三维多物质弹塑性流体动力学Euler方法的并行算法研究及程序测试

 并行计算是解决爆炸与冲击问题大规模数值模拟最有效的手段之一。针对Euler方法并行程序设计的复杂性,阐述了三维多物质弹塑性流体动力学程序MMIC-3D并行设计的总体策略,基于消息传递接口(MPI)设计出相应的PMMIC-3D并行程序,并提出了一套实用的程序测试方案。结合聚能射流形成过程的数值模拟算例,在八节点的集群上测试了加速比、并行效率及可扩放性,分析了影响并行性能的因素。     

基于JASMIN框架的辐射流体与粒子输运耦合计算

基于JASMIN框架的"联邦计算",将两个串行程序辐射流体RH2D与粒子输运Sn2D作为独立"邦元"耦合连接,形成的集成程序RHSn2D可以采用数千处理器并行模拟多物理耦合问题.集成程序RHSn2D中的邦元具有各自独立的网格划分与并行算法,同时借助框架技术,可以屏蔽邦元间的并行数据传递.算例表明,对于应用问题规模(90 720个网格单元,辐射流体100个Patch,粒子输运2 835个Patch,Sn方向48,16群),集成程序RHSn2D采用1 024个处理器可以达到36%的并行效率.     

三维面向对象的并行粒子模拟程序PLASIM3D

设计了基于区域分解的三维粒子模拟的并行算法,基于消息传递环境(MPI)编制了三维面向对象的并行粒子模拟程序PLASIM3D(PlasmaSimulator分别取前3个字母,3D表示三维).对激光与低密度等离子体薄靶相互作用问题作了粒子模拟计算,验证了该并行程序.最后在高性能并行机上测试并分析了并行性能,获得了接近线性的加速比.     

并行平台上的CFD第一部分:标准的建立

为解决分布式网络并行技术在计算流体力学应用中的主要问题,提出了"并行平台上的CFD通用界面标准”.应用该CFD通用界面标准简化了传统的CFD程序改为并行程序的工作,使不同的CFD核心程序可以方便而快捷的改为并行程序,应用于不同的算例中;提高了CFD并行程序的移植性,CFD核心层同并行边界层无需作任何改动就可应用到不同的并行平台上.在这个标准基础上开发了基于WINDOWSNT系统的WINSOCK并行平台,搭建了基于LINUX系统的PVM并行平台.     

基于MPI的分子对接并行算法

基于消息传递接口(Message Passing Interface,MPI),用两种不同的并行程序设计方法对Autodock程序进行修改.将修改后的程序应用于HIV-1蛋白酶(Protease)和小分子抑制剂XK263的对接体系,测试了并行程序的加速比和并行效率.结果表明,两种改进的并行Autodock程序都可以很好地完成计算,尤其是方案Ⅱ并行程序的加速比和并行效率更高.     

闪光X射线照相蒙特卡罗程序的并行化

研究了闪光X射线辐射照像蒙特卡罗程序(FXRMC)在MPI平台下的并行计算实现,给出了实现过程中并行随机数的产生方法。对并行程序的测试结果表明:并行程序与串行程序结果一致,加速比比较理想,呈线性增长,并行效率在16个处理器上可达80%以上。算例的结果说明了并行化可有效地解决程序计算散射技术性能时的耗时问题,从而有效化解FXRMC耗时和大规模计算的困难,提高了FXRMC程序的计算规模和计算速度,达到了研究要求。 (Institute of Fluid Physics, CAEP, P. O. Box 919-105, Mianyang 621900, China)     

三维声传播模型BELLHOP3D的信息传递接口并行优化

近些年,我国对海洋不断深入的探索对复杂环境中声场的快速预报提出了越来越高的需求。BELLHOP3D是一种基于射线法的三维声传播计算模型,在海洋声学中应用十分广泛。BELLHOP3D的计算效率比其他常用模型高,但是仍然有非常大的提升空间。该文使用信息传递接口对BELLHOP3D进行粗粒度的并行优化,并行后的程序计算结果稳定可靠,并行效率高,更适合在实际应用中实现快速的声场预报。并行BELLHOP3D程序可以在https://github.com/nj-zyq/BELLHOP3D＿MPI.git下载。     

刚体与SPH耦合的三维数值方法研究

 研究了刚性体与SPH（Smoothed Particle Hydrodynamics）方法相耦合的计算方法和技术,以此技术为核心,在原有三维综合拉氏程序CL3D的基础上,研制了适用于斜侵彻问题数值研究的刚性弹体与SPH靶体耦合计算的拉氏三维数值模拟程序CRSPH3D。针对金属材料编入了Johnson-Cook本构模型计算模块,针对混凝土材料编入了Johnson-Holmquist本构模型及与之配套的损伤模型和状态方程等的计算模块。用研制的计算程序对弹体分别正碰和斜碰穿透中厚铝靶和侵彻厚混凝土靶的过程进行了数值模拟计算和分析。结果表明,对于铝靶穿透问题,模拟计算得到的弹体剩余速度、弹体动能损失和贯穿图像与实验结果符合得较好。对于混凝土靶的侵彻问题也得到了与实验现象一致的、定性合理的图像。     

