Newton-Krylov methods are potentially well suited for the implicit solution of nonlinear problems whenever it is unreasonable to compute or store a true Jacobian. Krylov-Schwarz iterative methods are well suited for the parallel implicit solution of multidimensional systems of boundary value problems that arise in CFD. They provide good data locality so that even a high-latency workstation netw...

In this paper, we argue that because of recent technology advances, networks of workstations (NOWs) are poised to become the primary computing infrastructure for science and engineering, from low end interactive computing to demanding sequential and parallel applications. We identify three opportunities for NOWs that will benefit endusers: dramatically improving virtual memory and file system p...

A parallel programming system is presented in this paper. The system supports a speciication language, parScript, which can conveniently encode a fork-joint graph by which many parallel applications are repre-sentable. The system consists of a parScript-to-C compiler, and a runtime kernel. The compiler translates parScript speciication directly into C code executable either on a sequential work...

The increasing complexity of network models poses a growing computational burden. At the same time, computational neuroscientists are finding it easier to access parallel hardware, such as multiprocessor personal computers, workstation clusters, and massively parallel supercomputers. The practical question is how to move a working network model from a single processor to parallel hardware. Here...

Workstation and PC clusters interconnected by SCI (Scalable Coherent Interface) are very promising technologies for high performance cluster computing. Using commercial SBus to SCI interface cards and early system software and drivers, a two-workstation cluster has been constructed for initial testing and evaluation. The PVM system has been adapted to operate on this cluster using raw device ac...

Parallel computing on clusters of workstations is attractive because of the low costs in comparison to MPPs, but the speed of the local area network limits the class of applications that can be run efficiently. Fortunately, faster network technology is becoming available for the next generation of workstation clusters. This paper studies the effect of running challenging applications that commu...

The distributed computer system consisting of several workstations connected by local area network o ers a large amount of computational power inherently. In order to control such a system e ciently, the adequate load distribution method has to be applied, because it is often observed that some workstations have too many works although the others are idle relatively. Namely, the load imbalance ...

The realization of support for workstation clusters with the integrated tool environment Annai is discussed in this paper. Annai prototypes developed at CSCS supportparallel program developmentwith standard paradigms: implicitlyparallel High Performance Fortran (HPF) and explicit communication with the Message Passing Interface (MPI). The port to this new platform required only minor changes to...

The limits of sequential processing continue to be overcome with parallel and distributed architectures and algorithms. In particular, the use of parallel processing for high-performance signal processing and other grand-challenge applications is becoming the norm. However, the costs associated with such systems make it critical to be able to forecast system behavior before a hardware prototype...

Asynchronous parallel computing can result in high message generation rates, thus triggering network congestion. This paper investigates the network congestion problem that can result from asynchronous parallel programs' high message generation rates. First, we characterize the communication requirements of a large class of supercomputing applications falling under the category of xed-point pro...