检索结果-内蒙古大学图书馆

25th international conference on parallel and distributed Computing (Euro-Par)

作者： Ulm, Gregor Gustavsson, Emil Jirstrand, Mats Fraunhofer Chalmers Res Ctr Ind Math Chalmers Sci Pk S-41288 Gothenburg Sweden Fraunhofer Ctr Machine Learning Chalmers Sci Pk S-41288 Gothenburg Sweden

ISBN: (纸本)9783030483401;9783030483395

OODIDA (On-board/Off-board distributed Data Analytics) is a platform for distributing and executing concurrent data analytics tasks. It targets fleets of reference vehicles in the automotive industry and has a particular focus on rapid prototyping. Its underlying message-passing infrastructure has been implemented in Erlang/OTP. External Python applications perform data analytics tasks. Most work is performed by clients (on-board). A central cloud server performs supplementary tasks (off-board). OODIDA can be automatically packaged and deployed, which necessitates restarting parts of the system, or all of it. This is potentially disruptive. To address this issue, we added the ability to execute user-defined Python modules on clients as well as the server. These modules can be replaced without restarting any part of the system and they can even be replaced between iterations of an ongoing assignment. This facilitates use cases such as iterative A/B testing of machine learning algorithms or modifying experimental algorithms on-the-fly.

关键词： distributed computing Code replacement Erlang

来源：评论

学校读者我要写书评

暂无评论

Private Data System Enabling Self-Sovereign Storage Managed by Executable Choreographies 17th

Private Data System Enabling Self-Sovereign Storage Managed ...

引用

17th IFIP WG 6.1 international conference on distributed applications and Interoperable Systems (DAIS) Held as Part of the 12th international Federated conference on distributed Computing techniques (DisCoTec)

作者： Alboaie, Sinica Cosovan, Doina Alexandru Ioan Cuza Univ Iasi Romania Tech Univ Cluj Napoca Iasi Romania

ISBN: (纸本)9783319596655;9783319596648

With the increased use of Internet, governments and large companies store and share massive amounts of personal data in such a way that leaves no space for transparency. When a user needs to achieve a simple task like applying for college or a driving license, he needs to visit a lot of institutions and organizations, thus leaving a lot of private data in many places. The same happens when using the Internet. These privacy issues raised by the centralized architectures along with the recent developments in the area of serverless applications demand a decentralized private data layer under user control. We introduce the Private Data System (PDS), a distributed approach which enables self-sovereign storage and sharing of private data. The system is composed of nodes spread across the entire Internet managing local key-value databases. The communication between nodes is achieved through executable choreographies, which are capable of preventing information leakage when executing across different organizations with different regulations in place. The user has full control over his private data and is able to share and revoke access to organizations at any time. Even more, the updates are propagated instantly to all the parties which have access to the data thanks to the system design. Specifically, the processing organizations may retrieve and process the shared information, but are not allowed under any circumstances to store it on long term. PDS offers an alternative to systems that aim to ensure self-sovereignty of specific types of data through blockchain inspired techniques but face various problems, such as low performance. Both approaches propose a distributed database, but with different characteristics. While the blockchain-based systems are built to solve consensus problems, PDS's purpose is to solve the self-sovereignty aspects raised by the privacy laws, rules and principles.

关键词： Privacy Enhancing Technique Privacy by Design Privacy by Default Data self-sovereignty Privacy Private data distributed storage Executable choreography

来源：评论

学校读者我要写书评

暂无评论

A Balanced Allocation strategy for file assignment in parallel I/O systems

A Balanced Allocation strategy for file assignment in parall...

引用

IEEE international conference on Networking, Architecture and Storage

作者： Zhu, Yongqing Yu, Yang Wang, Wilson Yonghong Tan, Sophia Sufui Low, Tee Chong Data Storage Institute A STAR Singapore Singapore

ISBN: (纸本)9780769541341

In parallel I/O systems, fast response to disk access and load balancing are two important performance objectives pursued by end users and applications. These performances are largely determined by the data allocation strategies, or file assignment algorithms. However, most existing algorithms can only obtain one of the performance objectives, including the wellknown algorithms like Greedy, Sort Partition (SP) and Hybrid Partition (HP). New algorithms that can achieve both objectives are very necessary for parallel I/O systems. In this paper, we have proposed two new allocation algorithms for file assignment in parallel I/O systems: an offline Balanced Allocation with Sort (BAS) algorithm and an online Balanced Allocation with Sort for Batch (BASB) algorithm. Both algorithms aim to achieve the optimal mean response time and load balancing at the same time. The experiment results have shown that the BAS algorithm can get the optimal performance on response time among all compared algorithms and better performance on load balancing than SP. For online BASB algorithm, it can achieve the optimal performances on both response time and load balancing among all algorithms for comparison. © 2010 IEEE.

关键词： Online systems

来源：评论

学校读者我要写书评

暂无评论

Template for non-uniform parallel loops based on dynamic scheduling and prefetching techniques

Template for non-uniform parallel loops based on dynamic sch...

