@codexo/exojs-particles 0.15.2 → 0.16.0

package/dist/esm/gpu/ParticleGpuState.js

@@ -1,417 +1,569 @@
-import { wgslUniformByteSize } from '../modules/WgslContribution.js';
+import { getWgslUniformByteSize } from "../modules/WgslContribution.js";
+import particle_simulate_default from "./shaders/particle-simulate.wgsl.js";
+import { WebGpuComputePipeline, WebGpuStorageBuffer, WebGpuUniformBuffer, fillShaderSource, reflectComputeBindings } from "@codexo/exojs/renderer-sdk";
 
-/// <reference types="@webgpu/types" />
+//#region src/gpu/ParticleGpuState.ts
 /**
- * GPU-side mirror of one {@link ParticleSystem}. Owns:
- *
- * - **8 packed storage buffers** for the per-particle SoA data:
- *   positions/velocities/scales/rotInfo/timing as `vec2<f32>`, color and
- *   textureIndex as `u32`, plus the instance output buffer. Sits at the
- *   default WebGPU `maxStorageBuffersPerShaderStage = 8` limit.
- * - **One uniform buffer** for sim state (`dt`, `liveCount`).
- * - **One uniform buffer** for module configs (concatenated per-module
- *   structs with WGSL std140-ish alignment).
- * - **One uniform buffer** for frame UVs — `array<vec4<f32>, N>` where N
- *   is the system's frame count (or 1 when no atlas is declared). Each
- *   vec4 is `(uvMinX, uvMinY, uvMaxX, uvMaxY)` already flipY-adjusted.
- * - **N 1D textures** for modules that use lookup tables (Curve / ColorGradient).
- * - **Composite compute pipeline** built once at construction by
- *   concatenating the integration step + every registered module body +
- *   the pack-instances step into a single shader.
- *
- * The compute shader's pack-instances step reads `textureIndex[i]`, looks
- * up the matching frame UV, and writes a 40-byte interleaved record into
- * the instance output buffer (`STORAGE | VERTEX`). The renderer binds that
- * buffer directly as instanced vertex source — no readback.
- */
+* GPU-side mirror of one {@link ParticleSystem}. Owns:
+*
+* - **8 packed storage buffers** for the per-particle SoA data:
+*   positions/velocities/scales/rotInfo/timing as `vec2<f32>`, color and
+*   textureIndex as `u32`, plus the instance output buffer. Sits at the
+*   default WebGPU `maxStorageBuffersPerShaderStage = 8` limit. Built on
+*   the shared {@link WebGpuStorageBuffer} SDK primitive.
+* - **Three uniform buffers** (sim state `dt`/`liveCount`, module configs -
+*   concatenated per-module structs with WGSL std140-ish alignment - and
+*   frame UVs, `array<vec4<f32>, N>` where N is the system's frame count or
+*   1 when no atlas is declared, each vec4 `(uvMinX, uvMinY, uvMaxX, uvMaxY)`
+*   already flipY-adjusted), built on the shared {@link WebGpuUniformBuffer}
+*   SDK primitive.
+* - **N 1D textures** for modules that use lookup tables (Curve / ColorGradient).
+* - **Composite compute pipeline** built once at construction by
+*   concatenating the integration step + every registered module body +
+*   the pack-instances step into a single shader, via the shared
+*   {@link WebGpuComputePipeline} SDK primitive (two bind groups: group 0
+*   holds uniforms + module lookup textures/samplers, group 1 holds the
+*   8 SoA storage buffers). The bind-group *layouts* are derived straight
+*   from the shader's own `@group`/`@binding` declarations via
+*   {@link reflectComputeBindings} - no hand-written binding list kept in
+*   sync with the WGSL text by hand.
+*
+* The compute shader's pack-instances step reads `textureIndex[i]`, looks
+* up the matching frame UV, and writes a 40-byte interleaved record into
+* the instance output buffer (`STORAGE | VERTEX`). The renderer binds that
+* buffer directly as instanced vertex source - no readback.
+*/
 const workgroupSize = 64;
-const instanceBytes = 40; // 5 × f32 + 1 × u32 + 4 × f32 (uvMin.xy, uvMax.xy)
-class ParticleGpuState {
-    device;
-    capacity;
-    /** GPU buffer holding interleaved per-instance vertex data, written by compute, read as VERTEX by the renderer. */
-    instanceBuffer;
-    _positions;
-    _velocities;
-    _scales;
-    _rotInfo;
-    _timing;
-    _color;
-    _textureIndex;
-    _simUniformBuffer;
-    _simUniformData = new ArrayBuffer(16);
-    _simUniformView;
-    _moduleUniformBuffer;
-    _moduleUniformData;
-    _moduleUniformView;
-    _moduleSlots;
-    _framesUniformBuffer;
-    _framesUniformData;
-    _framesUniformView;
-    _frameCount;
-    _moduleTextures = new Map();
-    _samplerFiltering;
-    _samplerNonFiltering;
-    _pipeline;
-    _bindGroup0;
-    _bindGroup1;
-    constructor(device, capacity, modules, frames, texture) {
-        this.device = device;
-        this.capacity = capacity;
-        for (const m of modules) {
-            if (!m.wgsl) {
-                throw new Error(`ParticleGpuState: module ${m.constructor.name} has no wgsl() — ` + 'all registered UpdateModules must be GPU-eligible.');
-            }
-        }
-        // Module uniform layout.
-        const slots = [];
-        let uniformOffset = 0;
-        for (const m of modules) {
-            const c = m.wgsl();
-            const fields = c.uniforms ?? [];
-            const size = wgslUniformByteSize(fields);
-            uniformOffset = Math.ceil(uniformOffset / 16) * 16;
-            slots.push({
-                module: m,
-                contribution: c,
-                uniformByteOffset: uniformOffset,
-                uniformByteSize: size,
-            });
-            uniformOffset += size;
-        }
-        const totalUniformBytes = Math.max(16, Math.ceil(uniformOffset / 16) * 16);
-        this._moduleSlots = slots;
-        if (uniformOffset > 0) {
-            this._moduleUniformData = new ArrayBuffer(totalUniformBytes);
-            this._moduleUniformView = new DataView(this._moduleUniformData);
-            this._moduleUniformBuffer = device.createBuffer({
-                label: 'particle-module-uniforms',
-                size: totalUniformBytes,
-                usage: GPUBufferUsage.UNIFORM | GPUBufferUsage.COPY_DST,
-            });
-        }
-        else {
-            this._moduleUniformData = null;
-            this._moduleUniformView = null;
-            this._moduleUniformBuffer = null;
-        }
-        // Frames uniform buffer.
