检索结果-内蒙古大学图书馆

IEEE symposium on On-Line Testing (IOLTS)

作者： Tiago A. O. Alves Sandip Kundu Leandro A. J. Marzulo Felipe M. G. França Programa de Engenharia de Sistemas e Computação - COPPE Universidade Federal do Rio de Janeiro (UFRJ) Rio de Janeiro Brazil Department of Electrical and Computer Engineering University of Massachusetts Amherst Instituto de Matemática e Estatística Universidade do Estado do Rio de Janeiro (UERJ) Rio de Janeiro Brazil

The processor industry is well on its way towards manycore processors that comprise of large number of simple cores. The shift towards multi and manycores calls for new programming paradigms suitable for exploiting the inherent parallelism in applications. Dataflow execution was shown to be a good option for programming in such environments. It is well-known that as CMOS technology continues to scale, it becomes more prone to transient and permanent hardware faults. In this paper we present a novel mechanism for error detection and recovery that focuses on transient errors in dataflow execution. Due to the inherently parallel nature of dataflow, our solution is completely distributed and synchronizes only cores that have data dependencies between them, as opposed to prior work on error recovery that in general rely on global synchronization of the system. We evaluate the proposed solution via a software implementation on top of a dataflow runtime. Experimental results show that error detection overhead is highly related to the pressure on the memory bus. In memory bound applications, performance is found to deteriorate, while for other benchmarks, the observed overhead is less than 23%. We find no comparable previous work to contrast these results.

关键词： Transient analysis Benchmark testing Bit error rate Image edge detection parallel processing Runtime Hardware

来源：评论

学校读者我要写书评

暂无评论

Scheduling Coflows for Minimizing the Makespan in Identical parallel Networks

arXiv

引用

arXiv 2023年

作者： Chen, Chi-Yeh Chen, Jun Department of Computer Science and Information Engineering National Cheng Kung University Taiwan

With the rapid advancement of technology, parallel computing applications have become increasingly popular and are commonly executed in large data centers. These applications involve two phases: computation and communication, which are executed repeatedly to complete the work. However, due to the ever-increasing demand for computing power, large data centers are struggling to meet the massive communication demands. To address this problem, coflow has been proposed as a networking abstraction that captures communication patterns in data-parallel computing frameworks. This paper focuses on the coflow scheduling problem in identical parallel networks, where the primary objective is to minimize the makespan, which is the maximum completion time of coflows. It is considered one of the most significant NP-hard problems in large data centers. In this paper, we consider two problems: flow-level scheduling and coflow-level scheduling. In the flow-level scheduling problem, distinct flows can be transferred through different network cores, whereas in the coflow-level scheduling problem, all flows must be transferred through the same network core. To address the flow-level scheduling problem, this paper proposes two algorithms: (formula presented)-approximation algorithm and a (formula presented)-approximation algorithm, where m represents the number of network cores. For the coflow-level scheduling problem, this paper proposes a (2m)-approximation algorithm. Finally, we conduct simulations on our proposed algorithm and Weaver’s algorithm, as presented in Huang et al. (2020) in the 2020 IEEE international parallel and distributed processing symposium (IPDPS). We also validate the effectiveness of the proposed algorithms on heterogeneous parallel networks. Copyright © 2023, The Authors. All rights reserved.

关键词： Approximation algorithms

来源：评论

学校读者我要写书评

暂无评论

Performance Analysis of distributed Deep Learning Frameworks in a Multi-GPU Environment

Performance Analysis of Distributed Deep Learning Frameworks...

引用

international Conference on Ubiquitous Computing and Communications and international symposium on Cyberspace and Security (IUCC-CSS)

作者： Tulasi Kavarakuntla Liangxiu Han Huw Lloyd Annabel Latham Samson B. Akintoye Manchester Metropolitan University Manchester UK

ISBN: (纸本)9781665466684

Deep Learning frameworks, such as TensorFlow, MXNet, Chainer, provide many basic building blocks for designing effective neural network models for various applications (e.g. computer vision, speech recognition, natural language processing). However, run-time performance of these deep learning frameworks varies significantly even when training identical deep network models on the same GPUs. This study presents an experimental analysis and performance model for assessing deep learning models (Convolutional Neural Networks (CNNs), Multilayer Perceptrons (MLP), Autoencoder) on three frameworks: TensorFlow, MXNet, and Chainer, in a multi-GPU environment. We analyse factors that influence these frameworks' performance by computing the running time of each framework in our proposed model, taking load imbalance factor into account. The evaluation results highlight significiant differences in the scalability of the frameworks, and the importance of load balance in parallel distributed deep learning.

