# Objective
- Fixes#15460 (will open new issues for further `no_std` efforts)
- Supersedes #17715
## Solution
- Threaded in new features as required
- Made certain crates optional but default enabled
- Removed `compile-check-no-std` from internal `ci` tool since GitHub CI
can now simply check `bevy` itself now
- Added CI task to check `bevy` on `thumbv6m-none-eabi` to ensure
`portable-atomic` support is still valid [^1]
[^1]: This may be controversial, since it could be interpreted as
implying Bevy will maintain support for `thumbv6m-none-eabi` going
forward. In reality, just like `x86_64-unknown-none`, this is a
[canary](https://en.wiktionary.org/wiki/canary_in_a_coal_mine) target to
make it clear when `portable-atomic` no longer works as intended (fixing
atomic support on atomically challenged platforms). If a PR comes
through and makes supporting this class of platforms impossible, then
this CI task can be removed. I however wager this won't be a problem.
## Testing
- CI
---
## Release Notes
Bevy now has support for `no_std` directly from the `bevy` crate.
Users can disable default features and enable a new `default_no_std`
feature instead, allowing `bevy` to be used in `no_std` applications and
libraries.
```toml
# Bevy for `no_std` platforms
bevy = { version = "0.16", default-features = false, features = ["default_no_std"] }
```
`default_no_std` enables certain required features, such as `libm` and
`critical-section`, and as many optional crates as possible (currently
just `bevy_state`). For atomically-challenged platforms such as the
Raspberry Pi Pico, `portable-atomic` will be used automatically.
For library authors, we recommend depending on `bevy` with
`default-features = false` to allow `std` and `no_std` users to both
depend on your crate. Here are some recommended features a library crate
may want to expose:
```toml
[features]
# Most users will be on a platform which has `std` and can use the more-powerful `async_executor`.
default = ["std", "async_executor"]
# Features for typical platforms.
std = ["bevy/std"]
async_executor = ["bevy/async_executor"]
# Features for `no_std` platforms.
libm = ["bevy/libm"]
critical-section = ["bevy/critical-section"]
[dependencies]
# We disable default features to ensure we don't accidentally enable `std` on `no_std` targets, for example.
bevy = { version = "0.16", default-features = false }
```
While this is verbose, it gives the maximum control to end-users to
decide how they wish to use Bevy on their platform.
We encourage library authors to experiment with `no_std` support. For
libraries relying exclusively on `bevy` and no other dependencies, it
may be as simple as adding `#![no_std]` to your `lib.rs` and exposing
features as above! Bevy can also provide many `std` types, such as
`HashMap`, `Mutex`, and `Instant` on all platforms. See
`bevy::platform_support` for details on what's available out of the box!
## Migration Guide
- If you were previously relying on `bevy` with default features
disabled, you may need to enable the `std` and `async_executor`
features.
- `bevy_reflect` has had its `bevy` feature removed. If you were relying
on this feature, simply enable `smallvec` and `smol_str` instead.
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
In 0.11 you could easily access the inverse model matrix inside a WGSL
shader with `transpose(mesh.inverse_transpose_model)`. This was changed
in 0.12 when `inverse_transpose_model` was removed and it's now not as
straightfoward. I wrote a helper function for my own code and thought
I'd submit a pull request in case it would be helpful to others.
## Objective
`insert_or_spawn_batch` is due to be deprecated eventually (#15704), and
removing uses internally will make that easier.
## Solution
Replaced internal uses of `insert_or_spawn_batch` with
`try_insert_batch` (non-panicking variant because
`insert_or_spawn_batch` didn't panic).
All of the internal uses are in rendering code. Since retained rendering
was meant to get rid non-opaque entity IDs, I assume the code was just
using `insert_or_spawn_batch` because `insert_batch` didn't exist and
not because it actually wanted to spawn something. However, I am *not*
confident in my ability to judge rendering code.
# Objective
Component `require()` IDE integration is fully broken, as of #16575.
## Solution
This reverts us back to the previous "put the docs on Component trait"
impl. This _does_ reduce the accessibility of the required components in
rust docs, but the complete erasure of "required component IDE
experience" is not worth the price of slightly increased prominence of
requires in docs.
Additionally, Rust Analyzer has recently started including derive
attributes in suggestions, so we aren't losing that benefit of the
proc_macro attribute impl.
This reverts commit 0b5302d96a.
