检索结果-内蒙古大学图书馆

Resource optimised workflow scheduling in Hadoop using stochastic hill climbing technique

IET SOFTWARE 2017年第5期11卷 239-244页

作者： Rashmi, Shivaswamy Basu, Anirban East Point Coll Engn & Technol Dept Comp Sci & Engn Bangalore Karnataka India APS Coll Engn Dept Comp Sci & Engn Bangalore Karnataka India

Hadoop on datacentre is a popular analytical platform for enterprises. Cloud vendors host Hadoop clusters on the datacentre to provide high performance analytical computing facilities to its customers, who demand a parallel programming model to deal with huge data. Effective cost/time management and ingenious resource consumption among the concurrent users, must be the primary concern without which the key aspiration behind high performance cloud computing would suffer. Workflows portray such high performance applications in terms of individual jobs and dependencies between them. Workflows can be scheduled on virtual machines (VMs) in datacentre to make best possible use of resources. In the authors' earlier work, a mechanism to pack and execute the customer jobs as workflows on Hadoop platform was proposed which minimises the VM cost and also executes the workflow jobs within deadline. In this work, the authors try to optimise certain other parameters such as load on cloud, response time for workflows, resource usage effectiveness by applying soft computing methods. Stochastic hill climbing (SCH) is a soft computing approach used to solve many optimisation problems. In this study, they have employed the SHC approach to schedule workflow jobs to VMs and thereby optimise the above mentioned multiple parameters in cloud datacentre.

关键词： workflow management software stochastic processes data handling scheduling cloud computing parallel programming virtual machines operating systems (computers) resource optimised workflow scheduling stochastic hill climbing technique datacentre cloud vendors Hadoop clusters computing facilities parallel programming model resource consumption concurrent users cloud computing workflows portray virtual machines VM Hadoop platform SCH

来源：评论

学校读者我要写书评

暂无评论

Picos. A hardware runtime architecture support for OmpSs

引用

FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE 2015年 53卷 130-139页

作者： Yazdanpanah, Fahimeh Alvarez, Carlos Jimenez-Gonzalez, Daniel Badia, Rosa M. Valero, Mateo UPC Barcelona 08034 Spain BSC Barcelona 08034 Spain CSIC Artificial Intelligence Res Inst IIIA Barcelona Spain

OmpSs is a programming model that provides a simple and powerful way of annotating sequential programs to exploit heterogeneity and task parallelism based on runtime data dependency analysis, dataflow scheduling and out-of-order task execution;it has greatly influenced Version 4.0 of the OpenMP standard. The current implementation of OmpSs achieves those capabilities with a pure-software runtime library: Nanos++. Therefore, although powerful and easy to use, the performance benefits of exploiting fine-grained (pico) task parallelism are limited by the software runtime overheads. To overcome this handicap we propose Picos, an implementation of the Task Superscalar (TSS) architecture that provides hardware support to the OmpSs programming model. Picas is a novel hardware dataflow-based task scheduler that dynamically analyzes inter-task dependencies and identifies task-level parallelism at run-time. In this paper, we describe the Picos Hardware Design and the latencies of the main functionality of its components, based on the synthesis of their VHDL design. We have implemented a full cycle-accurate simulator based on those latencies to perform a design exploration of the characteristics and number of its components in a reasonable amount of time. Finally, we present a comparison of the Picas and Nanos++ runtime performance scalability with a set of real benchmarks. With Picos, a programmer can achieve ideal scalability using aggressive parallel strategies with a large number of fine granularity tasks. (C) 2015 Elsevier B.V. All rights reserved.

关键词： Hardware implementation Task scheduling Dataflow execution parallel programming model OmpSs OpenMP

来源：评论

学校读者我要写书评

暂无评论

Chunks and Tasks: A programming model for parallelization of dynamic algorithms

引用

parallel COMPUTING 2014年第7期40卷 328-343页

作者： Rubensson, Emanuel H. Rudberg, Elias Uppsala Univ Dept Informat Technol Div Comp Sci SE-75105 Uppsala Sweden

We propose Chunks and Tasks, a parallel programming model built on abstractions for both data and work. The application programmer specifies how data and work can be split into smaller pieces, chunks and tasks, respectively. The Chunks and Tasks library maps the chunks and tasks to physical resources. In this way we seek to combine user friendliness with high performance. An application programmer can express a parallel algorithm using a few simple building blocks, defining data and work objects and their relationships. No explicit communication calls are needed;the distribution of both work and data is handled by the Chunks and Tasks library. This makes efficient implementation of complex applications that require dynamic distribution of work and data easier. At the same time, Chunks and Tasks imposes restrictions on data access and task dependencies that facilitate the development of high performance parallel back ends. We discuss the fundamental abstractions underlying the programming model, as well as performance, determinism, and fault resilience considerations. We also present a pilot C++ library implementation for clusters of multicore machines and demonstrate its performance for irregular block-sparse matrix-matrix multiplication. (C) 2013 Elsevier B.V. All rights reserved.