关键词： Deep learning Training Analytical models Computational modeling Scalability Neural networks Speech recognition

来源：评论

学校读者我要写书评

暂无评论

PhD Forum: Efficient Privacy-Preserving processing via Memory-Centric Computing

PhD Forum: Efficient Privacy-Preserving Processing via Memor...

引用

Reliable distributed Systems

作者： Mpoki Mwaisela University of Neuchâtel Neuchâtel Switzerland

ISBN: (数字)9798331530037

ISBN: (纸本)9798331530044

Privacy-preserving computation techniques like homomorphic encryption (HE) and secure multi-party computation (SMPC) enhance data security by enabling processing on encrypted data. However, the significant computational and CPU-DRAM data movement overhead resulting from the underlying cryptographic algorithms impedes the adoption of these techniques in practice. Existing approaches focus on improving computational overhead using specialized hardware like GPUs and FPGAs, but these methods still suffer from the same processor-DRAM overhead. Novel hardware technologies that support in-memory processing have the potential to address this problem. Memory-centric computing, or processing-in-memory (PIM), brings computation closer to data by introducing low-power processors called data processing units (DPUs) into memory. Besides its in-memory computation capability, PIM provides extensive parallelism, resulting in significant performance improvement over state-of-the-art approaches. We propose a framework that uses recently available PIM hardware to achieve efficient privacy-preserving computation. Our design consists of a four-layer architecture: (1) an application layer that decouples privacy-preserving applications from the underlying protocols and hardware; (2) a protocol layer that implements existing secure computation protocols (HE and MPC); (3) a data orchestration layer that leverages data compression techniques to mitigate the data transfer overhead between DPUs and host memory; (4) a computation layer which implements DPU kernels on which secure computation algorithms are built.

关键词： Privacy Protocols Program processors Scalability Computer architecture parallel processing Hardware Multi-party computation Reliability Homomorphic encryption

来源：评论

学校读者我要写书评

暂无评论

Elastic Allocator: An Adaptive Task Scheduler for Streaming Query in the Cloud

Elastic Allocator: An Adaptive Task Scheduler for Streaming ...

引用

2014 IEEE 8th international symposium on Service Oriented System Engineering

作者： Zheng Han Rui Chu Haibo Mi Huaimin Wang Science and Technology on Parallel and Distributed Processing Laboratory National University of Defense Technology Changsha Hunan China

Many big data applications receive and process data in real time. These data, also known as data streams, are generated continuously and processed online in a low latency manner. Data stream is prone to change dramatically in volume, since its workload may have a variation of several orders between peak and valley periods. Fully provisioning resources for stream processing to handle the peak load is costly, while over-provisioning is wasteful when to deal with lightweight workload. Cloud computing emphasizes that resource should be utilized economically and elastically. An open question is how to allocate query task adaptively to keeping up the input rate of the data stream. Previous work focuses on using either local or global capacity information to improve the cluster CPU resource utilization, while the bandwidth utilization which is also critical to the system throughput is ignored or simplified. In this paper, we formalize the operator placement problem considering both the CPU and bandwidth usage, and introduce the Elastic Allocator. The Elastic Allocator uses a quantitative method to evaluate a node's capacity and bandwidth usage, and exploit both the local and global resource information to allocate the query task in a graceful manner to achieve high resource utilization. The experimental results and a simple prototype built on top of Storm finally demonstrate that Elastic Allocator is adaptive and feasible in cloud computing environment, and has an advantage of improving and balancing system resource utilization.

关键词： Bandwidth Resource management Clustering algorithms Storms Linear programming Cloud computing Computer architecture

来源：评论

学校读者我要写书评

暂无评论

An Implementation of Quarter Pixel Block Motion Estimation Using SIMD

An Implementation of Quarter Pixel Block Motion Estimation U...

