三维半导体器件漂移扩散模型的并行有限元方法研究

本文设计了一种新的三维自适应迎风稳定化有限元方法(SUPG-IP),并对比研究了几种半导体器件模拟的并行有限元方法.数值模拟结果表明:稳定化有限元方法适用于大偏压以及高掺杂器件模拟;而经典的Zlamal有限元方法更适用于计算半导体器件的电学响应曲线.我们基于三维并行自适应有限元平台PHG开发了半导体器件漂移扩散模型求解器DevSim,并对几种典型的半导体器件进行了模拟测试.计算结果与商业软件Sentaurus吻合较好,验证了算法的有效性.我们对PN结进行了超大规模网格并行模拟测试,网格达8亿单元并使用2048进程计算,展示了算法良好的并行可扩展性.     

大型结构并行有限元分析软件研究

有限元分析在很多领域得到了应用,但其较大的计算规模对使用造成了限制,对复杂对象的分析难以在单台计算机上完成。提出并实现了一种并行化有限元软件的开发模式,在保留成熟的商业性有限元分析软件各种优点的情况下对关键计算进行了并行化开发,以较少投入成功实现了应用于大型结构分析的并行有限元分析系统。完成了数据在分布存储计算环境下的有效存储、高效的大型稀疏矩阵的并行算法,通过试验验证系统有良好的分析精度和扩展性。     

NAPA软件的并行优化

大规模数值计算受到通信模式、并行算法、I/O速度等的多方面因素的制约,并行程序的好坏直接影响并行机性能的发挥,本文分别对上述影响并行性能的重要因素进行了分析并对NAPA软件进行了优化,测试中发现本文采用的并行算法性能比优化前提高了41.1%,此外,本文采用支持多视口的MPI I/O接口性能有明显提高.最后,本文分析了并行NAPA软件的可扩展性,并采用高超声速平板流动进行了测试,在Grid 97*49*49算例中,64个进程的情况下得到了较高的加速比(53.7)和并行效率(84%),表明,优化后的软件具有较好的并行效率和可扩展性.     

一种更有效的并行系统可扩展性模型

文中首先分析了等效率、等速度和等并行开销计算比三种并行系统可扩展性模型的特点,论证了等效率、等速度和等并行开销计算比三种条件的等价性,并指出这三种模型在描描可扩展性时的不直观及其局限性。然后提出了一种新的可扩展性模型。此模型直观地反映出并行系统在机器规模和问题规模扩展时,其性能的扩展特性。实例研究表明,该模型能更有效地解决下列问题：（1）定量研究并行系统的可扩展性;（2）全面地反映程序、机器、环境方面的因素对可扩展性的影响;（3）指导如何保持并行系统的可扩展性。     

一个基于PHG平台的并行有限元生物分子模拟解法器

本文介绍一个面向生物分子模拟的并行有限元解法器,该解法器基于三维并行自适应有限元软件平台PHG~([1]),计算并模拟在生物溶液系统在静电场下的扩散过程.该解法器的最新版本在已有算法的基础上~([2]),添加了整体求解、含时求解等一些新算法,规范并扩展了边界条件的选取,并整合多项辅助功能,现提供对于Poisson-Nernst-Planck(PNP)方程的两个含时算法和四个稳态算法,以及对于Smoluchowski-Poisson-Boltzmann(SPB)方程的一个稳态算法.解法器可模拟生物分子,离子通道和纳米管等模型,通过有限元方法计算静电场和离子浓度分布,并计算电流强度、反应速率等物理量,可研究离子通道的选择机理,酶的催化反应过程及反应速率等问题.相关软件、工具和进展见www.continuummodel.org.     

