_buffer`,
not `PreprocessWorkItemBuffers::new`. See the
`specialized_mesh_pipeline` example.
## Showcase
Occlusion culling example:
Bistro zoomed out, before occlusion culling:
Bistro zoomed out, after occlusion culling:
In this scene, occlusion culling reduces the number of meshes Bevy has
to render from 1591 to 585.
This commit adds support for *decal projectors* to Bevy, allowing for
textures to be projected on top of geometry. Decal projectors are
clusterable objects, just as punctual lights and light probes are. This
means that decals are only evaluated for objects within the conservative
bounds of the projector, and they don't require a second pass.
These clustered decals require support for bindless textures and as such
currently don't work on WebGL 2, WebGPU, macOS, or iOS. For an
alternative that doesn't require bindless, see PR #16600. I believe that
both contact projective decals in #16600 and clustered decals are
desirable to have in Bevy. Contact projective decals offer broader
hardware and driver support, while clustered decals don't require the
creation of bounding geometry.
A new example, `decal_projectors`, has been added, which demonstrates
multiple decals on a rotating object. The decal projectors can be scaled
and rotated with the mouse.
There are several limitations of this initial patch that can be
addressed in follow-ups:
1. There's no way to specify the Z-index of decals. That is, the order
in which multiple decals are blended on top of one another is arbitrary.
A follow-up could introduce some sort of Z-index field so that artists
can specify that some decals should be blended on top of others.
2. Decals don't take the normal of the surface they're projected onto
into account. Most decal implementations in other engines have a feature
whereby the angle between the decal projector and the normal of the
surface must be within some threshold for the decal to appear. Often,
artists can specify a fade-off range for a smooth transition between
oblique surfaces and aligned surfaces.
3. There's no distance-based fadeoff toward the end of the projector
range. Many decal implementations have this.
This addresses #2401.
## Showcase
This commit allows Bevy to use `multi_draw_indirect_count` for drawing
meshes. The `multi_draw_indirect_count` feature works just like
`multi_draw_indirect`, but it takes the number of indirect parameters
from a GPU buffer rather than specifying it on the CPU.
Currently, the CPU constructs the list of indirect draw parameters with
the instance count for each batch set to zero, uploads the resulting
buffer to the GPU, and dispatches a compute shader that bumps the
instance count for each mesh that survives culling. Unfortunately, this
is inefficient when we support `multi_draw_indirect_count`. Draw
commands corresponding to meshes for which all instances were culled
will remain present in the list when calling
`multi_draw_indirect_count`, causing overhead. Proper use of
`multi_draw_indirect_count` requires eliminating these empty draw
commands.
To address this inefficiency, this PR makes Bevy fully construct the
indirect draw commands on the GPU instead of on the CPU. Instead of
writing instance counts to the draw command buffer, the mesh
preprocessing shader now writes them to a separate *indirect metadata
buffer*. A second compute dispatch known as the *build indirect
parameters* shader runs after mesh preprocessing and converts the
indirect draw metadata into actual indirect draw commands for the GPU.
The build indirect parameters shader operates on a batch at a time,
rather than an instance at a time, and as such each thread writes only 0
or 1 indirect draw parameters, simplifying the current logic in
`mesh_preprocessing`, which currently has to have special cases for the
first mesh in each batch. The build indirect parameters shader emits
draw commands in a tightly packed manner, enabling maximally efficient
use of `multi_draw_indirect_count`.
Along the way, this patch switches mesh preprocessing to dispatch one
compute invocation per render phase per view, instead of dispatching one
compute invocation per view. This is preparation for two-phase occlusion
culling, in which we will have two mesh preprocessing stages. In that
scenario, the first mesh preprocessing stage must only process opaque
and alpha tested objects, so the work items must be separated into those
that are opaque or alpha tested and those that aren't. Thus this PR
splits out the work items into a separate buffer for each phase. As this
patch rewrites so much of the mesh preprocessing infrastructure, it was
simpler to just fold the change into this patch instead of deferring it
to the forthcoming occlusion culling PR.
