• The Transactions of the Korea Information Processing Society
• /
• v.6 no.4
• /
• pp.901-915
• /
• 1999

Parallel computers are known to be excellent in performance per cost also satisfying scalability and high performance. However parallel machines have enjoyed limited success because of difficulty in parallel programming and non-portability between parallel machines. Recently, researchers have sought to develop data parallel language that provides machine independent programming systems. Data parallel language such as High Performance Fortran provides a basis to write a parallel program based on a global name space by partitioning data and computation, generating message-passing function. In this paper, we describe the Parallel Programming Translator(PPTran), source-to-source data parallel compiler, generating MPI SPMD parallel program from HPF input program through four phases such as data dependence analysis, partitioning data, partitioning computation, and code generation with explicit message-passing and verify the performance of PPTran

• Bulletin of the Korean Mathematical Society
• /
• v.50 no.6
• /
• pp.2035-2051
• /
• 2013

In this paper, we study numerical schemes for solving multi-dimensional option pricing problem. We compare the direct solving method and the Operator Splitting Method(OSM) by using finite difference approximations. By varying parameters of the Black-Scholes equations for the maximum on the call option problem, we observed that there is no significant difference between the two methods on the convergence criterion except a huge difference in computation cost. Therefore, the two methods are compatible in practice and one can improve the time efficiency by combining the OSM with parallel computation technique. We show numerical examples including the Equity-Linked Security(ELS) pricing based on either two assets or three assets by using the OSM with the Monte-Carlo Simulation as the benchmark.

• Proceedings of the KIEE Conference
• /
• 2001.05a
• /
• pp.84-88
• /
• 2001

This paper proposes a method of solving a unit commitment problem using parallel tabu search (PTS). The TS is efficient optimization method using meta-heuristic. In this paper, to reduce the computation time for evaluating the neighborhoods, an evaluating method only on changed part and a path relinking method as diversification strategy are proposed. To show the usefulness of the proposed method, we simulated for 10 units system and 110 units system. Numerical results show improvements in the generation costs and the computation time compared with conventional methods. Numerical results show improvements in the generation cost and the computation time compared to previously obtained results.

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.31 no.3
• /
• pp.264-271
• /
• 1995

This paper presents the data parallel processing method between a PC and a micro-controller. To implement the method a dual port RAM for a real time data processing is used. In general an A/D & D/AC card is used to send or receive the data into or from the external plant and the PC does only the computation of the A/D and the D/A data because the A/D & D/AC card does not have the ability of computation. In this study, a data parallel processing method in which the PC and micro-controller own a common dual port RAM, is introduced, so that the PC can compute the A/D and D/A data and control the plant simultaneously. The PC system with a micro-controller and the common dual port RAM is designed and its effectiveness is investigated experimentally considering the performance of both the computation of data and the controlling and monitoring the external plant.

• Journal of Internet Computing and Services
• /
• v.12 no.3
• /
• pp.131-137
• /
• 2011

GPUs were originally designed for graphic processing, and GPGPUs are general-purpose GPUs for numerical computation with high performance and low electric power. In this paper, we implemented the parallel LU factorization program for GPGPUs. In CUDA, which is computational environment for Nvidia GPGPUs, domains are divided into blocks, and multi-threads compute each sub-blocks Simultaneously. In LU factorization program, computation order should be artificially decided due to the data dependence. To resolve the data dependancy, we suggested a parallel LU program for GPGPUs, and also explained parallel reduction algorithm for partial pivoting of LU factorization. We finally present performance analysis to show efficiency of the parallel LU factorization program based on multi-threads on GPGPUs.

• Proceedings of the KSME Conference
• /
• 2002.08a
• /
• pp.55-58
• /
• 2002

Parallel finite element code has been recently developed for the analysis of the incompressible Wavier-Stokes equations using domain decomposition method. Metis and MPI libraries are used for the domain partitioning of an unstructured mesh and the data communication between sub-domains, respectively. For unsteady computation of the incompressible Navier-Stokes equations, 4-step splitting method is combined with P1P1 finite element formulation. Smagorinsky and dynamic model are implemented for the simulation of turbulent flows. For the validation performance-estimation of the developed parallel code, three-dimensional Laplace equation has been solved. It has been found that the speed-up of 40 has been obtained from the present parallel code fir the bench mark problem. Lastly, the turbulent flows around the MIRA model and Tiburon model have been solved using 32 processors on IBM SMP cluster and unstructured mesh. The computed drag coefficient agrees better with the existing experiment as the mesh resolution of the region increases, where the variation of pressure is severe.

• Proceedings of the KIEE Conference
• /
• 2001.07a
• /
• pp.327-329
• /
• 2001

This paper presents the application of Parallel genetic algorithm and parallel tabu search to search an optimal solution of a unit commitment problem. The proposed method previously searches the solution globally using the parallel genetic algorithm, and then searches the solution locally using tabu search which has the good local search characteristic to reduce the computation time. This method combines the benefit of both method, and thus improves the performance. To show the usefulness of the proposed method, we simulated for 10 units system. Numerical results show the improvements of cost and computation time compared to previous obtained results.

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.50 no.11
• /
• pp.528-536
• /
• 2001

This paper proposes a method of solving a unit commitment problem using parallel tabu search(PTS) approach. Tabu search is a heuristic optimization method that has the tabu list to control the search process. To improve the searching ability of a global solution, we used a method of exchanging solutions among connected processors as a diversification strategy, and to reduce the computation time, a new evaluating method was proposed which evaluates only a changed par. To show the usefulness of the proposed method, we simulated for 10 units system and 110 units system. Numerical results show improvements in the generation costs and the computation time compared with other conventional methods.

• Journal of Mechanical Science and Technology
• /
• v.18 no.4
• /
• pp.718-730
• /
• 2004

The solution for the natural convection in internally finned horizontal annuli is obtained by using a numerical simulation of time-dependent and two-dimensional governing equations. The fins existing in annuli influence the flow pattern, temperature distribution and heat transfer rate. The variations of the On configuration suppress or accelerate the free convective effects compared to those of the smooth tubes. The effects of fin configuration, number of fins and ratio of annulus gap width to the inner cylinder radius on the fluid flow and heat transfer in annuli are demonstrated by the distribution of the velocity vector, isotherms and streamlines. The governing equations are solved efficiently by using a parallel implementation. The technique is adopted for reduction of the computation cost. The parallelization is performed with the domain decomposition technique and message passing between sub-domains on the basis of the MPI library. The results from parallel computation reveal in consistency with those of the sequential program. Moreover, the speed-up ratio shows linearity with the number of processor.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.864-867
• /
• 2005

In this paper we introduce a design of parallel architecture for wavelet transformation on FPGA. We implement wavelet transforms though lifting scheme and apply Daubechies4 transform equations. This technique has an advantage that we can obtain perfect reconstruction of the data. We divide our process to high pass filter and low pass filter. With this division, we can find coefficients from low and high pass filters simultaneously using parallel processing properties of FPGA to reduce processing time. From the equations, we have to design real number computation module, referred to IEEE754 standard. We choose 32 bit computation that is fine enough to reconstruct data. After that we arrange the real number module according to Daubechies4 transform though lifting scheme.