Abstract

Alle Ausführungen, die wörtlich oder sinngemäß übernommen wurden, sind als solche gekennzeichnet. In any large-scale simulation that is dealing with a huge number of objects, the broad-phase collision detection becomes a crucial part for achieving high performance. The focus of this paper will be on developing a time-efficient as well as memory-efficient implementation of a data structure called hierarchical hash grid, which is especially well-suited for coarse collision detection in simulations involving very large numbers of objects. As part of the pe physics engine, the hierarchical hash grid enabled the realization of a massively parallel rigid body simulation involving more than one billion objects (1.14 · 10 9). Further key features of the presented implementation are not only an average-case computational complexity of order O(N) as well as a space complexity of order O(N), but also a short actual runtime combined with low memory consumption. Moreover, the hierarchical hash grid has the ability to, at runtime, automatically and perfectly adapt its data structures to the objects of the underlying simulation. Also, arbitrary objects can be added and removed in constant time, O(1). In the course of this paper, the necessary theoretical background