Finally, this patch changes mesh preprocessing so that it runs
separately for indexed and non-indexed meshes. This is because draw
commands for indexed and non-indexed meshes have different sizes and
layouts. *The existing code is actually broken for non-indexed meshes*,
as it attempts to overlay the indirect parameters for non-indexed meshes
on top of those for indexed meshes. Consequently, right now the
parameters will be read incorrectly when multiple non-indexed meshes are
multi-drawn together. *This is a bug fix* and, as with the change to
dispatch phases separately noted above, was easiest to include in this
patch as opposed to separately.
## Migration Guide
* Systems that add custom phase items now need to populate the indirect
drawing-related buffers. See the `specialized_mesh_pipeline` example for
an example of how this is done.
We won't be able to retain render phases from frame to frame if the keys
are unstable. It's not as simple as simply keying off the main world
entity, however, because some main world entities extract to multiple
render world entities. For example, directional lights extract to
multiple shadow cascades, and point lights extract to one view per
cubemap face. Therefore, we key off a new type, `RetainedViewEntity`,
which contains the main entity plus a *subview ID*.
This is part of the preparation for retained bins.
---------
Co-authored-by: ickshonpe <david.curthoys@googlemail.com>
# Objective
- Closes https://github.com/bevyengine/bevy/issues/14322.
## Solution
- Implement fast 4-sample bicubic filtering based on this shader toy
https://www.shadertoy.com/view/4df3Dn, with a small speedup from a ghost
of tushima presentation.
## Testing
- Did you test these changes? If so, how?
- Ran on lightmapped example. Practically no difference in that scene.
- Are there any parts that need more testing?
- Lightmapping a better scene.
## Changelog
- Lightmaps now have a higher quality bicubic sampling method (off by
default).
---------
Co-authored-by: Patrick Walton <pcwalton@mimiga.net>
# Objective
Many instances of `clippy::too_many_arguments` linting happen to be on
systems - functions which we don't call manually, and thus there's not
much reason to worry about the argument count.
## Solution
Allow `clippy::too_many_arguments` globally, and remove all lint
attributes related to it.
Currently, our batchable binned items are stored in a hash table that
maps bin key, which includes the batch set key, to a list of entities.
Multidraw is handled by sorting the bin keys and accumulating adjacent
bins that can be multidrawn together (i.e. have the same batch set key)
into multidraw commands during `batch_and_prepare_binned_render_phase`.
This is reasonably efficient right now, but it will complicate future
work to retain indirect draw parameters from frame to frame. Consider
what must happen when we have retained indirect draw parameters and the
application adds a bin (i.e. a new mesh) that shares a batch set key
with some pre-existing meshes. (That is, the new mesh can be multidrawn
with the pre-existing meshes.) To be maximally efficient, our goal in
that scenario will be to update *only* the indirect draw parameters for
the batch set (i.e. multidraw command) containing the mesh that was
added, while leaving the others alone. That means that we have to
quickly locate all the bins that belong to the batch set being modified.
In the existing code, we would have to sort the list of bin keys so that
bins that can be multidrawn together become adjacent to one another in
the list. Then we would have to do a binary search through the sorted
list to find the location of the bin that was just added. Next, we would
have to widen our search to adjacent indexes that contain the same batch
set, doing expensive comparisons against the batch set key every time.
Finally, we would reallocate the indirect draw parameters and update the
stored pointers to the indirect draw parameters that the bins store.
By contrast, it'd be dramatically simpler if we simply changed the way
bins are stored to first map from batch set key (i.e. multidraw command)
to the bins (i.e. meshes) within that batch set key, and then from each
individual bin to the mesh instances. That way, the scenario above in
which we add a new mesh will be simpler to handle. First, we will look
up the batch set key corresponding to that mesh in the outer map to find
an inner map corresponding to the single multidraw command that will
draw that batch set. We will know how many meshes the multidraw command
is going to draw by the size of that inner map. Then we simply need to
reallocate the indirect draw parameters and update the pointers to those
parameters within the bins as necessary. There will be no need to do any
binary search or expensive batch set key comparison: only a single hash
lookup and an iteration over the inner map to update the pointers.
This patch implements the above technique. Because we don't have
retained bins yet, this PR provides no performance benefits. However, it
opens the door to maximally efficient updates when only a small number
of meshes change from frame to frame.
