检索结果-内蒙古大学图书馆

26th ieee International conference on network Protocols (INCP)

作者： Ben-Basat, Ran Chen, Xiaoqi Einziger, Gil Rottenstreich, Ori Technion Haifa Israel Princeton Univ Princeton NJ 08544 USA Nokia Bell Labs Holmdel NJ USA

ISBN: (纸本)9781538660430

Programmable network switches promise flexibility and high throughput, enabling applications such as load balancing and traffic engineering. network measurement is a fundamental building block for such applications, including tasks such as the identification of heavy hitters (largest flows) or the detection of traffic changes. However, high-throughput packet processing architectures place certain limitations on the programming model, such as restricted branching, limited capability for memory access, and a limited number of processing stages. These limitations restrict the types of measurement algorithms that can run on programmable switches. In this paper, we focus on the RMT programmable high-throughput switch architecture, and carefully examine its constraints on designing measurement algorithms. We demonstrate our findings while solving the heavy hitter problem. We introduce PRECISION, an algorithm that uses Probabilistic Recirculation to find top flows on a programmable switch. By recirculating a small fraction of packets, PRECISION simplifies the access to stateful memory to conform with RMT limitations and achieves higher accuracy than previous heavy hitter detection algorithms that avoid recirculation. We also analyze the effect of each architectural constraint on the measurement accuracy and provide insights for measurement algorithm designers.

关键词： Pipelines Computer architecture Throughput Frequency estimation programming Random access memory Approximation algorithms

来源：评论

学校读者我要写书评

暂无评论

Chameleon: A Generalized Reconfigurable open-Source Architecture for Deep Neural network Training

Chameleon: A Generalized Reconfigurable Open-Source Architec...

引用

ieee High Performance Extreme Computing conference (HPEC)

作者： Isakov, Mihailo Ehret, Alan Kinsy, Michel Boston Univ Dept Elect & Comp Engn Boston MA 02215 USA

ISBN: (纸本)9781538659892

We present Chameleon, an architecture for neural network training and inference design exploration on FPGAs. While there exists a great number of different network types and optimizations, it is not always clear how these differences impact the hardware implementation of neural networks. Chameleon is created with extensibility and experimentation in mind, supporting a number of activations, neuron types, and signal bitwidths. Furthermore, Chameleon is created to be modular, allowing designers to easily swap existing parts for improved ones, speeding up research iteration. While there exists a large number of inference architectures, we focus on speeding up training, as training time is the bottleneck for neural network architecture exploration. Chameleon therefore aims to help researchers better understand the bottlenecks in training deep neural networks and create models that circumvent these barriers. Finally, Chameleon is designed to be simple, without requiring a compiler or reconfiguration to function. This allows quick localized changes to the architecture and facilitates design exploration. We present synthesis results on an Altera Cyclone V SoC and show the design resource usage. We finish with an evaluation by training a network on the Wisconsin Breast Cancer dataset. The RTL and synthesis files for the architecture will be open-sourced upon publication at http://***/research/abc/chameleon/***.

关键词： neural networks training hardware architecture FPGA RTL design exploration open-source

来源：评论

学校读者我要写书评

暂无评论

Large-Scale Whale Call Classification Using Deep Convolutional Neural network architectures 8

Large-Scale Whale Call Classification Using Deep Convolution...

引用

ieee International conference on Signal Processing, Communications and Computing (ICSPCC)

作者： Wang, Dezhi Zhang, Lilun Lu, Zengquan Xu, Kele Natl Univ Def Technol Coll Meteorol & Oceanog Changsha Hunan Peoples R China Natl Univ Def Technol Sch Comp Changsha Hunan Peoples R China

ISBN: (纸本)9781538679463

As the rapid development of deep learning techniques, extensive interest has been taken into the applications of deep learning methods on challenging problems of different domains. In view of the recent success of convolutional neural network (CNN) in various tasks of audio analysis, a comparative performance study of different the-state-of-the-art CNN architectures on a large-scale whale-call classification task is investigated in this paper. On the basis of deep neural network models, distinctive features of whale sub-populations are extracted to obtain higher level abstract representations for the accurate classification, which is significantly superior to the traditional classification approaches using manual features based on expert knowledge. In particular, a large open-source acoustic dataset recorded by audio sensors carried by whales in different locations is employed for performance comparison. Based on the experiments, it is found that the advancement of popular CNN architectures significantly improve the accuracy on the whale call classification task. The accuracy and computational efficiency varies with the change of the CNN architectures. Xception provides the best performance among all four CNN architectures while an ensemble of CNN models can produce even better results.

关键词： whale call classification convolutional neural network deep learning

来源：评论

学校读者我要写书评

暂无评论

All-optical Routers Modeled through the Matrix Method with NVidia CUDA Development Framework

All-optical Routers Modeled through the Matrix Method with N...

