High‐frequency simulations and compact models compared with measurements for passive on‐chip components

Dedicated on‐chip passive test structures to test the quality of high‐frequency simulations and compact models are described. Based upon the simulations a compact (SPICE) model is calculated. The paper shows the quantitative correlation between measurements and simulations for a resistor, a coplanar line and a metal insulator metal capacitor. Agreement of less than 10% is obtained between measurements and simulations. Copyright © 2005 John Wiley & Sons, Ltd.     

A Fast Domain Decomposition High Order Poisson Solver

We present a fast high-order Poisson solver for implementation on parallel computers. The method uses deferred correction, such that high-order accuracy is obtained by solving a sequence of systems with a narrow stencil on the left-hand side. These systems are solved by a domain decomposition method. The method is direct in the sense that for any given order of accuracy, the number of arithmetic operations is fixed. Numerical experiments show that these high-order solvers easily outperform standard second-order ones. The very fast algorithm in combination with the coarser grid allowed for by the high-order method, also makes it quite possible to compete with adaptive methods and irregular grids for problems with solutions containing widely different scales.     

Enhancing CLP branch and bound techniques for scheduling problems

In this paper, we propose a constraint logic programming (CLP) approach to the solution of a job shop scheduling problem in the field of production planning in orthopaedic hospital departments. A pure CLP on finite domain (CLP(FD)) approach to the problem has been developed, leading to disappointing results. In fact, although CLP(FD) has been recognized as a suitable tool for solving combinatorial problems, it presents some drawbacks for optimization problems. The main reason concerns the fact that CLP(FD) solvers do not effectively handle the objective function and cost‐based reasoning through the simple branch and bound scheme they embed. Therefore, we have proposed an improvement of the standard CLP branch and bound algorithm by exploiting some well‐known operations research results. The branch and bound we integrate in a CLP environment is based on the optimal solution of a relaxation of the original problem. In particular, the relaxation used for the job shop scheduling problem considered is the well‐known shifted bottleneck procedure considering single machine problems. The idea is to decompose the original problem into subproblems and solve each of them independently. Clearly, the solutions of each subproblem may violate constraints among different subproblems which are not taken into account. However, these solutions can be exploited in order to improve the pruning of the search space and to guide the search by defining cost‐based heuristics. The resulting algorithm achieves a significant improvement with respect to the pure CLP(FD) approach that enables the solution of problems which are one order of magnitude greater than those solved by a pure CLP(FD) algorithm. In addition, the resulting code is less dependent on the input data configuration. Copyright © 2001 John Wiley & Sons, Ltd.     

Theoretical comparison of the FETI and algebraically partitioned FETI methods,and performance comparisons with a direct sparse solver

In this paper, we prove that the Algebraic A‐FETI method corresponds to one particular instance of the original one‐level FETI method. We also report on performance comparisons on an Origin 2000 between the one‐ and two‐level FETI methods and an optimized sparse solver, for two industrial applications: the stress analysis of a thin shell structure, and that of a three‐dimensional structure modelled by solid elements. These comparisons suggest that for topologically two‐dimensional problems, sparse solvers are effective when the number of processors is relatively small. They also suggest that for three‐dimensional applications, scalable domain decomposition methods such as FETI deliver a superior performance on both sequential and parallel hardware configurations. Copyright © 1999 John Wiley & Sons, Ltd.     

AN UNSTRUCTURED FINITE-VOLUME ALGORITHM FOR NONLINEAR TWO-DIMENSIOAL SHALLOW WATER EQUATION

An unstructured finite-volume numerical algorithm was presented for solution of the two-dimensional shallow water equations, based on triangular or arbitrary quadrilateral meshes. The Roe type approximate Riemann solver was used to the system. A second-order TVD scheme with the van Leer limiter was used in the space discretization and a two-step Runge-Kutta approach was used in the time discretization. An upwind, as opposed to a pointwise, treatment of the slope source terms was adopted and the semi-implicit treatment was used for the friction source terms. Verification for two-dimension dam-break problems are carried out by comparing the present results with others and very good agreement is shown.     

