Right now, we always include distance fog in the shader, which is
unfortunate as it's complex code and is rare. This commit changes it to
be a `#define` instead. I haven't confirmed that removing distance fog
meaningfully reduces VGPR usage, but it can't hurt.
This commit makes Bevy use change detection to only update
`RenderMaterialInstances` and `RenderMeshMaterialIds` when meshes have
been added, changed, or removed. `extract_mesh_materials`, the system
that extracts these, now follows the pattern that
`extract_meshes_for_gpu_building` established.
This improves frame time of `many_cubes` from 3.9ms to approximately
3.1ms, which slightly surpasses the performance of Bevy 0.14.
(Resubmitted from #16878 to clean up history.)
---------
Co-authored-by: Charlotte McElwain <charlotte.c.mcelwain@gmail.com>
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
# Objective
`bevy_ecs`'s `system` module is something of a grab bag, and *very*
large. This is particularly true for the `system_param` module, which is
more than 2k lines long!
While it could be defensible to put `Res` and `ResMut` there (lol no
they're in change_detection.rs, obviously), it doesn't make any sense to
put the `Resource` trait there. This is confusing to navigate (and
painful to work on and review).
## Solution
- Create a root level `bevy_ecs/resource.rs` module to mirror
`bevy_ecs/component.rs`
- move the `Resource` trait to that module
- move the `Resource` derive macro to that module as well (Rust really
likes when you pun on the names of the derive macro and trait and put
them in the same path)
- fix all of the imports
## Notes to reviewers
- We could probably move more stuff into here, but I wanted to keep this
PR as small as possible given the absurd level of import changes.
- This PR is ground work for my upcoming attempts to store resource data
on components (resources-as-entities). Splitting this code out will make
the work and review a bit easier, and is the sort of overdue refactor
that's good to do as part of more meaningful work.
## Testing
cargo build works!
## Migration Guide
`bevy_ecs::system::Resource` has been moved to
`bevy_ecs::resource::Resource`.
# Objective
The existing `RelationshipSourceCollection` uses `Vec` as the only
possible backing for our relationships. While a reasonable choice,
benchmarking use cases might reveal that a different data type is better
or faster.
For example:
- Not all relationships require a stable ordering between the
relationship sources (i.e. children). In cases where we a) have many
such relations and b) don't care about the ordering between them, a hash
set is likely a better datastructure than a `Vec`.
- The number of children-like entities may be small on average, and a
`smallvec` may be faster
## Solution
- Implement `RelationshipSourceCollection` for `EntityHashSet`, our
custom entity-optimized `HashSet`.
-~~Implement `DoubleEndedIterator` for `EntityHashSet` to make things
compile.~~
- This implementation was cursed and very surprising.
- Instead, by moving the iterator type on `RelationshipSourceCollection`
from an erased RPTIT to an explicit associated type we can add a trait
bound on the offending methods!
- Implement `RelationshipSourceCollection` for `SmallVec`
## Testing
I've added a pair of new tests to make sure this pattern compiles
successfully in practice!
## Migration Guide
`EntityHashSet` and `EntityHashMap` are no longer re-exported in
`bevy_ecs::entity` directly. If you were not using `bevy_ecs` / `bevy`'s
`prelude`, you can access them through their now-public modules,
`hash_set` and `hash_map` instead.
## Notes to reviewers
The `EntityHashSet::Iter` type needs to be public for this impl to be
allowed. I initially renamed it to something that wasn't ambiguous and
re-exported it, but as @Victoronz pointed out, that was somewhat
unidiomatic.
In
1a8564898f,
I instead made the `entity_hash_set` public (and its `entity_hash_set`)
sister public, and removed the re-export. I prefer this design (give me
module docs please), but it leads to a lot of churn in this PR.
Let me know which you'd prefer, and if you'd like me to split that
change out into its own micro PR.
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.
# Objective
allow setting ambient light via component on cameras.
arguably fixes#7193
note i chose to use a component rather than an entity since it was not
clear to me how to manage multiple ambient sources for a single
renderlayer, and it makes for a very small changeset.
