A task level intelligent computing architecture, called as smart workstation cluster distributed parallel computing model, is presented for transient stability constrained total transfer capability (TTC) evaluation of...
详细信息
A task level intelligent computing architecture, called as smart workstation cluster distributed parallel computing model, is presented for transient stability constrained total transfer capability (TTC) evaluation of large scale interconnected power system with respect to a specified contingency set. The proposed intelligent distributed parallel computing mode can effectively integrate the heterogeneous distributed computing resources around Internet and implement the dynamic load balancing so as to make full use of the whole system computing performance. Furthermore, the dynamic scalability and fault-tolerance of the proposed computing architecture are analyzed and developed as well. The case studies have been carried out on a real-sized Chinese power system, and results demonstrate the practicability and effectiveness of the proposed model.
Making the best use of modern computational resources for distributed applications requires expert knowledge of low-level programming tools, or a productive high-level and high-performance programming framework. Unfor...
详细信息
Making the best use of modern computational resources for distributed applications requires expert knowledge of low-level programming tools, or a productive high-level and high-performance programming framework. Unfortunately, even state-of-the-art high-level frameworks still require the developer to conduct a tedious manual tuning step to find the work partitioning which gives the best application execution performance. Here, we present a novel framework, with which developers can easily create high-performance dataflow applications, without the tedious tuning process. We compare the performance of our approach to that of three distributed programming frameworks which differ significantly in their programming paradigm, their support for multi-core CPUs and accelerators, and their load-balancing approach. These three frameworks are DataCutter, a component-based dataflow framework, KAAPI, a framework using asynchronous function calls, and MR-MPI, a MapReduce implementation. By highly optimizing the implementations of three applications on the four frameworks and comparing the execution time performance of the runtime engines, we show their strengths and weaknesses. We show that our approach achieves good performance for a wide range of applications, with a much-reduced development cost. (c) 2012 Elsevier B.V. All rights reserved.
We show that stopwatch automata are equivalent with timed shuffle expressions, an extension of timed regular expressions with the shuffle operation. Since the emptiness problem is undecidable for stopwatch automata, a...
详细信息
We show that stopwatch automata are equivalent with timed shuffle expressions, an extension of timed regular expressions with the shuffle operation. Since the emptiness problem is undecidable for stopwatch automata, and hence also for timed shuffle expressions, we introduce a decidable subclass of stopwatch automata called partitioned stopwatch automata. We give for this class an equivalent subclass of timed shuffle expressions and investigate closure properties by showing that partitioned stopwatch automata are closed under union, concatenation, star, shuffle and renaming, but not under intersection. We also show that partitioned stopwatch automata are equivalent with distributed time-asynchronous automata, which are asynchronous compositions of timed automata in which time may evolve independently.
This paper examines the application of service oriented architecture (SOA) in finite element analysis. SOA is a technology for designing and developing interoperable services. These services can reside on the same com...
详细信息
This paper examines the application of service oriented architecture (SOA) in finite element analysis. SOA is a technology for designing and developing interoperable services. These services can reside on the same computer or, more commonly, on distributed computers. The paper demonstrates how SOA can be used within the context of scientific computing. The implementation and application of SOA to equation solvers and finite element analysis is described. There are advantages in terms of software engineering, as it facilitates the separation of areas of complexity. SOA can be used on standalone computers, intranets and on the internet. The data transfer costs are examined. It is shown that SOA principles can be used to design applets that make use of finite element analysis and a simple example of this is described. (c) 2012 Elsevier Ltd. All rights reserved.
We prove that, for the black hole search problem in networks of arbitrary but known topology, the pebble model of agent interaction is computationally as powerful as the whiteboard model;furthermore the complexity is ...
详细信息
We prove that, for the black hole search problem in networks of arbitrary but known topology, the pebble model of agent interaction is computationally as powerful as the whiteboard model;furthermore the complexity is exactly the same. More precisely, we prove that a team of two asynchronous agents, each endowed with a single identical pebble (that can be placed only on nodes, and with no more than one pebble per node), can locate the black hole in an arbitrary network of known topology;this can be done with I similar to(nlog n) moves, where n is the number of nodes, even when the links are not FIFO. These results are obtained with a novel algorithmic technique, ping-pong, for agents using pebbles.
