检索结果-内蒙古大学图书馆

DARPA Information Survivability Conference and Exposition, DISCEX 2000

作者： Pandey, R. Hashii, B. Lal, M. Parallel and Distributed Computing Laboratory Computer Science Department University of California DavisCA95616 United States

ISBN: (纸本)0769504906

There is increasing interest in computing models that support extensibility of systems through code migration. Although appealing both from the system design and extensibility points of view, extensible systems are vulnerable to an external program's aberrant execution behaviors. We examine the problems of resource access control and resource consumption. We propose solutions for these problems and analyze their effectiveness. © 2000 IEEE.

关键词：

来源：评论

学校读者我要写书评

暂无评论

CPU resource control for mobile programs 1

CPU resource control for mobile programs

引用

1st International Symposium on Agent Systems and Applications and 3rd International Symposium on Mobile Agents, ASA/MA 1999

作者： Lal, Manoj Pandey, Raju Computer Science Department Parallel and Distributed Computing Laboratory University of California DavisCA95616 United States

ISBN: (纸本)0769503403

There is considerable interest in developing runtime infrastructures for programs that can migrate from one host to another. Mobile programs are appealing because they support efficient utilization of network resources and extensibility of information servers. This paper presents a scheduling scheme for allocating resources to a mix of real-Time and non real-Time mobile programs. Within this framework, both mobile programs and hosts can specify constraints on how CPU should be allocated. On the basis of the constraints, the scheme constructs a scheduling graph on which it applies several scheduling algorithms. In case of conflicts between mobile program and host specified constraints, the schemes implements a policy that resolves the conflicts in favor of the host. The resulting scheduling scheme is adaptive, flexible, and enforces both program and host specified constraints.

关键词： Scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Area-NeRF: Area-based Neural Radiance Fields 2

Area-NeRF: Area-based Neural Radiance Fields

引用

2nd International Conference on Image Processing, Computer Vision and Machine Learning, ICICML 2023

作者： Ye, Zonxin Li, Wenyu Qiao, Peng Dou, Yong National University of Defense Technology National Key Laboratory of Parallel and Distributed Computing School of Computer Changsha China

ISBN: (纸本)9798350331417

Neural Radiance Field (NeRF) has received widespread attention for its photo-realistic novel view synthesis quality. Current methods mainly represent the scene based on point sampling of ray casting, ignoring the influence of the observed area changing with distance. In addition, The current sampling strategies are all focused on the distribution of sampling points on the ray, without paying attention to the sampling of the ray. We found that the current ray sampling strategy for scenes with the camera moving forward severely reduces the convergence speed. In this work, we extend the point representation to area representation by using relative positional encoding, and propose a ray sampling strategy that is suitable for camera trajectory moving forward. We validated the effectiveness of our method on multiple public datasets. © 2023 IEEE.

关键词： NeRF neural radiance field neural rendering novel view synthesis

来源：评论

学校读者我要写书评

暂无评论

Local-Adaptive Transformer for Multivariate Time Series Anomaly Detection and Diagnosis

Local-Adaptive Transformer for Multivariate Time Series Anom...

引用

2023 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2023

作者： Zhou, Xiaohui Wang, Yijie Xu, Hongzuo Liu, Mingyu Zhang, Ruyi College of Computer National University of Defense Technology National Key Laboratory of Parallel and Distributed Computing Changsha China

ISBN: (纸本)9798350337020

Time series data are pervasive in varied real-world applications, and accurately identifying anomalies in time series is of great importance. Many current methods are insufficient to model long-term dependence, whereas some anomalies can be only identified through long temporal contextual information. This may finally lead to disastrous outcomes due to false negatives of these anomalies. Prior arts employ Transformers (i.e., a neural network architecture that has powerful capability in modeling long-term dependence and global association) to alleviate this problem;however, Transformers are insensitive in sensing local context, which may neglect subtle anomalies. Therefore, in this paper, we propose a local-adaptive Transformer based on cross-correlation for time series anomaly detection, which unifies both global and local information to capture comprehensive time series patterns. Specifically, we devise a cross-correlation mechanism by employing causal convolution to adaptively capture local pattern variation, offering diverse local information into the long-term temporal learning process. Furthermore, a novel optimization objective is utilized to jointly optimize reconstruction of the entire time series and matrix derived from cross-correlation mechanism, which prevents the cross-correlation from becoming trivial in the training phase. The generated cross-correlation matrix reveals underlying interactions between dimensions of multivariate time series, which provides valuable insights into anomaly diagnosis. Extensive experiments on six real-world datasets demonstrate that our model outperforms state-of-the-art competing methods and achieves 6.8%-27.5% F1 score improvement. Our method also has good anomaly interpretability and is effective for anomaly diagnosis. © 2023 IEEE.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A clustering-based approach for mining dockerfile evolutionary trajectories