# Objective
- Fixes#18158
- #17482 introduced rendering changes and was merged a bit too fast
## Solution
- Revert #17482 so that it can be redone and rendering changes discussed
before being merged. This will make it easier to compare changes with
main in the known "valid" state
This is not an issue with the work done in #17482 that is still
interesting
# Objective
Transparently uses simple `EnvironmentMapLight`s to mimic
`AmbientLight`s. Implements the first part of #17468, but I can
implement hemispherical lights in this PR too if needed.
## Solution
- A function `EnvironmentMapLight::solid_color(&mut Assets<Image>,
Color)` is provided to make an environment light with a solid color.
- A new system is added to `SimulationLightSystems` that maps
`AmbientLight`s on views or the world to a corresponding
`EnvironmentMapLight`.
I have never worked with (or on) Bevy before, so nitpicky comments on
how I did things are appreciated :).
## Testing
Testing was done on a modified version of the `3d/lighting` example,
where I removed all lights except the ambient light. I have not included
the example, but can if required.
## Migration
`bevy_pbr::AmbientLight` has been deprecated, so all usages of it should
be replaced by a `bevy_pbr::EnvironmentMapLight` created with
`EnvironmentMapLight::solid_color` placed on the camera. There is no
alternative to ambient lights as resources.
This commit makes the
`mark_meshes_as_changed_if_their_materials_changed` system use the new
`AssetChanged<MeshMaterial3d>` query filter in addition to
`Changed<MeshMaterial3d>`. This ensures that we update the
`MeshInputUniform`, which contains the bindless material slot. Updating
the `MeshInputUniform` fixes problems that occurred when the
`MeshBindGroupAllocator` reallocated meshes in such a way as to change
their bindless slot.
Closes#18102.
# Objective
Fixes https://github.com/bevyengine/bevy/issues/17590.
## Solution
`prepare_volumetric_fog_uniforms` adds a uniform for each combination of
fog volume and view. But it only allocated enough uniforms for one fog
volume per view.
## Testing
Ran the `volumetric_fog` example with 1/2/3/4 fog volumes. Also checked
the `fog_volumes` and `scrolling_fog` examples (without multiple
volumes). Win10/Vulkan/Nvidia.
To test multiple views I tried adding fog volumes to the `split_screen`
example. This doesn't quite work - the fog should be centred on the fox,
but instead it's centred on the window. Same result with and without the
PR, so I'm assuming it's a separate bug.
# Objective
As discussed in #14275, Bevy is currently too prone to panic, and makes
the easy / beginner-friendly way to do a large number of operations just
to panic on failure.
This is seriously frustrating in library code, but also slows down
development, as many of the `Query::single` panics can actually safely
be an early return (these panics are often due to a small ordering issue
or a change in game state.
More critically, in most "finished" products, panics are unacceptable:
any unexpected failures should be handled elsewhere. That's where the
new
With the advent of good system error handling, we can now remove this.
Note: I was instrumental in a) introducing this idea in the first place
and b) pushing to make the panicking variant the default. The
introduction of both `let else` statements in Rust and the fancy system
error handling work in 0.16 have changed my mind on the right balance
here.
## Solution
1. Make `Query::single` and `Query::single_mut` (and other random
related methods) return a `Result`.
2. Handle all of Bevy's internal usage of these APIs.
3. Deprecate `Query::get_single` and friends, since we've moved their
functionality to the nice names.
4. Add detailed advice on how to best handle these errors.
Generally I like the diff here, although `get_single().unwrap()` in
tests is a bit of a downgrade.
## Testing
I've done a global search for `.single` to track down any missed
deprecated usages.
As to whether or not all the migrations were successful, that's what CI
is for :)
## Future work
~~Rename `Query::get_single` and friends to `Query::single`!~~
~~I've opted not to do this in this PR, and smear it across two releases
in order to ease the migration. Successive deprecations are much easier
to manage than the semantics and types shifting under your feet.~~
Cart has convinced me to change my mind on this; see
https://github.com/bevyengine/bevy/pull/18082#discussion_r1974536085.
## Migration guide
`Query::single`, `Query::single_mut` and their `QueryState` equivalents
now return a `Result`. Generally, you'll want to:
1. Use Bevy 0.16's system error handling to return a `Result` using the
`?` operator.
2. Use a `let else Ok(data)` block to early return if it's an expected
failure.
