You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
845 lines
31 KiB
845 lines
31 KiB
4 years ago
|
use std::{iter, mem, num::NonZeroU32, ops::Range, rc::Rc};
|
||
|
|
||
|
#[path = "../framework.rs"]
|
||
|
mod framework;
|
||
|
|
||
|
use bytemuck::{Pod, Zeroable};
|
||
|
use wgpu::util::DeviceExt;
|
||
|
|
||
|
#[repr(C)]
|
||
|
#[derive(Clone, Copy, Pod, Zeroable)]
|
||
|
struct Vertex {
|
||
|
_pos: [i8; 4],
|
||
|
_normal: [i8; 4],
|
||
|
}
|
||
|
|
||
|
fn vertex(pos: [i8; 3], nor: [i8; 3]) -> Vertex {
|
||
|
Vertex {
|
||
|
_pos: [pos[0], pos[1], pos[2], 1],
|
||
|
_normal: [nor[0], nor[1], nor[2], 0],
|
||
|
}
|
||
|
}
|
||
|
|
||
|
fn create_cube() -> (Vec<Vertex>, Vec<u16>) {
|
||
|
let vertex_data = [
|
||
|
// top (0, 0, 1)
|
||
|
vertex([-1, -1, 1], [0, 0, 1]),
|
||
|
vertex([1, -1, 1], [0, 0, 1]),
|
||
|
vertex([1, 1, 1], [0, 0, 1]),
|
||
|
vertex([-1, 1, 1], [0, 0, 1]),
|
||
|
// bottom (0, 0, -1)
|
||
|
vertex([-1, 1, -1], [0, 0, -1]),
|
||
|
vertex([1, 1, -1], [0, 0, -1]),
|
||
|
vertex([1, -1, -1], [0, 0, -1]),
|
||
|
vertex([-1, -1, -1], [0, 0, -1]),
|
||
|
// right (1, 0, 0)
|
||
|
vertex([1, -1, -1], [1, 0, 0]),
|
||
|
vertex([1, 1, -1], [1, 0, 0]),
|
||
|
vertex([1, 1, 1], [1, 0, 0]),
|
||
|
vertex([1, -1, 1], [1, 0, 0]),
|
||
|
// left (-1, 0, 0)
|
||
|
vertex([-1, -1, 1], [-1, 0, 0]),
|
||
|
vertex([-1, 1, 1], [-1, 0, 0]),
|
||
|
vertex([-1, 1, -1], [-1, 0, 0]),
|
||
|
vertex([-1, -1, -1], [-1, 0, 0]),
|
||
|
// front (0, 1, 0)
|
||
|
vertex([1, 1, -1], [0, 1, 0]),
|
||
|
vertex([-1, 1, -1], [0, 1, 0]),
|
||
|
vertex([-1, 1, 1], [0, 1, 0]),
|
||
|
vertex([1, 1, 1], [0, 1, 0]),
|
||
|
// back (0, -1, 0)
|
||
|
vertex([1, -1, 1], [0, -1, 0]),
|
||
|
vertex([-1, -1, 1], [0, -1, 0]),
|
||
|
vertex([-1, -1, -1], [0, -1, 0]),
|
||
|
vertex([1, -1, -1], [0, -1, 0]),
|
||
|
];
|
||
|
|
||
|
let index_data: &[u16] = &[
|
||
|
0, 1, 2, 2, 3, 0, // top
|
||
|
4, 5, 6, 6, 7, 4, // bottom
|
||
|
8, 9, 10, 10, 11, 8, // right
|
||
|
12, 13, 14, 14, 15, 12, // left
|
||
|
16, 17, 18, 18, 19, 16, // front
|
||
|
20, 21, 22, 22, 23, 20, // back
|
||
|
];
|
||
|
|
||
|
(vertex_data.to_vec(), index_data.to_vec())
|
||
|
}
|
||
|
|
||
|
fn create_plane(size: i8) -> (Vec<Vertex>, Vec<u16>) {
|
||
|
let vertex_data = [
|
||
|
vertex([size, -size, 0], [0, 0, 1]),
|
||
|
vertex([size, size, 0], [0, 0, 1]),
|
||
|
vertex([-size, -size, 0], [0, 0, 1]),
|
||
|
vertex([-size, size, 0], [0, 0, 1]),
|
||
|
];
|
||
|
|
||
|
let index_data: &[u16] = &[0, 1, 2, 2, 1, 3];
|
||
|
|
||
|
(vertex_data.to_vec(), index_data.to_vec())
|
||
|
}
|
||
|
|
||
|
struct Entity {
|
||
|
mx_world: cgmath::Matrix4<f32>,
|
||
|
rotation_speed: f32,
|
||
|
color: wgpu::Color,
|
||
|
vertex_buf: Rc<wgpu::Buffer>,
|
||
|
index_buf: Rc<wgpu::Buffer>,
|
||
|
index_format: wgpu::IndexFormat,
|
||
|
index_count: usize,
|
||
|
uniform_offset: wgpu::DynamicOffset,
|
||
|
}
|
||
|
|
||
|
struct Light {
|
||
|
pos: cgmath::Point3<f32>,
|
||
|
color: wgpu::Color,
|
||
|
fov: f32,
|
||
|
depth: Range<f32>,
|
||
|
target_view: wgpu::TextureView,
|
||
|
}
|
||
|
|
||
|
#[repr(C)]
|
||
|
#[derive(Clone, Copy, Pod, Zeroable)]
|
||
|
struct LightRaw {
|
||
|
proj: [[f32; 4]; 4],
|
||
|
pos: [f32; 4],
|
||
|
color: [f32; 4],
|
||
|
}
|
||
|
|
||
|
impl Light {
|
||
|
fn to_raw(&self) -> LightRaw {
|
||
|
use cgmath::{Deg, EuclideanSpace, Matrix4, PerspectiveFov, Point3, Vector3};
|
||
|
|
||
|
let mx_view = Matrix4::look_at(self.pos, Point3::origin(), Vector3::unit_z());
|
||
|