The main churn that this patch causes is that the *batch set key* (which
uniquely specifies a multidraw command) and *bin key* (which uniquely
specifies a mesh *within* that multidraw command) are now separate,
instead of the batch set key being embedded *within* the bin key.
In order to isolate potential regressions, I think that at least #16890,
#16836, and #16825 should land before this PR does.
## Migration Guide
* The *batch set key* is now separate from the *bin key* in
`BinnedPhaseItem`. The batch set key is used to collect multidrawable
meshes together. If you aren't using the multidraw feature, you can
safely set the batch set key to `()`.
# Objective
- Contributes to #11478
## Solution
- Made `bevy_utils::tracing` `doc(hidden)`
- Re-exported `tracing` from `bevy_log` for end-users
- Added `tracing` directly to crates that need it.
## Testing
- CI
---
## Migration Guide
If you were importing `tracing` via `bevy::utils::tracing`, instead use
`bevy::log::tracing`. Note that many items within `tracing` are also
directly re-exported from `bevy::log` as well, so you may only need
`bevy::log` for the most common items (e.g., `warn!`, `trace!`, etc.).
This also applies to the `log_once!` family of macros.
## Notes
- While this doesn't reduce the line-count in `bevy_utils`, it further
decouples the internal crates from `bevy_utils`, making its eventual
removal more feasible in the future.
- I have just imported `tracing` as we do for all dependencies. However,
a workspace dependency may be more appropriate for version management.
A previous PR, #14599, attempted to enable lightmaps in deferred mode,
but it still used the `OpaqueNoLightmap3dBinKey`, which meant that it
would be broken if multiple lightmaps were used. This commit fixes that
issue, and allows bindless lightmaps to work with deferred rendering as
well.
Currently, `check_visibility` is parameterized over a query filter that
specifies the type of potentially-visible object. This has the
unfortunate side effect that we need a separate system,
`mark_view_visibility_as_changed_if_necessary`, to trigger view
visibility change detection. That system is quite slow because it must
iterate sequentially over all entities in the scene.
This PR moves the query filter from `check_visibility` to a new
component, `VisibilityClass`. `VisibilityClass` stores a list of type
IDs, each corresponding to one of the query filters we used to use.
Because `check_visibility` is no longer specialized to the query filter
at the type level, Bevy now only needs to invoke it once, leading to
better performance as `check_visibility` can do change detection on the
fly rather than delegating it to a separate system.
This commit also has ergonomic improvements, as there's no need for
applications that want to add their own custom renderable components to
add specializations of the `check_visibility` system to the schedule.
Instead, they only need to ensure that the `ViewVisibility` component is
properly kept up to date. The recommended way to do this, and the way
that's demonstrated in the `custom_phase_item` and
`specialized_mesh_pipeline` examples, is to make `ViewVisibility` a
required component and to add the type ID to it in a component add hook.
This patch does this for `Mesh3d`, `Mesh2d`, `Sprite`, `Light`, and
`Node`, which means that most app code doesn't need to change at all.
Note that, although this patch has a large impact on the performance of
visibility determination, it doesn't actually improve the end-to-end
frame time of `many_cubes`. That's because the render world was already
effectively hiding the latency from
`mark_view_visibility_as_changed_if_necessary`. This patch is, however,
necessary for *further* improvements to `many_cubes` performance.
`many_cubes` trace before:
`many_cubes` trace after:
## Migration Guide
* `check_visibility` no longer takes a `QueryFilter`, and there's no
need to add it manually to your app schedule anymore for custom
rendering items. Instead, entities with custom renderable components
should add the appropriate type IDs to `VisibilityClass`. See
`custom_phase_item` for an example.
This commit resolves most of the failures seen in #16670. It contains
two major fixes:
1. The prepass shaders weren't updated for bindless mode, so they were
accessing `material` as a single element instead of as an array. I added
the needed `BINDLESS` check.