-        this._frameCount = Math.max(1, frames.length);
-        this._framesUniformData = new ArrayBuffer(this._frameCount * 16);
-        this._framesUniformView = new Float32Array(this._framesUniformData);
-        this._framesUniformBuffer = device.createBuffer({
-            label: 'particle-frames-uniforms',
-            size: this._framesUniformData.byteLength,
-            usage: GPUBufferUsage.UNIFORM | GPUBufferUsage.COPY_DST,
-        });
-        this._writeFrames(frames, texture);
-        const vec2Bytes = capacity * 8;
-        const u32Bytes = capacity * 4;
-        this._positions = device.createBuffer({
-            label: 'particle-positions',
-            size: vec2Bytes,
-            usage: GPUBufferUsage.STORAGE | GPUBufferUsage.COPY_DST,
-        });
-        this._velocities = device.createBuffer({
-            label: 'particle-velocities',
-            size: vec2Bytes,
-            usage: GPUBufferUsage.STORAGE | GPUBufferUsage.COPY_DST,
-        });
-        this._scales = device.createBuffer({
-            label: 'particle-scales',
-            size: vec2Bytes,
-            usage: GPUBufferUsage.STORAGE | GPUBufferUsage.COPY_DST,
-        });
-        this._rotInfo = device.createBuffer({
-            label: 'particle-rotInfo',
-            size: vec2Bytes,
-            usage: GPUBufferUsage.STORAGE | GPUBufferUsage.COPY_DST,
-        });
-        this._timing = device.createBuffer({
-            label: 'particle-timing',
-            size: vec2Bytes,
-            usage: GPUBufferUsage.STORAGE | GPUBufferUsage.COPY_DST,
-        });
-        this._color = device.createBuffer({
-            label: 'particle-color',
-            size: u32Bytes,
-            usage: GPUBufferUsage.STORAGE | GPUBufferUsage.COPY_DST,
-        });
-        this._textureIndex = device.createBuffer({
-            label: 'particle-textureIndex',
-            size: u32Bytes,
-            usage: GPUBufferUsage.STORAGE | GPUBufferUsage.COPY_DST,
-        });
-        this.instanceBuffer = device.createBuffer({
-            label: 'particle-instance-output',
-            size: capacity * instanceBytes,
-            usage: GPUBufferUsage.STORAGE | GPUBufferUsage.VERTEX | GPUBufferUsage.COPY_DST,
-        });
-        this._simUniformView = new DataView(this._simUniformData);
-        this._simUniformBuffer = device.createBuffer({
-            label: 'particle-sim-uniforms',
-            size: 16,
-            usage: GPUBufferUsage.UNIFORM | GPUBufferUsage.COPY_DST,
-        });
-        // r32float textures aren't filterable in core WebGPU (would require
-        // the optional `float32-filterable` feature). Use `nearest` for
-        // r32float curve LUTs (256 taps is fine without interpolation) and
-        // `linear` for rgba8unorm gradients which support filtering natively.
-        this._samplerFiltering = device.createSampler({
-            label: 'particle-lookup-sampler-filtering',
-            minFilter: 'linear',
-            magFilter: 'linear',
-            addressModeU: 'clamp-to-edge',
-        });
-        this._samplerNonFiltering = device.createSampler({
-            label: 'particle-lookup-sampler-non-filtering',
-            minFilter: 'nearest',
-            magFilter: 'nearest',
-            addressModeU: 'clamp-to-edge',
-        });
-        // Allocate textures for modules that need them.
-        for (const slot of slots) {
-            const c = slot.contribution;
-            if (!c.textures)
-                continue;
-            for (const t of c.textures) {
-                const tex = device.createTexture({
-                    label: `particle-tex-${c.key}-${t.name}`,
-                    size: { width: 256, height: 1, depthOrArrayLayers: 1 },
-                    format: t.format,
-                    usage: GPUTextureUsage.TEXTURE_BINDING | GPUTextureUsage.COPY_DST,
-                    dimension: '1d',
-                });
-                this._moduleTextures.set(`${c.key}_${t.name}`, tex);
-            }
-        }
-        const wgsl = this._buildShader(slots);
-        const bindGroup0Layout = this._buildBindGroup0Layout(slots);
-        const bindGroup1Layout = device.createBindGroupLayout({
-            label: 'particle-soa-bgl',
-            entries: [
-                { binding: 0, visibility: GPUShaderStage.COMPUTE, buffer: { type: 'storage' } },
-                { binding: 1, visibility: GPUShaderStage.COMPUTE, buffer: { type: 'storage' } },
-                { binding: 2, visibility: GPUShaderStage.COMPUTE, buffer: { type: 'storage' } },
-                { binding: 3, visibility: GPUShaderStage.COMPUTE, buffer: { type: 'storage' } },
-                { binding: 4, visibility: GPUShaderStage.COMPUTE, buffer: { type: 'storage' } },
-                { binding: 5, visibility: GPUShaderStage.COMPUTE, buffer: { type: 'storage' } },
-                { binding: 6, visibility: GPUShaderStage.COMPUTE, buffer: { type: 'read-only-storage' } }, // textureIndex (matches WGSL `var<storage, read>`)
-                { binding: 7, visibility: GPUShaderStage.COMPUTE, buffer: { type: 'storage' } },
-            ],
-        });
-        const pipelineLayout = device.createPipelineLayout({
-            label: 'particle-compute-layout',
-            bindGroupLayouts: [bindGroup0Layout, bindGroup1Layout],
-        });
-        const shaderModule = device.createShaderModule({
-            label: 'particle-compute-shader',
-            code: wgsl,
-        });
-        this._pipeline = device.createComputePipeline({
-            label: 'particle-compute-pipeline',
-            layout: pipelineLayout,
-            compute: {
-                module: shaderModule,
-                entryPoint: 'main',
-            },
-        });
-        this._bindGroup0 = this._buildBindGroup0(bindGroup0Layout, slots);
-        this._bindGroup1 = device.createBindGroup({
-            label: 'particle-soa-bg',
-            layout: bindGroup1Layout,
-            entries: [
-                { binding: 0, resource: { buffer: this._positions } },
-                { binding: 1, resource: { buffer: this._velocities } },
-                { binding: 2, resource: { buffer: this._scales } },
-                { binding: 3, resource: { buffer: this._rotInfo } },
-                { binding: 4, resource: { buffer: this._timing } },
-                { binding: 5, resource: { buffer: this._color } },
-                { binding: 6, resource: { buffer: this._textureIndex } },
-                { binding: 7, resource: { buffer: this.instanceBuffer } },
-            ],
-        });
-        // Modules upload their lookup textures.
