检索结果-内蒙古大学图书馆

International Symposium on Parallel and Distributed Processing (IPDPS)

作者： Zhihui Du Hongyang Sun Yuxiong He Yu He David A. Bader Huazhe Zhang Tsinghua National Laboratory for Information Science and Technology Department of Computer Science and Technology Tsinghua University Beijing China School of Computer Engineering Nanyang Technological University Singapore Microsoft Research Redmonds WA USA College of Computing Georgia Institute of Technology Atlanta GA USA School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing China

ISBN: (纸本)9781467360661

High response quality is critical for many best-effort interactive services, and at the same time, reducing energy consumption can directly reduce the operational cost of service providers. In this paper, we study the quality-energy tradeoff for such services by using a composite performance metric that captures their relative importance in practice: Service providers usually grant top priority to quality guarantee and explore energy saving secondly. We consider scheduling on multicore systems with core-level DVFS support and a power budget. Our solution consists of two steps. First, we employ an equal sharing principle for both job and power distribution. Specifically, we present a “Cumulative Round-Robin” policy to distribute the jobs onto the cores, and a “Water-Filling” policy to distribute the power dynamically among the cores. Second, we exploit the concave quality function of many best-effort applications, and develop Online-QE, a myopic optimal online algorithm for scheduling jobs on a single-core system. Combining the two steps together, we present a heuristic online algorithm, called DES (Dynamic Equal Sharing), for scheduling best-effort interactive services on multicore systems. The simulation results based on a web search engine application show that DES takes advantage of the core-level DVFS architecture and exploits the concave quality function of best-effort applications to achieve high service quality with low energy consumption.

关键词： Multicore processing Schedules Heuristic algorithms Optimal scheduling Energy consumption Dynamic scheduling

来源：评论

学校读者我要写书评

暂无评论

Fault Tolerance and Recovery for Group Communication Services in Distributed Networks

引用

Journal of Computer science & technology 2012年第2期27卷 298-312页

作者：王跃华周忠 Ling Liu 吴威 State Key Laboratory of Virtual Reality Technology and Systems Beihang UniversityBeijing 100191China School of Computer Science and Engineering Beihang UniversityBeijing 100191China College of Computing Georgia Institute of TechnologyAtlanta 30332U.S.A.

Group communication services （GCSs） are becoming increasingly important as a wide field of promising applications has emerged to serve millions of users distributed across the ***,it is challenging to make the service fault tolerance and scalable to fulfill the voluminous demand of users in a distributed network （DN）.While many reliable group communication protocols have been dedicated to addressing such a challenge so as to accommodate the changes in the network,they are often costly or require complicated strategies to handle the service interruptions caused by node departures or link failures,which hinders the service *** this paper,we present two schemes to address the *** first one is a location-aware replication scheme called NS,which makes replicas in a dispersed fashion that enables the services on nodes to gain immunity of failures with different patterns （e.g.,network partition and single point failure） while keeping replication overhead *** second one is a novel failure recovery scheme that exploits the independence between service recovery and structure recovery in time domain to achieve quick failure *** simulation results indicate that the two proposed schemes outperform the existing schemes and simple alternative schemes in service success rate,recovery latency,and communication cost.

关键词： fault tolerance failure recovery replication location group communication

来源：评论

学校读者我要写书评

暂无评论

The HART II international workshop: an assessment of the state of the art in CFD/CSD prediction

引用

CEAS Aeronautical Journal 2013年第4期4卷 345-372页

作者： Smith, Marilyn J. Lim, Joon W. van der Wall, Berend G. Baeder, James D. Biedron, Robert T. Boyd, D. Douglas Jayaraman, Buvana Jung, Sung N. Min, Byung-Young School of Aerospace Engineering Georgia Institute of Technology Atlanta USA Ames Research Center US Army Aeroflightdynamics Moffett Field USA Institute of Flight Systems German Aerospace Center (DLR) Braunschweig Germany Department of Aerospace Engineering University of Maryland College Park USA NASA Langley Research Center Computational AeroSciences Branch Hampton USA NASA Langley Research Center Aeroacoustics Branch Hampton USA Science and Technology Corporation Moffett Field USA Department of Aerospace Information Engineering Konkuk University Seoul Republic of Korea Aerodynamics Department United Technologies Research Ctr East Hartford USA

