Compare commits

...

27 Commits

Author SHA1 Message Date
f35d44c295 project ideal point 2025-12-19 15:40:13 -08:00
7c55d5cc31 wonky edge shader 2025-12-19 14:58:46 -08:00
ece883a865 freakin huge circle 2025-12-19 14:58:28 -08:00
4b771f826a edge shader 2025-12-19 14:51:38 -08:00
21bb84265f do less sin 2025-12-19 14:51:38 -08:00
db2f5ef3f3 plumb vert pos 2025-12-18 13:37:03 -08:00
d23e5c965a add needless complexity 2025-12-18 12:46:40 -08:00
c3995fc81e ez modulo 2025-12-18 12:15:47 -08:00
1bcd9f6eb1 0.5 world-space unit circle 2025-12-18 12:09:54 -08:00
5b78a68406 Revert "drop resolution for thick lines"
This reverts commit 0ddd41c1b6.
2025-12-18 12:01:21 -08:00
8aa9b5b14c fix mesh shader 2025-12-18 11:58:49 -08:00
5dc0d067fb 0.1 screen sizes 2025-12-18 11:32:32 -08:00
78e87192d6 disable culling 2025-12-18 11:04:00 -08:00
67a82aae9a draw a circle at the mesh center for now 2025-12-18 11:04:00 -08:00
6a65741117 draw_mesh_tasks 2025-12-18 11:04:00 -08:00
ee5d1e22c1 raise limits 2025-12-18 11:04:00 -08:00
bfb2bc18be fix shader 2025-12-18 11:04:00 -08:00
de8b9431e0 raise limits 2025-12-18 11:04:00 -08:00
31c6507de6 raise limits 2025-12-18 11:04:00 -08:00
bb729f997f stage visibility 2025-12-18 11:04:00 -08:00
8fa629aa12 disable edge pipeline 2025-12-18 11:04:00 -08:00
090a64ea11 raise limits 2025-12-18 11:04:00 -08:00
a2035fcaeb fix shader 2025-12-18 11:03:59 -08:00
323738b661 enable experimental feature 2025-12-18 11:03:59 -08:00
53ebbaeff6 write shader 2025-12-18 11:03:59 -08:00
ac28d27e9c wgpu bleeding edge 2025-12-18 11:03:59 -08:00
f15583f239 wip mesh shader circle 2025-12-18 11:03:30 -08:00
4 changed files with 290 additions and 37 deletions

