FluidX3D - Simulating Rayleigh-Bénard Convection in Real Time and 3D on a GPU
Rayleigh-Bénard convection occurs for example in a hot pan filled with oil or in the Earths mantle, whenever a vertical temperature gradient is present and the parameters like viscosity, thermal conductivity and the thermal expansion coefficient are just right. At the bottom the fluid is heated, expands, becomes less dense and thus rises. When it reaches the cold top plate, it becomes more dense again and sinks to the bottom. In 3D this results in convection cells.
Performing such 3D simulations in real time has not been possible before the rise of GPUs. Traditional LBM codes running on CPUs are approximately 100x slower - far away from real time.
FluidX3D is a real time 3D fluid simulation based using the lattice Boltzmann method. It is written in OpenCL C (GPU code) and optimized to the physical limit (video memory bandwith, ~520GB/s).
All demonstrations but the last one are simulated and visualized in real time on a Nvidia Titan Xp. Graphics are done with the OpenCL C version of Line3D, the fastest graph
2 views
275
69
5 months ago 00:00:11 1
Aerodynamics of a cow
12 months ago 00:01:00 1
Bell 222 Helicopter in FluidX3D CFD - 10 Billion Cells, 71TB vizualized - 6.4 hours on 8x MI200 GPUs
1 year ago 00:04:59 1
Prigozhin’s Aircraft Crash Analysis
1 year ago 00:00:50 58
FluidX3D on GigaIO SuperNODE - Concorde 40 Billion Cell CFD Simulation - 33h on 32x MI210 64GB GPUs
2 years ago 00:01:01 1
Visualizing 144 TeraByte of CFD data for fun
2 years ago 00:02:00 43
X-wing - largest single-GPU CFD simulation ever done (AMD MI250)
2 years ago 00:01:00 48
How aerodynamic is Boeing 747 at Re=50k? (Nvidia A100)
3 years ago 00:02:20 2
FluidX3D - Simulating Rayleigh-Bénard Convection in Real Time and 3D on a GPU