Over the past decade, there have been significant advancements in the accuracy of rotor aeroelastic simulations with the application of computational fluid dynamics methods coupled with computational structural dynamics codes (CFD/CSD). The HART II international workshop database, which includes descent operating conditions with strong blade–vortex interactions (BVI), provides a unique opportunity to assess the ability of CFD/CSD to capture these physics. In addition to a baseline case with BVI, two additional cases with 3/rev higher harmonic blade root pitch control are available for comparison. The collaboration during the workshop permits assessment of structured, unstructured, and hybrid overset CFD/CSD methods from across the globe on the dynamics, aerodynamics, and wake structure. Evaluation of the plethora of CFD/CSD methods indicates that the most important numerical variables associated with most accurately capturing BVI include the use of either a two-equation RANS model or detached eddy simulation-based turbulence model and a sufficiently small time step. An appropriate trade-off between grid fidelity and spatial accuracy schemes also appears to be important for capturing BVI on the advancing side of the rotor disk. Overall, the CFD/CSD methods generally fall within the same accuracy; cost-effective hybrid Navier-Stokes/Lagrangian wake methods tend to correlate less accurately with experiment and have larger data scatter than the full CFD/CSD methods for most parameters evaluated. The importance of modeling the fuselage is observed, and other requirements are discussed.

关键词：

来源：评论

学校读者我要写书评

暂无评论

GREEN-IT: An approach to energy savings using energy aware network management system

GREEN-IT: An approach to energy savings using energy aware n...

引用

18th National Conference on Communications, NCC 2012

作者： Chaudhari, Santosh K. Nottath, Subhash Subramanian, Mani Murthy, Hema A. Computer Science and Engineering Indian Institute of Technology Madras Chennai - 600036 India College of Computing Georgia Institute of Technology Atlanta GA 30332-0280 United States

ISBN: (纸本)9781467308168

During peak hours, the focus of an Internet service provider (ISP) is on meeting user demand and maintaining the quality of service of a network. However this leads to under-utilization of network devices/components in different parts of the network. Since the power consumption of network devices are constant and independent of their actual workloads, the under-utilized network devices/components continue to consume the same power as that of a component working at full workload. We propose an energy aware network management system (NMS) framework based on the centralized decision management system (CDMS) reported in [1]. The system uses Bayesian belief network (BBN) to predict the utilization in different ports of network routers. The information already available in the NMS is sufficient for the proposed framework. For implementing the proposed scheme, an extended interface management information base (MIB) is proposed, which includes an operational state. Simulations are performed on two standard networks - IndiaNet and ARPANet, and the results show that with the proposed approach the CDMS is able to save approximately 10-16% of power at the cost of 1-5% end-to-end delay overhead. The cost of re-routing and Simple Network Management Protocol (SNMP) messages is minimal and acceptable considering the total power saved by CDMS. © 2012 IEEE.

关键词： Bayesian networks

来源：评论

学校读者我要写书评

暂无评论

Advertising keywords recommendation for short-text web pages using wikipedia

Advertising keywords recommendation for short-text web pages...

引用

作者： Zhang, Weinan Wang, Dingquan Xue, Gui-Rong Zha, Hongyuan Department of Computer Science and Engineering Shanghai Jiao Tong University Dongchuan Road Shanghai 200240 China *** Zhejiang China College of Computing Georgia Institute of Technology Atlanta GA 30032 United States

