• Journal of Mechanical Science and Technology
• /
• 제18권5호
• /
• pp.814-819
• /
• 2004

The conceptual design of a rotorcraft system involves many different analysis disciplines. The decomposition of such a system into several subsystems can make analysis and design more efficient in terms of the total computation time. Adaptive parallel decomposition makes the structure of the overall design problem suitable to apply the multidisciplinary design optimization methodologies and it can exploit parallel computing. This study proposes a decomposition method which adaptively determines the number and sequence of analyses in each sub-problem corresponding to the available number of processors in parallel. A rotorcraft design problem is solved and as a result, the adaptive parallel decomposition method shows better performance than other previous methods for the selected design problem.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• 제24권2호
• /
• pp.161-197
• /
• 2020

Total variation minimization is standard in mathematical imaging and there have been numerous researches over the last decades. In order to process large-scale images in real-time, it is essential to design parallel algorithms that utilize distributed memory computers efficiently. The aim of this paper is to illustrate recent advances of domain decomposition methods for total variation minimization as parallel algorithms. Domain decomposition methods are suitable for parallel computation since they solve a large-scale problem by dividing it into smaller problems and treating them in parallel, and they already have been widely used in structural mechanics. Differently from problems arising in structural mechanics, energy functionals of total variation minimization problems are in general nonlinear, nonsmooth, and nonseparable. Hence, designing efficient domain decomposition methods for total variation minimization is a quite challenging issue. We describe various existing approaches on domain decomposition methods for total variation minimization in a unified view. We address how the direction of research on the subject has changed over the past few years, and suggest several interesting topics for further research.

• 한국전산유체공학회:학술대회논문집
• /
• 한국전산유체공학회 1998년도 추계 학술대회논문집
• /
• pp.147-152
• /
• 1998

2-D flow fields are studied by using a shared memory parallel computer with a parallel flow analysis program which uses domain decomposition method and MPI library for data exchange at overlapped interface. Especially, effects of directional domain decomposition on parallel efficiency are studied for 2-D Lid-Driven cavity flow and flow through square cavity. It is known from the present study that domain decomposition along the main flow direction gives better parallel efficiency in 1-D partitioning than along the other direction. 2-D partitioning, however, is less sensitive to flow directions and gives good parallel efficiency for most of the cases considered.

• 대한수학회논문집
• /
• 제10권3호
• /
• pp.735-750
• /
• 1995

We present a domain decomposition algorithm for solving large sparse linear systems of equations arising from queuing networks. Such techniques are attractive since the problems in subdomains can be solved independently by parallel processors. Many of the methods proposed so far use some form of the preconditioned conjugate gradient method to deal with one large interface problem between subdomains. However, in this paper, we propose a "nested" domain decomposition method where the subsystems governing the interfaces are small enough so that they are easily solvable by direct methods on machines with many parallel processors. Convergence of the algorithms is also shown.lso shown.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2000년도 추계학술대회논문집A
• /
• pp.818-823
• /
• 2000

In practical design studies, most of designers solve multidisciplinary problems with complex design structure. These multidisciplinary problems have hundreds of analysis and thousands of variables. The sequence of process to solve these problems affects the speed of total design cycle. Thus it is very important for designer to reorder original design processes to minimize total cost and time. This is accomplished by decomposing large multidisciplinary problem into several multidisciplinary analysis subsystem (MDASS) and processing it in parallel. This paper proposes new strategy for parallel decomposition of multidisciplinary problem to raise design efficiency by using genetic algorithm and shows the relationship between decomposition and multidisciplinary design optimization (MDO) methodology.

• 한국전산유체공학회지
• /
• 제3권1호
• /
• pp.22-29
• /
• 1998

Parallel implementation is conducted for a SIMPLER finite volume model. The present parallelism is based on domain decomposition and explicit message passing using MPI and SHMEM. Two parallel solvers to tridiagonal matrix equation are employed. The implementation is verified on the Cray T3E system for a benchmark problem of natural convection in a sidewall-heated cavity. The test results illustrate good scalability of the present parallel models. Performance issues are elaborated in view of convergence as well as conventional parallel overheads and single processor performance. The effectiveness of a localized matrix solution algorithm is demonstrated.

