检索结果-内蒙古大学图书馆

IEEE International Conference on Cluster Computing

作者： Ashwin Mandayam Aji Antonio J. Peña Pavan Balaji Wu-chun Feng AMD Research Advanced Micro Devices Inc. +Math. and Comp. Sci. Division Argonne National Laboratory Department of Computer Science Virginia Tech

OpenCL is a portable interface that can be used to program cluster nodes with heterogeneous compute devices. The OpenCL specification tightly binds its workflow abstraction, or "command queue," to a specific device for the entire program. For best performance, the user has to find the ideal queue -- device mapping at command queue creation time, an effort that requires a thorough understanding of the match between the characteristics of all the underlying device architectures and the kernels in the program. In this paper, we propose to add scheduling attributes to the OpenCL context and command queue objects that can be leveraged by an intelligent runtime scheduler to automatically perform ideal queue - device mapping. Our proposed extensions enable the average OpenCL programmer to focus on the algorithm design rather than scheduling and automatically gain performance without sacrificing programmability. As an example, we design and implement an OpenCL runtime for task-parallel workloads, called MultiCL, which efficiently schedules command queues across devices. Within MultiCL, we implement several key optimizations to reduce runtime overhead. Our case studies include the SNU-NPB OpenCL benchmark suite and a real-world seismology simulation. We show that, on average, users have to apply our proposed scheduler extensions to only four source lines of code in existing OpenCL applications in order to automatically benefit from our runtime optimizations. We also show that MultiCL always maps command queues to the optimal device set with negligible runtime overhead.

关键词： Kernel Runtime Performance evaluation Processor scheduling Context Schedules Dynamic scheduling

来源：评论

学校读者我要写书评

暂无评论

Fast parallel algorithms for coloring random graphs 17th

引用

17th International Workshop on Graph-Theoretic Concepts in computer science, WG 1991

作者： Kedem, Zvi M. Palem, Krishna V. Pantziou, Grammati E. Spirakis, Paul G. Zaroliagis, Christos D. Dept of Computer Science Courant Institute of Math. Sciences NYU 251 Mercer St New YorkNY10012 United States IBM Research Division T.J. Watson Research Center P.O. Box 704 Yorktown HeightsNY10598 United States Computer Technology Institute P.O. Box 1122 Patras26110 Greece Computer Sc and Eng Dept University of Patras Greece

ISBN: (纸本)9783540551218

We improve here the expected performance of parallel algorithms for graph coloring. This is achieved through new adaptive techniques that may be useful for the average-case analysis of many graph algorithms. We apply our techniques to: (a) the class Gn,p of random graphs. We present a parallel algorithm which colors the graph with a number of colors at most twice its chromatic number and runs in time O(log4 n/ log log n) almost surely, for p = Ω((log(3) n)2/ log(2) n). The number of processors used is O(m) where m is the number of edges of the graph. (b) the class of all fc-colorable graphs, uniformly chosen. We present a parallel algorithm which actually constructs the coloring in expected parallel time O(log2 n), for constant k, by using O(m) processors on the average. This problem is not known to have a polynomial time algorithm in the worst case. © Springer-Verlag Berlin Heidelberg 1992.

关键词： Parallel algorithms

来源：评论

学校读者我要写书评

暂无评论

Communication analysis of parallel 3D FFT for flat cartesian meshes on large Blue Gene systems 08

引用

Proceedings of the 15th international conference on High performance computing

作者： Anthony Chan Pavan Balaji William Gropp Rajeev Thakur ASCI FLASH Center University of Chicago Math. and Comp. Sci. Division Argonne National Laboratory Dept. of Computer Science University of Illinois Urbana-Champaign

ISBN: (纸本)9783540898931

Parallel 3D FFT is a commonly used numerical method inscientific computing. P3DFFT is a recently proposed implementationof parallel 3D FFT that is designed to allow scalability to massivelylarge systems such as Blue Gene. While there has been recent workthat demonstrates such scalability on regular cartesian meshes (equallength in each dimension), its performance and scalability for flat cartesianmeshes (much smaller length in one dimension) is still a concern. Inthis paper, we perform studies on a 16-rack (16384-node) Blue Gene/Lsystem that demonstrates that a combination of the network topologyand the communication pattern of P3DFFT can result in early networksaturation and consequently performance loss. We also show that remappingprocesses on nodes and rotating the mesh by taking the communicationproperties of P3DFFT into consideration, can help alleviate thisproblem and improve performance by up to 48% in some special cases.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Adaptive fault-tolerant wormhole routing algorithms with low virtual channel requirements

Adaptive fault-tolerant wormhole routing algorithms with low...

