• 정보과학회 논문지
• /
• 제42권11호
• /
• pp.1332-1338
• /
• 2015

최근 고품질, 초고해상도 실시간 렌더링 지원을 위하여 다중 GPU 렌더링에 대한 관심이 커지고 있다. 실시간 렌더링에서 여러 개의 GPU로 고성능을 달성하기 위해서는 GPU 간의 데이터 전송 지연과 프레임 합성 부하를 고려해야 한다. 이 논문은 이러한 부하를 최소화하고 다중 GPU의 효율을 향상하기 위해 split frame 렌더링의 동기화를 묵시적 질의 기반으로 향상하는 기법을 제안한다. 또한, 이러한 묵시적 동기화 기반 프레임 합성을 지원하기 위한 메시지 큐 기반의 렌더링 스케줄링 알고리즘도 제안한다. 본 알고리즘을 적용한 실험은 본 알고리즘이 기존 알고리즘 대비 200% 이상 효율을 향상함을 확인하였다.

• Journal of applied mathematics & informatics
• /
• 제8권1호
• /
• pp.165-180
• /
• 2001

This paper presents a new difference scheme for numerical solution of stiff system of ODE’s. The present study is mainly motivated to develop an absolutely stable numerical method with a high order of approximation. In this work a double implicit A-stable difference scheme with the sixth order of approximation is suggested. Another purpose of this study is to introduce automatic choice of the integration step size of the difference scheme which is derived from the proposed scheme and the one step scheme of the fourth order of approximation. The algorithm was tested by means of solving the Kreiss problem and a chemical kinetics problem. The behavior of the gas explosive mixture (H₂+ O₂) in a closed space with a mobile piston is considered in test problem 2. It is our conclusion that a hydrogen-operated engine will permit to decrease the emitted levels of hazardous atmospheric pollutants.

• Structural Engineering and Mechanics
• /
• 제28권5호
• /
• pp.549-570
• /
• 2008

Numerical integration is an efficient approach for nonlinear dynamic analysis. In this paper, general category of the implicit integration errors will be discussed. In order to decrease the errors, Dynamic Relaxation method with modified time step (MFT) will be used. This procedure leads to an alternative algorithm which is very general and can be utilized with any implicit integration scheme. For numerical verification of the proposed technique, some single and multi degrees of freedom nonlinear dynamic systems will be analyzed. Moreover, results are compared with both exact and other available solutions. Suitable accuracy, high efficiency, simplicity, vector operations and automatic procedures are the main merits of the new algorithm in solving nonlinear dynamic problems.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1986년도 한국자동제어학술회의논문집; 한국과학기술대학, 충남; 17-18 Oct. 1986
• /
• pp.212-215
• /
• 1986

For distillation columns, dynamic models which consider variable pressure and vapor holdup were studied. A most rigorous model which used the vapor hydraulic equation was studied with implicit methods. Vapor holdup must be considered in high pressure columns in order to predict dynamic responses accurately. The effect of pressure changes on the tray was only important for the vacuum column, particularly when heat input disturbances occurred. The rigorous vapor hydraulic model was shown to be useful, despite the fact that it is extremely stiff, provided an implicit integration algorithm (LSODES) is employed.

• Steel and Composite Structures
• /
• 제43권1호
• /
• pp.1-17
• /
• 2022

The free and forced nonlinear dynamic behaviors of Porous Functionally Graded Material (PFGM) plates are examined by means of a High-Order Implicit Algorithm (HOIA). The formulation is developed using the Third-order Shear Deformation Theory (TSDT). Unlike previous works, the formulation is written without resorting to any homogenization technique neither rule of mixture nor considering FGM as a laminated composite, and the distribution of the porosity is assumed to be gradually variable through the thickness of the PFGM plates. Using the Hamilton principle, we establish the governing equations of motion. The Finite Element Method (FEM) is used to compute approximations of the resulting equations; FEM is adopted using a four-node quadrilateral finite element with seven Degrees Of Freedom (DOF) per node. Nonlinear equations are solved by a HOIA. The accuracy and the performance of the proposed approach are verified by presenting comparisons with literature results for vibration natural frequencies and dynamic response of PFGM plates under external loading. The influences of porosity volume fraction, porosity distribution, slenderness ratio and other parameters on the vibrations of PFGM plate are explored. The results demonstrate the significant impact of different physical and geometrical parameters on the vibration behavior of the PFGM plate.