2. If the mesh didn't support batch set keys (i.e. `get_batch_set_key()`
returns `None`), and multidraw was enabled, the batching logic would try
to multidraw all the meshes in a bin together instead of disabling
multidraw. This is because we checked whether the `Option<BatchSetKey>`
for the previous batch was equal to the `Option<BatchSetKey>` for the
next batch to determine whether objects could be multidrawn together,
which would return true if batch set keys were absent, causing an entire
bin to be multidrawn together. This patch fixes the logic so that
multidraw is only enabled if the batch set keys match *and are `Some`*.
Additionally, this commit adds batch key support for bins that use
`Opaque3dNoLightmapBinKey`, which in practice means prepasses.
Consequently, this patch enables multidraw for the prepass when GPU
culling is enabled.
When testing this patch, try adding `GpuCulling` to the camera in the
`deferred_rendering` and `ssr` examples. You can see that these examples
break without this patch and work properly with it.
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
This commit makes skinned meshes batchable on platforms other than WebGL
2. On supported platforms, it replaces the two uniform buffers used for
joint matrices with a pair of storage buffers containing all matrices
for all skinned meshes packed together. The indices into the buffer are
stored in the mesh uniform and mesh input uniform. The GPU mesh
preprocessing step copies the indices in if that step is enabled.
On the `many_foxes` demo, I observed a frame time decrease from 15.470ms
to 11.935ms. This is the result of reducing the `submit_graph_commands`
time from an average of 5.45ms to 0.489ms, an 11x speedup in that
portion of rendering.
This is what the profile looks like for `many_foxes` after these
changes.
---------
Co-authored-by: François Mockers <mockersf@gmail.com>
This commit makes `StandardMaterial` use bindless textures, as
implemented in PR #16368. Non-bindless mode, as used for example in
Metal and WebGL 2, remains fully supported via a plethora of `#ifdef
BINDLESS` preprocessor definitions.
Unfortunately, this PR introduces quite a bit of unsightliness into the
PBR shaders. This is a result of the fact that WGSL supports neither
passing binding arrays to functions nor passing individual *elements* of
binding arrays to functions, except directly to texture sample
functions. Thus we're unable to use the `sample_texture` abstraction
that helped abstract over the meshlet and non-meshlet paths. I don't
think there's anything we can do to help this other than to suggest
improvements to upstream Naga.
Currently, the prepass has no support for visibility ranges, so
artifacts appear when using dithering visibility ranges in conjunction
with a prepass. This patch fixes that problem.
Note that this patch changes the prepass to use sparse bind group
indices instead of sequential ones. I figured this is cleaner, because
it allows for greater sharing of WGSL code between the forward pipeline
and the prepass pipeline.
The `visibility_range` example has been updated to allow the prepass to
be toggled on and off.
This patch adds the infrastructure necessary for Bevy to support
*bindless resources*, by adding a new `#[bindless]` attribute to
`AsBindGroup`.
Classically, only a single texture (or sampler, or buffer) can be
attached to each shader binding. This means that switching materials
requires breaking a batch and issuing a new drawcall, even if the mesh
is otherwise identical. This adds significant overhead not only in the
driver but also in `wgpu`, as switching bind groups increases the amount
of validation work that `wgpu` must do.
*Bindless resources* are the typical solution to this problem. Instead
of switching bindings between each texture, the renderer instead
supplies a large *array* of all textures in the scene up front, and the
material contains an index into that array. This pattern is repeated for
buffers and samplers as well. The renderer now no longer needs to switch
binding descriptor sets while drawing the scene.
Unfortunately, as things currently stand, this approach won't quite work
for Bevy. Two aspects of `wgpu` conspire to make this ideal approach
unacceptably slow:
1. In the DX12 backend, all binding arrays (bindless resources) must
have a constant size declared in the shader, and all textures in an
array must be bound to actual textures. Changing the size requires a
recompile.
2. Changing even one texture incurs revalidation of all textures, a
process that takes time that's linear in the total size of the binding
array.
This means that declaring a large array of textures big enough to
encompass the entire scene is presently unacceptably slow. For example,
if you declare 4096 textures, then `wgpu` will have to revalidate all
4096 textures if even a single one changes. This process can take
multiple frames.
