# Objective
- Make use of storage buffers, where they are available, for clustered forward bindings to support far more point lights in a scene
- Fixes#3605
- Based on top of #4079
This branch on an M1 Max can keep 60fps with about 2150 point lights of radius 1m in the Sponza scene where I've been testing. The bottleneck is mostly assigning lights to clusters which grows faster than linearly (I think 1000 lights was about 1.5ms and 5000 was 7.5ms). I have seen papers and presentations leveraging compute shaders that can get this up to over 1 million. That said, I think any further optimisations should probably be done in a separate PR.
## Solution
- Add `RenderDevice` to the `Material` and `SpecializedMaterial` trait `::key()` functions to allow setting flags on the keys depending on feature/limit availability
- Make `GpuPointLights` and `ViewClusterBuffers` into enums containing `UniformVec` and `StorageBuffer` variants. Implement the necessary API on them to make usage the same for both cases, and the only difference is at initialisation time.
- Appropriate shader defs in the shader code to handle the two cases
## Context on some decisions / open questions
- I'm using `max_storage_buffers_per_shader_stage >= 3` as a check to see if storage buffers are supported. I was thinking about diving into 'binding resource management' but it feels like we don't have enough use cases to understand the problem yet, and it is mostly a separate concern to this PR, so I think it should be handled separately.
- Should `ViewClusterBuffers` and `ViewClusterBindings` be merged, duplicating the count variables into the enum variants?
Co-authored-by: Carter Anderson <mcanders1@gmail.com>
# Objective
- Support compressed textures including 'universal' formats (ETC1S, UASTC) and transcoding of them to
- Support `.dds`, `.ktx2`, and `.basis` files
## Solution
- Fixes https://github.com/bevyengine/bevy/issues/3608 Look there for more details.
- Note that the functionality is all enabled through non-default features. If it is desirable to enable some by default, I can do that.
- The `basis-universal` crate, used for `.basis` file support and for transcoding, is built on bindings against a C++ library. It's not feasible to rewrite in Rust in a short amount of time. There are no Rust alternatives of which I am aware and it's specialised code. In its current state it doesn't support the wasm target, but I don't know for sure. However, it is possible to build the upstream C++ library with emscripten, so there is perhaps a way to add support for web too with some shenanigans.
- There's no support for transcoding from BasisLZ/ETC1S in KTX2 files as it was quite non-trivial to implement and didn't feel important given people could use `.basis` files for ETC1S.
This makes it possible for materials to configure front or
back face culling, or disable culling.
Initially I looked at specializing the Mesh which currently
controls this state but conceptually it seems more appropriate
to control this at the material level, not the mesh level.
_Just for reference this also seems to be consistent with Unity
where materials/shaders can configure the culling mode between
front/back/off - as opposed to configuring any culling state
when importing a mesh._
After some archaeology, trying to understand how this might
relate to the existing 'double_sided' option, it was determined
that double_sided is a more high level lighting option originally
from Filament that will cause the normals for back faces to be
flipped.
For sake of avoiding complexity, but keeping control this
currently keeps the options orthogonal, and adds some clarifying
documentation for `double_sided`. This won't affect any existing
apps since there hasn't been a way to disable backface culling
up until now, so the option was essentially redundant.
double_sided support could potentially be updated to imply
disabling of backface culling.
For reference https://github.com/bevyengine/bevy/pull/3734/commits also looks at exposing cull mode control. I think the main difference here is that this patch handles RenderPipelineDescriptor specialization directly within the StandardMaterial implementation instead of communicating info back to the Mesh via the `queue_material_meshes` system.
With the way material.rs builds up the final RenderPipelineDescriptor first by calling specialize for the MeshPipeline followed by specialize for the material then it seems like we have a natural place to override anything in the descriptor that's first configured for the mesh state.
This PR makes a number of changes to how meshes and vertex attributes are handled, which the goal of enabling easy and flexible custom vertex attributes:
* Reworks the `Mesh` type to use the newly added `VertexAttribute` internally
* `VertexAttribute` defines the name, a unique `VertexAttributeId`, and a `VertexFormat`
* `VertexAttributeId` is used to produce consistent sort orders for vertex buffer generation, replacing the more expensive and often surprising "name based sorting"
* Meshes can be used to generate a `MeshVertexBufferLayout`, which defines the layout of the gpu buffer produced by the mesh. `MeshVertexBufferLayouts` can then be used to generate actual `VertexBufferLayouts` according to the requirements of a specific pipeline. This decoupling of "mesh layout" vs "pipeline vertex buffer layout" is what enables custom attributes. We don't need to standardize _mesh layouts_ or contort meshes to meet the needs of a specific pipeline. As long as the mesh has what the pipeline needs, it will work transparently.
