检索结果-内蒙古大学图书馆

advances in Computer Science and Information Technology 1

丛书名： Communications in Computer and Information Science

1000年

作者： Natarajan Meghanathan Brajesh Kumar Kaushik Dhinaharan Nagamalai

ISBN: (数字)9783642178573

ISBN: (纸本)9783642178566

This volume constitutes the first of three parts of the refereed proceedings of the First International Conference on Computer Science and Information Technology, CCSIT 2010, held in Bangalore, India, in January 2011. The 59 revised full papers presented in this volume were carefully reviewed and selected. The papers are organized in topical sections on distributed and parallel systems and algorithms; DSP, image processing, pattern recognition, and multimedia; software engineering; database and data Mining; as well as soft computing, such as AI, neural networks, fuzzy systems, etc.

关键词： Computer Communication Networks Artificial Intelligence Information Systems Applications (incl. Internet) Algorithm Analysis and Problem Complexity Software Engineering Information Storage and Retrieval

来源：评论

学校读者我要写书评

暂无评论

A configurable multiprocessor and dynamic load balancing for parallel LU factorization

A configurable multiprocessor and dynamic load balancing for...

引用

International Symposium on parallel and distributed Processing (IPDPS)

作者： X. Wang S.G. Ziavras Department of Electrical and Computer Engineering New Jersey Institute of Technology Newark NJ USA

Summary form only given. The exponentially increasing complexity of many scientific applications and the high cost of supercomputing force us to explore new, sustainable, and affordable high-performance computing platforms. Recent significant advances in FPGA technology and the inherent advantages of configurable logic have brought about new research efforts in the configurable computing field: parallel processing on configurable chips. We explore here parallel LU factorization of large sparse block-diagonal-bordered (BDB) matrices on a configurable multiprocessor that we have designed and implemented. A dynamic load balancing strategy is proposed and analyzed. Performance results for IEEE power test systems are provided. Our research provides evidence that configurable logic can be a viable alternative to high-performance scientific computing.

关键词： Load management Field programmable gate arrays Sparse matrices parallel processing Costs Concurrent computing Logic Scientific computing Equations Supercomputers

来源：评论

学校读者我要写书评

暂无评论

PyTorch distributed: Experiences on Accelerating Data parallel Training

引用

proceedings OF THE VLDB ENDOWMENT 2020年第12期13卷 3005-3018页

作者： Li, Shen Zhao, Yanli Varma, Rohan Salpekar, Omkar Noordhuis, Pieter Li, Teng Paszke, Adam Smith, Jeff Vaughan, Brian Damania, Pritam Chintala, Soumith Facebook AI Birmingham AL 35203 USA Facebook Menlo Pk CA USA Univ Warsaw Warsaw Poland

This paper presents the design, implementation, and evaluation of the PyTorch distributed data parallel module. PyTorch is a widely-adopted scientific computing package used in deep learning research and applications. Recent advances in deep learning argue for the value of large datasets and large models, which necessitates the ability to scale out model training to more computational resources. Data parallelism has emerged as a popular solution for distributed training thanks to its straightforward principle and broad applicability. In general, the technique of distributed data parallelism replicates the model on every computational resource to generate gradients independently and then communicates those gradients at each iteration to keep model replicas consistent. Despite the conceptual simplicity of the technique, the subtle dependencies between computation and communication make it non-trivial to optimize the distributed training efficiency. As of v1.5, PyTorch natively provides several techniques to accelerate distributed data parallel, including bucketing gradients, overlapping computation with communication, and skipping gradient synchronization. Evaluations show that, when configured appropriately, the PyTorch distributed data parallel module attains near-linear scalability using 256 GPUs.

关键词： Large dataset

来源：评论

学校读者我要写书评

暂无评论

parallel Training of a Set of Online Sequential Extreme Learning Machines 41

Parallel Training of a Set of Online Sequential Extreme Lear...

引用

41st International Conference of the Chilean Computer Science Society, SCCC 2022

作者： Gelvez-Almeida, Elkin Barrientos, Ricardo J. Vilches-Ponce, Karina Mora, Marco Doctorado en Modelamiento Matemático Aplicado Universidad Católica Del Maule Talca Chile Facultad de Ciencias Básicas Universidad Simón Bolívar San José de Cúcuta Colombia Laboratorio LITRP Facultad de Ciencias de la Ingeniería Universidad Católica Del Maule Talca Chile Facultad de Ciencias Básicas Universidad Católica Del Maule Talca Chile

