SPH-Fluid

In our final project for CSE6230: High Performance Parallel Computing - Tools and Applications, we implemented a Smoothed Particle Hydrodynamics (SPH) fluid simulation leveraging both CUDA and Taichi technologies. Our project features the Weakly Compressible SPH (WCSPH) model, a rapid fixed-radius neighbor search optimized for GPU performance, and a lockless ring buffer that facilitates concurrent operations between the CPU and GPU.

Result

References

Authors | Smoothed Particle Hydrodynamics Techniques for the Physics Based Simulation of Fluids and Solids (physics-simulation.org)

S4117-fast-fixed-radius-nearest-neighbor-gpu.pdf (gputechconf.com)

taichiCourse01 (TaichiCourse) (github.com)

11. Ring Library — Data Plane Development Kit 24.03.0 documentation (dpdk.org)

A Parallel SPH Implementation on Multi‐Core CPUs - Ihmsen - 2011 - Computer Graphics Forum - Wiley Online Library