To work around this problem, this PR groups bindless resources into
small *slabs* and maintains a free list for each. The size of each slab
for the bindless arrays associated with a material is specified via the
`#[bindless(N)]` attribute. For instance, consider the following
declaration:
```rust
#[derive(AsBindGroup)]
#[bindless(16)]
struct MyMaterial {
#[buffer(0)]
color: Vec4,
#[texture(1)]
#[sampler(2)]
diffuse: Handle<Image>,
}
```
The `#[bindless(N)]` attribute specifies that, if bindless arrays are
supported on the current platform, each resource becomes a binding array
of N instances of that resource. So, for `MyMaterial` above, the `color`
attribute is exposed to the shader as `binding_array<vec4<f32>, 16>`,
the `diffuse` texture is exposed to the shader as
`binding_array<texture_2d<f32>, 16>`, and the `diffuse` sampler is
exposed to the shader as `binding_array<sampler, 16>`. Inside the
material's vertex and fragment shaders, the applicable index is
available via the `material_bind_group_slot` field of the `Mesh`
structure. So, for instance, you can access the current color like so:
```wgsl
// `uniform` binding arrays are a non-sequitur, so `uniform` is automatically promoted
// to `storage` in bindless mode.
@group(2) @binding(0) var<storage> material_color: binding_array<Color, 4>;
...
@fragment
fn fragment(in: VertexOutput) -> @location(0) vec4<f32> {
let color = material_color[mesh[in.instance_index].material_bind_group_slot];
...
}
```
Note that portable shader code can't guarantee that the current platform
supports bindless textures. Indeed, bindless mode is only available in
Vulkan and DX12. The `BINDLESS` shader definition is available for your
use to determine whether you're on a bindless platform or not. Thus a
portable version of the shader above would look like:
```wgsl
#ifdef BINDLESS
@group(2) @binding(0) var<storage> material_color: binding_array<Color, 4>;
#else // BINDLESS
@group(2) @binding(0) var<uniform> material_color: Color;
#endif // BINDLESS
...
@fragment
fn fragment(in: VertexOutput) -> @location(0) vec4<f32> {
#ifdef BINDLESS
let color = material_color[mesh[in.instance_index].material_bind_group_slot];
#else // BINDLESS
let color = material_color;
#endif // BINDLESS
...
}
```
Importantly, this PR *doesn't* update `StandardMaterial` to be bindless.
So, for example, `scene_viewer` will currently not run any faster. I
intend to update `StandardMaterial` to use bindless mode in a follow-up
patch.
A new example, `shaders/shader_material_bindless`, has been added to
demonstrate how to use this new feature.
Here's a Tracy profile of `submit_graph_commands` of this patch and an
additional patch (not submitted yet) that makes `StandardMaterial` use
bindless. Red is those patches; yellow is `main`. The scene was Bistro
Exterior with a hack that forces all textures to opaque. You can see a
1.47x mean speedup.
## Migration Guide
* `RenderAssets::prepare_asset` now takes an `AssetId` parameter.
* Bin keys now have Bevy-specific material bind group indices instead of
`wgpu` material bind group IDs, as part of the bindless change. Use the
new `MaterialBindGroupAllocator` to map from bind group index to bind
group ID.
# Objective
- Fixes#16078
## Solution
- Rename things to clarify that we _want_ unclipped depth for
directional light shadow views, and need some way of disabling the GPU's
builtin depth clipping
- Use DEPTH_CLIP_CONTROL instead of the fragment shader emulation on
supported platforms
- Pass only the clip position depth instead of the whole clip position
between vertex->fragment shader (no idea if this helps performance or
not, compiler might optimize it anyways)
- Meshlets
- HW raster always uses DEPTH_CLIP_CONTROL since it targets a more
limited set of platforms
- SW raster was not handling DEPTH_CLAMP_ORTHO correctly, it ended up
pretty much doing nothing.
- This PR made me realize that SW raster technically should have depth
clipping for all views that are not directional light shadows, but I
decided not to bother writing it. I'm not sure that it ever matters in
practice. If proven otherwise, I can add it.
## Testing
- Did you test these changes? If so, how?
- Lighting example. Both opaque (no fragment shader) and alpha masked
geometry (fragment shader emulation) are working with
depth_clip_control, and both work when it's turned off. Also tested
meshlet example.