-        for (const slot of slots) {
-            if (!slot.module.uploadTextures)
-                continue;
-            const moduleTextures = new Map();
-            for (const t of slot.contribution.textures ?? []) {
-                const tex = this._moduleTextures.get(`${slot.contribution.key}_${t.name}`);
-                if (tex !== undefined) {
-                    moduleTextures.set(t.name, tex);
-                }
-            }
-            slot.module.uploadTextures(device, moduleTextures);
-        }
-    }
-    dispatch(system, dt) {
-        const liveCount = system.liveCount;
-        if (liveCount <= 0) {
-            return;
-        }
-        this._writeSimUniforms(dt, liveCount);
-        this._writeModuleUniforms(dt);
-        const encoder = this.device.createCommandEncoder({ label: 'particle-compute' });
-        const pass = encoder.beginComputePass({ label: 'particle-compute-pass' });
-        pass.setPipeline(this._pipeline);
-        pass.setBindGroup(0, this._bindGroup0);
-        pass.setBindGroup(1, this._bindGroup1);
-        pass.dispatchWorkgroups(Math.ceil(liveCount / workgroupSize));
-        pass.end();
-        this.device.queue.submit([encoder.finish()]);
-    }
-    destroy() {
-        this._positions.destroy();
-        this._velocities.destroy();
-        this._scales.destroy();
-        this._rotInfo.destroy();
-        this._timing.destroy();
-        this._color.destroy();
-        this._textureIndex.destroy();
-        this.instanceBuffer.destroy();
-        this._simUniformBuffer.destroy();
-        this._framesUniformBuffer.destroy();
-        this._moduleUniformBuffer?.destroy();
-        for (const tex of this._moduleTextures.values()) {
-            tex.destroy();
-        }
-        this._moduleTextures.clear();
-    }
-    _writeFrames(frames, texture) {
-        const view = this._framesUniformView;
-        const w = texture.width;
-        const h = texture.height;
-        const flipY = texture.flipY;
-        if (frames.length === 0) {
-            // Single-frame fallback — full texture.
-            view[0] = 0;
-            view[1] = flipY ? 1 : 0;
-            view[2] = 1;
-            view[3] = flipY ? 0 : 1;
-        }
-        else {
-            for (let i = 0; i < frames.length; i++) {
-                const f = frames[i];
-                const o = i * 4;
-                const minU = f.left / w;
-                const maxU = f.right / w;
-                const topV = f.top / h;
-                const bottomV = f.bottom / h;
-                view[o + 0] = minU;
-                view[o + 1] = flipY ? bottomV : topV;
-                view[o + 2] = maxU;
-                view[o + 3] = flipY ? topV : bottomV;
-            }
-        }
-        this.device.queue.writeBuffer(this._framesUniformBuffer, 0, this._framesUniformData);
-    }
-    /**
-     * Push the listed CPU SoA slots to the GPU. Called by `ParticleSystem`
-     * with newly-spawned slots and just-expired slots (lifetime sentinel).
-     * Slots not in the dirty set are left alone — GPU keeps the integrated
-     * state from previous compute dispatches.
-     *
-     * Each dirty slot triggers 7 small `queue.writeBuffer` calls (one per
-     * SoA channel). For typical spawn rates (≤200/s) this is negligible
-     * (≤1400 calls/s); contiguous-range batching is a future optimisation.
-     */
-    uploadDirty(system, slots) {
-        const queue = this.device.queue;
-        const scratch2 = this._dirtyScratchVec2;
-        const scratch1 = this._dirtyScratchU32;
-        for (const slot of slots) {
-            const byteOffset2 = slot * 8;
-            const byteOffset1 = slot * 4;
-            scratch2[0] = system.posX[slot];
-            scratch2[1] = system.posY[slot];
-            queue.writeBuffer(this._positions, byteOffset2, scratch2.buffer, 0, 8);
-            scratch2[0] = system.velX[slot];
-            scratch2[1] = system.velY[slot];
-            queue.writeBuffer(this._velocities, byteOffset2, scratch2.buffer, 0, 8);
-            scratch2[0] = system.scaleX[slot];
-            scratch2[1] = system.scaleY[slot];
-            queue.writeBuffer(this._scales, byteOffset2, scratch2.buffer, 0, 8);
-            scratch2[0] = system.rotations[slot];
-            scratch2[1] = system.rotationSpeeds[slot];
-            queue.writeBuffer(this._rotInfo, byteOffset2, scratch2.buffer, 0, 8);
-            scratch2[0] = system.elapsed[slot];
-            scratch2[1] = system.lifetime[slot];
-            queue.writeBuffer(this._timing, byteOffset2, scratch2.buffer, 0, 8);
-            scratch1[0] = system.color[slot];
-            queue.writeBuffer(this._color, byteOffset1, scratch1.buffer, 0, 4);
-            scratch1[0] = system.textureIndex[slot];
-            queue.writeBuffer(this._textureIndex, byteOffset1, scratch1.buffer, 0, 4);
-        }
-    }
-    _dirtyScratchVec2 = new Float32Array(2);
-    _dirtyScratchU32 = new Uint32Array(1);
-    _writeSimUniforms(dt, liveCount) {
-        this._simUniformView.setFloat32(0, dt, true);
-        this._simUniformView.setUint32(4, liveCount, true);
-        this.device.queue.writeBuffer(this._simUniformBuffer, 0, this._simUniformData);
-    }
-    _writeModuleUniforms(dt) {
-        if (this._moduleUniformView === null || this._moduleUniformBuffer === null || this._moduleUniformData === null) {
-            return;
-        }
-        for (const slot of this._moduleSlots) {
-            slot.module.writeUniforms?.(this._moduleUniformView, slot.uniformByteOffset, dt);
-        }
-        this.device.queue.writeBuffer(this._moduleUniformBuffer, 0, this._moduleUniformData);
-    }
-    _buildBindGroup0Layout(slots) {
-        const entries = [
-            { binding: 0, visibility: GPUShaderStage.COMPUTE, buffer: { type: 'uniform' } },
-            { binding: 1, visibility: GPUShaderStage.COMPUTE, buffer: { type: 'uniform' } },
-        ];
-        if (this._moduleUniformBuffer !== null) {
-            entries.push({ binding: 2, visibility: GPUShaderStage.COMPUTE, buffer: { type: 'uniform' } });
-        }
-        let textureBindingIndex = this._moduleUniformBuffer !== null ? 3 : 2;
-        for (const slot of slots) {
-            for (const t of slot.contribution.textures ?? []) {
-                const filterable = t.format !== 'r32float';
-                entries.push({
-                    binding: textureBindingIndex++,
-                    visibility: GPUShaderStage.COMPUTE,
-                    texture: {
-                        viewDimension: '1d',