Advertising keywords recommendation is an indispensable component for online advertising with the keywords selected from the target Web pages used for contextual advertising or sponsored search. Several ranking-based algorithms have been proposed for recommending advertising keywords. However, for most of them performance is still lacking, especially when dealing with short-text target Web pages, that is, those containing insufficient textual information for ranking. In some cases, short-text Web pages may not even contain enough keywords for selection. A natural alternative is then to recommend relevant keywords not present in the target Web pages. In this article, we propose a novel algorithm for advertising keywords recommendation for short-text Web pages by leveraging the contents of Wikipedia, a user-contributed online encyclopedia. Wikipedia contains numerous entities with related entities on a topic linked to each other. Given a target Web page, we propose to use a content-biased PageRank on the Wikipedia graph to rank the related entities. Furthermore, in order to recommend high-quality advertising keywords, we also add an advertisement-biased factor into our model. With these two biases, advertising keywords that are both relevant to a target Web page and valuable for advertising are recommended. In our experiments, several state-of-the-art approaches for keyword recommendation are compared. The experimental results demonstrate that our proposed approach produces substantial improvement in the precision of the top 20 recommended keywords on short-text Web pages over existing approaches. © 2012 ACM.

关键词： Websites

来源：评论

学校读者我要写书评

暂无评论

Improving the Performance of Dynamical Simulations Via Multiple Right-Hand Sides

Improving the Performance of Dynamical Simulations Via Multi...

引用

International Symposium on Parallel and Distributed Processing (IPDPS)

作者： Xing Liu Edmond Chow Karthikeyan Vaidyanathan Mikhail Smelyanskiy School of Computational Science and Engineering College of Computing Georgia Institute of Technology Atlanta GA USA Parallel Computing Laboratory Intel Corporation Santa Clara CA USA

This paper presents an algorithmic approach for improving the performance of many types of stochastic dynamical simulations. The approach is to redesign existing algorithms that use sparse matrix-vector products (SPMV) with single vectors to instead use a more efficient kernel, the generalized SPMV (GSPMV), which computes with multiple vectors simultaneously. In this paper, we show how to redesign a dynamical simulation to exploit GSPMV in way that is not initially obvious because only one vector is available at a time. We study the performance of GSPMV as a function of the number of vectors, and demonstrate the use of GSPMV in the Stokesian dynamics method for the simulation of the motion of macromolecules in the cell. Specifically, for our application, we find that with modern multicore Intel microprocessors in clusters of up to 64 nodes, we can typically multiply by 8 to 16 vectors in only twice the time required to multiply by a single vector. After redesigning the Stokesian dynamics algorithm to exploit GSPMV, we measure a 30 percent speedup in performance in single-node, data parallel simulations.

关键词： Vectors Sparse matrices Heuristic algorithms computational modeling Mathematical model Equations Kernel

来源：评论

学校读者我要写书评

暂无评论

Stock market volatility prediction: A service-oriented multi-kernel learning approach

Stock market volatility prediction: A service-oriented multi...

引用

2012 IEEE 9th International Conference on Services computing, SCC 2012

作者： Wang, Feng Liu, Ling Dou, Chenxiao State Key Lab. of Software Engineering Wuhan University Wuhan China College of Computing Georgia Institute of Technology Atlanta GA United States School of Computer Science Wuhan University Wuhan China

ISBN: (纸本)9780769547534

Stock market is an important and active part of nowadays financial markets. Stock time series volatility analysis is regarded as one of the most challenging time series forecasting due to the hard-to-predict volatility observed in worldwide stock markets. In this paper we argue that the stock market state is dynamic and invisible but it will be influenced by some visible stock market information. Existing research on financial time series analysis and stock market volatility prediction can be classified into two categories: in depth study of one market factor on the stock market volatility prediction or prediction by combining historical price fluctuations with either trading volume or news. In this paper we present a service-oriented multi-kernel based learning framework (MKL) for stock volatility analysis. Our MKL service framework promotes a two-tier learning architecture. In the top tier, we develop a suite of data preparation and data transformation techniques to provide a source-specific modeling, which transforms and normalizes a source specific input dataset into the MKL ready data representation. Then we apply data alignment techniques to prepare the datasets from multiple information sources based on the classification model we choose for cross-source correlation analysis. In the next tier, we develop model integration methods to perform three analytic tasks: (i) building one sub-kernel per source, (ii) learning and tuning the weights for sub-kernels through weight adjustment methods and (iii) performing multi-kernel based cross-correlation analysis of market volatility. To validate the effectiveness of our service oriented MKL approach, we performed experiments on HKEx 2001 stock market datasets with three important market information sources: historical prices, trading volumes and stock related news articles. Our experiments show that 1) multi-kernel learning method has a higher degree of accuracy and a lower degree of false prediction, compared to exist