引用

CHILEAN conference on Electrical, Electronics Engineering, Information and Communication Technologies (CHILECON)

作者： David Zabala-Blanco Cesar Augusto Azurdia-Meza Ali Dehghan Firoozabadi Pablo Palacios Játiva Department of Computing and Industries Universidad Católica del Maule Talca Chile Department of Electrical Engineering Universidad de Chile Santiago Chile Department of Electricity Universidad Tecnologica Metropolitana Santiago Chile

ISBN: (纸本)9781728131863

Optical packet switching (OPS) is a promising technology in order to satisfy the ever-increasing need for bandwidth. With this technology, data packets, which consist of a header and a payload, are assembled and transmitted over a wavelength division multiplexing network. Headers are processed electronically in the routers that comprise the network. This processing must be carried out as quickly and efficiently as possible in order to avoid packet loss. On the other hand, parallel computing has recently been extended and popularized thanks to the NVidia compute unified device architecture (CUDA) development framework. In this programming paradigm, several threads run the same code simultaneously in order to reduce the overall processing time. This work shows that an optical routing algorithm for OPS can perform better under parallel execution, depending on the amount of data to be processed. A routing method based on simple matrices is presented, and the computation time between a traditional sequential programming languages (C++), and CUDA C is presented. Other performance metrics related to the router dimensioning are also considered.

关键词： optical switches packet switching parallel architectures telecommunication computing telecommunication network routing wavelength division multiplexing Matrix methods telecommunication computing Parallel architectures telecommunication network routing Packet switching Optical switches Wavelength division multiplexing Routers Header computational time parallel execution

来源：评论

学校读者我要写书评

暂无评论

Comparison of Different Convolutional Neural network architectures for Satellite Image Segmentation 23

Comparison of Different Convolutional Neural Network Archite...

引用

23rd conference of open Innovations Association FRUCT

作者： Khryashehev, Vladimir Ivanovsky, Leonid Pavlov, Vladimir Ostrovskaya, Anna Rubtsov, Anton PG Demidov Yaroslavl State Univ Yaroslavl Russia RUDN Univ Peoples Friendship Univ Russia Moscow Russia Russian Space Syst Moscow Russia

ISBN: (纸本)9789526865362

Convolutional neural networks for detection geoobjects on the satellite images from DSTL, Landsat-8 and PlanetScope databases were analyzed. Three modification of convolutional neural network architecture for implementing the recognition algorithm was used. Images obtained from the Landsat-8 and PlanetScope satellites are used for estimation of automatic object detection quality. To analyze the accuracy of the object detection algorithm, the selected regions were compared with the areas by previously marked by experts. An important result of the study was the improvement of the detector for the class "Forest". Segmentation of satellite images has found application at urban planning, forest management, climate modelling, etc.

关键词： Erbium Economic indicators

来源：评论

学校读者我要写书评

暂无评论

Ultra-Reliable and Energy Efficient Wireless Sensor networks 19

Ultra-Reliable and Energy Efficient Wireless Sensor Networks

引用

ieee 19th Wireless and Microwave Technology conference (WAMICON)

作者： Sulieman, Nabeel I. Gitlin, Richard D. Univ S Florida Dept Elect Engn Tampa FL 33620 USA

ISBN: (纸本)9781538612675

Near instant link failure recovery and lower energy consumption can be achieved via the synergistic combination of Diversity and modified Triangular network Coding (eDC-NC), an open-loop coding technique, in a wide variety of network architectures. In this paper, eDC-NC is applied to Wireless Sensor networks (WSNs) to enable very rapid recovering from wireless link/node failures and provide low computational and energy cost, which are very important metrics for WSNs. It is shown that utilizing eDC-NC coding in WSNs can provide ultra-reliability with very rapid fault recovery, decrease energy consumption, and increase the network throughput for broadcasting applications.

关键词： Diversity Coding Triangular network Coding reliability throughput Wireless Sensor network

来源：评论

学校读者我要写书评

暂无评论

programming Support for Sharing Resources Across Heterogeneous Mobile Devices 5

Programming Support for Sharing Resources Across Heterogeneo...

引用

5th ACM/ieee International conference on Mobile Software Engineering and Systems (MOBILESoft)

作者： Song, Zheng Chadha, Sanchit Byalik, Antuan Tilevich, Eli Virginia Tech Software Innovat Lab Blacksburg VA 24061 USA