let projection = PerspectiveFov {
|
||
|
fovy: Deg(self.fov).into(),
|
||
|
aspect: 1.0,
|
||
|
near: self.depth.start,
|
||
|
far: self.depth.end,
|
||
|
};
|
||
|
let mx_correction = framework::OPENGL_TO_WGPU_MATRIX;
|
||
|
let mx_view_proj =
|
||
|
mx_correction * cgmath::Matrix4::from(projection.to_perspective()) * mx_view;
|
||
|
LightRaw {
|
||
|
proj: *mx_view_proj.as_ref(),
|
||
|
pos: [self.pos.x, self.pos.y, self.pos.z, 1.0],
|
||
|
color: [
|
||
|
self.color.r as f32,
|
||
|
self.color.g as f32,
|
||
|
self.color.b as f32,
|
||
|
1.0,
|
||
|
],
|
||
|
}
|
||
|
}
|
||
|
}
|
||
|
|
||
|
#[repr(C)]
|
||
|
#[derive(Clone, Copy, Pod, Zeroable)]
|
||
|
struct ForwardUniforms {
|
||
|
proj: [[f32; 4]; 4],
|
||
|
num_lights: [u32; 4],
|
||
|
}
|
||
|
|
||
|
#[repr(C)]
|
||
|
#[derive(Clone, Copy, Pod, Zeroable)]
|
||
|
struct EntityUniforms {
|
||
|
model: [[f32; 4]; 4],
|
||
|
color: [f32; 4],
|
||
|
}
|
||
|
|
||
|
#[repr(C)]
|
||
|
struct ShadowUniforms {
|
||
|
proj: [[f32; 4]; 4],
|
||
|
}
|
||
|
|
||
|
struct Pass {
|
||
|
pipeline: wgpu::RenderPipeline,
|
||
|
bind_group: wgpu::BindGroup,
|
||
|
uniform_buf: wgpu::Buffer,
|
||
|
}
|
||
|
|
||
|
struct Example {
|
||
|
entities: Vec<Entity>,
|
||
|
lights: Vec<Light>,
|
||
|
lights_are_dirty: bool,
|
||
|
shadow_pass: Pass,
|
||
|
forward_pass: Pass,
|
||
|
forward_depth: wgpu::TextureView,
|
||
|
entity_bind_group: wgpu::BindGroup,
|
||
|
light_uniform_buf: wgpu::Buffer,
|
||
|
entity_uniform_buf: wgpu::Buffer,
|
||
|
}
|
||
|
|
||
|
impl Example {
|
||
|
const MAX_LIGHTS: usize = 10;
|
||
|
const SHADOW_FORMAT: wgpu::TextureFormat = wgpu::TextureFormat::Depth32Float;
|
||
|
const SHADOW_SIZE: wgpu::Extent3d = wgpu::Extent3d {
|
||
|
width: 512,
|
||
|
height: 512,
|
||
|
depth: Self::MAX_LIGHTS as u32,
|
||
|
};
|
||
|
const DEPTH_FORMAT: wgpu::TextureFormat = wgpu::TextureFormat::Depth32Float;
|
||
|
|
||
|
fn generate_matrix(aspect_ratio: f32) -> cgmath::Matrix4<f32> {
|
||
|
let mx_projection = cgmath::perspective(cgmath::Deg(45f32), aspect_ratio, 1.0, 20.0);
|
||
|
let mx_view = cgmath::Matrix4::look_at(
|
||
|
cgmath::Point3::new(3.0f32, -10.0, 6.0),
|
||
|
cgmath::Point3::new(0f32, 0.0, 0.0),
|
||
|
cgmath::Vector3::unit_z(),
|
||
|
);
|
||
|
let mx_correction = framework::OPENGL_TO_WGPU_MATRIX;
|
||
|
mx_correction * mx_projection * mx_view
|
||
|
}
|
||
|
}
|
||
|
|
||
|
impl framework::Example for Example {
|
||
|
fn optional_features() -> wgpu::Features {
|
||
|
wgpu::Features::DEPTH_CLAMPING
|
||
|
}
|
||
|
|
||
|
fn init(
|
||
|
sc_desc: &wgpu::SwapChainDescriptor,
|
||
|
device: &wgpu::Device,
|
||
|
_queue: &wgpu::Queue,
|
||
|
) -> Self {
|
||
|
// Create the vertex and index buffers
|
||
|
let vertex_size = mem::size_of::<Vertex>();
|
||
|
let (cube_vertex_data, cube_index_data) = create_cube();
|
||
|
let cube_vertex_buf = Rc::new(device.create_buffer_init(
|
||
|
&wgpu::util::BufferInitDescriptor {
|
||
|
label: Some("Cubes Vertex Buffer"),
|
||
|
contents: bytemuck::cast_slice(&cube_vertex_data),
|
||
|
usage: wgpu::BufferUsage::VERTEX,
|
||
|
},
|
||
|
));
|
||
|
|
||
|
let cube_index_buf = Rc::new(device.create_buffer_init(
|
||
|
&wgpu::util::BufferInitDescriptor {
|
||
|
label: Some("Cubes Index Buffer"),
|
||
|
contents: bytemuck::cast_slice(&cube_index_data),
|
||
|
usage: wgpu::BufferUsage::INDEX,
|
||
|
},
|
||
|
));
|
||
|
|
||
|
let (plane_vertex_data, plane_index_data) = create_plane(7);
|
||
|
let plane_vertex_buf = device.create_buffer_init(&wgpu::util::BufferInitDescriptor {