-                        sampleType: filterable ? 'float' : 'unfilterable-float',
-                    },
-                });
-                entries.push({
-                    binding: textureBindingIndex++,
-                    visibility: GPUShaderStage.COMPUTE,
-                    sampler: { type: filterable ? 'filtering' : 'non-filtering' },
-                });
-            }
-        }
-        return this.device.createBindGroupLayout({ label: 'particle-uniforms-bgl', entries });
-    }
-    _buildBindGroup0(layout, slots) {
-        const entries = [
-            { binding: 0, resource: { buffer: this._simUniformBuffer } },
-            { binding: 1, resource: { buffer: this._framesUniformBuffer } },
-        ];
-        if (this._moduleUniformBuffer !== null) {
-            entries.push({ binding: 2, resource: { buffer: this._moduleUniformBuffer } });
-        }
-        let textureBindingIndex = this._moduleUniformBuffer !== null ? 3 : 2;
-        for (const slot of slots) {
-            for (const t of slot.contribution.textures ?? []) {
-                const tex = this._moduleTextures.get(`${slot.contribution.key}_${t.name}`);
-                const filterable = t.format !== 'r32float';
-                const sampler = filterable ? this._samplerFiltering : this._samplerNonFiltering;
-                entries.push({ binding: textureBindingIndex++, resource: tex.createView({ dimension: '1d' }) });
-                entries.push({ binding: textureBindingIndex++, resource: sampler });
-            }
-        }
-        return this.device.createBindGroup({ label: 'particle-uniforms-bg', layout, entries });
-    }
-    _buildShader(slots) {
-        const sections = [];
-        sections.push(`
+const instanceBytes = 40;
+/** 8 x f32 (position, velocity, rotation, scale, elapsed) + 2 x u32 (color, slot). */
+const deathRecordBytes = 40;
+const deathStagingSlots = 3;
+const minDeathStagingRecords = 32;
+const deathRecordFloats = 10;
+var ParticleGpuState = class {
+	device;
+	capacity;
+	/** GPU buffer holding interleaved per-instance vertex data, written by compute, read as VERTEX by the renderer. */
+	instanceBuffer;
+	_positions;
+	_velocities;
+	_scales;
+	_rotInfo;
+	_timing;
+	_color;
+	_instanceStorageBuffer;
+	/**
+	* Death records the compute shader appended, plus the atomic append counter
+	* in its first four bytes. Allocated only while the system has death modules,
+	* so a system without them pays neither the memory nor the copy.
+	*
+	* Detecting a death and reporting one have separate lifetimes, and the path
+	* between them has four stages: the shader appends here; records the copy has
+	* not taken yet stay here as a device-side backlog; a copy moves a batch into
+	* a staging slot; and the mapped batches are handed on in submission order.
+	*/
+	_deathBuffer = null;
+	_deathCounterReset = /* @__PURE__ */ new Uint32Array(1);
+	/**
+	* Readback slots for the death buffer. A slot stays unavailable from the
+	* submit that copies into it until its map resolves, so a single slot would
+	* make every death that happens while a readback is in flight either a
+	* validation error or a lost record. Three slots cover the queue depth a
+	* frame loop runs at; a fourth death batch waits on the device instead.
+	*/
+	_deathStaging = [];
+	/** Copies submitted into a staging slot, oldest first, awaiting delivery. */
+	_stagedDeaths = [];
+	/**
+	* Whether the death buffer holds records no readback has claimed yet. The
+	* append counter is reset only once a copy has taken them, so a step that
+	* finds no free staging slot leaves its records in place and the next step
+	* appends behind them.
+	*/
+	_deathBufferDirty = false;
+	/** Tail of the delivery chain, so batches are reported in submission order. */
+	_deathDelivery = Promise.resolve();
+	_simUniformBuffer;
+	_simUniformData = /* @__PURE__ */ new ArrayBuffer(16);
+	_simUniformView;
+	_moduleUniformBuffer = null;
+	_moduleUniformData = null;
+	_moduleUniformView = null;
+	_moduleSlots = [];
+	_framesUniformBuffer;
+	_framesUniformData;
+	_framesUniformView;
+	_frameCount;
+	_moduleTextures = /* @__PURE__ */ new Map();
+	_samplerFiltering;
+	_samplerNonFiltering;
+	_pipelineWrapper = null;
+	_bindGroup0 = null;
+	_bindGroup1 = null;
+	_reportsDeaths = false;
+	_destroyed = false;
+	constructor(device, capacity, modules, frames, texture, reportsDeaths = false) {
+		this.device = device;
+		this.capacity = capacity;
+		this._frameCount = Math.max(1, frames.length);
+		this._framesUniformData = /* @__PURE__ */ new ArrayBuffer(this._frameCount * 16);
+		this._framesUniformView = new Float32Array(this._framesUniformData);
+		this._framesUniformBuffer = new WebGpuUniformBuffer(device, this._framesUniformData.byteLength, "particle-frames-uniforms");
+		this._writeFrames(frames, texture);
+		const vec2Bytes = capacity * 8;
+		const vec4Bytes = capacity * 16;
+		const u32Bytes = capacity * 4;
+		this._positions = new WebGpuStorageBuffer(device, vec2Bytes, "particle-positions");
+		this._velocities = new WebGpuStorageBuffer(device, vec2Bytes, "particle-velocities");
+		this._scales = new WebGpuStorageBuffer(device, vec2Bytes, "particle-scales");
+		this._rotInfo = new WebGpuStorageBuffer(device, vec4Bytes, "particle-rotInfo");
+		this._timing = new WebGpuStorageBuffer(device, vec2Bytes, "particle-timing");
+		this._color = new WebGpuStorageBuffer(device, u32Bytes, "particle-color");
+		this._instanceStorageBuffer = new WebGpuStorageBuffer(device, capacity * instanceBytes, "particle-instance-output", GPUBufferUsage.VERTEX);
+		this.instanceBuffer = this._instanceStorageBuffer.buffer;
+		this._simUniformView = new DataView(this._simUniformData);
+		this._simUniformBuffer = new WebGpuUniformBuffer(device, 16, "particle-sim-uniforms");
+		this._samplerFiltering = device.createSampler({
+			label: "particle-lookup-sampler-filtering",
+			minFilter: "linear",
+			magFilter: "linear",
+			addressModeU: "clamp-to-edge"
+		});
+		this._samplerNonFiltering = device.createSampler({
+			label: "particle-lookup-sampler-non-filtering",
+			minFilter: "nearest",
+			magFilter: "nearest",
+			addressModeU: "clamp-to-edge"
+		});
+		this.setProgram(modules, reportsDeaths);
+	}
+	/**
+	* Compiles the pipeline for `modules` and binds it to the buffers this state
+	* already owns.
+	*
+	* The simulation lives in those buffers, not in the program, so changing the
+	* module list rebuilds the shader, its uniforms and its lookup textures while
+	* every live particle keeps the position and velocity the GPU last integrated.
