1 gorilla vs 100 ships

Extremely performant! I wonder how many substeps are needed? I’m curious on how you make the simulation stable with a large timestep.

This is single timestep. There is a 1/4 res coarse grid to help with pressure and the ships have clusters using position based shape matching.

Real ones know: Sumo Gummies hit different! 🔥 Hear it straight from the squad—10,000MG of THCA in every jar. Who’s stepping into the ring next? 💪 🎥 Watch the experience 👇 Drop your Sumo survival stories! #SumoGummies #HALFBAKD #THCA

impressive

Amazing! do you use some engine ?

Thanks! This is mostly custom, I was using DirectXTK12 which is a bit of a lower level helper library. I hope to bring this to the web soon with WebAssembly and WebGPU

So cool and performant very impressive

stupido .i.

Is the water body simulation used here?use the previous grid density and velocity gradient determinant to guess the future density. This fuses 2 P2Gs into 1，How is this achieved?

Is the mpm kernel function here a cubic linear difference or is it more complicated?