一种有限元求解的多核并行方案

来源 :中国计算力学大会2018暨国际华人计算力学大会2018 | 被引量 : 0次 | 上传用户:liujj08
下载到本地 , 更方便阅读
声明 : 本文档内容版权归属内容提供方 , 如果您对本文有版权争议 , 可与客服联系进行内容授权或下架
论文部分内容阅读
  有限元分析是工程设计及科学研究中最重要和最成功的数值方法,以其为核心的CAD/CAE 软件对国家的经济建设和国家安全起到了非常关键的作用,而自主研发此类软件并提高有限元分析效率一直以来都是工程师和科研人员关注的焦点问题.目前,现代计算平台正在经历一个本质性的转型阶段.Intel、IBM、Sun 以及AMD 公司均推出了在单个芯片上集成多个执行核的CPU 产品,并快速地普及至大多数个人电脑中.同时,并行应用环境OpenMP 的编程接口提供了一组与平台无关的编译指导语句、函数调用和环境变量,可以显式地指导编译器如何以及何时利用应用程序中的并行性,使用简单,可移植性非常好,大大方便了开发人员编写高效率的并行程序.
其他文献
Literature studies demonstrate that,for nanostructures under transverse bending,the effective Youngs modulus is appreciably greater(in magnitude)than that of the same elements under axial loads.
Graphene reinforced metal matrix composite have attracted much attention in the late years,which provides a new direction for the applications of graphene and other nanomaterials.Graphene/aluminum(Gr/
As an ideal candidate for structural elements or functional components in nanoscale systems,the carbon nanotube(CNT)is expected to be accurately and conveniently controlled,including their mechanical
Oxygen diffusion is the main form of transportation in the preparation process of polycrystalline superconductor YBa2Cu3O7(YBCO)thin films.Study of the diffusion of oxygen ions in the grain boundaries
Radial deformation of carbon nanotubes(CNTs)adhered to substrates can strongly influence their physical properties and the performance of CNT-based nanodevices.
本文建立了钢材的基本组成部分Fe-C 合金在原子尺度的微观模型,利用分子动力学方法模拟Fe-C 合金的单轴拉伸过程,研究温度和应变率对Fe-C 合金拉伸力学性能的影响。结果 表明:Fe-C 合金的弹性模量、屈服强度和峰值应变都随着温度的升高而逐渐降低;应变率的变化对弹性模量没有影响,屈服强度和峰值应变则随着应变率的增大而增大。根据分子动力学模拟结果,提出了Fe-C 合金拉伸力学性能温度和应变率效应
This study investigates the vibration characteristics of composite thick shells reinforced with single walled carbon nanotubes(SWCNTs)distributed in both uniform and functionally graded(FG)patterns.
In this paper,Spectro-Geometric Method(SGM)is applied to solve the free vibration of nonlocal Timoshenko beams with various boundary condition,which includes classical boundary condition and elastic b
Multi-scale simulation of the crystallization includes two parts: one is the temperature and flow simulation on the macroscale,the other is the morphological simulation on the mesoscale.It is known th
This pater presents a multi-scale concurrent optimization method for lattice structure base on a fully parallel and easy-to-implement topology optimization framework.The optimization aims at obtaining