|
||
|
label: Some("Plane Vertex Buffer"),
|
||
|
contents: bytemuck::cast_slice(&plane_vertex_data),
|
||
|
usage: wgpu::BufferUsage::VERTEX,
|
||
|
});
|
||
|
|
||
|
let plane_index_buf = device.create_buffer_init(&wgpu::util::BufferInitDescriptor {
|
||
|
label: Some("Plane Index Buffer"),
|
||
|
contents: bytemuck::cast_slice(&plane_index_data),
|
||
|
usage: wgpu::BufferUsage::INDEX,
|
||
|
});
|
||
|
|
||
|
struct CubeDesc {
|
||
|
offset: cgmath::Vector3<f32>,
|
||
|
angle: f32,
|
||
|
scale: f32,
|
||
|
rotation: f32,
|
||
|
}
|
||
|
let cube_descs = [
|
||
|
CubeDesc {
|
||
|
offset: cgmath::vec3(-2.0, -2.0, 2.0),
|
||
|
angle: 10.0,
|
||
|
scale: 0.7,
|
||
|
rotation: 0.1,
|
||
|
},
|
||
|
CubeDesc {
|
||
|
offset: cgmath::vec3(2.0, -2.0, 2.0),
|
||
|
angle: 50.0,
|
||
|
scale: 1.3,
|
||
|
rotation: 0.2,
|
||
|
},
|
||
|
CubeDesc {
|
||
|
offset: cgmath::vec3(-2.0, 2.0, 2.0),
|
||
|
angle: 140.0,
|
||
|
scale: 1.1,
|
||
|
rotation: 0.3,
|
||
|
},
|
||
|
CubeDesc {
|
||
|
offset: cgmath::vec3(2.0, 2.0, 2.0),
|
||
|
angle: 210.0,
|
||
|
scale: 0.9,
|
||
|
rotation: 0.4,
|
||
|
},
|
||
|
];
|
||
|
|
||
|
let entity_uniform_size = mem::size_of::<EntityUniforms>() as wgpu::BufferAddress;
|
||
|
let num_entities = 1 + cube_descs.len() as wgpu::BufferAddress;
|
||
|
assert!(entity_uniform_size <= wgpu::BIND_BUFFER_ALIGNMENT);
|
||
|
//Note: dynamic offsets also have to be aligned to `BIND_BUFFER_ALIGNMENT`.
|
||
|
let entity_uniform_buf = device.create_buffer(&wgpu::BufferDescriptor {
|
||
|
label: None,
|
||
|
size: num_entities * wgpu::BIND_BUFFER_ALIGNMENT,
|
||
|
usage: wgpu::BufferUsage::UNIFORM | wgpu::BufferUsage::COPY_DST,
|
||
|
mapped_at_creation: false,
|
||
|
});
|
||
|
|
||
|
let index_format = wgpu::IndexFormat::Uint16;
|
||
|
|
||
|
let mut entities = vec![{
|
||
|
use cgmath::SquareMatrix;
|
||
|
Entity {
|
||
|
mx_world: cgmath::Matrix4::identity(),
|
||
|
rotation_speed: 0.0,
|
||
|
color: wgpu::Color::WHITE,
|
||
|
vertex_buf: Rc::new(plane_vertex_buf),
|
||
|
index_buf: Rc::new(plane_index_buf),
|
||
|
index_format,
|
||
|
index_count: plane_index_data.len(),
|
||
|
uniform_offset: 0,
|
||
|
}
|
||
|
}];
|
||
|
|
||
|
for (i, cube) in cube_descs.iter().enumerate() {
|
||
|
use cgmath::{Decomposed, Deg, InnerSpace, Quaternion, Rotation3};
|
||
|
|
||
|
let transform = Decomposed {
|
||
|
disp: cube.offset.clone(),
|
||
|
rot: Quaternion::from_axis_angle(cube.offset.normalize(), Deg(cube.angle)),
|
||
|
scale: cube.scale,
|
||
|
};
|
||
|
entities.push(Entity {
|
||
|
mx_world: cgmath::Matrix4::from(transform),
|
||
|
rotation_speed: cube.rotation,
|
||
|
color: wgpu::Color::GREEN,
|
||
|
vertex_buf: Rc::clone(&cube_vertex_buf),
|
||
|
index_buf: Rc::clone(&cube_index_buf),
|
||
|
index_format,
|
||
|
index_count: cube_index_data.len(),
|
||
|
uniform_offset: ((i + 1) * wgpu::BIND_BUFFER_ALIGNMENT as usize) as _,
|
||
|
});
|
||
|
}
|
||
|
|
||
|
let local_bind_group_layout =
|
||
|
device.create_bind_group_layout(&wgpu::BindGroupLayoutDescriptor {
|
||
|
entries: &[wgpu::BindGroupLayoutEntry {
|
||
|
binding: 0,
|
||
|
visibility: wgpu::ShaderStage::VERTEX | wgpu::ShaderStage::FRAGMENT,
|
||
|
ty: wgpu::BindingType::Buffer {
|
||
|
ty: wgpu::BufferBindingType::Uniform,
|
||
|
has_dynamic_offset: true,
|
||
|
min_binding_size: wgpu::BufferSize::new(entity_uniform_size),
|
||
|
},
|
||
|
count: None,
|
||
|
}],
|
||
|
label: None,
|
||
|
});
|
||
|
let entity_bind_group = device.create_bind_group(&wgpu::BindGroupDescriptor {