ISBN: (纸本)9781665456746

Size databases have constantly increased from advances in technology and the Internet, so processing this vast amount of information has been a great challenge. The neural network Extreme Learning Machine (ELM) have been widely accepted in the scientific community due to their simplicity and good generalization capacity. This model consists of randomly assigning the weights of the hidden layer, and analytically calculating the weights of the output layer through the Moore- Penrose generalized inverse. High-Performance computing has emerged as an excellent alternative for tackling problems involving large-scale databases and reducing processing times. The use of parallel computing tools in Extreme Learning Machines and their variants, especially the Online Sequential Extreme Learning Machine (OS-ELM), has proven to be a good alternative to tackle regression and classification problems with largescale databases. In this paper, we present a parallel training methodology consisting of several Online Sequential Extreme Learning Machines running on different cores of the Central Processing Unit, with a balanced fingerprint database having 2,000,000 samples distributed in five classes. The results show that training and validation times decrease as the number of processes increases since the number of samples to train in each process decreases. In addition, by having several Online Sequential Extreme Learning Machines trained, new samples can beclassified on any of them. © 2022 IEEE.

关键词： Program processors

来源：评论

学校读者我要写书评

暂无评论

A GPGPU algorithm for c-Approximate r-nearest neighbor search in high dimensions

A GPGPU algorithm for c-Approximate r-nearest neighbor searc...

引用

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

作者： Carraher, Lee A. Wilsey, Philip A. Annexstein, Fred S. School of Electronic Computing Systems University of Cincinnati Cincinnati United States

ISBN: (纸本)9780769549798

Nearest Neighbor search is one of the simplest and most intuitive ideas in data mining. Due to it's simplicity and diverse utility, Nearest Neighbor search is often found to be the workhorse of a variety of data mining, machine learning, and computer vision algorithms. For very high dimensional data, the naive linear search tends to be optimal. This is due to the so called curse of dimensionality in database search. To accelerate the search time, may researchers have turned to the power and efficiency of GPGPU computing;mainly for computing the data intensive distance metrics. Though promising speedups are achieved through this method, it fails to capitalize on many of the recent advances in algorithms for bounded error approximate nearest neighbor searches. Approximate nearest neighbor methods give satisfactory results, especially in very large and possibly redundant datasets, while boasting sub-linear search complexities. In this paper we present a c-Approximate r-nearest neighbor search algorithm for CUDA using Locality Sensitive Hash with Nearest Neighbor search (LSH-NN). We implement this system in CUDA and test it on real world image SIFT vector data. Our tests show that we are able to achieve significant speedup over the serial version, with good approximations on scalability, while achieving the near-optimal search complexity of LSH-NN. © 2013 IEEE.

关键词： Nearest neighbor search

来源：评论

学校读者我要写书评

暂无评论

Critical issues in the design of large-scale distributed systems

Critical issues in the design of large-scale distributed sys...

引用

proceedings of the IEEE Workshop on advances in parallel and distributed Systems

作者： S. Howell N.-D. Hoang C. Nguyen N. Karangelen Dahlgren Division Naval Surface Warfare Center USA Trident Systems Inc.

The requirements for the design of large-scale, complex, software-intensive applications include real-time, time-critical, performance, dependability, security, and financial limitations. These requirements have put dramatic burdens on the systems engineers in designing these applications. In order to eliminate the system engineering shortcomings, technology must be developed to: (1) provide a rigorous means for systems engineers to identify and specify the necessary information for the design, (2) provide annotation capabilities for the system design to capture design goals/criteria that relate to the requirements and Measures of Effectiveness (MOEs), and (3) provide mechanisms to optimize the design given the design goals. A framework is described here which provides this ability to assess and optimize system designs and their implementations.

关键词： Large-scale systems Design engineering Design optimization Systems engineering and theory Hardware Space technology Security distributed computing Computer architecture Software safety

来源：评论

学校读者我要写书评

暂无评论

Metrics for distributed C/sup 3/ systems: Rationale and approach

Metrics for distributed C/sup 3/ systems: Rationale and appr...

引用

proceedings of the IEEE Workshop on advances in parallel and distributed Systems

作者： D. Bhatt M. Denz J. Norton K. Kim Honeywell Technology Center Minneapolis MN USA Rome Laboratories NY USA Department of ECE University of California Irvine CA USA

The empirical nature of distributed computing techniques used in developing Command, Control, and Communication (C/sup 3/) systems requires a comprehensive definition, implementation, and usage of metrics to ascertain how the system meets its functional and performance requirements. The additional objectives of adaptive fault-resistance and resource management need new metrics that are dynamically implementable. The authors present rationale/needs of metrics and an approach to their development and usage. This approach is based upon a framework that covers various aspects of metrics including usage across system life-cycle, metrics views across system functions and levels of abstraction, and instrumentation architecture/tools for implementation.

