WIP

Cargo.lock

@@ -366,6 +366,20 @@ dependencies = [
  "cfg_aliases",
 ]
 
+[[package]]
+name = "brush-next-nodes"
+version = "0.0.0"
+dependencies = [
+ "bytemuck",
+ "core-types",
+ "dyn-any",
+ "glam",
+ "node-macro",
+ "raster-types",
+ "wgpu",
+ "wgpu-executor",
+]
+
 [[package]]
 name = "brush-nodes"
 version = "0.1.0"
@@ -2113,6 +2127,7 @@ version = "0.1.0"
 dependencies = [
  "base64",
  "blending-nodes",
+ "brush-next-nodes",
  "brush-nodes",
  "bytemuck",
  "core-types",

Cargo.toml

@@ -80,6 +80,7 @@ vector-types = { path = "node-graph/libraries/vector-types" }
 graphic-types = { path = "node-graph/libraries/graphic-types" }
 rendering = { path = "node-graph/libraries/rendering" }
 brush-nodes = { path = "node-graph/nodes/brush" }
+brush-next-nodes = { path = "node-graph/nodes/brush-next" }
 blending-nodes = { path = "node-graph/nodes/blending" }
 graphene-core = { path = "node-graph/nodes/gcore" }
 graphic-nodes = { path = "node-graph/nodes/graphic" }

node-graph/nodes/brush-next/Cargo.toml

@@ -0,0 +1,23 @@
+[package]
+name = "brush-next-nodes"
+description = "Next-generation GPU brush rendering nodes for Graphene"
+version.workspace = true
+edition.workspace = true
+authors.workspace = true
+license.workspace = true
+
+[features]
+default = []
+
+[dependencies]
+# Local dependencies
+dyn-any = { workspace = true }
+core-types = { workspace = true }
+raster-types = { workspace = true, features = ["wgpu"] }
+wgpu-executor = { workspace = true }
+node-macro = { workspace = true }
+
+# Workspace dependencies
+glam = { workspace = true }
+wgpu = { workspace = true }
+bytemuck = { workspace = true }

node-graph/nodes/brush-next/src/lib.rs

@@ -0,0 +1 @@
+pub mod render_brush;

