bevy/crates/bevy_pbr/src/material.rs

593 lines
21 KiB
Rust
Raw Normal View History

use crate::{
AlphaMode, DrawMesh, MeshPipeline, MeshPipelineKey, MeshUniform, SetMeshBindGroup,
SetMeshViewBindGroup,
};
use bevy_app::{App, Plugin};
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
use bevy_asset::{AddAsset, AssetEvent, AssetServer, Assets, Handle};
separate tonemapping and upscaling passes (#3425) Attempt to make features like bloom https://github.com/bevyengine/bevy/pull/2876 easier to implement. **This PR:** - Moves the tonemapping from `pbr.wgsl` into a separate pass - also add a separate upscaling pass after the tonemapping which writes to the swap chain (enables resolution-independant rendering and post-processing after tonemapping) - adds a `hdr` bool to the camera which controls whether the pbr and sprite shaders render into a `Rgba16Float` texture **Open questions:** - ~should the 2d graph work the same as the 3d one?~ it is the same now - ~The current solution is a bit inflexible because while you can add a post processing pass that writes to e.g. the `hdr_texture`, you can't write to a separate `user_postprocess_texture` while reading the `hdr_texture` and tell the tone mapping pass to read from the `user_postprocess_texture` instead. If the tonemapping and upscaling render graph nodes were to take in a `TextureView` instead of the view entity this would almost work, but the bind groups for their respective input textures are already created in the `Queue` render stage in the hardcoded order.~ solved by creating bind groups in render node **New render graph:** ![render_graph](https://user-images.githubusercontent.com/22177966/147767249-57dd4229-cfab-4ec5-9bf3-dc76dccf8e8b.png) <details> <summary>Before</summary> ![render_graph_old](https://user-images.githubusercontent.com/22177966/147284579-c895fdbd-4028-41cf-914c-e1ffef60e44e.png) </details> Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-10-26 20:13:59 +00:00
use bevy_core_pipeline::{
core_3d::{AlphaMask3d, Opaque3d, Transparent3d},
tonemapping::Tonemapping,
};
Make `Resource` trait opt-in, requiring `#[derive(Resource)]` V2 (#5577) *This PR description is an edited copy of #5007, written by @alice-i-cecile.* # Objective Follow-up to https://github.com/bevyengine/bevy/pull/2254. The `Resource` trait currently has a blanket implementation for all types that meet its bounds. While ergonomic, this results in several drawbacks: * it is possible to make confusing, silent mistakes such as inserting a function pointer (Foo) rather than a value (Foo::Bar) as a resource * it is challenging to discover if a type is intended to be used as a resource * we cannot later add customization options (see the [RFC](https://github.com/bevyengine/rfcs/blob/main/rfcs/27-derive-component.md) for the equivalent choice for Component). * dependencies can use the same Rust type as a resource in invisibly conflicting ways * raw Rust types used as resources cannot preserve privacy appropriately, as anyone able to access that type can read and write to internal values * we cannot capture a definitive list of possible resources to display to users in an editor ## Notes to reviewers * Review this commit-by-commit; there's effectively no back-tracking and there's a lot of churn in some of these commits. *ira: My commits are not as well organized :')* * I've relaxed the bound on Local to Send + Sync + 'static: I don't think these concerns apply there, so this can keep things simple. Storing e.g. a u32 in a Local is fine, because there's a variable name attached explaining what it does. * I think this is a bad place for the Resource trait to live, but I've left it in place to make reviewing easier. IMO that's best tackled with https://github.com/bevyengine/bevy/issues/4981. ## Changelog `Resource` is no longer automatically implemented for all matching types. Instead, use the new `#[derive(Resource)]` macro. ## Migration Guide Add `#[derive(Resource)]` to all types you are using as a resource. If you are using a third party type as a resource, wrap it in a tuple struct to bypass orphan rules. Consider deriving `Deref` and `DerefMut` to improve ergonomics. `ClearColor` no longer implements `Component`. Using `ClearColor` as a component in 0.8 did nothing. Use the `ClearColorConfig` in the `Camera3d` and `Camera2d` components instead. Co-authored-by: Alice <alice.i.cecile@gmail.com> Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com> Co-authored-by: devil-ira <justthecooldude@gmail.com> Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-08-08 21:36:35 +00:00
use bevy_derive::{Deref, DerefMut};
use bevy_ecs::{
entity::Entity,
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
event::EventReader,
prelude::World,
Exclusive Systems Now Implement `System`. Flexible Exclusive System Params (#6083) # Objective The [Stageless RFC](https://github.com/bevyengine/rfcs/pull/45) involves allowing exclusive systems to be referenced and ordered relative to parallel systems. We've agreed that unifying systems under `System` is the right move. This is an alternative to #4166 (see rationale in the comments I left there). Note that this builds on the learnings established there (and borrows some patterns). ## Solution This unifies parallel and exclusive systems under the shared `System` trait, removing the old `ExclusiveSystem` trait / impls. This is accomplished by adding a new `ExclusiveFunctionSystem` impl similar to `FunctionSystem`. It is backed by `ExclusiveSystemParam`, which is similar to `SystemParam`. There is a new flattened out SystemContainer api (which cuts out a lot of trait and type complexity). This means you can remove all cases of `exclusive_system()`: ```rust // before commands.add_system(some_system.exclusive_system()); // after commands.add_system(some_system); ``` I've also implemented `ExclusiveSystemParam` for `&mut QueryState` and `&mut SystemState`, which makes this possible in exclusive systems: ```rust fn some_exclusive_system( world: &mut World, transforms: &mut QueryState<&Transform>, state: &mut SystemState<(Res<Time>, Query<&Player>)>, ) { for transform in transforms.iter(world) { println!("{transform:?}"); } let (time, players) = state.get(world); for player in players.iter() { println!("{player:?}"); } } ``` Note that "exclusive function systems" assume `&mut World` is present (and the first param). I think this is a fair assumption, given that the presence of `&mut World` is what defines the need for an exclusive system. I added some targeted SystemParam `static` constraints, which removed the need for this: ``` rust fn some_exclusive_system(state: &mut SystemState<(Res<'static, Time>, Query<&'static Player>)>) {} ``` ## Related - #2923 - #3001 - #3946 ## Changelog - `ExclusiveSystem` trait (and implementations) has been removed in favor of sharing the `System` trait. - `ExclusiveFunctionSystem` and `ExclusiveSystemParam` were added, enabling flexible exclusive function systems - `&mut SystemState` and `&mut QueryState` now implement `ExclusiveSystemParam` - Exclusive and parallel System configuration is now done via a unified `SystemDescriptor`, `IntoSystemDescriptor`, and `SystemContainer` api. ## Migration Guide Calling `.exclusive_system()` is no longer required (or supported) for converting exclusive system functions to exclusive systems: ```rust // Old (0.8) app.add_system(some_exclusive_system.exclusive_system()); // New (0.9) app.add_system(some_exclusive_system); ``` Converting "normal" parallel systems to exclusive systems is done by calling the exclusive ordering apis: ```rust // Old (0.8) app.add_system(some_system.exclusive_system().at_end()); // New (0.9) app.add_system(some_system.at_end()); ``` Query state in exclusive systems can now be cached via ExclusiveSystemParams, which should be preferred for clarity and performance reasons: ```rust // Old (0.8) fn some_system(world: &mut World) { let mut transforms = world.query::<&Transform>(); for transform in transforms.iter(world) { } } // New (0.9) fn some_system(world: &mut World, transforms: &mut QueryState<&Transform>) { for transform in transforms.iter(world) { } } ```
2022-09-26 23:57:07 +00:00
schedule::IntoSystemDescriptor,
system::{
lifetimeless::{Read, SQuery, SRes},
Make `Resource` trait opt-in, requiring `#[derive(Resource)]` V2 (#5577) *This PR description is an edited copy of #5007, written by @alice-i-cecile.* # Objective Follow-up to https://github.com/bevyengine/bevy/pull/2254. The `Resource` trait currently has a blanket implementation for all types that meet its bounds. While ergonomic, this results in several drawbacks: * it is possible to make confusing, silent mistakes such as inserting a function pointer (Foo) rather than a value (Foo::Bar) as a resource * it is challenging to discover if a type is intended to be used as a resource * we cannot later add customization options (see the [RFC](https://github.com/bevyengine/rfcs/blob/main/rfcs/27-derive-component.md) for the equivalent choice for Component). * dependencies can use the same Rust type as a resource in invisibly conflicting ways * raw Rust types used as resources cannot preserve privacy appropriately, as anyone able to access that type can read and write to internal values * we cannot capture a definitive list of possible resources to display to users in an editor ## Notes to reviewers * Review this commit-by-commit; there's effectively no back-tracking and there's a lot of churn in some of these commits. *ira: My commits are not as well organized :')* * I've relaxed the bound on Local to Send + Sync + 'static: I don't think these concerns apply there, so this can keep things simple. Storing e.g. a u32 in a Local is fine, because there's a variable name attached explaining what it does. * I think this is a bad place for the Resource trait to live, but I've left it in place to make reviewing easier. IMO that's best tackled with https://github.com/bevyengine/bevy/issues/4981. ## Changelog `Resource` is no longer automatically implemented for all matching types. Instead, use the new `#[derive(Resource)]` macro. ## Migration Guide Add `#[derive(Resource)]` to all types you are using as a resource. If you are using a third party type as a resource, wrap it in a tuple struct to bypass orphan rules. Consider deriving `Deref` and `DerefMut` to improve ergonomics. `ClearColor` no longer implements `Component`. Using `ClearColor` as a component in 0.8 did nothing. Use the `ClearColorConfig` in the `Camera3d` and `Camera2d` components instead. Co-authored-by: Alice <alice.i.cecile@gmail.com> Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com> Co-authored-by: devil-ira <justthecooldude@gmail.com> Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-08-08 21:36:35 +00:00
Commands, Local, Query, Res, ResMut, Resource, SystemParamItem,
},
world::FromWorld,
};
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
use bevy_reflect::TypeUuid;
use bevy_render::{
extract_component::ExtractComponentPlugin,
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
mesh::{Mesh, MeshVertexBufferLayout},
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
prelude::Image,
render_asset::{PrepareAssetLabel, RenderAssets},
render_phase::{
AddRenderCommand, DrawFunctions, EntityRenderCommand, RenderCommandResult, RenderPhase,
SetItemPipeline, TrackedRenderPass,
},
render_resource::{
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
AsBindGroup, AsBindGroupError, BindGroup, BindGroupLayout, OwnedBindingResource,
PipelineCache, RenderPipelineDescriptor, Shader, ShaderRef, SpecializedMeshPipeline,
SpecializedMeshPipelineError, SpecializedMeshPipelines,
},
renderer::RenderDevice,
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
texture::FallbackImage,
view::{ExtractedView, Msaa, VisibleEntities},
Make `RenderStage::Extract` run on the render world (#4402) # Objective - Currently, the `Extract` `RenderStage` is executed on the main world, with the render world available as a resource. - However, when needing access to resources in the render world (e.g. to mutate them), the only way to do so was to get exclusive access to the whole `RenderWorld` resource. - This meant that effectively only one extract which wrote to resources could run at a time. - We didn't previously make `Extract`ing writing to the world a non-happy path, even though we want to discourage that. ## Solution - Move the extract stage to run on the render world. - Add the main world as a `MainWorld` resource. - Add an `Extract` `SystemParam` as a convenience to access a (read only) `SystemParam` in the main world during `Extract`. ## Future work It should be possible to avoid needing to use `get_or_spawn` for the render commands, since now the `Commands`' `Entities` matches up with the world being executed on. We need to determine how this interacts with https://github.com/bevyengine/bevy/pull/3519 It's theoretically possible to remove the need for the `value` method on `Extract`. However, that requires slightly changing the `SystemParam` interface, which would make it more complicated. That would probably mess up the `SystemState` api too. ## Todo I still need to add doc comments to `Extract`. --- ## Changelog ### Changed - The `Extract` `RenderStage` now runs on the render world (instead of the main world as before). You must use the `Extract` `SystemParam` to access the main world during the extract phase. Resources on the render world can now be accessed using `ResMut` during extract. ### Removed - `Commands::spawn_and_forget`. Use `Commands::get_or_spawn(e).insert_bundle(bundle)` instead ## Migration Guide The `Extract` `RenderStage` now runs on the render world (instead of the main world as before). You must use the `Extract` `SystemParam` to access the main world during the extract phase. `Extract` takes a single type parameter, which is any system parameter (such as `Res`, `Query` etc.). It will extract this from the main world, and returns the result of this extraction when `value` is called on it. For example, if previously your extract system looked like: ```rust fn extract_clouds(mut commands: Commands, clouds: Query<Entity, With<Cloud>>) { for cloud in clouds.iter() { commands.get_or_spawn(cloud).insert(Cloud); } } ``` the new version would be: ```rust fn extract_clouds(mut commands: Commands, mut clouds: Extract<Query<Entity, With<Cloud>>>) { for cloud in clouds.value().iter() { commands.get_or_spawn(cloud).insert(Cloud); } } ``` The diff is: ```diff --- a/src/clouds.rs +++ b/src/clouds.rs @@ -1,5 +1,5 @@ -fn extract_clouds(mut commands: Commands, clouds: Query<Entity, With<Cloud>>) { - for cloud in clouds.iter() { +fn extract_clouds(mut commands: Commands, mut clouds: Extract<Query<Entity, With<Cloud>>>) { + for cloud in clouds.value().iter() { commands.get_or_spawn(cloud).insert(Cloud); } } ``` You can now also access resources from the render world using the normal system parameters during `Extract`: ```rust fn extract_assets(mut render_assets: ResMut<MyAssets>, source_assets: Extract<Res<MyAssets>>) { *render_assets = source_assets.clone(); } ``` Please note that all existing extract systems need to be updated to match this new style; even if they currently compile they will not run as expected. A warning will be emitted on a best-effort basis if this is not met. Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-07-08 23:56:33 +00:00
Extract, RenderApp, RenderStage,
};
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
use bevy_utils::{tracing::error, HashMap, HashSet};
use std::hash::Hash;
use std::marker::PhantomData;
2020-02-11 17:31:49 +00:00
/// Materials are used alongside [`MaterialPlugin`] and [`MaterialMeshBundle`](crate::MaterialMeshBundle)
/// to spawn entities that are rendered with a specific [`Material`] type. They serve as an easy to use high level
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
/// way to render [`Mesh`] entities with custom shader logic.