## Solution
- make ambient light a component as well as a resource
- extract it
- use the component if present on a camera, fallback to the resource
## Testing
i added
```rs
if index == 1 {
commands.entity(camera).insert(AmbientLight{
color: Color::linear_rgba(1.0, 0.0, 0.0, 1.0),
brightness: 1000.0,
..Default::default()
});
}
```
at line 84 of the split_screen example
---------
Co-authored-by: François Mockers <mockersf@gmail.com>
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
Rework / build on #17043 to simplify the implementation. #17043 should
be merged first, and the diff from this PR will get much nicer after it
is merged (this PR is net negative LOC).
## Solution
1. Command and EntityCommand have been vastly simplified. No more marker
components. Just one function.
2. Command and EntityCommand are now generic on the return type. This
enables result-less commands to exist, and allows us to statically
distinguish between fallible and infallible commands, which allows us to
skip the "error handling overhead" for cases that don't need it.
3. There are now only two command queue variants: `queue` and
`queue_fallible`. `queue` accepts commands with no return type.
`queue_fallible` accepts commands that return a Result (specifically,
one that returns an error that can convert to
`bevy_ecs::result::Error`).
4. I've added the concept of the "default error handler", which is used
by `queue_fallible`. This is a simple direct call to the `panic()` error
handler by default. Users that want to override this can enable the
`configurable_error_handler` cargo feature, then initialize the
GLOBAL_ERROR_HANDLER value on startup. This is behind a flag because
there might be minor overhead with `OnceLock` and I'm guessing this will
be a niche feature. We can also do perf testing with OnceLock if someone
really wants it to be used unconditionally, but I don't personally feel
the need to do that.
5. I removed the "temporary error handler" on Commands (and all code
associated with it). It added more branching, made Commands bigger /
more expensive to initialize (note that we construct it at high
frequencies / treat it like a pointer type), made the code harder to
follow, and introduced a bunch of additional functions. We instead rely
on the new default error handler used in `queue_fallible` for most
things. In the event that a custom handler is required,
`handle_error_with` can be used.
6. EntityCommand now _only_ supports functions that take
`EntityWorldMut` (and all existing entity commands have been ported).
Removing the marker component from EntityCommand hinged on this change,
but I strongly believe this is for the best anyway, as this sets the
stage for more efficient batched entity commands.
7. I added `EntityWorldMut::resource` and the other variants for more
ergonomic resource access on `EntityWorldMut` (removes the need for
entity.world_scope, which also incurs entity-lookup overhead).
## Open Questions
1. I believe we could merge `queue` and `queue_fallible` into a single
`queue` which accepts both fallible and infallible commands (via the
introduction of a `QueueCommand` trait). Is this desirable?
# 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.
# Objective
Remove some outdated docs from 0.15 that mention a removed function.
## Solution
In `pbr_functions.wgsl`, I think it's fine to just remove the mention.
In `meshlet/asset.rs`, I think it would be nice to still have a note on
how texture samples should be done. Unfortunately, there isn't a nice
abstraction for it any more. Current workaround, for reference:
b386d08d0f/crates/bevy_pbr/src/render/pbr_fragment.wgsl (L184-L208)
For now, I've just removed the mention.
# Objective
the `get` function on [`InstanceInputUniformBuffer`] seems very
error-prone. This PR hopes to fix this.
## Solution
Do a few checks to ensure the index is in bounds and that the `BDI` is
not removed.
Return `Option<BDI>` instead of `BDI`.
## Testing
- Did you test these changes? If so, how?
added a test to verify that the instance buffer works correctly
## Future Work
Performance decreases when using .binary_search(). However this is
likely due to the fact that [`InstanceInputUniformBuffer::get`] for now
is never used, and only get_unchecked.
## Migration Guide
`InstanceInputUniformBuffer::get` now returns `Option<BDI>` instead of
`BDI` to reduce panics. If you require the old functionality of
`InstanceInputUniformBuffer::get` consider using
`InstanceInputUniformBuffer::get_unchecked`.
---------
Co-authored-by: Tim Overbeek <oorbeck@gmail.com>
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.
Derived `Default` for all public unit structs that already derive from
`Component`. This allows them to be used more easily as required
components.
To avoid clutter in tests/examples, only public components were
affected, but this could easily be expanded to affect all unit
components.
