agentleFS
Bbeierle12Bbeierle12/Skill-MCP-Claude/skills/particles-gpu/SKILL.md

particles-gpu

Skill9 starsChanged 6 months agoMITView on GitHub

GPU-based particle systems using instanced rendering, buffer attributes, Points geometry, and custom shaders. Use when rendering thousands to millions of particles efficiently, creating particle effects like snow, rain, stars, or abstract visualizations.

1 other project uses a version of this.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Add per-particle data like color, size, velocity:

    line 125

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

GPU Particles

Render massive particle counts (10k-1M+) efficiently using GPU instancing and custom shaders.

Quick Start

import { useRef, useMemo } from 'react';
import { useFrame } from '@react-three/fiber';
import * as THREE from 'three';

function Particles({ count = 10000 }) {
  const points = useRef<THREE.Points>(null!);
  
  const positions = useMemo(() => {
    const pos = new Float32Array(count * 3);
    for (let i = 0; i < count; i++) {
      pos[i * 3] = (Math.random() - 0.5) * 10;
      pos[i * 3 + 1] = (Math.random() - 0.5) * 10;
      pos[i * 3 + 2] = (Math.random() - 0.5) * 10;
    }
    return pos;
  }, [count]);
  
  return (
    <points ref={points}>
      <bufferGeometry>
        <bufferAttribute
          attach="attributes-position"
          count={count}
          array={positions}
          itemSize={3}
        />
      </bufferGeometry>
      <pointsMaterial size={0.05} color="#ffffff" />
    </points>
  );
}

Rendering Approaches

ApproachParticle CountComplexityUse Case
Points10k - 500kLowSimple particles, stars
Instanced Mesh1k - 100kMedium3D geometry particles
Custom Shader100k - 10MHighMaximum control

Points Geometry

Simplest approach—each particle is a screen-facing point sprite.

Basic Points

function BasicPoints({ count = 5000 }) {
  const positions = useMemo(() => {
    const pos = new Float32Array(count * 3);
    for (let i = 0; i < count; i++) {
      const theta = Math.random() * Math.PI * 2;
      const phi = Math.acos(2 * Math.random() - 1);
      const r = Math.cbrt(Math.random()) * 5;
      
      pos[i * 3] = r * Math.sin(phi) * Math.cos(theta);
      pos[i * 3 + 1] = r * Math.sin(phi) * Math.sin(theta);
      pos[i * 3 + 2] = r * Math.cos(phi);
    }
    return pos;
  }, [count]);
  
  return (
    <points>
      <bufferGeometry>
        <bufferAttribute
          attach="attributes-position"
          count={count}
          array={positions}
          itemSize={3}
        />
      </bufferGeometry>
      <pointsMaterial
        size={0.1}
        sizeAttenuation={true}
        transparent={true}
        opacity={0.8}
        depthWrite={false}
        blending={THREE.AdditiveBlending}
      />
    </points>
  );
}

Points with Texture

function TexturedPoints({ count = 5000 }) {
  const texture = useTexture('/particle.png');
  
  return (
    <points>
      <bufferGeometry>
        {/* ... positions ... */}
      </bufferGeometry>
      <pointsMaterial
        size={0.5}
        map={texture}
        transparent={true}
        alphaTest={0.01}
        depthWrite={false}
        blending={THREE.AdditiveBlending}
      />
    </points>
  );
}

Custom Attributes

Add per-particle data like color, size, velocity:

function ColoredParticles({ count = 10000 }) {
  const { positions, colors, sizes } = useMemo(() => {
    const pos = new Float32Array(count * 3);
    const col = new Float32Array(count * 3);
    const siz = new Float32Array(count);
    
    for (let i = 0; i < count; i++) {
      // Position
      pos[i * 3] = (Math.random() - 0.5) * 10;
      pos[i * 3 + 1] = (Math.random() - 0.5) * 10;
      pos[i * 3 + 2] = (Math.random() - 0.5) * 10;
      
      // Color (HSL to RGB)
      const color = new THREE.Color();
      color.setHSL(Math.random(), 0.8, 0.5);
      col[i * 3] = color.r;
      col[i * 3 + 1] = color.g;
      col[i * 3 + 2] = color.b;
      
      // Size
      siz[i] = 0.05 + Math.random() * 0.1;
    }
    
    return { positions: pos, colors: col, sizes: siz };
  }, [count]);
  
  return (
    <points>
      <bufferGeometry>
        <bufferAttribute
          attach="attributes-position"
          count={count}
          array={positions}
          itemSize={3}
        />
        <bufferAttribute
          attach="attributes-color"
          count={count}
          array={colors}
          itemSize={3}
        />
        <bufferAttribute
          attach="attributes-size"
          count={count}
          array={sizes}
          itemSize={1}
        />
      </bufferGeometry>
      <pointsMaterial
        vertexColors
        size={0.1}
        sizeAttenuation
        transparent
        depthWrite={false}
      />
    </points>
  );
}