3. Use `unwrap()` or `Ok` destructuring inside of tests.
The old `Query::get_single` (etc) methods which did this have been
deprecated.
## Objective
Alternative to #18001.
- Now that systems can handle the `?` operator, `get_entity` returning
`Result` would be more useful than `Option`.
- With `get_entity` being more flexible, combined with entity commands
now checking the entity's existence automatically, the panic in `entity`
isn't really necessary.
## Solution
- Changed `Commands::get_entity` to return `Result<EntityCommands,
EntityDoesNotExistError>`.
- Removed panic from `Commands::entity`.
Even though opaque deferred entities aren't placed into the `Opaque3d`
bin, we still want to cache them as though they were, so that we don't
have to re-queue them every frame. This commit implements that logic,
reducing the time of `queue_material_meshes` to near-zero on Caldera.
Currently, the structure-level `#[uniform]` attribute of `AsBindGroup`
creates a binding array of individual buffers, each of which contains
data for a single material. A more efficient approach would be to
provide a single buffer with an array containing all of the data for all
materials in the bind group. Because `StandardMaterial` uses
`#[uniform]`, this can be notably inefficient with large numbers of
materials.
This patch introduces a new attribute on `AsBindGroup`, `#[data]`, which
works identically to `#[uniform]` except that it concatenates all the
data into a single buffer that the material bind group allocator itself
manages. It also converts `StandardMaterial` to use this new
functionality. This effectively provides the "material data in arrays"
feature.
# 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.**
# Objective
- Fixes#17960
## Solution
- Followed the [edition upgrade
guide](https://doc.rust-lang.org/edition-guide/editions/transitioning-an-existing-project-to-a-new-edition.html)
## Testing
- CI
---
## Summary of Changes
### Documentation Indentation
When using lists in documentation, proper indentation is now linted for.
This means subsequent lines within the same list item must start at the
same indentation level as the item.
```rust
/* Valid */
/// - Item 1
/// Run-on sentence.
/// - Item 2
struct Foo;
/* Invalid */
/// - Item 1
/// Run-on sentence.
/// - Item 2
struct Foo;
```
### Implicit `!` to `()` Conversion
`!` (the never return type, returned by `panic!`, etc.) no longer
implicitly converts to `()`. This is particularly painful for systems
with `todo!` or `panic!` statements, as they will no longer be functions
returning `()` (or `Result<()>`), making them invalid systems for
functions like `add_systems`. The ideal fix would be to accept functions
returning `!` (or rather, _not_ returning), but this is blocked on the
[stabilisation of the `!` type
itself](https://doc.rust-lang.org/std/primitive.never.html), which is
not done.
The "simple" fix would be to add an explicit `-> ()` to system
signatures (e.g., `|| { todo!() }` becomes `|| -> () { todo!() }`).
However, this is _also_ banned, as there is an existing lint which (IMO,
incorrectly) marks this as an unnecessary annotation.
So, the "fix" (read: workaround) is to put these kinds of `|| -> ! { ...
}` closuers into variables and give the variable an explicit type (e.g.,
`fn()`).
```rust
// Valid
let system: fn() = || todo!("Not implemented yet!");
app.add_systems(..., system);
// Invalid
app.add_systems(..., || todo!("Not implemented yet!"));
```
### Temporary Variable Lifetimes
The order in which temporary variables are dropped has changed. The
simple fix here is _usually_ to just assign temporaries to a named
variable before use.
### `gen` is a keyword
We can no longer use the name `gen` as it is reserved for a future
generator syntax. This involved replacing uses of the name `gen` with
`r#gen` (the raw-identifier syntax).
### Formatting has changed
Use statements have had the order of imports changed, causing a
substantial +/-3,000 diff when applied. For now, I have opted-out of
this change by amending `rustfmt.toml`
```toml
style_edition = "2021"
```
This preserves the original formatting for now, reducing the size of
this PR. It would be a simple followup to update this to 2024 and run
`cargo fmt`.
### New `use<>` Opt-Out Syntax
Lifetimes are now implicitly included in RPIT types. There was a handful
of instances where it needed to be added to satisfy the borrow checker,
but there may be more cases where it _should_ be added to avoid
breakages in user code.
### `MyUnitStruct { .. }` is an invalid pattern
Previously, you could match against unit structs (and unit enum
variants) with a `{ .. }` destructuring. This is no longer valid.
