检索结果-内蒙古大学图书馆

2009 International Conference on Machine Learning and Cybernetics(2009机器学习与控制论国际会议)

作者： JIN-CAI CHANGE YU-HUAN CUI AI-MIN YANG CHUN-FENG LIU College of Science Hebei Ploytechnic University Tangshan 063009 China College of Science Hebei Ploytechnic University Tangshan 063009 China School of Management Hebei

Through the research of the parallel computational model based on the principal and subordinate mode and the basic theory of Gmres algorithm in Krylov subspace, this essay raises a new parallel PCGMRES algorithm which possesses PC pattern, and shows the computing examples for linear equations. After the comparison with the result from the parallel GMRES (m) algorithm, it shows that this designed parallel algorithm can reduce the iteration frequency, shorten the computing time and obtain better speedup ratio and computing efficiency at the premise of assuring the computation precision.s.

关键词： Key Krylov Subspace PCGMRES algorithm parallel algorithm

来源：评论

学校读者我要写书评

暂无评论

Efficient Reduction from Block Hessenberg Form to Hessenberg Form Using Shared Memory

引用

10th Nordic International Conference on Applied parallel Computing - State of the Art in Scientific and parallel Computing (PARA)

作者： Karlsson, Lars Kagstrom, Bo Umea Univ Dept Comp Sci Umea Sweden Umea Univ HPC2N Umea Sweden

ISBN: (纸本)9783642281440;9783642281457

A new cache-efficient algorithm for reduction from block Hessenberg form to Hessenberg form is presented and evaluated. The algorithm targets parallel computers with shared memory. One level of look-ahead in combination with a dynamic load-balancing scheme significantly reduces the idle time and allows the use of coarse-grained tasks. The coarse tasks lead to high-performance computations on each processor/core. Speedups close to 13 over the sequential unblocked algorithm have been observed on a dual quad-core machine using one thread per core.

关键词： Hessenberg reduction block Hessenberg form parallel algorithm dynamic load-balancing blocked algorithm high performance

来源：评论

学校读者我要写书评

暂无评论

Implementations of Main algorithms for Generalized Eigenproblem on GPU Accelerator

Implementations of Main Algorithms for Generalized Eigenprob...

引用

3rd International Conference on Swarm Intelligence (ICSI)

作者： Zhao, Yonghua Zhang, Jian Chi, Xuebin Chinese Acad Sci Supercomp Ctr Comp Network Informat Ctr Beijing 100190 Peoples R China

ISBN: (纸本)9783642310201

A generalized eigensystem problem is usually transformed, utilizing Cholesky decomposition, to a standard eigenproblem. The latter is then solved efficiently by a matrix reduction approach based on Householder tridiagonalization method. We present parallel implementation of an integrated transformation-reduction algorithm on GPU accelerator using CUBLAS. Experimental results clearly demonstrate the potential of data-parallel coprocessors for scientific computations. When comparing against the CPU implementation, the GPU implementations achieve above 16-fold and 26-fold speedups in double precision for reduction and transformation respectively.

关键词： parallel algorithm eigenproblem GPGPU symmetric matrix

来源：评论

学校读者我要写书评

暂无评论

SPUM: A Screen Partition Update Method for Embedded Multi-window Systems

SPUM: A Screen Partition Update Method for Embedded Multi-wi...

引用

4th International Symposium on Information Science and Engineering (ISISE)

作者： Jiang Yan Zeng Xue-wen Sun Peng Zhu Xiao-yong Chinese Acad Sci Inst Acoust Natl Network New Media Engn Res Ctr Beijing Peoples R China

ISBN: (纸本)9781467356800

This paper proposes a screen partition update method (SPUM) for embedded multi-window systems with the purpose of improving their display performance. In this method the whole screen is partitioned into multiple independent sub-regions according to the position and size information of application windows at first and the overlap degree of each sub-region is calculated afterwards. Each window has an associated bitmap used to mark which sub-regions on the whole screen are contained by this window and which are not. When one application window updates, sub-regions of this window are updated step by step. In order to reduce the probability of conflict, the free sub-region with bigger overlap degree is updated preferentially. This method increases the probability of parallel update. When we apply the SPUM algorithm into an actual DirectFB graphics system, the total window update time cost is reduced by 35% and the conflict number is decreased by 72% in our experiment. Further experiment shows that with the increase of refresh rate the performance improvement introduced by the algorithm is more notable.