|
||
|
layout: &local_bind_group_layout,
|
||
|
entries: &[wgpu::BindGroupEntry {
|
||
|
binding: 0,
|
||
|
resource: wgpu::BindingResource::Buffer {
|
||
|
buffer: &entity_uniform_buf,
|
||
|
offset: 0,
|
||
|
size: wgpu::BufferSize::new(entity_uniform_size),
|
||
|
},
|
||
|
}],
|
||
|
label: None,
|
||
|
});
|
||
|
|
||
|
// Create other resources
|
||
|
let shadow_sampler = device.create_sampler(&wgpu::SamplerDescriptor {
|
||
|
label: Some("shadow"),
|
||
|
address_mode_u: wgpu::AddressMode::ClampToEdge,
|
||
|
address_mode_v: wgpu::AddressMode::ClampToEdge,
|
||
|
address_mode_w: wgpu::AddressMode::ClampToEdge,
|
||
|
mag_filter: wgpu::FilterMode::Linear,
|
||
|
min_filter: wgpu::FilterMode::Linear,
|
||
|
mipmap_filter: wgpu::FilterMode::Nearest,
|
||
|
compare: Some(wgpu::CompareFunction::LessEqual),
|
||
|
..Default::default()
|
||
|
});
|
||
|
|
||
|
let shadow_texture = device.create_texture(&wgpu::TextureDescriptor {
|
||
|
size: Self::SHADOW_SIZE,
|
||
|
mip_level_count: 1,
|
||
|
sample_count: 1,
|
||
|
dimension: wgpu::TextureDimension::D2,
|
||
|
format: Self::SHADOW_FORMAT,
|
||
|
usage: wgpu::TextureUsage::RENDER_ATTACHMENT | wgpu::TextureUsage::SAMPLED,
|
||
|
label: None,
|
||
|
});
|
||
|
let shadow_view = shadow_texture.create_view(&wgpu::TextureViewDescriptor::default());
|
||
|
|
||
|
let mut shadow_target_views = (0..2)
|
||
|
.map(|i| {
|
||
|
Some(shadow_texture.create_view(&wgpu::TextureViewDescriptor {
|
||
|
label: Some("shadow"),
|
||
|
format: None,
|
||
|
dimension: Some(wgpu::TextureViewDimension::D2),
|
||
|
aspect: wgpu::TextureAspect::All,
|
||
|
base_mip_level: 0,
|
||
|
level_count: None,
|
||
|
base_array_layer: i as u32,
|
||
|
array_layer_count: NonZeroU32::new(1),
|
||
|
}))
|
||
|
})
|
||
|
.collect::<Vec<_>>();
|
||
|
let lights = vec![
|
||
|
Light {
|
||
|
pos: cgmath::Point3::new(7.0, -5.0, 10.0),
|
||
|
color: wgpu::Color {
|
||
|
r: 0.5,
|
||
|
g: 1.0,
|
||
|
b: 0.5,
|
||
|
a: 1.0,
|
||
|
},
|
||
|
fov: 60.0,
|
||
|
depth: 1.0..20.0,
|
||
|
target_view: shadow_target_views[0].take().unwrap(),
|
||
|
},
|
||
|
Light {
|
||
|
pos: cgmath::Point3::new(-5.0, 7.0, 10.0),
|
||
|
color: wgpu::Color {
|
||
|
r: 1.0,
|
||
|
g: 0.5,
|
||
|
b: 0.5,
|
||
|
a: 1.0,
|
||
|
},
|
||
|
fov: 45.0,
|
||
|
depth: 1.0..20.0,
|
||
|
target_view: shadow_target_views[1].take().unwrap(),
|
||
|
},
|
||
|
];
|
||
|
let light_uniform_size =
|
||
|
(Self::MAX_LIGHTS * mem::size_of::<LightRaw>()) as wgpu::BufferAddress;
|
||
|
let light_uniform_buf = device.create_buffer(&wgpu::BufferDescriptor {
|
||
|
label: None,
|
||
|
size: light_uniform_size,
|
||
|
usage: wgpu::BufferUsage::UNIFORM
|
||
|
| wgpu::BufferUsage::COPY_SRC
|
||
|
| wgpu::BufferUsage::COPY_DST,
|
||
|
mapped_at_creation: false,
|
||
|
});
|
||
|
|
||
|
let vertex_attr = wgpu::vertex_attr_array![0 => Char4, 1 => Char4];
|
||
|
let vb_desc = wgpu::VertexBufferDescriptor {
|
||
|
stride: vertex_size as wgpu::BufferAddress,
|
||
|
step_mode: wgpu::InputStepMode::Vertex,
|
||
|
attributes: &vertex_attr,
|
||
|
};
|
||
|
|
||
|
let shadow_pass = {
|
||
|
let uniform_size = mem::size_of::<ShadowUniforms>() as wgpu::BufferAddress;
|
||
|
// Create pipeline layout
|
||
|
let bind_group_layout =
|
||
|
device.create_bind_group_layout(&wgpu::BindGroupLayoutDescriptor {
|
||
|
label: None,
|
||
|
entries: &[wgpu::BindGroupLayoutEntry {
|
||
|
binding: 0, // global
|
||
|
visibility: wgpu::ShaderStage::VERTEX,
|
||
|
ty: wgpu::BindingType::Buffer {