View File

@@ -8,7 +8,17 @@ use wgpu::{util::DeviceExt,AstcBlock,AstcChannel};
use crate::model::{self as model_graphics,IndexedGraphicsMeshOwnedRenderConfig,IndexedGraphicsMeshOwnedRenderConfigId,GraphicsMeshOwnedRenderConfig,GraphicsModelColor4,GraphicsModelOwned,GraphicsVertex,DebugGraphicsVertex};
pub fn required_limits()->wgpu::Limits{
wgpu::Limits::default()
let mut limits=wgpu::Limits::default();
limits.max_task_invocations_per_dimension=1;
limits.max_task_invocations_per_workgroup=1;
limits.max_mesh_invocations_per_dimension=1;
limits.max_mesh_invocations_per_workgroup=1;
limits.max_task_mesh_workgroup_total_count=1;
limits.max_task_mesh_workgroups_per_dimension=1;
limits.max_task_payload_size=4;
limits.max_mesh_output_vertices=2*(3+2+4+8);
limits.max_mesh_output_primitives=2*(1+2+4+8)+2;
limits
}
struct Indices{
@@ -38,17 +48,21 @@ struct GraphicsModel{
struct DebugGraphicsSubmesh{
verts:Vec<strafesnet_physics::model::MeshVertId>,
edges:Vec<Indices>,
edges:Vec<[strafesnet_physics::model::MeshVertId;2]>,
faces:Vec<Indices>,
}
struct DebugGraphicsMesh{
verts:Indices,
vertices:Vec<DebugGraphicsVertex>,
submeshes:Vec<DebugGraphicsSubmesh>,
vertex_buf:wgpu::Buffer,
}
struct DebugGraphicsModel{
debug_mesh_id:u32,
bind_group:wgpu::BindGroup,
// 32 bytes used to tell the mesh shader where to draw
// Vert: [vec4,_]
// Edge: [vec4,vec4]
debug_buf:wgpu::Buffer,
}
struct GraphicsSamplers{
@@ -215,7 +229,7 @@ impl GraphicsState{
let submeshes=if let Ok(physics_mesh)=strafesnet_physics::model::PhysicsMesh::try_from(mesh){
physics_mesh.submesh_views().into_iter().map(|submesh_view|DebugGraphicsSubmesh{
verts:submesh_view.verts().to_owned(),
edges:submesh_view.edge_vert_ids_iter().map(|edge_verts|indices!(edge_verts)).collect(),
edges:submesh_view.edge_vert_ids_iter().collect(),
faces:submesh_view.face_vert_ids_iter().map(|face_verts|{
// triangulate
let mut indices=Vec::new();
@@ -235,7 +249,7 @@ impl GraphicsState{
};
DebugGraphicsMesh{
verts:indices!((0..vertices.len() as u32).map(strafesnet_physics::model::MeshVertId::new)),
vertices,
submeshes,
vertex_buf,
}
@@ -253,6 +267,11 @@ impl GraphicsState{
contents:bytemuck::cast_slice(&model_uniforms),
usage:wgpu::BufferUsages::UNIFORM|wgpu::BufferUsages::COPY_DST,
});
let debug_buf=device.create_buffer_init(&wgpu::util::BufferInitDescriptor{
label:Some(format!("Debug Model{} EV Buf",model_id).as_str()),
contents:bytemuck::cast_slice(&[0u8;32]),
usage:wgpu::BufferUsages::UNIFORM|wgpu::BufferUsages::COPY_DST,
});
let bind_group=device.create_bind_group(&wgpu::BindGroupDescriptor{
layout:&self.bind_group_layouts.debug_model,
entries:&[
@@ -260,12 +279,17 @@ impl GraphicsState{
binding:0,
resource:model_buf.as_entire_binding(),
},
wgpu::BindGroupEntry{
binding:1,
resource:debug_buf.as_entire_binding(),
},
],
label:Some(format!("Debug Model{} Bind Group",model_id).as_str()),
});
DebugGraphicsModel{
debug_mesh_id:model.mesh.get(),
bind_group,
debug_buf,
}
}).collect();
@@ -649,7 +673,7 @@ impl GraphicsState{
entries:&[
wgpu::BindGroupLayoutEntry{
binding:0,
visibility:wgpu::ShaderStages::VERTEX,
visibility:wgpu::ShaderStages::VERTEX|wgpu::ShaderStages::MESH,
ty:wgpu::BindingType::Buffer{
ty:wgpu::BufferBindingType::Uniform,
has_dynamic_offset:false,
@@ -716,7 +740,17 @@ impl GraphicsState{
entries:&[
wgpu::BindGroupLayoutEntry{
binding:0,
visibility:wgpu::ShaderStages::VERTEX_FRAGMENT,
visibility:wgpu::ShaderStages::VERTEX_FRAGMENT|wgpu::ShaderStages::MESH,