The large amount of Semantic Web data and its fast growth pose a significant computational challenge in performing efficient and scalable reasoning. On a large scale, the resources of single machines are no longer suf...
详细信息
The large amount of Semantic Web data and its fast growth pose a significant computational challenge in performing efficient and scalable reasoning. On a large scale, the resources of single machines are no longer sufficient and we are required to distribute the process to improve performance. In this article, we propose a distributed technique to perform materialization under the RDFS and OWL ter Horst semantics using the MapReduce programming model. We will show that a straightforward implementation is not efficient and does not scale. Our technique addresses the challenge of distributed reasoning through a set of algorithms which, combined, significantly increase performance. We have implemented WebPIE (Web-scale Inference Engine) and we demonstrate its performance on a cluster of up to 64 nodes. We have evaluated our system using very large real-world datasets (Bio2RDF, LLD, LDSR) and the LUBM synthetic benchmark, scaling up to 100 billion triples. Results show that our implementation scales linearly and vastly outperforms current systems in terms of maximum data size and inference speed. (C) 2011 Elsevier B.V. All rights reserved.
Interactive visualization and simulation of astrophysical phenomena help astronomers and enable digital planetariums and television documentaries to take their spectators on a journey into deep space to explore the as...
详细信息
Interactive visualization and simulation of astrophysical phenomena help astronomers and enable digital planetariums and television documentaries to take their spectators on a journey into deep space to explore the astronomical wonders of our universe in 3D.
The increasing use of parallel/distributed applications demands a continuous support to take significant advantages from parallel power. This includes the evolution of performance analysis and tuning tools which autom...
详细信息
The increasing use of parallel/distributed applications demands a continuous support to take significant advantages from parallel power. This includes the evolution of performance analysis and tuning tools which automatically allows for obtaining a better behavior of the applications. Different approaches and tools have been proposed and they are continuously evolving to cover the requirements and expectations of users. One such tool is MATE (Monitoring Analysis and Tuning Environment), which provides automatic and dynamic tuning for parallel/distributed applications. The knowledge used by MATE to analyze and take decisions is based on performance models which include a set of performance parameters and a set of mathematical expressions modeling the solution of the performance problem. These elements are used by the tuning environment to conduct the monitoring and analysis steps, respectively. The tuning phase depends on the results of the performance analysis. This paper presents a methodology to specify performance models. Each performance model specification can be automatically and transparently translated into a piece of software code encapsulating the knowledge to be straightforwardly included in MATE. Applying this methodology, the user does not have to be involved in the implementation details of MATE, which makes the usage of the tool more transparent. Copyright (c) 2011 John Wiley & Sons, Ltd.
To solve the problem that the common podcast system based on traditional transcoding system costs much and is very difficult to support concurrent users, in the paper, a novel transcoding system based on distributed f...
详细信息
To solve the problem that the common podcast system based on traditional transcoding system costs much and is very difficult to support concurrent users, in the paper, a novel transcoding system based on distributed farming computing architecture was designed. It can save a great deal cost of hardware resource and improve its transcoding efficiency at the statement of large concurrent accesses. By the comparison of performance between the traditional transcoding system and the proposed in the paper with the simulation, it has been proved that the podcast system adopting the new transcoding method has so much good feature as higher ratio of cost and performance, shorter user waiting time and faster video accumulation speed.
We address the enumeration and the leader election problems over partially anonymous and multi-hop broadcast networks. We consider an asynchronous communication model where each process broadcasts a message and all it...
详细信息
We address the enumeration and the leader election problems over partially anonymous and multi-hop broadcast networks. We consider an asynchronous communication model where each process broadcasts a message and all its neighbours receive this message after arbitrary and unpredictable time. In this paper, we present necessary conditions that must be satisfied by any graph to solve these problems and we show that these conditions are sufficient by providing an enumeration algorithm on the one hand and a leader election algorithm on the other hand. For both problems, we highlight the importance of the initial knowledge. Considering the enumeration problem, each process only knows the size of the graph and, contrary to related works, the number of its neighbouring processes is unknown. Whereas for the election problem, we show that this combination of knowledge is not sufficient. Our algorithm assumes that each process initially knows a map of the network (without knowing its position in this map). From the complexity viewpoint, our algorithms offer polynomial complexities (memory at each process, number and size of exchanged messages).
暂无评论