引用

Proceedings of the 1996 international conference on Supercomputing

作者： Orlando, Salvatore Perego, Raffaele Universita Ca' Foscari di Venezia Venezia Mestre Italy

In this paper we present an efficient template for the implementation on distributed-memory multiprocessors of non-uniform parallel loops, i.e. loops whose independent iterations are characterized by highly varying execution times. The template relies upon a static blocking distribution of array data sets to exploit locality, and a hybrid scheduling policy to smooth uneven processor finishing time. It initially adopts a static scheduling technique based on the owner computes rule. As soon as a workload imbalance is detected, it exploits a dynamic receiver initiated technique to move work towards unloaded processors. Prefetching is used to reduce overheads due to the communications needed to monitor the load, move iterations, and restore the consistency of migrated data. Accurate performance costs of the technique can be derived, thus allowing the template to be used by a compiler to generate well-balanced code for non-uniform parallel loops. Experiments were conducted on a 64-node Cray T3D, and the performance of the proposed template was compared with the one obtained by using the CRAFT-Fortran language (an HPF-like language).

关键词： Data processing

来源：评论

学校读者我要写书评

暂无评论

LoCaF: Detecting Real-World States with Lousy Wireless Cameras

LoCaF: Detecting Real-World States with Lousy Wireless Camer...

引用

8th IEEE international conference on distributed Computing in Sensor Systems (DCOSS)/IWSN/WiSARN/PWSN

作者： Meyer, Benjamin Mietz, Richard Roemer, Kay Univ Lubeck Inst Comp Engn Lubeck Germany

ISBN: (纸本)9780769547077

The Internet of Things (IoT) integrates wireless sensors to provide online and real-time access to the state of things and places. However, many interesting real-world states are difficult to detect with traditional scalar sensors. Tiny wireless camera sensor nodes are an interesting alternative as a single camera can observe a large area in great detail. However, low image resolution, poor image quality, and low frame rates as well as varying lighting conditions in outdoor scenarios make the detection of real-world states using these lousy cameras a challenging problem. In this paper we introduce a framework that addresses this problem by providing an end-to-end solution that includes energy-efficient image capture, image enhancement to mitigate low picture quality, object detection with low frame rates, inference of high-level states, and publishing of these states on the IoT. The framework can be flexibly configured by end-users without programming skills and supports a variety of different applications.

关键词： Frame Rate Cameras World government wireless PROGRAMMING techniques image resolution End users Publishing Sensors Image quality Frameworks

来源：评论

学校读者我要写书评

暂无评论

A simple and efficient parallel approach to large-scale railway freight data analysis

A simple and efficient parallel approach to large-scale rail...

引用

IEEE international conference on High Performance Computing and Communication

作者： Xie, Yi Li, Xiaolin Liu, Fengzhi Institute of Computing Technology Chinese Academy Sciences Beijing China Civil and Environment School University of Science and Technology Beijing China

ISBN: (纸本)9780769542140

Analysis of rail freight data is essential to make a better strategy for railway freight traffic organization. However, rail freight data are national-widely stored in data center of railway bureaus. Furthermore, the size of the data is very large and grows rapidly. Although analytical data processing has received much attention these years, little attention is paid to transparently processing data stored in remote data sources, particularly where the data are generated. In this paper, we present a simple and efficient parallel approach, RFDPS, to large-scale rail freight data processing. RFDPS provides a novel virtualization data model to integrate geographically distributed freight data. Besides, it can parallelize and optimize data processing automatically across multiple data center of railway bureaus. We run typical rail freight analysis applications in production environment to evaluate the efficiency of RFDPS. The results show that the performance of RFDPS is improved exponentially compared to the approach China's Ministry of Railways adopted before. © 2010 IEEE.

关键词： Data integration

来源：评论

学校读者我要写书评

暂无评论

Energy Efficiency Impact of processing in Memory: A Comprehensive Review of Workloads on the UPMEM Architecture

Energy Efficiency Impact of Processing in Memory: A Comprehe...

引用

29th international conference on parallel and distributed Computing (Euro-Par)

作者： Falevoz, Yann Legriel, Julien UPMEM 6 Pl Robert Schuman F-38000 Grenoble France

ISBN: (纸本)9783031488023;9783031488030

processing-in-Memory (PIM) architectures have emerged as a promising solution for data-intensive applications, providing significant speedup by processing data directly within the memory. However, the impact of PIM on energy efficiency is not well characterized. In this paper, we provide a comprehensive review of workloads ported to the first PIM product available on the market, namely the UPMEM architecture, and quantify the impact on each workload in terms of energy efficiency. Less than the half of the reviewed papers provide insights on the impact of PIM on energy efficiency, and the evaluation methods differ from one paper to the other. To provide a comprehensive overview, we propose a methodology for estimating energy consumption and efficiency for both the PIM and baseline systems at data center level, enabling a direct comparison of the two systems. Our results show that PIM can provide significant energy savings for data intensive workloads. We also identify key factors that impact the energy efficiency of UPMEM PIM, including the workload characteristics. Overall, this paper provides valuable insights for researchers and practitioners looking to optimize energy efficiency in data-intensive applications using UPMEM PIM architecture.