node-graph/nodes/brush-next/src/render_brush.rs

@@ -0,0 +1,230 @@
+use bytemuck::{Pod, Zeroable};
+use core_types::list::{Item, List};
+use core_types::{ATTR_TRANSFORM, Ctx, ProtoNodeIdentifier};
+use glam::{DAffine2, DVec2, UVec2, Vec2};
+use raster_types::{GPU, Raster};
+use std::marker::PhantomData;
+use std::sync::Arc;
+use wgpu::util::DeviceExt;
+use wgpu_executor::{AsyncWgpuPipeline, WgpuExecutor, WgpuPipelineCache};
+
+const WGPU_EXECUTOR_IDENTIFIER: ProtoNodeIdentifier = ProtoNodeIdentifier::new("graphene_std::platform_application_io::WgpuExecutorNode");
+
+#[node_macro::node(category("Raster: Brush"))]
+pub async fn render_brush_next<'a: 'n>(
+	_ctx: impl Ctx,
+	#[scope(brush_pipeline::IDENTIFIER)] pipeline: WgpuPipelineCache,
+	start: DVec2,
+	end: DVec2,
+	#[default(8)] stamps: u32,
+	#[default(16.)] radius: f64,
+) -> List<Raster<GPU>> {
+	let pad = DVec2::splat(radius);
+	let bbox_min = start.min(end) - pad;
+	let bbox_max = start.max(end) + pad;
+	let bbox_size = (bbox_max - bbox_min).max(DVec2::ONE);
+	let size = bbox_size.ceil().as_uvec2().max(UVec2::ONE);
+
+	let args = BrushPipelineArgs {
+		start: start - bbox_min,
+		end: end - bbox_min,
+		stamps,
+		radius: radius as f32,
+		size,
+		_marker: PhantomData,
+	};
+	let texture = pipeline.run::<BrushPipeline>(&args).await;
+	let raster = Raster::<GPU>::new_gpu(texture.as_ref().clone());
+	let transform = DAffine2::from_translation(bbox_min) * DAffine2::from_scale(DVec2::new(size.x as f64, size.y as f64));
+	List::new_from_item(Item::new_from_element(raster).with_attribute(ATTR_TRANSFORM, transform))
+}
+
+#[node_macro::node(category(""), inject_scope)]
+async fn brush_pipeline<'a: 'n>(_ctx: impl Ctx, #[scope(WGPU_EXECUTOR_IDENTIFIER)] executor: &'a WgpuExecutor, #[data] pipeline: WgpuPipelineCache) -> WgpuPipelineCache {
+	executor.pipeline_init::<BrushPipeline>(pipeline);
+	pipeline.clone()
+}
+
+#[repr(C)]
+#[derive(Copy, Clone, Debug, Pod, Zeroable)]
+struct Stamp {
+	center_px: [f32; 2],
+	radius_px: f32,
+	_pad: f32,
+}
+
+#[repr(C)]
+#[derive(Copy, Clone, Debug, Pod, Zeroable)]
+struct Globals {
+	viewport_size: [f32; 2],
+	_pad: [f32; 2],
+}
+
+pub struct BrushPipeline {
+	pipeline: wgpu::RenderPipeline,
+	bind_group_layout: wgpu::BindGroupLayout,
+}
+
+pub struct BrushPipelineArgs<'a> {
+	pub start: DVec2,
+	pub end: DVec2,
+	pub stamps: u32,
+	pub radius: f32,
+	pub size: UVec2,
+	pub _marker: PhantomData<&'a ()>,
+}
+
+impl AsyncWgpuPipeline for BrushPipeline {
+	type Args<'a> = BrushPipelineArgs<'a>;
+	type Out = Arc<wgpu::Texture>;
+
+	fn create(executor: &WgpuExecutor) -> Self {
+		let device = &executor.context().device;
+		let shader = device.create_shader_module(wgpu::include_wgsl!("render_brush.wgsl"));
+
+		let bind_group_layout = device.create_bind_group_layout(&wgpu::BindGroupLayoutDescriptor {
+			label: Some("brush_bind_group_layout"),
+			entries: &[wgpu::BindGroupLayoutEntry {
+				binding: 0,
+				visibility: wgpu::ShaderStages::VERTEX,
+				ty: wgpu::BindingType::Buffer {
+					ty: wgpu::BufferBindingType::Uniform,
+					has_dynamic_offset: false,
+					min_binding_size: None,
+				},
+				count: None,
+			}],
+		});
+
+		let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
+			label: Some("brush_pipeline_layout"),
+			bind_group_layouts: &[Some(&bind_group_layout)],
+			immediate_size: 0,
+		});
+
+		let instance_layout = wgpu::VertexBufferLayout {
+			array_stride: std::mem::size_of::<Stamp>() as wgpu::BufferAddress,
+			step_mode: wgpu::VertexStepMode::Instance,
+			attributes: &[
+				wgpu::VertexAttribute {
+					offset: 0,
+					shader_location: 0,
+					format: wgpu::VertexFormat::Float32x2,
+				},
+				wgpu::VertexAttribute {
+					offset: std::mem::size_of::<[f32; 2]>() as wgpu::BufferAddress,
+					shader_location: 1,
+					format: wgpu::VertexFormat::Float32,
+				},
+			],
+		};
+
+		let pipeline = device.create_render_pipeline(&wgpu::RenderPipelineDescriptor {
+			label: Some("brush_pipeline"),
+			layout: Some(&pipeline_layout),
+			vertex: wgpu::VertexState {
+				module: &shader,
+				entry_point: Some("vs_main"),
+				compilation_options: Default::default(),
+				buffers: &[instance_layout],
+			},
+			fragment: Some(wgpu::FragmentState {
+				module: &shader,
+				entry_point: Some("fs_main"),
+				compilation_options: Default::default(),
+				targets: &[Some(wgpu::ColorTargetState {
+					format: wgpu::TextureFormat::Rgba8Unorm,
+					blend: Some(wgpu::BlendState::PREMULTIPLIED_ALPHA_BLENDING),
+					write_mask: wgpu::ColorWrites::ALL,
+				})],
+			}),
+			primitive: wgpu::PrimitiveState {
+				topology: wgpu::PrimitiveTopology::TriangleList,
+				..Default::default()
+			},
+			depth_stencil: None,
+			multisample: wgpu::MultisampleState::default(),
+			multiview_mask: None,
+			cache: None,
+		});
+
+		Self { pipeline, bind_group_layout }
+	}
+
+	async fn run<'a>(&'a self, executor: &'a WgpuExecutor, args: &'a Self::Args<'_>) -> Self::Out {
+		let &BrushPipelineArgs { start, end, stamps, radius, size, .. } = args;
+
+		let size = size.max(UVec2::ONE);
+		let output = executor.request_texture(size).await;
+
+		let device = &executor.context().device;
+		let queue = &executor.context().queue;
+
+		let count = stamps.max(1);
+		let mut instances = Vec::with_capacity(count as usize);
+		for i in 0..count {
+			let t = if count == 1 { 0.0 } else { i as f64 / (count - 1) as f64 };
+			let center = start.lerp(end, t);
+			instances.push(Stamp {
+				center_px: [center.x as f32, center.y as f32],
+				radius_px: radius,
+				_pad: 0.0,
+			});
+		}
+
+		let instance_buffer = device.create_buffer_init(&wgpu::util::BufferInitDescriptor {
+			label: Some("brush_instance_buffer"),
+			contents: bytemuck::cast_slice(&instances),
+			usage: wgpu::BufferUsages::VERTEX,
+		});
+
+		let globals = Globals {
+			viewport_size: Vec2::new(size.x as f32, size.y as f32).to_array(),
+			_pad: [0.0; 2],
+		};
+
+		let globals_buffer = device.create_buffer_init(&wgpu::util::BufferInitDescriptor {
+			label: Some("brush_globals_buffer"),
+			contents: bytemuck::bytes_of(&globals),
+			usage: wgpu::BufferUsages::UNIFORM,
+		});
+
+		let bind_group = device.create_bind_group(&wgpu::BindGroupDescriptor {
+			label: Some("brush_bind_group"),
+			layout: &self.bind_group_layout,
+			entries: &[wgpu::BindGroupEntry {
+				binding: 0,
+				resource: globals_buffer.as_entire_binding(),
+			}],
+		});
+
+		let output_view = output.create_view(&wgpu::TextureViewDescriptor::default());
+
+		let mut encoder = device.create_command_encoder(&wgpu::CommandEncoderDescriptor { label: Some("brush_encoder") });
+
+		{
+			let mut pass = encoder.begin_render_pass(&wgpu::RenderPassDescriptor {
+				label: Some("brush_pass"),
+				color_attachments: &[Some(wgpu::RenderPassColorAttachment {
+					view: &output_view,
+					resolve_target: None,
+					ops: wgpu::Operations {
+						load: wgpu::LoadOp::Clear(wgpu::Color::TRANSPARENT),
+						store: wgpu::StoreOp::Store,
+					},
+					depth_slice: None,
+				})],
+				..Default::default()
+			});
+
+			pass.set_pipeline(&self.pipeline);
+			pass.set_bind_group(0, &bind_group, &[]);
+			pass.set_vertex_buffer(0, instance_buffer.slice(..));
+			pass.draw(0..3, 0..count);
+		}
+
+		queue.submit([encoder.finish()]);
+
+		output
+	}
+}