关键词： Distributed memory parallelization Dynamic data distribution Dynamic load balancing Fault tolerance parallel programming model Determinism

来源：评论

学校读者我要写书评

暂无评论

Supporting asynchronization in OpenMP for event-driven programming

引用

parallel COMPUTING 2019年 82卷 57-74页

作者： Fan, Xing Sinnen, Oliver Giacaman, Nasser Univ Auckland Dept Elect & Comp Engn Auckland New Zealand

The event-driven programming pattern is pervasive in a wide range of modern software applications. Unfortunately, it is not easy to achieve good performance and responsiveness when developing event-driven applications. Traditional approaches require a great amount of programmer effort to restructure and refactor code, to achieve the performance speedup from parallelism and asynchronization. Not only does this restructuring require a lot of development time, it also makes the code harder to debug and understand. We propose an asynchronous programming model based on the philosophy of OpenMP, which does not require code restructuring of the original sequential code. This asynchronous programming model is complementary to the existing OpenMP fork-join model. The coexistence of the two models has potential to decrease developing time for parallel event-driven programs, since it avoids major code refactoring. In addition to its programming simplicity, evaluations show that this approach achieves good performance improvements consistent with more traditional event-driven parallelization. (C) 2018 Elsevier B.V. All rights reserved.

关键词： OpenMP parallel programming model Event-driven programming Asynchronous programming

来源：评论

学校读者我要写书评

暂无评论

Active Data: A programming model to manage data life cycle across heterogeneous systems and infrastructures

引用

FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE 2015年 53卷 25-42页

作者： Simonet, Anthony Fedak, Gilles Ripeanu, Matei Univ Lyon LIP ENS Lyon Inria F-69007 Lyon France Univ British Columbia Vancouver BC V6T 1Z4 Canada

The Big Data challenge consists in managing, storing, analyzing and visualizing these huge and ever growing data sets to extract sense and knowledge. As the volume of data grows exponentially, the management of these data becomes more complex in proportion. A key point is to handle the complexity of the data life cycle, i.e. the various operations performed on data: transfer, archiving, replication, deletion, etc. Indeed, data-intensive applications span over a large variety of devices and e-infrastructures which implies that many systems are involved in data management and processing. We propose Active Data, a programming model to automate and improve the expressiveness of data management applications. We first define the concept of data life cycle and introduce a formal model that allows to expose data life cycle across heterogeneous systems and infrastructures. The Active Data programming model allows code execution at each stage of the data life cycle: routines provided by programmers are executed when a set of events (creation, replication, transfer, deletion) happen to any data. We implement and evaluate the model with four use cases: a storage cache to Amazon-S3, a cooperative sensor network, an incremental implementation of the MapReduce programming model and automated data provenance tracking across heterogeneous systems. Altogether, these scenarios illustrate the adequateness of the model to program applications that manage distributed and dynamic data sets. We also show that applications that do not leverage on data life cycle can still benefit from Active Data to improve their performances. (C) 2015 Elsevier B.V. All rights reserved.

关键词： parallel programming model Distributed and heterogeneous systems Data life cycle

来源：评论

学校读者我要写书评

暂无评论

Mitigating the NUMA effect on task-based runtime systems

引用

JOURNAL OF SUPERCOMPUTING 2023年第13期79卷 14287-14312页

作者： Maronas, Marcos Navarro, Antoni Ayguade, Eduard Beltran, Vicenc Barcelona Supercomp Ctr Barcelona Spain Univ Politecn Cataluna Barcelona Spain

Processors with multiple sockets or chiplets are becoming more conventional. These kinds of processors usually expose a single shared address space. However, due to hardware restrictions, they adopt a NUMA approach, where each processor accesses local memory faster than remote memories. Reducing data motion is crucial to improve the overall performance. Thus, computations must run as close as possible to where the data resides. We propose a new approach that mitigates the NUMA effect on NUMA systems. Our solution is based on the OmpSs-2 programming model, a task-based parallel programming model, similar to OpenMP. We first provide a simple API to allocate memory in NUMA systems using different policies. Then, combining user-given information that specifies dependences between tasks, and information collected in a global directory when allocating data, we extend our runtime library to perform NUMA-aware work scheduling. Our heuristic considers data location, distance between NUMA nodes, and the load of each NUMA node to seamlessly minimize data motion costs and load imbalance. Our evaluation shows that our NUMA support can significantly mitigate the NUMA effect by reducing the amount of remote accesses, and so improving performance on most benchmarks, reaching up to 2x speedup in a 2-NUMA machine, and up to 7.1x in a 8-NUMA machine.