引用

2004 international symposium on distributed Computing and applications to Business, Engineering and Science

作者： Xinchen Zhang Ruimin Hu Deren Li Zhongyuan Wang The Key Laboratory of Multimedia and Network Communications Engineering, Hubei Province,Wuhan University Wuhan,Hubei 430072,China

For improving video coding efficiency,sub-pixel motion estimation(ME) is used extensively in the existing video coding *** quarter pixel ME is one of the high complexity tools in H.264/*** this paper,a parallel quarter block motion estimation algorithm that not only accelerates the process of sub-pixel motion estimation but also maintains accuracy as that of the original algorithm is *** Intel P4 CPU,the SIMD(single instruction multiple data) technique is commonly used to provide an execution *** implementation of this algorithm using parallel processing on P4 platform is *** proposed algorithm satisfies in particular the requirements of low-rate real-timed video *** results show that the optimized video encoder is more than 13.5 times faster than the original reference software while keeping the accuracy of the latter approximately.

关键词： Sub-pixel Block Motion Estimation Computation Complexity parallel Algorithm SIMD H.264/AVC

来源：评论

学校读者我要写书评

暂无评论

FLOWGAN:Unbalanced Network Encrypted Traffic Identification Method Based on GAN

FLOWGAN:Unbalanced Network Encrypted Traffic Identification ...

引用

IEEE international Conference on Big Data and Cloud Computing (BdCloud)

作者： ZiXuan Wang Pan Wang Xiaokang Zhou ShuHang Li MoXuan Zhang School of Modern Posts University of Nanjing University of Posts and Telecommunications Nanjing China Faculty of Data Science Shiga University Hikone Japan RIKEN Center for Advanced Intelligence Project Tokyo Japan Schools of International Education Jinling Institute of Technology Nanjing China

ISBN: (数字)9781728143286

ISBN: (纸本)9781728143293

It is crucial to accurately identify the type of traffic and application so that it can enable various policy-driven network management and security monitoring. However, with the increasing adoption of Internet applications use encryption protocols to transmit data, traffic classification is becoming more difficult. Although existing machine learning methods and novel deep learning methods have many advantages, which can solve the drawbacks of port and payload based methods, but there are still some shortcomings, one of which is the imbalanced property of network traffic data. In this paper, we proposed a GAN based method called FlowGAN to tackle with the problem of class imbalance for traffic classification. As an instance of Generative Adversarial Network (GAN), FlowGAN leverages the superiority of GAN's data augmentation to produce synthetic traffic data for classes with few samples. Furthermore, we trained a classical deep learning model, Multilayer perceptron (MLP) based network traffic classifier to evaluate the performance of FlowGAN. Based on the public dataset 'ISCX', our experimental results show that our proposed FlowGAN can outperform an unbalanced dataset and balancing dataset by the oversampling method in terms of data augmentation. Based on the public dataset ISCX, our experimental results show that the recognition performance of FlowGAN on small samples, compared with the unbalanced dataset, Precision, Recall, and F1-score increased by 13.2%, 17.0%, and 15.6% on average, compared with the balanced dataset Precision, Recall, F1-score increased by 2.15%, 2.06%, 2.12% on average.

关键词： Cryptography Gallium nitride Generative adversarial networks Training Generators Deep learning Feature extraction

来源：评论

学校读者我要写书评

暂无评论

parallel Reservoir Integrated Simulation Platform For One Million Grid Blocks Case

Parallel Reservoir Integrated Simulation Platform For One Mi...

引用

2004 international symposium on distributed Computing and applications to Business, Engineering and Science

作者： Pan Feng,Cao Jianwen,Sun Jiachang Research & Development Center for parallel Software, Institute of Software,Chinese Academy of Sciences P.O.Box 8718,Beijing,P.R.China