///
/// Materials must implement [`AsBindGroup`] to define how data will be transferred to the GPU and bound in shaders.
/// [`AsBindGroup`] can be derived, which makes generating bindings straightforward. See the [`AsBindGroup`] docs for details.
///
/// Materials must also implement [`TypeUuid`] so they can be treated as an [`Asset`](bevy_asset::Asset).
///
/// # Example
///
/// Here is a simple Material implementation. The [`AsBindGroup`] derive has many features. To see what else is available,
/// check out the [`AsBindGroup`] documentation.
/// ```
/// # use bevy_pbr::{Material, MaterialMeshBundle};
/// # use bevy_ecs::prelude::*;
/// # use bevy_reflect::TypeUuid;
/// # use bevy_render::{render_resource::{AsBindGroup, ShaderRef}, texture::Image, color::Color};
/// # use bevy_asset::{Handle, AssetServer, Assets};
///
/// #[derive(AsBindGroup, TypeUuid, Debug, Clone)]
/// #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"]
/// pub struct CustomMaterial {
/// // Uniform bindings must implement `ShaderType`, which will be used to convert the value to
/// // its shader-compatible equivalent. Most core math types already implement `ShaderType`.
/// #[uniform(0)]
/// color: Color,
/// // Images can be bound as textures in shaders. If the Image's sampler is also needed, just
/// // add the sampler attribute with a different binding index.
/// #[texture(1)]
/// #[sampler(2)]
/// color_texture: Handle<Image>,
/// }
///
/// // All functions on `Material` have default impls. You only need to implement the
/// // functions that are relevant for your material.
/// impl Material for CustomMaterial {
/// fn fragment_shader() -> ShaderRef {
/// "shaders/custom_material.wgsl".into()
/// }
/// }
///
/// // Spawn an entity using `CustomMaterial`.
/// fn setup(mut commands: Commands, mut materials: ResMut<Assets<CustomMaterial>>, asset_server: Res<AssetServer>) {
Spawn now takes a Bundle (#6054) # Objective Now that we can consolidate Bundles and Components under a single insert (thanks to #2975 and #6039), almost 100% of world spawns now look like `world.spawn().insert((Some, Tuple, Here))`. Spawning an entity without any components is an extremely uncommon pattern, so it makes sense to give spawn the "first class" ergonomic api. This consolidated api should be made consistent across all spawn apis (such as World and Commands). ## Solution All `spawn` apis (`World::spawn`, `Commands:;spawn`, `ChildBuilder::spawn`, and `WorldChildBuilder::spawn`) now accept a bundle as input: ```rust // before: commands .spawn() .insert((A, B, C)); world .spawn() .insert((A, B, C); // after commands.spawn((A, B, C)); world.spawn((A, B, C)); ``` All existing instances of `spawn_bundle` have been deprecated in favor of the new `spawn` api. A new `spawn_empty` has been added, replacing the old `spawn` api. By allowing `world.spawn(some_bundle)` to replace `world.spawn().insert(some_bundle)`, this opened the door to removing the initial entity allocation in the "empty" archetype / table done in `spawn()` (and subsequent move to the actual archetype in `.insert(some_bundle)`). This improves spawn performance by over 10%: ![image](https://user-images.githubusercontent.com/2694663/191627587-4ab2f949-4ccd-4231-80eb-80dd4d9ad6b9.png) To take this measurement, I added a new `world_spawn` benchmark. Unfortunately, optimizing `Commands::spawn` is slightly less trivial, as Commands expose the Entity id of spawned entities prior to actually spawning. Doing the optimization would (naively) require assurances that the `spawn(some_bundle)` command is applied before all other commands involving the entity (which would not necessarily be true, if memory serves). Optimizing `Commands::spawn` this way does feel possible, but it will require careful thought (and maybe some additional checks), which deserves its own PR. For now, it has the same performance characteristics of the current `Commands::spawn_bundle` on main. **Note that 99% of this PR is simple renames and refactors. The only code that needs careful scrutiny is the new `World::spawn()` impl, which is relatively straightforward, but it has some new unsafe code (which re-uses battle tested BundlerSpawner code path).** --- ## Changelog - All `spawn` apis (`World::spawn`, `Commands:;spawn`, `ChildBuilder::spawn`, and `WorldChildBuilder::spawn`) now accept a bundle as input - All instances of `spawn_bundle` have been deprecated in favor of the new `spawn` api - World and Commands now have `spawn_empty()`, which is equivalent to the old `spawn()` behavior. ## Migration Guide ```rust // Old (0.8): commands .spawn() .insert_bundle((A, B, C)); // New (0.9) commands.spawn((A, B, C)); // Old (0.8): commands.spawn_bundle((A, B, C)); // New (0.9) commands.spawn((A, B, C)); // Old (0.8): let entity = commands.spawn().id(); // New (0.9) let entity = commands.spawn_empty().id(); // Old (0.8) let entity = world.spawn().id(); // New (0.9) let entity = world.spawn_empty(); ```
2022-09-23 19:55:54 +00:00
/// commands.spawn(MaterialMeshBundle {
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
/// material: materials.add(CustomMaterial {
/// color: Color::RED,
/// color_texture: asset_server.load("some_image.png"),
/// }),
/// ..Default::default()
/// });
/// }
/// ```
/// In WGSL shaders, the material's binding would look like this:
///
/// ```wgsl
/// @group(1) @binding(0)
/// var<uniform> color: vec4<f32>;
/// @group(1) @binding(1)
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
/// var color_texture: texture_2d<f32>;
/// @group(1) @binding(2)
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
/// var color_sampler: sampler;
/// ```
pub trait Material: AsBindGroup + Send + Sync + Clone + TypeUuid + Sized + 'static {
/// Returns this material's vertex shader. If [`ShaderRef::Default`] is returned, the default mesh vertex shader
/// will be used.
fn vertex_shader() -> ShaderRef {
ShaderRef::Default
2020-02-11 17:31:49 +00:00
}
2020-08-01 00:10:29 +00:00
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
/// Returns this material's fragment shader. If [`ShaderRef::Default`] is returned, the default mesh fragment shader
/// will be used.
#[allow(unused_variables)]
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
fn fragment_shader() -> ShaderRef {
ShaderRef::Default
}
/// Returns this material's [`AlphaMode`]. Defaults to [`AlphaMode::Opaque`].
#[inline]
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
fn alpha_mode(&self) -> AlphaMode {
AlphaMode::Opaque
2020-08-01 00:10:29 +00:00
}
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
#[inline]
/// Add a bias to the view depth of the mesh which can be used to force a specific render order
/// for meshes with equal depth, to avoid z-fighting.
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
fn depth_bias(&self) -> f32 {
0.0
}
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
/// Customizes the default [`RenderPipelineDescriptor`] for a specific entity using the entity's
/// [`MaterialPipelineKey`] and [`MeshVertexBufferLayout`] as input.
