That is amazing.

Spiffy, but I wonder how much of an effect it would have on your framerates in a real gaming environment? Not just under lab conditions.

In the words of Neo:

whoa

My thoughts, too. Doubt we'll see this level of simulation in an actual game this coming generation. Pretty though

What Water Will Look Like in Videogames of the Future has been added as a related post - related requested by pumkinandstorm on that post.

Would like to see one of the scenes compared to what real water would do in the same container/shape.

Does it really matter? The goal is not to accurately model what water would do, but to look similar enough that it fools the observer.

Well lets see...if it EXACTLY REPLICATES what real water would do then I'd say that would "fool the user". So that should be the goal. So don't kid yourself. The goal is to accurately model what real water would do.

Sure, and with unlimited computing power that's exactly what you'd do.

Meanwhile, back in the real world there are serious computational constraints placed on any real time rendering system.

So, once again, the goal is not to accurately model water, the goal is to look like you're accurately modelling water.

The same thing happens with physics engines in games.

So... simulating the motions/interactions of spheres and then just interpolating their highest/lowest points somehow? (IANA3DP)

If you want to accurately model what real water/fluids would do, you've got to model at the molecular (if not atomic) level. Games just have to fool you, which is a way lower bar.