대용량 위성영상의 무감독 분류를 위한 k-Means Clustering 알고리즘의 병렬처리: 다중코어와 PC-Cluster를 이용한 Hybrid 방식

한수희,송정현 한국측량학회 2019 한국측량학회지 Vol.37 No.6

In this study, parallel processing codes of k-means clustering algorithm were developed and implemented in a PC-cluster for unsupervised classification of large satellite images. We implemented intra-node code using multicores of CPU (Central Processing Unit) based on OpenMP (Open Multi-Processing), inter-nodes code using a PC-cluster based on message passing interface, and hybrid code using both. The PC-cluster consists of one master node and eight slave nodes, and each node is equipped with eight multicores. Two operating systems, Microsoft Windows and Canonical Ubuntu, were installed in the PC-cluster in turn and tested to compare parallel processing performance. Two multispectral satellite images were tested, which are a medium-capacity LANDSAT 8 OLI (Operational Land Imager) image and a high-capacity Sentinel 2A image. To evaluate the performance of parallel processing, speedup and efficiency were measured. Overall, the speedup was over N / 2 and the efficiency was over 0.5. From the comparison of the two operating systems, the Ubuntu system showed two to three times faster performance. To confirm that the results of the sequential and parallel processing coincide with the other, the center value of each band and the number of classified pixels were compared, and result images were examined by pixel to pixel comparison. It was found that care should be taken to avoid false sharing of OpenMP in intra-node implementation. To process large satellite images in a PC-cluster, code and hardware should be designed to reduce performance degradation caused by file I / O. Also, it was found that performance can differ depending on the operating system installed in a PC-cluster. 본 연구에서는 대용량 위성영상의 무감독분류를 위해 k-means clustering 알고리즘의 병렬처리 코드를 개발하여 PC-cluster에서 구현하였다. 이를 위해 OpenMP (Open Multi-Processing)를 기반으로 CPU (Central Processing Unit)의 다중코어를 이용하는 intra-node 코드와 message passing interface를 기반으로 PC-cluster를 이용하는 inter-nodes 코드, 그리고 이 둘을 병용하는 hybrid 코드를 구현하였다. 본 연구에 사용한 PC-cluster는 한 대의 마스터 노드와 여덟 대의 슬래이브 노드로 구성되어 있고 각 노드에는 여덟 개의 다중코어가 장착되어 있다. PC-cluster에는 Microsoft Windows와 Canonical Ubuntu의 두 가지 운영체제를 설치하여 병렬처리 성능을 비교하였다. 실험에 사용한 자료는 두 가지 다중분광 위성영상으로서 중용량인 LANDSAT 8 OLI (Operational Land Imager) 영상과 대용량인 Sentinel 2A 영상이다. 병렬처리의 성능을 평가하기 위하여 speedup과 efficiency를 측정한 결과 전반적으로 speedup은 N/2 이상, efficiency는 0.5 이상으로 나타났다. Microsoft Windows와 Canonical Ubuntu를 비교한 결과 Ubuntu가 2-3배의 빠른 결과를 나타내었다. 순차처리와 병렬처리 결과가 일치하는지 확인하기 위해 각 클래스의 밴드별 중심값과 분류된 화소의 수를 비교하고 결과 영상간 화소대 화소 비교도 수행하였다. Intra-node 코드를 구현할 때에는 OpenMP에 의한 false sharing이 발생하지 않도록 주의해야 하고, PC-cluster에서 대용량 위성영상을 처리하기 위해서는 파일 I/O에 의한 성능저하를 줄일 수 있도록 코드 및 하드웨어를 설계해야 함을 알 수 있었다. 또한 PC-cluster에 설치된 운영체제에 따라서도 성능 차이가 발생함을 알 수 있었다.

PC Cluster에 대한 연구

김기철 서울시립대학교정보기술연구소 2001 정보기술연구소 논문집 Vol.3 No.-

This paper addresses a mapping problem on PC cluster. The virtual process topology addressed is hypercube and the physical processor topology addressed is switch based PC clusters with irregular topology. Mapping hypercubes onto switch based PC clusters with irregular topology is of practical importance since there are many parallel algorithms developed for hypercubes and switch based PC cluster with irregular topology is the most available type of PC cluster. Using the proposed mapping, many existing algorithms developed for hypercubes can be performed on low cost switch based PC clusters with irregular topology.

32개 노우드로 구성된 PC-Cluster 컴퓨터의 성능 연구

박재현,변민수,박준관,문종춘,서용권 동아대학교 해양자원연구소 2000 硏究論文集 Vol.13 No.1

Numerical test for the performance of the parallel computer of a PC-cluster type named 'Jupitor' has been performed. We used a Poisson equation as a model equation. It was shown that there exists an optimum number of nodes which gives the best efficiency(performance per node), Further, the overall performance is also dependent on the domain partitioning. Key Words : PC-Cluster Computer(PC클러스터컴퓨터), Poisson Equation( 프와송 방정식), Parallel Computation(병렬계산), PCGM(Preconditioned Conjugate Gradient Method)

온라인 선로상정사고 분산처리를 위한 SIMD구조의 PC 클러스터링

張世煥(Se-Hwan Jang),金眞鎬(Jin-Ho Kim),朴俊灝(June Ho Park) 대한전기학회 2008 전기학회논문지 Vol.57 No.7