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
#[allow(unused_variables)]
#[inline]
fn specialize(
Use storage buffers for clustered forward point lights (#3989) # 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>
2022-04-07 16:16:35 +00:00
pipeline: &MaterialPipeline<Self>,
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
descriptor: &mut RenderPipelineDescriptor,
layout: &MeshVertexBufferLayout,
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
key: MaterialPipelineKey<Self>,
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
) -> Result<(), SpecializedMeshPipelineError> {
Ok(())
}
2020-08-01 00:10:29 +00:00
}
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
/// Adds the necessary ECS resources and render logic to enable rendering entities using the given [`Material`]
/// asset type.
pub struct MaterialPlugin<M: Material>(PhantomData<M>);
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
impl<M: Material> Default for MaterialPlugin<M> {
fn default() -> Self {
Self(Default::default())
}
}
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
impl<M: Material> Plugin for MaterialPlugin<M>
where
M::Data: PartialEq + Eq + Hash + Clone,
{
fn build(&self, app: &mut App) {
app.add_asset::<M>()
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
.add_plugin(ExtractComponentPlugin::<Handle<M>>::extract_visible());
if let Ok(render_app) = app.get_sub_app_mut(RenderApp) {
render_app
.add_render_command::<Transparent3d, DrawMaterial<M>>()
.add_render_command::<Opaque3d, DrawMaterial<M>>()
.add_render_command::<AlphaMask3d, DrawMaterial<M>>()
.init_resource::<MaterialPipeline<M>>()
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
.init_resource::<ExtractedMaterials<M>>()
.init_resource::<RenderMaterials<M>>()
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
.init_resource::<SpecializedMeshPipelines<MaterialPipeline<M>>>()
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
.add_system_to_stage(RenderStage::Extract, extract_materials::<M>)
.add_system_to_stage(
RenderStage::Prepare,
prepare_materials::<M>.after(PrepareAssetLabel::PreAssetPrepare),
)
.add_system_to_stage(RenderStage::Queue, queue_material_meshes::<M>);
}
}
}
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
/// A key uniquely identifying a specialized [`MaterialPipeline`].
pub struct MaterialPipelineKey<M: Material> {
pub mesh_key: MeshPipelineKey,
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
pub bind_group_data: M::Data,
}
impl<M: Material> Eq for MaterialPipelineKey<M> where M::Data: PartialEq {}
impl<M: Material> PartialEq for MaterialPipelineKey<M>
where
M::Data: PartialEq,
{
fn eq(&self, other: &Self) -> bool {
self.mesh_key == other.mesh_key && self.bind_group_data == other.bind_group_data
}
}
impl<M: Material> Clone for MaterialPipelineKey<M>
where
M::Data: Clone,
{
fn clone(&self) -> Self {
Self {
mesh_key: self.mesh_key,
bind_group_data: self.bind_group_data.clone(),
}
}
}
impl<M: Material> Hash for MaterialPipelineKey<M>
where
M::Data: Hash,
{
fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
self.mesh_key.hash(state);
self.bind_group_data.hash(state);
}
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
}
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
/// Render pipeline data for a given [`Material`].
Make `Resource` trait opt-in, requiring `#[derive(Resource)]` V2 (#5577) *This PR description is an edited copy of #5007, written by @alice-i-cecile.* # Objective Follow-up to https://github.com/bevyengine/bevy/pull/2254. The `Resource` trait currently has a blanket implementation for all types that meet its bounds. While ergonomic, this results in several drawbacks: * it is possible to make confusing, silent mistakes such as inserting a function pointer (Foo) rather than a value (Foo::Bar) as a resource * it is challenging to discover if a type is intended to be used as a resource * we cannot later add customization options (see the [RFC](https://github.com/bevyengine/rfcs/blob/main/rfcs/27-derive-component.md) for the equivalent choice for Component). * dependencies can use the same Rust type as a resource in invisibly conflicting ways * raw Rust types used as resources cannot preserve privacy appropriately, as anyone able to access that type can read and write to internal values * we cannot capture a definitive list of possible resources to display to users in an editor ## Notes to reviewers * Review this commit-by-commit; there's effectively no back-tracking and there's a lot of churn in some of these commits. *ira: My commits are not as well organized :')* * I've relaxed the bound on Local to Send + Sync + 'static: I don't think these concerns apply there, so this can keep things simple. Storing e.g. a u32 in a Local is fine, because there's a variable name attached explaining what it does. * I think this is a bad place for the Resource trait to live, but I've left it in place to make reviewing easier. IMO that's best tackled with https://github.com/bevyengine/bevy/issues/4981. ## Changelog `Resource` is no longer automatically implemented for all matching types. Instead, use the new `#[derive(Resource)]` macro. ## Migration Guide Add `#[derive(Resource)]` to all types you are using as a resource. If you are using a third party type as a resource, wrap it in a tuple struct to bypass orphan rules. Consider deriving `Deref` and `DerefMut` to improve ergonomics. `ClearColor` no longer implements `Component`. Using `ClearColor` as a component in 0.8 did nothing. Use the `ClearColorConfig` in the `Camera3d` and `Camera2d` components instead. Co-authored-by: Alice <alice.i.cecile@gmail.com> Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com> Co-authored-by: devil-ira <justthecooldude@gmail.com> Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-08-08 21:36:35 +00:00
#[derive(Resource)]
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
pub struct MaterialPipeline<M: Material> {
pub mesh_pipeline: MeshPipeline,
pub material_layout: BindGroupLayout,
pub vertex_shader: Option<Handle<Shader>>,
pub fragment_shader: Option<Handle<Shader>>,
marker: PhantomData<M>,
}
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
impl<M: Material> SpecializedMeshPipeline for MaterialPipeline<M>
where
M::Data: PartialEq + Eq + Hash + Clone,
{
type Key = MaterialPipelineKey<M>;
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
fn specialize(
&self,
key: Self::Key,
layout: &MeshVertexBufferLayout,
) -> Result<RenderPipelineDescriptor, SpecializedMeshPipelineError> {
let mut descriptor = self.mesh_pipeline.specialize(key.mesh_key, layout)?;
if let Some(vertex_shader) = &self.vertex_shader {
descriptor.vertex.shader = vertex_shader.clone();
}
if let Some(fragment_shader) = &self.fragment_shader {
descriptor.fragment.as_mut().unwrap().shader = fragment_shader.clone();
}
// MeshPipeline::specialize's current implementation guarantees that the returned
// specialized descriptor has a populated layout
let descriptor_layout = descriptor.layout.as_mut().unwrap();
descriptor_layout.insert(1, self.material_layout.clone());
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
M::specialize(self, &mut descriptor, layout, key)?;
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
Ok(descriptor)
}
}
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
impl<M: Material> FromWorld for MaterialPipeline<M> {
fn from_world(world: &mut World) -> Self {
let asset_server = world.resource::<AssetServer>();
let render_device = world.resource::<RenderDevice>();
MaterialPipeline {
mesh_pipeline: world.resource::<MeshPipeline>().clone(),
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
material_layout: M::bind_group_layout(render_device),
vertex_shader: match M::vertex_shader() {
ShaderRef::Default => None,
ShaderRef::Handle(handle) => Some(handle),
ShaderRef::Path(path) => Some(asset_server.load(path)),
},
fragment_shader: match M::fragment_shader() {
ShaderRef::Default => None,
ShaderRef::Handle(handle) => Some(handle),
ShaderRef::Path(path) => Some(asset_server.load(path)),
},
marker: PhantomData,
}
}
}
type DrawMaterial<M> = (
SetItemPipeline,
SetMeshViewBindGroup<0>,
SetMaterialBindGroup<M, 1>,
SetMeshBindGroup<2>,
DrawMesh,
);
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
/// Sets the bind group for a given [`Material`] at the configured `I` index.
pub struct SetMaterialBindGroup<M: Material, const I: usize>(PhantomData<M>);
impl<M: Material, const I: usize> EntityRenderCommand for SetMaterialBindGroup<M, I> {
type Param = (SRes<RenderMaterials<M>>, SQuery<Read<Handle<M>>>);
fn render<'w>(
_view: Entity,
item: Entity,
(materials, query): SystemParamItem<'w, '_, Self::Param>,
pass: &mut TrackedRenderPass<'w>,
) -> RenderCommandResult {
let material_handle = query.get(item).unwrap();
let material = materials.into_inner().get(material_handle).unwrap();
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
pass.set_bind_group(I, &material.bind_group, &[]);
RenderCommandResult::Success
}
}
#[allow(clippy::too_many_arguments)]
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
pub fn queue_material_meshes<M: Material>(
opaque_draw_functions: Res<DrawFunctions<Opaque3d>>,
alpha_mask_draw_functions: Res<DrawFunctions<AlphaMask3d>>,
transparent_draw_functions: Res<DrawFunctions<Transparent3d>>,
material_pipeline: Res<MaterialPipeline<M>>,
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
mut pipelines: ResMut<SpecializedMeshPipelines<MaterialPipeline<M>>>,
mut pipeline_cache: ResMut<PipelineCache>,
msaa: Res<Msaa>,
render_meshes: Res<RenderAssets<Mesh>>,
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
render_materials: Res<RenderMaterials<M>>,
material_meshes: Query<(&Handle<M>, &Handle<Mesh>, &MeshUniform)>,
mut views: Query<(
&ExtractedView,
&VisibleEntities,
separate tonemapping and upscaling passes (#3425) Attempt to make features like bloom https://github.com/bevyengine/bevy/pull/2876 easier to implement. **This PR:** - Moves the tonemapping from `pbr.wgsl` into a separate pass - also add a separate upscaling pass after the tonemapping which writes to the swap chain (enables resolution-independant rendering and post-processing after tonemapping) - adds a `hdr` bool to the camera which controls whether the pbr and sprite shaders render into a `Rgba16Float` texture **Open questions:** - ~should the 2d graph work the same as the 3d one?~ it is the same now - ~The current solution is a bit inflexible because while you can add a post processing pass that writes to e.g. the `hdr_texture`, you can't write to a separate `user_postprocess_texture` while reading the `hdr_texture` and tell the tone mapping pass to read from the `user_postprocess_texture` instead. If the tonemapping and upscaling render graph nodes were to take in a `TextureView` instead of the view entity this would almost work, but the bind groups for their respective input textures are already created in the `Queue` render stage in the hardcoded order.~ solved by creating bind groups in render node **New render graph:** ![render_graph](https://user-images.githubusercontent.com/22177966/147767249-57dd4229-cfab-4ec5-9bf3-dc76dccf8e8b.png) <details> <summary>Before</summary> ![render_graph_old](https://user-images.githubusercontent.com/22177966/147284579-c895fdbd-4028-41cf-914c-e1ffef60e44e.png) </details> Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-10-26 20:13:59 +00:00
Option<&Tonemapping>,
&mut RenderPhase<Opaque3d>,
&mut RenderPhase<AlphaMask3d>,
&mut RenderPhase<Transparent3d>,
)>,
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
) where
M::Data: PartialEq + Eq + Hash + Clone,
{
separate tonemapping and upscaling passes (#3425) Attempt to make features like bloom https://github.com/bevyengine/bevy/pull/2876 easier to implement. **This PR:** - Moves the tonemapping from `pbr.wgsl` into a separate pass - also add a separate upscaling pass after the tonemapping which writes to the swap chain (enables resolution-independant rendering and post-processing after tonemapping) - adds a `hdr` bool to the camera which controls whether the pbr and sprite shaders render into a `Rgba16Float` texture **Open questions:** - ~should the 2d graph work the same as the 3d one?~ it is the same now - ~The current solution is a bit inflexible because while you can add a post processing pass that writes to e.g. the `hdr_texture`, you can't write to a separate `user_postprocess_texture` while reading the `hdr_texture` and tell the tone mapping pass to read from the `user_postprocess_texture` instead. If the tonemapping and upscaling render graph nodes were to take in a `TextureView` instead of the view entity this would almost work, but the bind groups for their respective input textures are already created in the `Queue` render stage in the hardcoded order.~ solved by creating bind groups in render node **New render graph:** ![render_graph](https://user-images.githubusercontent.com/22177966/147767249-57dd4229-cfab-4ec5-9bf3-dc76dccf8e8b.png) <details> <summary>Before</summary> ![render_graph_old](https://user-images.githubusercontent.com/22177966/147284579-c895fdbd-4028-41cf-914c-e1ffef60e44e.png) </details> Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-10-26 20:13:59 +00:00
