Repository navigation
☕ CappuSim v1.0 - Official Release
We are excited to announce the v1.0 release of CappuSim, a high-performance, modular, and extensible Lattice Boltzmann Method (LBM) simulator built with Rust and OpenCL.
After a cycle of rigorous development and validation, CappuSim stands as a robust, accurate, and efficient tool for Computational Fluid Dynamics (CFD) simulations. This version provides a stable platform for research, development, and practical applications in aerodynamics and other fluid flow phenomena.
Updates in v1.0 - Official Release
-
Drastically improved overall performance
-
Bug fixes
-
Added Precision selection
-
Constante External Force
-
Validated against classical CFD scenarios
-
Improved Built-in Benchmark Tool
-
Supports
D2Q9,D3Q7,D3Q15,D3Q19,D3Q27models -
Simulation Examples
- Couette Flow
- Poiseuille Flow
- Lid-Driven Cavity
- Von Kármán Vortex Street
- Airfoil
-
Output Formats
.csv.vtkfor visualization in ParaView- Exports physical variables
-
Benchmark Ready
- Standard Benchmark example
- MLUPs calculation
- Scalable to large 3D grids
--
Getting Started
git clone https://github.com/gustavoverneck/CappuSim
cd CappuSim
cargo build --release
cargo run --release🐞 Bug Reports & Feedback
This is the first stable release, but the work is ongoing. If you encounter any bugs, have suggestions for new features, or want to contribute, your feedback is highly welcome!
Please open an issue on the GitHub repository's Issues tab to report any problems.
Roadmap (no promises)
- Thermal Flows: Implementing a second distribution function to simulate heat transfer and natural convection (via the Boussinesq approximation).
- Advanced Collision Models: Integrating MRT (Multiple-Relaxation-Time) and TRT (Two-Relaxation-Time) for enhanced stability.
- Multiphase Fluids: Adding capabilities to simulate interactions between multiple fluids (e.g., Shan-Chen model).
- GUI: Developing a simple graphical interface for easier setup and real-time visualization.
- Better solids input
Notes
- Currently tested with OpenCL 1.2+ on NVIDIA and AMD GPUs.
- Focused on correctness, modularity, and ease of experimentation.