This article first provides a brief introduction to the numerical reservoir simulation and a parallel numerical reservoir integrated simulation platform from RDCPS(Research & Development Center for parallel Software,Institute of Software,Chinese Academy of Sciences),including Pre-processing,Simulator(for a Three-Dimensional & Three-Phase Black Oil models),Post processing,seamlessly integrated with parallel *** then present key technologies of the simulator,such as the nonlinear and linear solvers,communications among processors,parallel I/O,etc., and corresponding ***,some results with the platform to solve one million grid blocks cases from Chinese oil fields will be given in the article,which can show that the simulator has a very robust portability,high-speed for deadline and good scalability for the tested *** application software,our object is always focusing on meeting deadlines from oil ***,for one million grid blocks' case with 20～30 years production,its elapsed time with 16 processors is less than 12 hours on parallel computers based on Myrinet or QsNet,namely"to submit a case just before off-duty and get its result just before on-duty".A decreasing line of elapsed time appears for a one million grid blocks *** developing trace of the simulator along with parallel computers can be also inferred.

关键词： parallel numerical reservoir simulation fine residual-oil distribution one million grid blocks problem integrated simulation platform parallel computer deadline grid computing

来源：评论

学校读者我要写书评

暂无评论

Performance evaluation of VBR traffic in InfiniBand

Performance evaluation of VBR traffic in InfiniBand

引用

Canadian Conference on Electrical and Computer Engineering (CCECE)

作者： F.J. Alfaro J.L. Sanchez L. Orozco J. Duato Department de Informática Escuela Politécnica Superior Uni Albacete Spain School of Information Technology and Engineering University of Ottawa Ottawa ONT Canada Department de Information de Sistemas y Computadores University Politéchnica de Valencia Valencia Spain

The InfiniBand architecture (IBA) is becoming an industry standard both for communication between processing nodes and I/O devices and for interprocessor communication. It replaces the traditional I/O bus with a switch-based interconnect for connecting processing nodes and I/O devices. It is being developed by the InfiniBand/sup SM/ Trade Association (IBTA) to provide the levels of reliability, availability, performance, scalability, and quality of service (QoS) necessary for present and future server systems. For this, IBA provides a series of mechanisms that are able to guarantee QoS to the applications. Alfaro, Sanchez and Das (see Proceedings of international parallel and distributed processing symposium, April 2002) proposed a strategy to compute the InfiniBand arbitration tables. We only evaluated our proposal for CBR traffic with fixed mean bandwidth requirements. We evaluate our strategy to compute the InfiniBand arbitration tables with VBR traffic. Performance results show that, this class of traffic also gets their QoS requirements.

关键词： Quality of service Communication industry Communication standards Communication switching Joining processes Samarium Availability Scalability distributed processing Concurrent computing

来源：评论

学校读者我要写书评

暂无评论

Slipstream execution mode for CMP-based multiprocessors

Slipstream execution mode for CMP-based multiprocessors

引用

IEEE symposium on High-Performance Computer Architecture

作者： K.Z. Ibrahim G.T. Byrd E. Rotenberg Department of Electrical and Computer Engineering North Carolina State University USA

Scalability of applications on distributed shared-memory (DSM) multiprocessors is limited by communication overheads. At some point, using more processors to increase parallelism yields diminishing returns or even degrades performance. When increasing concurrency is futile, we propose an additional mode of execution, called slipstream mode, that instead enlists extra processors to assist parallel tasks by reducing perceived overheads. We consider DSM multiprocessors built from dual-processor chip multiprocessor (CMP) nodes with shared L2 cache. A task is allocated on one processor of each CMP node. The other processor of each node executes a reduced version of the same task. The reduced version skips shared-memory stores and synchronization, running ahead of the true task. Even with the skipped operations, the reduced task makes accurate forward progress and generates an accurate reference stream, because branches and addresses depend primarily on private data. Slipstream execution mode yields two benefits. First, the reduced task prefetches data on behalf of the true task. Second, reduced tasks provide a detailed picture of future reference behavior, enabling a number of optimizations aimed at accelerating coherence events, e.g., self-invalidation. For multiprocessor systems with up to 16 CMP nodes, slipstream mode outperforms running one or two conventional tasks per CMP in 7 out of 9 parallel scientific benchmarks. Slipstream mode is 12-19% faster with prefetching only and up to 29% faster with self-invalidation enabled.

关键词： Concurrent computing parallel processing Scalability Degradation Prefetching Performance gain Acceleration Coherence Multiprocessing systems Character generation

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：