for (
view,
visible_entities,
tonemapping,
mut opaque_phase,
mut alpha_mask_phase,
mut transparent_phase,
) in &mut views
{
let draw_opaque_pbr = opaque_draw_functions.read().id::<DrawMaterial<M>>();
let draw_alpha_mask_pbr = alpha_mask_draw_functions.read().id::<DrawMaterial<M>>();
let draw_transparent_pbr = transparent_draw_functions.read().id::<DrawMaterial<M>>();
separate tonemapping and upscaling passes (#3425) Attempt to make features like bloom https://github.com/bevyengine/bevy/pull/2876 easier to implement. **This PR:** - Moves the tonemapping from `pbr.wgsl` into a separate pass - also add a separate upscaling pass after the tonemapping which writes to the swap chain (enables resolution-independant rendering and post-processing after tonemapping) - adds a `hdr` bool to the camera which controls whether the pbr and sprite shaders render into a `Rgba16Float` texture **Open questions:** - ~should the 2d graph work the same as the 3d one?~ it is the same now - ~The current solution is a bit inflexible because while you can add a post processing pass that writes to e.g. the `hdr_texture`, you can't write to a separate `user_postprocess_texture` while reading the `hdr_texture` and tell the tone mapping pass to read from the `user_postprocess_texture` instead. If the tonemapping and upscaling render graph nodes were to take in a `TextureView` instead of the view entity this would almost work, but the bind groups for their respective input textures are already created in the `Queue` render stage in the hardcoded order.~ solved by creating bind groups in render node **New render graph:** ![render_graph](https://user-images.githubusercontent.com/22177966/147767249-57dd4229-cfab-4ec5-9bf3-dc76dccf8e8b.png) <details> <summary>Before</summary> ![render_graph_old](https://user-images.githubusercontent.com/22177966/147284579-c895fdbd-4028-41cf-914c-e1ffef60e44e.png) </details> Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-10-26 20:13:59 +00:00
let mut view_key =
MeshPipelineKey::from_msaa_samples(msaa.samples) | MeshPipelineKey::from_hdr(view.hdr);
Fix color banding by dithering image before quantization (#5264) # Objective - Closes #5262 - Fix color banding caused by quantization. ## Solution - Adds dithering to the tonemapping node from #3425. - This is inspired by Godot's default "debanding" shader: https://gist.github.com/belzecue/ - Unlike Godot: - debanding happens after tonemapping. My understanding is that this is preferred, because we are running the debanding at the last moment before quantization (`[f32, f32, f32, f32]` -> `f32`). This ensures we aren't biasing the dithering strength by applying it in a different (linear) color space. - This code instead uses and reference the origin source, Valve at GDC 2015 ![Screenshot from 2022-11-10 13-44-46](https://user-images.githubusercontent.com/2632925/201218880-70f4cdab-a1ed-44de-a88c-8759e77197f1.png) ![Screenshot from 2022-11-10 13-41-11](https://user-images.githubusercontent.com/2632925/201218883-72393352-b162-41da-88bb-6e54a1e26853.png) ## Additional Notes Real time rendering to standard dynamic range outputs is limited to 8 bits of depth per color channel. Internally we keep everything in full 32-bit precision (`vec4<f32>`) inside passes and 16-bit between passes until the image is ready to be displayed, at which point the GPU implicitly converts our `vec4<f32>` into a single 32bit value per pixel, with each channel (rgba) getting 8 of those 32 bits. ### The Problem 8 bits of color depth is simply not enough precision to make each step invisible - we only have 256 values per channel! Human vision can perceive steps in luma to about 14 bits of precision. When drawing a very slight gradient, the transition between steps become visible because with a gradient, neighboring pixels will all jump to the next "step" of precision at the same time. ### The Solution One solution is to simply output in HDR - more bits of color data means the transition between bands will become smaller. However, not everyone has hardware that supports 10+ bit color depth. Additionally, 10 bit color doesn't even fully solve the issue, banding will result in coherent bands on shallow gradients, but the steps will be harder to perceive. The solution in this PR adds noise to the signal before it is "quantized" or resampled from 32 to 8 bits. Done naively, it's easy to add unneeded noise to the image. To ensure dithering is correct and absolutely minimal, noise is adding *within* one step of the output color depth. When converting from the 32bit to 8bit signal, the value is rounded to the nearest 8 bit value (0 - 255). Banding occurs around the transition from one value to the next, let's say from 50-51. Dithering will never add more than +/-0.5 bits of noise, so the pixels near this transition might round to 50 instead of 51 but will never round more than one step. This means that the output image won't have excess variance: - in a gradient from 49 to 51, there will be a step between each band at 49, 50, and 51. - Done correctly, the modified image of this gradient will never have a adjacent pixels more than one step (0-255) from each other. - I.e. when scanning across the gradient you should expect to see: ``` |-band-| |-band-| |-band-| Baseline: 49 49 49 50 50 50 51 51 51 Dithered: 49 50 49 50 50 51 50 51 51 Dithered (wrong): 49 50 51 49 50 51 49 51 50 ``` ![Screenshot from 2022-11-10 14-12-36](https://user-images.githubusercontent.com/2632925/201219075-ab3f46be-d4e9-4869-b66b-a92e1706f49e.png) ![Screenshot from 2022-11-10 14-11-48](https://user-images.githubusercontent.com/2632925/201219079-ec5d2add-817d-487a-8fc1-84569c9cda73.png) You can see from above how correct dithering "fuzzes" the transition between bands to reduce distinct steps in color, without adding excess noise. ### HDR The previous section (and this PR) assumes the final output is to an 8-bit texture, however this is not always the case. When Bevy adds HDR support, the dithering code will need to take the per-channel depth into account instead of assuming it to be 0-255. Edit: I talked with Rob about this and it seems like the current solution is okay. We may need to revisit once we have actual HDR final image output. --- ## Changelog ### Added - All pipelines now support deband dithering. This is enabled by default in 3D, and can be toggled in the `Tonemapping` component in camera bundles. Banding is a graphical artifact created when the rendered image is crunched from high precision (f32 per color channel) down to the final output (u8 per channel in SDR). This results in subtle gradients becoming blocky due to the reduced color precision. Deband dithering applies a small amount of noise to the signal before it is "crunched", which breaks up the hard edges of blocks (bands) of color. Note that this does not add excess noise to the image, as the amount of noise is less than a single step of a color channel - just enough to break up the transition between color blocks in a gradient. Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-11-11 19:43:45 +00:00
if let Some(Tonemapping::Enabled { deband_dither }) = tonemapping {
if !view.hdr {
separate tonemapping and upscaling passes (#3425) Attempt to make features like bloom https://github.com/bevyengine/bevy/pull/2876 easier to implement. **This PR:** - Moves the tonemapping from `pbr.wgsl` into a separate pass - also add a separate upscaling pass after the tonemapping which writes to the swap chain (enables resolution-independant rendering and post-processing after tonemapping) - adds a `hdr` bool to the camera which controls whether the pbr and sprite shaders render into a `Rgba16Float` texture **Open questions:** - ~should the 2d graph work the same as the 3d one?~ it is the same now - ~The current solution is a bit inflexible because while you can add a post processing pass that writes to e.g. the `hdr_texture`, you can't write to a separate `user_postprocess_texture` while reading the `hdr_texture` and tell the tone mapping pass to read from the `user_postprocess_texture` instead. If the tonemapping and upscaling render graph nodes were to take in a `TextureView` instead of the view entity this would almost work, but the bind groups for their respective input textures are already created in the `Queue` render stage in the hardcoded order.~ solved by creating bind groups in render node **New render graph:** ![render_graph](https://user-images.githubusercontent.com/22177966/147767249-57dd4229-cfab-4ec5-9bf3-dc76dccf8e8b.png) <details> <summary>Before</summary> ![render_graph_old](https://user-images.githubusercontent.com/22177966/147284579-c895fdbd-4028-41cf-914c-e1ffef60e44e.png) </details> Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-10-26 20:13:59 +00:00
view_key |= MeshPipelineKey::TONEMAP_IN_SHADER;
Fix color banding by dithering image before quantization (#5264) # Objective - Closes #5262 - Fix color banding caused by quantization. ## Solution - Adds dithering to the tonemapping node from #3425. - This is inspired by Godot's default "debanding" shader: https://gist.github.com/belzecue/ - Unlike Godot: - debanding happens after tonemapping. My understanding is that this is preferred, because we are running the debanding at the last moment before quantization (`[f32, f32, f32, f32]` -> `f32`). This ensures we aren't biasing the dithering strength by applying it in a different (linear) color space. - This code instead uses and reference the origin source, Valve at GDC 2015 ![Screenshot from 2022-11-10 13-44-46](https://user-images.githubusercontent.com/2632925/201218880-70f4cdab-a1ed-44de-a88c-8759e77197f1.png) ![Screenshot from 2022-11-10 13-41-11](https://user-images.githubusercontent.com/2632925/201218883-72393352-b162-41da-88bb-6e54a1e26853.png) ## Additional Notes Real time rendering to standard dynamic range outputs is limited to 8 bits of depth per color channel. Internally we keep everything in full 32-bit precision (`vec4<f32>`) inside passes and 16-bit between passes until the image is ready to be displayed, at which point the GPU implicitly converts our `vec4<f32>` into a single 32bit value per pixel, with each channel (rgba) getting 8 of those 32 bits. ### The Problem 8 bits of color depth is simply not enough precision to make each step invisible - we only have 256 values per channel! Human vision can perceive steps in luma to about 14 bits of precision. When drawing a very slight gradient, the transition between steps become visible because with a gradient, neighboring pixels will all jump to the next "step" of precision at the same time. ### The Solution One solution is to simply output in HDR - more bits of color data means the transition between bands will become smaller. However, not everyone has hardware that supports 10+ bit color depth. Additionally, 10 bit color doesn't even fully solve the issue, banding will result in coherent bands on shallow gradients, but the steps will be harder to perceive. The solution in this PR adds noise to the signal before it is "quantized" or resampled from 32 to 8 bits. Done naively, it's easy to add unneeded noise to the image. To ensure dithering is correct and absolutely minimal, noise is adding *within* one step of the output color depth. When converting from the 32bit to 8bit signal, the value is rounded to the nearest 8 bit value (0 - 255). Banding occurs around the transition from one value to the next, let's say from 50-51. Dithering will never add more than +/-0.5 bits of noise, so the pixels near this transition might round to 50 instead of 51 but will never round more than one step. This means that the output image won't have excess variance: - in a gradient from 49 to 51, there will be a step between each band at 49, 50, and 51. - Done correctly, the modified image of this gradient will never have a adjacent pixels more than one step (0-255) from each other. - I.e. when scanning across the gradient you should expect to see: ``` |-band-| |-band-| |-band-| Baseline: 49 49 49 50 50 50 51 51 51 Dithered: 49 50 49 50 50 51 50 51 51 Dithered (wrong): 49 50 51 49 50 51 49 51 50 ``` ![Screenshot from 2022-11-10 14-12-36](https://user-images.githubusercontent.com/2632925/201219075-ab3f46be-d4e9-4869-b66b-a92e1706f49e.png) ![Screenshot from 2022-11-10 14-11-48](https://user-images.githubusercontent.com/2632925/201219079-ec5d2add-817d-487a-8fc1-84569c9cda73.png) You can see from above how correct dithering "fuzzes" the transition between bands to reduce distinct steps in color, without adding excess noise. ### HDR The previous section (and this PR) assumes the final output is to an 8-bit texture, however this is not always the case. When Bevy adds HDR support, the dithering code will need to take the per-channel depth into account instead of assuming it to be 0-255. Edit: I talked with Rob about this and it seems like the current solution is okay. We may need to revisit once we have actual HDR final image output. --- ## Changelog ### Added - All pipelines now support deband dithering. This is enabled by default in 3D, and can be toggled in the `Tonemapping` component in camera bundles. Banding is a graphical artifact created when the rendered image is crunched from high precision (f32 per color channel) down to the final output (u8 per channel in SDR). This results in subtle gradients becoming blocky due to the reduced color precision. Deband dithering applies a small amount of noise to the signal before it is "crunched", which breaks up the hard edges of blocks (bands) of color. Note that this does not add excess noise to the image, as the amount of noise is less than a single step of a color channel - just enough to break up the transition between color blocks in a gradient. Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-11-11 19:43:45 +00:00
