检索结果-内蒙古大学图书馆

IEEE International Conference on Cloud computing Technology and science (CloudCom)

作者： Peng Lu Young Choon Lee Vincent Gramoli Luke M. Leslie Albert Y. Zomaya Center for Distributed and High Performance Computing The University of Sydney Australia School of Information Technologies NICTA and the University of Sydney Australia Department of Computer Science University of Illinois at Urbana-Champaign USA

Resource capacity is often over provisioned to primarily deal with short periods of peak load. Shaping these peaks by shifting them to low utilization periods (valleys) is referred to as "resource consumption shaping". While originally aimed at the data center level, the resource consumption shaping we consider focuses on local resources, like CPU or I/O as we have identified that individual jobs also incur load peaks and valleys on these resources. In this paper, we present Local Resource Shaper (LRS), which limits fairness in resource sharing between co-located MapReduce tasks. LRS enables Hadoop to maximize resource utilization and minimize resource contention independently of job type. Co-located MapReduce tasks are often prone to resource contention (i.e., Load peak) due to similar resource usage patterns particularly with traditional fair resource sharing. In essence, LRS differentiates co-located tasks through active and passive slots that serve as containers for interchangeable map or reduce tasks. LRS lets an active slot consume as much resources as possible, and a passive slot make use of any unused resources. LRS leverages such slot differentiation with its new scheduler, Interleave. Our results show that LRS always outperforms the best static slot configuration with three Hadoop schedulers in terms of both resource utilization and performance.

关键词： Resource management Scheduling Benchmark testing Educational institutions Heart beat Containers Linux

来源：评论

学校读者我要写书评

暂无评论

How file access patterns influence interference among cluster applications

How file access patterns influence interference among cluste...

引用

IEEE International Conference on Cluster computing

作者： Chih-Song Kuo Aamer Shah Akihiro Nomura Satoshi Matsuoka Felix Wolf Department of Computer Science RWTH Aachen University Aachen Germany German Research School for Simulation Sciences Laboratory for Parallel Programming Aachen Germany Tokyo Institute of Technology Global Scientific Information and Computing Center Tokyo Japan

On large-scale clusters, tens to hundreds of applications can simultaneously access a parallel file system, leading to contention and in its wake to degraded application performance. However, the degree of interference depends on the specific file access pattern. On the basis of synchronized time-slice profiles, we compare the interference potential of different file access patterns. We consider both micro-benchmarks, to study the effects of certain patterns in isolation, and realistic applications to gauge the severity of such interference under production conditions. In particular, we found that writing large files simultaneously with small files can slow down the latter at small chunk sizes but the former at larger chunk sizes. We further show that such effects can seriously affect the runtime of real applications-up to a factor of five in one instance. In the future, both our insights and profiling techniques can be used to automatically classify the interference potential between applications and to adjust scheduling decisions accordingly.

关键词： Interference Servers Throughput Benchmark testing Market research Degradation Aggregates

来源：评论

学校读者我要写书评

暂无评论

A mathematical model for empowerment of Beowulf clusters for exascale computing

A mathematical model for empowerment of Beowulf clusters for...

引用

2013 11th International Conference on High Performance computing and Simulation, HPCS 2013

作者： Mirtaheri, Seyedeh Leili Khaneghah, Ehsan Mousavi Grandinetti, Lucio Sharifi, Mohsen Center of High Performance Computing for Parallel and Distributed Processing University of Calabria Rende Italy School of Computer Engineering Iran University of Science and Technology Tehran Iran

ISBN: (纸本)9781479908363

High-performance computing (HPC) clusters are currently faced with two major challenges - namely, the dynamic nature of new generation of applications and the heterogeneity of platforms - if they are going to be useful for exascale computing. Processes running these applications may well demand unpredictable requirements and changes to system configuration and capabilities at runtime, thereby requiring fast system response without sacrificing the transparency and integrity of the reconfigured empowered system that is running on a heterogeneous platform. While a challenge in and of itself, platform heterogeneity is both useful and instrumental in the handling of unpredictable requests. The realization of such a dynamically reconfigurable and heterogeneous HPC cluster system for exascale computing requires a model to guide running processes to determine if they need empowerment of the current cluster, and if yes, by how much. To show the feasibility of empowerment of traditional HPC clusters for exascale computing, we have selected Beowulf as a noble candidate cluster and present a mathematical model for the empowerment of Beowulf clusters for exascale computing (EBEC). We have developed the model in line with Beowulf's cluster approach and by using vector space algebra. In contrast to traditional hardware-oriented approaches to improvise the performance of clusters, we use a software approach to the development of the proposed model by emphasizing processes, which act as the creators of the cluster and thus should decide on system (re)configuration, as the principal building blocks of the system. We have also adopted a new approach to heterogeneity by considering heterogeneity at different levels including hardware, system software, application software, and system functionality. In addition to support for heterogeneity and dynamic reconfiguration, the proposed model includes support for scalability that is crucial to exascale computing too. © 2013 IEEE.

