This paper presents the development of a general-purpose parallel ocean circulation model, for use on a wide range of computer platforms, from traditional scalar machines to workstation clusters and massively parallel processors. Parallelism is provided, as a modular option, via high-level message-passing routines, thus hiding the technical intricacies from the user. An initial implementation h...

The availability of high-speed networks and increasingly powerful commodity microprocessors is making the usage of clusters, or networks, of computers an appealing platform for cost e€ective parallel computing. However, the ease of developing ecient high-performance parallel software to exploit these platforms presents a major challenge. Advances in distributed object software technology have ...

In this paper, the implementation of open-source parallel-version FDTD (Finite-Difference-Time-Domain) software, MEEP, on Texas A&M supercomputers and commercial finite element package, COMSOL, on a single workstation for the design design of nano-optical device is reported. The the computer architecture and performance of both numerical methods on the same design will be briefly described.

In this paper we develop techniques that are suitable for the parallel implementation of Fuzzy ARTMAP networks. Speedup and learning performance results are provided for execution on a DECmpp/Sx-1208 parallel processor consisting of a DEC RISC Workstation Front-End (FE) and MasPar MP-1 Back-End (BE) with 8,192 processors. Experiments of the parallel implementation were conducted on the Letters ...

This paper describes the parallelization of a deterministic iterative method for the placement of standard cell circuits in VLSI design. The programming environment is the implementation of the parallel programming language ParMod-C for workstation clusters. The essential steps for transforming the existing sequential program to a parallel program are shown. An emphasis is laid on the discussio...

The increase of processing speed is an important goal in the development of image recognition systems, especially in case of the recognition of complex objects. The use of parallel computers offers possibilities for the acceleration of the necessary algorithms. However, only few research work has been done in the area of high level image recognition. In our contribution parallel knowledge based...

Performing computations using networks of workstations is increasingly becoming an alternative to using a supercomputer. This approach is motivated by the the vast quantities of unused idle-time available in workstation networks. Unlike computing on a tightly coupled parallel computer, where a xed number of processor nodes is used within a computation, the number of useable nodes in a workstati...

The eecient usage of parallel computers and workstation clusters for biologically motivated simulations depends rst of all on a dynamic redistribution of the workload. For the development of a parallel algorithm for the Penna model of aging we have used a dynamic load balancing library , called PLB. It turns out that PLB manages a nearly balanced load situation during runtime taking only a low ...

This paper describes the current state of Prophet — a system that provides run-time scheduling support for parallel applications in heterogeneous workstation networks. Prior work on Prophet [10][12] demonstrated that scheduling SPMD applications could be effectively automated with excellent performance. Enhancements have been made to Prophet to broaden its use to other application types includi...

We describe a 3D vision system implemented upon a locally developed transputer-based hybrid parallel processing engine named MARVIN (Multiprocessor ARchitecture for VIsioN), hosted by a SUN workstation. In addition to the recovery of scene descriptions from edge based binocular stereo, the system incorporates a model matching algorithm which is able to accurately locate modelled objects within ...