+	*/
+	setProgram(modules, reportsDeaths) {
+		for (const m of modules) if (!m.wgsl) throw new Error(`ParticleGpuState: module ${m.constructor.name} has no wgsl() - all registered UpdateModules must be GPU-eligible.`);
+		this._destroyProgram();
+		this._reportsDeaths = reportsDeaths;
+		const slots = [];
+		let uniformOffset = 0;
+		for (const m of modules) {
+			const c = m.wgsl();
+			const fields = c.uniforms ?? [];
+			const size = getWgslUniformByteSize(fields);
+			uniformOffset = Math.ceil(uniformOffset / 16) * 16;
+			slots.push({
+				module: m,
+				contribution: c,
+				uniformByteOffset: uniformOffset,
+				uniformByteSize: size
+			});
+			uniformOffset += size;
+		}
+		const totalUniformBytes = Math.max(16, Math.ceil(uniformOffset / 16) * 16);
+		this._moduleSlots = slots;
+		if (uniformOffset > 0) {
+			this._moduleUniformData = new ArrayBuffer(totalUniformBytes);
+			this._moduleUniformView = new DataView(this._moduleUniformData);
+			this._moduleUniformBuffer = new WebGpuUniformBuffer(this.device, totalUniformBytes, "particle-module-uniforms");
+		}
+		if (reportsDeaths && this._deathBuffer === null) this._deathBuffer = new WebGpuStorageBuffer(this.device, 4 + this.capacity * deathRecordBytes, "particle-deaths", GPUBufferUsage.COPY_SRC);
+		for (const slot of slots) {
+			const c = slot.contribution;
+			if (!c.textures) continue;
+			for (const t of c.textures) {
+				const tex = this.device.createTexture({
+					label: `particle-tex-${c.key}-${t.name}`,
+					size: {
+						width: 256,
+						height: 1,
+						depthOrArrayLayers: 1
+					},
+					format: t.format,
+					usage: GPUTextureUsage.TEXTURE_BINDING | GPUTextureUsage.COPY_DST,
+					dimension: "1d"
+				});
+				this._moduleTextures.set(`${c.key}_${t.name}`, tex);
+			}
+		}
+		const wgsl = this._buildShader(slots);
+		this._pipelineWrapper = WebGpuComputePipeline.create(this.device, {
+			wgsl,
+			workgroupSize,
+			bindingGroups: reflectComputeBindings(wgsl, { nonFilteringResources: this._nonFilteringResourceNames(slots) }),
+			label: "particle-compute"
+		});
+		this._bindGroup0 = this._pipelineWrapper.createBindGroup(0, this._buildBindGroup0Entries(slots), "particle-uniforms-bg");
+		this._bindGroup1 = this._pipelineWrapper.createBindGroup(1, this._buildSoaBindGroupEntries(), "particle-soa-bg");
+		for (const slot of slots) {
+			if (!slot.module.uploadTextures) continue;
+			const moduleTextures = /* @__PURE__ */ new Map();
+			for (const t of slot.contribution.textures ?? []) {
+				const tex = this._moduleTextures.get(`${slot.contribution.key}_${t.name}`);
+				if (tex !== void 0) moduleTextures.set(t.name, tex);
+			}
+			slot.module.uploadTextures(this.device, moduleTextures);
+		}
+	}
+	/** Releases everything the current program owns, leaving the simulation buffers untouched. */
+	_destroyProgram() {
+		this._moduleUniformBuffer?.destroy();
+		this._moduleUniformBuffer = null;
+		this._moduleUniformData = null;
+		this._moduleUniformView = null;
+		this._moduleSlots = [];
+		for (const tex of this._moduleTextures.values()) tex.destroy();
+		this._moduleTextures.clear();
+		this._pipelineWrapper = null;
+		this._bindGroup0 = null;
+		this._bindGroup1 = null;
+	}
+	/**
+	* Runs one simulation step over `[0, dispatchCount)`.
+	*
+	* `pendingDeaths` is how many expired particles the caller is still holding,
+	* including the ones it marked for this step. The shader appends exactly one
+	* record per marked particle, so the readback copies the used prefix of the
+	* death buffer without first having to read its counter back - the count is
+	* already known on this side.
+	*
+	* Returns whether a death batch was staged for readback. Deaths stay on the
+	* device while every staging slot is in flight, so a step can report deaths
+	* without staging any, and a later step can stage a batch spanning several
+	* steps. A caller keeping per-death state has to hold it, and keep counting it
+	* into `pendingDeaths`, until a step stages - then hand over all of it.
+	*/
+	dispatch(dt, dispatchCount, pendingDeaths = 0) {
+		const pipeline = this._pipelineWrapper;
+		const bindGroup0 = this._bindGroup0;
+		const bindGroup1 = this._bindGroup1;
+		if (pipeline === null || bindGroup0 === null || bindGroup1 === null) return false;
+		const simulating = dispatchCount > 0;
+		const reporting = pendingDeaths > 0 && this._reportsDeaths && this._deathBuffer !== null;
+		if (!simulating && !reporting) return false;
+		if (simulating) {
+			this._writeSimUniforms(dt, dispatchCount);
+			this._writeModuleUniforms(dt);
+		}
+		if (reporting && !this._deathBufferDirty) {
+			this._deathCounterReset[0] = 0;
+			this._deathBuffer.write(this._deathCounterReset, 0);
+		}
+		const encoder = this.device.createCommandEncoder({ label: "particle-compute" });
+		if (simulating) {
+			const pass = encoder.beginComputePass({ label: "particle-compute-pass" });
+			pipeline.dispatch(pass, dispatchCount, [bindGroup0, bindGroup1]);
+			pass.end();
+		}
+		let staged = false;
+		if (reporting) {
+			this._deathBufferDirty = true;
+			const count = Math.min(pendingDeaths, this.capacity);
+			const slot = this._acquireDeathStaging(count);
+			if (slot !== null) {
+				encoder.copyBufferToBuffer(this._deathBuffer.buffer, 4, slot.buffer, 0, count * deathRecordBytes);
+				slot.busy = true;
+				this._stagedDeaths.push({
+					slot,
+					count
+				});
+				this._deathBufferDirty = false;
+				staged = true;
+			}
+		}
+		this.device.queue.submit([encoder.finish()]);
+		return staged;
+	}
+	/**
+	* Resolves the oldest staged death batch and hands its records to `receive`
+	* once the copy has landed.
+	*
+	* The mapping is asynchronous by nature, so a death is delivered no earlier
+	* than the frame after it happened. Batches are delivered in the order they
+	* were submitted, and nothing blocks on a map: further steps keep staging
+	* into other slots while one is in flight. A device that goes away mid-map
+	* resolves to nothing rather than throwing: the deaths are lost with the
+	* simulation that produced them.