ty:wgpu::BindingType::Buffer{
ty:wgpu::BufferBindingType::Uniform,
has_dynamic_offset:false,
min_binding_size:None,
},
count:None,
},
wgpu::BindGroupLayoutEntry{
binding:1,
visibility:wgpu::ShaderStages::MESH,
ty:wgpu::BindingType::Buffer{
ty:wgpu::BufferBindingType::Uniform,
has_dynamic_offset:false,
@@ -880,7 +914,7 @@ impl GraphicsState{
&camera_bind_group_layout,
&debug_model_bind_group_layout,
],
push_constant_ranges:&[],
immediate_size:0,
});
let sky_pipeline_layout=device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor{
label:None,
@@ -957,12 +991,12 @@ impl GraphicsState{
multiview_mask:None,
cache:None,
});
let mut debug_model_pipeline=wgpu::RenderPipelineDescriptor{
label:None,
let debug_model_pipeline_face=device.create_render_pipeline(&wgpu::RenderPipelineDescriptor{
label:Some("Debug Face Pipeline"),
layout:Some(&debug_model_pipeline_layout),
vertex:wgpu::VertexState{
module:&shader,
entry_point:Some("vs_debug"),
entry_point:Some("vs_debug_face"),
buffers:&[wgpu::VertexBufferLayout{
array_stride:size_of::<DebugGraphicsVertex>() as wgpu::BufferAddress,
step_mode:wgpu::VertexStepMode::Vertex,
@@ -981,6 +1015,7 @@ impl GraphicsState{
compilation_options:wgpu::PipelineCompilationOptions::default(),
}),
primitive:wgpu::PrimitiveState{
topology:wgpu::PrimitiveTopology::TriangleList,
front_face:wgpu::FrontFace::Cw,
cull_mode:Some(wgpu::Face::Front),
..Default::default()
@@ -993,18 +1028,55 @@ impl GraphicsState{
bias:wgpu::DepthBiasState::default(),
}),
multisample:wgpu::MultisampleState::default(),
multiview_mask:None,
cache:None,
});
let mut debug_model_pipeline=wgpu::MeshPipelineDescriptor{
label:None,//filled in below
layout:Some(&debug_model_pipeline_layout),
task:Some(wgpu::TaskState{
module:&shader,
entry_point:Some("ts_main"),
compilation_options:wgpu::PipelineCompilationOptions::default(),
}),
mesh:wgpu::MeshState{
module:&shader,
entry_point:None,//filled in below
compilation_options:wgpu::PipelineCompilationOptions::default(),
},
fragment:Some(wgpu::FragmentState{
module:&shader,
entry_point:Some("fs_debug"),
targets:&[Some(wgpu::ColorTargetState{
format:config.view_formats[0],
blend:Some(wgpu::BlendState::ALPHA_BLENDING),
write_mask:wgpu::ColorWrites::default(),
})],
compilation_options:wgpu::PipelineCompilationOptions::default(),
}),
primitive:wgpu::PrimitiveState{
topology:wgpu::PrimitiveTopology::TriangleList,
front_face:wgpu::FrontFace::Cw,
cull_mode:None,
..Default::default()
},
depth_stencil:Some(wgpu::DepthStencilState{
format:Self::DEPTH_FORMAT,
depth_write_enabled:true,
depth_compare:wgpu::CompareFunction::Always,
stencil:wgpu::StencilState::default(),
bias:wgpu::DepthBiasState::default(),
}),
multisample:wgpu::MultisampleState::default(),
multiview:None,
cache:None,
};
debug_model_pipeline.label=Some("Debug Face Pipeline");
debug_model_pipeline.primitive.topology=wgpu::PrimitiveTopology::TriangleList;
let debug_model_pipeline_face=device.create_render_pipeline(&debug_model_pipeline);
debug_model_pipeline.label=Some("Debug Edge Pipeline");
debug_model_pipeline.primitive.topology=wgpu::PrimitiveTopology::LineList;
let debug_model_pipeline_edge=device.create_render_pipeline(&debug_model_pipeline);
debug_model_pipeline.label=Some("Debug Vert Pipeline");
debug_model_pipeline.primitive.topology=wgpu::PrimitiveTopology::PointList;
let debug_model_pipeline_vert=device.create_render_pipeline(&debug_model_pipeline);
debug_model_pipeline.mesh.entry_point=Some("ms_debug_vert");
