• 한국전산유체공학회:학술대회논문집
• /
• 1999.11a
• /
• pp.115-126
• /
• 1999

A two-dimensional Navier-Stokes solver using hybrid meshes is parallelized with a domain decompostion method. The focus of this paper is placed on minimizing the amount of effort in parallelizing the serial version of the solver, and this is achieved by adding an additional layer of cells to each decomposed domain. Most subroutines of the serial solver are used without modification, and the information exchange between neighboring domains is achieved using MPI(Message Passing Interface) library. Load balancing among the processors and scheduling of the message passing are implemented to reduce the overhead of parallelization, and the speed-up achieved by parallelization is measured on the transonic invisicd and turbulent flow problems. The parallelization efficiencies of the explicit Runge-Kutta scheme and the implicit point-SGS scheme are compared and the effects of various factors on the results are also studied.

• Atmosphere
• /
• v.22 no.1
• /
• pp.109-115
• /
• 2012

We parallelized the CFD_NIMR model, which is a numerical meteorological model, for best performance on both of distributed and shared memory parallel computers. This hybrid parallelization uses MPI (Message Passing Interface) to apply horizontal 2-dimensional sub-domain out of the 3-dimensional computing domain for distributed memory system, as well as uses OpenMP (Open Multi-Processing) to apply vertical 1-dimensional sub-domain for utilizing advantage of shared memory structure. We validated the parallel model with the original sequential model, and the parallel CFD_NIMR model shows efficient speedup on the distributed and shared memory system.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.33 no.3
• /
• pp.171-178
• /
• 2020

In this study, we develop MPI-OpenMP hybrid parallelization for multibody peridynamic simulations. Peridynamics is suitable for analyzing complicated dynamic fractures and various discontinuities. However, compared with a conventional finite element method, nonlocal interactions in peridynamics cost more time and memory. In multibody peridynamic analysis, the costs increase due to the additional interactions that occur when computing the nonlocal contact and ghost interlayer models between adjacent bodies. The costs become excessive when further refinement and smaller time steps are required in cases of high-velocity impact fracturing or similar instances. Thus, high computational efficiency and performance can be achieved by parallelization and optimization of multibody peridynamic simulations. The analytical code is developed using an Intel Fortran MPI compiler and OpenMP in NURION of the KISTI HPC center and parallelized through MPI-OpenMP hybrid parallelization. Further parallelization is conducted by hybridizing with OpenMP threads in each MPI process. We also try to minimize communication operations by model-based decomposition of MPI processes. The numerical results for the impact fracturing of multiple bodies show that the computing performance improves significantly with MPI-OpenMP hybrid parallelization.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2022.01a
• /
• pp.391-392
• /
• 2022

본 논문에서는 SparkHDE-EM이라는 생태학적 모델 알고리즘에 기반한 하이브리드 DE를 제안한다. 그리고 Spark 기반 아일랜드 모델을 도입하여 다양한 DE 변종의 병렬화를 구현한다. 또한 Monod 모델을 활용하여 자원 간의 균형을 유지하는 방법을 제안한다.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.21 no.1
• /
• pp.113-124
• /
• 2008

Distributed operation of overall structural design process, by which product optimization and process parallelization are simultaneously implemented, is presented in this paper. The database-interacted hybrid method, which selectively takes the accustomed procedure of the conventional method in the framework of the optimal design, is utilized here. The staged application of design constraints reduces the computational burden for large complex optimization problems. Two kinds of numeric and graphic processes are simultaneously implemented by concurrent engineering approach in the distributed environment of PC networks. The former is based on finite element optimization method and the latter is represented by AutoCAD using AutoLISP programming language. Numerical computation and database interaction on servers and graphic works on independent clients are communicated through message passing. The numerical experiments for some steel truss models show the validity and usability of the method. This study has sufficient adaptability and expandability, in that it is based on general methodologies and industry standard platforms.

