Core Concepts

Understanding these fundamental concepts will help you make the most of Entropy Zero.

The Simulation Loop

Entropy Zero uses a fixed-timestep simulation loop for deterministic results:

┌─────────────────────────────────────────┐
│  Accumulate Time                        │
│  ↓                                      │
│  While (accumulated >= timestep)        │
│    → Apply Forces                       │
│    → Detect Collisions                  │
│    → Resolve Constraints                │
│    → Integrate Motion                   │
│    → accumulated -= timestep            │
│  ↓                                      │
│  Interpolate for Rendering              │
└─────────────────────────────────────────┘

Entity-Component Architecture

Simulated objects are composed of components:

• Transform — Position, rotation, scale
• RigidBody — Mass, velocity, angular momentum
• Collider — Shape for collision detection
• Constraint — Joints, springs, distance limits

Spatial Partitioning

For performance, Entropy Zero uses adaptive spatial partitioning:

• Small scenes → Uniform grid
• Large scenes → BVH (Bounding Volume Hierarchy)
• Mixed scales → Hybrid approach

Parallelism

The simulation leverages Rayon for parallel processing:

• Broadphase collision detection runs in parallel
• Force calculations are batched and parallelized
• Constraint solving uses parallel Gauss-Seidel iterations