关键词： processing in memory (PIM) UPMEM architecture Data-centric architectures Workload optimization High-performance computing Computational efficiency Energy consumption Energy efficiency

来源：评论

学校读者我要写书评

暂无评论

A Case of System-Wide Power Management for Scientific applications

A Case of System-Wide Power Management for Scientific Applic...

引用

15th IEEE international conference on Cluster Computing (CLUSTER)

作者： Liu, Zhuo Lofstead, Jay Wang, Teng Yu, Weikuan Auburn Univ Auburn AL 36849 USA Sandia Natl Labs Livermore CA 94550 USA

ISBN: (纸本)9781479908981

The advance of high-performance computing systems towards exascale will be constrained by the systems' energy consumption levels. Large numbers of processing components, memory, interconnects, and storage components must all be considered to achieve exascale performance within a targeted energy bound. While application-aware power allocation schemes for computing resources are well studied, a portable and scalable budget-constrained power management scheme for scientific applications on exascale systems is still required. Execution activities within scientific applications can be categorized as CPU-bound, I/O-bound and communication-bound. Such activities tend to be clustered into 'phases', offering opportunities to manage their power consumption separately. Our experiments have demonstrated that their performance and energy consumption are affected differently by CPU frequency, an opportunity to fine tune CPU frequency for a minimal impact on the total execution time but significant savings on the energy consumption. By exploiting this opportunity, we present a phase-aware hierarchical power management framework that can opportunistically deliver good tradeoffs between system power consumption and application performance under a power budget. Our hierarchical power management framework consists of two main techniques: Phase-Aware CPU Frequency Scaling (PAFS) and opportunistic provisioning for power-constrained performance optimization. We have performed a systematic evaluation using both simulations and representative scientific applications on real systems. Our results show that our techniques can achieve 4.3%47% better energy efficiency for large-scale scientific applications.

关键词： energy conservation parallel processing power aware computing

来源：评论

学校读者我要写书评

暂无评论

Accelerating The Optimal Trade-Off Circular Harmonic Function Filter Design on Multicore Systems 16

Accelerating The Optimal Trade-Off Circular Harmonic Functio...

引用

7th ACM/SPEC international conference on Performance Engineering (ICPE)

作者： Jain, Anubhav Kalele, Amit Tata Consultancy Serv Ctr Excellence Parallelizat & Optimizat New Delhi India Tata Consultancy Serv Ctr Excellence Parallelizat & Optimizat Pune Maharashtra India

ISBN: (纸本)9781450340809

Optimal correlation filters are widely used in signal processing and pattern recognition applications. Correlation filters are a set of synthesized spatial filters that produce controlled response with sharp peaks. While providing excellent discrimination capabilities correlation filters offer shift, rotation and scale invariance for 2D images. Correlation filters are optimized to enhance the recognition of consistent parts while suppressing the varying patterns. Synthesizing the correlation filters for pattern recognition applications involves several complex mathematical operations and requires high computation resources especially for high resolution images and videos. In this paper, we show that near real time performance can be achieved for the design of the OTCHF filter with help of optimization and parallelization on multicore GPUs and CPUs.

关键词： Correlation filter Circular Harmonic Function Filter Muticore CPU GPU Performance Optimization parallel Computing HPC

来源：评论

学校读者我要写书评

暂无评论

Intelligent data staging with overlapped execution of grid applications

引用

FUTURE GENERATION COMPUTER SYSTEMS-THE international JOURNAL OF ESCIENCE 2008年第5期24卷 425-433页

作者： Machida, Yuya Takizawa, Shin'ichiro Nakada, Hidemoto Matsuoka, Satoshi Tokyo Inst Technol Tokyo 1528550 Japan Natl Inst Adv Ind Sci & Technol Tsukuba Ibaraki 3058568 Japan Res Org Informat & Syst Natl Inst Informat Chiyoda Ku Tokyo 1018430 Japan

Existing data grid scheduling systems handle huge data I/O via replica location services coupled with simple staging and decoupled from scheduling of computing tasks. However, when the application/workflow scales, we observe considerable degradations in performance, compared to processing within a tightly-coupled cluster. For example, when numerous nodes access the same set of files simultaneously, extreme performance degradation occurs even if replicas are used, due to bottlenecks that show in the infrastructure. Instead of resorting to expensive solutions such as parallel file systems, we propose tightly coupling replica and data transfer management with computation scheduling for alleviating such situations. In particular, we propose three techniques: (1) data-staging requests aggregation and O(1) replication across multiple nodes using a multireplication framework, (2) replica-centric scheduling, which reuses previously used data for minimizing staging time and (3) overlapped execution of data staging and compute bound tasks. Early benchmark results implemented in our prototype Condor-like grid scheduling system demonstrate that the techniques are quite effective in eliminating much of the overhead in data transfers and achieving 100% of CPU utilization. (c) 2007 Elsevier B.V. All rights reserved.

关键词： grid computing job scheduling data grid data-intensive application

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：