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Robust Kernel Nonnegative Matrix Factorization

Robust Kernel Nonnegative Matrix Factorization

引用

IEEE International Conference on Data Mining Workshops (ICDM Workshops)

作者： Zhichen Xia Chris Ding Edmond Chow College of Computing Georgia Institute of Technology Department of Computer Science and Engineering University of Texas at Arlington

ISBN: (纸本)9781467351645

Kernel methods and Nonnegative matrix factorization (NMF) are both widely used in data mining and machine learning. The previous one is best known for its capability of transforming data into high dimension feature space, while the latter one is well known for its natural interpretations and good performance. In this paper, we propose a robust kernel NMF approach using L 2, 1 norm loss function. Compared with the standard NMF algorithm, the new robust kernel NMF updating algorithm is as elegant and as simple, but with the newly added robustness to handle significantly corrupted datasets because of using L 2, 1 norm. Experiments on normal and occluded datasets indicate that robust kernel NMF always perform better than k-means and standard NMF.

关键词： Kernel Robustness Standards Linear programming Convergence Clustering algorithms Noise

来源：评论

学校读者我要写书评

暂无评论

ACTIVE LEARNING FOR TAG RECOMMENDATION UTILIZING ON-LINE PHOTOS LACKING TAGS

ACTIVE LEARNING FOR TAG RECOMMENDATION UTILIZING ON-LINE PHO...

引用

IEEE International Conference on Image Processing

作者： Yajun Gao Baoxin Li College of Computing Georgia Institute of Technology USA Computer Science and Engineering Arizona State University USA

Recommending text tags for on-line photos is useful for Internet photo services. Typical solutions to this problem require analysis of the correlation among different attributes of the photos, including the correlation between the textual features and visual features computed from a photo. However, most on-line photos have very few tags or even no tags, and thus they contribute little or none to the analysis of tag-photo correlation, which is a key component in those schemes that rely on such analysis for tag recommendation. To address this practical challenge, we propose an active learning method for incorporating photos with no or few tags so as to enhance the correlation analysis for improved performance in tag recommendation. We demonstrate the effectiveness of the proposed approach using a dataset of more than 33,000 photos collected from Flickr.

关键词： Training Correlation Visualization Semantics Measurement Tagging computational modeling

来源：评论

学校读者我要写书评

暂无评论

SDHT： Efficient One-Hop DHT Lookup Framework for P2P Live Streaming

引用

China Communications 2012年第8期9卷 88-104页

作者： Chen Zhuo Feng Gang Zhou Yang Lu Yi National Key Laboratory of Science and Technology on Communications University of Electronic Science and Technology of China Chengdu 611731 P. R. China College of Computing Georgia Institute of Technology Atlanta GA 30332 USA College of Computer Science and Engineering Chongqing University of Technology Chongqing 400054 P. R. China

Video data location plays a key role for Peer-to-Peer （P2P） live streaming applications. In this paper, we propose a new one-hop Distributed Hash Table （DHT） lookup frarrework called Strearre ing-DHT （SDHT） to provide efficient video data location service. By adopting an enhanced events dissemination mechanism-EDRA＋, the accuracy of routing table at peers can be guaranteed. More importantly, in order to enhance the perforlmnce of video data lookup operation without incurring extra overhead, we design a so-called Distributed Index Mapping and Management Mechanism （DIMM） for SDHT. Both analytical modeling and intensive simulation experiments are conducted to demonstrate the effectiveness of SDHT framework. Numerical results show that almost 90% requested video data can be retrieved within one second in SDHT based systems, and SDHT needs only 26% average bandwidth consumption when compared with similar one-hop DHT solutions such as D1HT. This indicates that SDHT framework is an appropriate data lookup solution for time-sensitive network applications such as P2P live streaming.

关键词： peer-to-peer distributed hash table streaming application live streaming EDRA

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：