CUDA并行程序的内存访问优化技术研究

对统一计算设备架构CUDA技术进行了研究,分析了CUDA体系结构及其内存访问机制的显著特点,总结了CUDA并行程序常见的内存访问问题,针对全局内存的非对齐访问和共享内存的访问冲突,提出了相应的内存访问优化策略;最后,利用直方图均衡算法对此优化技术进行了测试,对比了优化前后的程序执行时间;实验结果表明,利用此优化技术可以大大缩短CUDA程序的执行时间,并且图像像素越大,优化效果越好。     

Performance evaluation of CUDA programming for 5-axis machining multi-scale simulation

5-Axis milling simulations in CAM software are mainly used to detect collisions between the tool and the part. They are very limited in terms of surface topography investigations to validate machining strategies as well as machining parameters such as chordal deviation, scallop height and tool feed. Z-buffer or N-buffer machining simulations provide more precise simulations but require long computation time, especially when using realistic cutting tools models including cutting edges geometry. Thus, the aim of this paper is to evaluate Nvidia CUDA architecture to speed-up Z-buffer or N-buffer machining simulations. Several strategies for parallel computing are investigated and compared to single-threaded and multi-threaded CPU, relatively to the complexity of the simulation. Simulations are conducted with two different configurations including Nvidia Quadro 4000 and Geforce GTX 560 graphic cards.     

CUDA架构下的三维弹性静力学边界元并行计算

针对传统边界元法计算量大、计算效率低的问题,以三维弹性静力学的边界元法为对象,将基于CUDA的GPU并行计算应用到其边界元计算中,提出了基于CUDA架构的GPU并行算法.该算法首先对不同类型的边界元系数积分进行并行性分析,描述了相关的GPU并行算法,然后阐述了边界元方程组的求解方法及其并行策略.实验结果表明,文中算法较传统算法具有显著的加速效果.     

锥束CT的FDK算法与CUDA实现

计算机层析成像技术,在医学和工业等诸多领域中有着广泛应用。在三维锥束CT图像重建算法中,基于圆形轨道和二维平板探测器的FDK算法最为著名。传统CPU上实现的FDK算法,计算复杂性主要集中在所谓的反投影阶段,占据了整个重建时间的99%。给出了基于CUDA统一计算架构的FDK算法的GPU实现,对于整个重建过程获得了超过百倍的加速。     

Modified genetic algorithms for solving fuzzy flow shop scheduling problems and their implementation with CUDA

In this paper we propose an improved algorithm to search optimal solutions to the flow shop scheduling problems with fuzzy processing times and fuzzy due dates. A longest common substring method is proposed to combine with the random key method. Numerical simulation shows that longest common substring method combined with rearranging mating method improves the search efficiency of genetic algorithm in this problem. For application in large-sized problems, we also enhance this modified algorithm by CUDA based parallel computation. Numerical experiments show that the performances of the CUDA program on GPU compare favorably to the traditional programs on CPU. Based on the modified algorithm invoking with CUDA scheme, we can search satisfied solutions to the fuzzy flow shop scheduling problems with high performance.     

现代战争模拟研究的新方向：社会仿真

随着世界社会经济和科技水平的不断发展,特别是现代信息技术的发展,社会因素对战争的影响越来越突出,因此,战争模拟应该把视野扩展到传统研究之外的社会领域.主要阐述社会仿真概念的基本内涵,建模仿真基本方法,然后重点介绍社会仿真研究领域的基础设施建模研究、大规模群体行为建模研究以及国内外典型的大规模社会仿真系统,最后,介绍在该领域开展的部分研究工作,并指出我国在该领域展开研究工作的重要意义.     

A review of vision-based systems for soccer video analysis

This paper presents a survey of soccer video analysis systems for different applications: video summarization, provision of augmented information, high-level analysis. Computer vision techniques have been adapted to be applicable in the challenging soccer context. Different semantic levels of interpretation are required according to the complexity of the corresponding applications. For each application area we analyze the computer vision methodologies, their strengths and weaknesses and we investigate whether these approaches can be applied to extensive and real time soccer video analysis.     