+	*/
+	readDeaths(receive) {
+		const batch = this._stagedDeaths.shift();
+		if (batch === void 0 || this._destroyed) return Promise.resolve();
+		const previous = this._deathDelivery;
+		let settled;
+		this._deathDelivery = new Promise((resolve) => {
+			settled = resolve;
+		});
+		return this._deliverDeaths(batch, previous, receive).finally(() => settled());
+	}
+	async _deliverDeaths(batch, previous, receive) {
+		const { slot, count } = batch;
+		const bytes = count * deathRecordBytes;
+		try {
+			await slot.buffer.mapAsync(GPUMapMode.READ, 0, bytes);
+		} catch {
+			slot.busy = false;
+			return;
+		}
+		if (this._destroyed) {
+			slot.busy = false;
+			return;
+		}
+		const records = [];
+		const mapped = slot.buffer.getMappedRange(0, bytes);
+		const floats = new Float32Array(mapped);
+		const uints = new Uint32Array(mapped);
+		for (let i = 0; i < count; i++) {
+			const base = i * deathRecordFloats;
+			records.push({
+				x: floats[base + 0],
+				y: floats[base + 1],
+				velocityX: floats[base + 2],
+				velocityY: floats[base + 3],
+				rotation: floats[base + 4],
+				scaleX: floats[base + 5],
+				scaleY: floats[base + 6],
+				elapsed: floats[base + 7],
+				color: uints[base + 8],
+				slot: uints[base + 9]
+			});
+		}
+		slot.buffer.unmap();
+		slot.busy = false;
+		records.sort((a, b) => a.slot - b.slot);
+		await previous;
+		if (this._destroyed) return;
+		receive(records);
+	}
+	/**
+	* Claims a staging slot able to hold `records`, or null while every slot of
+	* the ring is still in flight.
+	*/
+	_acquireDeathStaging(records) {
+		let reusable = null;
+		for (const slot of this._deathStaging) {
+			if (slot.busy) continue;
+			if (slot.records >= records) return slot;
+			reusable = slot;
+		}
+		if (reusable === null && this._deathStaging.length >= deathStagingSlots) return null;
+		const size = Math.max(records, (reusable?.records ?? 0) * 2, minDeathStagingRecords);
+		const buffer = this.device.createBuffer({
+			label: "particle-deaths-staging",
+			size: size * deathRecordBytes,
+			usage: GPUBufferUsage.COPY_DST | GPUBufferUsage.MAP_READ
+		});
+		if (reusable !== null) {
+			reusable.buffer.destroy();
+			reusable.buffer = buffer;
+			reusable.records = size;
+			return reusable;
+		}
+		const slot = {
+			buffer,
+			records: size,
+			busy: false
+		};
+		this._deathStaging.push(slot);
+		return slot;
+	}
+	_releaseDeathStaging() {
+		for (const slot of this._deathStaging) slot.buffer.destroy();
+		this._deathStaging.length = 0;
+		this._stagedDeaths.length = 0;
+		this._deathBufferDirty = false;
+	}
+	destroy() {
+		this._destroyed = true;
+		this._destroyProgram();
+		this._positions.destroy();
+		this._velocities.destroy();
+		this._scales.destroy();
+		this._rotInfo.destroy();
+		this._timing.destroy();
+		this._color.destroy();
+		this._instanceStorageBuffer.destroy();
+		this._simUniformBuffer.destroy();
+		this._framesUniformBuffer.destroy();
+		this._deathBuffer?.destroy();
+		this._deathBuffer = null;
+		this._releaseDeathStaging();
+	}
+	_writeFrames(frames, texture) {
+		const view = this._framesUniformView;
+		const w = texture.width;
+		const h = texture.height;
+		const flipY = texture.flipY;
+		if (frames.length === 0) {
+			view[0] = 0;
+			view[1] = flipY ? 1 : 0;
+			view[2] = 1;
+			view[3] = flipY ? 0 : 1;
+		} else for (let i = 0; i < frames.length; i++) {
+			const f = frames[i];
+			const o = i * 4;
+			const minU = f.left / w;
+			const maxU = f.right / w;
+			const topV = f.top / h;
+			const bottomV = f.bottom / h;
+			view[o + 0] = minU;
+			view[o + 1] = flipY ? bottomV : topV;
+			view[o + 2] = maxU;
+			view[o + 3] = flipY ? topV : bottomV;
+		}
+		this._framesUniformBuffer.write(this._framesUniformView);
+	}
+	/**
+	* Push the listed CPU SoA slots to the GPU. Called by `ParticleSystem`
+	* with newly-spawned slots and just-expired slots (lifetime sentinel).
+	* Slots not in the dirty set are left alone - GPU keeps the integrated
+	* state from previous compute dispatches.
+	*
+	* Each dirty slot triggers 7 small writes (one per SoA channel, via
+	* {@link WebGpuStorageBuffer.write}). For typical spawn rates (≤200/s)
+	* this is negligible (≤1400 calls/s); contiguous-range batching is a
+	* future optimisation.
+	*/
+	uploadDirty(system, slots) {
+		const scratch2 = this._dirtyScratchVec2;
+		const scratch4 = this._dirtyScratchVec4;
+		const scratch1 = this._dirtyScratchU32;
+		const storage = system._storage;
+		for (const slot of slots) {
+			const byteOffset2 = slot * 8;
+			const byteOffset1 = slot * 4;
+			scratch2[0] = storage.posX[slot];
+			scratch2[1] = storage.posY[slot];
+			this._positions.write(scratch2, byteOffset2);
+			scratch2[0] = storage.velX[slot];
+			scratch2[1] = storage.velY[slot];
+			this._velocities.write(scratch2, byteOffset2);
+			scratch2[0] = storage.scaleX[slot];
+			scratch2[1] = storage.scaleY[slot];
+			this._scales.write(scratch2, byteOffset2);
+			scratch4[0] = storage.rotations[slot];
+			scratch4[1] = storage.rotationSpeeds[slot];
+			scratch4[2] = storage.frame[slot];
+			scratch4[3] = 0;
+			this._rotInfo.write(scratch4, slot * 16);
+			scratch2[0] = storage.elapsed[slot];
+			scratch2[1] = storage.lifetime[slot];
+			this._timing.write(scratch2, byteOffset2);
+			scratch1[0] = storage.color[slot];
+			this._color.write(scratch1, byteOffset1);
+		}
+	}
+	/**
+	* Marks `slot` expired for the device without touching anything else about
+	* it. Only the lifetime lane is written: a full slot upload would push the
+	* CPU's stale position and velocity over the values the device integrated,
+	* and those are exactly what the death record is supposed to carry.
