The texture-based volume rendering is a memory-intensive algorithm. Its performance relies heavily on the performance of the texture cache. However, most existing texture-based volume rendering methods blindly map computational resources to texture memory and result in incoherent memory access patterns, causing low cache hit rates in certain cases. The distance between samples taken by threads of an atomic scheduling unit (e.g. a warp of 32 threads in CUDA) of the GPU is a crucial factor that affects the texture cache performance. Based on this fact, we present a new sampling strategy, called Warp Marching, for the ray-casting algorithm of texture-based volume rendering. The effects of different sample organizations and different thread-pixel mappings in the ray-casting algorithm are thoroughly analyzed. Also, a pipeline manner color blending approach is introduced and the power of warp-level GPU operations is leveraged to improve the efficiency of parallel executions on the GPU. In addition, the rendering performance of the Warp Marching is view-independent, and it outperforms existing empty space skipping techniques in scenarios that need to render large dynamic volumes in a low resolution image. Through a series of micro-benchmarking and real-life data experiments, we rigorously analyze our sampling strategies and demonstrate significant performance enhancements over existing sampling methods.

Ray casting algorithm is a major component of the direct volume rendering, which exhibits inherent parallelism, making it suitable for graphics processing units (GPUs). However, blindly mapping the ray casting algorithm on a GPU's complex parallel architecture can result in a magnitude of performance loss. In this paper, a novel computation-to-core mapping strategy, called Warp Marching, for the texture-based iso-surface volume rendering is introduced. We evaluate and compare this new strategy with the most commonly used existing mapping strategy. Texture cache performance and load balancing are the two major evaluation factors since they have significant consequences on the overall rendering performance. Through a series of real-life data experiments, we conclude that the texture cache performances of these two computation-to-core mapping strategies are significantly affected by the viewing direction; and the Warp Marching performs better in balancing workloads among threads and concurrent hardware components of a GPU.

In many research domains, such as mechanical engineering, game development and virtual reality, a typical output model is usually produced in a multi-object manner for efficient data management. To have a complete description of the whole model, tens of thousands of, or even millions of, such objects are created, which makes the entire dataset exceptionally complex. Consequently, visualizing the model becomes a computationally intensive process that impedes a real-time rendering and interaction.

ABSTRACT We present a crowd model informed by common ground theory to accommodate high‐level socially aware behavioral realism of characters in crowd simulations. In our approach, group members maintain group cohesiveness by communicating and adapting their behaviors to each other. The resulting character behaviors in animations form a consequential chain interpreted as a coherent story by observers. We demonstrate that our model produces more believable animations from the viewpoint of human observers through a series of user studies. Copyright © 2013 John Wiley & Sons, Ltd.

Large crowds are seldom made up solely of a mass of individuals. They typically also include large collections of small groups. However the most of existing approaches to crowd modeling treat a crowd either as a collection of isolated individuals, each maintaining its own goal, or as an aggregated entity in which large number of individuals share the same goal and behavior pattern.

Inaccurate perception of distances is a known problem within virtual environments. We hypothesize that the inclusion of virtual characters within these environments can improve an observer's ability to judge distances and achieve accurate spatial understanding. We have conducted an empirical study using a desktop display of a small scale environment to evaluate the validity of this concept. We investigated whether the presence and quantity of virtual human characters, as well as the naturalness of their locomotion animations, could improve egocentric and exocentric distance estimations. Preliminary results suggest that static or properly animated characters could improve exocentric estimations, and properly animated characters could reduce egocentric distance compression errors.

Abstract Rendering massive 3D models in real‐time has long been recognized as a very challenging problem because of the limited computational power and memory space available in a workstation. Most existing rendering techniques, especially level of detail (LOD) processing, have suffered from their sequential execution natures. We present a GPU‐based approach which enables interactive rendering of large 3D models with hundreds of millions of triangles. Our work contributes to the massive rendering research in two ways. First, we present a simple and efficient mesh simplification algorithm towards GPU architecture. Second, we propose a novel GPU out‐of‐core approach that adopts a frame‐to‐frame coherence scheme in order to minimize the high communication cost between CPU and GPU. Our results show that the parallel algorithm of mesh simplification and the GPU out‐of‐core approach significantly improve the overall rendering performance.

In volume rendering, most optical models currently in use are based on the assumptions that a volumetric object is a collection of particles and that the macro behavior of particles, when they interact with light rays, can be predicted based on the behavior of each individual particle. However, such models are not capable of characterizing the collective optical effect of a collection of particles which dominates the appearance of the boundaries of dense objects. In this paper, we propose a generalized optical model that combines particle elements and surface elements together to characterize both the behavior of individual particles and the collective effect of particles. The framework based on a new model provides a more powerful and flexible tool for hybrid rendering of isosurfaces and transparent clouds of particles in a single scene. It also provides a more rational basis for shading, so the problem of normal-based shading in homogeneous regions encountered in conventional volume rendering can be easily avoided. The model can be seen as an extension to the classical model. It can be implemented easily, and most of the advanced numerical estimation methods previously developed specifically for the particle-based optical model, such as preintegration, can be applied to the new model to achieve high-quality rendering results.

Through the algorithmic design patterns of data parallelism and task parallelism, the graphics processing unit (GPU) offers the potential to vastly accelerate discovery and innovation across a multitude of disciplines. For example, the exponential growth in data volume now presents an obstacle for high-throughput data mining in fields such as neuroscience and bioinformatics. As such, we present a characterization of a MapReduced-based data-mining application on a general-purpose GPU (GPGPU). Using neuroscience as the application vehicle, the results of our multi-dimensional performance evaluation show that a ldquoone-size-fits-allrdquo approach maps poorly across different GPGPU cards. Rather, a high-performance implementation on the GPGPU should factor in the 1) problem size, 2) type of GPU, 3) type of algorithm, and 4) data-access method when determining the type and level of parallelism. To guide the GPGPU programmer towards optimal performance within such a broad design space, we provide eight general performance characterizations of our data-mining application.

Temporal data mining algorithms are becoming increasingly important in many application domains including computational neuroscience, especially the analysis of spike train data. While application scientists have been able to readily gather multi-neuronal datasets, analysis capabilities have lagged behind, due to both lack of powerful algorithms and inaccessibility to powerful hardware platforms. The advent of GPU architectures such as Nvidia's GTX 280 offers a cost-effective option to bring these capabilities to the neuroscientist's desktop. Rather than port existing algorithms onto this architecture, we advocate the need for algorithm transformation, i.e., rethinking the design of the algorithm in a way that need not necessarily mirror its serial implementation strictly. We present a novel implementation of a frequent episode discovery algorithm by revisiting 'in-the-large' issues such as problem decomposition as well as 'in-the-small' issues such as data layouts and memory access patterns. This is non-trivial because frequent episode discovery does not lend itself to GPU-friendly data-parallel mapping strategies. Applications to many datasets and comparisons to CPU as well as prior GPU implementations showcase the advantages of our approach.