关键词： Instruments System testing Fault tolerant systems Monitoring System performance Debugging Runtime Measurement Force control Control systems

来源：评论

学校读者我要写书评

暂无评论

Scalable parallel and cluster computing

Scalable parallel and cluster computing

引用

International Conference on parallel and distributed Systems (ICPADS)

作者： Kai Hwang

Summary form only given, as follows. In this talk, Dr. Hwang will present two approaches to scalable parallel computing: The first approach is based on scaling STAP benchmarks on Cray T3D, IBM SP2, and Intel Paragon. STAP was originally developed by the MIT Lincoln Laboratory for real-time signal processing. The benchmark suite was recently parallelized and tested on several available MPPs (massively parallel processors). The second approach is to develop an SSI (Single-System-Image) cluster of UNIX workstations and SMP servers for distributed multimedia and Intranet applications. The talk will evaluate scalable multiprocessors and network-based cluster architectures. Specific R/D topics to be covered include the following: - STAP benchmark evaluation of T3D, SP2, Paragon, and Power Challenge/XL - Architectural - Hardware and software support for SSI in a cluster of workstations and servers. - distributed cluster computing in multimedia and Intranet applications. assessment of recent advances in CC-NUMA and Clusters. Scalable parallel computer platforms provide not only an incremental growth path, but also performance scaling and enhanced availability. These features are very much desired in today's business and scientific applications. Dr. Hwang will share the research findings and benchmarking experiences from his research groups at USC and HKU. He will also comment on the research, development, and application trends in parallel and cluster computing.

关键词： Concurrent computing Application software Benchmark testing Workstations Network servers parallel processing Signal processing Computer architecture Hardware

来源：评论

学校读者我要写书评

暂无评论

8th International Conference on Soft computing for Problem Solving, SocProS 2018

8th International Conference on Soft Computing for Problem S...

引用

8th International Conference on Soft computing for Problem Solving, SocProS 2018

ISBN: (纸本)9789811500343

The proceedings contain 163 papers. The special focus in this conference is on Soft computing for Problem Solving. The topics include: Optimization of Target Oriented Network Intelligence Collection for the Social Web by Using k-Beam Search;compressed Air Energy Storage Driven by Wind Power Plant for Water Desalination Through Reverse Osmosis Process;fully Fuzzy Semi-linear Dynamical System Solved by Fuzzy Laplace Transform Under Modified Hukuhara Derivative;comparison of Performance of Four-Element Microstrip Array Antenna Using Electromagnetic Bandgap Structures;model Development for Strength Properties of Laterized Concrete Using Artificial Neural Network Principles;wind Power Forecasting Using parallel Random Forest Algorithm;leukemia Cell Segmentation from Microscopic Blood Smear Image Using C-Mode;Implementation of Exploration in TONIC Using Non-stationary Volatile Multi-arm Bandits;fuzzy Logic-Based Model for Predicting Surface Roughness of Friction Drilled Holes;load Bearing Capacity for a Ferrofluid Squeeze Film in Double Layered Porous Rough Conical Plates;Face Recognition and Classification Using GoogleNET Architecture;Enhancing Public Health Surveillance by Measurement of Similarity Using Rough Sets and GIS;data Analytics Implemented over E-commerce Data to Evaluate Performance of Supervised Learning Approaches in Relation to Customer Behavior;optimal Renewable Energy Resource Based distributed Generation Allocation in a Radial Distribution System;PV Module Temperature Estimation by Using ANFIS;modified Artificial Potential Field Approaches for Mobile Robot Navigation in Unknown Environments;Analysis of BASNs Battery Performance at Different Temperature Conditions Using Artificial Neural Networks (ANN);ASIC Implementation of Fixed-Point Iterative, parallel, and Pipeline CORDIC Algorithm;adaptive Sensor Ranking Based on Utility Using Logistic Regression.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A global computing environment for networked resources

A global computing environment for networked resources

引用

International Conference on parallel Processing (ICPP)

作者： H. Topcuoglu S. Hariri Department of Electrical Engineering and Computer Science Syracuse University Syracuse NY USA

Current advances in high-speed networks and WWW technologies have made network computing a cost-effective, high-performance computing alternative. New software tools are being developed to utilize efficiently the network computing environment. Our project, called Virtual distributed computing Environment (VDCE), is a high-performance computing environment that allows users to write and evaluate networked applications for different hardware and software configurations using a web interface. In this paper we present the software architecture of VDCE by emphasizing application development and specification, scheduling, and execution/runtime aspects.

关键词： Computer networks distributed computing Application software High-speed networks World Wide Web Software tools Computer interfaces Hardware Software architecture Processor scheduling

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：