ISBN: (纸本)9781450357128

Modern mobile users commonly use multiple heterogeneous mobile devices, including smartphones, tablets, and wearables. Enabling these devices to seamlessly share their computational, network, and sensing resources has great potential benefit. Sharing resources across collocated mobile devices creates mobile device clouds (MDCs), commonly used to optimize application performance and to enable novel applications. However, enabling heterogeneous mobile devices to share their resources presents a number of difficulties, including the need to coordinate and steer the execution of devices with dissimilar network interfaces, application programming models, and system architectures. In this paper, we describe a solution that systematically empowers heterogeneous mobile devices to seamlessly, reliably, and efficiently share their resources. We present a programming model and runtime support for heterogeneous mobile device-to-device resource sharing. Our solution comprises a declarative domain-specific language for device-to-device cooperation, supported by a powerful runtime infrastructure. we evaluated our solution by conducting a controlled user study and running performance/energy efficiency benchmarks. The evaluation results indicate that our solution can become a practical tool for enhancing the capabilities of modern mobile applications by leveraging the resources of nearby mobile devices.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

Reliable and Resilient Coordinated Multi Point Fronthaul networks 19

Reliable and Resilient Coordinated Multi Point Fronthaul Net...

引用

ieee 19th Wireless and Microwave Technology conference (WAMICON)

作者： Sulieman, Nabeel I. Balevi, Eren Gitlin, Richard D. Univ S Florida Dept Elect Engn Tampa FL 33620 USA

ISBN: (纸本)9781538612675

Very fast link failure recovery and high throughput can be achieved via the synergistic combination of Diversity and network Coding (DC-NC), an open-loop coding technique, in a wide variety of network architectures. In this paper, DC-NC is applied to enable recovering from a wireless link/node failure in the fronthaul portion of downlink Coordinated Multi Point (CoMP) 5G wireless network in a C-RAN environment. Our results demonstrate that by utilizing DC-NC coding in CoMP systems, resource consumption is reduced by about one-third and ultra-reliability with near-instantaneous fault recovery is achieved.

关键词： 5G CoMP Diversity Coding fronthaul network Coding reliability throughput

来源：评论

学校读者我要写书评

暂无评论

Scalability Issues and Solutions for Software Defined networks

引用

ieee JOURNAL ON SELECTED AREAS IN COMMUNICATIONS 2018年第12期36卷 2595-2602页

作者： Hohlfeld, Oliver Kempf, James Reisslein, Martin Schmid, Stefan Shah, Nadir Rhein Westfal TH Aachen COMSYS Network Architectures Grp D-52074 Aachen Germany Ericsson Res Santa Clara CA 95054 USA Arizona State Univ Sch Elect Comp & Energy Engn Tempe AZ 85287 USA Univ Vienna Fac Comp Sci A-1090 Vienna Austria COMSATS Inst Informat Technol Dept Comp Sci Wah Cantt 47040 Pakistan

Software Defined networking (in short SDN, which is also an acronym for Software Defined network), has emerged as a response to the limitations and complexities of traditional network architectures. At the heart of SDN lies the idea to consolidate the control over network devices into a logically centralized (software) controller separated from the data plane. The separation of the control plane and the data plane is realized via an open programming interface between the data plane switches and the SDN controller. The decoupling allows the control plane to evolve independently of the data plane, which enables faster innovation since software often exceeds hardware in innovation speed. Furthermore, logical centralization has the potential to simplify network operation and management by providing a single focal point where the consequences of management actions can be assessed, and possibly rejected if they would lead to some violation of operational constraints. openFlow, the standard SDN protocol today, is based on a simple match-action paradigm which results in great flexibilities, e.g., in terms of traffic engineering, definition of flows, as well as in-band network control functionalities.

关键词： Special issues and sections Software defined networking Scalability

来源：评论

学校读者我要写书评

暂无评论

CellS: A cell-inspired efficient software framework for Intelligent System 5

CellS: A cell-inspired efficient software framework for Inte...

引用

ieee International conference on Consumer Electronics-Taiwan (ICCE-TW)

作者： Wu, Mu-Che Chen, Ching-Han Natl Cent Univ Taoyuan Taiwan

ISBN: (纸本)9781538663011

Imagine next generation system can talk with people, move around, aware of and respond to changes in the environment, and learn from data. This system needs a lot of sensors and effectors in its hardware. In its software, it must have the capability to handle sensor data autonomously and to generate appropriate responses. This kind of system exactly is BDI (Belief, Desire, and Intention) system. However, traditional BDI system is too difficult to develop a software system, and it needs more cost in maintaining and developing software. To reduce the cost and increase efficiency, we propose a cell-inspired efficient software framework called the CellS. The framework implementation adopted the concept of traditional BDI system refactored by the new cell layer. The cell design can illustrate the BDI concept specifically. Software architectures continue evolvement, from structured programming to object-oriented programming, from single processor programming to multiprocessor programming, from passive calls to active execution. Our framework adopted dynamic parallel mechanism, autonomic BDI and Java language. It suits to all java development environments, and developer just need to learn basic BDI concept. The CellS has the following advantages: 1) high performance;and 2) high scalability. An empirical experimental evaluation demonstrates that the CellS work well in performance.

关键词： Software Brain cells Task analysis Computer architecture programming Java Computational modeling"

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：