With the transition to multicore processors almost complete, the parallel processing community is seeking efficient ways to port legacy message passing applications on shared memory and multicore processors. MPJ Expre...
详细信息
ISBN:
(纸本)9781450302692
With the transition to multicore processors almost complete, the parallel processing community is seeking efficient ways to port legacy message passing applications on shared memory and multicore processors. MPJ Express is our reference implementation of Message Passing Interface (MPI)-like bindings for the Java language. Starting with the current release, the MPJ Express software can be configured in two modes: the multicore and the cluster mode. In the multicore mode, parallel Java applications execute on shared memory or multicore processors. In the cluster mode, Java applications parallelized using MPJ Express can be executed on distributed memory platforms like compute clusters and clouds. The multicore device has been implemented using Java threads in order to satisfy two main design goals of portability and performance. We also discuss the challenges of integrating the multicore device in the MPJ Express software. This turned out to be a challenging task because the parallel application executes in a single JVM in the multicore mode. On the contrary in the cluster mode, the parallel user application executes in multiple JVMs. Due to these inherent architectural differences between the two modes, the MPJ Express runtime is modified to ensure correct semantics of the parallel program. Towards the end, we compare performance of MPJ Express (multicore mode) with other C and Java message passing libraries - -including mpiJava, MPJ/Ibis, MPICH2, MPJ Express (cluster mode) - -on shared memory and multicore processors. We found out that MPJ Express performs signicantly better in the multicore mode than in the cluster mode. Not only this but the MPJ Express software also performs better in comparison to other Java messaging libraries including mpiJava and MPJ/Ibis when used in the multicore mode on shared memory or multicore processors. We also demonstrate effectiveness of the MPJ Express multicore device in Gadget-2, which is a massively parallel astrophysics N-body
A new paradigm to design computer systems called evolving reconfigurable architecture (ERA) is proposed. Several holistic principles are introduced and pursued through the ERA design cycle. We show how reconfigurabili...
详细信息
ISBN:
(纸本)9781424474226;9780769540887
A new paradigm to design computer systems called evolving reconfigurable architecture (ERA) is proposed. Several holistic principles are introduced and pursued through the ERA design cycle. We show how reconfigurability is implemented at the level of hardware elements and system software structure. Two new models: (i) graph-logic and (ii) control-data-predicate are introduced and applied in the ERA. We propose a scheme to transform sequential programs for execution on ERA, we also show that ERA has exceptional flexibility and reconfigurability and fits embedded, high performance and safety critical applications. Finally, further progress of ERA is discussed.
Cooperative distributed Problem Solving refers to a loosely coupled network of problem solvers working together to solve problems that may be beyond the capabilities of individual agents. Individual nodes may not have...
详细信息
Cooperative distributed Problem Solving refers to a loosely coupled network of problem solvers working together to solve problems that may be beyond the capabilities of individual agents. Individual nodes may not have sufficient expertise, resources or information to solve the entire problem but every node is capable of solving some part. A group of autonomous agents in a Multi-agent System often work in a cooperative manner and form coalitions to obtain higher payoff than they can get individually. The coalition formation framework of Multi-agent systems can be extended to a distributed Problem Solving formulation where agents strive to achieve higher group benefits. This paper tries to demonstrate how a group of autonomous agents can solve a package delivery domain problem with reasonably good parameters in a purely distributed manner. Unlike traditional distributed problem solving experiments, here the agents are distributed across the problem domain and all control decisions are decentralized with no agent acting as a central dispatcher/ scheduler.
Software architecture design plays a crucial role for both the software and business success. Most of the existing methods conduct software architecture design driven by the key software requirements, facing the chall...
详细信息
ISBN:
(纸本)9781424474226;9780769540887
Software architecture design plays a crucial role for both the software and business success. Most of the existing methods conduct software architecture design driven by the key software requirements, facing the challenges of creating architectures from scratch and aligning with the high-level business goals. In this paper we propose the BASAD (Business And Software Architecture co-Design) method to facilitate alleviating the difficulty of software architecture design and achieving the high-level business goals. The method addresses the business and software architecture design in an integrated process from a systems engineering perspective, leveraging the SysML and UML modeling languages and the automated model transformation from the business architectures to the software architectures, so that provide pragmatic support for the architecture co-design.