* Mesh-based pipelines now specialize on `&MeshVertexBufferLayout` via the new `SpecializedMeshPipeline` trait (which behaves like `SpecializedPipeline`, but adds `&MeshVertexBufferLayout`). The integrity of the pipeline cache is maintained because the `MeshVertexBufferLayout` is treated as part of the key (which is fully abstracted from implementers of the trait ... no need to add any additional info to the specialization key).
* Hashing `MeshVertexBufferLayout` is too expensive to do for every entity, every frame. To make this scalable, I added a generalized "pre-hashing" solution to `bevy_utils`: `Hashed<T>` keys and `PreHashMap<K, V>` (which uses `Hashed<T>` internally) . Why didn't I just do the quick and dirty in-place "pre-compute hash and use that u64 as a key in a hashmap" that we've done in the past? Because its wrong! Hashes by themselves aren't enough because two different values can produce the same hash. Re-hashing a hash is even worse! I decided to build a generalized solution because this pattern has come up in the past and we've chosen to do the wrong thing. Now we can do the right thing! This did unfortunately require pulling in `hashbrown` and using that in `bevy_utils`, because avoiding re-hashes requires the `raw_entry_mut` api, which isn't stabilized yet (and may never be ... `entry_ref` has favor now, but also isn't available yet). If std's HashMap ever provides the tools we need, we can move back to that. Note that adding `hashbrown` doesn't increase our dependency count because it was already in our tree. I will probably break these changes out into their own PR.
* Specializing on `MeshVertexBufferLayout` has one non-obvious behavior: it can produce identical pipelines for two different MeshVertexBufferLayouts. To optimize the number of active pipelines / reduce re-binds while drawing, I de-duplicate pipelines post-specialization using the final `VertexBufferLayout` as the key. For example, consider a pipeline that needs the layout `(position, normal)` and is specialized using two meshes: `(position, normal, uv)` and `(position, normal, other_vec2)`. If both of these meshes result in `(position, normal)` specializations, we can use the same pipeline! Now we do. Cool!
To briefly illustrate, this is what the relevant section of `MeshPipeline`'s specialization code looks like now:
```rust
impl SpecializedMeshPipeline for MeshPipeline {
type Key = MeshPipelineKey;
fn specialize(
&self,
key: Self::Key,
layout: &MeshVertexBufferLayout,
) -> RenderPipelineDescriptor {
let mut vertex_attributes = vec![
Mesh::ATTRIBUTE_POSITION.at_shader_location(0),
Mesh::ATTRIBUTE_NORMAL.at_shader_location(1),
Mesh::ATTRIBUTE_UV_0.at_shader_location(2),
];
let mut shader_defs = Vec::new();
if layout.contains(Mesh::ATTRIBUTE_TANGENT) {
shader_defs.push(String::from("VERTEX_TANGENTS"));
vertex_attributes.push(Mesh::ATTRIBUTE_TANGENT.at_shader_location(3));
}
let vertex_buffer_layout = layout
.get_layout(&vertex_attributes)
.expect("Mesh is missing a vertex attribute");
```
Notice that this is _much_ simpler than it was before. And now any mesh with any layout can be used with this pipeline, provided it has vertex postions, normals, and uvs. We even got to remove `HAS_TANGENTS` from MeshPipelineKey and `has_tangents` from `GpuMesh`, because that information is redundant with `MeshVertexBufferLayout`.
This is still a draft because I still need to:
* Add more docs
* Experiment with adding error handling to mesh pipeline specialization (which would print errors at runtime when a mesh is missing a vertex attribute required by a pipeline). If it doesn't tank perf, we'll keep it.
* Consider breaking out the PreHash / hashbrown changes into a separate PR.
* Add an example illustrating this change
* Verify that the "mesh-specialized pipeline de-duplication code" works properly
Please dont yell at me for not doing these things yet :) Just trying to get this in peoples' hands asap.
Alternative to #3120Fixes#3030
Co-authored-by: Carter Anderson <mcanders1@gmail.com>
What is says on the tin.
This has got more to do with making `clippy` slightly more *quiet* than it does with changing anything that might greatly impact readability or performance.
that said, deriving `Default` for a couple of structs is a nice easy win
This adds "high level" `Material` and `SpecializedMaterial` traits, which can be used with a `MaterialPlugin<T: SpecializedMaterial>`. `MaterialPlugin` automatically registers the appropriate resources, draw functions, and queue systems. The `Material` trait is simpler, and should cover most use cases. `SpecializedMaterial` is like `Material`, but it also requires defining a "specialization key" (see #3031). `Material` has a trivial blanket impl of `SpecializedMaterial`, which allows us to use the same types + functions for both.
This makes defining custom 3d materials much simpler (see the `shader_material` example diff) and ensures consistent behavior across all 3d materials (both built in and custom). I ported the built in `StandardMaterial` to `MaterialPlugin`. There is also a new `MaterialMeshBundle<T: SpecializedMaterial>`, which `PbrBundle` aliases to.