引用

Science China(Information Sciences) 2019年第1期62卷 211-213页

作者： Yang ZHANG Huaimin WANG Vladimir FILKOV Key Laboratory of Parallel and Distributed Computing National University of Defense Technology College of Computer National University of Defense Technology DECAL Lab University of California Computer Science Department University of California

Dear editor,Docker1), as a de-facto industry standard [1], enables the packaging of an application with all its dependencies and execution environment in a light-weight, self-contained unit, i.e., *** launching the container from Docker image, developers can easily share the same operating system, libraries, and binaries [2]. As the configuration file, the dockerfile plays an important role,

关键词： A clustering-based approach for mining dockerfile evolutionary trajectories

来源：评论

学校读者我要写书评

暂无评论

Using JIT compilation and configurable runtime systems for efficient deployment of java programs on ubiquitous devices

引用

International Conference on Ubiquitous computing, Ubicomp 2001

作者： Teodorescu, Radu Pandey, Raju Parallel and Distributed Computing Laboratory Computer Science Department University of California Davis United States

ISBN: (纸本)3540426140

As the proliferation of ubiquitous devices moves computation away from the conventional desktop computer boundaries, distributed systems design is being exposed to new challenges. A distributed system supporting a ubiquitous computing application must deal with a wider spectrum of hardware architectures featuring structural and functional differences, and resources limitations. Due to its architecture independent infrastructure and object-oriented programming model, the Java programming environment can support flexible solutions for addressing the diversity among these devices. Unfortunately, Java solutions are often associated with high costs in terms of resource consumption, which limits the range of devices that can benefit from this approach. In this paper, we present an architecture that deals with the cost and complexity of running Java programs by partitioning the process of Java program execution between system nodes and remote devices. The system nodes prepare a Java application for execution on a remote device by generating device-specific native code and application-specific runtime system on the fly. The resulting infrastructure provides the flexibility of a high-level programming model and the architecture independence specific to Java. At the same time the amount of resources consumed by an application on the targeted device are comparable to that of a native implementation. © Springer-Verlag Berlin Heidelberg 2001.

关键词： Object oriented programming

来源：评论

学校读者我要写书评

暂无评论

Powerlyra: Differentiated graph computation and partitioning on skewed graphs 15

Powerlyra: Differentiated graph computation and partitioning...

引用

10th European Conference on Computer Systems, EuroSys 2015

作者： Chen, Rong Shi, Jiaxin Chen, Yanzhe Chen, Haibo Shanghai Key Laboratory of Scalable Computing and Systems Institute of Parallel and Distributed Systems Shanghai Jiao Tong University China

ISBN: (纸本)9781450332385

Natural graphs with skewed distribution raise unique challenges to graph computation and partitioning. Existing graph-parallel systems usually use a "one size fits all" design that uniformly processes all vertices, which either suffer from notable load imbalance and high contention for high-degree vertices (e.g., Pregel and GraphLab), or incur high communication cost and memory consumption even for low-degree vertices (e.g., PowerGraph and GraphX). In this paper, we argue that skewed distribution in natural graphs also calls for differentiated processing on highdegree and low-degree vertices. We then introduce Power-Lyra, a new graph computation engine that embraces the best of both worlds of existing graph-parallel systems, by dynamically applying different computation and partitioning strategies for different vertices. PowerLyra further provides an efficient hybrid graph partitioning algorithm (hybrid-cut) that combines edge-cut and vertex-cut with heuristics. Based on PowerLyra, we design locality-conscious data layout optimization to improve cache locality of graph accesses during communication. PowerLyra is implemented as a separate computation engine of PowerGraph, and can seamlessly support various graph algorithms. A detailed evaluation on two clusters using graph-analytics and MLDM (machine learning and data mining) applications show that PowerLyra outperforms PowerGraph by up to 5.53X (from 1.24X) and 3.26X (from 1.49X) for real-world and synthetic graphs accordingly, and is much faster than other systems like GraphX and Giraph, yet with much less memory consumption. A porting of hybrid-cut to GraphX further confirms the efficiency and generality of PowerLyra. Copyright © 2015 ACM 978-1-4503-3238-5/15/04 $15.00.