### Pretty much every use of `ref` and `mut` are gone
Pattern binding has changed to the point where these terms are largely
unused now. They still serve a purpose, but it is far more niche now.
### `iter::repeat(...).take(...)` is bad
New lint recommends using the more explicit `iter::repeat_n(..., ...)`
instead.
## Migration Guide
The lifetimes of functions using return-position impl-trait (RPIT) are
likely _more_ conservative than they had been previously. If you
encounter lifetime issues with such a function, please create an issue
to investigate the addition of `+ use<...>`.
## Notes
- Check the individual commits for a clearer breakdown for what
_actually_ changed.
---------
Co-authored-by: François Mockers <francois.mockers@vleue.com>
# Objective
`Eq`/`PartialEq` are currently implemented for `MeshMaterial{2|3}d` only
through the derive macro. Since we don't have perfect derive yet, the
impls are only present for `M: Eq` and `M: PartialEq`. On the other
hand, I want to be able to compare material components for my toy
reactivity project.
## Solution
Switch to manual `Eq`/`PartialEq` impl.
## Testing
Boy I hope this didn't break anything!
PR #17898 regressed this, causing much of #17970. This commit fixes the
issue by freeing and reallocating materials in the
`MaterialBindGroupAllocator` on change. Note that more efficiency is
possible, but I opted for the simple approach because (1) we should fix
this bug ASAP; (2) I'd like #17965 to land first, because that unlocks
the biggest potential optimization, which is not recreating the bind
group if it isn't necessary to do so.
We might not be able to prepare a material on the first frame we
encounter a mesh using it for various reasons, including that the
material hasn't been loaded yet or that preparing the material is
exceeding the per-frame cap on number of bytes to load. When this
happens, we currently try to find the material in the
`MaterialBindGroupAllocator`, fail, and then fall back to group 0, slot
0, the default `MaterialBindGroupId`, which is obviously incorrect.
Worse, we then fail to dirty the mesh and reextract it when we *do*
finish preparing the material, so the mesh will continue to be rendered
with an incorrect material.
This patch fixes both problems. In `collect_meshes_for_gpu_building`, if
we fail to find a mesh's material in the `MeshBindGroupAllocator`, then
we detect that case, bail out, and add it to a list,
`MeshesToReextractNextFrame`. On subsequent frames, we process all the
meshes in `MeshesToReextractNextFrame` as though they were changed. This
ensures both that we don't render a mesh if its material hasn't been
loaded and that we start rendering the mesh once its material does load.
This was first noticed in the intermittent Pixel Eagle failures in the
`testbed_3d` patch in #17898, although the problem has actually existed
for some time. I believe it just so happened that the changes to the
allocator in that PR caused the problem to appear more commonly than it
did before.
This patch fixes two bugs in the new non-bindless material allocator
that landed in PR #17898:
1. A debug assertion to prevent double frees had been flipped: we
checked to see whether the slot was empty before freeing, while we
should have checked to see whether the slot was full.
2. The non-bindless allocator returned `None` when querying a slab that
hadn't been prepared yet instead of returning a handle to that slab.
This resulted in a 1-frame delay when modifying materials. In the
`animated_material` example, this resulted in the meshes never showing
up at all, because that example changes every material every frame.
Together with #17979, this patch locally fixes the problems with
`animated_material` on macOS that were reported in #17970.
Two-phase occlusion culling can be helpful for shadow maps just as it
can for a prepass, in order to reduce vertex and alpha mask fragment
shading overhead. This patch implements occlusion culling for shadow
maps from directional lights, when the `OcclusionCulling` component is
present on the entities containing the lights. Shadow maps from point
lights are deferred to a follow-up patch. Much of this patch involves
expanding the hierarchical Z-buffer to cover shadow maps in addition to
standard view depth buffers.
The `scene_viewer` example has been updated to add `OcclusionCulling` to
the directional light that it creates.
This improved the performance of the rend3 sci-fi test scene when
enabling shadows.
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
Deferred rendering currently doesn't support occlusion culling. This PR
implements it in a straightforward way, mirroring what we already do for
the non-deferred pipeline.
On the rend3 sci-fi base test scene, this resulted in roughly a 2×
speedup when applied on top of my other patches. For that scene, it was
useful to add another option, `--add-light`, which forces the addition
of a shadow-casting light, to the scene viewer, which I included in this
patch.
