Theory Accelerated, led by Matt Puchala, former TD at ILM and Naughty Dog, has released Paradigm 1.0. This GPU-accelerated sparse liquid solver promises to be faster than Houdini's native tools. Unlike their other product, Axiom (focused on smoke and fire), Paradigm focuses exclusively on liquids, offering an efficient alternative for complex simulations.
FLIP Technology on GPU with Multi-Architecture Support 🚀
Paradigm is a fully GPU-based FLIP solver using OpenCL, compatible with AMD, Intel, NVIDIA GPUs, and Apple Silicon processors. It allows simulating millions of particles in minutes, reducing hours of work to minutes. Built on Axiom's technology, it includes tools for small-scale simulations like pouring and splashing, with support for surface tension, viscosity, and adhesion. When exporting, it can mesh particles into an SDF surface. For open water, it includes the Paradigm Tank node, although it lacks a full wave solver.
The Liquid Revolution: Now Your GPU Works While You Drink Coffee ☕
Finally, visual effects artists can simulate millions of particles in the time it used to take just to brew coffee. Paradigm promises to turn hours of computation into minutes, though that might mean less time for technical excuses. The curious thing is that, despite being separate products, they share technology with Axiom. So if your smoke was already fast, get ready for your liquids to be too. Now we just need someone to develop a solver for TDs' dark circles. 😄