Custom Shader Particles

Maximum control over particle appearance and animation:

const vertexShader = `
  attribute float aSize;
  attribute vec3 aColor;
  attribute float aAlpha;
  
  uniform float uTime;
  uniform float uPixelRatio;
  
  varying vec3 vColor;
  varying float vAlpha;
  
  void main() {
    vColor = aColor;
    vAlpha = aAlpha;
    
    vec4 mvPosition = modelViewMatrix * vec4(position, 1.0);
    
    // Size attenuation
    gl_PointSize = aSize * uPixelRatio * (300.0 / -mvPosition.z);
    gl_Position = projectionMatrix * mvPosition;
  }
`;

const fragmentShader = `
  varying vec3 vColor;
  varying float vAlpha;
  
  void main() {
    // Circular particle
    float dist = length(gl_PointCoord - 0.5);
    if (dist > 0.5) discard;
    
    // Soft edge
    float alpha = 1.0 - smoothstep(0.4, 0.5, dist);
    
    gl_FragColor = vec4(vColor, alpha * vAlpha);
  }
`;

function ShaderParticles({ count = 50000 }) {
  const points = useRef<THREE.Points>(null!);
  
  const { positions, sizes, colors, alphas } = useMemo(() => {
    const pos = new Float32Array(count * 3);
    const siz = new Float32Array(count);
    const col = new Float32Array(count * 3);
    const alp = new Float32Array(count);
    
    for (let i = 0; i < count; i++) {
      pos[i * 3] = (Math.random() - 0.5) * 20;
      pos[i * 3 + 1] = (Math.random() - 0.5) * 20;
      pos[i * 3 + 2] = (Math.random() - 0.5) * 20;
      
      siz[i] = 10 + Math.random() * 20;
      
      const color = new THREE.Color();
      color.setHSL(0.6 + Math.random() * 0.2, 0.8, 0.5);
      col[i * 3] = color.r;
      col[i * 3 + 1] = color.g;
      col[i * 3 + 2] = color.b;
      
      alp[i] = 0.3 + Math.random() * 0.7;
    }
    
    return { positions: pos, sizes: siz, colors: col, alphas: alp };
  }, [count]);
  
  useFrame(({ clock }) => {
    points.current.material.uniforms.uTime.value = clock.elapsedTime;
  });
  
  return (
    <points ref={points}>
      <bufferGeometry>
        <bufferAttribute attach="attributes-position" count={count} array={positions} itemSize={3} />
        <bufferAttribute attach="attributes-aSize" count={count} array={sizes} itemSize={1} />
        <bufferAttribute attach="attributes-aColor" count={count} array={colors} itemSize={3} />
        <bufferAttribute attach="attributes-aAlpha" count={count} array={alphas} itemSize={1} />
      </bufferGeometry>
      <shaderMaterial
        vertexShader={vertexShader}
        fragmentShader={fragmentShader}
        uniforms={{
          uTime: { value: 0 },
          uPixelRatio: { value: Math.min(window.devicePixelRatio, 2) }
        }}
        transparent
        depthWrite={false}
        blending={THREE.AdditiveBlending}
      />
    </points>
  );
}

Animated Particles

Position Animation in Shader

// Vertex shader with animation
attribute vec3 aVelocity;
attribute float aPhase;

uniform float uTime;

void main() {
  vec3 pos = position;
  
  // Simple oscillation
  pos.y += sin(uTime * 2.0 + aPhase) * 0.5;
  
  // Velocity-based movement
  pos += aVelocity * uTime;
  
  // Wrap around bounds
  pos = mod(pos + 10.0, 20.0) - 10.0;
  
  vec4 mvPosition = modelViewMatrix * vec4(pos, 1.0);
  gl_PointSize = 10.0 * (300.0 / -mvPosition.z);
  gl_Position = projectionMatrix * mvPosition;
}

CPU Animation (for dynamic systems)

function AnimatedParticles({ count = 10000 }) {
  const points = useRef<THREE.Points>(null!);
  
  const velocities = useMemo(() => {
    const vel = new Float32Array(count * 3);
    for (let i = 0; i < count; i++) {
      vel[i * 3] = (Math.random() - 0.5) * 0.02;
      vel[i * 3 + 1] = (Math.random() - 0.5) * 0.02;
      vel[i * 3 + 2] = (Math.random() - 0.5) * 0.02;
    }
    return vel;
  }, [count]);
  
  useFrame(() => {
    const positions = points.current.geometry.attributes.position.array as Float32Array;
    
    for (let i = 0; i < count; i++) {
      positions[i * 3] += velocities[i * 3];
      positions[i * 3 + 1] += velocities[i * 3 + 1];
      positions[i * 3 + 2] += velocities[i * 3 + 2];
      
      // Wrap around
      for (let j = 0; j < 3; j++) {
        if (positions[i * 3 + j] > 5) positions[i * 3 + j] = -5;
        if (positions[i * 3 + j] < -5) positions[i * 3 + j] = 5;
      }
    }
    
    points.current.geometry.attributes.position.needsUpdate = true;
  });
  