+	*/
+	uploadExpiry(slot) {
+		this._expiryScratch[0] = -1;
+		this._timing.write(this._expiryScratch, slot * 8 + 4);
+	}
+	_expiryScratch = /* @__PURE__ */ new Float32Array(1);
+	_dirtyScratchVec2 = /* @__PURE__ */ new Float32Array(2);
+	_dirtyScratchVec4 = /* @__PURE__ */ new Float32Array(4);
+	_dirtyScratchU32 = /* @__PURE__ */ new Uint32Array(1);
+	_writeSimUniforms(dt, liveCount) {
+		this._simUniformView.setFloat32(0, dt, true);
+		this._simUniformView.setUint32(4, liveCount, true);
+		this._simUniformBuffer.write(this._simUniformView);
+	}
+	_writeModuleUniforms(dt) {
+		if (this._moduleUniformView === null || this._moduleUniformBuffer === null || this._moduleUniformData === null) return;
+		for (const slot of this._moduleSlots) slot.module.writeUniforms?.(this._moduleUniformView, slot.uniformByteOffset, dt);
+		this._moduleUniformBuffer.write(this._moduleUniformView);
+	}
+	/**
+	* WGSL variable names of module lookup textures/samplers whose format isn't natively
+	* filterable (`r32float` Curve LUTs) - fed to {@link reflectComputeBindings} so it declares
+	* `'unfilterable-float'`/`'non-filtering'` for those specific bindings instead of the default
+	* `'float'`/`'filtering'`, an ambiguity the WGSL text itself can't resolve (see that
+	* function's doc comment).
+	*/
+	_nonFilteringResourceNames(slots) {
+		const names = /* @__PURE__ */ new Set();
+		for (const slot of slots) for (const t of slot.contribution.textures ?? []) if (t.format === "r32float") {
+			names.add(`u_${slot.contribution.key}_${t.name}`);
+			names.add(`u_${slot.contribution.key}_${t.name}_sampler`);
+		}
+		return names;
+	}
+	/** Group-0 bind-group entries, matching group 0's bindings (reflected from the shader text) one-to-one. */
+	_buildBindGroup0Entries(slots) {
+		const entries = [{
+			binding: 0,
+			buffer: this._simUniformBuffer.buffer
+		}, {
+			binding: 1,
+			buffer: this._framesUniformBuffer.buffer
+		}];
+		if (this._moduleUniformBuffer !== null) entries.push({
+			binding: 2,
+			buffer: this._moduleUniformBuffer.buffer
+		});
+		let textureBindingIndex = this._moduleUniformBuffer !== null ? 3 : 2;
+		for (const slot of slots) for (const t of slot.contribution.textures ?? []) {
+			const tex = this._moduleTextures.get(`${slot.contribution.key}_${t.name}`);
+			const sampler = t.format !== "r32float" ? this._samplerFiltering : this._samplerNonFiltering;
+			entries.push({
+				binding: textureBindingIndex++,
+				textureView: tex.createView({ dimension: "1d" })
+			});
+			entries.push({
+				binding: textureBindingIndex++,
+				sampler
+			});
+		}
+		return entries;
+	}
+	/** Group-1 bind-group entries: the 8 SoA storage buffers, matching group 1's bindings (reflected from the shader text) one-to-one. */
+	_buildSoaBindGroupEntries() {
+		const entries = [
+			{
+				binding: 0,
+				buffer: this._positions.buffer
+			},
+			{
+				binding: 1,
+				buffer: this._velocities.buffer
+			},
+			{
+				binding: 2,
+				buffer: this._scales.buffer
+			},
+			{
+				binding: 3,
+				buffer: this._rotInfo.buffer
+			},
+			{
+				binding: 4,
+				buffer: this._timing.buffer
+			},
+			{
+				binding: 5,
+				buffer: this._color.buffer
+			},
+			{
+				binding: 6,
+				buffer: this._instanceStorageBuffer.buffer
+			}
+		];
+		if (this._reportsDeaths && this._deathBuffer !== null) entries.push({
+			binding: 7,
+			buffer: this._deathBuffer.buffer
+		});
+		return entries;
+	}
+	_buildShader(slots) {
+		const sections = [];
+		sections.push(`
 struct SimUniforms {
     dt: f32,
     liveCount: u32,
@@ -426,110 +578,116 @@ struct FrameUniforms {
 @group(0) @binding(0) var<uniform> sim: SimUniforms;
 @group(0) @binding(1) var<uniform> frameUv: FrameUniforms;
         `);
-        const moduleStructFields = [];
-        for (const slot of slots) {
-            const c = slot.contribution;
-            const fields = c.uniforms ?? [];
-            if (fields.length === 0) {
-                continue;
-            }
-            sections.push(this._renderModuleStruct(c.key, fields));
-            moduleStructFields.push(`u_${c.key}: ${c.key}Uniforms,`);
-        }
-        if (moduleStructFields.length > 0) {
-            sections.push(`
+		const moduleStructFields = [];
+		for (const slot of slots) {
+			const c = slot.contribution;
+			const fields = c.uniforms ?? [];
+			if (fields.length === 0) continue;
+			sections.push(this._renderModuleStruct(c.key, fields));
+			moduleStructFields.push(`u_${c.key}: ${c.key}Uniforms,`);
+		}
+		if (moduleStructFields.length > 0) sections.push(`
 struct ModuleUniforms {
-${moduleStructFields.map(s => `    ${s}`).join('\n')}
+${moduleStructFields.map((s) => `    ${s}`).join("\n")}
 }
 
 @group(0) @binding(2) var<uniform> modules: ModuleUniforms;
             `);
-        }
-        let textureBindingIndex = moduleStructFields.length > 0 ? 3 : 2;
-        for (const slot of slots) {
-            for (const t of slot.contribution.textures ?? []) {
-                sections.push(`
+		let textureBindingIndex = moduleStructFields.length > 0 ? 3 : 2;
+		for (const slot of slots) for (const t of slot.contribution.textures ?? []) sections.push(`
 @group(0) @binding(${textureBindingIndex++}) var u_${slot.contribution.key}_${t.name}: texture_1d<f32>;
 @group(0) @binding(${textureBindingIndex++}) var u_${slot.contribution.key}_${t.name}_sampler: sampler;
                 `);
-            }
-        }
-        sections.push(`
+		sections.push(`
 @group(1) @binding(0) var<storage, read_write> positions: array<vec2<f32>>;
 @group(1) @binding(1) var<storage, read_write> velocities: array<vec2<f32>>;
 @group(1) @binding(2) var<storage, read_write> scales: array<vec2<f32>>;
-@group(1) @binding(3) var<storage, read_write> rotInfo: array<vec2<f32>>;
+@group(1) @binding(3) var<storage, read_write> rotInfo: array<vec4<f32>>;
 @group(1) @binding(4) var<storage, read_write> timing: array<vec2<f32>>;
 @group(1) @binding(5) var<storage, read_write> color: array<u32>;
-@group(1) @binding(6) var<storage, read>       textureIndex: array<u32>;
-@group(1) @binding(7) var<storage, read_write> instanceOutput: array<u32>;
+@group(1) @binding(6) var<storage, read_write> instanceOutput: array<u32>;
         `);
-        // Module preludes (helper functions/constants). Concatenated in
-        // registration order; modules sharing the same key are emitted only
-        // once (the contribution body strings are still inlined per-instance,
-        // but the prelude function definitions can't be duplicated).
