• Journal of KIISE:Computer Systems and Theory
• /
• v.29 no.1
• /
• pp.35-41
• /
• 2002

It is extremely important to minimize network access time in constructing a high-performance PC cluster system For an SCI based PC cluster it is possilbe to reduce the network access time by maintaining network cache in each cluster node, This paper presents a CC-NUMA card that utilizes network cache for SCI based PC clustering The CC-NUMA card is directly plugged into the PCI solot of each node, and contains shared memory network cache, and interconnection modules. The network cache is maintained for the shared memory on the PCI bus of cluster nodes. The coherency mechanism between the network cache and the shared memory is based on the IEEE SCI standard. A CC-NUMA prototype card is developed to evaluate the performance of the system. According to the experiments. the cluster system with the CC-NUMA card showed considerable improvements compared with an SCI based clustser without network cache.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.37 no.6
• /
• pp.445-452
• /
• 2019

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.

• 한국전산유체공학회:학술대회논문집
• /
• 2005.04a
• /
• pp.41-46
• /
• 2005

A new PC cluster was designed and constructed based on Windows XP Operating system. Primary target of the present design was the high node density per rack by using the general PC parts those are cost-effective and readily available in the market. Other major design points were system cooling and the convenient maintenance using standard PC parts. Presently 24 nodes per rack seems to be optimum considering the specification of the network switching device, system cooling and power supply, but 40 nodes can be accommodated within a single rack at maximum. Windows XP was selected as a high-performance computing environment considering the cost and the convenience in acquisition, maintenance and education. Both fast-Ethernet and Gigabit Ethernet network connection were tested and compared with previous data, especially for Linux doter using Myrinet. The result shows that there is no significant difference between the operating systems and the Fast-Ethernet and/or Gigabit Ethernet are good solution for the high-performance PC cluster considering the cost and performance.

• Journal of KIISE:Computer Systems and Theory
• /
• v.29 no.10
• /
• pp.539-549
• /
• 2002

PC cluster-based Information Retrieval (IR) systems improve their performances by parallel processing of query terms using cluster nodes. However TCP/IP based communication used to exchange data between cluster nodes prevents the performance from being improved further. The user-level communication mechanisms solve the problem by eliminating the time-consuming kernel access in exchanging data between cluster nodes. The Virtual Interface Architecture (VIA) is one of the representative user-level communication mechanisms which provide low latency and high bandwidth. In this paper, we propose a VIA-based parallel IR system on a PC cluster. The IR system is implemented using the following three communication methods: Sealable Coherent Interface (SCI) based VIA, MPI on SCI based VIA, MPI on Fast Ethernet based VIA. Through experiments, the performances of the three methods are analyzed in various aspects.

• Journal of KIISE:Computer Systems and Theory
• /
• v.31 no.12
• /
• pp.716-725
• /
• 2004

It is extremely important to minimize network access time in constructing a high-performance PC cluster system. For PC cluster systems, it is possible to reduce network access time by maintaining network cache in each cluster node. This paper presents a Network Cache Coherent NUMA (NCC-NUMA) system to utilize network cache by locating shared memory on the PCI bus, and the NCC-NUMA card which is core module of the NCC-NUMA system is developed. The NCC-NUMA card is directly plugged into the PCI slot of each node, and contains shared memory, network cache, shared memory control module and network control module. The network cache is maintained for the shared memory on the PCI bus of cluster nodes. The coherency mechanism between the network cache and the shared memory is based on the IEEE SCI standard. According to the SPLASH-2 benchmark experiments, the NCC-NUMA system showed improvements of 56% compared with an SCI-based cluster without network cache.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.27 no.1
• /
• pp.35-41
• /
• 2017

This paper presents the development of a monitoring and control system for a vehicle cluster using a model-based design technique. For MBD(model-based design), MATLAB GUI(Graphic User Interface), M programs, simulink, state flow, and tool boxes are used to monitor a number of data such as warning, interrupts, and etc. connected to a real vehicle cluster. As a monitoring tool, a PC(Personal Computer) station interworks with the real vehicle cluster through the interface commands of tool boxes. Thus, unlike existing text-based designs, the MBD based vehicle cluster system provides very easy algorithm updates and addition, since it offers a number of blocks and state flow programs for each functional actions. Furthermore, the proposed MBD technique reduces the required time and cost for the development and modification of a vehicle cluster, because of verification and validation of the cluster algorithm on the monitor through a PC.

