太阳风系统仿真与关键技术研究
|
摘要
太阳风是从太阳日冕层经过加速射出，以200-800km/s运动的带电粒子流。快速变化的太阳风严重影响着地球的空间环境，如地磁暴、电离层暴等。这些现象严重干扰无线通信、地面相关设施等。在人类航天活动方面，太阳风对航天航空探索有着显著的影响。为了探索太阳风对空间环境的影响；对卫星、无线通信产生的干扰；对地面电力网、管道和其它大型结构的破坏，本课题通过对太阳风形成、运动、传输等方面的研究，建立太阳风数学物理模型，进行太阳风仿真模拟实验，发现太阳风的活动与运动规律。通过分析和探索太阳风对卫星通讯系统、地面相关设施等产生的各种影响因素，以期减少太阳风对地球和人类带来的灾难，建立相关的预警和监测系统。
太阳风系统仿真是一个涉及大气物理学、地球探测与信息技术、天文学、仿真学、高性能计算、数据存储技术、计算机网络技术、高级程序设计等多个领域的交叉前沿课题。本课题首先研究太阳风、现代仿真技术、Enzo宇宙进化仿真等相关理论。其次，建立数学物理模型和仿真模型。第三，构建太阳风系统仿真体系。最后，探索与研究太阳风的活动与运动规律，建立相关的预警和监测系统。
本课题建立了太阳风粒子系统模型，解决了太阳风粒子系统仿真驱动；构建太阳风系统仿真体系架构，分析研究了太阳风系统仿真关键技术；建立了太阳风粒子系统和地磁系统仿真器；首次提出了基于CPU+GPU异构计算分布式云仿真器设计理念；建立了仿真的任务调度模型；建立了太阳风和太阳高能粒子第三级的科学数据数据映射模型。
太阳风粒子系统模型是太阳风系统仿真需要解决的关键技术之一，根据太阳风固有的属性，建立适用于本系统仿真太阳风粒子系统驱动方程数学物理模型，是需要研究的首要内容。通过研究太阳风的质子、带电、气态等特性，基于现有的研究成果，采用数学建模的方式，为太阳风粒子系统建立了基于引力场、电磁场、理想气体等理论的数学物理方程，驱动太阳粒子的机械、电磁和分子热运动，是本课题的研究内容之一。该模型能够有效地解决太阳风粒子系统仿真器的引擎驱动问题，仿真实验结果表明，该模型能够较好地满足数据精度要求。
太阳风系统仿真需要借助计算机系统软、硬件来完成，构建太阳风系统仿真体系，是完成太阳风系统仿真需要解决的另一关键问题。太阳风系统仿真与工业控制的仿真有所不同，简单通过MATLAB/Simulink进行实现不太适合，其仿真与Enzo宇宙进化仿真有所类似。为了完成太阳风系统仿真的研发，通过研究现代计算机仿真技术及其所涉及的高性能计算、超大规模数据存储和仿真终端的异构计算等，首次提出并建立了太阳风仿真体系所需的仿真系统架构。系统架构所涉及的高性能计算采用了基于网络通讯的远程计算云，存储采用了基于网络链路的云存储等模型设计，为研发该系统提供了基础框架，其涉及的硬件异构现象，通过软件设计得到了有效的控制，其松耦合的设计理念为项目研发进度控制提供了异步开发策略，较好地解决了项目进度差异带来的项目开发时延，对其他类似项目的研发提供了项目级控制参考策略。
在仿真器的设计中，为了减轻仿真器的压力和超大规模终端仿真的需求，采用分布式渲染技术，利用CPU+GPU异构计算技术，首次提出并研发了基于RIA的云仿真客户端，分别实现了太阳风粒子系统和大地电磁云客户端仿真器。该仿真器能够提供在单台计算机上完成多仿真器同时运行的实时仿真，也提供分离在各个节点独立仿真的运行模式，该方法具有一定的独创性，为基于CPU+GPU异构计算分布式云仿真器的设计提供了一定的参考。
在太阳风系统仿真中，为了解决在业务量增长时，能够充分地利用计算系统资源，采用将系统任务进行分解的方法，有机地将任务分配到各个计算单元进行计算，因此研究太阳风系统仿真中的任务调度模型是本课题的内容之一。在研究异构环境下太阳风仿真任务调度NP问题研究中，利用仿真任务可分解的特性，在现有成果和理论上通过修正和更新任务动态分解调度模型，解决了高性能计算仿真中任务的调度问题。通过对任务调度模型优化和将子任务分解为2层m叉树，引入复杂度计算能力比，较好地解决了仿真中负载平衡和资源优化问题，为子任务可分解异构环境任务调度问题的研究提供了参考。
为了能够快速获得第三级太阳风粒子可分析处理的结构化数据，采用将数据映射到内存和语言集成查询技术，以解决数据的完整映射和快速访问。首次提出并设计数据映射模型，将采用PDS数据标签的数据产品分别映射到模型的值域和属性域，解决本地或网络数据内存映射问题。采用双倍缓冲和贝塞尔插值技术对数据能谱进行了实时绘制。结果表明该模型具有高效性、完备性和高吻合度。该模型能够较好地解决异构数据快速结构化数据访问时数据源动力不足问题，为处理和分析太阳高能粒子的频谱、成份和通量及随时间、空间变化的分布特征等提供了基础。
The solar wind is a stream of charged particles which is ejected from the solarcorona after acceleration and moves at the speed of200km to800km per second. Therapid changing solar wind has a serious effect on the earth’s space environment. Suchphenomena as geomagnetic storm, ionosphere storm, etc. gravely obstruct wirelesscommunication and relevant ground facilities. In human aerospace activities, the solarwind has a significant influence on the aerospace exploration. In this thesis, the authorexplores what effects the solar wind has on space environment, how it disturbs thesatellite and wireless communication and how it destroys the ground power network,pipelines and other large structures. On the basis of research conducted on its formation,movement, transmission and some other aspects, the physical and mathematical modelof the solar wind is established to performan simulation experiments for the purpose offinding out the activity and motion law. Through analysis and exploration of theinfluences of the solar wind on the satellite communication system, relevant groundfacilities etc., a early warning and monitoring system is established so as to reduce thepossible disasters brought by the solar wind to the earth and human beings.