This commit restructures the multidrawable batch set builder for better
performance in various ways:
* The bin traversal is optimized to make the best use of the CPU cache.
* The inner loop that iterates over the bins, which is the hottest part
of `batch_and_prepare_binned_render_phase`, has been shrunk as small as
possible.
* Where possible, multiple elements are added to or reserved from GPU
buffers as a batch instead of one at a time.
* Methods that LLVM wasn't inlining have been marked `#[inline]` where
doing so would unlock optimizations.
This code has also been refactored to avoid duplication between the
logic for indexed and non-indexed meshes via the introduction of a
`MultidrawableBatchSetPreparer` object.
Together, this improved the `batch_and_prepare_binned_render_phase` time
on Caldera by approximately 2×.
Eventually, we should optimize the batchable-but-not-multidrawable and
unbatchable logic as well, but these meshes are much rarer, so in the
interests of keeping this patch relatively small I opted to leave those
to a follow-up.
# Objective
Fix incorrect mesh culling where objects (particularly directional
shadows) were being incorrectly culled during the early preprocessing
phase. The issue manifested specifically on Apple M1 GPUs but not on
newer devices like the M4. The bug was in the
`view_frustum_intersects_obb` function, where including the w component
(plane distance) in the dot product calculations led to false positive
culling results. This caused objects to be incorrectly culled before
shadow casting could begin.
## Issue Details
The problem of missing shadows is reproducible on Apple M1 GPUs as of
this commit (bisected):
```
00722b8d0 Make indirect drawing opt-out instead of opt-in, enabling multidraw by default. (#16757)
```
and as recent as this commit:
```
c818c9214 Add option to animate materials in many_cubes (#17927)
```
- The frustum culling calculation incorrectly included the w component
(plane distance) when transforming basis vectors
- The relative radius calculation should only consider directional
transformation (xyz), not positional information (w)
- This caused false positive culling specifically on M1 devices likely
due to different device-specific floating-point behavior
- When objects were incorrectly culled, `early_instance_count` never
incremented, leading to missing geometry in the shadow pass
## Testing
- Tested on M1 and M4 devices to verify the fix
- Verified shadows and geometry render correctly on both platforms
- Confirmed the solution matches the existing Rust implementation's
behavior for calculating the relative radius:
c818c92143/crates/bevy_render/src/primitives/mod.rs (L77-L87)
- The fix resolves a mathematical error in the frustum culling
calculation while maintaining correct culling behavior across all
platforms.
---
## Showcase
`c818c9214`
<img width="1284" alt="c818c9214"
src="https://github.com/user-attachments/assets/fe1c7ea9-b13d-422e-b12d-f1cd74475213"
/>
`mate-h/frustum-cull-fix`
<img width="1283" alt="frustum-cull-fix"
src="https://github.com/user-attachments/assets/8a9ccb2a-64b6-4d5e-a17d-ac4798da5b51"
/>
The `check_visibility` system currently follows this algorithm:
1. Store all meshes that were visible last frame in the
`PreviousVisibleMeshes` set.
2. Determine which meshes are visible. For each such visible mesh,
remove it from `PreviousVisibleMeshes`.
3. Mark all meshes that remain in `PreviousVisibleMeshes` as invisible.
This algorithm would be correct if the `check_visibility` were the only
system that marked meshes visible. However, it's not: the shadow-related
systems `check_dir_light_mesh_visibility` and
`check_point_light_mesh_visibility` can as well. This results in the
following sequence of events for meshes that are in a shadow map but
*not* visible from a camera:
A. `check_visibility` runs, finds that no camera contains these meshes,
and marks them hidden, which sets the changed flag.
B. `check_dir_light_mesh_visibility` and/or
`check_point_light_mesh_visibility` run, discover that these meshes
are visible in the shadow map, and marks them as visible, again
setting the `ViewVisibility` changed flag.
C. During the extraction phase, the mesh extraction system sees that
`ViewVisibility` is changed and re-extracts the mesh.
This is inefficient and results in needless work during rendering.
This patch fixes the issue in two ways:
* The `check_dir_light_mesh_visibility` and
`check_point_light_mesh_visibility` systems now remove meshes that they
discover from `PreviousVisibleMeshes`.