一种改进的基于CUDA的纹理映射和光线投射结合的体绘制算法

针对传统的基于GPU的光线投射算法绘制效率较低的问题,利用CUDA架构的并行计算特性和对三维纹理的处理能力进行改进和优化.将体数据映射为三维纹理,利用CUDA三维数组进行存储与绑定,纹理拾取的浮点返回值利用线性滤波进行平滑.在传输函数的设计中引入中心差分梯度幅值增强对体数据边界面的绘制效果.每条光线的求交及颜色积累采用并行计算,按照由前向后进行颜色及不透明度累积.设置不透明度阈值,采用不透明度提前终止加速绘制.实验结果表明,绘制速度较传统的基于GPU算法有10％的速度提升,绘制效果也有很大的改善.     

A comparative study of partitioning methods for crowd simulations

The simulation of large crowds of autonomous agents with realistic behavior is still a challenge for several computer research communities. In order to handle large crowds, some scalable architectures have been proposed. Nevertheless, the effective use of distributed systems requires the use of partitioning methods that can properly distribute the workload generated by agents among the existing distributed resources.In this paper, we analyze the use of irregular shape regions (convex hulls) for solving the partitioning problem. We have compared a partitioning method based on convex hulls with two techniques that use rectangular regions. The performance evaluation results show that the convex hull method outperforms the rest of the considered methods in terms of both fitness function values and execution times, regardless of the movement pattern followed by the agents. These results show that the shape of the regions in the partition can improve the performance of the partitioning method, rather than the heuristic method used.     

基于物理的真实感火灾疏散仿真

人员的疏散是目前防火防灾研究的热点,但是传统的人员疏散研究多集中于人群疏散的离散仿真及可视化显示,无法逼真地表现个体的疏散行为.从火灾时人员疏散的物理特点出发,考虑不同人在火灾情况下的行为特征,采取几何连续的疏散空间,可较准确地计算人员安全疏散的时间及路径;建立了建筑物、火焰和角色动画的三维模型,动态交互地模拟火灾疏散过程,仿真效果逼真.     

Statistical analysis and modeling of the temperature-dependent sleep behavior of drosophila

The sleep behavior of drosophila is analyzed under different temperatures. The activity per minute of the flies is recorded automatically. Sleep for a fruit fly is defined as the periods without any activity and longer than 5 minutes. Several parameters such as total sleep time, circadian sleep profile, quality of sleep are analyzed. The sleep behaviors are significantly different for flies at different temperature. Interestingly, the durations of daytime sleep periods show a common scale-free power law distribution. We propose a stochastic model to simulate the activities of the population of neurons which regulate the dynamics of sleep–wake process to explain the distribution of daytime sleep.     

Virtual humans: thirty years of research, what next?

In this paper, we present research results and future challenges in creating realistic and believable Virtual Humans. To realize these modeling goals, real-time realistic representation is essential, but we also need interactive and perceptive Virtual Humans to populate the Virtual Worlds. Three levels of modeling should be considered to create these believable Virtual Humans: 1) realistic appearance modeling, 2) realistic, smooth and flexible motion modeling, and 3) realistic high-level behaviors modeling. At first, the issues of creating virtual humans with better skeleton and realistic deformable bodies are illustrated. To give a level of believable behavior, challenges are laid on generating on the fly flexible motion and complex behaviours of Virtual Humans inside their environments using a realistic perception of the environment. Interactivity and group behaviours are also important parameters to create believable Virtual Humans which have challenges in creating believable relationship between real and virtual humans based on emotion and personality, and simulating realistic and believable behaviors of groups and crowds. Finally, issues in generating realistic virtual clothed and haired people are presented.     