Fixes#17052.
# Objective
Use the latest version of `typos` and fix the typos that it now detects
# Additional Info
By the way, `typos` has a "low priority typo suggestions issue" where we
can throw typos we find that `typos` doesn't catch.
(This link may go stale) https://github.com/crate-ci/typos/issues/1200
# Objective
Some sort calls and `Ord` impls are unnecessarily complex.
## Solution
Rewrite the "match on cmp, if equal do another cmp" as either a
comparison on tuples, or `Ordering::then_with`, depending on whether the
compare keys need construction.
`sort_by` -> `sort_by_key` when symmetrical. Do the same for
`min_by`/`max_by`.
Note that `total_cmp` can only work with `sort_by`, and not on tuples.
When sorting collected query results that contain
`Entity`/`MainEntity`/`RenderEntity` in their `QueryData`, with that
`Entity` in the sort key:
stable -> unstable sort (all queried entities are unique)
If key construction is not simple, switch to `sort_by_cached_key` when
possible.
Sorts that are only performed to discover the maximal element are
replaced by `max_by_key`.
Dedicated comparison functions and structs are removed where simple.
Derive `PartialOrd`/`Ord` when useful.
Misc. closure style inconsistencies.
## Testing
- Existing tests.
This commit makes the following changes:
* `IndirectParametersBuffer` has been changed from a `BufferVec` to a
`RawBufferVec`. This won about 20us or so on Bistro by avoiding `encase`
overhead.
* The methods on the `GetFullBatchData` trait no longer have the
`entity` parameter, as it was unused.
* `PreprocessWorkItem`, which specifies a transform-and-cull operation,
now supplies the mesh instance uniform output index directly instead of
having the shader look it up from the indirect draw parameters.
Accordingly, the responsibility of writing the output index to the
indirect draw parameters has been moved from the CPU to the GPU. This is
in preparation for retained indirect instance draw commands, where the
mesh instance uniform output index may change from frame to frame, while
the indirect instance draw commands will be cached. We won't want the
CPU to have to upload the same indirect draw parameters again and again
if a batch didn't change from frame to frame.
* `batch_and_prepare_binned_render_phase` and
`batch_and_prepare_sorted_render_phase` now allocate indirect draw
commands for an entire batch set at a time when possible, instead of one
batch at a time. This change will allow us to retain the indirect draw
commands for whole batch sets.
* `GetFullBatchData::get_batch_indirect_parameters_index` has been
replaced with `GetFullBatchData::write_batch_indirect_parameters`, which
takes an offset and writes into it instead of allocating. This is
necessary in order to use the optimization mentioned in the previous
point.
* At the WGSL level, `IndirectParameters` has been factored out into
`mesh_preprocess_types.wgsl`. This is because we'll need a new compute
shader that zeroes out the instance counts in preparation for a new
frame. That shader will need to access `IndirectParameters`, so it was
moved to a separate file.
* Bins are no longer raw vectors but are instances of a separate type,
`RenderBin`. This is so that the bin can eventually contain its retained
batches.
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.
# Objective
- Running example `load_gltf` when not using bindless gives this error
```
ERROR bevy_render::render_resource::pipeline_cache: failed to process shader:
error: no definition in scope for identifier: 'slot'
┌─ crates/bevy_pbr/src/render/pbr_fragment.wgsl:153:13
│
153 │ slot,
│ ^^^^ unknown identifier
│
= no definition in scope for identifier: 'slot'
```
- since https://github.com/bevyengine/bevy/pull/16825
## Solution
- Set `slot` to the expected value when not mindless
- Also use it for `uv_b`
## Testing
- Run example `load_gltf` on a Mac or in wasm
# Objective
Fixes: #16578
## Solution
This is a patch fix, proper fix requires a breaking change.
Added `Panic` enum variant and using is as the system meta default.
Warn once behavior can be enabled same way disabling panic (originally
disabling wans) is.
To fix an issue with the current architecture, where **all** combinator
system params get checked together,
combinator systems only check params of the first system.
This will result in old, panicking behavior on subsequent systems and
will be fixed in 0.16.
## Testing
Ran unit tests and `fallible_params` example.
