Workstation Cluster Environment Research Articles

Strength Check of Aircraft Parts Based on Multi-GPU Clusters for Fast Calculation of Sparse Linear Equations

In order to improve the cost-effectiveness ratio, the next-generation vehicle needs to meet the requirements of reuse, while adopting a lighter structural weight, so it is necessary to realize the strength calculation and condition monitoring of key components in the digital twin. Most of the current monitoring methods are based on the characteristics of various data acquisition systems, but they need the support of a large number of flight data. The disadvantages of the above strategy can be avoided by reducing the structure of aircraft components to a finite element model and quickly checking the key components in the health management system. In order to solve the problem of fast calculation of the finite element model of the key components of the aircraft, a parallel algorithm and framework of large-scale sparse matrix preprocessing conjugate gradient method based on CUDA(Compute Unified Device Architecture) technology is proposed in the multi GPU(Graphics Processing Unit) workstation cluster environment. Once the sparse matrix is too large to be processed in a single workstation, this paper discusses how to realize the optimized data segmentation in the distributed multi-GPU computing environment. For the problem of iterative solution of matrix preprocessing, two preprocessing strategies of matrix bandwidth reduction parallelization and incomplete Cholesky decomposition are proposed, and asynchronous task concurrency and load balancing strategies are designed on the architecture. The calculation of some examples in the standard sparse matrix database shows that the algorithm and architecture proposed in this paper have the ability to solve large-scale sparse matrix quickly and efficiently, and can complete the fast strength verification of vehicle components.

A parallel cost‐based abductive reasoning system on workstation cluster

AbstractThis paper proposes a method for dynamic load balancing which works efficiently in the workstation cluster environment. A parallel abductive reasoning system based on the proposed method is also proposed. In the workstation cluster environment, parallel processing is performed using workstations with various processing abilities and loads. Thus, dynamic load balancing must consider the performance and loading conditions of the workstations. This paper proposes a method of estimating the processing ability of the workstation during operation, and also a dynamic load balancing algorithm to balance the estimated conditions. Using the proposed method, a parallel abductive reasoning system which can derive the most preferable explanation for a given observation is implemented. The results of experiments on the system are also reported. © 2006 Wiley Periodicals, Inc. Syst Comp Jpn, 37(3): 80–89, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.10190

Supporting parallel computing on a distributed object architecture

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 effective parallel computing. However, the ease of developing efficient high-performance parallel software to exploit these platforms presents a major challenge. Advances in distributed object software technology have made the management of distributed computing resources easier than before. This also brings many benefits for parallel computing. Firstly, distributed object technology facilitates the encapsulating of parallel computing resources into a uniform model despite their differences in implementations that are based on different languages executing on different platforms. Secondly, mature object-oriented analysis, design method, as well as component idea embodied in distributed object technology can enhance the reusability of parallel software. To support parallel computing in a distributed object-based computing platform, a uniform high performance distributed object architecture layer is necessary. In this paper, we propose a distributed object-based framework called DoHPC to support parallel computing on distributed object architectures. We present the use of dependence analysis technique to exploit intra-object parallelism and an interoperability model for supporting distributed parallel objects. Experimental results on a Fujitsu AP3000 workstation cluster consisting of a cluster of 32 UltraSPARC workstations show that the implementation of inter-object parallelism on a workstation cluster environment is efficient. With intra-object parallel computation speedup efficiency is greater than 90% and with overhead of less than 10% for large problem, and the interoperability model improves speedup by 20%.

Real time signal processing in the clinical setting.

We routinely use a variety of real time signal acquisition, enhancement, and display techniques in the operating room to provide the surgeon with functional information. This enables reduction of surgical morbidity in cases which present a significant risk to the nervous system. Here we present regression based signal processing algorithms which produce considerable signal-to-noise-ratio enhancement with corresponding reduction in the time required to obtain an interpretable neurophysiological signal. We also present the approach we have applied to fault tolerance and distributed data display for our workstation cluster environment.

An incremental elastic-plastic Finite Element solver in a workstation cluster environment Part I. Formulations and parallel processing

We discuss solution schemes for the incremental elastic-plastic structural problem, discretized by means of the Finite Element method. Attention is focused on their formulation and implementation in a parallel computing environment defined by a cluster of workstations connected by means of a network. The availability of parallel computers allows one to consider possible formulations and solution strategies so far not considered competitive with the classical Newton-like schemes implying the definition of an elastic-plastic tangent stiffness matrix. The solution strategies herein considered are based on the explicit integration of the actual elastic-plastic rate problem. This, in turn, is phrased in terms of two different formulations, whose relative advantages—particularly with respect to their integration in parallel—are discussed. A u̇ − gl (displacemen plastic multiplier) formulation of the structural rate theory of plasticity [1], integrated by means of an explicit, element-by-element scheme, seems to be the most promising one.

An incremental elastic-plastic Finite Element solver in a workstation cluster environment Part II. Performance of a first implementation

This work illustrates results obtained by implementing in a parallel computer environment the gl rate formulation of the theory of plasticity and its integration scheme as illustrated in the preceding Part I. The Preconditioned Conjugate Gradient is used as a tool for repeatedly solving linear systems of equations. Although the investigation about the performance of the workstation cluster as a parallel virtual computer is still far from being completed, it is already possible to conclude that, in such an environment, several techniques proposed for reducing the computer time required by the iterative solver are not applicable. One purpose of the work is therefore to give guidelines in terms of expected performances of the Conjugate Gradient method when applied to stiff problems, in which the condition number may shoot up to the billions. Even if the implemented computer code is not based on the most convenient rate formulation in terms of parallelizability, as shown in Part I of this work, the obtained results indicate anyway that, for some categories of structural problems, the development of a parallel, element-by-element computer code is a promising line of work.