bevy/crates/bevy_render/src/renderer/render_device.rs
robtfm 2cd0bd7575 improve compile time by type-erasing wgpu structs (#5950)
# Objective

structs containing wgpu types take a long time to compile. this is particularly bad for generics containing the wgpu structs (like the depth pipeline builder with `#[derive(SystemParam)]` i've been working on).

we can avoid that by boxing and type-erasing in the bevy `render_resource` wrappers.

type system magic is not a strength of mine so i guess there will be a cleaner way to achieve this, happy to take feedback or for it to be taken as a proof of concept if someone else wants to do a better job.

## Solution

- add macros to box and type-erase in debug mode
- leave current impl for release mode

timings:


<html xmlns:v="urn:schemas-microsoft-com:vml"
xmlns:o="urn:schemas-microsoft-com:office:office"
xmlns:x="urn:schemas-microsoft-com:office:excel"
xmlns="http://www.w3.org/TR/REC-html40">

<head>

<meta name=ProgId content=Excel.Sheet>
<meta name=Generator content="Microsoft Excel 15">
<link id=Main-File rel=Main-File
href="file:///C:/Users/robfm/AppData/Local/Temp/msohtmlclip1/01/clip.htm">
<link rel=File-List
href="file:///C:/Users/robfm/AppData/Local/Temp/msohtmlclip1/01/clip_filelist.xml">
<!--table
	{mso-displayed-decimal-separator:"\.";
	mso-displayed-thousand-separator:"\,";}
@page
	{margin:.75in .7in .75in .7in;
	mso-header-margin:.3in;
	mso-footer-margin:.3in;}
tr
	{mso-height-source:auto;}
col
	{mso-width-source:auto;}
br
	{mso-data-placement:same-cell;}
td
	{padding-top:1px;
	padding-right:1px;
	padding-left:1px;
	mso-ignore:padding;
	color:black;
	font-size:11.0pt;
	font-weight:400;
	font-style:normal;
	text-decoration:none;
	font-family:Calibri, sans-serif;
	mso-font-charset:0;
	mso-number-format:General;
	text-align:general;
	vertical-align:bottom;
	border:none;
	mso-background-source:auto;
	mso-pattern:auto;
	mso-protection:locked visible;
	white-space:nowrap;
	mso-rotate:0;}
.xl65
	{mso-number-format:0%;}
.xl66
	{vertical-align:middle;
	white-space:normal;}
.xl67
	{vertical-align:middle;}
-->
</head>

<body link="#0563C1" vlink="#954F72">



current |   |   |  
-- | -- | -- | --
  | Total time: | 64.9s |  
  | bevy_pbr v0.9.0-dev | 19.2s |  
  | bevy_render v0.9.0-dev | 17.0s |  
  | bevy_sprite v0.9.0-dev | 15.1s |  
  | DepthPipelineBuilder | 18.7s |  
  |   |   |  
with type-erasing |   |   | diff
  | Total time: | 49.0s | -24%
  | bevy_render v0.9.0-dev | 12.0s | -38%
  | bevy_pbr v0.9.0-dev | 8.7s | -49%
  | bevy_sprite v0.9.0-dev | 6.1s | -60%
  | DepthPipelineBuilder | 1.2s | -94%