---------
Co-authored-by: François Mockers <mockersf@gmail.com>
Co-authored-by: François Mockers <francois.mockers@vleue.com>
This commit fixes the following regressions:
1. Transmission-specific calls to shader lighting functions didn't pass
the `enable_diffuse` parameter, breaking the `transmission` example.
2. The combination of bindless `StandardMaterial` and bindless lightmaps
caused us to blow past the 128 texture limit on M1/M2 chips in some
cases, in particular the `depth_of_field` example.
https://github.com/gfx-rs/wgpu/issues/3334 should fix this, but in the
meantime this patch reduces the number of bindless lightmaps from 16 to
4 in order to stay under the limit.
3. The renderer was crashing on startup on Adreno 610 chips. This PR
simply disables bindless on Adreno 610 and lower.
# Objective
`EntityHashMap` and `EntityHashSet` iterators do not implement
`EntitySetIterator`.
## Solution
Make them newtypes instead of aliases. The methods that create the
iterators can then produce their own newtypes that carry the `Hasher`
generic and implement `EntitySetIterator`. Functionality remains the
same otherwise.
There are some other small benefits, f.e. the removal of `with_hasher`
associated functions, and the ability to implement more traits
ourselves.
`MainEntityHashMap` and `MainEntityHashSet` are currently left as the
previous type aliases, because supporting general `TrustedEntityBorrow`
hashing is more complex. However, it can also be done.
## Testing
Pre-existing `EntityHashMap` tests.
## Migration Guide
Users of `with_hasher` and `with_capacity_and_hasher` on
`EntityHashMap`/`Set` must now use `new` and `with_capacity`
respectively.
If the non-newtyped versions are required, they can be obtained via
`Deref`, `DerefMut` or `into_inner` calls.
# Objective
When preparing `GpuImage`s, we currently discard the
`depth_or_array_layers` of the `Image`'s size by converting it into a
`UVec2`.
Fixes#16715.
## Solution
Change `GpuImage::size` to `Extent3d`, and just pass that through when
creating `GpuImage`s.
Also copy the `aspect_ratio`, and `size` (now `size_2d` for
disambiguation from the field) functions from `Image` to `GpuImage` for
ease of use with 2D textures.
I originally copied all size-related functions (like `width`, and
`height`), but i think they are unnecessary considering how visible the
`size` field on `GpuImage` is compared to `Image`.
## Testing
Tested via `cargo r -p ci` for everything except docs, when generating
docs it keeps spitting out a ton of
```
error[E0554]: `#![feature]` may not be used on the stable release channel
--> crates/bevy_dylib/src/lib.rs:1:21
|
1 | #![cfg_attr(docsrs, feature(doc_auto_cfg))]
|
```
Not sure why this is happening, but it also happens without my changes,
so it's almost certainly some strange issue specific to my machine.
## Migration Guide
- `GpuImage::size` is now an `Extent3d`. To easily get 2D size, use
`size_2d()`.
The only thing that was preventing `extract_meshes_for_gpu_building` and
`extract_mesh_materials` from running in parallel was the
`ResMut<RenderMeshMaterialIds>`. This lookup can be safely moved to the
`collect_meshes_for_gpu_building` phase, which runs after the extraction
phase.
This results in a small win on `many_cubes`. `extract_mesh_materials` is
currently nonretained, so it's still slow, but running it in parallel is
an easy win.
Before:
After:
This PR adds support for *mixed lighting* to Bevy, whereby some parts of
the scene are lightmapped, while others take part in real-time lighting.
(Here *real-time lighting* means lighting at runtime via the PBR shader,
as opposed to precomputed light using lightmaps.) It does so by adding a
new field, `affects_lightmapped_meshes` to `IrradianceVolume` and
`AmbientLight`, and a corresponding field
`affects_lightmapped_mesh_diffuse` to `DirectionalLight`, `PointLight`,
`SpotLight`, and `EnvironmentMapLight`. By default, this value is set to
true; when set to false, the light contributes nothing to the diffuse
irradiance component to meshes with lightmaps.
Note that specular light is unaffected. This is because the correct way
to bake specular lighting is *directional lightmaps*, which we have no
support for yet.