|
||
|
ty: wgpu::BufferBindingType::Uniform,
|
||
|
has_dynamic_offset: false,
|
||
|
min_binding_size: wgpu::BufferSize::new(uniform_size),
|
||
|
},
|
||
|
count: None,
|
||
|
}],
|
||
|
});
|
||
|
let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
|
||
|
label: Some("shadow"),
|
||
|
bind_group_layouts: &[&bind_group_layout, &local_bind_group_layout],
|
||
|
push_constant_ranges: &[],
|
||
|
});
|
||
|
|
||
|
let uniform_buf = device.create_buffer(&wgpu::BufferDescriptor {
|
||
|
label: None,
|
||
|
size: uniform_size,
|
||
|
usage: wgpu::BufferUsage::UNIFORM | wgpu::BufferUsage::COPY_DST,
|
||
|
mapped_at_creation: false,
|
||
|
});
|
||
|
|
||
|
// Create bind group
|
||
|
let bind_group = device.create_bind_group(&wgpu::BindGroupDescriptor {
|
||
|
layout: &bind_group_layout,
|
||
|
entries: &[wgpu::BindGroupEntry {
|
||
|
binding: 0,
|
||
|
resource: uniform_buf.as_entire_binding(),
|
||
|
}],
|
||
|
label: None,
|
||
|
});
|
||
|
|
||
|
// Create the render pipeline
|
||
|
let vs_module = device.create_shader_module(&wgpu::include_spirv!("bake.vert.spv"));
|
||
|
|
||
|
let pipeline = device.create_render_pipeline(&wgpu::RenderPipelineDescriptor {
|
||
|
label: Some("shadow"),
|
||
|
layout: Some(&pipeline_layout),
|
||
|
vertex_stage: wgpu::ProgrammableStageDescriptor {
|
||
|
module: &vs_module,
|
||
|
entry_point: "main",
|
||
|
},
|
||
|
fragment_stage: None,
|
||
|
rasterization_state: Some(wgpu::RasterizationStateDescriptor {
|
||
|
front_face: wgpu::FrontFace::Ccw,
|
||
|
cull_mode: wgpu::CullMode::Back,
|
||
|
polygon_mode: wgpu::PolygonMode::Fill,
|
||
|
depth_bias: 2, // corresponds to bilinear filtering
|
||
|
depth_bias_slope_scale: 2.0,
|
||
|
depth_bias_clamp: 0.0,
|
||
|
clamp_depth: device.features().contains(wgpu::Features::DEPTH_CLAMPING),
|
||
|
}),
|
||
|
primitive_topology: wgpu::PrimitiveTopology::TriangleList,
|
||
|
color_states: &[],
|
||
|
depth_stencil_state: Some(wgpu::DepthStencilStateDescriptor {
|
||
|
format: Self::SHADOW_FORMAT,
|
||
|
depth_write_enabled: true,
|
||
|
depth_compare: wgpu::CompareFunction::LessEqual,
|
||
|
stencil: wgpu::StencilStateDescriptor::default(),
|
||
|
}),
|
||
|
vertex_state: wgpu::VertexStateDescriptor {
|
||
|
index_format: Some(index_format),
|
||
|
vertex_buffers: &[vb_desc.clone()],
|
||
|
},
|
||
|
sample_count: 1,
|
||
|
sample_mask: !0,
|
||
|
alpha_to_coverage_enabled: false,
|
||
|
});
|
||
|
|
||
|
Pass {
|
||
|
pipeline,
|
||
|
bind_group,
|
||
|
uniform_buf,
|
||
|
}
|
||
|
};
|
||
|
|
||
|
let forward_pass = {
|
||
|
// Create pipeline layout
|
||
|
let bind_group_layout =
|
||
|
device.create_bind_group_layout(&wgpu::BindGroupLayoutDescriptor {
|
||
|
entries: &[
|
||
|
wgpu::BindGroupLayoutEntry {
|
||
|
binding: 0, // global
|
||
|
visibility: wgpu::ShaderStage::VERTEX | wgpu::ShaderStage::FRAGMENT,
|
||
|
ty: wgpu::BindingType::Buffer {
|
||
|
ty: wgpu::BufferBindingType::Uniform,
|
||
|
has_dynamic_offset: false,
|
||
|
min_binding_size: wgpu::BufferSize::new(mem::size_of::<
|
||
|
ForwardUniforms,
|
||
|
>(
|
||
|
)
|
||
|
as _),
|
||
|
},
|
||
|
count: None,
|
||
|
},
|
||
|
wgpu::BindGroupLayoutEntry {
|
||
|
binding: 1, // lights
|
||
|
visibility: wgpu::ShaderStage::VERTEX | wgpu::ShaderStage::FRAGMENT,
|
||
|
ty: wgpu::BindingType::Buffer {
|
||
|
ty: wgpu::BufferBindingType::Uniform,
|
||
|
has_dynamic_offset: false,
|
||
|
min_binding_size: wgpu::BufferSize::new(light_uniform_size),
|
||
|
},
|
||
|
count: None,
|
||
|
},
|
||
|