关键词： embedded system screen partition graphics system parallel algorithm

来源：评论

学校读者我要写书评

暂无评论

parallelization of the Seismic Ray Trace algorithm

Parallelization of the Seismic Ray Trace Algorithm

引用

9th International Conference on parallel Processing and Applied Mathematics (PPAM)

作者： Szostek, Kamil Lesniak, Andrzej AGH Univ Sci & Technol Fac Geol Geophys & Environm Protect Dept Geoinformat & Appl Comp Sci PL-30059 Krakow Poland

ISBN: (纸本)9783642314995;9783642315008

This article presents the parallelization of seismic ray trace algorithm. The chosen Urdaneta's algorithm is shortly described. It provides wavelength dependent smoothing and frequency dependant scattering thanks to the implementation of Lomax's method for approximating broad-band wave propagation. It also includes Vinje et al. wavefront propagation technique that provides fairly constant density of rays. Then the parallelized algorithm is preliminarily tested on synthetic data and the results are presented and discussed.

关键词： raytrace 2D seismic modeling parallel algorithm

来源：评论

学校读者我要写书评

暂无评论

algorithms of Wavelet Compression of Linear Spline Spaces

引用

VESTNIK ST PETERSBURG UNIVERSITY-MATHEMATICS 2012年第2期45卷 82-92页

作者： Makarov, A. A. St Petersburg State Univ Univ Skaya Nab 7-9 St Petersburg 199034 Russia

Splines and wavelets have been finding increasing use in the theory of information. Wavelet decompositions are used in designing efficient algorithms for processing (compression) of large information flows. If one succeeds in establishing the embeddability of spaces of splines on a sequence of sparsing/refining grids, in representing the chain of embedded spaces as a direct sum of wavelet spaces, and in realizing the base functions with the minimum length of their support, then this suggests a wavelet decomposition of the information flow, leading, in turn, to substantial savings in the computational cost. This being so, it proves possible to resolve the initial information flow into components to single out the principal and refining information flows, depending on the needs. For uniform grids on the real line, wavelet decompositions are well known. In this case, there applies the powerful technique of harmonic analysis, as well as the lifting scheme or the wavelet scheme. However, many applications require considering bounded intervals and nonuniform grids. For example, for efficient compression of nonuniform flows of information (featuring singularities or rapidly fluctuating characteristics), it is expedient to employ an adaptive nonuniform grid, which takes account of the singularities of the flow being processed. This renders possible to improve approximation of functions without complicating the computations. The previously obtained results pertained to splines on infinite grids. Making both the grid and the corresponding numerical flow infinite renders theoretical studies simpler;however, in practice, one has to deal with finite flows. This paper continues the studies initiated for finite-dimensional spaces. The purpose of this work is to built a wavelet decomposition (compression) on a nonuniform grid and develop the corresponding decomposition and reconstruction algorithms for infinite flows (with a grid on an open interval) and finite flows (with a grid o

关键词： approximation theory splines wavelets data compression parallel algorithm

来源：评论

学校读者我要写书评

暂无评论

Digital and Analog Compatible TV Transmitter Power Measurement

Digital and Analog Compatible TV Transmitter Power Measureme...

