View-dependent level of detail for the parallel rendering of complex models

The ever increasing complexity of 3D models generated by 3D scanners and through modeling requires ever faster graphics cards to render them at interactive rates. Unfortunately, the rate at which the performance of graphics cards increases is outpaced by the advances in scanning technology, making today's 3D models too large to render interactively. Because of the sheer amount of geometry to render, datasets such as the USGS Earth model are also too large to fit into memory, making the rendering even more challenging.; In this dissertation we present a new multi-resolution, multi-grained level of detail hierarchy along with new algorithms used to traverse this hierarchy. The new hierarchy builds on previous work by permitting any element in the hierarchy two degrees of freedom as opposed to a single degree for previous methods, permitting a more flexible traversal of the hierarchy.; The adaptability of the system is especially useful in the current computing environment, with its ever-increasing variety of options in terms of number and power of CPUs and GPUs. Because any two computers can have any combination of both CPU and GPU, our method allows the hierarchy to be reused and adapted at runtime rather then requiring a time-consuming re-computation of the data.; The Level of Detail system presented in this dissertation also adds new capabilities to single-computer rendering in the form of support for both multi-CPU and multi-GPU architectures, achieving considerable performance improvements by utilizing the parallelism available in the latest generation of computers.; The design of the system also makes it capable of rendering to a tiled display wall using a rendering cluster. The adaptive hierarchy and regular grid layout used in our method allows the entire cluster to use the same hierarchy, regardless of the load on the machine or the machine configuration.; We demonstrate our approach on both large triangle meshes and terrains with up to billions of vertices.