关键词： Data mining

来源：评论

学校读者我要写书评

暂无评论

A comparative study of star graphs and rotator graphs

A comparative study of star graphs and rotator graphs

引用

23rd International Conference on parallel Processing, ICPP 1994

作者： Ponnuswamy, Subburajan Chaudhary, Vipin Parallel and Distributed Computing Laboratory Dept. of ECE Wayne State University Detroit MI 48202 United States

ISBN: (纸本)0849324939

Star graph is an extensively studied Cayley graph, considered to be an attractive alternative to the popular binary cube. The rotator graphs are a set of directed Cayley graphs introduced recently. In this paper we compare the structural and algorithmic aspects of star graphs with that of rotator graphs. In the process we present some new results for star graphs and rotator graphs. We present a formula for the number of nodes at any distance from the identity permutation in star graphs. The minimum bisection width of star and rotator graphs is obtained. Partitioning and fault tolerant parameters for both star and rotator graphs are analyzed. The node disjoint parallel paths and hence the upper bound on the fault diameter of rotator graphs are presented. We compare the minimal path routing in star and rotator graphs using simulation results. © 1994 IEEE.

关键词： Directed graphs

来源：评论

学校读者我要写书评

暂无评论

Mbapp: Efficient Memory-Balanced Pipeline parallelism for Large Model Fine-Tuning on Commodity GPU Servers 24

Mbapp: Efficient Memory-Balanced Pipeline Parallelism for La...

引用

5th International Conference on Computer Information and Big Data Applications, CIBDA 2024

作者： Liu, Yujie Lai, Zhiquan Li, Dongsheng National Key Laboratory of Parallel and Distributed Computing College of Computer National University of Defense Technology Changsha410000 China

ISBN: (纸本)9798400718106

Large-scale models have demonstrated outstanding performance across various downstream tasks. Pipeline parallelism is essential for fine-tuning large models on commodity GPU servers, as it plays a crucial role in making their exceptional performance more accessible and widespread. Existing approaches have encountered challenges in attaining effective memory-balanced pipeline parallelism. This paper presents a novel memory load-balanced pipeline parallel solution that aims to distribute memory usage evenly across stages on commodity GPU servers by leveraging NVLink bridges. The solution presents a novel approach for transferring data from GPUs to CPUs through the PCI-e link between adjacent GPUs interconnected by the NVLink bridge. Moreover, our methodology orchestrates data transfer operations to reduce offloading latency during the fine-tuning of large models. The evaluation demonstrates that our approach enhances switching efficiency by 1.46x to 1.76x and improves throughput by 2.04x to 3.59x compared to the PyTorch offloading technique. © 2024 ACM.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Rethinking benchmarking for NVM-based file systems 16

Rethinking benchmarking for NVM-based file systems

引用

7th ACM SIGOPS Asia-Pacific Workshop on Systems, APSys 2016

作者： Dong, Mingkai Yu, Qianqian Zhou, Xiaozhou Hong, Yang Chen, Haibo Zang, Binyu Shanghai Key Laboratory of Scalable Computing and Systems Institute of Parallel and Distributed Systems Shanghai Jiao Tong University China

ISBN: (纸本)9781450342650

Fast, byte-addressable NVM promises near-cache latency and near memory bus throughput for file systems, which is evident by recent development of a number of NVM file systems (NVMFS). However, a key approach to measuring their performance, file system benchmarks, remains unchanged as those for disk-based file systems (DFS). In this paper, we analyze the pitfalls of using DFS benchmarks to evaluate NVMFS and propose several changes to existing DFS benchmarks to better characterize the performance of NVMFS. © 2016 ACM.

关键词： Benchmarking

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：