• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• 제13권1호
• /
• pp.13-20
• /
• 2009

The IPSAP which is a finite element analysis program has been developed for high parallel performance computing. This program consists of various analysis modules - stress, vibration and thermal analysis module, etc. The M orthogonal block Lanczos algorithm with shiftinvert transformation is used for solving eigenvalue problems in the vibration module. And the multifrontal algorithm which is one of the most efficient direct linear equation solvers is applied to factorization and triangular system solving phases in this block Lanczos iteration routine. In this study, the performance enhancement procedures of the IPSAP are composed of the following stages: 1) communication volume minimization of the factorization phase by modifying parallel matrix subroutines. 2) idling time minimization in triangular system solving phase by partial inverse of the frontal matrix and the LCM (least common multiple) concept.

• Structural Engineering and Mechanics
• /
• 제38권4호
• /
• pp.405-416
• /
• 2011

An efficient two-level domain decomposition parallel algorithm is suggested to solve large-DOF structural problems with nonlinear material models generating unsymmetric tangent matrices, such as a group of plastic-damage material models. The parallel version of the stabilized bi-conjugate gradient method is developed to solve unsymmetric coarse problems iteratively. In the present approach the coarse DOF system is solved parallelly on each processor rather than the whole system equation to minimize the data communication between processors, which is appropriate to maintain the computing performance on a non-supercomputer level cluster system. The performance test results show that the suggested algorithm provides scalability on computing performance and an efficient approach to solve large-DOF nonlinear structural problems on a cluster system.

• 전자공학회논문지B
• /
• 제30B권5호
• /
• pp.120-128
• /
• 1993

In this paper, an implementation of high-speed parallel processing system for neural network design on the multicomputer network is presented. Linear speedup expandability is increased by reducing the synchronization penalty and the communication overhead. Also, we presented the parallel processing models and their performance evaluation models for each of the parallization methods of the neural network. The results of the experiments for the character recognition of the neural network bases on the proposed system show that the proposed approach has the higher linear speedup expandability than the other systems. The proposed parallel processing models and the performance evaluation models could be used effectively for the design and the performance estimation of the neural network on the multicomputer network.

• 대한기계학회논문집A
• /
• 제21권12호
• /
• pp.2096-2105
• /
• 1997

A parallel approach using a network of engineering workstations is presented for the efficient computation in the elastoplastic analysis of strip rolling process. The domain decomposition method coupled with the frontal solver for elimination of internal degrees of freedom in each subdomain is used. PVM is used for message passing and synchronization between processors. A 2-D plane strain problem and the strip rolling process are analyzed to demonstrate the performance of the algorithm and factors that have a great effect on efficiency are discussed. In spite of much communication time on the network the result illustrates the advantages of this parallel algorithm over its corresponding sequential algorithm.

• 대한전기학회논문지
• /
• 제39권10호
• /
• pp.1057-1065
• /
• 1990

A parallel processing algorithm for the discrete Kalman filter, which is one of the most commonly used filtering techniques in modern control, signal processing, and communication, is proposed. To decrease the number of computations critical in the Kalman filter, previously proposed parallel algorithms are of the hierarchical structure by distributed processing of measurements, or of the systolic structure to disperse the computational burden. In this paper, a new parallel Kalman filter employing a structure similar to recursive doubling is proposed. Estimated valuse of state variables by the new algorithm converge faster to the true values because the new algorithm can process data twice faster than the conventional Kalman filter. Moreover, it maintains the optimality of the conventional Kalman filter.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1990년도 하계학술대회 논문집
• /
• pp.64-67
• /
• 1990

A parallel processing algorithm for discrete Kalman filter, which is one of the most commonly used filtering technique in modern control, signal processing, and communication. is proposed. Previously proposed parallel algorithms to decrease the number of computations needed in the Kalman filter are the hierachical structures by distributed processing of measurements, or the systolic structures to disperse the computational burden. In this paper, a new parallel Kalman filter employing a structure similar to recursive doubling is proposed. Estimated values of state variables by the new algorithm converge with two times faster data processing speed than that of the conventional Kalman filter. Moreover it maintains the optimality of the conventional Kalman filter.

• 정보처리학회논문지:컴퓨터 및 통신 시스템
• /
• 제4권10호
• /
• pp.337-342
• /
• 2015

이 논문에서는 소수 추출 방법인 Sieve of Eratosthenes 알고리즘을 병렬화하여 실행 시간과 에너지 소모 면에서 개선된 효과를 얻고자 실험을 진행하였다. 성능 개선을 위해 부하 균등화를 정교하게 조절하도록 파이프라인 작업 방식을 도입하였고, 멀티코어 컴퓨터 클러스터에 하이브리드 병렬 프로그래밍 모델을 활용하여 효과를 높였다. 소규모 컴퓨터 클러스터와 저전력 컴퓨터에서 구현, 실험한 결과 이전 방식보다 연산 속도가 향상되었고, 에너지 사용량도 감소함을 확인하였다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2003년도 하계학술대회 논문집 D
• /
• pp.2358-2361
• /
• 2003

In this paper, parallel control method that enables coordinating operations of multiple mini ice storage air conditioner is proposed based on LonWorks system. The proposed parallel control method is a kind of master/slave control method based on LonTalk network communication. In the control method, master controller leading whole system operations is determined by date to avoid the disproportion of outdoor unit operations. In addition to the above, duty shift control method is proposed to avoid redundant operations of outdoor units and to equalize the operation frequencies of them. Through some simulation tests, the proposed parallel control method and the duty shift control method are evaluated and the efficiencies of them are validated by comparing with existing methods.

• 전력전자학회논문지
• /
• 제18권3호
• /
• pp.256-262
• /
• 2013

In the large system such DC distribution for building, the method that a number of modules converters operation in parallel is commonly used. When parallel operation, circulating current is directly related to the loss of the entire system. Accordingly, each module to share the same current is the most important for the safety of the power system. In this paper, control method for reducing circulating current in parallel operation is proposed. furthermore response and operation of steady-state with parallel system was verified by simulation and experiment results.

• 전자공학회논문지C
• /
• 제35C권4호
• /
• pp.38-49
• /
• 1998

A fossil power plant can be modeled by a lot of algebraic equations and differential equations. When we simulate a large, complicated fossil power plant by a computer such as workstation or PC, it takes much time until overall equations are completely calculated. Therefore, new processing systems which have high computing speed is ultimately needed for real-time or high-speed(faster than real-time) simulators. This paper presents an enhanced strategy in which high computing power can be provided by parallel processing of DSP processors with communication links. DSP system is designed for general purpose. Parallel DSP system can be easily expanded by just connecting new DSP modules to the system. General urpose DSP modules and a VME interface module was developed. New model and techniques for the task allocation are also presented which take into account the special characteristics of parallel I/O and computation. As a realistic cost function of task allocation, we suggested 'simulation period' which represents the period of simulation output intervals. Based on the development of parallel DSP system and realistic task allocation techniques, we cound achieve good efficiency of parallel processing and faster simulation speed than real-time.