if *deband_dither {
view_key |= MeshPipelineKey::DEBAND_DITHER;
}
separate tonemapping and upscaling passes (#3425) Attempt to make features like bloom https://github.com/bevyengine/bevy/pull/2876 easier to implement. **This PR:** - Moves the tonemapping from `pbr.wgsl` into a separate pass - also add a separate upscaling pass after the tonemapping which writes to the swap chain (enables resolution-independant rendering and post-processing after tonemapping) - adds a `hdr` bool to the camera which controls whether the pbr and sprite shaders render into a `Rgba16Float` texture **Open questions:** - ~should the 2d graph work the same as the 3d one?~ it is the same now - ~The current solution is a bit inflexible because while you can add a post processing pass that writes to e.g. the `hdr_texture`, you can't write to a separate `user_postprocess_texture` while reading the `hdr_texture` and tell the tone mapping pass to read from the `user_postprocess_texture` instead. If the tonemapping and upscaling render graph nodes were to take in a `TextureView` instead of the view entity this would almost work, but the bind groups for their respective input textures are already created in the `Queue` render stage in the hardcoded order.~ solved by creating bind groups in render node **New render graph:** ![render_graph](https://user-images.githubusercontent.com/22177966/147767249-57dd4229-cfab-4ec5-9bf3-dc76dccf8e8b.png) <details> <summary>Before</summary> ![render_graph_old](https://user-images.githubusercontent.com/22177966/147284579-c895fdbd-4028-41cf-914c-e1ffef60e44e.png) </details> Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-10-26 20:13:59 +00:00
}
}
Add ViewRangefinder3d to reduce boilerplate when enqueuing standard 3D PhaseItems. (#5014) # Objective Reduce the boilerplate code needed to make draw order sorting work correctly when queuing items through new common functionality. Also fix several instances in the bevy code-base (mostly examples) where this boilerplate appears to be incorrect. ## Solution - Moved the logic for handling back-to-front vs front-to-back draw ordering into the PhaseItems by inverting the sort key ordering of Opaque3d and AlphaMask3d. The means that all the standard 3d rendering phases measure distance in the same way. Clients of these structs no longer need to know to negate the distance. - Added a new utility struct, ViewRangefinder3d, which encapsulates the maths needed to calculate a "distance" from an ExtractedView and a mesh's transform matrix. - Converted all the occurrences of the distance calculations in Bevy and its examples to use ViewRangefinder3d. Several of these occurrences appear to be buggy because they don't invert the view matrix or don't negate the distance where appropriate. This leads me to the view that Bevy should expose a facility to correctly perform this calculation. ## Migration Guide Code which creates Opaque3d, AlphaMask3d, or Transparent3d phase items _should_ use ViewRangefinder3d to calculate the distance value. Code which manually calculated the distance for Opaque3d or AlphaMask3d phase items and correctly negated the z value will no longer depth sort correctly. However, incorrect depth sorting for these types will not impact the rendered output as sorting is only a performance optimisation when drawing with depth-testing enabled. Code which manually calculated the distance for Transparent3d phase items will continue to work as before.
2022-07-05 06:13:39 +00:00
let rangefinder = view.rangefinder3d();
for visible_entity in &visible_entities.entities {
if let Ok((material_handle, mesh_handle, mesh_uniform)) =
material_meshes.get(*visible_entity)
{
if let Some(material) = render_materials.get(material_handle) {
if let Some(mesh) = render_meshes.get(mesh_handle) {
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
let mut mesh_key =
MeshPipelineKey::from_primitive_topology(mesh.primitive_topology)
separate tonemapping and upscaling passes (#3425) Attempt to make features like bloom https://github.com/bevyengine/bevy/pull/2876 easier to implement. **This PR:** - Moves the tonemapping from `pbr.wgsl` into a separate pass - also add a separate upscaling pass after the tonemapping which writes to the swap chain (enables resolution-independant rendering and post-processing after tonemapping) - adds a `hdr` bool to the camera which controls whether the pbr and sprite shaders render into a `Rgba16Float` texture **Open questions:** - ~should the 2d graph work the same as the 3d one?~ it is the same now - ~The current solution is a bit inflexible because while you can add a post processing pass that writes to e.g. the `hdr_texture`, you can't write to a separate `user_postprocess_texture` while reading the `hdr_texture` and tell the tone mapping pass to read from the `user_postprocess_texture` instead. If the tonemapping and upscaling render graph nodes were to take in a `TextureView` instead of the view entity this would almost work, but the bind groups for their respective input textures are already created in the `Queue` render stage in the hardcoded order.~ solved by creating bind groups in render node **New render graph:** ![render_graph](https://user-images.githubusercontent.com/22177966/147767249-57dd4229-cfab-4ec5-9bf3-dc76dccf8e8b.png) <details> <summary>Before</summary> ![render_graph_old](https://user-images.githubusercontent.com/22177966/147284579-c895fdbd-4028-41cf-914c-e1ffef60e44e.png) </details> Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-10-26 20:13:59 +00:00
| view_key;
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
let alpha_mode = material.properties.alpha_mode;
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
if let AlphaMode::Blend = alpha_mode {
mesh_key |= MeshPipelineKey::TRANSPARENT_MAIN_PASS;
}
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
let pipeline_id = pipelines.specialize(
&mut pipeline_cache,
&material_pipeline,
MaterialPipelineKey {
mesh_key,
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
bind_group_data: material.key.clone(),
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
},
&mesh.layout,
);
let pipeline_id = match pipeline_id {
Ok(id) => id,
Err(err) => {
error!("{}", err);
continue;
}
};
Add ViewRangefinder3d to reduce boilerplate when enqueuing standard 3D PhaseItems. (#5014) # Objective Reduce the boilerplate code needed to make draw order sorting work correctly when queuing items through new common functionality. Also fix several instances in the bevy code-base (mostly examples) where this boilerplate appears to be incorrect. ## Solution - Moved the logic for handling back-to-front vs front-to-back draw ordering into the PhaseItems by inverting the sort key ordering of Opaque3d and AlphaMask3d. The means that all the standard 3d rendering phases measure distance in the same way. Clients of these structs no longer need to know to negate the distance. - Added a new utility struct, ViewRangefinder3d, which encapsulates the maths needed to calculate a "distance" from an ExtractedView and a mesh's transform matrix. - Converted all the occurrences of the distance calculations in Bevy and its examples to use ViewRangefinder3d. Several of these occurrences appear to be buggy because they don't invert the view matrix or don't negate the distance where appropriate. This leads me to the view that Bevy should expose a facility to correctly perform this calculation. ## Migration Guide Code which creates Opaque3d, AlphaMask3d, or Transparent3d phase items _should_ use ViewRangefinder3d to calculate the distance value. Code which manually calculated the distance for Opaque3d or AlphaMask3d phase items and correctly negated the z value will no longer depth sort correctly. However, incorrect depth sorting for these types will not impact the rendered output as sorting is only a performance optimisation when drawing with depth-testing enabled. Code which manually calculated the distance for Transparent3d phase items will continue to work as before.
2022-07-05 06:13:39 +00:00
let distance = rangefinder.distance(&mesh_uniform.transform)
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
+ material.properties.depth_bias;
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
match alpha_mode {
AlphaMode::Opaque => {
opaque_phase.add(Opaque3d {
entity: *visible_entity,
draw_function: draw_opaque_pbr,
pipeline: pipeline_id,
Add ViewRangefinder3d to reduce boilerplate when enqueuing standard 3D PhaseItems. (#5014) # Objective Reduce the boilerplate code needed to make draw order sorting work correctly when queuing items through new common functionality. Also fix several instances in the bevy code-base (mostly examples) where this boilerplate appears to be incorrect. ## Solution - Moved the logic for handling back-to-front vs front-to-back draw ordering into the PhaseItems by inverting the sort key ordering of Opaque3d and AlphaMask3d. The means that all the standard 3d rendering phases measure distance in the same way. Clients of these structs no longer need to know to negate the distance. - Added a new utility struct, ViewRangefinder3d, which encapsulates the maths needed to calculate a "distance" from an ExtractedView and a mesh's transform matrix. - Converted all the occurrences of the distance calculations in Bevy and its examples to use ViewRangefinder3d. Several of these occurrences appear to be buggy because they don't invert the view matrix or don't negate the distance where appropriate. This leads me to the view that Bevy should expose a facility to correctly perform this calculation. ## Migration Guide Code which creates Opaque3d, AlphaMask3d, or Transparent3d phase items _should_ use ViewRangefinder3d to calculate the distance value. Code which manually calculated the distance for Opaque3d or AlphaMask3d phase items and correctly negated the z value will no longer depth sort correctly. However, incorrect depth sorting for these types will not impact the rendered output as sorting is only a performance optimisation when drawing with depth-testing enabled. Code which manually calculated the distance for Transparent3d phase items will continue to work as before.
2022-07-05 06:13:39 +00:00
distance,
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
});
}
AlphaMode::Mask(_) => {
alpha_mask_phase.add(AlphaMask3d {
entity: *visible_entity,
draw_function: draw_alpha_mask_pbr,
pipeline: pipeline_id,
Add ViewRangefinder3d to reduce boilerplate when enqueuing standard 3D PhaseItems. (#5014) # Objective Reduce the boilerplate code needed to make draw order sorting work correctly when queuing items through new common functionality. Also fix several instances in the bevy code-base (mostly examples) where this boilerplate appears to be incorrect. ## Solution - Moved the logic for handling back-to-front vs front-to-back draw ordering into the PhaseItems by inverting the sort key ordering of Opaque3d and AlphaMask3d. The means that all the standard 3d rendering phases measure distance in the same way. Clients of these structs no longer need to know to negate the distance. - Added a new utility struct, ViewRangefinder3d, which encapsulates the maths needed to calculate a "distance" from an ExtractedView and a mesh's transform matrix. - Converted all the occurrences of the distance calculations in Bevy and its examples to use ViewRangefinder3d. Several of these occurrences appear to be buggy because they don't invert the view matrix or don't negate the distance where appropriate. This leads me to the view that Bevy should expose a facility to correctly perform this calculation. ## Migration Guide Code which creates Opaque3d, AlphaMask3d, or Transparent3d phase items _should_ use ViewRangefinder3d to calculate the distance value. Code which manually calculated the distance for Opaque3d or AlphaMask3d phase items and correctly negated the z value will no longer depth sort correctly. However, incorrect depth sorting for these types will not impact the rendered output as sorting is only a performance optimisation when drawing with depth-testing enabled. Code which manually calculated the distance for Transparent3d phase items will continue to work as before.