- Are there any parts that need more testing?
- Performance. I can't figure out a good test scene.
- How can other people (reviewers) test your changes? Is there anything
specific they need to know?
- Toggle depth_clip_control_supported in prepass/mod.rs line 323 to turn
this PR on or off.
- If relevant, what platforms did you test these changes on, and are
there any important ones you can't test?
- Native
---
## Migration Guide
- `MeshPipelineKey::DEPTH_CLAMP_ORTHO` is now
`MeshPipelineKey::UNCLIPPED_DEPTH_ORTHO`
- The `DEPTH_CLAMP_ORTHO` shaderdef has been renamed to
`UNCLIPPED_DEPTH_ORTHO_EMULATION`
- `clip_position_unclamped: vec4<f32>` is now `unclipped_depth: f32`
# Objective
- wgpu 0.20 made workgroup vars stop being zero-init by default. this
broke some applications (cough foresight cough) and now we workaround
it. wgpu exposes a compilation option that zero initializes workgroup
memory by default, but bevy does not expose it.
## Solution
- expose the compilation option wgpu gives us
## Testing
- ran examples: 3d_scene, compute_shader_game_of_life, gpu_readback,
lines, specialized_mesh_pipeline. they all work
- confirmed fix for our own problems
---
</details>
## Migration Guide
- add `zero_initialize_workgroup_memory: false,` to
`ComputePipelineDescriptor` or `RenderPipelineDescriptor` structs to
preserve 0.14 functionality, add `zero_initialize_workgroup_memory:
true,` to restore bevy 0.13 functionality.
# Objective
- Make the meshlet fill cluster buffers pass slightly faster
- Address https://github.com/bevyengine/bevy/issues/15920 for meshlets
- Added PreviousGlobalTransform as a required meshlet component to avoid
extra archetype moves, slightly alleviating
https://github.com/bevyengine/bevy/issues/14681 for meshlets
- Enforce that MeshletPlugin::cluster_buffer_slots is not greater than
2^25 (glitches will occur otherwise). Technically this field controls
post-lod/culling cluster count, and the issue is on pre-lod/culling
cluster count, but it's still valid now, and in the future this will be
more true.
Needs to be merged after https://github.com/bevyengine/bevy/pull/15846
and https://github.com/bevyengine/bevy/pull/15886
## Solution
- Old pass dispatched a thread per cluster, and did a binary search over
the instances to find which instance the cluster belongs to, and what
meshlet index within the instance it is.
- New pass dispatches a workgroup per instance, and has the workgroup
loop over all meshlets in the instance in order to write out the cluster
data.
- Use a push constant instead of arrayLength to fix the linked bug
- Remap 1d->2d dispatch for software raster only if actually needed to
save on spawning excess workgroups
## Testing
- Did you test these changes? If so, how?
- Ran the meshlet example, and an example with 1041 instances of 32217
meshlets per instance. Profiled the second scene with nsight, went from
0.55ms -> 0.40ms. Small savings. We're pretty much VRAM bandwidth bound
at this point.
- How can other people (reviewers) test your changes? Is there anything
specific they need to know?
- Run the meshlet example
## Changelog (non-meshlets)
- PreviousGlobalTransform now implements the Default trait
# Objective
- Fixes https://github.com/bevyengine/bevy/issues/15871
(Camera is done in #15946)
## Solution
- Do the same as #15904 for other extraction systems
- Added missing `SyncComponentPlugin` for DOF, TAA, and SSAO
(According to the
[documentation](https://dev-docs.bevyengine.org/bevy/render/sync_component/struct.SyncComponentPlugin.html),
this plugin "needs to be added for manual extraction implementations."
We may need to check this is done.)
## Testing
Modified example locally to add toggles if not exist.