let debug_model_pipeline_vert=device.create_mesh_pipeline(&debug_model_pipeline);
debug_model_pipeline.label=Some("Debug Edge Pipeline");
debug_model_pipeline.mesh.entry_point=Some("ms_debug_edge");
let debug_model_pipeline_edge=device.create_mesh_pipeline(&debug_model_pipeline);
let camera=GraphicsCamera::default();
let camera_uniforms=camera.to_uniform_data(glam::Vec3::ZERO,glam::Vec2::ZERO);
@@ -1117,6 +1189,49 @@ impl GraphicsState{
.copy_from_slice(bytemuck::cast_slice(&model_uniforms));
}
*/
// upload the edge or vertex for teh mesh shader to highlight
if let Some(hit)=&frame_state.hit{
if let Some(closest_fev)=&hit.closest_fev{
let model_id:model::ModelId=hit.convex_mesh_id.model_id.into();
if let Some(model)=self.debug_models.get(model_id.get() as usize){
let mesh=&self.debug_meshes[model.debug_mesh_id as usize];
match closest_fev{
strafesnet_physics::model::FEV::Face(_face)=>{
// face is rendered normally
},
strafesnet_physics::model::FEV::Edge(edge)=>{
let [v0_id,v1_id]=mesh.submeshes[hit.convex_mesh_id.submesh_id.get() as usize].edges[edge.get() as usize];
let v0_pos=mesh.vertices[v0_id.get() as usize].pos;
let v1_pos=mesh.vertices[v1_id.get() as usize].pos;
let debug_data=[glam::Vec3A::from_array(v0_pos).extend(1.0).to_array(),glam::Vec3A::from_array(v1_pos).extend(1.0).to_array()];
let debug_slice=bytemuck::cast_slice(&debug_data);
self.staging_belt
.write_buffer(
&mut encoder,
&model.debug_buf,
0,
wgpu::BufferSize::new(debug_slice.len() as wgpu::BufferAddress).unwrap(),
)
.copy_from_slice(debug_slice);
},
strafesnet_physics::model::FEV::Vert(vert)=>{
let vert_id=mesh.submeshes[hit.convex_mesh_id.submesh_id.get() as usize].verts[vert.get() as usize].get();
let pos=mesh.vertices[vert_id as usize].pos;
let debug_data=[glam::Vec3A::from_array(pos).extend(1.0).to_array()];
let debug_slice=bytemuck::cast_slice(&debug_data);
self.staging_belt
.write_buffer(
&mut encoder,
&model.debug_buf,
0,
wgpu::BufferSize::new(debug_slice.len() as wgpu::BufferAddress).unwrap(),
)
.copy_from_slice(debug_slice);
},
}
}
}
}
self.staging_belt.finish();
{
@@ -1166,7 +1281,7 @@ impl GraphicsState{
rpass.draw(0..3,0..1);
// render a single debug_model in red
if let Some(hit)=frame_state.hit{
if let Some(hit)=&frame_state.hit{
if let Some(closest_fev)=&hit.closest_fev{
let model_id:model::ModelId=hit.convex_mesh_id.model_id.into();
if let Some(model)=self.debug_models.get(model_id.get() as usize){
@@ -1181,18 +1296,15 @@ impl GraphicsState{
//TODO: loop over triangle strips
rpass.draw_indexed(0..indices.count,0,0..1);
},
strafesnet_physics::model::FEV::Edge(edge)=>{
strafesnet_physics::model::FEV::Edge(_edge)=>{
rpass.set_pipeline(&self.pipelines.debug_edge);
let indices=&mesh.submeshes[hit.convex_mesh_id.submesh_id.get() as usize].edges[edge.get() as usize];
rpass.set_index_buffer(indices.buf.slice(..),indices.format);
rpass.draw_indexed(0..indices.count,0,0..1);
// the data has already been primed by the staging belt
rpass.draw_mesh_tasks(1, 1, 1);
},
strafesnet_physics::model::FEV::Vert(vert)=>{
strafesnet_physics::model::FEV::Vert(_vert)=>{
rpass.set_pipeline(&self.pipelines.debug_vert);
let indices=&mesh.verts;
rpass.set_index_buffer(indices.buf.slice(..),indices.format);
let vert_id=mesh.submeshes[hit.convex_mesh_id.submesh_id.get() as usize].verts[vert.get() as usize].get();
rpass.draw_indexed(0..1,vert_id as i32,0..1);
// the data has already been primed by the staging belt
rpass.draw_mesh_tasks(1, 1, 1);
},
}
}