There are two general ways I expect this field to be used:
1. When diffuse indirect light is baked into lightmaps, irradiance
volumes and reflection probes shouldn't contribute any diffuse light to
the static geometry that has a lightmap. That's because the baking tool
should have already accounted for it, and in a higher-quality fashion,
as lightmaps typically offer a higher effective texture resolution than
the light probe does.
2. When direct diffuse light is baked into a lightmap, punctual lights
shouldn't contribute any diffuse light to static geometry with a
lightmap, to avoid double-counting. It may seem odd to bake *direct*
light into a lightmap, as opposed to indirect light. But there is a use
case: in a scene with many lights, avoiding light leaks requires shadow
mapping, which quickly becomes prohibitive when many lights are
involved. Baking lightmaps allows light leaks to be eliminated on static
geometry.
A new example, `mixed_lighting`, has been added. It demonstrates a sofa
(model from the [glTF Sample Assets]) that has been lightmapped offline
using [Bakery]. It has four modes:
1. In *baked* mode, all objects are locked in place, and all the diffuse
direct and indirect light has been calculated ahead of time. Note that
the bottom of the sphere has a red tint from the sofa, illustrating that
the baking tool captured indirect light for it.
2. In *mixed direct* mode, lightmaps capturing diffuse direct and
indirect light have been pre-calculated for the static objects, but the
dynamic sphere has real-time lighting. Note that, because the diffuse
lighting has been entirely pre-calculated for the scenery, the dynamic
sphere casts no shadow. In a real app, you would typically use real-time
lighting for the most important light so that dynamic objects can shadow
the scenery and relegate baked lighting to the less important lights for
which shadows aren't as important. Also note that there is no red tint
on the sphere, because there is no global illumination applied to it. In
an actual game, you could fix this problem by supplementing the
lightmapped objects with an irradiance volume.
3. In *mixed indirect* mode, all direct light is calculated in
real-time, and the static objects have pre-calculated indirect lighting.
This corresponds to the mode that most applications are expected to use.
Because direct light on the scenery is computed dynamically, shadows are
fully supported. As in mixed direct mode, there is no global
illumination on the sphere; in a real application, irradiance volumes
could be used to supplement the lightmaps.
4. In *real-time* mode, no lightmaps are used at all, and all punctual
lights are rendered in real-time. No global illumination exists.
In the example, you can click around to move the sphere, unless you're
in baked mode, in which case the sphere must be locked in place to be
lit correctly.
## Showcase
Baked mode:
Mixed direct mode:
Mixed indirect mode (default):
Real-time mode:
## Migration guide
* The `AmbientLight` resource, the `IrradianceVolume` component, and the
`EnvironmentMapLight` component now have `affects_lightmapped_meshes`
fields. If you don't need to use that field (for example, if you aren't
using lightmaps), you can safely set the field to true.
* `DirectionalLight`, `PointLight`, and `SpotLight` now have
`affects_lightmapped_mesh_diffuse` fields. If you don't need to use that
field (for example, if you aren't using lightmaps), you can safely set
the field to true.
[glTF Sample Assets]:
https://github.com/KhronosGroup/glTF-Sample-Assets/tree/main
[Bakery]:
https://geom.io/bakery/wiki/index.php?title=Bakery_-_GPU_Lightmapper
This commit allows Bevy to bind 16 lightmaps at a time, if the current
platform supports bindless textures. Naturally, if bindless textures
aren't supported, Bevy falls back to binding only a single lightmap at a
time. As lightmaps are usually heavily atlased, I doubt many scenes will
use more than 16 lightmap textures.
This has little performance impact now, but it's desirable for us to
reap the benefits of multidraw and bindless textures on scenes that use
lightmaps. Otherwise, we might have to break batches in order to switch
those lightmaps.
Additionally, this PR slightly reduces the cost of binning because it
makes the lightmap index in `Opaque3dBinKey` 32 bits instead of an
`AssetId`.
## Migration Guide
* The `Opaque3dBinKey::lightmap_image` field is now
`Opaque3dBinKey::lightmap_slab`, which is a lightweight identifier for
an entire binding array of lightmaps.