This paper introduces the PC clustering of the SIMD structure for a distributed processing of on-line contingency to assess a static security of a power system. To execute on-line contingency analysis of a large-scale power system, we need to use high-speed execution device. Therefore, we constructed PC-cluster system using PC clustering method of the SIMD structure and applied to a power system, which relatively shows high quality on the high-speed execution and has a low price. SIMD(single instruction stream, multiple data stream) is a structure that processes are controlled by one signal. The PC cluster system is consisting of 8 PCs. Each PC employs the 2 ㎓ Pentium 4 CPU and is connected with the others through ethernet switch based fast ethernet. Also, we consider N-1 line contingency that have high potentiality of occurrence realistically. We propose the distributed process algorithm of the SIMD structure for reducing too much execution time on the on-line N-1 line contingency analysis in the large-scale power system. And we have verified a usefulness of the proposed algorithm and the constructed PC cluster system through IEEE 39 and 118 bus system.

PC 클러스터 기반의 Multi-HPSO를 이용한 안전도 제약의 경제급전

장세환(Se-Hwan Jang),김진호(Jin-Ho Kim),박종배(Jong-Bae Park),박준호(June Ho Park) 대한전기학회 2009 전기학회논문지 Vol.58 No.9

This paper proposes an approach of Multi_HPSO based on the PC cluster system to reduce or remove the stagnation on an early convergence effect of PSO, reduce an execution time and improve a search ability on an optimal solution. Hybrid PSO(HPSO) is combines the PSO(Particle Swarm Optimization) with the mutation of conventional GA(Genetic Algorithm). The conventional PSO has operated a search process in a single swarm. However, Multi_PSO operates a search process through multiple swarms, which increments diversity of expected solutions and reduces the execution time. Multiple Swarms are composed of unsynchronized PC clusters. We apply to SCED(security constrained economic dispatch) problem, a nonlinear optimization problem, which considers line flow constraints and N-1 line contingency constraints. To consider N-1 line contingency in power system, we have chosen critical line contingency through a process of Screening and Selection based on PI(performace Index). We have applied to IEEE 118 bus system for verifying a usefulness of the proposed approaches.

32개 노우드로 구성된 PC-Cluster 컴퓨터의 성능 연구

박재현(J.H.Park),변민수(M.S.Byoun),박준관(J.G.Park),문종춘(J.C.Moon),서용권(Y.K.Suh) 동아대학교 해양자원연구소 2000 동아대학교 해양자원연구소 연구논문집 Vol.13 No.-

  Numerical test for the performance of the parallel computer of a PC-cluster type named "Jupitor" has been performed. We used a Poisson equation as a model equation. It was shown that there exists an optimum number of nodes which gives the best efficiency(performance per node). Further, the overall performance is also dependent on the domain partitioning.

다수의 PC를 이용한 디지털 비디오 홀로그램의 고속 생성

박한훈(Hanhoon Park),김창섭(Changseob Kim),박종일(Jong-Il Park) 한국방송·미디어공학회 2017 방송공학회논문지 Vol.22 No.4

A Load Balanced PC-Cluster for Video-On-Demand Server Systems

Liang-Teh Lee,Hung-Yuan Chang,Der-Fu Tao,Siang-Lin Yang 보안공학연구지원센터(IJGDC) 2008 International Journal of Grid and Distributed Comp Vol.1 No.1

병렬계산에 의한 비축대칭 3차원 스핀업 유동해석

박재현,최윤환,서용권 동아대학교 생산기술연구소 2001 生産技術硏究所硏究論文集 Vol.6 No.2

In this study, spin-up flows in a rectangular container are analysed by using three-dimensional computation. In the numerical computation, we use the parallel computer system of PC-cluster type. We compared our results with those obtained by two-dimensional computation. Effect of velocity and vorticity on the flow is studied. The result shows that two-dimensional solution is in good agreement with the 3-D result. Attention is given to the region where the 3-D flow is significant.

Distribution System Reconfiguration Using the PC Cluster based Parallel Adaptive Evolutionary Algorithm

Mun Kyeong-Jun,Lee Hwa-Seok,Park June Ho,Hwang Gi-Hyun,Yoon Yoo-Soo The Korean Institute of Electrical Engineers 2005 KIEE International Transactions on Power Engineeri Vol.a5 No.3

This paper presents an application of the parallel Adaptive Evolutionary Algorithm (AEA) to search an optimal solution of a reconfiguration in distribution systems. The aim of the reconfiguration is to determine the appropriate switch position to be opened for loss minimization in radial distribution systems, which is a discrete optimization problem. This problem has many constraints and it is very difficult to find the optimal switch position because of its numerous local minima. In this investigation, a parallel AEA was developed for the reconfiguration of the distribution system. In parallel AEA, a genetic algorithm (GA) and an evolution strategy (ES) in an adaptive manner are used in order to combine the merits of two different evolutionary algorithms: the global search capability of GA and the local search capability of ES. In the reproduction procedure, proportions of the population by GA and ES are adaptively modulated according to the fitness. After AEA operations, the best solutions of AEA processors are transferred to the neighboring processors. For parallel computing, a PC-cluster system consisting of 8 PCs·was developed. Each PC employs the 2 GHz Pentium IV CPU, and is connected with others through switch based fast Ethernet. The new developed algorithm has been tested and is compared to distribution systems in the reference paper to verify the usefulness of the proposed method. From the simulation results, it is found that the proposed algorithm is efficient and robust for distribution system reconfiguration in terms of the solution quality, speedup, efficiency, and computation time.