An agglomeration multigrid strategy is developed and implemented for the solution of three-dimensional steady viscous flows. The method enables convergence acceleration with minimal additional memory overhead and is c...
详细信息
An agglomeration multigrid strategy is developed and implemented for the solution of three-dimensional steady viscous flows. The method enables convergence acceleration with minimal additional memory overhead and is completely automated in that it can deal with grids of arbitrary construction. The multigrid technique is validated by comparing the delivered convergence rates with those obtained by a previously developed overset-mesh multigrid approach and by demonstrating grid-independent convergence rates for aerodynamic problems on very large grids. Prospects for further increases in multigrid efficiency for high-Reynolds-number viscous flows on highly stretched meshes are discussed.
The sound radiated by isotropic turbulence is computed using inertial range scaling expressions for the relevant two-time and two-point correlations. The result depends on whether the decay of Eulerian time correlatio...
详细信息
The sound radiated by isotropic turbulence is computed using inertial range scaling expressions for the relevant two-time and two-point correlations. The result depends on whether the decay of Eulerian time correlations is dominated by large scale sweeping or by local straining: the straining hypothesis leads to an expression for total acoustic power given originally by Proudman, whereas the sweeping hypothesis leads to a more recent result due to Lilley. (C) 1996 American Institute of Physics.
An implicit method for the computation of unsteady Rows on unstructured grids is presented. Following a finite difference approximation for the time derivative, the resulting nonlinear system of equations is solved at...
详细信息
An implicit method for the computation of unsteady Rows on unstructured grids is presented. Following a finite difference approximation for the time derivative, the resulting nonlinear system of equations is solved at each time step by using an agglomeration multigrid procedure. The method allows for arbitrarily large time steps and is efficient in terms of computational effort and storage. Inviscid and viscous unsteady flows are computed to validate the procedure. The issue of the mass matrix which arises with vertex-centered finite volume schemes is addressed. The present formulation allows the mass matrix to be inverted indirectly. A mesh point movement and reconnection procedure is described that allows the grids to evolve with the motion of bodies. As an example of flow over bodies in relative motion, flow over a multi-element airfoil system undergoing deployment is computed. (C) 1996 Academic Press, Inc.
An enhanced diffusion-reaction system (DRS) is proposed as a statistical model for the evolution of multiple scalars undergoing mixing and reaction in an isotropic turbulence field. The DRS model is dose enough to the...
An enhanced diffusion-reaction system (DRS) is proposed as a statistical model for the evolution of multiple scalars undergoing mixing and reaction in an isotropic turbulence field. The DRS model is dose enough to the scalar equations in a reacting how that other statistical models of turbulent mixing, which decouple the velocity field from scalar mixing and reaction (e.g. mapping closure model, assumed-pdf models), cannot ditinguish the model (DRS) equations from the exact equations. Numerical simulations of the DRS equations are performed for three scalars evolving from non-premixed initial conditions. A simple one-step reversible reaction is considered. The data from the simulations are used to (i) study the effect of chemical conversion on the evolution of scalar statistics, and (ii) evaluate other models (mapping-closure model, assumed multivariate beta-pdf model) leading to inferences about engineering approaches to turbulent combustion modeling.
Parallel direct execution simulation is an important tool for performance and scalability analysis of large message passing parallel programs executing on top of a virtual computer. However, detailed simulation of mes...
Compressible mixing layers are analyzed using a dilatational covariances model based on a pseudo-sound constitutive relation. The calculations are used to evaluate the different physical phenomena affecting compressib...
详细信息
Developing efficient programs for distributed systems is difficult because computations must be efficiently distributed and managed on multiple processors. In particular, the programmer must partition functions and da...
详细信息
Developing efficient programs for distributed systems is difficult because computations must be efficiently distributed and managed on multiple processors. In particular, the programmer must partition functions and data in an attempt to find a reasonable balance between parallelism and overhead. Furthermore, it is very expensive to code an algorithm only to find out that the implementation is not efficient. As a result, it is often necessary to determine and examine those characteristics of an algorithm that can be used to predict its suitability for a distributed computing system. In earlier work (Wilson, 1994 and Wilson and Gonzalez, 1994), we presented a framework for the study of synchronization and communication effects on the theoretical performance of common homogeneous algorithmic structures. In particular, we examined the synchronous, asynchronous, nearest-neighbor, and asynchronous master-slave structures in terms of expected execution times. In this paper, we examine the effects of synchronization and communication on the expected execution times of heterogeneous algorithmic structures. Specifically, we consider structures containing two different types of tasks, where the execution times of the tasks follow one of two different uniform distributions or one of two different normal distributions. Furthermore, we compare the expected execution times of the heterogeneous algorithmic structures with times for corresponding homogeneous structures. Finally, we develop bounds for the expected execution times of the heterogeneous structures and compare those bounds to simulated execution times.
In earlier work, we examined the theoretical effects of synchronization and communication on execution times for different categories of algorithmic structures. The paper demonstrates how two candidate problems for di...
详细信息
In earlier work, we examined the theoretical effects of synchronization and communication on execution times for different categories of algorithmic structures. The paper demonstrates how two candidate problems for distributed systems can be represented by different algorithmic structures. Specifically, we examine different representations of a partial differential equation (PDE) solver and the travelling salesman problem (TSP). Furthermore, we present results from execution of the algorithms on the Pleiades/ESP System and the Intel Paragon. Using the results from problems executed on real systems, we confirm that our theoretical framework for obtaining bounds on execution time can be used to accurately estimate the performance of a distributed algorithm.
Parallel direct execution simulation is an important tool for performance and scalability analysis of large message passing parallel programs executing on top of a virtual computer. However, detailed simulation of mes...
详细信息
Parallel direct execution simulation is an important tool for performance and scalability analysis of large message passing parallel programs executing on top of a virtual computer. However, detailed simulation of message-passing codes requires a great deal of computation. We are therefore interested in pursuing implementation techniques which can decrease this cost. One idea is to implement the application virtual processes as lightweight threads rather than traditional Unix processes, reducing both on-processor communication costs and context switching costs. We describe an initial implementation of a thread-based parallel direct execution simulator. We discuss the advantages of such an approach and present preliminary results that indicate a significant improvement over the process-based approach.
暂无评论