This patch replaces the undocumented `NoGpuCulling` component with a new
component, `NoIndirectDrawing`, effectively turning indirect drawing on
by default. Indirect mode is needed for the recently-landed multidraw
feature (#16427). Since multidraw is such a win for performance, when
that feature is supported the small performance tax that indirect mode
incurs is virtually always worth paying.
To ensure that custom drawing code such as that in the
`custom_shader_instancing` example continues to function, this commit
additionally makes GPU culling take the `NoFrustumCulling` component
into account.
This PR is an alternative to #16670 that doesn't break the
`custom_shader_instancing` example. **PR #16755 should land first in
order to avoid breaking deferred rendering, as multidraw currently
breaks it**.
## Migration Guide
* Indirect drawing (GPU culling) is now enabled by default, so the
`GpuCulling` component is no longer available. To disable indirect mode,
which may be useful with custom render nodes, add the new
`NoIndirectDrawing` component to your camera.
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>
Updating dependencies; adopted version of #15696. (Supercedes #15696.)
Long answer: hashbrown is no longer using ahash by default, meaning that
we can't use the default-hasher methods with ahasher. So, we have to use
the longer-winded versions instead. This takes the opportunity to also
switch our default hasher as well, but without actually enabling the
default-hasher feature for hashbrown, meaning that we'll be able to
change our hasher more easily at the cost of all of these method calls
being obnoxious forever.
One large change from 0.15 is that `insert_unique_unchecked` is now
`unsafe`, and for cases where unsafe code was denied at the crate level,
I replaced it with `insert`.
## Migration Guide
`bevy_utils` has updated its version of `hashbrown` to 0.15 and now
defaults to `foldhash` instead of `ahash`. This means that if you've
hard-coded your hasher to `bevy_utils::AHasher` or separately used the
`ahash` crate in your code, you may need to switch to `foldhash` to
ensure that everything works like it does in Bevy.
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.
This patch makes shadows use multidraw when the camera they'll be drawn
to has the `GpuCulling` component. This results in a significant
reduction in drawcalls; Bistro Exterior drops to 3 drawcalls for each
shadow cascade.
Note that PR #16670 will remove the `GpuCulling` component, making
shadows automatically use multidraw. Beware of that when testing this
patch; before #16670 lands, you'll need to manually add `GpuCulling` to
your camera in order to see any performance benefits.
PR #15756 made us create temporary render entities for all visible
objects, even if they had no render world counterpart. This regressed
our `many_cubes` time from about 3.59 ms/frame to 4.66 ms/frame.
This commit changes that behavior to use `Entity::PLACEHOLDER` instead
of creating a temporary render entity. This improves our `many_cubes`
time from 5.66 ms/frame to 3.96 ms/frame, a 43% speedup.
I tested 3D, 2D gizmos, and UI and they seem to work.
See the following graph of `many_cubes` frame time (lower is better). PR
#15756 is the one in October.
This commit removes the logic that attempted to keep the
`MeshInputUniform` buffer contiguous. Not only was it slow and complex,
but it was also incorrect, which caused #16686 and #16690. I changed the
logic to simply maintain a free list of unused slots in the buffer and
preferentially fill them when pushing new mesh input uniforms.
Closes#16686.
Closes#16690.
# Objective
- A `Trigger` has multiple associated `Entity`s - the entity observing
the event, and the entity that was targeted by the event.
- The field `entity: Entity` encodes no semantic information about what
the entity is used for, you can already tell that it's an `Entity` by
the type signature!
## Solution
- Rename `trigger.entity()` to `trigger.target()`
---
## Changelog
- `Trigger`s are associated with multiple entities. `Trigger::entity()`
has been renamed to `Trigger::target()` to reflect the semantics of the
entity being returned.
## Migration Guide
- Rename `Trigger::entity()` to `Trigger::target()`.
- Rename `ObserverTrigger::entity` to `ObserverTrigger::target`
This commit adds support for *multidraw*, which is a feature that allows
multiple meshes to be drawn in a single drawcall. `wgpu` currently
implements multidraw on Vulkan, so this feature is only enabled there.
Multiple meshes can be drawn at once if they're in the same vertex and
index buffers and are otherwise placed in the same bin. (Thus, for
example, at present the materials and textures must be identical, but
see #16368.) Multidraw is a significant performance improvement during
the draw phase because it reduces the number of rebindings, as well as
the number of drawcalls.