</body>

</html>

the depth pipeline builder is a binary with body: 
```rust
use std::{marker::PhantomData, hash::Hash};
use bevy::{prelude::*, ecs::system::SystemParam, pbr::{RenderMaterials, MaterialPipeline, ShadowPipeline}, render::{renderer::RenderDevice, render_resource::{SpecializedMeshPipelines, PipelineCache}, render_asset::RenderAssets}};

fn main() {
    println!("Hello, world p!\n");
}

#[derive(SystemParam)]
pub struct DepthPipelineBuilder<'w, 's, M: Material> 
where M::Data: Eq + Hash + Clone,
{
    render_device: Res<'w, RenderDevice>,
    material_pipeline: Res<'w, MaterialPipeline<M>>,
    material_pipelines: ResMut<'w, SpecializedMeshPipelines<MaterialPipeline<M>>>,
    shadow_pipeline: Res<'w, ShadowPipeline>,
    pipeline_cache: ResMut<'w, PipelineCache>,
    render_meshes: Res<'w, RenderAssets<Mesh>>,
    render_materials: Res<'w, RenderMaterials<M>>,
    msaa: Res<'w, Msaa>,
    #[system_param(ignore)]
    _p: PhantomData<&'s M>,
}
```
2022-11-18 22:04:23 +00:00

