Parallel Methods for Real-Time Visualization of Snow

Snow is a familiar scene in the Nordic countries during the winter months. This paper discusses some of the complex numerical algorithms behind snow simulations. Previous methods for snow simulation have either covered only a very limited aspect of snow, or have been unsuitable for real-time performance. Here, some of these methods are combined into a model for real-time snow simulation that handles snowflake motion through the air, wind simulation, as well as accumulation of snow on objects including the ground. With our goal towards achieving realtime performance with more than 25 frames per second, some new parallel methods for the snow model are introduced. The algorithms are first parallelized by dividing the data structures among threads. This scheme is then improved by overlapping inherently sequential algorithms with computations for the following frame, to eliminate processor idle time. SMP and multi-core systems are considered.