Many large-scale scientific applications feature distributedcomputing workflows of complex structures that must be executed and transferred in shared wide-area networks consisting of unreliable nodes and links. Mappi...
详细信息
Many large-scale scientific applications feature distributedcomputing workflows of complex structures that must be executed and transferred in shared wide-area networks consisting of unreliable nodes and links. Mapping these computing workflows in such faulty network environments for optimal latency while ensuring certain fault tolerance is crucial to the success of eScience that requires both performance and reliability. We construct analytical cost models and formulate workflow mapping as an optimization problem under failure rate constraint. We propose a distributed heuristic mapping solution based on recursive critical path to achieve minimum end-to-end delay and satisfy a pre-specified overall failure rate for a guaranteed level of fault tolerance. The performance superiority of the proposed mapping solution is illustrated by extensive simulation-based comparisons with existing mapping algorithms.
The topologies of Internet and Internet-based information systems have complex network properties. Designing Internet-based virtual computing environment topology with appropriate properties is significant for both th...
详细信息
The topologies of Internet and Internet-based information systems have complex network properties. Designing Internet-based virtual computing environment topology with appropriate properties is significant for both the resource sharing and system performance. We analyses the topology properties of the typical P2P systems, and proposes a new topology generating method, which includes three phases, birth, growth and maturity, and supports multi-node concurrent joining in. The iVCE topology generation method can produce stable structure, with load balancing capability. Analysis of the generated topologies shows that the degree of their super-node obeys normal distribution law, the average path length between nodes shows small-world properties.
Scheduling of tasks in a heterogeneous computing (HC) environment is a critical task. It is also a well-known NP-complete problem, and hence several researchers have presented a number of heuristics for the same. The ...
详细信息
Scheduling of tasks in a heterogeneous computing (HC) environment is a critical task. It is also a well-known NP-complete problem, and hence several researchers have presented a number of heuristics for the same. The paper begins with introducing a new heuristic called Sympathy, and later a variant called Segmented Sympathy. A new Genetic Algorithm based heuristic using the Segmented Sympathy heuristic is proposed, which is aimed at improving over the speed and makespan of the implementation by Braun et al. Finally, the results of Simulation reveal that the proposed Genetic Algorithm gave up to 8.34% and on an average 3.42% better makespans. The new heuristic is also about 160% faster with respect to the execution time.
This paper proposes a novel platform for object locating application in the Internet of Things environment. In this platform, objects and inquirers access and query locations using uniform service entry interfaces in ...
详细信息
This paper proposes a novel platform for object locating application in the Internet of Things environment. In this platform, objects and inquirers access and query locations using uniform service entry interfaces in heterogeneous services. To build a virtual storage system, services entries integrate enterprise database clusters and a DHT peer-to-peer network built with inquirers' devices. The DHT network is originally designed for accurate object locating, to enable fuzzy object locating we construct a hierarchical storage overlay network based on the DHT network. This LBS platform simplifies the object locating operation for ordinary inquirers greatly, moreover it provides huge virtual computing and storage resources for small companies and individual developers.
Fault tolerance in distributedsystems relies heavily on some form of replication. Replication can also be used to reduce the access latency and the bandwidth consumption in large scale distributedsystems. However, i...
详细信息
Fault tolerance in distributedsystems relies heavily on some form of replication. Replication can also be used to reduce the access latency and the bandwidth consumption in large scale distributedsystems. However, in case of large volumes of data, the replica placing strategy and the consistency algorithms become key factors for the performance of the data replication strategy. We present a simulation model designed to realistically evaluate replication solutions for large scale distributedsystems. The model was implemented as an extension of the MONARC simulator. In this context, we present a scalable architecture designed to facilitate the adoption of data replication strategies in large scale distributedsystems. The solution combines a hybrid replication model with a proposed fault tolerant strategy for data consistency. We present evaluation results of this strategy using the MONARC simulator.
暂无评论