ABSTRACT
Real-time simulation of fluid dynamics has been around for a while now, but it has not made its way into many games because of its performance characteristics, which have never been at a level acceptable enough to be deemed “game-ready.” In game development, there is this imaginary scale used to determine whether or not a piece of visual effects (VFX) will make it into a game: on one end of this scale is ultra-high visual quality-which is used to describe the purest of physically based effects-and at the other end is low-quality/performance. All real-time VFX are subject to this scale, and all lie somewhere around the middle of the two extremes; this can be thought of as the performance/quality tradeoff. When it is desirable to take an effect from the physically based perfection side over to the performance (game-ready) side, then some work has to be done in order to significantly reduce the cost of that effect while still maintaining as much quality as possible.
