摘要:

AbstractA parallel preconditioned conjugate gradient (PCG) algorithm is derived by using a two‐level twisted factorization. The algorithm is optimal for a four‐processor parallel computer (quadputer). Numerical results obtained on a quadputer with distributed memory are presen

ISSN:1040-3108    

DOI:10.1002/cpe.4330040202

出版商:John Wiley&Sons, Ltd    

年代:1992

数据来源: WILEY

摘要:

AbstractAmong the many techniques in computer graphics, ray tracing is prized because it can render realistic images, albeit at great computational expense. Ray tracing's large computation requirements, coupled with its inherently parallel nature, make ray tracing algorithms attractive candidates for parallel implementation. In this paper we illustrate the utility and the importance of a suite of performance analysis tools when exploring the performance of several approaches to ray tracing on a distributed memory parallel system. These ray tracing algorithm variations introduce parallelism based on both ray and object partitions.Traditional timing analysis can quantify the performance effects of different algorithm choices (i.e. when an algorithm is best matched to a given problem), but it cannot identify the causes of these performance differences. We argue, by example, that a performance instrumentation system is needed that can trace the execution of distributed memory parallel programs by recording the occurrence of parallel program events. The resulting event traces can be used to compile summary stapistics that provide a global view of program performance. In addition, visualization tools permit the graphic display of event traces. Visual presentation of performance data is particularly useful, indeed, necessary for large‐scale, parallel computations; assimilating the enormous volume of performance data mandates visual displa

ISSN:1040-3108    

DOI:10.1002/cpe.4330040203

出版商:John Wiley&Sons, Ltd    

年代:1992

数据来源: WILEY

摘要:

AbstractThis paper describes the design, the implementation, and the performance results of a routing algorithm which provides deadlock‐free communication in a tightly coupled message‐passing concurrent computer. The algorithm is adaptive, isolated and uses the store‐and‐forward technique. It allows message communication between two processes regardless of where they are physically located on the network. The routing algorithm has many positive characteristics including provable deadlock freedom, guaranteed message arrival, and automatic local congestion reduction. It can be used as a basis for the design of high‐level communication primitives. An Occam implementation on a network of inmos Transputers is discussed. The experimental results show that the routing algorithm is effective to support process to process communication on a concurrent

ISSN:1040-3108    

DOI:10.1002/cpe.4330040204

出版商:John Wiley&Sons, Ltd    

年代:1992

数据来源: WILEY

摘要:

AbstractThis paper discusses efforts to develop parallel algorithms that can be used to solve large‐scale assignment problems typical of military battle management systems. Attempts to parallellze the classical Hungarian method are explored. Drawing on experimental data, a regresslon model is derived that relates the Hungarian method's run time to a cubic function of the number of parallel processors used in the solution. Four parallel heuristics for solving the assignment problem are developed and analysed. Two of them perform well in a parallel environment. The first, based on Vagel's approximation, can be used to identify a feasible, near‐optimal assignment. The second algorithm partitions the assignment problem into independent subproblems across the parallel array. The resulting solution, although infeasible in the strict definition of the assignment problem, is seen to be reasonably good and can be obtained very rapidly. Either of these two heuristics are of potential use in the stringent computing environment of a real‐time resource management s

ISSN:1040-3108    

DOI:10.1002/cpe.4330040205

出版商:John Wiley&Sons, Ltd    

年代:1992

数据来源: WILEY

摘要:

AbstractThe modelling of the highly coupled and non‐linear dynamic models of robot manipulators is computationally expensive in terms of computer hardware. Hence this problem has always presented a major obstacle in on‐line dynamic control applications. However, the recent advent of the VLSI single on‐chip computer (Transputer) makes it possible to implement new algorithms that compute these dynamic models within real‐time constraints. This paper describes a solution of this problem by employing a parallel‐processing approach. The dynamic model of a robot manipulator is divided into different tasks. Further, each task is divided into several subtasks. A heuristic scheduling algorithm is used to produce different near‐optimum task allocations. Real‐time implementations of the proposed tasks allocations are given to demonstrate the efficiency and superiority of the parallel‐proc

ISSN:1040-3108    

DOI:10.1002/cpe.4330040206

出版商:John Wiley&Sons, Ltd    

年代:1992

数据来源: WILEY

ISSN:1040-3108    

DOI:10.1002/cpe.4330040201

出版商:John Wiley&Sons, Ltd    

年代:1992

数据来源: WILEY