关键词： NUMA-awareness OmpSs-2 parallel programming model Scheduling Task-aware

来源：评论

学校读者我要写书评

暂无评论

Distributed shared arrays: A distributed virtual machine with mobility support for reconfiguration

引用

CLUSTER COMPUTING-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS 2006年第3期9卷 237-255页

作者： Fu, Song Xu, Cheng-Zhong Wims, Brian Basharahil, Ramzi Wayne State Univ Dept Elect & Comp Engn Detroit MI 48202 USA

Distributed Shared Arrays (DSA) is a distributed virtual machine that supports Java-compliant multithreaded programming with mobility support for system reconfiguration in distributed environments. The DSA programming model allows programmers to explicitly control data distribution so as to take advantage of the deep memory hierarchy, while relieving them from error-prone orchestration of communication and synchronization at run-time. The DSA system is developed as an integral component of mobility support middleware for Grid computing so that DSA-based virtual machines can be reconfigured to adapt to the varying resource supplies or demand over the course of a computation. The DSA runtime system also features a directorybased cache coherence protocol in support of replication of user-defined sharing granularity and a communication proxy mechanism for reducing network contention. System reconfiguration is achieved by a DSA service migration mechanism, which moves the DSA service and residing computational agents between physical servers for load balancing and fault resilience. We demonstrate the programmability of the model in a number of parallel applications and evaluate its performance by application benchmark programs, in particular, the impact of the coherence granularity and service migration overhead.

关键词： distributed shared arrays (DSA) distributed virtual machine parallel programming model DSA service migration

来源：评论

学校读者我要写书评

暂无评论

programming Highly parallel Reconfigurable Architectures for Symmetric and Asymmetric Cryptographic Applications

引用

JOURNAL OF COMPUTERS 2007年第9期2卷 50-59页

作者： Agosta, Giovanni Breveglieri, Luca Pelosi, Gerardo Sykora, Martino Politecn Milan DEI Via Ponzio 34-5 I-20133 Milan Italy

Tiled architectures are emerging as an architectural platform that allows high levels of instruction level parallelism. Traditional compiler parallelization techniques are usually employed to generate programs for these architectures. However, for specific application domains, the compiler is not able to effectively exploit the domain knowledge. In this paper, we propose a new programming model that, by means of the definition of software function units, allows domain-specific features to be explicitly modeled, achieving good performances while reducing development times with respect to low-level programming. Identity-based cryptographic algorithms are known to be computationally intensive and difficult to parallelize automatically. Recent advances have led to the adoption of embedded cryptographic coprocessors to speed up both traditional and identity-based public key algorithms. We show the effectiveness of the proposed programming model by applying it to the case of computationally intensive cryptographic algorithms in both identity-based and traditional algorithms. Custom-designed coprocessors have high development costs and times with respect to general purpose or DSP coprocessors. Therefore, the proposed methodology can be effectively employed to reduce time to market while preserving performances. It also represents a starting point for the definition of cryptography-oriented programming languages. We prove that tiled architecture well compare w.r.t. competitors implementations such as StrongARM and FPGAs.

关键词： identity-based cryptography tiled architectures parallel programming model reconfigurable architectures multiobjective exploration

来源：评论

学校读者我要写书评

暂无评论

CAOPLE: A programming Language for Microservices SaaS

CAOPLE: A Programming Language for Microservices SaaS

引用

10th IEEE International Symposium on Service-Oriented System Engineering (IEEE SOSE)

作者： Xu, Chengzhi Zhu, Hong Bayley, Ian Lightfoot, David Green, Mark Marshall, Peter Hubei Univ Technol Sch Comp Sci Wuhan Peoples R China Oxford Brookes Univ Dept Comp & Comm Tech Oxford England

ISBN: (纸本)9781509022533

The microservices architecture is widely regarded as a promising approach to service-oriented systems. However, developing applications in the microservices architecture presents three main challenges: (a) how to program systems that consists of a large number of services running in parallel and distributed over a cluster of computers;(b) how to reduce the communication overhead caused by executing a large number of small services;(c) how to support the flexible deployment of services to a network to achieve system load balance. This paper presents a programming language called CAOPLE and reports the implementation of the language on a virtual machine called CAVM-2. The paper demonstrates how this approach meets these challenges.

关键词： Service-oriented software Microservices architecture Virtual machine programming languages Cloud computing parallel programming model Agent orientation

来源：评论

学校读者我要写书评

暂无评论

A multicast inter-task communication protocol for embedded multiprocessor systems 05

A multicast inter-task communication protocol for embedded m...

引用

International Conference on Hardware/Software Codesign and System Synthesis

作者： Reyes, V Bautista, T Marrero, G Núñez, A Kruijtzer, W Univ Las Palmas GC Inst Appl Microelect IUMA Las Palmas Gran Canaria 35017 Spain

ISBN: (纸本)1595931619

Recently, a new programming model and platform interface for MPSoC design and integration called TTL (Task Transaction Level) has been developed and advocated as a standard. In this paper, a specific implementation of the TTL interface named ITCP (Inter-Task Communication Protocol) is presented. ITCP is well suited for both hardware and software implementations and supports features such as multitasking and multicast communication. A configurable SystemC model of the ITCP protocol and its integration in a system-level design methodology is disclosed in this work. Moreover, details of a multi-task ITCP software shell implementation for an ARM9 with eCos RTOS are also given in the paper.

关键词： task transaction level multiprocessor design platform interface parallel programming model

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

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

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：