2022-07-05 06:13:39 +00:00
distance,
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
});
}
AlphaMode::Blend => {
transparent_phase.add(Transparent3d {
entity: *visible_entity,
draw_function: draw_transparent_pbr,
pipeline: pipeline_id,
Add ViewRangefinder3d to reduce boilerplate when enqueuing standard 3D PhaseItems. (#5014) # Objective Reduce the boilerplate code needed to make draw order sorting work correctly when queuing items through new common functionality. Also fix several instances in the bevy code-base (mostly examples) where this boilerplate appears to be incorrect. ## Solution - Moved the logic for handling back-to-front vs front-to-back draw ordering into the PhaseItems by inverting the sort key ordering of Opaque3d and AlphaMask3d. The means that all the standard 3d rendering phases measure distance in the same way. Clients of these structs no longer need to know to negate the distance. - Added a new utility struct, ViewRangefinder3d, which encapsulates the maths needed to calculate a "distance" from an ExtractedView and a mesh's transform matrix. - Converted all the occurrences of the distance calculations in Bevy and its examples to use ViewRangefinder3d. Several of these occurrences appear to be buggy because they don't invert the view matrix or don't negate the distance where appropriate. This leads me to the view that Bevy should expose a facility to correctly perform this calculation. ## Migration Guide Code which creates Opaque3d, AlphaMask3d, or Transparent3d phase items _should_ use ViewRangefinder3d to calculate the distance value. Code which manually calculated the distance for Opaque3d or AlphaMask3d phase items and correctly negated the z value will no longer depth sort correctly. However, incorrect depth sorting for these types will not impact the rendered output as sorting is only a performance optimisation when drawing with depth-testing enabled. Code which manually calculated the distance for Transparent3d phase items will continue to work as before.
2022-07-05 06:13:39 +00:00
distance,
Mesh vertex buffer layouts (#3959) 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 #3120 Fixes #3030 Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-02-23 23:21:13 +00:00
});
}
}
}
}
}
}
}
}
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
/// Common [`Material`] properties, calculated for a specific material instance.
pub struct MaterialProperties {
/// The [`AlphaMode`] of this material.
pub alpha_mode: AlphaMode,
/// Add a bias to the view depth of the mesh which can be used to force a specific render order
/// for meshes with equal depth, to avoid z-fighting.
pub depth_bias: f32,
}
/// Data prepared for a [`Material`] instance.
pub struct PreparedMaterial<T: Material> {
pub bindings: Vec<OwnedBindingResource>,
pub bind_group: BindGroup,
pub key: T::Data,
pub properties: MaterialProperties,
}
Make `Resource` trait opt-in, requiring `#[derive(Resource)]` V2 (#5577) *This PR description is an edited copy of #5007, written by @alice-i-cecile.* # Objective Follow-up to https://github.com/bevyengine/bevy/pull/2254. The `Resource` trait currently has a blanket implementation for all types that meet its bounds. While ergonomic, this results in several drawbacks: * it is possible to make confusing, silent mistakes such as inserting a function pointer (Foo) rather than a value (Foo::Bar) as a resource * it is challenging to discover if a type is intended to be used as a resource * we cannot later add customization options (see the [RFC](https://github.com/bevyengine/rfcs/blob/main/rfcs/27-derive-component.md) for the equivalent choice for Component). * dependencies can use the same Rust type as a resource in invisibly conflicting ways * raw Rust types used as resources cannot preserve privacy appropriately, as anyone able to access that type can read and write to internal values * we cannot capture a definitive list of possible resources to display to users in an editor ## Notes to reviewers * Review this commit-by-commit; there's effectively no back-tracking and there's a lot of churn in some of these commits. *ira: My commits are not as well organized :')* * I've relaxed the bound on Local to Send + Sync + 'static: I don't think these concerns apply there, so this can keep things simple. Storing e.g. a u32 in a Local is fine, because there's a variable name attached explaining what it does. * I think this is a bad place for the Resource trait to live, but I've left it in place to make reviewing easier. IMO that's best tackled with https://github.com/bevyengine/bevy/issues/4981. ## Changelog `Resource` is no longer automatically implemented for all matching types. Instead, use the new `#[derive(Resource)]` macro. ## Migration Guide Add `#[derive(Resource)]` to all types you are using as a resource. If you are using a third party type as a resource, wrap it in a tuple struct to bypass orphan rules. Consider deriving `Deref` and `DerefMut` to improve ergonomics. `ClearColor` no longer implements `Component`. Using `ClearColor` as a component in 0.8 did nothing. Use the `ClearColorConfig` in the `Camera3d` and `Camera2d` components instead. Co-authored-by: Alice <alice.i.cecile@gmail.com> Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com> Co-authored-by: devil-ira <justthecooldude@gmail.com> Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-08-08 21:36:35 +00:00
#[derive(Resource)]
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
struct ExtractedMaterials<M: Material> {
extracted: Vec<(Handle<M>, M)>,
removed: Vec<Handle<M>>,
}
impl<M: Material> Default for ExtractedMaterials<M> {
fn default() -> Self {
Self {
extracted: Default::default(),
removed: Default::default(),
}
}
}
/// Stores all prepared representations of [`Material`] assets for as long as they exist.
Make `Resource` trait opt-in, requiring `#[derive(Resource)]` V2 (#5577) *This PR description is an edited copy of #5007, written by @alice-i-cecile.* # Objective Follow-up to https://github.com/bevyengine/bevy/pull/2254. The `Resource` trait currently has a blanket implementation for all types that meet its bounds. While ergonomic, this results in several drawbacks: * it is possible to make confusing, silent mistakes such as inserting a function pointer (Foo) rather than a value (Foo::Bar) as a resource * it is challenging to discover if a type is intended to be used as a resource * we cannot later add customization options (see the [RFC](https://github.com/bevyengine/rfcs/blob/main/rfcs/27-derive-component.md) for the equivalent choice for Component). * dependencies can use the same Rust type as a resource in invisibly conflicting ways * raw Rust types used as resources cannot preserve privacy appropriately, as anyone able to access that type can read and write to internal values * we cannot capture a definitive list of possible resources to display to users in an editor ## Notes to reviewers * Review this commit-by-commit; there's effectively no back-tracking and there's a lot of churn in some of these commits. *ira: My commits are not as well organized :')* * I've relaxed the bound on Local to Send + Sync + 'static: I don't think these concerns apply there, so this can keep things simple. Storing e.g. a u32 in a Local is fine, because there's a variable name attached explaining what it does. * I think this is a bad place for the Resource trait to live, but I've left it in place to make reviewing easier. IMO that's best tackled with https://github.com/bevyengine/bevy/issues/4981. ## Changelog `Resource` is no longer automatically implemented for all matching types. Instead, use the new `#[derive(Resource)]` macro. ## Migration Guide Add `#[derive(Resource)]` to all types you are using as a resource. If you are using a third party type as a resource, wrap it in a tuple struct to bypass orphan rules. Consider deriving `Deref` and `DerefMut` to improve ergonomics. `ClearColor` no longer implements `Component`. Using `ClearColor` as a component in 0.8 did nothing. Use the `ClearColorConfig` in the `Camera3d` and `Camera2d` components instead. Co-authored-by: Alice <alice.i.cecile@gmail.com> Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com> Co-authored-by: devil-ira <justthecooldude@gmail.com> Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-08-08 21:36:35 +00:00
#[derive(Resource, Deref, DerefMut)]
pub struct RenderMaterials<T: Material>(pub HashMap<Handle<T>, PreparedMaterial<T>>);
impl<T: Material> Default for RenderMaterials<T> {
fn default() -> Self {
Self(Default::default())
}
}
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
/// This system extracts all created or modified assets of the corresponding [`Material`] type
/// into the "render world".
fn extract_materials<M: Material>(
mut commands: Commands,
Make `RenderStage::Extract` run on the render world (#4402) # Objective - Currently, the `Extract` `RenderStage` is executed on the main world, with the render world available as a resource. - However, when needing access to resources in the render world (e.g. to mutate them), the only way to do so was to get exclusive access to the whole `RenderWorld` resource. - This meant that effectively only one extract which wrote to resources could run at a time. - We didn't previously make `Extract`ing writing to the world a non-happy path, even though we want to discourage that. ## Solution - Move the extract stage to run on the render world. - Add the main world as a `MainWorld` resource. - Add an `Extract` `SystemParam` as a convenience to access a (read only) `SystemParam` in the main world during `Extract`. ## Future work It should be possible to avoid needing to use `get_or_spawn` for the render commands, since now the `Commands`' `Entities` matches up with the world being executed on. We need to determine how this interacts with https://github.com/bevyengine/bevy/pull/3519 It's theoretically possible to remove the need for the `value` method on `Extract`. However, that requires slightly changing the `SystemParam` interface, which would make it more complicated. That would probably mess up the `SystemState` api too. ## Todo I still need to add doc comments to `Extract`. --- ## Changelog ### Changed - The `Extract` `RenderStage` now runs on the render world (instead of the main world as before). You must use the `Extract` `SystemParam` to access the main world during the extract phase. Resources on the render world can now be accessed using `ResMut` during extract. ### Removed - `Commands::spawn_and_forget`. Use `Commands::get_or_spawn(e).insert_bundle(bundle)` instead ## Migration Guide The `Extract` `RenderStage` now runs on the render world (instead of the main world as before). You must use the `Extract` `SystemParam` to access the main world during the extract phase. `Extract` takes a single type parameter, which is any system parameter (such as `Res`, `Query` etc.). It will extract this from the main world, and returns the result of this extraction when `value` is called on it. For example, if previously your extract system looked like: ```rust fn extract_clouds(mut commands: Commands, clouds: Query<Entity, With<Cloud>>) { for cloud in clouds.iter() { commands.get_or_spawn(cloud).insert(Cloud); } } ``` the new version would be: ```rust fn extract_clouds(mut commands: Commands, mut clouds: Extract<Query<Entity, With<Cloud>>>) { for cloud in clouds.value().iter() { commands.get_or_spawn(cloud).insert(Cloud); } } ``` The diff is: ```diff --- a/src/clouds.rs +++ b/src/clouds.rs @@ -1,5 +1,5 @@ -fn extract_clouds(mut commands: Commands, clouds: Query<Entity, With<Cloud>>) { - for cloud in clouds.iter() { +fn extract_clouds(mut commands: Commands, mut clouds: Extract<Query<Entity, With<Cloud>>>) { + for cloud in clouds.value().iter() { commands.get_or_spawn(cloud).insert(Cloud); } } ``` You can now also access resources from the render world using the normal system parameters during `Extract`: ```rust fn extract_assets(mut render_assets: ResMut<MyAssets>, source_assets: Extract<Res<MyAssets>>) { *render_assets = source_assets.clone(); } ``` Please note that all existing extract systems need to be updated to match this new style; even if they currently compile they will not run as expected. A warning will be emitted on a best-effort basis if this is not met. Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-07-08 23:56:33 +00:00