Generalized Aggregation Multilevel solver

The paper presents a Generalized Aggregation Multilevel (GAM) solver, which automatically constructs nearly optimal auxiliary coarse models based on the information available in the source grid only. GAM solver is a hybrid solution scheme where approximation space of each aggregate (group of neighbouring elements) is adaptively and automatically selected depending on the spectral characteristics of individual aggregates. Adaptive features include automated construction of auxiliary aggregated model by tracing ‘stiff’ and ‘soft’ elements, adaptive selection of intergrid transfer operators, and adaptive smoothing. An obstacle test consisting of nine industry problems, such as ring–strut–ring structure, casting setup in airfoil, nozzle for turbines, turbine blade and diffuser casing as well as on poor conditioned shell problems, such as High Speed Civil Transport, automobile body and canoe, was designed to test the performance of GAM solver. Comparison to the state of the art direct and iterative (PCG with Incomplete Cholesky preconditioner) is carried out. Numerical experiments indicate that GAM solver possesses an optimal rate of convergence by which the CPU time grows linearly with the problem size, and at the same time, robustness is not compromised, as its performance is almost insensitive to problem conditioning. © 1997 John Wiley & Sons, Ltd.     

战术导弹CFD软件求解器开发与应用

介绍了战术导弹CFD软件求解器MISSsolver．该求解器采用有限体积法离散求解任意坐标系下的Euler／N—S方程,获得绕导弹的复杂流场和全弹／部件的气动特性;采用多块结构网格技术,并包含多种时间、空间差分格式,提高数值模拟的范围与精度;具有通用的前后处理接口,方便用户使用．本文主要介绍MISSsolver求解器基本结构、功能,以及应用．     

面向GPU平台的并行结构化稀疏三角方程组求解器

稀疏三角线性方程组求解(SpTRSV)是预条件子部分的重要操作, 其中结构化SpTRSV问题, 在以迭代方法求解偏微分方程组的科学计算程序中, 是一种较为常见的问题类型, 而且通常是科学计算程序的需要解决的一个性能瓶颈. 针对GPU平台, 目前以CUSPARSE为代表的商用GPU数学库, 采用分层调度(level-scheduling)方法并行化SpTRSV操作. 该方法不仅预处理耗时较长, 而且在处理结构化SpTRSV问题时会出现较为严重GPU线程闲置问题. 针对结构化SpTRSV问题, 提出一种面向结构化SpTRSV问题的并行算法. 该算法利用结构化SpTRSV问题的特殊非零元分布规律进行任务划分, 避免对输入问题的非零元结构进行预处理分析. 并对现有分层调度方法的逐元素处理策略进行改进, 在有效缓解GPU线程闲置问题的基础上, 还隐藏了部分矩阵非零元素的访存延迟. 还根据算法的任务划分特点, 采用状态变量压缩技术, 显著提高算法状态变量操作的缓存命中率. 在此基础上, 还结合谓词执行等GPU硬件特性, 对算法实现进行全面的优化. 所提算法在NVIDIA V100 GPU上的实测性能, 相比CUSPARSE平均有2.71倍的加速效果, 有效访存带宽最高可达225.2 GB/s. 改进后的逐元素处理策略, 配合针对GPU硬件的一系列调优手段, 优化效果显著, 将算法的有效访存带宽提高了约1.15倍.     

一种求解一维理想磁流体方程组的保正拉氏方法

拉氏方法在计算流体力学中扮演了一个十分重要的角色，并且十分适合于处理含有强磁场的物理问题，例如 Z 箍缩、托卡马克、惯性约束聚变等等。在这些物理问题中密度和热力学压力总是非负的。然而，运用数值格式对上述方程进行逼近时，得到的近似解并不能总是保持这种正性。为了处理这一问题，首先构建了一种拉氏 HLLD 近似黎曼解，这一近似黎曼解在合适的信号速度下可以保持保正性质。运用这一黎曼解，提出了一种求解一维理想可压缩磁流体方程组的守恒保正拉氏格式。最后，给出一些数值算例来证明方法的保正性。     

利用规划求解工具验证在线冶金数模系统

针对在线冶金数模系统结构的整体性和复杂性,利用规划求解设计一套简化的数学运算工具验证在线系统,经与原始模型反复测试,运算结果与在线系统严格吻合,具有结构灵活简单、易于调整,便于验证等特点,对于闪速炉冶金控制模型原理的深入研究有着深远的意义.