Pseudo-fluid Simulation

[maximile]

http://www.youtube.com/watch?v=V1xiVyJ3EOM

No crazy-cool techniques or anything here, just lots of circles and time. And hardly any friction. Thought you might like to see it.

BTW, it's *not* real time. About 5 or 6 seconds per frame at its slowest. Perhaps not helped by World of Warcraft running in the background, though.

[lucas]

That is very, very awesome. How big are the circles?

[slembcke]

Yeah, I've tried something similar. I was definitely impressed that it worked at all to be honest. Simulating water out of little solid circles just doesn't seem like it would work, but it definitely does a pretty good job.

Though it does make me wonder how fast you could get such a thing to run if you simplified the collision detection, and simulated them only as solid particles using vertlet or something. Could be fun to try sometime.

[maximile]

The smallest circles are about 0.05 radius, which with my setup corresponds to about 1 pixel. I realise now that I could probably have sped it up if I had set the

Code: Select all

 cpSpaceResizeActiveHash()

accordingly. Oh well.

[Ralith]

I'd be really interested to see if this could be optimized to provide a fairly realistic realtime 2D water system.