PySPH C++ Port

A C++ port of the Python pysph screen-space fluid rendering demo.

Features

• SPH Fluid Simulation - Smoothed Particle Hydrodynamics with CUDA

  • Grid-based neighbor search using spatial hashing
  • Thrust-accelerated sorting
  • M4 (cubic spline) kernel for density estimation
  • Pressure and viscosity forces

• Screen-Space Fluid Rendering (Simon Green's technique)

  • Three render modes: Points, Balls, Advanced
  • Depth map with curvature flow smoothing
  • Thickness map with Gaussian blur
  • Final compositing with Beer's law absorption and Fresnel specular

Requirements

• Windows 10/11
• Visual Studio 2019 or 2022
• CUDA 12.x (with CUDA build customizations for Visual Studio)
• vcpkg with the following packages:
  • freeglut:x64-windows-static
  • glew:x64-windows-static
  • imgui[freeglut-binding,opengl3-binding]:x64-windows-static

Building

1. Install vcpkg dependencies:

   cd ../vcpkg
   ./vcpkg install freeglut:x64-windows-static glew:x64-windows-static imgui[freeglut-binding,opengl3-binding]:x64-windows-static

2. Open pysph-cpp.sln in Visual Studio

3. Build (Release x64 recommended for performance)

Usage

pysph-cpp.exe [num_particles]

Default is 8000 particles. Start with a lower number and increase based on your GPU.

Controls

• Left mouse drag: Rotate view
• Right mouse drag: Translate view
• Mouse wheel: Zoom
• H: Toggle UI
• Space: Pause/Resume simulation
• M: Cycle render mode (Points → Balls → Advanced)
• R: Reset simulation
• ESC: Exit

Render Modes

1. Points: Simple GL_POINTS rendering
2. Balls: Point sprites rendered as lit spheres
3. Advanced: Full screen-space fluid rendering with:
   • Depth smoothing (curvature flow)
   • Thickness-based transparency
   • Beer's law color absorption (water-like blue)
   • Fresnel reflections

Parameters (in Advanced mode)

• Smooth Depth: Enable/disable curvature flow smoothing
• Blur Thickness: Enable/disable Gaussian blur on thickness map
• Smoothing Iterations: Number of curvature flow iterations (0-100)
• Z Contribution: Adaptive smoothing factor based on depth gradient

Credits

• Original Python implementation: [pysph project]
• Screen-space rendering technique: Simon Green (NVIDIA)
• SPH theory: Matthias Müller et al.