摘要:

We analyse the properties of linear transformations used in the Data Dependence method. Actually the Dependence Graph (DG) is mapped to a Space Time Graph (STG) with special properties. The transformation is valid if 4 conditions are satisfied: the timing condition, the space time condition, the regular data flow condition and the existence condition. The obtained STG has not been analysed in the literature previously.

ISSN:1063-7192    

DOI:10.1080/10637199608915546

出版商:Taylor & Francis Group    

年代:1996

数据来源: Taylor

摘要:

In this paper we propose a new, more realistic queueing modelling of Discrete Event Dynamic Systems. Being based on the analysis of the real-world complex discrete event systems, the model allows hot only the exploitation of their inherent parallelism, but also the implementation and the efficiency analysis of deadlock detection/recovery and shared resource conflict solving algorithms. Unlike in most approaches developed so far, we consider the case in which the completion of a task is achieved by sequentially running its subtasks on a heterogeneous system composed of homogenous clusters. Moreover, our model makes distinction between processing clusters and system resources. Each cluster is composed of an infinite length queue and of many identical processing units, and is individualised by the type of processing units it contains and by its position in the subtask processing sequence. A parallel simulator based on this model was implemented on a Sequent Balance shared-memory parallel computer. As the simulator allows the interfacing of its input with the Petri Net representation of the simulated system, our approach also represents a new step in the direction of a unitary modelling theory of Discrete Event Dynamic Systems.

ISSN:1063-7192    

DOI:10.1080/10637199608915547

出版商:Taylor & Francis Group    

年代:1996

数据来源: Taylor

摘要:

This work deals with the investigation of the performance of parallel iterative algorithms, used for the solution of linear systems obtained from the discretization of Elliptic PDEs using the Finite Element Collocation method. To increase parallelicity the initial matrix is reordered and then the algorithm is mapped on a distributed memory parallel computer. The case of the Star architecture with master-slave communication of the processors is studied. Earlier work of the authors is improved here and at the same time a new solution approach is developed so that the application of these iterative algorithms on more general problems is feasible. Speedup, processor utilization and efficiency measures are presented. The theoretical optimum is almost reached even though fixed number of processors was available.

ISSN:1063-7192    

DOI:10.1080/10637199608915548

出版商:Taylor & Francis Group    

年代:1996

数据来源: Taylor

摘要:

In this paper, we have given a Parallel Alternating Group Explicit (AGE) Algorithm for Parabolic Equations with a small parameter. The Implementation of this algorithm on LINEAR ARRAY architecture is prescribed. The Complexity analysis is also presented. The solutions of this algorithm are compared with the Exact solution.

ISSN:1063-7192    

DOI:10.1080/10637199608915549

出版商:Taylor & Francis Group    

年代:1996

数据来源: Taylor

摘要:

This paper provides new systolic algorithms and a neural network for parallel inference. Systolic algorithms to partition the logic computation for orthogonal arrays and triangular arrays are presented. The neural network for inference is proposed then. For demonstrating the algorithms, a few examples executing on the systolic arrays and the neural networks are given. Finally, we work out the computational complexity of the algorithms and some suggestions.

ISSN:1063-7192    

DOI:10.1080/10637199608915550

出版商:Taylor & Francis Group    

年代:1996

数据来源: Taylor