wgpu::BindGroupLayoutEntry {
|
||
|
binding: 2,
|
||
|
visibility: wgpu::ShaderStage::FRAGMENT,
|
||
|
ty: wgpu::BindingType::Texture {
|
||
|
multisampled: false,
|
||
|
sample_type: wgpu::TextureSampleType::Depth,
|
||
|
view_dimension: wgpu::TextureViewDimension::D2Array,
|
||
|
},
|
||
|
count: None,
|
||
|
},
|
||
|
wgpu::BindGroupLayoutEntry {
|
||
|
binding: 3,
|
||
|
visibility: wgpu::ShaderStage::FRAGMENT,
|
||
|
ty: wgpu::BindingType::Sampler {
|
||
|
comparison: true,
|
||
|
filtering: false,
|
||
|
},
|
||
|
count: None,
|
||
|
},
|
||
|
],
|
||
|
label: None,
|
||
|
});
|
||
|
let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
|
||
|
label: Some("main"),
|
||
|
bind_group_layouts: &[&bind_group_layout, &local_bind_group_layout],
|
||
|
push_constant_ranges: &[],
|
||
|
});
|
||
|
|
||
|
let mx_total = Self::generate_matrix(sc_desc.width as f32 / sc_desc.height as f32);
|
||
|
let forward_uniforms = ForwardUniforms {
|
||
|
proj: *mx_total.as_ref(),
|
||
|
num_lights: [lights.len() as u32, 0, 0, 0],
|
||
|
};
|
||
|
let uniform_buf = device.create_buffer_init(&wgpu::util::BufferInitDescriptor {
|
||
|
label: Some("Uniform Buffer"),
|
||
|
contents: bytemuck::bytes_of(&forward_uniforms),
|
||
|
usage: wgpu::BufferUsage::UNIFORM | wgpu::BufferUsage::COPY_DST,
|
||
|
});
|
||
|
|
||
|
// Create bind group
|
||
|
let bind_group = device.create_bind_group(&wgpu::BindGroupDescriptor {
|
||
|
layout: &bind_group_layout,
|
||
|
entries: &[
|
||
|
wgpu::BindGroupEntry {
|
||
|
binding: 0,
|
||
|
resource: uniform_buf.as_entire_binding(),
|
||
|
},
|
||
|
wgpu::BindGroupEntry {
|
||
|
binding: 1,
|
||
|
resource: light_uniform_buf.as_entire_binding(),
|
||
|
},
|
||
|
wgpu::BindGroupEntry {
|
||
|
binding: 2,
|
||
|
resource: wgpu::BindingResource::TextureView(&shadow_view),
|
||
|
},
|
||
|
wgpu::BindGroupEntry {
|
||
|
binding: 3,
|
||
|
resource: wgpu::BindingResource::Sampler(&shadow_sampler),
|
||
|
},
|
||
|
],
|
||
|
label: None,
|
||
|
});
|
||
|
|
||
|
// Create the render pipeline
|
||
|
let vs_module = device.create_shader_module(&wgpu::include_spirv!("forward.vert.spv"));
|
||
|
let fs_module = device.create_shader_module(&wgpu::include_spirv!("forward.frag.spv"));
|
||
|
|
||
|
let pipeline = device.create_render_pipeline(&wgpu::RenderPipelineDescriptor {
|
||
|
label: Some("main"),
|
||
|
layout: Some(&pipeline_layout),
|
||
|
vertex_stage: wgpu::ProgrammableStageDescriptor {
|
||
|
module: &vs_module,
|
||
|
entry_point: "main",
|
||
|
},
|
||
|
fragment_stage: Some(wgpu::ProgrammableStageDescriptor {
|
||
|
module: &fs_module,
|
||
|
entry_point: "main",
|
||
|
}),
|
||
|
rasterization_state: Some(wgpu::RasterizationStateDescriptor {
|
||
|
front_face: wgpu::FrontFace::Ccw,
|
||
|
cull_mode: wgpu::CullMode::Back,
|
||
|
..Default::default()
|
||
|
}),
|
||
|
primitive_topology: wgpu::PrimitiveTopology::TriangleList,
|
||
|
color_states: &[sc_desc.format.into()],
|
||
|
depth_stencil_state: Some(wgpu::DepthStencilStateDescriptor {
|
||
|
format: Self::DEPTH_FORMAT,
|
||
|
depth_write_enabled: true,
|
||
|
depth_compare: wgpu::CompareFunction::Less,
|
||
|
stencil: wgpu::StencilStateDescriptor::default(),
|
||
|
}),
|
||
|
vertex_state: wgpu::VertexStateDescriptor {
|
||
|
index_format: Some(index_format),
|
||
|
vertex_buffers: &[vb_desc],
|
||
|
},
|
||
|
sample_count: 1,
|
||
|
sample_mask: !0,
|
||
|
alpha_to_coverage_enabled: false,
|
||
|
});
|
||
|
|
||
|
Pass {
|
||
|
pipeline,
|
||
|
bind_group,
|
||
|
uniform_buf,
|
||