* Step (3) above has been moved from `check_visibility` to a separate
system, `mark_newly_hidden_entities_invisible`. This system runs after
all visibility-determining systems, ensuring that
`PreviousVisibleMeshes` contains only those meshes that truly became
invisible on this frame.
This fix dramatically improves the performance of [the Caldera
benchmark], when combined with several other patches I've submitted.
[the Caldera benchmark]:
https://github.com/DGriffin91/bevy_caldera_scene
PR #17688 broke motion vector computation, and therefore motion blur,
because it enabled retention of `MeshInputUniform`s, and
`MeshInputUniform`s contain the indices of the previous frame's
transform and the previous frame's skinned mesh joint matrices. On frame
N, if a `MeshInputUniform` is retained on GPU from the previous frame,
the `previous_input_index` and `previous_skin_index` would refer to the
indices for frame N - 2, not the index for frame N - 1.
This patch fixes the problems. It solves these issues in two different
ways, one for transforms and one for skins:
1. To fix transforms, this patch supplies the *frame index* to the
shader as part of the view uniforms, and specifies which frame index
each mesh's previous transform refers to. So, in the situation described
above, the frame index would be N, the previous frame index would be N -
1, and the `previous_input_frame_number` would be N - 2. The shader can
now detect this situation and infer that the mesh has been retained, and
can therefore conclude that the mesh's transform hasn't changed.
2. To fix skins, this patch replaces the explicit `previous_skin_index`
with an invariant that the index of the joints for the current frame and
the index of the joints for the previous frame are the same. This means
that the `MeshInputUniform` never has to be updated even if the skin is
animated. The downside is that we have to copy joint matrices from the
previous frame's buffer to the current frame's buffer in
`extract_skins`.
The rationale behind (2) is that we currently have no mechanism to
detect when joints that affect a skin have been updated, short of
comparing all the transforms and setting a flag for
`extract_meshes_for_gpu_building` to consume, which would regress
performance as we want `extract_skins` and
`extract_meshes_for_gpu_building` to be able to run in parallel.
To test this change, use `cargo run --example motion_blur`.
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, Bevy rebuilds the buffer containing all the transforms for
joints every frame, during the extraction phase. This is inefficient in
cases in which many skins are present in the scene and their joints
don't move, such as the Caldera test scene.
To address this problem, this commit switches skin extraction to use a
set of retained GPU buffers with allocations managed by the offset
allocator. I use fine-grained change detection in order to determine
which skins need updating. Note that the granularity is on the level of
an entire skin, not individual joints. Using the change detection at
that level would yield poor performance in common cases in which an
entire skin is animated at once. Also, this patch yields additional
performance from the fact that changing joint transforms no longer
requires the skinned mesh to be re-extracted.
Note that this optimization can be a double-edged sword. In
`many_foxes`, fine-grained change detection regressed the performance of
`extract_skins` by 3.4x. This is because every joint is updated every
frame in that example, so change detection is pointless and is pure
overhead. Because the `many_foxes` workload is actually representative
of animated scenes, this patch includes a heuristic that disables
fine-grained change detection if the number of transformed entities in
the frame exceeds a certain fraction of the total number of joints.
Currently, this threshold is set to 25%. Note that this is a crude
heuristic, because it doesn't distinguish between the number of
transformed *joints* and the number of transformed *entities*; however,
it should be good enough to yield the optimum code path most of the
time.
Finally, this patch fixes a bug whereby skinned meshes are actually
being incorrectly retained if the buffer offsets of the joints of those
skinned meshes changes from frame to frame. To fix this without
retaining skins, we would have to re-extract every skinned mesh every
frame. Doing this was a significant regression on Caldera. With this PR,
by contrast, mesh joints stay at the same buffer offset, so we don't
have to update the `MeshInputUniform` containing the buffer offset every
frame. This also makes PR #17717 easier to implement, because that PR
uses the buffer offset from the previous frame, and the logic for
calculating that is simplified if the previous frame's buffer offset is
guaranteed to be identical to that of the current frame.
On Caldera, this patch reduces the time spent in `extract_skins` from
1.79 ms to near zero. On `many_foxes`, this patch regresses the
performance of `extract_skins` by approximately 10%-25%, depending on
the number of foxes. This has only a small impact on frame rate.
The GPU can fill out many of the fields in `IndirectParametersMetadata`
using information it already has:
* `early_instance_count` and `late_instance_count` are always
initialized to zero.