View File

@@ -34,8 +34,8 @@ pub fn new(
Instruction::Resize(size,user_settings)=>{
println!("Resizing to {:?}",size);
let t0=std::time::Instant::now();
config.width=640;
config.height=480;
config.width=size.width.max(1);
config.height=size.height.max(1);
surface.configure(&device,&config);
graphics.resize(&device,&config,&user_settings);
println!("Resize took {:?}",t0.elapsed());

View File

@@ -3,7 +3,7 @@ fn optional_features()->wgpu::Features{
|wgpu::Features::TEXTURE_COMPRESSION_ETC2
}
fn required_features()->wgpu::Features{
wgpu::Features::TEXTURE_COMPRESSION_BC
wgpu::Features::TEXTURE_COMPRESSION_BC|wgpu::Features::EXPERIMENTAL_MESH_SHADER
}
fn required_downlevel_capabilities()->wgpu::DownlevelCapabilities{
wgpu::DownlevelCapabilities{
@@ -125,7 +125,7 @@ impl<'a> SetupContextPartial3<'a>{
required_limits:needed_limits,
memory_hints:wgpu::MemoryHints::Performance,
trace:wgpu::Trace::Off,
experimental_features:wgpu::ExperimentalFeatures::disabled(),
experimental_features:unsafe{wgpu::ExperimentalFeatures::enabled()},
},
))
.expect("Unable to find a suitable GPU adapter!");