|
}
|
||
|
};
|
||
|
|
||
|
let depth_texture = device.create_texture(&wgpu::TextureDescriptor {
|
||
|
size: wgpu::Extent3d {
|
||
|
width: sc_desc.width,
|
||
|
height: sc_desc.height,
|
||
|
depth: 1,
|
||
|
},
|
||
|
mip_level_count: 1,
|
||
|
sample_count: 1,
|
||
|
dimension: wgpu::TextureDimension::D2,
|
||
|
format: Self::DEPTH_FORMAT,
|
||
|
usage: wgpu::TextureUsage::RENDER_ATTACHMENT,
|
||
|
label: None,
|
||
|
});
|
||
|
|
||
|
Example {
|
||
|
entities,
|
||
|
lights,
|
||
|
lights_are_dirty: true,
|
||
|
shadow_pass,
|
||
|
forward_pass,
|
||
|
forward_depth: depth_texture.create_view(&wgpu::TextureViewDescriptor::default()),
|
||
|
light_uniform_buf,
|
||
|
entity_uniform_buf,
|
||
|
entity_bind_group,
|
||
|
}
|
||
|
}
|
||
|
|
||
|
fn update(&mut self, _event: winit::event::WindowEvent) {
|
||
|
//empty
|
||
|
}
|
||
|
|
||
|
fn resize(
|
||
|
&mut self,
|
||
|
sc_desc: &wgpu::SwapChainDescriptor,
|
||
|
device: &wgpu::Device,
|
||
|
queue: &wgpu::Queue,
|
||
|
) {
|
||
|
// update view-projection matrix
|
||
|
let mx_total = Self::generate_matrix(sc_desc.width as f32 / sc_desc.height as f32);
|
||
|
let mx_ref: &[f32; 16] = mx_total.as_ref();
|
||
|
queue.write_buffer(
|
||
|
&self.forward_pass.uniform_buf,
|
||
|
0,
|
||
|
bytemuck::cast_slice(mx_ref),
|
||
|
);
|
||
|
|
||
|
let depth_texture = device.create_texture(&wgpu::TextureDescriptor {
|
||
|
size: wgpu::Extent3d {
|
||
|
width: sc_desc.width,
|
||
|
height: sc_desc.height,
|
||
|
depth: 1,
|
||
|
},
|
||
|
mip_level_count: 1,
|
||
|
sample_count: 1,
|
||
|
dimension: wgpu::TextureDimension::D2,
|
||
|
format: Self::DEPTH_FORMAT,
|
||
|
usage: wgpu::TextureUsage::RENDER_ATTACHMENT,
|
||
|
label: None,
|
||
|
});
|
||
|
self.forward_depth = depth_texture.create_view(&wgpu::TextureViewDescriptor::default());
|
||
|
}
|
||
|
|
||
|
fn render(
|
||
|
&mut self,
|
||
|
frame: &wgpu::SwapChainTexture,
|
||
|
device: &wgpu::Device,
|
||
|
queue: &wgpu::Queue,
|
||
|
_spawner: &impl futures::task::LocalSpawn,
|
||
|
) {
|
||
|
// update uniforms
|
||
|
for entity in self.entities.iter_mut() {
|
||
|
if entity.rotation_speed != 0.0 {
|
||
|
let rotation = cgmath::Matrix4::from_angle_x(cgmath::Deg(entity.rotation_speed));
|
||
|
entity.mx_world = entity.mx_world * rotation;
|
||
|
}
|
||
|
let data = EntityUniforms {
|
||
|
model: entity.mx_world.into(),
|
||
|
color: [
|
||
|
entity.color.r as f32,
|
||
|
entity.color.g as f32,
|
||
|
entity.color.b as f32,
|
||
|
entity.color.a as f32,
|
||
|
],
|
||
|
};
|
||
|
queue.write_buffer(
|
||
|
&self.entity_uniform_buf,
|
||
|
entity.uniform_offset as wgpu::BufferAddress,
|
||
|
bytemuck::bytes_of(&data),
|
||
|
);
|
||
|
}
|
||
|
|
||
|
if self.lights_are_dirty {
|
||
|
self.lights_are_dirty = false;
|
||
|
for (i, light) in self.lights.iter().enumerate() {
|
||
|
queue.write_buffer(
|
||
|
&self.light_uniform_buf,
|
||
|
(i * mem::size_of::<LightRaw>()) as wgpu::BufferAddress,
|
||
|
bytemuck::bytes_of(&light.to_raw()),
|
||
|
);
|
||
|
}
|
||
|
}
|
||
|
|
||
|
let mut encoder =
|
||
|
device.create_command_encoder(&wgpu::CommandEncoderDescriptor { label: None });
|
||
|
|
||
|
encoder.push_debug_group("shadow passes");
|
||
|
for (i, light) in self.lights.iter().enumerate() {
|
||
|
encoder.push_debug_group(&format!(
|
||
|
"shadow pass {} (light at position {:?})",
|
||
|
i, light.pos
|
||
|
));
|
||
|
|
||
|
// The light uniform buffer already has the projection,
|
||
|
// let's just copy it over to the shadow uniform buffer.