引用

International Symposium on Parallel Architectures, Algorithms and Networks (ISPAN)

作者： S. Chalasani R.V. Boppana Electrical & Comp. Engineering Department University of Wisconsin Madison Madison WI USA Division of Math. and Computer Science University of Texas San Antonio San Antonio TX USA

We present simple methods to enhance a recently proposed class of fully adaptive algorithms for fault tolerant wormhole routing. These algorithms are based on J. Duato's (1993) theory and are being used in several research projects. We show that with three virtual channels per physical channel, multiple rectangular shaped fault blocks can be tolerated in two dimensional meshes. There is no restriction on the number of faults, and maintaining fault information locally is sufficient. The logic for fault tolerant routing is used only in the presence of faults, and there is no performance degradation in the absence of faults. The proposed technique incorporates fault tolerance into wormhole algorithms with simple logic and low virtual channel requirements.< >

关键词： Fault tolerance Routing Logic System recovery Algorithm design and analysis Adaptive algorithm computer science Degradation computer networks Concurrent computing

来源：评论

学校读者我要写书评

暂无评论

Reconstruction of 2D PET data with monte carlo generated natural pixels

Reconstruction of 2D PET data with monte carlo generated nat...

引用

2004 2nd IEEE International Symposium on Biomedical Imaging: Macro to Nano

作者： Vandenberghe, S. Byrne, C.L. Soares, E.J. Lemahieu, I. Glick, S. FWO Flanders MEDISIP Ghent University University of Massachusetts Medical School Worcester MA Dept. of Mathematical Sciences University of Massachussetts Lowell Lowell MA Dept. of Math. and Computer Science College of the Holy Cross Worcester MA ELIS Department MEDISIP Ghent University Ghent Belgium Division of Nuclear Medicine University of Massachussetts Medical School Worcester MA

ISBN: (纸本)0780383885

In discrete detector PET, natural pixels are image basis functions calculated from responses of detector pairs. By using reconstruction with natural pixels the discretization of the object into a predefined grid can be avoided. Instead of trying to find the object x from system matrix A and measured data p by solving Ax = p, one solves Mq = AATq = p. The backprojection of the final q results in the estimated image of the object xˆ. In previous work natural pixels were calculated by intersections of strip functions or combinations of strip functions. However, in PET systems the detector response is not correctly described by a strip function. In PET the detector response is better described by a gaussian tube. It is quite difficult to calculate the intersections of all possible Lines Of Response (LOR) in the scanner. This paper proposes an easy and efficient way to generate the matrix M directly by Monte Carlo simulation. Using the natural pixel matrix M has several advantages over using the conventional system matrix A. The discretisation of the object into pixels, which requires an angular dependent projector and backprojector, is avoided. Due to rotational symmetry in the PET scanner the matrix M is block circulant and only the first block row needs to be stored. The block circulant property of the matrix allows use of fast direct methods to calculate the solution. Iterative methods to solve the system are preferred because of numerical stability. ©2004 IEEE.

关键词： Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Automated Discovery of Active Motifs in Multiple RNA Secondary Structures 2

Automated Discovery of Active Motifs in Multiple RNA Seconda...

引用

2nd International Conference on Knowledge Discovery and Data Mining, KDD 1996

作者： Wang, Jason T. L. Shapiro, Bruce A. Shasha, Dennis Zhang, Kaizhong Chang, Chia-Yo Computer and Information Sci. Dept. New Jersey Institute of Technology NewarkNJ07102 United States Division of Basic Sciences National Institutes of Health FrederickMD21702 United States Courant Institute of Math. Sciences New York University New YorkNY10012 United States Computer Science Dept. University of Western Ontario London OntarioN6A 5B7 Canada

ISBN: (纸本)1577350049

In this paper we present a method for discovering approximately common motifs (also known as active motifs) in multiple RNA secondary structures. The secondary structures can be represented as ordered trees (i.e., the order among siblings matters). Motifs in these trees are connected subgraphs that can differ in both substitutions and deletions/insertions. The proposed method consists of two steps: (1) find candidate motifs in a small sample of the secondary structures;(2) search all of the secondary structures to determine how frequently these motifs occur (within the allowed approximation) in the secondary structures. To reduce the running time, we develop two optimization heuristics based on sampling and pattern matching techniques. Experimental results obtained by running these algorithms on both generated data and RNA secondary structures show the good performance of the algorithms. To demonstrate the utility of our algorithms, we discuss their applications to conducting the phylogenetic study of RNA sequences obtained from GenBank. © 1996 AAAI (***). All Rights Reserved.

关键词： RNA

来源：评论

学校读者我要写书评

暂无评论

Free bits, PCPs and non-approximability-towards tight results

Free bits, PCPs and non-approximability-towards tight result...