Theory and techniques of data mining in CGF behavior modeling

For the large quantity of data,rules and models generated in the course of computer generated forces (CGFs) behavior modeling,the common analytical methods are statistical methods based on the tactical rules,tactical doctrine and empirical knowledge.However,from the viewpoint of data mining,we can find many of these analytical methods are also each-and-every different data mining methods.In this paper,we survey the data mining theory and techniques that have appeared in the course of CGF behavior modeling f...     

硅橡胶材料填充改性的相行为模拟研究

结合推广的Flory-Huggins理论和Monte carlo方法,模拟了由气相白炭黑和改性气相白炭黑填充的硅橡胶的相行为,获得了填料与聚硅氧烷的相互作用能分布曲线,配位数及相图。相图和热力学分析表明,经六甲基二硅氮烷改性的气相白炭黑与聚硅氧烷的相容性和共混体系的稳定性得以明显提高。     

用游戏的方式强化传播力的可能性

游戏是引发行为交流的一种比较好的方式,其良好的娱乐性能抵消人们对于大量信息的抵触心里。信息传播是视觉传达设计的最基本职能之一,从信息传播方式入手,实验用游戏的策略来引导行为交流的产生,并引入到视觉传达设计中去。以促成特定环境内人与人之间的交流,在娱乐的气氛中传播信息,从而增强信息传播过程的趣味性,实现强化传播力的效果。     

RTS游戏中用户行为的神经网络预测模型

即时战略游戏(简称RTS游戏)中,用户的行为由于游戏自身庞大的决策空间而难以预测.针对这个问题,提出了通过对RTS游戏的对战记录进行分析,建立5种结构的神经网络模型来预测用户行为的方法.模型考虑了不同时间片的状态对于决策行为的影响,设计了单时间片输入和双时间片输入的神经网络,并与基于动态贝叶斯网络的模型进行了比较.实验结果表明,基于单时间片输入的神经网络模型能够更加快速地完成训练过程并达到满意的预测准确度.     

Discrete kinetic theory for 2D modeling of a moving crowd: Application to the evacuation of a non-connected bounded domain

This paper concerns the mathematical modeling of the motion of a crowd in a non connected bounded domain, based on kinetic and stochastic game theories. The proposed model is a mesoscopic probabilistic approach that retains features obtained from both micro- and macro-scale representations; pedestrian interactions with various obstacles being managed from a probabilistic perspective. A proof of the existence and uniqueness of the proposed mathematical model’s solution is given for large times. A numerical resolution scheme based on the splitting method is implemented and then applied to crowd evacuation in a non connected bounded domain with one rectangular obstacle. The evacuation time of the room is then calculated by our technique, according to the dimensions and position of a square-shaped obstacle, and finally compared to the time obtained by a deterministic approach by means of randomly varying some of its parameters.     

A grid-based spatial data model for the simulation and analysis of individual behaviours in micro-spatial environments

As crowd simulation in micro-spatial environment is more widely applied in urban planning and management, the construction of an appropriate spatial data model that supports such applications becomes essential. To address the requirements necessary to building a model of crowd simulation and people–place relationship analysis in micro-spatial environments, the concept of the grid as a basic unit of people–place data association is presented in this article. Subsequently, a grid-based spatial data model is developed for modelling spatial data using Geographic Information System (GIS). The application of the model for crowd simulations in indoor and outdoor spatial environments is described. There are four advantages of this model: first, both the geometrical characteristics of geographic entities and behaviour characteristics of individuals within micro-spatial environments are involved; second, the object-oriented model and spatial topological relationships are fused; third, the integrated expression of indoor and outdoor environments can be realised; and fourth, crowd simulation models, such as Multi-agent System (MAS) and Cellular Automata (CA), can be further fused for intelligent simulation and the analysis of individual behaviours. Lastly, this article presents an experimental implementation of the data model, individual behaviours are simulated and analysed to illustrate the potential of the proposed model.