基于PHG平台的非结构四面体网格欧拉方程间断有限元并行求解器

针对计算流体力学对高性能计算的需求,基于三维并行自适应有限元程序开发平台PHG (Par allel Hierarchical Grid)开发了在非结构四面体网格上求解可压缩流欧拉方程的间断有限元法并行求解器(Libdgphg库).该求解器以C++函数库的形式实现数值方法中各项功能.实施了模态基一次间断有限元,采用低耗散的MLP (Multi-dimensional Limiting Process)限制器来抑制间断附近的数值振荡.由于MLP限制器需要所有与当前单元共享顶点的邻近单元的信息,模板较宽,这给程序设计带来一定的困难.我们通过引入辅助向量收集共享顶点的所有单元中的最大、最小单元积分平均值,并归属到单元数据结构上,从而利用PHG内在的通信机制实现MPI分区间的信息交换.通过几个数值算例测试了 Libdgphg库的数值结果以及并行性能.算例表明:该求解器能得到理论精度阶和较高分辨率,同时有良好的并行性能,在千核测试中可达到60%以上的并行效率,可用于流体问题的大规模计算.     

MIOS: 面向大规模CCNUMA系统的多实例操作系统

MIOS是一个面向大规模CCNUMA系统设计的新型高可扩展操作系统.MIOS创新地采用了多实例内核结构,每个内核实例执行相同代码,分别独立运行和管理一个处理器,多核间通过分布存储管理构成高可扩展的一致性系统映像空间,支持弱共享进程、线程并行模型.MIOS针对大规模CCNUMA系统特点和高性能并行科学计算应用的需求,采用了显式共享数据分布、层次式任务调度、自适应任务间通信以及寄存器锁等优化.在大规模CCNUMA体系结构的银河深度并行计算机上的测试表明,MIOS对MPI应用具有同传统操作系统类似的性能,并可以有效支持2048处理器规模的OMP应用高效运行,具有良好的系统可扩展性.     

基于Beowulf集群的可扩展性模型的研究

可扩展性是衡量并行算法与并行系统匹配程度的一项重要指标.分析了传统的等并行开销计算比评价准则,指出其优缺点.为了适用于基于Beowulf集群的分布式并行计算环境,对传统的等并行开销计算比评价准则进行改进,得到Beowulf环境下的新的可扩展函数.该可扩展函数能够直观地反映基于Beowulf集群的分布式并行系统在机器规模和问题规模扩展时,其性能的扩展特性.用该评价准则分析并论证了编制的并行算法与BeowuIf集群相结合的可扩展性.     

“共创联盟”引领开放创新

目前“,联盟”会员数量超过4万个,吸引了众多开源爱好者的加盟。“联盟”收集整理了近15000个成熟的开源软件,验证了150多个成熟的开源软件工程工具。近10000个整改过的程序应用到了北京、上海、广州、等地的863软件专业孵化器。“联盟”成功开发了开源软件资源库和开源创新研究支撑平台,形成了国内最大的开源软件的下载中心和开源软件协同开发平台,支持了近千个创新项目开发。     

材料微观结构演化大规模分子动力学软件比较

分子动力学模拟是研究材料辐照效应下微观结构演化的重要工具。对4款主流材料计算大规模分子动力学开源软件LAMMPS、Ls1-MarDyn、IMD和CoMD进行了详细的介绍,从数据结构、计算方法、并行分解方式、原子存储等多个方面进行了横向分析比较。通过设计算例测试了各软件的计算效率、并行性能和内存使用情况。针对压力容器关键材料辐照缺陷演化计算特点,提出了设计节省存储空间的面向单晶的新型数据结构的设想,为后续实现大时空尺度分子动力学模拟研究提供了研究基础。     

The concurrent element level processing for nonlinear dynamic analysis on a massively parallel computer

The goal of this paper is to explore parallel methodologies with the desired flexibility, generality and accuracy for nonlinear dynamic finite element analysis on massively parallel computer. This paper tests the generality of the concurrent element processing approach and proposes a basic software design strategy to fully take advantage of features available in massively parallel computers having a hierarchical ring architecture. As a testbed, a large scale general purpose code, DYNA3D as used and modified as appropriate to test proposed parallel design concepts on a KSRI parallel computer.     