• Journal of applied mathematics & informatics
• /
• v.4 no.1
• /
• pp.1-16
• /
• 1997

In this paper we study efficient parallel implementation for hybrid iterative methods BICGSTAB and BICGSTAB $(\ell)$ with ${Well}=2$ on the CRAY C90 and the efficiency of their parallel performance is evaluated. numerical experiments suggest that on the CRAY C90 a parallel inner product algorithm called PDOTB be used for the par-allelization of hybrid iterative methods containing sensitive values of inner products. Lastly it is shown that the number of iterations in which parallel hybrid iterative methods satisfy a certain convergence criterion depends on the number of processors to be used.

• Nuclear Engineering and Technology
• /
• v.53 no.12
• /
• pp.3902-3909
• /
• 2021

SACOS series of subchannel analysis codes have been developed by XJTU-NuTheL for many years and are being used for the thermal-hydraulic safety analysis of various reactor cores. To achieve fine whole core pin-level analysis, the input preprocessing and parallel capabilities of the code have been developed in this study. Preprocessing is suitable for modeling rectangular and hexagonal assemblies with less error-prone input; parallelization is established based on the domain decomposition method with the hybrid of MPI and OpenMP. For domain decomposition, a more flexible method has been proposed which can determine the appropriate task division of the core domain according to the number of processors of the server. By performing the calculation time evaluation for the several PWR assembly problems, the code parallelization has been successfully verified with different number of processors. Subsequent analysis results for rectangular- and hexagonal-assembly core imply that the code can be used to model and perform pin-level core safety analysis with acceptable computational efficiency.

• KIPS Transactions on Computer and Communication Systems
• /
• v.4 no.10
• /
• pp.337-342
• /
• 2015

We develop a new parallelization method for Sieve of Eratosthenes algorithm, which enhances both computation speed and energy efficiency. A pipeline scheduling is included for better load balancing after proper workload partitioning. They run on multicore CPUs with hybrid parallel programming model which uses both message passing and multithreading computation. Experimental results performed on both small scale clusters and a PC with a mobile processor show significant improvement in execution time and energy consumptions.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.416-417
• /
• 2022

In this paper, we propose a hybrid DE based on an ecological model algorithm called SparkHDE-EM. This model implements the parallelization of various DE variants by introducing an island model based on Spark, and utilizes the Monod model to maintain a balance between resources.

• Journal of the KSME
• /
• v.30 no.3
• /
• pp.267-274
• /
• 1990

기술분야의 발전이 거듭됨과 병행하여, 산업계의 설계 해석 부서에서 CAE의 중요성은 컴퓨팅 시 스템과 더불어 더욱 강조되고 있다. 슈펴 컴퓨팅 파워의 필요성은, 첫째, applied mechanics의 advanced 해석분야, 둘째, 대형 엔지니어링 문제해석 결과처리 (data acquisition & management), 셋째, real time response와 integrated 엔지니어링 시스템개발 (CAD/CAM 데이터와의 연계성) 등의 측면에서 나타난다고 볼 수 있다. 원가, 안정성 그리고 신뢰성은 모두 만족시키는 효율적인 방안으로서는 대형 수퍼 컴퓨터와 마이크로 컴퓨터의 중간 역할을 할 수 있는 hybrid type의 미니 수퍼 시스템, 즉 departmental highly-parallel 시스템의 등장이 필수적이라고 할 수 있다. 또한, 산업계의 설계 해석 지원과 관련, 구조, 유체, 동력학 등의 분야별 응용 문제와 numerical formulation등의 특성에 적합한 시스템 configuration과 프로세싱이 개발되어야 한다. 이는 실 수요자의 상황에 맞는 "목적성 전용 machine"의 등장이 가능해 질 것으로 판단된다. 향 후, vectorization 그리고 parallelization 된 소프트웨어의 가용성이 극복해야 할 큰 과제로 남아 있 다고 본다.