Unisim - N-Body Physics Simulation

A high-performance n-body physics simulation application built with C++ and GTK4. Simulate gravitational interactions between thousands of bodies using various numerical integrators and force computation methods.

Features

• Multiple Integrators: Euler, Verlet, Leapfrog, Runge-Kutta (RK4)
• Force Computation: Brute Force (O(N²)), Barnes-Hut (O(N log N)), Fast Multipole (O(N))
• Rendering: 2D top-down, 3D, and 3D OpenGL views
• Initial Conditions: Random, Spiral Galaxy, Elliptical Galaxy, Galaxy Collision, Solar System, Binary Star System
• Compute Backends: CPU (always available), Metal (macOS), CUDA (optional)

Prerequisites

macOS

brew install gtk4

Linux

Install GTK4 development packages using your distribution's package manager:

# Ubuntu/Debian
sudo apt-get install libgtk-4-dev libepoxy-dev

# Fedora
sudo dnf install gtk4-devel libepoxy-devel

Requirements

• CMake 3.10 or higher
• C++17 compatible compiler
• pkg-config

Building

mkdir build
cd build
cmake ..
make

The executable unisim will be created in the build/ directory.

Optional: CUDA Support

If CUDA is installed and detected by CMake, it will be automatically enabled. The CUDA backend will appear as an option in the UI if available.

Running

From the build directory:

./unisim

Or from any directory:

/path/to/build/unisim

Operation Instructions

Basic Controls

• Play/Pause Button: Start or pause the simulation
• Reset Button: Regenerate initial conditions with current settings
• Reset Camera: Reset the viewport to default position and zoom

Simulation Parameters

• Number of Bodies: Adjust the number of simulated bodies (affects performance)
• Time Step (dt): Control simulation time step (smaller = more accurate, slower)

Backend Selection

Choose the compute backend:

• CPU: Standard CPU computation (always available)
• Metal: GPU acceleration on macOS (if available)
• CUDA: GPU acceleration on systems with NVIDIA GPUs (if available)

Integrator Selection

Choose a numerical integration method:

• Euler: Simple first-order method (fast, less accurate)
• Verlet: Second-order symplectic method (good energy conservation)
• Leapfrog: Second-order symplectic method (velocity-centered)
• Runge-Kutta (RK4): Fourth-order method (accurate, slower)

Force Method Selection

Choose how gravitational forces are computed:

• Brute Force: O(N²) - Accurate but slow for many bodies
• Barnes-Hut: O(N log N) - Tree-based approximation, faster for large N
• Fast Multipole: O(N) - Multipole expansion method, fastest for very large N

Renderer Selection

Choose the visualization mode:

• 2D: Top-down or side view projection
• 3D: Simple 3D rendering
• 3D OpenGL: Hardware-accelerated 3D rendering (recommended)

Initial Conditions

Select the starting configuration:

• Random: Random distribution of bodies
• Spiral Galaxy: Simulated spiral galaxy formation
• Elliptical Galaxy: Elliptical galaxy configuration
• Galaxy Collision: Two galaxies colliding
• Solar System: Our solar system simulation
• Binary Star System: Binary star orbit

View Options (3D OpenGL)

• Show Vectors: Display velocity vectors
• Show Trajectories: Show body path trails
• Show Glow: Add glow effect to bodies
• Show Starfield: Display background starfield
• Show Grid: Display reference grid
• Body Size: Adjust body rendering size
• Grid Width: Adjust grid spacing

Scenario Generator

Use the "Scenario Generator" button to create custom scenarios with preset parameters for different initial conditions.

Camera Controls (3D View)

• Mouse Drag: Rotate camera around the simulation
• Scroll Wheel: Zoom in/out
• Right-Click Drag: Pan the view

Testing

Run the physics regression tests:

cd build
ctest

Or run the test directly:

./build/physics_regression

Performance Tips

• Use Barnes-Hut or Fast Multipole for simulations with >1000 bodies
• Verlet integrator provides a good balance of speed and accuracy
• Metal backend (macOS) significantly improves performance for large simulations
• Reduce Time Step for better accuracy with fast-moving bodies
• Lower Number of Bodies if the simulation runs slowly

Architecture

For detailed architecture documentation, see ARCHITECTURE.md.

Project Structure

src/
├── simulation/        # Core physics (Vector3D, Body, Universe)
├── integrators/       # Numerical integrators
├── force_computers/   # Force computation methods
├── rendering/         # Visualization renderers
├── initial_conditions/# Initial condition generators
├── compute/          # Compute backends (CPU, GPU)
└── ui/               # GTK4 user interface

Disclaimer

I vibecoded this.

License

MIT