Simulation of the solar wind is a frontier cross subject dealing with a wide range offields such as atmospheric physics, geodetection and information technology,astronomy, simulation, high performance computing, data storage technology,computer network technology, advanced programming, etc. In this thesis, research isfirstly conducted into some theories relating to the solar wind, modern simulationtechnology and Enzo evolution of the universe simulation. Secondly, the physical andmathematical model of the solar wind and the simulation model are establishedrespectively. Thirdly, the solar wind simulation system is constructed. And in the end,the activity and motion law of the solar wind is explored and the corresponding earlywarning and monitoring system is established.
In this paper, a solar wind particle system model is established to ensure simulation driving; the solar wind particle simulation system is constructed and the keytechnology is analyzed and discussed; a solar wind particles system simulator andgeomagnetic system simulator are created; the distributed cloud simulator designconcept, based on heterogeneous computing of CPU+GPU is put forward; thesimulation scheduling models are built up; the solar wind and solar energetic particledata mapping model(the third level) is formed.
The solar wind particles system model is one of the key techniques to be solved inthe solar wind system simulation. Establishing a physical and mathematical modelapplicable to the system simulation of solar wind particles system driving equationaccording to the inherent properties of the solar wind, becomes a task of the firstimportance. Through the study of its properties such as the protons, being charged, gasand other characteristics, based on the existing research results, the physical andmathematical equation founded on the theories of electromagnetic field, gravitationalfield and ideal gas is worked out to probe into the mechanical, electromagnetic andmolecular thermal motion of solar particles by means of mathematical modeling. Thismodel can effectively solve the drive engine problem of the solar wind systemsimulator. The simulation experiment result shows that the model meets therequirement of data accuracy.
The solar wind simulation system will be established with the help of computersoftware and hardware, but how to establish it is another key problem to be resolved.The solar wind simulation, which is different from the industry-controlled ones, can'tbe simply realized by MATLAB/Simulink. It is as similar as Enzo cosmic evolutionsimulation. And for this purpose, the author studies modern computer simulationtechnology, such as high performance computing, large scale data storage andheterogeneous computing of terminal emulation and so on, and firstly puts forwardsimulation system framework that can be used in solar wind simulation system. Exactlyspeaking, the simulation system framework consists of four parts: High performancecomputing employs the remote computing cloud based on network communication; thestorage providing a basic frame for the research and development of simulation systemuses cloud storage model design based on a network; the hardware isomerism iscontrolled effectively by the software; and the loosely coupled design concept providesasynchronous development strategy for the progress control of project research anddevelopment, it not only solves the problem of project development delay due to theproject schedule difference, but also provides us project level control strategies for theresearch and development of the other similar projects.