204 lines
6.5 KiB
Rust

use crate::render_resource::{
BindGroup, BindGroupLayout, Buffer, ComputePipeline, RawRenderPipelineDescriptor,
RenderPipeline, Sampler, Texture,
};
use bevy_ecs::system::Resource;
use wgpu::{util::DeviceExt, BufferAsyncError, BufferBindingType};
use super::RenderQueue;
use crate::render_resource::resource_macros::*;
render_resource_wrapper!(ErasedRenderDevice, wgpu::Device);
/// This GPU device is responsible for the creation of most rendering and compute resources.
#[derive(Resource, Clone)]
pub struct RenderDevice {
device: ErasedRenderDevice,
}
impl From<wgpu::Device> for RenderDevice {
fn from(device: wgpu::Device) -> Self {
Self {
device: ErasedRenderDevice::new(device),
}
}
}
impl RenderDevice {
/// List all [`Features`](wgpu::Features) that may be used with this device.
///
/// Functions may panic if you use unsupported features.
#[inline]
pub fn features(&self) -> wgpu::Features {
self.device.features()
}
/// List all [`Limits`](wgpu::Limits) that were requested of this device.
///
/// If any of these limits are exceeded, functions may panic.
#[inline]
pub fn limits(&self) -> wgpu::Limits {
self.device.limits()
}
/// Creates a [`ShaderModule`](wgpu::ShaderModule) from either SPIR-V or WGSL source code.
#[inline]
pub fn create_shader_module(&self, desc: wgpu::ShaderModuleDescriptor) -> wgpu::ShaderModule {
self.device.create_shader_module(desc)
}
/// Check for resource cleanups and mapping callbacks.
///
/// no-op on the web, device is automatically polled.
#[inline]
pub fn poll(&self, maintain: wgpu::Maintain) {
self.device.poll(maintain);
}
/// Creates an empty [`CommandEncoder`](wgpu::CommandEncoder).
#[inline]
pub fn create_command_encoder(
&self,
desc: &wgpu::CommandEncoderDescriptor,
) -> wgpu::CommandEncoder {
self.device.create_command_encoder(desc)
}
/// Creates an empty [`RenderBundleEncoder`](wgpu::RenderBundleEncoder).
#[inline]
pub fn create_render_bundle_encoder(
&self,
desc: &wgpu::RenderBundleEncoderDescriptor,
) -> wgpu::RenderBundleEncoder {
self.device.create_render_bundle_encoder(desc)
}
/// Creates a new [`BindGroup`](wgpu::BindGroup).
#[inline]
pub fn create_bind_group(&self, desc: &wgpu::BindGroupDescriptor) -> BindGroup {
let wgpu_bind_group = self.device.create_bind_group(desc);
BindGroup::from(wgpu_bind_group)
}
/// Creates a [`BindGroupLayout`](wgpu::BindGroupLayout).
#[inline]
pub fn create_bind_group_layout(
&self,
desc: &wgpu::BindGroupLayoutDescriptor,
) -> BindGroupLayout {
BindGroupLayout::from(self.device.create_bind_group_layout(desc))
}
/// Creates a [`PipelineLayout`](wgpu::PipelineLayout).
#[inline]
pub fn create_pipeline_layout(
&self,
desc: &wgpu::PipelineLayoutDescriptor,
) -> wgpu::PipelineLayout {
self.device.create_pipeline_layout(desc)
}
/// Creates a [`RenderPipeline`].
#[inline]
pub fn create_render_pipeline(&self, desc: &RawRenderPipelineDescriptor) -> RenderPipeline {
let wgpu_render_pipeline = self.device.create_render_pipeline(desc);
RenderPipeline::from(wgpu_render_pipeline)
}
/// Creates a [`ComputePipeline`].
#[inline]
pub fn create_compute_pipeline(
&self,
desc: &wgpu::ComputePipelineDescriptor,
) -> ComputePipeline {
let wgpu_compute_pipeline = self.device.create_compute_pipeline(desc);
ComputePipeline::from(wgpu_compute_pipeline)
}
/// Creates a [`Buffer`].
pub fn create_buffer(&self, desc: &wgpu::BufferDescriptor) -> Buffer {
let wgpu_buffer = self.device.create_buffer(desc);
Buffer::from(wgpu_buffer)
}
/// Creates a [`Buffer`] and initializes it with the specified data.
pub fn create_buffer_with_data(&self, desc: &wgpu::util::BufferInitDescriptor) -> Buffer {
let wgpu_buffer = self.device.create_buffer_init(desc);
Buffer::from(wgpu_buffer)
}
/// Creates a new [`Texture`] and initializes it with the specified data.
///
/// `desc` specifies the general format of the texture.
/// `data` is the raw data.
pub fn create_texture_with_data(
&self,
render_queue: &RenderQueue,
desc: &wgpu::TextureDescriptor,
data: &[u8],
) -> Texture {
let wgpu_texture = self
.device
.create_texture_with_data(render_queue.as_ref(), desc, data);
Texture::from(wgpu_texture)
}
/// Creates a new [`Texture`].
///
/// `desc` specifies the general format of the texture.
pub fn create_texture(&self, desc: &wgpu::TextureDescriptor) -> Texture {
let wgpu_texture = self.device.create_texture(desc);
Texture::from(wgpu_texture)
}
/// Creates a new [`Sampler`].
///
/// `desc` specifies the behavior of the sampler.
pub fn create_sampler(&self, desc: &wgpu::SamplerDescriptor) -> Sampler {
let wgpu_sampler = self.device.create_sampler(desc);
Sampler::from(wgpu_sampler)
}
/// Initializes [`Surface`](wgpu::Surface) for presentation.
///
/// # Panics
///
/// - A old [`SurfaceTexture`](wgpu::SurfaceTexture) is still alive referencing an old surface.
/// - Texture format requested is unsupported on the surface.
pub fn configure_surface(&self, surface: &wgpu::Surface, config: &wgpu::SurfaceConfiguration) {
surface.configure(&self.device, config);
}
/// Returns the wgpu [`Device`](wgpu::Device).
pub fn wgpu_device(&self) -> &wgpu::Device {
&self.device
}
pub fn map_buffer(
&self,
buffer: &wgpu::BufferSlice,
map_mode: wgpu::MapMode,
callback: impl FnOnce(Result<(), BufferAsyncError>) + Send + 'static,
) {
buffer.map_async(map_mode, callback);
}
pub fn align_copy_bytes_per_row(row_bytes: usize) -> usize {
let align = wgpu::COPY_BYTES_PER_ROW_ALIGNMENT as usize;
let padded_bytes_per_row_padding = (align - row_bytes % align) % align;
row_bytes + padded_bytes_per_row_padding
}
pub fn get_supported_read_only_binding_type(
&self,
buffers_per_shader_stage: u32,
) -> BufferBindingType {
if self.limits().max_storage_buffers_per_shader_stage >= buffers_per_shader_stage {
BufferBindingType::Storage { read_only: true }
} else {
BufferBindingType::Uniform
}
}
}