• 한국전산유체공학회지
• /
• 제3권2호
• /
• pp.17-26
• /
• 1998

The three-dimensional incompressible Navier-Stokes equations have been solved by a node-centered finite volume method with unstructured hybrid grids. The pressure-velocity coupling is handled by the artificial compressibility algorithm and convective fluxes are obtained by Roe's flux difference splitting scheme with linear reconstruction of the solutions. Euler implicit method with Jacobi matrix solver is used for the time-integration. The viscous terms are discretised in a manner to handle any kind of grids such as tetragedra, prisms, pyramids, hexahedra, or mixed-element grid. Inviscid bump flow is solved to check the accuracy of high order convective flux discretisation. And viscous flows around a circular cylinder and a sphere are studied to show the efficiency and accuracy of the solver.

• 전기학회논문지
• /
• 제61권8호
• /
• pp.1123-1129
• /
• 2012

This paper presents the fast steady state analysis using time-stepping finite element method for a high power three-phase transformer. The high power transformer spends huge computational cost of the time-stepping finite element method. It is because that the high power transformer requires a lot of time to reach steady state by its large inductance component. In order to reduce computational cost, in this paper, the adaptive time-step control algorithm combined with the embedded 2nd 4th singly diagonally implicit Runge-Kutta method and the analysis strategy using variation of the winding resistance are studied, and their numerical results are compared with those from the typical time-stepping finite element method.

• 연구논문집
• /
• 통권27호
• /
• pp.15-34
• /
• 1997

The 3-dimensional unsteady compressible flows around the high speed train have been simulated for the train entering a tunnel and for passing another train. The simulation method employs the implicit approximation-factorization finite difference algorithm for the inviscid Euler equations in general curvilinear coordinates. A moving grid scheme is applied in order to resolve the train movement relative to the tunnel and the other train. The velo-city and pressure fields and pressure drag are calculated to study the effects of tunnel and the other train. The side directional force which is time dependent is also computed for the passing train. Pressure distribution shows that the compression wave is generated in front of the train noise just after the tunnel entrance and proceeds along the inside of tunnel.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2005년도 학술발표회 논문집
• /
• pp.245-249
• /
• 2005

The numerical simulation of a two-phase flow In a discrete fracture network model is presented in this paper, The purpose of this work is to consider density-driven flows induced by the density difference between hot autochthonous heavy brines and injected cold water. Mechanical consequences of high pressure waves on the fracture permeability and heat exchanges between fluids and rock matrix are neglected in this study. The finite volume method is employed to discretize spatially and the system is solved by using an IMPES(Implicit Pressure-Explicit Saturation) scheme. In order to solve the strong non-linearity of the system, the Newton-Raphson algorithm is used. The well-known Buckeley-Leverett problem is adapted to validate results calculated from the model and a relatively good agreement is obtained.

• Earthquakes and Structures
• /
• 제13권3호
• /
• pp.279-288
• /
• 2017

Having the ability to keep on yielding stable solutions in problems involving high potential of instability, composite time integration methods have become very popular among scientists. These methods try to split a time step into multiple sub-steps so that each sub-step can be solved using different time integration methods with different behaviors. This paper proposes a new composite time integration in which a time step is divided into two sub-steps; the first sub-step is solved using the well-known Newmark method and the second sub-step is solved using Simpson's Rule of integration. An unconditional stability region is determined for the constant parameters to be chosen from. Also accuracy analysis is perform on the proposed method and proved that minor period elongation as well as a reasonable amount of numerical dissipation is produced in the responses obtained by the proposed method. Finally, in order to provide a practical assessment of the method, several benchmark problems are solved using the proposed method.