In order to alleviate the simulator and the demand for ultra-large-scale terminalemulation in simulator-designing, the author puts forward and develops cloudsimulation client based on RIA using distributed rendering technology and CPU+GPUheterogeneous computing technology. According to this, the solar wind particlessystem and the magnetotelluric cloud client emulator are established respectively. Theemulator enables many emulators to run simultaneously on a single computer, but alsoprovide independent simulation running mode separated in each node. This originalresearch is a reference for the design of distributed cloud emulator based on CPU+GPUheterogeneous computing.
Another important content of this thesis is the development of task schedulingmodel. This model is introduced to deal with an increasing number of tasks by makingfull use of computing resources. That is, the tasks will be discomposed to eachcomputing unit. As is known, the simulation tasks can be discomposed. In the study ofNP, a problem of the solar wind task scheduling under the heterogeneous environment,efforts are made to solve the scheduling problem in high performance computingsimulation tasks by revising and updating the dynamic task decomposition andscheduling model according to the current findings and theories. In this system, thetasks are discomposed and resources are optimized through optimizing the schedulingmodel, dividing the sub tasks into2layers of mass and introducing the complexitycalculating power ratio, thus providing a reference for study of this field.
In order to rapidly obtain disposable and structured data of the third grade solarwind particles, data are mapped in the memory and integrated language query so thatcomplete mapping and instant access are achievable. In this thesis, the author for thefirst time advances and designs the data mapping model. In this model data tagged withPDS are mapped to the range and attribute domain of the model to solve the local ornetwork data memory mapping problem. Besides, double buffer and Besselinterpolation technology are applied to draw a real-time data spectrum. The resultsshow that the model is of high efficiency, completeness and conformity. This model canbe used to solve the source power shortage problem of data when the heterogeneousdata and fast structured data are accessed. This research provides a sound basis forprocessing and analyzing the spectrum, composition and flux of high-energy solarparticles and their distribution characteristics which vary with time and space.
太阳风系统仿真是一个涉及大气物理学、地球探测与信息技术、天文学、仿真学、高性能计算、数据存储技术、计算机网络技术、高级程序设计等多个领域的交叉前沿课题。本课题首先研究太阳风、现代仿真技术、Enzo宇宙进化仿真等相关理论。其次，建立数学物理模型和仿真模型。第三，构建太阳风系统仿真体系。最后，探索与研究太阳风的活动与运动规律，建立相关的预警和监测系统。
本课题建立了太阳风粒子系统模型，解决了太阳风粒子系统仿真驱动；构建太阳风系统仿真体系架构，分析研究了太阳风系统仿真关键技术；建立了太阳风粒子系统和地磁系统仿真器；首次提出了基于CPU+GPU异构计算分布式云仿真器设计理念；建立了仿真的任务调度模型；建立了太阳风和太阳高能粒子第三级的科学数据数据映射模型。
太阳风粒子系统模型是太阳风系统仿真需要解决的关键技术之一，根据太阳风固有的属性，建立适用于本系统仿真太阳风粒子系统驱动方程数学物理模型，是需要研究的首要内容。通过研究太阳风的质子、带电、气态等特性，基于现有的研究成果，采用数学建模的方式，为太阳风粒子系统建立了基于引力场、电磁场、理想气体等理论的数学物理方程，驱动太阳粒子的机械、电磁和分子热运动，是本课题的研究内容之一。该模型能够有效地解决太阳风粒子系统仿真器的引擎驱动问题，仿真实验结果表明，该模型能够较好地满足数据精度要求。
太阳风系统仿真需要借助计算机系统软、硬件来完成，构建太阳风系统仿真体系，是完成太阳风系统仿真需要解决的另一关键问题。太阳风系统仿真与工业控制的仿真有所不同，简单通过MATLAB/Simulink进行实现不太适合，其仿真与Enzo宇宙进化仿真有所类似。为了完成太阳风系统仿真的研发，通过研究现代计算机仿真技术及其所涉及的高性能计算、超大规模数据存储和仿真终端的异构计算等，首次提出并建立了太阳风仿真体系所需的仿真系统架构。系统架构所涉及的高性能计算采用了基于网络通讯的远程计算云，存储采用了基于网络链路的云存储等模型设计，为研发该系统提供了基础框架，其涉及的硬件异构现象，通过软件设计得到了有效的控制，其松耦合的设计理念为项目研发进度控制提供了异步开发策略，较好地解决了项目进度差异带来的项目开发时延，对其他类似项目的研发提供了项目级控制参考策略。