关键词： Vector spaces

来源：评论

学校读者我要写书评

暂无评论

TRANSDUCTIVE NONNEGATIVE MATRIX FACTORIZATION FOR SEMI-SUPERVISED HIGH-PERFORMANCE SPEECH SEPARATION

TRANSDUCTIVE NONNEGATIVE MATRIX FACTORIZATION FOR SEMI-SUPER...

引用

IEEE International Conference on Acoustics, Speech and Signal Processing

作者： Naiyang Guan Long Lan Dacheng Tao Zhigang Luo Xuejun Yang Science and Technology on Parallel and Distributed Processing Laboratory School of Computer Science National University of Defense Technology Changsha Hunan 410073 China Centre for Quantum Computation and Intelligent Systems FEIT University of Technology Sydney Sydney NSW 2007 Australia State Key Laboratory of High Performance Computing National University of Defense Technology Changsha Hunan 410073 China

ISBN: (纸本)9781479928941

Regarding the non-negativity property of the magnitude spectrogram of speech signals, nonnegative matrix factorization (NMF) has obtained promising performance for speech separation by independently learning a dictionary on the speech signals of each known speaker. However, traditional NM-F fails to represent the mixture signals accurately because the dictionaries for speakers are learned in the absence of mixture signals. In this paper, we propose a new transductive NMF algorithm (TNMF) to jointly learn a dictionary on both speech signals of each speaker and the mixture signals to be separated. Since TNMF learns a more descriptive dictionary by encoding the mixture signals than that learned by NMF, it significantly boosts the separation performance. Experiments results on a popular TIMIT dataset show that the proposed TNMF-based methods outperform traditional NMF-based methods for separating the monophonic mixtures of speech signals of known speakers.

关键词： Speech Dictionaries as Topic Loudspeakers NMF protocol signals mixtures factorization HIGH-PERFORMANCE

来源：评论

学校读者我要写书评

暂无评论

Consumer participation in quality improvements for chronic disease care: Development and evaluation of an interactive patient-centered survey to identify preferred service initiatives

引用

Journal of Medical Internet Research 2014年第12期16卷 e292页

作者： Fradgley, Elizabeth A. Paul, Christine L. Bryant, Jamie Roos, Ian A. Henskens, Frans A. Paul, David J. Priority Research Centre for Health Behaviour School of Medicine and Public Health Hunter Medical Research Institute University of Newcastle Callaghan 2305 Australia Youth Research Center University of Melbourne Melbourne Australia Distributed Computing Research Group School of Electrical Engineering and Computer Science University of Newcastle Callaghan Australia

Background: With increasing attention given to the quality of chronic disease care, a measurement approach that empowers consumers to participate in improving quality of care and enables health services to systematically introduce patient-centered initiatives is needed. A Web-based survey with complex adaptive questioning and interactive survey items would allow consumers to easily identify and prioritize detailed service initiatives. Objective: The aim was to develop and test a Web-based survey capable of identifying and prioritizing patient-centered initiatives in chronic disease outpatient services. Testing included (1) test-retest reliability, (2) patient-perceived acceptability of the survey content and delivery mode, and (3) average completion time, completion rates, and Flesch-Kincaid reading score. Methods: In Phase I, the Web-based Consumer Preferences Survey was developed based on a structured literature review and iterative feedback from expert groups of service providers and consumers. The touchscreen survey contained 23 general initiatives, 110 specific initiatives available through adaptive questioning, and a relative prioritization exercise. In Phase II, a pilot study was conducted within 4 outpatient clinics to evaluate the reliability properties, patient-perceived acceptability, and feasibility of the survey. Eligible participants were approached to complete the survey while waiting for an appointment or receiving intravenous therapy. The age and gender of nonconsenters was estimated to ascertain consent bias. Participants with a subsequent appointment within 14 days were asked to complete the survey for a second time. Results: A total of 741 of 1042 individuals consented to participate (71.11% consent), 529 of 741 completed all survey content (78.9% completion), and 39 of 68 completed the test-retest component. Substantial or moderate reliability (Cohen's kappa>0.4) was reported for 16 of 20 general initiatives with observed percentage agreement ra

关键词： Ambulatory care Cardiology Chronic disease Consumer participation Health care surveys Medical oncology Neurology Patient-centered care

来源：评论

学校读者我要写书评

暂无评论

parallel macro pipelining on the intel SCC many-core computer

Parallel macro pipelining on the intel SCC many-core compute...