- [x] DOF - toggling DOF component and perspective in `depth_of_field`
example
- [x] TAA - toggling `Camera.is_active` and TAA component
- [x] clusters - not entirely sure, toggling `Camera.is_active` in
`many_lights` example (no crash/glitch even without this PR)
- [x] previous_view - toggling `Camera.is_active` in `skybox` (no
crash/glitch even without this PR)
- [x] lights - toggling `Visibility` of `DirectionalLight` in `lighting`
example
- [x] SSAO - toggling `Camera.is_active` and SSAO component in `ssao`
example
- [x] default UI camera view - toggling `Camera.is_active` (nop without
#15946 because UI defaults to some camera even if `DefaultCameraView` is
not there)
- [x] volumetric fog - toggling existence of volumetric light. Looks
like optimization, no change in behavior/visuals
# Objective
- Fixes https://github.com/bevyengine/bevy/issues/13552
## Solution
- Thanks for the guidance from @DGriffin91, the current solution is to
transmit the light_map through the emissive channel to avoid increasing
the bandwidth of deferred shading.
- <del>Store lightmap sample result into G-Buffer and pass them into the
`Deferred Lighting Pipeline`, therefore we can get the correct indirect
lighting via the `apply_pbr_lighting` function.</del>
- <del>The original G-Buffer lacks storage for lightmap data, therefore
a new buffer is added. We can only use Rgba16Uint here due to the
32-byte limit on the render targets.</del>
## Testing
- Need to test all the examples that contains a prepass, with both the
forward and deferred rendering mode.
- I have tested the ones below.
- `lightmaps` (adjust the code based on the issue and check the
rendering result)
- `transmission` (it contains a prepass)
- `ssr` (it also uses the G-Bufffer)
- `meshlet` (forward and deferred)
- `pbr`
## Showcase
By updating the `lightmaps` example to use deferred rendering, this pull
request enables correct rendering result of the Cornell Box.
```
diff --git a/examples/3d/lightmaps.rs b/examples/3d/lightmaps.rs
index 564a3162b..11a748fba 100644
--- a/examples/3d/lightmaps.rs
+++ b/examples/3d/lightmaps.rs
@@ -1,12 +1,14 @@
//! Rendering a scene with baked lightmaps.
-use bevy::pbr::Lightmap;
+use bevy::core_pipeline::prepass::DeferredPrepass;
+use bevy::pbr::{DefaultOpaqueRendererMethod, Lightmap};
use bevy::prelude::*;
fn main() {
App::new()
.add_plugins(DefaultPlugins)
.insert_resource(AmbientLight::NONE)
+ .insert_resource(DefaultOpaqueRendererMethod::deferred())
.add_systems(Startup, setup)
.add_systems(Update, add_lightmaps_to_meshes)
.run();
@@ -19,10 +21,12 @@ fn setup(mut commands: Commands, asset_server: Res<AssetServer>) {
..default()
});
- commands.spawn(Camera3dBundle {
- transform: Transform::from_xyz(-278.0, 273.0, 800.0),
- ..default()
- });
+ commands
+ .spawn(Camera3dBundle {
+ transform: Transform::from_xyz(-278.0, 273.0, 800.0),
+ ..default()
+ })
+ .insert(DeferredPrepass);
}
fn add_lightmaps_to_meshes(
```
<img width="1280" alt="image"
src="https://github.com/user-attachments/assets/17fd3367-61cc-4c23-b956-e7cfc751af3c">
## Emissive Issue
**The emissive light object appears incorrectly rendered because the
alpha channel of emission is set to 1 in deferred rendering and 0 in
forward rendering, leading to different emissive light result. Could
this be a bug?**
```wgsl
// pbr_deferred_functions.wgsl - pbr_input_from_deferred_gbuffer
let emissive = rgb9e5::rgb9e5_to_vec3_(gbuffer.g);
if ((pbr.material.flags & STANDARD_MATERIAL_FLAGS_UNLIT_BIT) != 0u) {
pbr.material.base_color = vec4(emissive, 1.0);
pbr.material.emissive = vec4(vec3(0.0), 1.0);
} else {
pbr.material.base_color = vec4(pow(base_rough.rgb, vec3(2.2)), 1.0);
pbr.material.emissive = vec4(emissive, 1.0);
}
// pbr_functions.wgsl - apply_pbr_lighting
emissive_light = emissive_light * mix(1.0, view_bindings::view.exposure, emissive.a);
```
---------
Co-authored-by: JMS55 <47158642+JMS55@users.noreply.github.com>
