A combination of the lattice Boltzmann method and the most recently developed dynamic mode decomposition is proposed for stability analysis. The simulations are performed on a graphical processing unit. Stability of t...A combination of the lattice Boltzmann method and the most recently developed dynamic mode decomposition is proposed for stability analysis. The simulations are performed on a graphical processing unit. Stability of the flow past a cylinder at supercritical state, Re = 50, is studied by the combination for both the exponential growing and the limit cycle regimes. The Ritz values, energy spectrum, and modes for both regimes are presented and compared with the Koopman eigenvalues. For harmonic-like periodic flow in the limit cycle, global analysis from the combination gives the same results as those from the Koopman analysis. For transient flow as in the exponential growth regime, the combination can provide more reasonable results. It is demonstrated that the combination of the lattice Boltzmann method and the dynamic mode decomposition is powerful and can be used for stability analysis for more complex flows.展开更多
在图形处理器(GPU)上实现对角稀疏矩阵向量乘法(SpMV)可以充分利用GPU的并行计算能力,并加速矩阵向量乘法;然而,相关主流算法存在零元填充数据多、计算效率低的问题。针对上述问题,提出一种对角SpMV算法DIA-Dynamic(DIAgonal-Dynamic)...在图形处理器(GPU)上实现对角稀疏矩阵向量乘法(SpMV)可以充分利用GPU的并行计算能力,并加速矩阵向量乘法;然而,相关主流算法存在零元填充数据多、计算效率低的问题。针对上述问题,提出一种对角SpMV算法DIA-Dynamic(DIAgonal-Dynamic)。首先,设计一种全新的动态划分策略,根据矩阵的不同特征进行分块,在保证GPU高计算效率的同时大幅减少零元填充,去除冗余计算量;其次,提出一种对角稀疏矩阵存储格式BDIA(Block DIAgonal)存储分块数据,并调整数据布局,提高GPU上的访存性能;最后,基于GPU的底层进行条件分支优化,以减少分支判断,并使用动态共享内存解决向量的不规则访问问题。DIA-Dynamic与前沿Tile SpMV算法相比,平均加速比达到了1.88;与前沿BRCSD(Diagonal Compressed Storage based on Row-Blocks)-Ⅱ算法相比,平均零元填充减少了43%,平均加速比达到了1.70。实验结果表明,DIA-Dynamic能够有效提高GPU上对角SpMV的计算效率,缩短计算时间,提升程序性能。展开更多
为解决Web端建筑信息模型(BIM,Building Information Modeling)场景数据加载技术面向复杂模型数据时存在的加载时间过长、用户体验不佳的问题,提出一种Web端基于工业基础类(IFC,Industry Foundation Classes)标准的面向需求的动态加载...为解决Web端建筑信息模型(BIM,Building Information Modeling)场景数据加载技术面向复杂模型数据时存在的加载时间过长、用户体验不佳的问题,提出一种Web端基于工业基础类(IFC,Industry Foundation Classes)标准的面向需求的动态加载方法。以IFC模型文件作为研究对象,在遵从建筑语义前提下,以建筑构件为粒度,将层次关系、几何特性、材质、属性等信息拆分存储;结合构件可见性和几何相关性,设计出基于图形处理器(GPU,Graphics Processing Unit)加速的面向需求的动态加载方法;搭建实验测试环境,选取若干IFC模型文件,进行方法验证。以初始加载构件个数、内存占用和初始加载时间作为性能评价指标,与使用BIMServer开源服务器平台加载的方法相比,文章所提方法的初始加载组件数量减少了约71%,内存占用减少了约40%,初始加载时间缩短了约78%,有效减少了用户因加载而等待的时间,改善了用户交互体验,可为铁路行业开展Web端BIM大场景应用提供快速加载技术支持。展开更多
Lattice-Boltzmann methods are versatile numerical modeling techniques capable of reproducing a wide variety of fluid-mechanical behavior.These methods are well suited to parallel implementation,particularly on the sin...Lattice-Boltzmann methods are versatile numerical modeling techniques capable of reproducing a wide variety of fluid-mechanical behavior.These methods are well suited to parallel implementation,particularly on the single-instruction multiple data(SIMD)parallel processing environments found in computer graphics processing units(GPUs).Although recent programming tools dramatically improve the ease with which GPUbased applications can be written,the programming environment still lacks the flexibility available to more traditional CPU programs.In particular,it may be difficult to develop modular and extensible programs that require variable on-device functionality with current GPU architectures.This paper describes a process of automatic code generation that overcomes these difficulties for lattice-Boltzmann simulations.It details the development of GPU-based modules for an extensible lattice-Boltzmann simulation package-LBHydra.The performance of the automatically generated code is compared to equivalent purpose written codes for both single-phase,multiphase,and multicomponent flows.The flexibility of the new method is demonstrated by simulating a rising,dissolving droplet moving through a porous medium with user generated lattice-Boltzmann models and subroutines.展开更多
文摘A combination of the lattice Boltzmann method and the most recently developed dynamic mode decomposition is proposed for stability analysis. The simulations are performed on a graphical processing unit. Stability of the flow past a cylinder at supercritical state, Re = 50, is studied by the combination for both the exponential growing and the limit cycle regimes. The Ritz values, energy spectrum, and modes for both regimes are presented and compared with the Koopman eigenvalues. For harmonic-like periodic flow in the limit cycle, global analysis from the combination gives the same results as those from the Koopman analysis. For transient flow as in the exponential growth regime, the combination can provide more reasonable results. It is demonstrated that the combination of the lattice Boltzmann method and the dynamic mode decomposition is powerful and can be used for stability analysis for more complex flows.