This feature is currently only enabled when GPU culling is used: i.e.
when `GpuCulling` is present on a camera. Therefore, if you run for
example `scene_viewer`, you will not see any performance improvements,
because `scene_viewer` doesn't add the `GpuCulling` component to its
camera.
Additionally, the multidraw feature is only implemented for opaque 3D
meshes and not for shadows or 2D meshes. I plan to make GPU culling the
default and to extend the feature to shadows in the future. Also, in the
future I suspect that polyfilling multidraw on APIs that don't support
it will be fruitful, as even without driver-level support use of
multidraw allows us to avoid expensive `wgpu` rebindings.
The bindless PR (#16368) broke some examples:
* `specialized_mesh_pipeline` and `custom_shader_instancing` failed
because they expect to be able to render a mesh with no material, by
overriding enough of the render pipeline to be able to do so. This PR
fixes the issue by restoring the old behavior in which we extract meshes
even if they have no material.
* `texture_binding_array` broke because it doesn't implement
`AsBindGroup::unprepared_bind_group`. This was tricky to fix because
there's a very good reason why `texture_binding_array` doesn't implement
that method: there's no sensible way to do so with `wgpu`'s current
bindless API, due to its multiple levels of borrowed references. To fix
the example, I split `MaterialBindGroup` into
`MaterialBindlessBindGroup` and `MaterialNonBindlessBindGroup`, and
allow direct custom implementations of `AsBindGroup::as_bind_group` for
the latter type of bind groups. To opt in to the new behavior, return
the `AsBindGroupError::CreateBindGroupDirectly` error from your
`AsBindGroup::unprepared_bind_group` implementation, and Bevy will call
your custom `AsBindGroup::as_bind_group` method as before.
## Migration Guide
* Bevy will now unconditionally call
`AsBindGroup::unprepared_bind_group` for your materials, so you must no
longer panic in that function. Instead, return the new
`AsBindGroupError::CreateBindGroupDirectly` error, and Bevy will fall
back to calling `AsBindGroup::as_bind_group` as before.
This commit moves the front end of the rendering pipeline to a retained
model when GPU preprocessing is in use (i.e. by default, except in
constrained environments). `RenderMeshInstance` and `MeshUniformData`
are stored from frame to frame and are updated only for the entities
that changed state. This was rather tricky and requires some careful
surgery to keep the data valid in the case of removals.
This patch is built on top of Bevy's change detection. Generally, this
worked, except that `ViewVisibility` isn't currently properly tracked.
Therefore, this commit adds proper change tracking for `ViewVisibility`.
Doing this required adding a new system that runs after all
`check_visibility` invocations, as no single `check_visibility`
invocation has enough global information to detect changes.
On the Bistro exterior scene, with all textures forced to opaque, this
patch improves steady-state `extract_meshes_for_gpu_building` from
93.8us to 34.5us and steady-state `collect_meshes_for_gpu_building` from
195.7us to 4.28us. Altogether this constitutes an improvement from 290us
to 38us, which is a 7.46x speedup.
This patch is only lightly tested and shouldn't land before 0.15 is
released anyway, so I'm releasing it as a draft.
This commit allows the Bevy renderer to use the clustering
infrastructure for light probes (reflection probes and irradiance
volumes) on platforms where at least 3 storage buffers are available. On
such platforms (the vast majority), we stop performing brute-force
searches of light probes for each fragment and instead only search the
light probes with bounding spheres that intersect the current cluster.
This should dramatically improve scalability of irradiance volumes and
reflection probes.
The primary platform that doesn't support 3 storage buffers is WebGL 2,
and we continue using a brute-force search of light probes on that
platform, as the UBO that stores per-cluster indices is too small to fit
the light probe counts. Note, however, that that platform also doesn't
support bindless textures (indeed, it would be very odd for a platform
to support bindless textures but not SSBOs), so we only support one of
each type of light probe per drawcall there in the first place.
Consequently, this isn't a performance problem, as the search will only
have one light probe to consider. (In fact, clustering would probably
end up being a performance loss.)