引用

Annual IEEE Symposium on Foundations of computer science

作者： M. Bellare O. Goldreich M. Sudan Department of Computer Science & Engineering University of California San Diego La Jolla CA USA Computer Science and Applied Math. Department Weizmann Institute of Science Rehovot Israel IBM Research Division IBM Thomas J. Watson Research Center Yorktown Heights NY USA

The first part of this paper presents new proof systems and improved non-approximability results. In particular we present a proof system for NP using logarithmic randomness and two amortized free bits, so that Max clique is hard within N/sup 1/3/ and chromatic number within N/sup 1/5/. We also show hardness of 38/37 for Max-3-SAT, 27/26 for vertex cover, 82/81 for Max-cut, and 94/93 for Max-2-SAT. The second part of this paper presents a "reverse" of the FGLSS connection by showing that an NP-hardness result for the approximation of Max clique to within a factor of N/sup 1/(g+1/) would imply a probabilistic verifier for NP with logarithmic randomness and amortized free-bit complexity g. We also show that "existing techniques" won't yield proof systems of less than two bits in amortized free bit complexity. Finally, we initiate a comprehensive study of PCP and FPCP parameters, proving several triviality results and providing several useful transformations.

关键词： computer science Postal services Drives Cities and towns Error correction codes Polynomials History

来源：评论

学校读者我要写书评

暂无评论

Analyzing industrial furnace efficiency using comparative visualization in a virtual reality environment

American Society of Mechanical Engineers, Material Handling ...

引用

American Society of Mechanical Engineers, Material Handling division, MHD 1999年 5卷 191-199页

作者： Freitag, Lori Urness, Tim Math. and Computer Science Division Argonne National Laboratory Argonne IL 60439 United States Scientific Visualization Laboratory 1003 Cowley Hall University of Wisconsin - La Crosse La Crosse WI 54601 United States Argonne National Laboratory

We describe an interactive toolkit used to perform comparative analysis of two or more data sets arising from numerical simulations. Several techniques have been incorporated into this toolkit, including (1) successive visualization of individual data sets, (2) data comparison techniques such as computation and visualization of the differences between data sets, and (3) image comparison methods such as scalar field height profiles plotted in a common coordinate system. We describe each technique in detail and show example usage in an industrial application aimed at designing an efficient, low-NOx burner for industrial furnaces. Critical insights are obtained by interactively adjusted color maps, data culling, and data manipulation. New paradigms for scaling small values in the data comparison technique are described. The display device used for this application was the CAVE virtual reality theater, and we describe the user interface to the visualization toolkit and the benefits of immersive 3D visualization for comparative analysis.

关键词： Virtual reality

来源：评论

学校读者我要写书评

暂无评论

Calculated effect of alloy additions on the saturation magnetization of Fe0.80B0.20

引用

Materials Research Society Symposium - Proceedings 2003年 754卷 451-456页

作者： Nicholson, D.M.C. Wang, Yang Widom, Mike Computer Science and Math. Division Oak Ridge National Laboratory Oak Ridge TN 37830 United States Pittsburgh Supercomputing Center Pittsburgh PA 15213 United States Department of Physics Carnegie Mellon University Pittsburgh PA 15213 United States

The great number of different local environments in amorphous alloys leads to the evolution of complicated non collinear magnetic structures. Alloy additions can affect the magnetic structure in surprising ways. For example, replacement of a small amount of Fe with Co increases the saturation magnetization even though Co has a much smaller moment than Fe. The calculated behavior of the magnetic structure of (Fe(1-x)Mx)0.8B0.2 with M=Co, Cr, Zr, and Mn2Zr are presented.

关键词： Iron compounds

来源：评论

学校读者我要写书评

暂无评论

On the Sum of Squared η-μ Random Variates With Application to the Performance of Wireless Communication Systems

On the Sum of Squared η-μ Random Variates With Application...

引用

IEEE Vehicular Technology Conference

作者： Imran Shafique Ansari Ferkan Yilmaz Mohamed-Slim Alouini Computer Electrical and Mathematical Sciences and Engineering (CEMSE) Division at King Abdullah University of Science and Technology (KAUST) Al-Khawarizmi Applied Math. Building (Bldg. #1) Thuwal 23955-6900 Makkah Province Kingdom of Saudi Arabia

ISBN: (纸本)9781467363365

The probability density function (PDF) and cumulative distribution function of the sum of L independent but not necessarily identically distributed squared η-μ variates, applicable to the output statistics of maximal ratio combining (MRC) receiver operating over η-μ fading channels that includes the Hoyt and the Nakagami-m models as special cases, is presented in closed-form in terms of the Fox's H function. Further analysis, particularly on the bit error rate via PDF-based approach, is also represented in closed form in terms of the extended Fox's H function (H). The proposed new analytical results complement previous results and are illustrated by extensive numerical and Monte Carlo simulation results.

关键词： η-μ variates cellular mobile radio systems non-integer parameters diversity maximal ratio combining binary modulation schemes bit error rate Fox's H function Meijer's G function and extended Fox's H function Bit error rate diversity reception Wireless communication systems Mobile radio systems closed forms sums Cumulative distribution function Hydrogen

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：