在仿真器的设计中，为了减轻仿真器的压力和超大规模终端仿真的需求，采用分布式渲染技术，利用CPU+GPU异构计算技术，首次提出并研发了基于RIA的云仿真客户端，分别实现了太阳风粒子系统和大地电磁云客户端仿真器。该仿真器能够提供在单台计算机上完成多仿真器同时运行的实时仿真，也提供分离在各个节点独立仿真的运行模式，该方法具有一定的独创性，为基于CPU+GPU异构计算分布式云仿真器的设计提供了一定的参考。
在太阳风系统仿真中，为了解决在业务量增长时，能够充分地利用计算系统资源，采用将系统任务进行分解的方法，有机地将任务分配到各个计算单元进行计算，因此研究太阳风系统仿真中的任务调度模型是本课题的内容之一。在研究异构环境下太阳风仿真任务调度NP问题研究中，利用仿真任务可分解的特性，在现有成果和理论上通过修正和更新任务动态分解调度模型，解决了高性能计算仿真中任务的调度问题。通过对任务调度模型优化和将子任务分解为2层m叉树，引入复杂度计算能力比，较好地解决了仿真中负载平衡和资源优化问题，为子任务可分解异构环境任务调度问题的研究提供了参考。
为了能够快速获得第三级太阳风粒子可分析处理的结构化数据，采用将数据映射到内存和语言集成查询技术，以解决数据的完整映射和快速访问。首次提出并设计数据映射模型，将采用PDS数据标签的数据产品分别映射到模型的值域和属性域，解决本地或网络数据内存映射问题。采用双倍缓冲和贝塞尔插值技术对数据能谱进行了实时绘制。结果表明该模型具有高效性、完备性和高吻合度。该模型能够较好地解决异构数据快速结构化数据访问时数据源动力不足问题，为处理和分析太阳高能粒子的频谱、成份和通量及随时间、空间变化的分布特征等提供了基础。
The solar wind is a stream of charged particles which is ejected from the solarcorona after acceleration and moves at the speed of200km to800km per second. Therapid changing solar wind has a serious effect on the earth’s space environment. Suchphenomena as geomagnetic storm, ionosphere storm, etc. gravely obstruct wirelesscommunication and relevant ground facilities. In human aerospace activities, the solarwind has a significant influence on the aerospace exploration. In this thesis, the authorexplores what effects the solar wind has on space environment, how it disturbs thesatellite and wireless communication and how it destroys the ground power network,pipelines and other large structures. On the basis of research conducted on its formation,movement, transmission and some other aspects, the physical and mathematical modelof the solar wind is established to performan simulation experiments for the purpose offinding out the activity and motion law. Through analysis and exploration of theinfluences of the solar wind on the satellite communication system, relevant groundfacilities etc., a early warning and monitoring system is established so as to reduce thepossible disasters brought by the solar wind to the earth and human beings.
Simulation of the solar wind is a frontier cross subject dealing with a wide range offields such as atmospheric physics, geodetection and information technology,astronomy, simulation, high performance computing, data storage technology,computer network technology, advanced programming, etc. In this thesis, research isfirstly conducted into some theories relating to the solar wind, modern simulationtechnology and Enzo evolution of the universe simulation. Secondly, the physical andmathematical model of the solar wind and the simulation model are establishedrespectively. Thirdly, the solar wind simulation system is constructed. And in the end,the activity and motion law of the solar wind is explored and the corresponding earlywarning and monitoring system is established.
In this paper, a solar wind particle system model is established to ensure simulation driving; the solar wind particle simulation system is constructed and the keytechnology is analyzed and discussed; a solar wind particles system simulator andgeomagnetic system simulator are created; the distributed cloud simulator designconcept, based on heterogeneous computing of CPU+GPU is put forward; thesimulation scheduling models are built up; the solar wind and solar energetic particledata mapping model(the third level) is formed.
