Vectorization,threading, and cache‐blocking considerations for hydrocodes on emerging architectures

The computational efficiency of existing hydrocodes is expected to suffer as computer architectures advance beyond the traditional parallel central processing unit (CPU) model 1 . Concerning new computer architectures, sources of relative performance degradation might include reduced memory bandwidth per core, increased resource contention due to concurrency, increased single instruction, multiple data (SIMD) length, and increasingly complex memory hierarchies. Concerning existing codes, any performance degradation will be influenced by a lack of attention to performance in their design and implementation. This work reports on considerations for improving computational performance in preparation for current and expected changes to computer architecture. The algorithms studied will include increasingly complex prototypes for radiation hydrodynamics codes, such as gradient routines and diffusion matrix assembly (e.g., in 1 - 6 ). The meshes considered for the algorithms are structured or unstructured meshes. The considerations applied for performance improvements are meant to be general in terms of architecture (not specifically graphical processing unit (GPUs) or multi‐core machines, for example) and include techniques for vectorization, threading, tiling, and cache blocking. Out of a survey of optimization techniques on applications such as diffusion and hydrodynamics, we make general recommendations with a view toward making these techniques conceptually accessible to the applications code developer. Published 2015. This article is a U.S. Government work and is in the public domain in the USA.     

A meshless finite point method for three‐dimensional analysis of compressible flow problems involving moving boundaries and adaptivity

A finite point method for solving compressible flow problems involving moving boundaries and adaptivity is presented. The numerical methodology is based on an upwind‐biased discretization of the Euler equations, written in arbitrary Lagrangian–Eulerian form and integrated in time by means of a dual‐time steeping technique. In order to exploit the meshless potential of the method, a domain deformation approach based on the spring network analogy is implemented, and h‐adaptivity is also employed in the computations. Typical movable boundary problems in transonic flow regime are solved to assess the performance of the proposed technique. In addition, an application to a fluid–structure interaction problem involving static aeroelasticity illustrates the capability of the method to deal with practical engineering analyses. The computational cost and multi‐core performance of the proposed technique is also discussed through the examples provided. Copyright © 2013 John Wiley & Sons, Ltd.     

Mesh regularization for an ALE code based on the limitation of the Lagrangian mesh velocity

The Lagrangian approach is usually used for the simulation of flow with strong shock waves. Moreover, this approach is particularly well suited to treatment of material interfaces in the case of multimaterial flows.Unfortunately, this formulation leads to very large deformations in the mesh. The arbitrary Lagrangian‐Eulerian method overcomes this drawback by using a mesh regularization that is based on an analysis of cell geometry. The regularization step may be considered as a method used to correct the nonconvex and potentially tangled cells that constitute the mesh. In this paper, we present a new approach to mesh regularization. Instead of using a purely geometric criterion, we propose that the mesh evolution is computed on the basis of the flow vorticity. This approach is called the large Eddy limitation method, and it is aimed here to be used in finite volume direct arbitrary Lagrangian‐Eulerian methods. The large Eddy limitation method is general, which means that it is not restricted to applications in the finite volume framework dedicated to fluid flow simulation; for instance, it could also be naturally applied to the finite element framework.     

Improved production implicit continuous‐fluid Eulerian method for compressible flow problems in Uintah

The implicit continuous‐fluid Eulerian (ICE) method is a successful and widely used semi‐implicit finite‐volume method that applies to flows that range from supersonic to subsonic regimes. The classical ICE method has been expanded to problems in multiphase flow, which spans a wide area of science and engineering. The ICE method is utilized by the Center for the Simulation of Accidental Fires and Explosions code Uintah written at the University of Utah to simulate explosions, fires and other fluid and fluid‐structure interaction phenomena. The ICE method used in Uintah (referred to here as Production ICE) is described in many papers by Kashiwa at Los Alamos National Laboratory and Harman at University of Utah. However, Production ICE does not perform as well as many current methods for compressible flow problems governed by the Euler equations. We show, via examples, that changing the nonconservation form of the solver in Production ICE to a conservation form improves the numerical solutions. In addition, the use of slope limiters makes it possible to suppress the nonphysical oscillations generated by the ICE method in conservation form. This new form of ICE is referred to as IMPICE, the IMproved Production ICE method. The accuracy of IMPICE for one‐dimensional Euler equations is investigated by using a number of test cases. Copyright © 2011 John Wiley & Sons, Ltd.     

二维拉氏辐射流体力学人为解构造方法

人为构造解方法是复杂多物理过程耦合程序正确性验证的重要方法之一,适用于二维拉氏大变形网格的流体、辐射耦合人为解模型较为少见。针对拉氏辐射流体力学程序正确性验证的需要,从二维拉氏辐射流体力学方程组出发,基于坐标变换技术,给出了拉氏空间到欧氏空间的物理变量导数关系式,开展了辐射流体耦合的人为解构造方法研究,构造了一类质量方程无源项的二维人为解模型,并应用于非结构拉氏程序LAD2D辐射流体力学计算的正确性考核,为流体运动网格上的辐射扩散计算提供了一种有效手段。数值结果显示观测到的数值模拟收敛阶与理论分析一致。

     

Development of a three-dimensional solver based on the local domain-free discretization and immersed boundary method and its application for incompressible flow problems

Recently, the author and two other coauthors have proposed a two-dimensional hybrid local domain-free discretization and immersed boundary method (LDFD-IBM), which can be used to solve the flow problem with complex geometries. In this paper, the LDFD-IBM is extended to solve a three-dimensional unsteady incompressible flow with the complex computational domain. The technical issues related to the implementation of the LDFD-IBM in three-dimensional problems are discussed in detail, particularly for the discretization of Navier-Stokes equations, mesh strategies for a three-dimensional flow, and the fast algorithm on the identification of the status of mesh nodes (ie, to identify if the mesh node is located in the solid domain, in the fluid domain, or near the immersed boundary). Numerical tests show that the LDFD-IBM can accurately solve three-dimensional incompressible problems with ease.     

Three‐dimensional computation for flow‐induced vibrations in in‐line and cross‐flow directions of a circular cylinder

We present numerical results for in‐line and cross‐flow vibrations of a circular cylinder, which is immersed in a uniform flow and is elastically supported by damper‐spring systems to compute vibrations of a rigid cylinder. In the case of a circular cylinder with a low Scruton number, it is well‐known that two types of self‐excited vibrations appear in the in‐line direction in the range of low reduced velocities. On the other hand, a cross‐flow vibration of the circular cylinder can be excited in the range of high reduced velocities. Therefore, we compute the flow‐induced vibrations of the circular cylinder in the wide range of the reduced velocities at low and high Scruton numbers and discuss about excitation mechanisms in the in‐line and cross‐flow directions. Copyright © 2011 John Wiley & Sons, Ltd.