Graphics processing units (GPUs) have been revolutionizing the way computer graphics and visualization are practiced. Driven by the computer-games industry and its demand for efficient hardware support for 3D graphics, GPUs have dramatically increased in performance and functionality within only a few years. Although graphics hardware is primarily designed for the fast rendering of 3D scenes, it can also be used for other types of computations. In fact, GPUs have evolved to programmable processors that can facilitate applications beyond traditional real-time 3D rendering.
This book addresses scientific visualization as one application area that significantly benefits from the use of GPUs. In general, scientific visualization has become an important tool for visual analysis in many scientific, engineering, and medical disciplines. For example, scientists and engineers regularly use visualization to interpret simulations of air or water flow in computational fluid dynamics. Another example is the medical imaging of 3D CT (computer tomography) or MRI (magnetic resonance imaging) scans. The interactive exploration of data sets is becoming increasingly more important with the growing amount and complexity of those data sets: the user – as an expert in his or her field – uses visualization as a tool to investigate the data and extract insight from it.
This book focuses on efficient visualization techniques, which are the prerequisite for interactive exploration. High performance is primarily achieved by algorithms specifically designed for GPUs and their special features. Other aspects discussed in this work include parallelization on cluster computers with several GPUs, adaptive rendering methods, multi-resolution models, and non-photorealistic rendering techniques for visualization. This book also addresses the effectiveness of visualization methods, which can be improved by taking into account perceptual aspects and user interaction. Covering both the theoretical foundations and practical implementations of algorithms, this text provides a basis to understand and reproduce modern GPU-based visualization approaches.