Summary form only given. Sparse matrices are matrices having a large number of zero elements. When such matrices are used, both computing time and memory space may be dramatically reduced by taking into account their ...
详细信息
Summary form only given. Sparse matrices are matrices having a large number of zero elements. When such matrices are used, both computing time and memory space may be dramatically reduced by taking into account their sparsity. It is well known that the sparse matrix-vector product (SMVP) where the matrix is sparse and the vector is dense is an important kernel in many scientific applications e.g. iterative methods for linear systems and/or eigen problem. The final aim of this work is to design a kind of user-"expert system" that can be used to improve performances in computing environments, particularly grids involving heterogeneous nodes, on which the SMVP kernel is distributed. In this paper, we study the unrolling as an optimization technique and we apply it to the SMVP when the CRS sparse matrix compression format (CSF) is used. After an analysis of the problem, we detail a series of experiments achieved on three different machines. A set of conclusions could be obtained, particularly, the fact that the compiler optimization does not always lead to the best performances. Indeed, specific manual optimizations through loop unrolling could be better.
Summary form only given. Over the past few years, the development of technology has allowed new advances in sensor networks that monitor the physical world. The data streams produced by sensor networks have different ...
详细信息
Summary form only given. Over the past few years, the development of technology has allowed new advances in sensor networks that monitor the physical world. The data streams produced by sensor networks have different characteristics from the data of traditional data processing, thus requiring new paradigms of data processing systems. Recently, a lot of research has been reported on data stream processing systems. However, there are many underlying assumptions about these systems that have not been explicitly specified. In this paper, our attempt is to provide a general model and architecture for data stream processing in sensor networks. This can serve as a reference architecture to better understand and categorize research in this area.
Summary form only given. As the model-checking (which is based on reachability analysis) becomes increasingly used in the industry, there is a big need for efficient new methods to deal with the large real-size design...
详细信息
Summary form only given. As the model-checking (which is based on reachability analysis) becomes increasingly used in the industry, there is a big need for efficient new methods to deal with the large real-size designs. This paper presents a novel method for improving the performance of reachability analysis using parallelization techniques. We propose a new method for partitioning the large state space modeling industrial designs with hundreds of millions of states and transitions. This method partitions the state space by performing a combination of abstraction-partition-refinement on its structure. The reachability analysis is distributed on processes located on the different network machines. Each one owns a partition and executes the reachability analysis on it. The algorithm for parallel reachability analysis is designed by a way reducing the communication overhead between the processes. The experimental results on large real designs show that this method improves the quality of partitions, the communication overhead and then the overall performance of the reachability analysis.
Tracking position control of a linear magnetic actuator implemented on distributed control network was studied. In the distributed control system involving a discrete-time controller and a linear magnetic actuator, ti...
详细信息
ISBN:
(纸本)0819460745
Tracking position control of a linear magnetic actuator implemented on distributed control network was studied. In the distributed control system involving a discrete-time controller and a linear magnetic actuator, time delay happens as the controller reads the sensor data or sends out the control input to the actuator located at a remote site. In many cases the time delay is not ignorable, and has significant effect on the control system performance. In order to compensate for the adverse effect of the time delay, the modified Smith predictor was introduced to the control system. The results of simulation and experiment show that the performance of linear magnetic actuator can be improved by augmenting the modified Smith predictor onto the existing control system. With the modified Smith predictor, the position error is reduced and the constant disturbance is suppressed.
This paper presents an implementation model of a sensor fusion system using multiple autonomous agents that interact with each other with a collaborative behaviour and implemented in reconfigurable hardware (hardware ...
详细信息
This paper presents an implementation model of a sensor fusion system using multiple autonomous agents that interact with each other with a collaborative behaviour and implemented in reconfigurable hardware (hardware agents). Such hardware implementations has the potential of greatly improving the performance of agent based sensor fusion operations when compared to equivalent software only implementations.
The JAKOBI (Java and component based industrial control) concept is a fully Java based architecture which provides an execution environment for distributed control applications based on real-time Java following the RT...
详细信息
The JAKOBI (Java and component based industrial control) concept is a fully Java based architecture which provides an execution environment for distributed control applications based on real-time Java following the RTSJ-specification. A special focus was on concepts providing the possibility to exchange parts of the control application during runtime. Problematic aspects of a RTSJ based implementation are discussed and a test system proving the applicability of these mechanisms is described.
Summary form only given. In this paper we propose a new peer-to-peer protocol for atomic reliable multicast based on view consistency. It is designed for nearly real-time collaborative applications which require consi...
详细信息
Summary form only given. In this paper we propose a new peer-to-peer protocol for atomic reliable multicast based on view consistency. It is designed for nearly real-time collaborative applications which require consistency of data in a group of participants. A typical example for this is drawing on a shared whiteboard. Collaborations frequently arise in an ad-hoc way in which setting up and operating a dedicated server is not feasible. Thus, our protocol is intended to be used in server-less environments and avoids dependence on central components. We implemented our protocol using the peer-to-peer middleware JXTA for use in TCP/IP-connected networks. This implementation offers its service to other applications via a lean Java API.
This paper defines a new wireless mobility architecture with quality of service (QoS) that uses the fixed network infrastructure for group communication. The mobile core-based tree (M-CBT) architecture uses multicast ...
详细信息
This paper defines a new wireless mobility architecture with quality of service (QoS) that uses the fixed network infrastructure for group communication. The mobile core-based tree (M-CBT) architecture uses multicast to provide fast handoff and reduce network utilization costs for group communication. The M-CBT architecture can be coupled with a multicast QoS routing protocol to provide QoS for mobile communication. The explore best path message (EBPM) protocol is a probe- based dynamic distributed multicast QoS routing protocol that is designed to quickly search the network for an optimal QoS path from a joining node to the multicast tree. This paper illustrates how mobile devices using M-CBT and EBPM can participate in group communications and quickly find a QoS path as they handoff to a new network.
We present SeNeTs and EnviSense. SeNeTs is a powerful software environment for test and validation of sensor network applications. It administrates large-scale sensor networks during execution without affecting commun...
详细信息
We present SeNeTs and EnviSense. SeNeTs is a powerful software environment for test and validation of sensor network applications. It administrates large-scale sensor networks during execution without affecting communication among sensor nodes. Moreover, SeNeTs allows efficient debugging of sensor network applications with sophisticated update mechanisms. EnviSense is a graphical frontend to easily configure and administrate sensor networks using SeNeTs. Both software systems are independent on the hardware platform and also able to run stand-alone.
暂无评论