node-graph/nodes/brush-next/src/render_brush.wgsl

@@ -0,0 +1,50 @@
+struct Globals {
+	viewport_size: vec2<f32>,
+	_pad: vec2<f32>,
+};
+
+@group(0) @binding(0)
+var<uniform> globals: Globals;
+
+struct VertexOutput {
+	@builtin(position) clip_position: vec4<f32>,
+	@location(0) center_px: vec2<f32>,
+	@location(1) radius_px: f32,
+};
+
+@vertex
+fn vs_main(
+	@builtin(vertex_index) vertex_index: u32,
+	@location(0) center_px: vec2<f32>,
+	@location(1) radius_px: f32,
+) -> VertexOutput {
+	let viewport = globals.viewport_size;
+
+	let offset_x = f32(i32(vertex_index) - 1) * 2.0;
+	let offset_y = f32(i32(vertex_index & 1u) * 2 - 1) * 2.0;
+
+	let pos_px = center_px + vec2<f32>(offset_x, offset_y) * radius_px;
+
+	let clip = vec2<f32>(
+		2.0 * pos_px.x / viewport.x - 1.0,
+		1.0 - 2.0 * pos_px.y / viewport.y,
+	);
+
+	var out: VertexOutput;
+	out.clip_position = vec4<f32>(clip, 0.0, 1.0);
+	out.center_px = center_px;
+	out.radius_px = radius_px;
+	return out;
+}
+
+@fragment
+fn fs_main(in: VertexOutput) -> @location(0) vec4<f32> {
+	let frag_px = in.clip_position.xy;
+	let d = distance(frag_px, in.center_px);
+	if (d > in.radius_px) {
+		discard;
+	}
+	let t = d / in.radius_px;
+	let alpha = 1.0 - smoothstep(0.0, 1.0, t);
+	return vec4<f32>(0.0, 0.0, 0.0, alpha);
+}

node-graph/nodes/gstd/Cargo.toml

@@ -49,6 +49,7 @@ rendering = { workspace = true }
 graphene-application-io = { workspace = true }
 raster-nodes = { workspace = true }
 brush-nodes = { workspace = true }
+brush-next-nodes = { workspace = true }
 graphene-core = { workspace = true }
 graphic-nodes = { workspace = true }
 repeat-nodes = { workspace = true }

node-graph/nodes/gstd/src/lib.rs

@@ -6,6 +6,7 @@ pub mod render_node;
 pub mod render_pixel_preview;
 pub mod text;
 pub use blending_nodes;
+pub use brush_next_nodes as brush_next;
 pub use brush_nodes as brush;
 pub use core_types::*;
 pub use graphene_application_io as application_io;