The solar wind particles system model is one of the key techniques to be solved inthe solar wind system simulation. Establishing a physical and mathematical modelapplicable to the system simulation of solar wind particles system driving equationaccording to the inherent properties of the solar wind, becomes a task of the firstimportance. Through the study of its properties such as the protons, being charged, gasand other characteristics, based on the existing research results, the physical andmathematical equation founded on the theories of electromagnetic field, gravitationalfield and ideal gas is worked out to probe into the mechanical, electromagnetic andmolecular thermal motion of solar particles by means of mathematical modeling. Thismodel can effectively solve the drive engine problem of the solar wind systemsimulator. The simulation experiment result shows that the model meets therequirement of data accuracy.
The solar wind simulation system will be established with the help of computersoftware and hardware, but how to establish it is another key problem to be resolved.The solar wind simulation, which is different from the industry-controlled ones, can'tbe simply realized by MATLAB/Simulink. It is as similar as Enzo cosmic evolutionsimulation. And for this purpose, the author studies modern computer simulationtechnology, such as high performance computing, large scale data storage andheterogeneous computing of terminal emulation and so on, and firstly puts forwardsimulation system framework that can be used in solar wind simulation system. Exactlyspeaking, the simulation system framework consists of four parts: High performancecomputing employs the remote computing cloud based on network communication; thestorage providing a basic frame for the research and development of simulation systemuses cloud storage model design based on a network; the hardware isomerism iscontrolled effectively by the software; and the loosely coupled design concept providesasynchronous development strategy for the progress control of project research anddevelopment, it not only solves the problem of project development delay due to theproject schedule difference, but also provides us project level control strategies for theresearch and development of the other similar projects.
In order to alleviate the simulator and the demand for ultra-large-scale terminalemulation in simulator-designing, the author puts forward and develops cloudsimulation client based on RIA using distributed rendering technology and CPU+GPUheterogeneous computing technology. According to this, the solar wind particlessystem and the magnetotelluric cloud client emulator are established respectively. Theemulator enables many emulators to run simultaneously on a single computer, but alsoprovide independent simulation running mode separated in each node. This originalresearch is a reference for the design of distributed cloud emulator based on CPU+GPUheterogeneous computing.
Another important content of this thesis is the development of task schedulingmodel. This model is introduced to deal with an increasing number of tasks by makingfull use of computing resources. That is, the tasks will be discomposed to eachcomputing unit. As is known, the simulation tasks can be discomposed. In the study ofNP, a problem of the solar wind task scheduling under the heterogeneous environment,efforts are made to solve the scheduling problem in high performance computingsimulation tasks by revising and updating the dynamic task decomposition andscheduling model according to the current findings and theories. In this system, thetasks are discomposed and resources are optimized through optimizing the schedulingmodel, dividing the sub tasks into2layers of mass and introducing the complexitycalculating power ratio, thus providing a reference for study of this field.
In order to rapidly obtain disposable and structured data of the third grade solarwind particles, data are mapped in the memory and integrated language query so thatcomplete mapping and instant access are achievable. In this thesis, the author for thefirst time advances and designs the data mapping model. In this model data tagged withPDS are mapped to the range and attribute domain of the model to solve the local ornetwork data memory mapping problem. Besides, double buffer and Besselinterpolation technology are applied to draw a real-time data spectrum. The resultsshow that the model is of high efficiency, completeness and conformity. This model canbe used to solve the source power shortage problem of data when the heterogeneousdata and fast structured data are accessed. This research provides a sound basis forprocessing and analyzing the spectrum, composition and flux of high-energy solarparticles and their distribution characteristics which vary with time and space.
引文
[1]能源知识.太阳风.能源与节能,2011,5:18.
[2] Cranmer S R et al. Spectroscopic constraints on models of ions-cyclotron resonance heating inthe polar solar corona and high-speed Solar wind. Astrophys.J.,1999,518-L937.
[3]束传政,王玲华,涂传诒.回旋波共振加热加速的多重离子流体模型.空间科学学报,2000,(03):201-207.
[4]冶萍,张靖周.求解N-S方程SIMPLE算法的半交错网格研究.南京航空航天大学学报.2004,36(1):11-15.
[5]阚圣哲,陈国兴,陈磊.基于Abaqus软件的并行计算集群平台构建与优化方法[J].防灾减灾工程学报,2009,29(06):644-651
[6]段孝国.分布式计算技术介绍[J].电脑知识与技术,2011.7(22):5463-5465.