|
||
|
encoder.copy_buffer_to_buffer(
|
||
|
&self.light_uniform_buf,
|
||
|
(i * mem::size_of::<LightRaw>()) as wgpu::BufferAddress,
|
||
|
&self.shadow_pass.uniform_buf,
|
||
|
0,
|
||
|
64,
|
||
|
);
|
||
|
|
||
|
encoder.insert_debug_marker("render entities");
|
||
|
{
|
||
|
let mut pass = encoder.begin_render_pass(&wgpu::RenderPassDescriptor {
|
||
|
label: None,
|
||
|
color_attachments: &[],
|
||
|
depth_stencil_attachment: Some(
|
||
|
wgpu::RenderPassDepthStencilAttachmentDescriptor {
|
||
|
attachment: &light.target_view,
|
||
|
depth_ops: Some(wgpu::Operations {
|
||
|
load: wgpu::LoadOp::Clear(1.0),
|
||
|
store: true,
|
||
|
}),
|
||
|
stencil_ops: None,
|
||
|
},
|
||
|
),
|
||
|
});
|
||
|
pass.set_pipeline(&self.shadow_pass.pipeline);
|
||
|
pass.set_bind_group(0, &self.shadow_pass.bind_group, &[]);
|
||
|
|
||
|
for entity in &self.entities {
|
||
|
pass.set_bind_group(1, &self.entity_bind_group, &[entity.uniform_offset]);
|
||
|
pass.set_index_buffer(entity.index_buf.slice(..), entity.index_format);
|
||
|
pass.set_vertex_buffer(0, entity.vertex_buf.slice(..));
|
||
|
pass.draw_indexed(0..entity.index_count as u32, 0, 0..1);
|
||
|
}
|
||
|
}
|
||
|
|
||
|
encoder.pop_debug_group();
|
||
|
}
|
||
|
encoder.pop_debug_group();
|
||
|
|
||
|
// forward pass
|
||
|
encoder.push_debug_group("forward rendering pass");
|
||
|
{
|
||
|
let mut pass = encoder.begin_render_pass(&wgpu::RenderPassDescriptor {
|
||
|
label: None,
|
||
|
color_attachments: &[wgpu::RenderPassColorAttachmentDescriptor {
|
||
|
attachment: &frame.view,
|
||
|
resolve_target: None,
|
||
|
ops: wgpu::Operations {
|
||
|
load: wgpu::LoadOp::Clear(wgpu::Color {
|
||
|
r: 0.1,
|
||
|
g: 0.2,
|
||
|
b: 0.3,
|
||
|
a: 1.0,
|
||
|
}),
|
||
|
store: true,
|
||
|
},
|
||
|
}],
|
||
|
depth_stencil_attachment: Some(wgpu::RenderPassDepthStencilAttachmentDescriptor {
|
||
|
attachment: &self.forward_depth,
|
||
|
depth_ops: Some(wgpu::Operations {
|
||
|
load: wgpu::LoadOp::Clear(1.0),
|
||
|
store: false,
|
||
|
}),
|
||
|
stencil_ops: None,
|
||
|
}),
|
||
|
});
|
||
|
pass.set_pipeline(&self.forward_pass.pipeline);
|
||
|
pass.set_bind_group(0, &self.forward_pass.bind_group, &[]);
|
||
|
|
||
|
for entity in &self.entities {
|
||
|
pass.set_bind_group(1, &self.entity_bind_group, &[entity.uniform_offset]);
|
||
|
pass.set_index_buffer(entity.index_buf.slice(..), entity.index_format);
|
||
|
pass.set_vertex_buffer(0, entity.vertex_buf.slice(..));
|
||
|
pass.draw_indexed(0..entity.index_count as u32, 0, 0..1);
|
||
|
}
|
||
|
}
|
||
|
encoder.pop_debug_group();
|
||
|
|
||
|
queue.submit(iter::once(encoder.finish()));
|
||
|
}
|
||
|
}
|
||
|
|
||
|
fn main() {
|
||
|
framework::run::<Example>("shadow");
|
||
|
}
|