mut events: Extract<EventReader<AssetEvent<M>>>,
assets: Extract<Res<Assets<M>>>,
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
) {
let mut changed_assets = HashSet::default();
let mut removed = Vec::new();
for event in events.iter() {
match event {
AssetEvent::Created { handle } | AssetEvent::Modified { handle } => {
Make `RenderStage::Extract` run on the render world (#4402) # Objective - Currently, the `Extract` `RenderStage` is executed on the main world, with the render world available as a resource. - However, when needing access to resources in the render world (e.g. to mutate them), the only way to do so was to get exclusive access to the whole `RenderWorld` resource. - This meant that effectively only one extract which wrote to resources could run at a time. - We didn't previously make `Extract`ing writing to the world a non-happy path, even though we want to discourage that. ## Solution - Move the extract stage to run on the render world. - Add the main world as a `MainWorld` resource. - Add an `Extract` `SystemParam` as a convenience to access a (read only) `SystemParam` in the main world during `Extract`. ## Future work It should be possible to avoid needing to use `get_or_spawn` for the render commands, since now the `Commands`' `Entities` matches up with the world being executed on. We need to determine how this interacts with https://github.com/bevyengine/bevy/pull/3519 It's theoretically possible to remove the need for the `value` method on `Extract`. However, that requires slightly changing the `SystemParam` interface, which would make it more complicated. That would probably mess up the `SystemState` api too. ## Todo I still need to add doc comments to `Extract`. --- ## Changelog ### Changed - The `Extract` `RenderStage` now runs on the render world (instead of the main world as before). You must use the `Extract` `SystemParam` to access the main world during the extract phase. Resources on the render world can now be accessed using `ResMut` during extract. ### Removed - `Commands::spawn_and_forget`. Use `Commands::get_or_spawn(e).insert_bundle(bundle)` instead ## Migration Guide The `Extract` `RenderStage` now runs on the render world (instead of the main world as before). You must use the `Extract` `SystemParam` to access the main world during the extract phase. `Extract` takes a single type parameter, which is any system parameter (such as `Res`, `Query` etc.). It will extract this from the main world, and returns the result of this extraction when `value` is called on it. For example, if previously your extract system looked like: ```rust fn extract_clouds(mut commands: Commands, clouds: Query<Entity, With<Cloud>>) { for cloud in clouds.iter() { commands.get_or_spawn(cloud).insert(Cloud); } } ``` the new version would be: ```rust fn extract_clouds(mut commands: Commands, mut clouds: Extract<Query<Entity, With<Cloud>>>) { for cloud in clouds.value().iter() { commands.get_or_spawn(cloud).insert(Cloud); } } ``` The diff is: ```diff --- a/src/clouds.rs +++ b/src/clouds.rs @@ -1,5 +1,5 @@ -fn extract_clouds(mut commands: Commands, clouds: Query<Entity, With<Cloud>>) { - for cloud in clouds.iter() { +fn extract_clouds(mut commands: Commands, mut clouds: Extract<Query<Entity, With<Cloud>>>) { + for cloud in clouds.value().iter() { commands.get_or_spawn(cloud).insert(Cloud); } } ``` You can now also access resources from the render world using the normal system parameters during `Extract`: ```rust fn extract_assets(mut render_assets: ResMut<MyAssets>, source_assets: Extract<Res<MyAssets>>) { *render_assets = source_assets.clone(); } ``` Please note that all existing extract systems need to be updated to match this new style; even if they currently compile they will not run as expected. A warning will be emitted on a best-effort basis if this is not met. Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-07-08 23:56:33 +00:00
changed_assets.insert(handle.clone_weak());
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
}
AssetEvent::Removed { handle } => {
changed_assets.remove(handle);
removed.push(handle.clone_weak());
}
}
}
let mut extracted_assets = Vec::new();
for handle in changed_assets.drain() {
Make `RenderStage::Extract` run on the render world (#4402) # Objective - Currently, the `Extract` `RenderStage` is executed on the main world, with the render world available as a resource. - However, when needing access to resources in the render world (e.g. to mutate them), the only way to do so was to get exclusive access to the whole `RenderWorld` resource. - This meant that effectively only one extract which wrote to resources could run at a time. - We didn't previously make `Extract`ing writing to the world a non-happy path, even though we want to discourage that. ## Solution - Move the extract stage to run on the render world. - Add the main world as a `MainWorld` resource. - Add an `Extract` `SystemParam` as a convenience to access a (read only) `SystemParam` in the main world during `Extract`. ## Future work It should be possible to avoid needing to use `get_or_spawn` for the render commands, since now the `Commands`' `Entities` matches up with the world being executed on. We need to determine how this interacts with https://github.com/bevyengine/bevy/pull/3519 It's theoretically possible to remove the need for the `value` method on `Extract`. However, that requires slightly changing the `SystemParam` interface, which would make it more complicated. That would probably mess up the `SystemState` api too. ## Todo I still need to add doc comments to `Extract`. --- ## Changelog ### Changed - The `Extract` `RenderStage` now runs on the render world (instead of the main world as before). You must use the `Extract` `SystemParam` to access the main world during the extract phase. Resources on the render world can now be accessed using `ResMut` during extract. ### Removed - `Commands::spawn_and_forget`. Use `Commands::get_or_spawn(e).insert_bundle(bundle)` instead ## Migration Guide The `Extract` `RenderStage` now runs on the render world (instead of the main world as before). You must use the `Extract` `SystemParam` to access the main world during the extract phase. `Extract` takes a single type parameter, which is any system parameter (such as `Res`, `Query` etc.). It will extract this from the main world, and returns the result of this extraction when `value` is called on it. For example, if previously your extract system looked like: ```rust fn extract_clouds(mut commands: Commands, clouds: Query<Entity, With<Cloud>>) { for cloud in clouds.iter() { commands.get_or_spawn(cloud).insert(Cloud); } } ``` the new version would be: ```rust fn extract_clouds(mut commands: Commands, mut clouds: Extract<Query<Entity, With<Cloud>>>) { for cloud in clouds.value().iter() { commands.get_or_spawn(cloud).insert(Cloud); } } ``` The diff is: ```diff --- a/src/clouds.rs +++ b/src/clouds.rs @@ -1,5 +1,5 @@ -fn extract_clouds(mut commands: Commands, clouds: Query<Entity, With<Cloud>>) { - for cloud in clouds.iter() { +fn extract_clouds(mut commands: Commands, mut clouds: Extract<Query<Entity, With<Cloud>>>) { + for cloud in clouds.value().iter() { commands.get_or_spawn(cloud).insert(Cloud); } } ``` You can now also access resources from the render world using the normal system parameters during `Extract`: ```rust fn extract_assets(mut render_assets: ResMut<MyAssets>, source_assets: Extract<Res<MyAssets>>) { *render_assets = source_assets.clone(); } ``` Please note that all existing extract systems need to be updated to match this new style; even if they currently compile they will not run as expected. A warning will be emitted on a best-effort basis if this is not met. Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-07-08 23:56:33 +00:00
if let Some(asset) = assets.get(&handle) {
extracted_assets.push((handle, asset.clone()));
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
}
}
commands.insert_resource(ExtractedMaterials {
extracted: extracted_assets,
removed,
});
}
/// All [`Material`] values of a given type that should be prepared next frame.
pub struct PrepareNextFrameMaterials<M: Material> {
assets: Vec<(Handle<M>, M)>,
}
impl<M: Material> Default for PrepareNextFrameMaterials<M> {
fn default() -> Self {
Self {
assets: Default::default(),
}
}
}
/// This system prepares all assets of the corresponding [`Material`] type
/// which where extracted this frame for the GPU.
fn prepare_materials<M: Material>(
mut prepare_next_frame: Local<PrepareNextFrameMaterials<M>>,
mut extracted_assets: ResMut<ExtractedMaterials<M>>,
mut render_materials: ResMut<RenderMaterials<M>>,
render_device: Res<RenderDevice>,
images: Res<RenderAssets<Image>>,
fallback_image: Res<FallbackImage>,
pipeline: Res<MaterialPipeline<M>>,
) {
let queued_assets = std::mem::take(&mut prepare_next_frame.assets);
for (handle, material) in queued_assets.into_iter() {
Better Materials: AsBindGroup trait and derive, simpler Material trait (#5053) # Objective This PR reworks Bevy's Material system, making the user experience of defining Materials _much_ nicer. Bevy's previous material system leaves a lot to be desired: * Materials require manually implementing the `RenderAsset` trait, which involves manually generating the bind group, handling gpu buffer data transfer, looking up image textures, etc. Even the simplest single-texture material involves writing ~80 unnecessary lines of code. This was never the long term plan. * There are two material traits, which is confusing, hard to document, and often redundant: `Material` and `SpecializedMaterial`. `Material` implicitly implements `SpecializedMaterial`, and `SpecializedMaterial` is used in most high level apis to support both use cases. Most users shouldn't need to think about specialization at all (I consider it a "power-user tool"), so the fact that `SpecializedMaterial` is front-and-center in our apis is a miss. * Implementing either material trait involves a lot of "type soup". The "prepared asset" parameter is particularly heinous: `&<Self as RenderAsset>::PreparedAsset`. Defining vertex and fragment shaders is also more verbose than it needs to be. ## Solution Say hello to the new `Material` system: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } } ``` Thats it! This same material would have required [~80 lines of complicated "type heavy" code](https://github.com/bevyengine/bevy/blob/v0.7.0/examples/shader/shader_material.rs) in the old Material system. Now it is just 14 lines of simple, readable code. This is thanks to a new consolidated `Material` trait and the new `AsBindGroup` trait / derive. ### The new `Material` trait The old "split" `Material` and `SpecializedMaterial` traits have been removed in favor of a new consolidated `Material` trait. All of the functions on the trait are optional. The difficulty of implementing `Material` has been reduced by simplifying dataflow and removing type complexity: ```rust // Old impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn alpha_mode(render_asset: &<Self as RenderAsset>::PreparedAsset) -> AlphaMode { render_asset.alpha_mode } } // New impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn alpha_mode(&self) -> AlphaMode { self.alpha_mode } } ``` Specialization is still supported, but it is hidden by default under the `specialize()` function (more on this later). ### The `AsBindGroup` trait / derive The `Material` trait now requires the `AsBindGroup` derive. This can be implemented manually relatively easily, but deriving it will almost always be preferable. Field attributes like `uniform` and `texture` are used to define which fields should be bindings, what their binding type is, and what index they should be bound at: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` In WGSL shaders, the binding looks like this: ```wgsl struct CoolMaterial { color: vec4<f32>; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; [[group(1), binding(1)]] var color_texture: texture_2d<f32>; [[group(1), binding(2)]] var color_sampler: sampler; ``` Note that the "group" index is determined by the usage context. It is not defined in `AsBindGroup`. Bevy material bind groups are bound to group 1. The following field-level attributes are supported: * `uniform(BINDING_INDEX)` * The field will be converted to a shader-compatible type using the `ShaderType` trait, written to a `Buffer`, and bound as a uniform. It can also be derived for custom structs. * `texture(BINDING_INDEX)` * This field's `Handle<Image>` will be used to look up the matching `Texture` gpu resource, which will be bound as a texture in shaders. The field will be assumed to implement `Into<Option<Handle<Image>>>`. In practice, most fields should be a `Handle<Image>` or `Option<Handle<Image>>`. If the value of an `Option<Handle<Image>>` is `None`, the new `FallbackImage` resource will be used instead. This attribute can be used in conjunction with a `sampler` binding attribute (with a different binding index). * `sampler(BINDING_INDEX)` * Behaves exactly like the `texture` attribute, but sets the Image's sampler binding instead of the texture. Note that fields without field-level binding attributes will be ignored. ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, this_field_is_ignored: String, } ``` As mentioned above, `Option<Handle<Image>>` is also supported: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Option<Handle<Image>>, } ``` This is useful if you want a texture to be optional. When the value is `None`, the `FallbackImage` will be used for the binding instead, which defaults to "pure white". Field uniforms with the same binding index will be combined into a single binding: ```rust #[derive(AsBindGroup)] struct CoolMaterial { #[uniform(0)] color: Color, #[uniform(0)] roughness: f32, } ``` In WGSL shaders, the binding would look like this: ```wgsl struct CoolMaterial { color: vec4<f32>; roughness: f32; }; [[group(1), binding(0)]] var<uniform> material: CoolMaterial; ``` Some less common scenarios will require "struct-level" attributes. These are the currently supported struct-level attributes: * `uniform(BINDING_INDEX, ConvertedShaderType)` * Similar to the field-level `uniform` attribute, but instead the entire `AsBindGroup` value is converted to `ConvertedShaderType`, which must implement `ShaderType`. This is useful if more complicated conversion logic is required. * `bind_group_data(DataType)` * The `AsBindGroup` type will be converted to some `DataType` using `Into<DataType>` and stored as `AsBindGroup::Data` as part of the `AsBindGroup::as_bind_group` call. This is useful if data needs to be stored alongside the generated bind group, such as a unique identifier for a material's bind group. The most common use case for this attribute is "shader pipeline specialization". The previous `CoolMaterial` example illustrating "combining multiple field-level uniform attributes with the same binding index" can also be equivalently represented with a single struct-level uniform attribute: ```rust #[derive(AsBindGroup)] #[uniform(0, CoolMaterialUniform)] struct CoolMaterial { color: Color, roughness: f32, } #[derive(ShaderType)] struct CoolMaterialUniform { color: Color, roughness: f32, } impl From<&CoolMaterial> for CoolMaterialUniform { fn from(material: &CoolMaterial) -> CoolMaterialUniform { CoolMaterialUniform { color: material.color, roughness: material.roughness, } } } ``` ### Material Specialization Material shader specialization is now _much_ simpler: ```rust #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] #[bind_group_data(CoolMaterialKey)] struct CoolMaterial { #[uniform(0)] color: Color, is_red: bool, } #[derive(Copy, Clone, Hash, Eq, PartialEq)] struct CoolMaterialKey { is_red: bool, } impl From<&CoolMaterial> for CoolMaterialKey { fn from(material: &CoolMaterial) -> CoolMaterialKey { CoolMaterialKey { is_red: material.is_red, } } } impl Material for CoolMaterial { fn fragment_shader() -> ShaderRef { "cool_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { if key.bind_group_data.is_red { let fragment = descriptor.fragment.as_mut().unwrap(); fragment.shader_defs.push("IS_RED".to_string()); } Ok(()) } } ``` Setting `bind_group_data` is not required for specialization (it defaults to `()`). Scenarios like "custom vertex attributes" also benefit from this system: ```rust impl Material for CustomMaterial { fn vertex_shader() -> ShaderRef { "custom_material.wgsl".into() } fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } fn specialize( pipeline: &MaterialPipeline<Self>, descriptor: &mut RenderPipelineDescriptor, layout: &MeshVertexBufferLayout, key: MaterialPipelineKey<Self>, ) -> Result<(), SpecializedMeshPipelineError> { let vertex_layout = layout.get_layout(&[ Mesh::ATTRIBUTE_POSITION.at_shader_location(0), ATTRIBUTE_BLEND_COLOR.at_shader_location(1), ])?; descriptor.vertex.buffers = vec![vertex_layout]; Ok(()) } } ``` ### Ported `StandardMaterial` to the new `Material` system Bevy's built-in PBR material uses the new Material system (including the AsBindGroup derive): ```rust #[derive(AsBindGroup, Debug, Clone, TypeUuid)] #[uuid = "7494888b-c082-457b-aacf-517228cc0c22"] #[bind_group_data(StandardMaterialKey)] #[uniform(0, StandardMaterialUniform)] pub struct StandardMaterial { pub base_color: Color, #[texture(1)] #[sampler(2)] pub base_color_texture: Option<Handle<Image>>, /* other fields omitted for brevity */ ``` ### Ported Bevy examples to the new `Material` system The overall complexity of Bevy's "custom shader examples" has gone down significantly. Take a look at the diffs if you want a dopamine spike. Please note that while this PR has a net increase in "lines of code", most of those extra lines come from added documentation. There is a significant reduction in the overall complexity of the code (even accounting for the new derive logic). --- ## Changelog ### Added * `AsBindGroup` trait and derive, which make it much easier to transfer data to the gpu and generate bind groups for a given type. ### Changed * The old `Material` and `SpecializedMaterial` traits have been replaced by a consolidated (much simpler) `Material` trait. Materials no longer implement `RenderAsset`. * `StandardMaterial` was ported to the new material system. There are no user-facing api changes to the `StandardMaterial` struct api, but it now implements `AsBindGroup` and `Material` instead of `RenderAsset` and `SpecializedMaterial`. ## Migration Guide The Material system has been reworked to be much simpler. We've removed a lot of boilerplate with the new `AsBindGroup` derive and the `Material` trait is simpler as well! ### Bevy 0.7 (old) ```rust #[derive(Debug, Clone, TypeUuid)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { color: Color, color_texture: Handle<Image>, } #[derive(Clone)] pub struct GpuCustomMaterial { _buffer: Buffer, bind_group: BindGroup, } impl RenderAsset for CustomMaterial { type ExtractedAsset = CustomMaterial; type PreparedAsset = GpuCustomMaterial; type Param = (SRes<RenderDevice>, SRes<MaterialPipeline<Self>>); fn extract_asset(&self) -> Self::ExtractedAsset { self.clone() } fn prepare_asset( extracted_asset: Self::ExtractedAsset, (render_device, material_pipeline): &mut SystemParamItem<Self::Param>, ) -> Result<Self::PreparedAsset, PrepareAssetError<Self::ExtractedAsset>> { let color = Vec4::from_slice(&extracted_asset.color.as_linear_rgba_f32()); let byte_buffer = [0u8; Vec4::SIZE.get() as usize]; let mut buffer = encase::UniformBuffer::new(byte_buffer); buffer.write(&color).unwrap(); let buffer = render_device.create_buffer_with_data(&BufferInitDescriptor { contents: buffer.as_ref(), label: None, usage: BufferUsages::UNIFORM | BufferUsages::COPY_DST, }); let (texture_view, texture_sampler) = if let Some(result) = material_pipeline .mesh_pipeline .get_image_texture(gpu_images, &Some(extracted_asset.color_texture.clone())) { result } else { return Err(PrepareAssetError::RetryNextUpdate(extracted_asset)); }; let bind_group = render_device.create_bind_group(&BindGroupDescriptor { entries: &[ BindGroupEntry { binding: 0, resource: buffer.as_entire_binding(), }, BindGroupEntry { binding: 0, resource: BindingResource::TextureView(texture_view), }, BindGroupEntry { binding: 1, resource: BindingResource::Sampler(texture_sampler), }, ], label: None, layout: &material_pipeline.material_layout, }); Ok(GpuCustomMaterial { _buffer: buffer, bind_group, }) } } impl Material for CustomMaterial { fn fragment_shader(asset_server: &AssetServer) -> Option<Handle<Shader>> { Some(asset_server.load("custom_material.wgsl")) } fn bind_group(render_asset: &<Self as RenderAsset>::PreparedAsset) -> &BindGroup { &render_asset.bind_group } fn bind_group_layout(render_device: &RenderDevice) -> BindGroupLayout { render_device.create_bind_group_layout(&BindGroupLayoutDescriptor { entries: &[ BindGroupLayoutEntry { binding: 0, visibility: ShaderStages::FRAGMENT, ty: BindingType::Buffer { ty: BufferBindingType::Uniform, has_dynamic_offset: false, min_binding_size: Some(Vec4::min_size()), }, count: None, }, BindGroupLayoutEntry { binding: 1, visibility: ShaderStages::FRAGMENT, ty: BindingType::Texture { multisampled: false, sample_type: TextureSampleType::Float { filterable: true }, view_dimension: TextureViewDimension::D2Array, }, count: None, }, BindGroupLayoutEntry { binding: 2, visibility: ShaderStages::FRAGMENT, ty: BindingType::Sampler(SamplerBindingType::Filtering), count: None, }, ], label: None, }) } } ``` ### Bevy 0.8 (new) ```rust impl Material for CustomMaterial { fn fragment_shader() -> ShaderRef { "custom_material.wgsl".into() } } #[derive(AsBindGroup, TypeUuid, Debug, Clone)] #[uuid = "f690fdae-d598-45ab-8225-97e2a3f056e0"] pub struct CustomMaterial { #[uniform(0)] color: Color, #[texture(1)] #[sampler(2)] color_texture: Handle<Image>, } ``` ## Future Work * Add support for more binding types (cubemaps, buffers, etc). This PR intentionally includes a bare minimum number of binding types to keep "reviewability" in check. * Consider optionally eliding binding indices using binding names. `AsBindGroup` could pass in (optional?) reflection info as a "hint". * This would make it possible for the derive to do this: ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[uniform] color: Color, #[texture] #[sampler] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or this ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { #[binding] color: Color, #[binding] color_texture: Option<Handle<Image>>, alpha_mode: AlphaMode, } ``` * Or even this (if we flip to "include bindings by default") ```rust #[derive(AsBindGroup)] pub struct CustomMaterial { color: Color, color_texture: Option<Handle<Image>>, #[binding(ignore)] alpha_mode: AlphaMode, } ``` * If we add the option to define custom draw functions for materials (which could be done in a type-erased way), I think that would be enough to support extra non-material bindings. Worth considering!
2022-06-30 23:48:46 +00:00
match prepare_material(
&material,
&render_device,
&images,
&fallback_image,
&pipeline,
) {
Ok(prepared_asset) => {
render_materials.insert(handle, prepared_asset);
}
Err(AsBindGroupError::RetryNextUpdate) => {
prepare_next_frame.assets.push((handle, material));
}
}
}
for removed in std::mem::take(&mut extracted_assets.removed) {
render_materials.remove(&removed);
}
for (handle, material) in std::mem::take(&mut extracted_assets.extracted) {
match prepare_material(
&material,
&render_device,
&images,
&fallback_image,
&pipeline,
) {
Ok(prepared_asset) => {
render_materials.insert(handle, prepared_asset);
}
Err(AsBindGroupError::RetryNextUpdate) => {
prepare_next_frame.assets.push((handle, material));
}
}
}
}
fn prepare_material<M: Material>(
material: &M,
render_device: &RenderDevice,
images: &RenderAssets<Image>,
fallback_image: &FallbackImage,
pipeline: &MaterialPipeline<M>,
) -> Result<PreparedMaterial<M>, AsBindGroupError> {
let prepared = material.as_bind_group(
&pipeline.material_layout,
render_device,
images,
fallback_image,
)?;
Ok(PreparedMaterial {
bindings: prepared.bindings,
bind_group: prepared.bind_group,
key: prepared.data,
properties: MaterialProperties {
alpha_mode: material.alpha_mode(),
depth_bias: material.depth_bias(),
},
})
}