View File

@@ -1,3 +1,5 @@
enable wgpu_mesh_shader;
struct Camera {
// from camera to screen
proj: mat4x4<f32>,
@@ -90,23 +92,162 @@ fn vs_entity_texture(
@binding(0)
var<uniform> model_instance: ModelInstance;
@group(1)
@binding(1)
var<uniform> ve_verts: array<vec4<f32>, 2>;
struct DebugEntityOutput {
@builtin(position) position: vec4<f32>,
@location(1) normal: vec3<f32>,
@location(2) view: vec3<f32>,
};
@vertex
fn vs_debug(
fn vs_debug_face(
@location(0) pos: vec3<f32>,
) -> DebugEntityOutput {
var position: vec4<f32> = model_instance.transform * vec4<f32>(pos, 1.0);
var result: DebugEntityOutput;
result.view = position.xyz - camera.view_inv[3].xyz;//col(3)
result.position = camera.proj * camera.view * position;
return result;
}
struct TaskPayload {
four_bytes: u32,
}
var<task_payload> taskPayload: TaskPayload;
@task
@payload(taskPayload)
@workgroup_size(1)
fn ts_main() -> @builtin(mesh_task_size) vec3<u32> {
taskPayload.four_bytes = 0;
return vec3(1, 1, 1);
}
struct VertexOutput {
@builtin(position) position: vec4<f32>,
@location(0) color: vec4<f32>,
}
struct PrimitiveOutput {
@builtin(triangle_indices) indices: vec3<u32>,
}
struct CircleOutput {
@builtin(vertices) vertices: array<VertexOutput, 24>,
@builtin(primitives) primitives: array<PrimitiveOutput, 22>,
@builtin(vertex_count) vertex_count: u32,
@builtin(primitive_count) primitive_count: u32,
}
var<workgroup> mesh_output: CircleOutput;
const tau: f32 = 3.141592653589793 * 2.0;
fn modulo(value:u32,modulus:u32)->u32{
return value-value/modulus*modulus;
}
@mesh(mesh_output)
@payload(taskPayload)
@workgroup_size(1)
fn ms_debug_vert(){
// circle with 24 vertices.
const LAYERS:u32 = 3;
const N:u32 = 3*(1<<LAYERS);
mesh_output.vertex_count = N;
mesh_output.primitive_count = N-2;
var vertex_world_position: vec4<f32> = model_instance.transform * ve_verts[0];
var vertex_screen_position: vec4<f32> = camera.proj * camera.view * vertex_world_position;
for (var i:u32 = 0; i<N/4; i++){
// draw a 1 unit redius circle
var theta: f32 = f32(i) * tau / f32(N);
var cos_sin: vec2<f32> = vec2(cos(theta), sin(theta));
var offset: vec2<f32> = 0.5 * cos_sin;
mesh_output.vertices[i].position = vertex_screen_position + vec4<f32>(offset, 0.0, 0.0);
mesh_output.vertices[i+N/4].position = vertex_screen_position + vec4<f32>(-offset.y, offset.x, 0.0, 0.0);
mesh_output.vertices[i+N/4*2].position = vertex_screen_position + vec4<f32>(-offset, 0.0, 0.0);
mesh_output.vertices[i+N/4*3].position = vertex_screen_position + vec4<f32>(offset.y, -offset.x, 0.0, 0.0);
}
// max area triangle indices
// the big triangle
mesh_output.primitives[0].indices = vec3<u32>(0, N/3, N/3*2);
// 3 layers of infill triangles to approximate circle better than 1 triangle.
// we start on the outer layer because it's easier to construct this way
var count:u32=N;
var base:u32=1;
for (var layer:u32 = 0; layer<LAYERS; layer++){
count=count>>1;
var step:u32=N/count;
for (var i:u32 = 0; i<count; i++){
mesh_output.primitives[base+i].indices = vec3<u32>(i*step, i*step+(step>>1), modulo(i*step+step,N));
}
base+=count;
}
}
@mesh(mesh_output)
@payload(taskPayload)
@workgroup_size(1)
fn ms_debug_edge(){
// draw two circles for now.
const LAYERS:u32 = 3;
const N:u32 = 2*(1<<LAYERS);
mesh_output.vertex_count = 2*(3+2+4+8);
mesh_output.primitive_count = 2*(1+2+4+8)+2;
var v0_world_position: vec4<f32> = model_instance.transform * ve_verts[0];
var v1_world_position: vec4<f32> = model_instance.transform * ve_verts[1];
var v0_screen_position: vec4<f32> = camera.proj * camera.view * v0_world_position;
var v1_screen_position: vec4<f32> = camera.proj * camera.view * v1_world_position;
var edge_dir_world: vec4<f32> = normalize(v0_world_position - v1_world_position);
var edge_dir_screen: vec4<f32> = camera.proj * camera.view * edge_dir_world;
for (var i:u32 = 0; i<=N/2; i++){
// two half circles that make a whole
var theta: f32 = f32(i) * tau / f32(N);
var cos_sin: vec2<f32> = vec2(cos(theta), sin(theta));
// construct basis vectors
var y_axis: vec2<f32> = edge_dir_screen.xy;
var x_axis: vec2<f32> = y_axis.yx;
x_axis.x = -x_axis.x;
var offset: vec4<f32> = vec4<f32>(0.5 * (x_axis * cos_sin.x + y_axis * cos_sin.y), 0.0, 0.0);;
mesh_output.vertices[i].position = v0_screen_position + offset;
mesh_output.vertices[N/2+1+i].position = v1_screen_position - offset;
}
// max area triangle indices
// number of primitives per circle half
const P:u32 = N/2;
// the big triangles between the circles
mesh_output.primitives[0].indices = vec3<u32>(0, N/2+1, P);
mesh_output.primitives[P].indices = vec3<u32>(N/2+1, 0, P + N/2+1);
// 3 layers of infill triangles to approximate circle better than 1 triangle.
// we start on the outer layer because it's easier to construct this way
var count:u32=P;
var base:u32=1;
for (var layer:u32 = 0; layer<LAYERS; layer++){
count=count>>1;
var step:u32=P/count;
for (var i:u32 = 0; i<count; i++){
var indices = vec3<u32>(i*step, i*step+(step>>1), i*step+step);
mesh_output.primitives[base+i].indices = indices;
mesh_output.primitives[P+base+i].indices = indices + N/2+1;
}
base+=count;
}
}
//group 2 is the skybox texture
@group(1)
@binding(0)