* `mesh_index` is already present in the work item buffer as the
`input_index` of the first work item in each batch.
This patch moves these fields to a separate buffer, the *GPU indirect
parameters metadata* buffer. That way, it avoids having to write them on
CPU during `batch_and_prepare_binned_render_phase`. This effectively
reduces the number of bits that that function must write per mesh from
160 to 64 (in addition to the 64 bits per mesh *instance*).
Additionally, this PR refactors `UntypedPhaseIndirectParametersBuffers`
to add another layer, `MeshClassIndirectParametersBuffers`, which allows
abstracting over the buffers corresponding indexed and non-indexed
meshes. This patch doesn't make much use of this abstraction, but
forthcoming patches will, and it's overall a cleaner approach.
This didn't seem to have much of an effect by itself on
`batch_and_prepare_binned_render_phase` time, but subsequent PRs
dependent on this PR yield roughly a 2× speedup.
Appending to these vectors is performance-critical in
`batch_and_prepare_binned_render_phase`, so `RawBufferVec`, which
doesn't have the overhead of `encase`, is more appropriate.
The `output_index` field is only used in direct mode, and the
`indirect_parameters_index` field is only used in indirect mode.
Consequently, we can combine them into a single field, reducing the size
of `PreprocessWorkItem`, which
`batch_and_prepare_{binned,sorted}_render_phase` must construct every
frame for every mesh instance, from 96 bits to 64 bits.
# Objective
Update typos, fix new typos.
1.29.6 was just released to fix an
[issue](https://github.com/crate-ci/typos/issues/1228) where January's
corrections were not included in the binaries for the last release.
Reminder: typos can be tossed in the monthly [non-critical corrections
issue](https://github.com/crate-ci/typos/issues/1221).
## Solution
I chose to allow `implementors`, because a good argument seems to be
being made [here](https://github.com/crate-ci/typos/issues/1226) and
there is now a PR to address that.
## Discussion
Should I exclude `bevy_mikktspace`?
At one point I think we had an informal policy of "don't mess with
mikktspace until https://github.com/bevyengine/bevy/pull/9050 is merged"
but it doesn't seem like that is likely to be merged any time soon.
I think these particular corrections in mikktspace are fine because
- The same typo mistake seems to have been fixed in that PR
- The entire file containing these corrections was deleted in that PR
## Typo of the Month
correspindong -> corresponding
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.
* Use texture atomics rather than buffer atomics for the visbuffer
(haven't tested perf on a raster-heavy scene yet)
* Unfortunately to clear the visbuffer we now need a compute pass to
clear it. Using wgpu's clear_texture function internally uses a buffer
-> image copy that's insanely expensive. Ideally it should be using
vkCmdClearColorImage, which I've opened an issue for
https://github.com/gfx-rs/wgpu/issues/7090. For now we'll have to stick
with a custom compute pass and all the extra code that brings.
* Faster resolve depth pass by discarding 0 depth pixels instead of
redundantly writing zero (2x faster for big depth textures like shadow
views)
## Objective
Get rid of a redundant Cargo feature flag.
## Solution
Use the built-in `target_abi = "sim"` instead of a custom Cargo feature
flag, which is set for the iOS (and visionOS and tvOS) simulator. This
has been stable since Rust 1.78.
In the future, some of this may become redundant if Wgpu implements
proper supper for the iOS Simulator:
https://github.com/gfx-rs/wgpu/issues/7057
CC @mockersf who implemented [the original
fix](https://github.com/bevyengine/bevy/pull/10178).
## Testing
- Open mobile example in Xcode.
- Launch the simulator.
- See that no errors are emitted.
- Remove the code cfg-guarded behind `target_abi = "sim"`.
- See that an error now happens.
(I haven't actually performed these steps on the latest `main`, because
I'm hitting an unrelated error (EDIT: It was
https://github.com/bevyengine/bevy/pull/17637). But tested it on
0.15.0).
---
## Migration Guide
> If you're using a project that builds upon the mobile example, remove
the `ios_simulator` feature from your `Cargo.toml` (Bevy now handles
this internally).
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.
# Objective
- Fixes#17797
## Solution
- `mesh` in `bevy_pbr::mesh_bindings` is behind a `ifndef
MESHLET_MESH_MATERIAL_PASS`. also gate `get_tag` which uses this `mesh`
## Testing
- Run the meshlet example
# Objective
Since previously we only had the alpha channel available, we stored the
mean of the transmittance in the aerial view lut, resulting in a grayer
fog than should be expected.