  // ... geometry setup
}

Instanced Mesh Particles

For 3D geometry particles (not just points):

function InstancedParticles({ count = 1000 }) {
  const mesh = useRef<THREE.InstancedMesh>(null!);
  const dummy = useMemo(() => new THREE.Object3D(), []);
  
  useEffect(() => {
    for (let i = 0; i < count; i++) {
      dummy.position.set(
        (Math.random() - 0.5) * 10,
        (Math.random() - 0.5) * 10,
        (Math.random() - 0.5) * 10
      );
      dummy.rotation.set(
        Math.random() * Math.PI,
        Math.random() * Math.PI,
        0
      );
      dummy.scale.setScalar(0.05 + Math.random() * 0.1);
      dummy.updateMatrix();
      mesh.current.setMatrixAt(i, dummy.matrix);
    }
    mesh.current.instanceMatrix.needsUpdate = true;
  }, [count, dummy]);
  
  useFrame(({ clock }) => {
    for (let i = 0; i < count; i++) {
      mesh.current.getMatrixAt(i, dummy.matrix);
      dummy.matrix.decompose(dummy.position, dummy.quaternion, dummy.scale);
      
      dummy.rotation.x += 0.01;
      dummy.rotation.y += 0.01;
      
      dummy.updateMatrix();
      mesh.current.setMatrixAt(i, dummy.matrix);
    }
    mesh.current.instanceMatrix.needsUpdate = true;
  });
  
  return (
    <instancedMesh ref={mesh} args={[undefined, undefined, count]}>
      <icosahedronGeometry args={[1, 0]} />
      <meshStandardMaterial color="#ff6b6b" />
    </instancedMesh>
  );
}

Buffer Geometry Patterns

Sphere Distribution

function spherePositions(count: number, radius: number) {
  const positions = new Float32Array(count * 3);
  
  for (let i = 0; i < count; i++) {
    const theta = Math.random() * Math.PI * 2;
    const phi = Math.acos(2 * Math.random() - 1);
    const r = Math.cbrt(Math.random()) * radius;  // Cube root for uniform volume
    
    positions[i * 3] = r * Math.sin(phi) * Math.cos(theta);
    positions[i * 3 + 1] = r * Math.sin(phi) * Math.sin(theta);
    positions[i * 3 + 2] = r * Math.cos(phi);
  }
  
  return positions;
}

Galaxy Spiral

function galaxyPositions(count: number, arms: number, spin: number) {
  const positions = new Float32Array(count * 3);
  
  for (let i = 0; i < count; i++) {
    const armIndex = i % arms;
    const armAngle = (armIndex / arms) * Math.PI * 2;
    
    const radius = Math.random() * 5;
    const spinAngle = radius * spin;
    const angle = armAngle + spinAngle;
    
    // Add randomness
    const randomX = (Math.random() - 0.5) * 0.5 * radius;
    const randomY = (Math.random() - 0.5) * 0.2;
    const randomZ = (Math.random() - 0.5) * 0.5 * radius;
    
    positions[i * 3] = Math.cos(angle) * radius + randomX;
    positions[i * 3 + 1] = randomY;
    positions[i * 3 + 2] = Math.sin(angle) * radius + randomZ;
  }
  
  return positions;
}

Grid Distribution

function gridPositions(countPerAxis: number, spacing: number) {
  const count = countPerAxis ** 3;
  const positions = new Float32Array(count * 3);
  const offset = (countPerAxis - 1) * spacing * 0.5;
  
  let index = 0;
  for (let x = 0; x < countPerAxis; x++) {
    for (let y = 0; y < countPerAxis; y++) {
      for (let z = 0; z < countPerAxis; z++) {
        positions[index * 3] = x * spacing - offset;
        positions[index * 3 + 1] = y * spacing - offset;
        positions[index * 3 + 2] = z * spacing - offset;
        index++;
      }
    }
  }
  
  return positions;
}

Performance Tips

TechniqueImpact
Use Points over InstancedMesh5-10x faster for simple particles
GPU animation (shader) vs CPU10-100x faster at scale
Disable depthWriteFaster blending
Use Float32ArrayRequired for buffers
Frustum culling (default on)Skip off-screen

Optimal Settings

<pointsMaterial
  transparent
  depthWrite={false}           // Faster blending
  blending={THREE.AdditiveBlending}  // Good for glowing particles
  sizeAttenuation              // Perspective-correct size
/>

File Structure

particles-gpu/
├── SKILL.md
├── references/
│   ├── buffer-patterns.md     # Distribution patterns
│   └── shader-examples.md     # Complete shader examples
└── scripts/
    ├── particles/
    │   ├── basic-points.tsx   # Simple points setup
    │   ├── shader-points.tsx  # Custom shader particles
    │   └── instanced.tsx      # Instanced mesh particles
    └── distributions/
        ├── sphere.ts          # Sphere distribution
        ├── galaxy.ts          # Galaxy spiral
        └── grid.ts            # Grid distribution

Reference

  • references/buffer-patterns.md — Position distribution patterns
  • references/shader-examples.md — Complete particle shaders

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.