Single Parallel Computer Research Articles

Current problems in parallelization of the computation algorithms and organization of parallel computations

The last decade witnessed an intensive development of the network technologies which enabled a multifold increase in the computer throughput. The matter primarily concerns design of the supercomputers, the most powerful modern computer systems based on the concept of parallel processing of massive data. For example, the supercomputers combining the commercial processors and the facilities of the high-speed LAN’s have become widespread. Another, cheaper path to using the network technologies lies in organizing the so-called distributed computational environments. The gist of this approach lies in uniting the already existing LAN computers, this union being regarded as a single parallel computer. Yet, as compared with the supercomputer based on the commercial processors, this computer system has a lower reliability on account of different membership of the network computer nodes, higher structural nonuniformity on account of different architectures of the nodes, and other distinctions. At that, the problems of establishing efficient means of data exchange between the individual nodes and optimizing the topological structure of the network come to the forefront just as it is the case with the supercomputers. As was already mentioned, the parallel data processing is the characteristic feature of the modern high-speed computer systems. That is why the dedicated software and its proper databases for parallel operation become topical. The supercomputers and distributed computer systems enable one to settle many complicated— and previously intractable or unsolvable—applied problems such as the control of the dynamic systems which in practice have huge dimensions and are labor-intensive. Solution of such problems by the existing computational facilities requires special algorithms taking into account the characteristics of this hardware and enabling parallelization of computations. All of the aforementioned problems such as the outlooks for design of the supercomputers, features of the distributed computer environments, methods of organization of the parallel databases, and optimization of the local-network architecture are considered to a certain extent in the papers published in this issue of Automation and Remote Control . Diverse parallel computation-based algorithms for different fields of mathematics such as discrete and stochastic optimization, mathematical programming with search of the global extremum, statistical modeling of the gas dynamics, direct and inverse problems of the mathematical physics, and so on were proposed. Consideration was also given to numerous applied problems whose solution makes use of the parallel computation technology. We expect that publication under one cover of the papers of renowned researchers that are devoted to a single subject will attract interest of the experts in parallel computations. Although some papers by themselves are beyond the scope of the subject-matter of this issue, together they assume a new sense and, in our opinion, will be of use to the readership of Avtomatika i telemekhanika. I.I. Eremin, A.I. Kibzun

Distributed Parallel Computing for Fluid-Structure Coupled Simulations on a Heterogeneous Parallel Computer Cluster

Distributed parallel computing for a fluid-structure coupled simulation has been performed on a heterogeneous parallel computer cluster. The fluid and the structure dynamics are simulated on different parallel computers connected by a high-speed local network. These dynamics are coupled by a loose coupling method exchanging the boundary data between the fluid and the structure domains through the network. The data communication among parallel computers is realized by using the new communication library, Stampi, which has been developed to enable communication in a heterogeneous environment. The performance evaluation on a heterogeneous parallel computer cluster has shown that the distributed parallel computing for fluid-structure coupled simulations has the advantage of increasing the performance compared with the parallel computing on a single parallel computer.

Fifth generation and VLSI architectures

Most Western Governments (USA, Japan, EEC, etc.) have now launched national programmes to develop computer systems for use in the 1990s. These so-called Fifth Generation computers are viewed as “knowledge” processing systems which support the symbolic computation underlying Artificial Intelligence applications. The major driving force in Fifth Generation computer design is to efficiently support very high level programming languages (i.e. VHLL architecture). Historycally, however, commercial VHLL architectures have been largely unsuccesful. The driving force in computer designs has principally been advances in hardware which at the present time means architectures to exploit very large scale integration (i.e. VLSI architecture). This paper examines VHLL architectures and VLSI architectures and their probable influences on Fifth Generation computers. Interestingly the major problem for both architecture classes is parallelism; how to orchestrate a single parallel computation so that it can be distributed across an ensemble of processors.

Utilising Workstation Clusters With PVM For TheSolution Of Large TLM Problems

The Transmission Line Modelling (TLM) method has been used to model a wide variety of electromagnetic and thermal diffusion problems in the timedomain. While it compares favourably to other methods in terms of its computational requirements, there is continual motivation to simultaneously reduce simulation times, increase feasible problem size and spatial resolution. There has been growing interest recently in the use of parallel computing architectures in a wide range of numerical simulation situations; the TLM method lends itself well to these approaches. However, traditional massively parallel computing systems are generally difficult to use and very expensive. This has limited their availability for most researchers, who mainly rely on essentially stand-alone workstations. The feasibility of using a cluster of workstations as a single parallel computation resource, based on the Parallel Virtual Machine (PVM) message passing library, and with particular reference to TLM software, is examined. Some of the issues involved in the optimal use of such a system, including data partitioning, load balancing, and programming methodologies are outlined. It is demonstrated that for the solution of large problems PVM permits a significant increase in computational throughput with minimal additional financial and programming costs.