SUT-DAM: An integrated software environment for multi-disciplinary geotechnical engineering

As computer simulation increasingly supports engineering design, the requirement for a computer software environment providing an integration platform for computational engineering software increases. A key component of an integrated environment is the use of computational engineering to assist and support solutions for complex design. In the present paper, an integrated software environment is demonstrated for multi-disciplinary computational modeling of structural and geotechnical problems. The SUT-DAM is designed in both popularity and functionality with the development of user-friendly pre- and post-processing software. Pre-processing software is used to create the model, generate an appropriate finite element grid, apply the appropriate boundary conditions, and view the total model. Post-processing provides visualization of the computed results. In SUT-DAM, a numerical model is developed based on a Lagrangian finite element formulation for large deformation dynamic analysis of saturated and unsaturated soils. An adaptive FEM strategy is used into the large displacement finite element formulation by employing an error estimator, adaptive mesh refinement, and data transfer operator. This consists in defining new appropriate finite element mesh within the updated, deformed geometry and interpolating (mapping) the pertinent variables from one mesh to another in order to continue the analysis. The SUT-DAM supports different yield criteria, including classical and advanced constitutive models, such as the Pastor–Zienkiewicz and cap plasticity models. The paper presents details of the environment and includes several examples of the integration of application software.     

Concepts and implementation of parallel finite element analysis

The design of complex engineering systems such as advanced aircraft structures and offshore platforms requires continually increasing levels of detail in supporting analysis. The finite element method is widely used as a computational method with which to model physical systems in various engineering problems. For detailed analyses of complex designs, structural models composed of several thousands of degrees of freedom are no longer uncommon. Such design activities require large order finite element and/or finite difference models and excessive computation demands in both calculation speed and information management. The computer simulation of the nonlinear dynamic response of structures and the implementation of parallel FEM systems on a high speed multiprocessor have received considerable attention in recent years. The driving forces of these activities included the reliable simulation of automotive and aircraft crash phenomena, and the increased performance of computers. Most existing major structural analysis software systems were designed 10–20 years ago and have been optimized for current sequential computers. Such systems often are not well structured to take maximum advantage of the recent and continuing revolution in parallel vector computing capabilities. These parallel vector computer architectures not only occur in the form of large supercomputers, but are now also occurring for minicomputers and even engineering workstations. To benefit from advances in parallel computers, software must be developed which takes maximum advantage of the parallel processing feature.     

A Large scale finite element analysis using domain decomposition method on a parallel computer

A parallel finite element analysis based on a domain decomposition technique (DDT) is considered. In the present DDT, an analysis domain is divided into a number of smaller subdomains without overlap. Finite element analyses of the subdomains are performed under the constraint of both displacement continuity and force equivalence among them. The constraint is satisfied through iterative calculations based on either the Uzawa algorithm or the Conjugate Gradient (CG) method. Owing to the iterative algorithm, a large scale finite element analysis can be divided into a number of smaller ones which can be carried out in parallel.

The DDT is implemented on a parallel computer network composed of a number of 32-bit microprocessors, transputers. The developed parallel calculation system named the ‘FEM server type system’ involves peculiar features such as network independence and dynamic workload balance.

The characteristics of the domain decomposition method such as computational speed and memory requirement are first examined in detail through the finite element calculations of homogeneous or inhomogeneous cracked plate subjected to a tensile load on a single CPU computer.

The ‘speedup’ and ‘performance’ features of the FEM server type system are discussed on a parallel computer system composed of up to 16 transputers, with changing network types and domain decompositions. It is clearly demonstrated that the present parallel computing system requires a much smaller amount of computational memory than the conventional finite element method and also that, due to the feature of dynamic workload balancing, high performance (over 90%) is achieved even in a large scale finite element calculation with irregular domain decomposition.     

3D finite element meshing from imaging data

This paper describes an algorithm to extract adaptive and quality 3D meshes directly from volumetric imaging data. The extracted tetrahedral and hexahedral meshes are extensively used in the finite element method (FEM). A top-down octree subdivision coupled with a dual contouring method is used to rapidly extract adaptive 3D finite element meshes with correct topology from volumetric imaging data. The edge contraction and smoothing methods are used to improve mesh quality. The main contribution is extending the dual contouring method to crack-free interval volume 3D meshing with boundary feature sensitive adaptation. Compared to other tetrahedral extraction methods from imaging data, our method generates adaptive and quality 3D meshes without introducing any hanging nodes. The algorithm has been successfully applied to constructing quality meshes for finite element calculations.     