Known potential improvements include:
1. We currently cull based on a conservative bounding sphere test and
not based on the oriented bounding box (OBB) of the light probe. This is
improvable, but in the interests of simplicity, I opted to keep the
bounding sphere test for now. The OBB improvement can be a follow-up.
2. This patch doesn't change the fact that each fragment only takes a
single light probe into account. Typical light probe implementations
detect the case in which multiple light probes cover the current
fragment and perform some sort of weighted blend between them. As the
light probe fetch function presently returns only a single light probe,
implementing that feature would require more code restructuring, so I
left it out for now. It can be added as a follow-up.
3. Light probe implementations typically have a falloff range. Although
this is a wanted feature in Bevy, this particular commit also doesn't
implement that feature, as it's out of scope.
4. This commit doesn't raise the maximum number of light probes past its
current value of 8 for each type. This should be addressed later, but
would possibly require more bindings on platforms with storage buffers,
which would increase this patch's complexity. Even without raising the
limit, this patch should constitute a significant performance
improvement for scenes that get anywhere close to this limit. In the
interest of keeping this patch small, I opted to leave raising the limit
to a follow-up.
## Changelog
### Changed
* Light probes (reflection probes and irradiance volumes) are now
clustered on most platforms, improving performance when many light
probes are present.
---------
Co-authored-by: Benjamin Brienen <Benjamin.Brienen@outlook.com>
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
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
PCSS still has some fundamental issues (#16155). We should resolve them
before "releasing" the feature.
## Solution
1. Rename the already-optional `pbr_pcss` cargo feature to
`experimental_pbr_pcss` to better communicate its state to developers.
2. Adjust the description of the `experimental_pbr_pcss` cargo feature
to better communicate its state to developers.
3. Gate PCSS-related light component fields behind that cargo feature,
to prevent surfacing them to developers by default.
# Objective
_If I understand it correctly_, we were checking mesh visibility, as
well as re-rendering point and spot light shadow maps for each view.
This makes it so that M views and N lights produce M x N complexity.
This PR aims to fix that, as well as introduce a stress test for this
specific scenario.
## Solution
- Keep track of what lights have already had mesh visibility calculated
and do not calculate it again;
- Reuse shadow depth textures and attachments across all views, and only
render shadow maps for the _first_ time a light is encountered on a
view;
- Directional lights remain unaltered, since their shadow map cascades
are view-dependent;
- Add a new `many_cameras_lights` stress test example to verify the
solution
## Showcase
110% speed up on the stress test
83% reduction of memory usage in stress test
### Before (5.35 FPS on stress test)
<img width="1392" alt="Screenshot 2024-09-11 at 12 25 57"
src="https://github.com/user-attachments/assets/136b0785-e9a4-44df-9a22-f99cc465e126">
### After (11.34 FPS on stress test)
<img width="1392" alt="Screenshot 2024-09-11 at 12 24 35"
src="https://github.com/user-attachments/assets/b8dd858f-5e19-467f-8344-2b46ca039630">
## Testing
- Did you test these changes? If so, how?
- On my game project where I have two cameras, and many shadow casting
lights I managed to get pretty much double the FPS.
- Also included a stress test, see the comparison above
- Are there any parts that need more testing?
- Yes, I would like help verifying that this fix is indeed correct, and
that we were really re-rendering the shadow maps by mistake and it's
indeed okay to not do that
- How can other people (reviewers) test your changes? Is there anything
specific they need to know?
- Run the `many_cameras_lights` example
- On the `main` branch, cherry pick the commit with the example (`git
cherry-pick --no-commit 1ed4ace01`) and run it
- If relevant, what platforms did you test these changes on, and are
there any important ones you can't test?
- macOS
---------
Co-authored-by: François Mockers <francois.mockers@vleue.com>
# Objective
Fixes#15940
## Solution
Remove the `pub use` and fix the compile errors.
Make `bevy_image` available as `bevy::image`.
## Testing
Feature Frenzy would be good here! Maybe I'll learn how to use it if I
have some time this weekend, or maybe a reviewer can use it.
## Migration Guide
Use `bevy_image` instead of `bevy_render::texture` items.
---------
Co-authored-by: chompaa <antony.m.3012@gmail.com>
Co-authored-by: Carter Anderson <mcanders1@gmail.com>