引用

3rd International Conference on Information Computing and Applications (ICICA 2012)

作者： Zhang, Sujuan Zhang, Yanli Liu, Yongchang Ma, Jun Hebei United Univ Network Informat Ctr Tangshan Peoples R China

ISBN: (纸本)9783642340406

The transmission power of analog TV transmitter is always measured as visual peak power, that is, the power level reaches while the synchronizing pulses are being transmitted, and so ordinary power meter cannot measure the value of analog TV transmitter power. The paper proposes a new measurement method;a parallel algorithm running in FPGA control high-speed AD, which can measure three analog TV RF signal powers simultaneously, and the paper also provides the signal type recognition and corresponding filter method to achieve digital and analog signals compatible.

关键词： analog signal digital signal power measurement FPGA parallel algorithm

来源：评论

学校读者我要写书评

暂无评论

Design and Implementation of a High-Performance Client/Server Voiceprint Recognition System

Design and Implementation of a High-Performance Client/Serve...

引用

IEEE International Conference on Information and Automation (ICIA)

作者： Gao Guanyu Kang Kai Guan ShengXiao Gao Guanhua Univ Sci & Technol China Dept Automat Hefei 230026 Peoples R China Northeastern Univ Software Coll Shenyang Peoples R China

ISBN: (纸本)9781467322379

This paper designs and implements an authentication system of client/server architecture based on voiceprint recognition. The client uses MFCC method to extract feature vectors from the speaker's voice and the server uses VQ method for recognition. We propose a new method of endpoint detection named short-term variance which can effectively pick up the voice signal from the original signal in actual environment. With the improved endpoint detection algorithm, the system can effectively resist the noise and achieve a higher recognition rate. In order to boost the server's performance, we implemented a new parallel algorithm for VQ codeword search on the SMP system. By using this method, the server improved the processor load rate and the speed of operation, as well as reduced the system response time. In the experiment, we evaluated the system recognition accuracy and efficiency of the VQ parallel algorithms.

关键词： voiceprint recognition endpoint detection client/server parallel algorithm multi-thread

来源：评论

学校读者我要写书评

暂无评论

algorithms of parallel calculations in task of tolerance ellipsoidal estimation of interval model parameters

引用

BULLETIN OF THE POLISH ACADEMY OF SCIENCES-TECHNICAL SCIENCES 2012年第1期60卷 159-164页

作者： Dyvak, M. Stakhiv, P. Pukas, A. Ternopil Natl Econ Univ Fac Comp Informat Technol UA-46020 Ternopol Ukraine Tech Univ Lodz Inst Comp Sci PL-93005 Lodz Poland

The methods of the tolerance ellipsoidal estimation for the tasks of synthesis of the tolerances to parameters of radio-electronic circuits and possibility of its parallelization are considered. These methods are the result of the task of estimation the solutions of an interval system of linear algebraic equations (ISLAE) which is built according to given criteria of optimality. The numerical algorithm is proposed for solving the tolerance ellipsoidal estimation tasks with a possibility of parallelization.

关键词： interval model parameters identification tolerance ellipsoidal estimation parallel algorithm

来源：评论

学校读者我要写书评

暂无评论

parallel block interface domain decomposition methods for the 2D convection-diffusion equation

引用

INTERNATIONAL JOURNAL OF COMPUTER MATHEMATICS 2012年第12期89卷 1704-1723页

作者： Tan, Kah Bee Ali, Norhashidah Hj. Mohd. Lai, Choi-Hong Univ Sains Malaysia Sch Math Sci George Town 11800 Malaysia Univ Greenwich Old Royal Naval Coll Sch Comp & Math Sci London SE10 9LS England

In this paper, a new block interface domain decomposition method (BI-DDM) with non-overlapping subdomains for the numerical solution of a two-dimensional convection-diffusion equation is presented. The block interface formulation is derived from the idea of using small groups of a certain number of mesh points where this group is treated explicitly similar to the way a single point is treated in the point method. The BI-DDM is incorporated with a correction phase which is able to economize further on the computing cost. The performance analysis of this method on several recently developed group iterative schemes implemented on a message-passing architecture are presented and discussed.

关键词： explicit group method domain decomposition message-passing interface convection-diffusion parallel algorithm

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：