• Journal of the Korea Society of Computer and Information
• /
• v.13 no.2
• /
• pp.221-233
• /
• 2008

In general, we use a Parallel processing computer manufactured specially for the purpose of parallel processing to do high performance computings. But we can depoly and use a PC cluster composed of several common PCs instead of the very expensive parallel processing computer. A common way to get a PC cluster is to adopt the star topology network connected by a switch hub. But in this paper, we grope efficient implementations of hypercube networks based on TCP/IP to connect 8 PCs directly for more useful parallel processing environment, and make evaluations on functionality and efficiency of them using ping, netperf, MPICH. The two proposed methods of implementation are IP configuration based on link and IP configuration based on node. The results of comparison between them show that there is not obvious difference in performance but the latter is more efficient in simplicity of routing table. For verification of functionality, we compare the parallel processing results of an application in them with the same in a star network based PC cluster. These results also show that the proposed hypercube networks support a perfect parallel processing environment respectively.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2004.05a
• /
• pp.1461-1464
• /
• 2004

본 논문에서는 LAN 상에서 유휴 PC들을 사용하여 PC Cluster를 구성한 후 이를 사용한 내용기반의 화상 정보 검색(CBIR) 시스템에 관한 연구를 수행하였다. LAN 상의 유휴 PC 들은 인터넷상의 연산 자원들보다 안정되고 신뢰성이 있기 때문에 복잡한 보안 기법을 사용하지 않아도 되며 또한 연산시간이 유휴시간으로 고정되어 있기 때문에 네트워크의 부하 및 노드의 부하를 고려하는 복잡한 부하 균등화 기법이나 스케쥴링 기법이 필요로 하지 않는 특징을 갖는다. 내용기반의 화상 정보 검색은 화상 데이터의 대용량 특징으로 인해 화상 특징 추출 및 유사도 계산을 위해 많은 연산을 필요로 한다. 특히 다양한 내용기반의 정보 검색 서비스를 지원하기 위해 다중 특징(Multiple Features)을 동시에 추출하고자 할 때 연산시간은 급격히 증가한다. 따라서 이러한 내용 기반의 화상 정보 검색 시스템을 구현하기 위해 저비용의 고성능의 PC Cluster를 사용하여 전체 연산시간을 단축하고 실시간 정보검색이 가능하도록 하는 연구를 수행 하였다.

• Proceedings of the KIEE Conference
• /
• 2008.07a
• /
• pp.303-304
• /
• 2008

For the state estimation problem, the weighted least squares (WLS) method and the fast decoupled method are widely used at present. However, these algorithms can converge to local optimal solutions. Recently, modern heuristic optimization methods such as Particle Swarm Optimization (PSO) have been introduced to overcome the disadvantage of the classical optimization problem. However, heuristic optimization methods based on populations require a lengthy computing time to find an optimal solution. In this paper, we used PSO to search for the optimal solution of state estimation in power systems. To overcome the shortcoming of heuristic optimization methods, we proposed parallel processing of the PSO algorithm based on the PC cluster system. the proposed approach was tested with the IEEE-118 bus systems. From the simulation results, we found that the parallel PSO based on the PC cluster system can be applicable for power system state estimation.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.30 no.6
• /
• pp.21-30
• /
• 2002

A preconditioned Navier-Stokes code was parallelized by the domain decomposition technique, and the accuracy of the parallelized code was verified through a comparison with the result of a sequential code and experimental data. Parallel performance of the code was examined on a Myrinet based PC-cluster and a Fast-Ethernet system. Speed-up ratio was examined as a major performance parameter depending on the number of processor and the network communication topology. In this test, Myrinet system shows a superior parallel performance to the Fast-Ethernet system as was expected. A test for the dependency on problem size also shows that network communication speed in a crucial factor for parallel performance, and the Myrinet based PC-cluster is a plausible candidate for high performance parallel computing system.