3维全电磁粒子模拟大规模并行程序NEPTUNE

 介绍了自主编制的3维全电磁粒子模拟大规模并行程序NEPTUNE的基本情况。该程序具备对多种典型高功率微波源器件的3维模拟能力,可以在数百乃至上千个CPU上稳定运行。该程序使用时域有限差分(FDTD)方法更新计算电磁场,采用Buneman-Boris算法更新粒子运动状态,运用质点网格法（PIC）处理粒子与电磁场的耦合关系,最后利用Boris方法求解泊松方程对电场散度进行修正,以确保计算精度。该程序初步具备复杂几何结构建模能力,可以对典型高功率微波器件中常见的一些复杂结构,如任意边界形状的轴对称几何体、正交投影面几何体,慢波结构、耦合孔洞、金属线和曲面薄膜等进行几何建模。该程序将理想导体边界、外加波边界、粒子发射与吸收边界及完全匹配层边界等物理边界应用于几何边界上,实现了数值计算的封闭求解。最后以算例的形式,介绍了使用NEPTUNE程序对磁绝缘线振荡器、相对论返波管、虚阴极振荡器及相对论速调管等典型高功率微波源器件进行的模拟计算情况,验证了模拟计算结果的可靠性,同时给出了并行效率的分布情况。     

基于JASMIN框架多物理耦合程序的性能优化及分析

基于并行应用支撑软件框架JASMIN的辐射流体与粒子输运耦合程序RHSn2D,采用最小邦元固定处理器数目的并行策略,计算实际模型的并行规模扩展至8192核,并行效率约为16%.集成程序时间分析,验证软件框架底层MPI并行环境聚合通信对于并行优化算法(尤其是辐射流体计算时间)的影响.     

抛物方程的多重非均匀网格模型及其应用

提出了抛物方程的多重非均匀网格模型,以准确求解三维空间存在多辐射源的电波传播问题。通过对不同辐射源建立不同的坐标系,并对其仿真空间采用不同的非均匀网格划分,构建了抛物方程的多重非均匀网格模型。在此基础上,实现了三维多辐射源问题的并行计算。实例仿真了空间存在四个辐射源的电波传播特性。结果表明,抛物方程的多重非均匀网格模型能够准确求解多源的空间电磁场分布特性,且在该算例中,并行技术使得抛物方程的计算速度提升了2.41倍,极大地提高了抛物方程对三维多源问题的求解效率。     

Parallel computation and numerical visualization for non-uniform particles included in gas and liquid flows

A parallel computation method has been proposed for the mixing and segregation of granular mixture included in gas and liquid flows. In this method, a three-dimensional (3D) computational volume is decomposed into multiple sub-blocks and their geometries are represented by 3D body-fitted coordinates. The fluid-particle interactions are treated by two types of models: a two-way model for liquid-solid flows and a one-way model in case of gas-solid flows. The computations of the particle motions in the multiple sub-blocks are executed simultaneously on the basis of the distinct element method (DEM). Since a graphic process is also executed as one of the parallel jobs, the particle distributions can be visualized during the computations. The computational method was applied to the gas-solid flows consisting of different diameters and densities in the horizontal and inclined cylinders rotating around their axes. From the comparison with the experimental results, nearly uniform mixing and particle segregation are successfully predicted in the oscillating liquid flows. In addition, it has been indicated that the particle pathline is very effective to visualize and understand the flow patterns of the particles with different properties. The result of the computations for the liquid-solid flows demonstrated that the vertical segregation of the non-uniform particles is reasonably reproduced.     

运动物体在颗粒物质中的动力学过程及最大穿透深度仿真研究

利用离散元法仿真了运动物体在颗粒物质中的三维动力学过程, 仿真采用周期边界条件, 并考虑了重力、接触力、阻尼力、摩擦力的影响. 将仿真结果和相关的三维实验结果进行了对比, 两者符合较好. 仿真结果表明穿透深度与运动物体的冲击速度、运动物体质量、颗粒介质床的密度均有关系. 运动物体质量越大, 速度越快, 则穿透越深, 而且穿透深度和质量呈线性关系. 仿真过程较为真实地再现了小颗粒的飞溅现象. 关键词： 颗粒物质 动力学过程 仿真 离散元法     

Parallel implementation of the FETI-DPEM algorithm for general 3D EM simulations

A parallel implementation of the electromagnetic dual-primal finite element tearing and interconnecting algorithm (FETI-DPEM) is designed for general three-dimensional (3D) electromagnetic large-scale simulations. As a domain decomposition implementation of the finite element method, the FETI-DPEM algorithm provides fully decoupled subdomain problems and an excellent numerical scalability, and thus is well suited for parallel computation. The parallel implementation of the FETI-DPEM algorithm on a distributed-memory system using the message passing interface (MPI) is discussed in detail along with a few practical guidelines obtained from numerical experiments. Numerical examples are provided to demonstrate the efficiency of the parallel implementation.     

