• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.14 no.2
• /
• pp.127-134
• /
• 2001

구조물이 점점 더 커짐에 따라 그들을 분석하고 설계하는 것이 더 복잡해지고 더 많은 시간이 요구된다. 현재 사용되는 단일 프로세서를 가진 컴퓨터는 그와 같은 구조물을 해석하기에 효율적이지 못하다고 여겨진다. 이 논문에서는 거대규모의 구조물을 분석하기 위하여 컴포넌트 모우드법(CMM)과 메시지전달 시스템(MPI)을 이용함으로써 표준 병렬기법과 고도로 효율적이고 이식성 있는 프로그램을 새로 개발하였다. 이 연구에서는 구조물의 동적 해석을 위해 병렬처리기법을 지닌 컴퓨터 프로그램이 제시되고 새로 개발된 프로그램이 신뢰도를 갖고 있다는 것이 입증된다. 또한 이 프로그램은 상업용 프로그램보다 훨씬 처리속도가 빠르고 병렬처리 컴퓨터에서도 사용될 수 있다는 것을 보여준다.

• Journal of Korean Society of Water and Wastewater
• /
• v.16 no.6
• /
• pp.670-678
• /
• 2002

Since Korean government imposed a stricter regulation on effluent T-N and T-P concentrations from wastewater treatment plant, a new process has to be developed to meet these rules and this process should remove T-N and T-P, economically, from weak wastewater that is typical for Korea's combined sewer system sewage. In this study, a computer simulator, BioWin from EnviroSim, Inc. was used. Three processes - A2/O, Modified Johannesburg, UCT- had been simulated under same operational conditions and a new process - Parallel BNR Process - had been developed based on these simulation results. The Parallel BNR process consists of two rows of reactors: One row has anaerobic and aerobic reactors in series, and the other row has RAS anoxic1 and RAS anoxic2 reactors in series. In order to ensure anaerobic state in anaerobic tank, a part of influent is fed to RAS anoxic1 tank in second row. This process had been simulated under same conditions of other three processes and the simulation results were compared. The results showed that three existing processes could not perform biological phosphorus removal when the average influent was fed at any operation temperatures. However, the Parallel BNR process was found that biological phosphorus removal could be performed when both design and average influent were fed at any operation temperatures. This process showed the T-N concentration in effluent had a maximum value of 15mg/L when design influent was fed at $13^{\circ}C$ and a minimum value of 14mg/L when average influent was fed at $20^{\circ}C$. Also, T-P concentrations had a maximum value of 1.3mg/L when average influent was fed at $20^{\circ}C$ and a minimum value of 1.1mg/L when design influent was fed at $13^{\circ}C$. Based on these results, we found that this process can remove nitrogen and phosphorus biologically under any operational conditions.

• Annual Conference of KIPS
• /
• 2011.11a
• /
• pp.369-372
• /
• 2011

양 방향성 필터는 이미지표면 평탄화와 잡음제거에 좋은 성능을 보이지만 특유의 연산 복잡도로 인하여 연산 시간이 오래 걸린다는 단점이 존재한다. 따라서 본 논문에서는 고도의 병렬수행을 바탕으로 하는 그래픽연산장치(GPU)에 적합하도록 수정된 효율적인 양 방향성 필터를 NVIDIA 의 CUDA 를 사용하여 GTX 285 GPU 에서 구현하였다. 영상의 전 영역을 참조하는 대신 인접하고 연속된 영역으로의 근사화, 적은 메모리 사용량, 빠른 접근속도를 가지며 충돌이 최소화된 공유메모리 버퍼, Warp 를 고려한 병합된 메모리 접근방법을 바탕으로 병렬화 하였다. 그 결과, 같은 방식의 순차실행 알고리즘 대비 최소 약 34 배에서 최대 약 76 배의 속도 개선과 30dB 내외의 PSNR 을 갖는 양 방향성 필터를 구현할 수 있었다.

• The KIPS Transactions:PartA
• /
• v.8A no.3
• /
• pp.235-244
• /
• 2001

Development of network and computer technology results in many studies to use physically distributed computers as a single resource. Generally, these studies have focused on developing environments based on message passing. These environments are mainly used to solve problems for scientific computation and process in parallel suing inside parallelism of the given problems. Therefore, these environments provide high parallelism generally, while it is difficult to program and use as well as it is required to have user accounts in the distributed computers. If a given problem is divided into completely independent subproblems, more efficient environment can be provided. We can find these problems in bio-informatics, 3D animatin, graphics, and etc., so the development of new environment for these problems can be considered to be very important. Therefore, we suggest new environment called InterCom based on a proxy computing, which can solve these problems efficiently, and explain the implementation of this environment. This environment consists of agent, server, and client. Merits of this environment are easy programing, no need of user accounts in the distributed computers, and easiness by compiling distributed code automatically.

• The Proceeding of the Korean Institute of Electromagnetic Engineering and Science
• /
• v.4 no.1
• /
• pp.28-37
• /
• 1993

The future Personal Communication Service(PCS) is the total personal information communication service to satisfy the importance of information in a highly information - oriented society. Compared with the conventional communication method, it is necessary to provide considerable intelligence to the commu- nication network in order to implement the PCS which guarantees the mobility of each individual. Consequen- tly, the PCS needs a fast large database system which stores the information of each individual. To achieve this goal, this paper presents the parallel implementation of deductive database systems as a main - memory database system.