• 전자공학회논문지B
• /
• 제33B권10호
• /
• pp.90-97
• /
• 1996

This paper presents an efficient parallel implementation and its performance evaluations of a SOFM neural netowrk on the multicomputer. We investigate the parallel performance as the size of a neural network N, the number of the patterns L, and the number of the processors p increase. We propose an analytica performance evaluation model for eac of the parallel implementations and verified the validity of the model through experiments. Analytical result show that the number of processors for a maximum speedup of the network decomposition nd the training-set decomposition increases in proportion to .root.N and .root.L, respectively. The performances of the both decompositions depend on the number of training patterns L and the size of the neural network N and, if L.geq.0.423N, the performance of trhe training-set decomposition is proved to be better than that of the network decomposition.

• 인터넷정보학회논문지
• /
• 제14권6호
• /
• pp.41-47
• /
• 2013

GP-GPU는 그래픽 처리를 위한 GPU의 다중쓰레드를 일반 수치 계산에 활용하여 초고속으로 계산하는 장치이다. GP-GPU에서는 CPU의 캐시메모리와는 달리 다중쓰레드가 공유하는 공유메모리의 형태로 캐시메모리를 제공하며, 공유메모리는 사용자 프로그램에서 직접 제어할 수 있다. 본 연구에서는 GP-GPU의 캐시메모리를 사용하여 계산 성능을 향상시키기 위한 블록 구조의 병렬 LU 분해 프로그램을 구현하였다. Nvidia CUDA C로 구현된 병렬 블록 LU 분해 프로그램은 동일한 GP-GPU 상에서 일반 LU 분해 프로그램에 비교하여 7~8배 이상의 속도 개선을 보였다.

• Journal of Mechanical Science and Technology
• /
• 제19권11호
• /
• pp.2007-2015
• /
• 2005

A large-scale analysis to evaluate complex material and structural behaviors is one of interesting topic in diverse engineering and scientific fields. Also, the utilization of massively parallel processors has been a recent trend of high performance computing. The objective of this paper is to introduce a parallel process system which consists of general purpose finite element analysis solver as well as parallelized PC cluster. The later was constructed using eight processing elements and the former was developed adopting both hierarchical domain decomposition method and balancing domain decomposition method. Then, to verify the efficiency of the established system, it was applied for structural analysis of steam generator in nuclear power plant. Since the prototypal evaluation results agreed well to the corresponding reference solutions it is believed that, after reinforcement of PC cluster by increasing number of processing elements, the promising parallel process system can be utilized as a useful tool for advanced structural integrity evaluation.

• 한국전산구조공학회논문집
• /
• 제22권1호
• /
• pp.35-44
• /
• 2009

본 논문에서는 대규모 3차원 구조해석에 필요한 병렬 유한요소해석을 위한 영역분할법의 적용에 대해 묘사하였다. 영역분할법을 사용한 병렬 유한요소법 시스템을 개발하였다. 절점 생성시, 절점들간의 거리가 특정절점에서의 공간함수와 같아지면 절점이 생성되어 진다. 이 절점공간함수는 퍼지지식처리에 의해 조절되어 진다. 기본적인 요소생성은 데로우니 삼각화 기법을 적용하였다. 자동요소생성 시스템을 이용한 영역분할법은 3차원 해석에 큰 도움이 된다. 공간함수와 유사하게 절점들간의 유한요소해석을 위한 병렬 수치 알고리즘으로서 영역분할법을 전체의 해석영역을 완전히 여러 개의 작은 영역으로 겹치지 않게 나누는 공역구배인 반복적 솔버와 결합시켰다. 개발된 시스템의 효용성에 대한 성능을 몇 가지 예를 통해 제시하였다.