Title:Adaptive Position-Based Fluids: Improving Performance of Fluid Simulations for Real-Time Applications

Abstract:The Position Based Fluids (PBF) method is a state-of-the-art approach for fluid simulations in the context of real-time applications like games. It uses an iterative solver concept that tries to maintain a constant fluid density (incompressibility) to realize incompressible fluids like water. However, larger fluid volumes that consist of several hundred thousand particles (e.g. for the simulation of oceans) require many iterations and a lot of simulation power. We present a lightweight and easy-to-integrate extension to PBF that adaptively adjusts the number of solver iterations on a fine-grained basis. Using a novel adaptive-simulation approach, we are able to achieve significant improvements in performance on our evaluation scenarios while maintaining high-quality results in terms of visualization quality, which makes it a perfect choice for game developers. Furthermore, our method does not weaken the advantages of prior work and seamlessly integrates into other position-based methods for physically-based simulations.