-        const seenPreludeKeys = new Set();
-        for (const slot of slots) {
-            const prelude = slot.contribution.prelude;
-            if (prelude === undefined || prelude.trim() === '')
-                continue;
-            if (seenPreludeKeys.has(slot.contribution.key))
-                continue;
-            seenPreludeKeys.add(slot.contribution.key);
-            sections.push(prelude);
-        }
-        const moduleBodies = slots.map(s => s.contribution.body).join('\n');
-        const frameCountConst = this._frameCount;
-        sections.push(`
-@compute @workgroup_size(${workgroupSize})
-fn main(@builtin(global_invocation_id) gid: vec3<u32>) {
-    let idx = gid.x;
-    if (idx >= sim.liveCount) { return; }
-
-    let dt = sim.dt;
+		if (this._reportsDeaths) sections.push(`
+struct DeathRecord {
+    x: f32,
+    y: f32,
+    velocityX: f32,
+    velocityY: f32,
+    rotation: f32,
+    scaleX: f32,
+    scaleY: f32,
+    elapsed: f32,
+    color: u32,
+    slot: u32,
+}
 
-    // Skip dead particles (lifetime sentinel < 0). Write zero-scale instance
-    // so the renderer doesn't accidentally draw them.
-    if (timing[idx].y < 0.0) {
-        let outBaseDead = idx * 10u;
-        for (var k: u32 = 0u; k < 10u; k++) { instanceOutput[outBaseDead + k] = 0u; }
-        return;
-    }
+struct DeathBuffer {
+    count: atomic<u32>,
+    records: array<DeathRecord>,
+}
 
-    // Integration.
-    positions[idx] = positions[idx] + velocities[idx] * dt;
-    rotInfo[idx].x = rotInfo[idx].x + rotInfo[idx].y * dt;
-    timing[idx].x = timing[idx].x + dt;
+@group(1) @binding(7) var<storage, read_write> deaths: DeathBuffer;
+            `);
+		const seenPreludeKeys = /* @__PURE__ */ new Set();
+		for (const slot of slots) {
+			const prelude = slot.contribution.prelude;
+			if (prelude === void 0 || prelude.trim() === "") continue;
+			if (seenPreludeKeys.has(slot.contribution.key)) continue;
+			seenPreludeKeys.add(slot.contribution.key);
+			sections.push(prelude);
+		}
+		const moduleBodies = slots.map((s) => s.contribution.body).join("\n");
+		sections.push(`
+${fillShaderSource(particle_simulate_default, {
+			workgroupSize,
+			moduleBodies,
+			frameCount: this._frameCount,
+			deathReport: this._deathReportSource()
+		})}        `);
+		return sections.join("\n\n");
+	}
+	/**
+	* The block that captures a particle's state the first time the shader sees
+	* its expiry sentinel, or nothing when the system has no death modules.
+	*
+	* The CPU marks an expired particle with `lifetime = -1` and the shader
+	* rewrites that to `-2` once captured, so exactly one record is appended per
+	* death even though a dead slot is visited every frame until it is reused.
+	*/
+	_deathReportSource() {
+		if (!this._reportsDeaths) return "";
+		return `
+    if (timing[idx].y > -1.5) {
+        // The CPU found this particle expired from an elapsed time it had
+        // already advanced for this frame, so the device owes it the matching
+        // integration step before the snapshot is taken. Without it a death
+        // reported from the GPU would sit one step behind the same death
+        // reported from the CPU pipeline.
+        positions[idx] = positions[idx] + velocities[idx] * dt;
+        rotInfo[idx].x = rotInfo[idx].x + rotInfo[idx].y * dt;
+        timing[idx].x = timing[idx].x + dt;
 
-    // Module bodies (in registration order).
-${moduleBodies}
+        let at = atomicAdd(&deaths.count, 1u);
 
-    // Resolve frame UVs.
-    let frameIndex = min(textureIndex[idx], ${frameCountConst}u - 1u);
-    let frameUvBounds = frameUv.frames[frameIndex];
+        deaths.records[at].x = positions[idx].x;
+        deaths.records[at].y = positions[idx].y;
+        deaths.records[at].velocityX = velocities[idx].x;
+        deaths.records[at].velocityY = velocities[idx].y;
+        deaths.records[at].rotation = rotInfo[idx].x;
+        deaths.records[at].scaleX = scales[idx].x;
+        deaths.records[at].scaleY = scales[idx].y;
+        deaths.records[at].elapsed = timing[idx].x;
+        deaths.records[at].color = color[idx];
+        deaths.records[at].slot = idx;
 
-    // Pack interleaved instance data (10 u32s per particle):
-    //   x, y, scaleX, scaleY, rotation (f32×5) + color (u32) + uvMin.xy (f32×2) + uvMax.xy (f32×2)
-    let outBase = idx * 10u;
-    instanceOutput[outBase + 0u] = bitcast<u32>(positions[idx].x);
-    instanceOutput[outBase + 1u] = bitcast<u32>(positions[idx].y);
-    instanceOutput[outBase + 2u] = bitcast<u32>(scales[idx].x);
-    instanceOutput[outBase + 3u] = bitcast<u32>(scales[idx].y);
-    instanceOutput[outBase + 4u] = bitcast<u32>(rotInfo[idx].x);
-    instanceOutput[outBase + 5u] = color[idx];
-    instanceOutput[outBase + 6u] = bitcast<u32>(frameUvBounds.x);
-    instanceOutput[outBase + 7u] = bitcast<u32>(frameUvBounds.y);
-    instanceOutput[outBase + 8u] = bitcast<u32>(frameUvBounds.z);
-    instanceOutput[outBase + 9u] = bitcast<u32>(frameUvBounds.w);
-}
-        `);
-        return sections.join('\n\n');
-    }
-    _renderModuleStruct(key, fields) {
-        const lines = fields.map(f => `    ${f.name}: ${f.type},`).join('\n');
-        return `struct ${key}Uniforms {\n${lines}\n}`;
-    }
-}
+        timing[idx].y = -2.0;
+    }`;
+	}
+	_renderModuleStruct(key, fields) {
+		return `struct ${key}Uniforms {\n${fields.map((f) => `    ${f.name}: ${f.type},`).join("\n")}\n}`;
+	}
+};
 
+//#endregion
 export { ParticleGpuState };
-//# sourceMappingURL=ParticleGpuState.js.map
+//# sourceMappingURL=ParticleGpuState.js.map