基于层次化数据结构的并行有限元计算

高效数据结构及并行算法是大规模并行有限元计算的关键。本文面向非结构网格,提出了一种层次化的数据结构,并在该数据结构基础上设计了并行有限元算法。最后,用数值算例验证了数据结构和并行算法的正确性和并行可扩展性。     

面向异构超算的结构分析高效并行计算方法

In order to exploit the efficient computing power of many integrated cores on heterogeneous cluster, a multi-level and multi-granularity collaborative parallel computing method is proposed for finite element structural mechanical analysis. Computing tasks are divided into three levels: inter-node parallelism, inter-device parallelism and inter-core parallelism. Through mapping decomposablecomput- ing jobs to different hardware layers of heterogeneous MIC system, the proposed method not only effectively resolves the load balancing problem between CPU and MIC devices, but also significantly reduces the communication overheads of the system. Different engineering simulation case experiments for large scale parallel computing were conducted on “Tianhe 2” supercomputer. Up to 39000 CPU+MIC cores were employed and the finite element size of the analysis was more than 100 million units. Test results show that the proposed method can achieve good speedup and parallel computing efficiency in large scale parallel computing of finite element structural analysis. The optimized adaptation of finite element structural analysis and heterogeneous MIC computing platform is realized, which can provide reference for parallel porting and performance optimization of similar applications.     

A brief review of FEM software technique

This paper gives the names and main features of the current nearly 100 large-scale general-purpose programs in use worldwide. The progress and development of current finite element methods (FEM) software technique are analysed, and the notion that the advance in FEM methodology, which is as yet a developing frontier science, will be gradual is presented. Thereby, the traditional structural analysis calculation can be transformed into a procedure of design decision for engineers. On this basis, database management techniques and computer-aided design (CAD), which is adopted widely in modern FEM software, are introduced. It is shown that the fully integrated building block structure, the graphical I/O feature and the auto-generating feature, which are the main characteristics of the modern FEM technique, will replace the traditional program structure and pre-post processing. This paper provides the constructions, features, managements, menus, manipulative techniques, etc., of a CAD system for finite element analysis, which can be used as reference for structural engineers to develop a structural analysis CAD/CAM in this field.     

大规模结构有限元分析程序在多核集群计算环境中的性能分析和优化

通过对基于MPI编程模型实现的开源有限元计算分析软件在多核集群计算平台中的程序性能的分析,找出程序瓶颈及其原因,实现了基于MPI编程模型的并行程序在多核计算环境中的性能优化。根据程序性能瓶颈的分析,提出了基于MPI/OpenMP混合并行编程模型的大规模线性/非线性方程组求解和多线程多进程同时进行消息通信的两种程序性能优化方案。不同计算规模的实验结果表明,在多核集群计算平台中,MPI/OpenMP混合编程模型实现的大规模非线性方程组求解器相对于单纯基于MPI编程模型实现的并行程序,其性能有2倍到3倍的提升;多线程多进程同时消息传递的优化方案虽然对程序能够起到性能优化作用,但是对解决程序消息通信瓶颈的问题不是最好的方法。两个方案总体性能分析结果表明,基于MPI/OpenMP混合编程模型实现的并行程序,在多核集群计算平台中能够更好地发挥硬件系统的计算能力。     

Parallel electromagnetic modeling of 3D microstrip discontinuities using FEM and PML

An approach is developed for 3D microstrip discontinuities using the finite element method (FEM) and the perfectly matched layers (PML). It is shown that iterative solvers are not suitable for this problem since the matrix equations from the PML‐used FEM modeling are rather ill‐conditioned and lead to a very slow or nonconvergent result. A newly developed package SuperLU of the sparse LU decomposition solver is incorporated into our developed approach running on a PC‐based parallel platform. Various implementation techniques are detailed. Numerical experiments clearly show that the developed approach is reliable, efficient, and suitable for modeling 3D microstrip discontinuities. © 2001 John Wiley & Sons, Inc. Int J RF and Microwave CAE 11: 38–47, 2001