引用

2013 IEEE 37th Annual computer Software and Applications Conference, COMPSAC 2013

作者： Suss, Tim Schoenrock, Andrew Meisner, Sebastian Plessl, Christian Johannes Gutenberg-University Mainz Zentrum fur Datenverarbeitung Mainz Germany Carleton University School of Computer Science Ottawa Canada University of Paderborn Paderborn Center for Parallel Computing Paderborn Germany

ISBN: (纸本)9780769549798

In this paper we present how Intel's Single-Chip-Cloud processor behaves for parallel macro pipeline applications. Subsets of the SCC's available cores can be arranged as a pipeline where each core processes one stage of the overall workload. Each of the independent cores processes a small part of a larger task and feeds the following core with new data after it finishes its work. Our case-study is a parallel rendering system which renders successive images and applies different filters on them. On normal graphics adapters this is usually done in multiple cycles, we do this in a single pipeline pass. We show that we can achieve a significant speedup by using multiple parallel pipelines on the SCC. We show that we can further improve performance by using SCC's controlling PC in conjunction with the SCC. We also identify aspects of the SCC that hinder the overall performance, mainly the lack of local memory banks for each core on the SCC. The results presented in this paper are not limited to only image processing, but users could expect similar experiences where macro pipelining is used in other applications on the SCC. © 2013 IEEE.

关键词： Pipelines

来源：评论

学校读者我要写书评

暂无评论

Multi-label Classification with Clustering for Image and Text Categorization

Multi-label Classification with Clustering for Image and Tex...

引用

International Congress on Image and Signal Processing

作者： Gulisong Nasierding Atul Sajjanhar School of Computer Science and Technology Xinjiang Normal University Parallel and Distributed Computing Lab School of Information Technology Deakin University

ISBN: (纸本)9781479927654

This paper explores effective multi-label classification methods for multi-semantic image and text categorization. We perform an experimental study of clustering based multi-label classification (CBMLC) for the target problem. Experimental evaluation is conducted for identifying the impact of different clustering algorithms and base classifiers on the predictive performance and efficiency of CBMLC. In the experimental setting, three widely used clustering algorithms and six popular multi-label classification algorithms are used and evaluated on multi-label image and text datasets. A multi-label classification evaluation metrics, micro F1-measure, is used for presenting predictive performances of the classifications. Experimental evaluation results reveal that clustering based multi-label learning algorithms are more effective compared to their non-clustering counterparts.

关键词： Multi-label classification Clustering Image and text categorization

来源：评论

学校读者我要写书评

暂无评论

A Dynamic Replication Mechanism to Reduce Response-Time of I/O Operations in High Performance computing Clusters

A Dynamic Replication Mechanism to Reduce Response-Time of I...

引用

IEEE International Conference on Social computing (SocialCom)

作者： Ehsan Mousavi Khaneghah Seyedeh Leili Mirtaheri Lucio Grandinetti Amir Saman Memaripour Mohsen Sharifi Center of High Performance Computing for Parallel and Distributed Processing University of Calabria Rende Italy School of Computer Engineering Iran University of Science and Technology Tehran Iran

ISBN: (纸本)9781479915194

Extraordinary large datasets of high performance computing applications require improvement in existing storage and retrieval mechanisms. Moreover, enlargement of the gap between data processing and I/O operations' throughput will bound the system performance to storage and retrieval operations and remarkably reduce the overall performance of high performance computing clusters. File replication is a way to improve the performance of I/O operations and increase network utilization by storing several copies of every file. Furthermore, this will lead to a more reliable and fault-tolerant storage cluster. In order to improve the response time of I/O operations, we have proposed a mechanism that estimates the required number of replicas for each file based on its popularity. Besides that, the remaining space of storage cluster is considered in the evaluation of replication factors and the number of replicas is adapted to the storage state. We have implemented the proposed mechanism using HDFS and evaluated it using MapReduce framework. Evaluation results prove its capability to improve the response time of read operations and increase network utilization. Consequently, this mechanism reduces the overall response time of read operations by considering files' popularity in replication process and adapts the replication factor to the cluster state.

关键词： Time factors High performance computing Bandwidth Reliability Throughput System performance History

来源：评论

学校读者我要写书评

暂无评论

A mathematical model for empowerment of Beowulf clusters for exascale computing

A mathematical model for empowerment of Beowulf clusters for...

引用

International Conference on High Performance computing & Simulation (HPCS)

作者： Seyedeh Leili Mirtaheri Ehsan Mousavi Khaneghah Lucio Grandinetti Mohsen Sharifi Center of High Performance Computing for Parallel and Distributed Processing University of Calabria Rende Italy School of Computer Engineering Iran University of Science and Technology Tehran Iran

关键词： Vectors Mathematical model Hardware Runtime computers Computational modeling

来源：评论

学校读者我要写书评

暂无评论

parallel Macro Pipelining on the Intel SCC Many-Core computer

Parallel Macro Pipelining on the Intel SCC Many-Core Compute...

引用

IEEE International Symposium on parallel and distributed Processing Workshops and Phd Forum (IPDPSW)

作者： Tim Süß Andrew Schoenrock Sebastian Meisner Christian Plessl Johannes Gutenberg Universitat Mainz Mainz Rheinland-Pfalz DE Carleton University School of Computer Science Ottawa Canada University of Paderborn Paderborn Center for Parallel Computing Paderborn Germany

ISBN: (纸本)9781479913725

关键词： Pipelines Pipeline processing Rendering (computer graphics) Image color analysis Random access memory Tiles computers

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：