## Solution
- Calculate transmittance to scene in `render_sky` with two samples from
the transmittance lut
- use dual-source blending to effectively have per-component alpha
blending
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
Because of mesh preprocessing, users cannot rely on
`@builtin(instance_index)` in order to reference external data, as the
instance index is not stable, either from frame to frame or relative to
the total spawn order of mesh instances.
## Solution
Add a user supplied mesh index that can be used for referencing external
data when drawing instanced meshes.
Closes#13373
## Testing
Benchmarked `many_cubes` showing no difference in total frame time.
## Showcase
https://github.com/user-attachments/assets/80620147-aafc-4d9d-a8ee-e2149f7c8f3b
---------
Co-authored-by: IceSentry <IceSentry@users.noreply.github.com>
# Objective
https://github.com/bevyengine/bevy/issues/17746
## Solution
- Change `Image.data` from being a `Vec<u8>` to a `Option<Vec<u8>>`
- Added functions to help with creating images
## Testing
- Did you test these changes? If so, how?
All current tests pass
Tested a variety of existing examples to make sure they don't crash
(they don't)
- If relevant, what platforms did you test these changes on, and are
there any important ones you can't test?
Linux x86 64-bit NixOS
---
## Migration Guide
Code that directly access `Image` data will now need to use unwrap or
handle the case where no data is provided.
Behaviour of new_fill slightly changed, but not in a way that is likely
to affect anything. It no longer panics and will fill the whole texture
instead of leaving black pixels if the data provided is not a nice
factor of the size of the image.
---------
Co-authored-by: IceSentry <IceSentry@users.noreply.github.com>
# Objective
https://github.com/bevyengine/bevy/pull/16966 tried to fix a bug where
`slot` wasn't passed to `parallaxed_uv` when not running under bindless,
but failed to account for meshlets. This surfaces on macOS where
bindless is disabled.
## Solution
Lift the slot variable out of the bindless condition so it's always
available.
Didn't remove WgpuWrapper. Not sure if it's needed or not still.
## Testing
- Did you test these changes? If so, how? Example runner
- Are there any parts that need more testing? Web (portable atomics
thingy?), DXC.
## Migration Guide
- Bevy has upgraded to [wgpu
v24](https://github.com/gfx-rs/wgpu/blob/trunk/CHANGELOG.md#v2400-2025-01-15).
- When using the DirectX 12 rendering backend, the new priority system
for choosing a shader compiler is as follows:
- If the `WGPU_DX12_COMPILER` environment variable is set at runtime, it
is used
- Else if the new `statically-linked-dxc` feature is enabled, a custom
version of DXC will be statically linked into your app at compile time.
- Else Bevy will look in the app's working directory for
`dxcompiler.dll` and `dxil.dll` at runtime.
- Else if they are missing, Bevy will fall back to FXC (not recommended)
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
Co-authored-by: IceSentry <c.giguere42@gmail.com>
Co-authored-by: François Mockers <francois.mockers@vleue.com>
# Objective
- publish script copy the license files to all subcrates, meaning that
all publish are dirty. this breaks git verification of crates
- the order and list of crates to publish is manually maintained,
leading to error. cargo 1.84 is more strict and the list is currently
wrong
## Solution
- duplicate all the licenses to all crates and remove the
`--allow-dirty` flag
- instead of a manual list of crates, get it from `cargo package
--workspace`
- remove the `--no-verify` flag to... verify more things?
# 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`.
Right now, we key the cached light change ticks off the `LightEntity`.
This uses the render world entity, which isn't stable between frames.
Thus in practice few shadows are retained from frame to frame. This PR
fixes the issue by keying off the `RetainedViewEntity` instead, which is
designed to be stable from frame to frame.
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.
Right now, meshes aren't grouped together based on the bindless texture
slab when drawing shadows. This manifests itself as flickering in
Bistro. I believe that there are two causes of this:
1. Alpha masked shadows may try to sample from the wrong texture,
causing the alpha mask to appear and disappear.
2. Objects may try to sample from the blank textures that we pad out the
bindless slabs with, causing them to vanish intermittently.
This commit fixes the issue by including the material bind group ID as
part of the shadow batch set key, just as we do for the prepass and main
pass.