文摘在图形处理器(GPU)上实现对角稀疏矩阵向量乘法(SpMV)可以充分利用GPU的并行计算能力,并加速矩阵向量乘法;然而,相关主流算法存在零元填充数据多、计算效率低的问题。针对上述问题,提出一种对角SpMV算法DIA-Dynamic(DIAgonal-Dynamic)。首先,设计一种全新的动态划分策略,根据矩阵的不同特征进行分块,在保证GPU高计算效率的同时大幅减少零元填充,去除冗余计算量;其次,提出一种对角稀疏矩阵存储格式BDIA(Block DIAgonal)存储分块数据,并调整数据布局,提高GPU上的访存性能;最后,基于GPU的底层进行条件分支优化,以减少分支判断,并使用动态共享内存解决向量的不规则访问问题。DIA-Dynamic与前沿Tile SpMV算法相比,平均加速比达到了1.88;与前沿BRCSD(Diagonal Compressed Storage based on Row-Blocks)-Ⅱ算法相比,平均零元填充减少了43%,平均加速比达到了1.70。实验结果表明,DIA-Dynamic能够有效提高GPU上对角SpMV的计算效率,缩短计算时间,提升程序性能。
文摘为解决Web端建筑信息模型(BIM,Building Information Modeling)场景数据加载技术面向复杂模型数据时存在的加载时间过长、用户体验不佳的问题,提出一种Web端基于工业基础类(IFC,Industry Foundation Classes)标准的面向需求的动态加载方法。以IFC模型文件作为研究对象,在遵从建筑语义前提下,以建筑构件为粒度,将层次关系、几何特性、材质、属性等信息拆分存储;结合构件可见性和几何相关性,设计出基于图形处理器(GPU,Graphics Processing Unit)加速的面向需求的动态加载方法;搭建实验测试环境,选取若干IFC模型文件,进行方法验证。以初始加载构件个数、内存占用和初始加载时间作为性能评价指标,与使用BIMServer开源服务器平台加载的方法相比,文章所提方法的初始加载组件数量减少了约71%,内存占用减少了约40%,初始加载时间缩短了约78%,有效减少了用户因加载而等待的时间,改善了用户交互体验,可为铁路行业开展Web端BIM大场景应用提供快速加载技术支持。
基金performed under the auspices of the U.S.Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344support by the National Science Foundation(NSF)under Grant No.DMS-0724560,Grant No.EAR-0838541,and Grant No.EAR-0941666the Department of Energy(DOE)under Grant No.DE-EE0002764.
文摘Lattice-Boltzmann methods are versatile numerical modeling techniques capable of reproducing a wide variety of fluid-mechanical behavior.These methods are well suited to parallel implementation,particularly on the single-instruction multiple data(SIMD)parallel processing environments found in computer graphics processing units(GPUs).Although recent programming tools dramatically improve the ease with which GPUbased applications can be written,the programming environment still lacks the flexibility available to more traditional CPU programs.In particular,it may be difficult to develop modular and extensible programs that require variable on-device functionality with current GPU architectures.This paper describes a process of automatic code generation that overcomes these difficulties for lattice-Boltzmann simulations.It details the development of GPU-based modules for an extensible lattice-Boltzmann simulation package-LBHydra.The performance of the automatically generated code is compared to equivalent purpose written codes for both single-phase,multiphase,and multicomponent flows.The flexibility of the new method is demonstrated by simulating a rising,dissolving droplet moving through a porous medium with user generated lattice-Boltzmann models and subroutines.