高速碰撞数值计算中的光滑粒子法

扼要讨论了改进的光滑粒子法的离散思想,给出了二维轴对称问题中连续介质力学守恒方程的离散过程及离散格式,提出了轴对称坐标下确定影响域内粒子数的方法.最后通过高速碰撞的系列算例说明,光滑粒子法不但适宜于计算大变形冲击力学问题,而且有着其它网格法所无法替代的优势.     

Improved l1-SPIRiT using 3D walsh transform-based sparsity basis

l1-SPIRiT is a fast magnetic resonance imaging (MRI) method which combines parallel imaging (PI) with compressed sensing (CS) by performing a joint l1-norm and l2-norm optimization procedure. The original l1-SPIRiT method uses two-dimensional (2D) Wavelet transform to exploit the intra-coil data redundancies and a joint sparsity model to exploit the inter-coil data redundancies. In this work, we propose to stack all the coil images into a three-dimensional (3D) matrix, and then a novel 3D Walsh transform-based sparsity basis is applied to simultaneously reduce the intra-coil and inter-coil data redundancies. Both the 2D Wavelet transform-based and the proposed 3D Walsh transform-based sparsity bases were investigated in the l1-SPIRiT method. The experimental results show that the proposed 3D Walsh transform-based l1-SPIRiT method outperformed the original l1-SPIRiT in terms of image quality and computational efficiency.     

A 3D image interleaving method for real-time autostereoscopic display based on GPU

Multiview autostereoscopic systems display three-dimensional (3D) images interleaved from parallax images to provide adequate viewing zones for viewers. A 3D image interleaving method based on parallel processing with Graphic Processing Unit (GPU) is proposed to accelerate the image interleaving for real-time autostereoscopic displays. The proposed method is implemented with Direct3D Application Programming Interface (API) library. Masks used to interleave the 3D images are generated according to some particular principles. Both multiple texture blending and alpha blending operations are used to interleave a 3D image. The experiments prove that high interleaving efficiency and good stereoscopic sense can be obtained. The proposed method meets the requirement of real-time multiview autostereoscopic displays and avoids extra shader program.     

改进的3维粒子模拟并行算法

 提出了两种改进的3维粒子模拟并行算法,改进的并行算法能在每个时间步减少一次进程同步。算法分析和数值模拟表明,由于粒子运动路径和发射的初始位置与随机函数有关,只有一种改进的并行算法能保证并行计算正确。在3维粒子模拟软件CHIPIC3D上实现了改进的并行算法,应用CHIPIC3D对一种相对论返波管进行了并行模拟,模拟结果表明改进的并行算法能取得更高的加速比和效率。     

Three-dimensional finite-element simulation of Zener pinning dynamics

The Zener pinning dynamics of a moving boundary interacting with one or more particles is described by a three-dimensional (3D) finite-element model. The model, based upon a variational formulation for boundary motion by viscous drag, is solved by a finite-element method to obtain the velocity at each node of triangular linear elements on the grain boundary. It is first applied to relatively simple and validated cases, for which analytical and numerical results are available. These cases correspond to an axisymmetrical geometry, in which the grain boundary interacts with a centred particle. A simple analytical pinning criterion is derived from these simulations. The model is then applied to general 3D cases, in which the grain boundary interacts with arbitrarily localized and sized particles. The aim of these 3D simulations is to quantify the influence of the position and the number of particles on the average grain-boundary velocity. It is shown, for example, that the drag effect is enhanced when the particle, or the cluster of particles, is off-centre and that pinning is less efficient with several particles than with a single particle producing the same Zener force.     

Measurement of temperature of a dusty plasma from its configuration

A new concept called “configurational temperature” is introduced in the context of dusty plasma, where the temperature of the dust particles submerged in the plasma can be measured directly from the positional information of the individual dust particles and the interaction potential between the dust grains. This method does not require the velocity information of individual particles, which is a key parameter to measure the dust temperature in the conventional method. The technique is initially tested using two-dimensional (2D) OpenMP parallel molecular dynamics and Monte Carlo simulation and then compared with the temperature evaluated from experimental data. The experiment have been carried out in the Dusty Plasma Experimental (DPEx) device, where a 2D stationary plasma crystal of melamine formaldehyde particles is formed in the cathode sheath of a DC glow discharge argon plasma. The kinetic temperature of the dust is calculated using the standard particle image velocimetry technique at different pressures. An extended simulation result for the three-dimensional case is also presented, which can be employed for the temperature measurement of a three-dimensional dust crystal in laboratory devices.