• The Journal of the Acoustical Society of Korea
• /
• v.12 no.1E
• /
• pp.39-47
• /
• 1993

2-Dimension IIR 디지털 필터를 시스토릭 어레이 구조로 실현하는 방법을 보였다. 시스토릭 어레이는 1-D IIR 디지털 필터로 부분 실현한 후 종속연결하여 구현하였다. 부분 실현한 시스토릭 어레이의 종속 연결은 신호 지연에 사용되는 요소를 감소 시킨다. 여기서 1-D 시스토릭 어레이는 local communication 접근에 의해 DG를 설계한후 SFG로의 사상을 통해 유도하였다. 유도된 구조는 매우 간단하며, 입력 샘플이 공급되어지면 매 샘플링 기간마다 새로운 출력을 얻는 매우 높은 데이터 처리율을 갖는다. 2-Dimension IIR 디지털 필터를 시스토릭 어레이로 실현함으로써 규칙적이고, modularity, local interconnection, 높은 농기형 다중처리의 특징을 갖기 때문에 VLSI 실현에 매우 적합하다. 또한 PE셀의 승산기 설계에서는 modified Booth's 알고리즘과 Ling's 알고리즘에 기초를 두고 고도의 병렬처리를 행할수 있도록 설계하였다.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.13 no.5
• /
• pp.1037-1042
• /
• 2018

The neuromorphic technology has been researched for decades, which learns and processes the information by imitating the human brain. The hardware implementations of neuromorphic systems are configured with highly parallel processing structures and a number of simple computational units. It can achieve high processing speed, low power consumption, and low hardware complexity. Recently, the interests of the neuromorphic technology for low power and small embedded systems have been increasing rapidly. To implement low-complexity hardware, it is necessary to reduce input data dimension without accuracy loss. This paper proposed a low-complexity artificial intelligent engine which consists of parallel neuron engines and a feature extractor. A artificial intelligent engine has a number of neuron engines and its controller to process multimodal sensor data. We verified the performance of the proposed neuron engine including the designed artificial intelligent engines, the feature extractor, and a Micro Controller Unit(MCU).

• Computational Structural Engineering
• /
• v.3 no.3
• /
• pp.62-65
• /
• 1990

RISC System/6000은 유닉스 시스템인 AIX를 오퍼레이팅 시스템으로 채택하였고, 기존의 RISC기술에 혁신적인 진보를 이룩하여 가격 대 성능비를 크게 높임과 동시에 시스템의 기능을 극도로 최적화 시킨 새로운 차원의 아이비엠의 고성능 시스템패밀리이다. 이 시스템은 새로운 RISC 시스템 구조인 POWER(Performance Optimization With Enhanced RISC) 개념과 제2세대 수퍼스칼라 기법 및 마이크로 채널 아키텍쳐로 설계되어 있다. 특히 하나의 사이클에서 4개 이상의 명령어를 병렬처리 하도록 설계된 수퍼스칼라 기능을 통하여 복잡한 그래픽 또는 이미지 처리 및 고도의 수치해석 기능이 뛰어나다. RISC시스템/6000은 과학기술계산업무나 멀티사용자의 일반 비즈니스용으로도 모두 뛰어난 범용 컴퓨터로 그래픽 프로세서의 선택과 함께 CAD/CAM이나 그래픽/애니메이션전용 시스템을 구성할 수 있으며, 최고 512 사용자에 이르는 멀티 사용자 시스템을 구성하여 사용할 수 있다. 이전의 유닉스 시스템에 있어서 큰 약점이었던 사용자 인터페이스와 멀티 사용자 및 테스킹이 크게 강화 되었으며, 기존의 IBM 시스템 및 타 기종과도 네트워크 구성이 용이하고 수백여종의 과학기술 적용업무를 이용할 수 있다.

• Geophysics and Geophysical Exploration
• /
• v.12 no.1
• /
• pp.8-16
• /
• 2009

The holistic inversion approach for frequency domain airborne electromagnetic data has previously been employed to simultaneously calibrate, process and invert raw frequency-domain data where prior information was available. Analternative formulation has been developed, which is suitable in the case where explicit prior information is not available. It incorporates: a multi-layer vertically-smooth conductivity model; a simplified bias parameterisation; horizontal smoothing with respect to elevation; and cluster computer parallelisation. Without using any prior data, an inversion of 8.0 million data for 3.4 million parameters yields results that are consistent with independently derived calibration parameters, downhole logs and groundwater elevation data. We conclude that the success of the holistic inversion method is not dependent on a sophisticated conceptual model or the direct inclusion of survey-area specific prior information. In addition, acquisition costs could potentially be reduced by employing the holistic approach which largely eliminates the need for high altitude zero-level measurements.

• Journal of KIISE
• /
• v.41 no.9
• /
• pp.686-698
• /
• 2014

Recently, large-scale streaming sensor data have emerged due to explosive supply of smart phones, diffusion of IoT and Cloud computing technology, and generalization of IoT devices. Also, researches on combination of semantic web technology are being actively pushed forward by increasing of requirements for creating new value of data through data sharing and mash-up in large-scale environments. However, we are faced with big issues due to large-scale and streaming data in the inference field for creating a new knowledge. For this reason, we propose the RDFS rule based parallel reasoning scheme to service by processing large-scale streaming sensor data with the semantic web technology. In the proposed scheme, we run in parallel each job of Rete network algorithm, the existing rule inference algorithm and sharing data using the HBase, a hadoop database, as a public storage. To achieve this, we implement our system and evaluate performance through the AWS data of the weather center as large-scale streaming sensor data.