[7] Behrouz A. Forouzan,Firouz Mosharraf. Foundations of Computer Science[M].2th ed.London:Gaynor Redvers-Mutton,2008:4-5.
[8]刘久星,孙永强. PRAM,BSP和LogP并行模型之间的关系及其比较.小型微型计算机系统,1999,20(11):824-827.
[9]罗达强.探析Windows Azure Platform微软云计算平台.硅谷,2010(6):9-10.
[10] Grama,M.,等.并行计算导论[M].张武,等,译.2版.北京:机械工业出版社,2004:1-6.
[11] David A. Bader.面向千万亿次计算[M].都志辉,等译.北京：清华大学出版社,2008:59-72.
[12] P. Hut. Dense stellar systems as laboratiories for fundamental physics. In ArXiv Astrophysicse-prints,2006.
[13] John D Evans, Robort R Kessler. A communication-ordered task graph allocation algorithm.IEEE transaction on Computers, March1981, C(3):207-214.
[14]ollweg, Joseph V., Heating and acceleration of the solar wind in coronal holes: Cyclotronresonances. Advances in Space Research, v30, n3, p469,2002.
[15]Li, Xing. Acceleration of the fast solar wind through minor ions. European Space Agency,(Special Publication) ESA SP, n547, p369-374.
[16]Marsch, Eckart. Status of knowledge after helios, ulysses and soho of the microstate of thecoronal and solar-wind plasma. European Space Agency,(Special Publication) ESA SP, n SP-641,October2006.
[17]Chouliaras.Customization of an embedded RISC CPU with SIMD extensions for videoencoding: A case study.Integration, the VLSI Journal, v41, n1, p135-152, January2008.
[18] Mohanty, Saraju P..GPU-CPU multi-core for real-time signal processing. Digest ofTechnical Papers-IEEE International Conference on Consumer Electronics,2009.
[19]Smelyanskiy, Mikhail (Intel Corporation); Holmes, David; Chhugani, Jatin; etc..Mappinghigh-fidelity volume rendering for medical imaging to CPU, GPU and many-core architectures.IEEE Transactions on Visualization and Computer Graphics, v15, n6, p1563-1570.
[20] Teodoro, George. Coordinating the use of GPU and CPU for improving performance ofcompute intensive applications. Proceedings-IEEE International Conference on ClusterComputing, ICCC,2009,2009IEEE International Conference on Cluster Computing andWorkshops, CLUSTER '09Language: English
[21]涂传诒.太阳风起源和形成的研究进展.中国科学A辑,2000,(S1),5-8.
[22] P. Hut. Dense stellar systems as laboratiories for fundamental physics. In ArXiv Astrophysicse-prints,2006.
[23]石勇,魏奉思,冯学尚等.1997年1月事件行星际扰动在三维背景太阳风中传播的数值模拟.地球物理学报,2001,(03),303-310。
[24]敖先志,沈迹,涂传诒.太阳风质子各向异性速度分布的产生机制.中国科学A辑,2002,(10),940-944.
[25]刘小龙,赵华,C.T.Russell等.上游太阳风中高能电子向同步轨道区的传输.空间科学学报,2003,(03),174-180.
[26]薛金涛,王敬芳,钱昕等.单流体二维太阳风结构的磁流体力学模拟.空间科学学报,2003,(01),1-6.
[27]叶占银,魏奉思,王赤.两冕流间CME事件数值模拟的改进.空间科学学报,2004,(01),1-7.
[28]冯学尚,向长青,钟鼎坤等.三维太阳风结构的Ulysses观测和MHD模拟的比较研究.科学通报,2005,(08),820-826.
[29]彭忠.太阳风-磁层-电离层耦合的全球MHD数组模拟研究[D].合肥:中国科学技术大学,2009:2-14.
[30]许丽华,刘淼. MPI并行编程环境的研究.现代计算机,2003,(04):28-31.
[31]崔雪冰,张延红,李国徽.基于通用计算的GPU-CPU协作计算模式研究微.电子学与计算机,2009,(08):30-33.
[32]王栋栋,庄雷. CPU与GPU并行计算的火焰模拟.计算机应用,2009,(06):1702-1706,1710
[1] Horne, Richard B., Wave acceleration of electrons in the Van Allen radiation belts. NaturePublishing Group,2005,437:227-230.
[2]吴锤结.太阳“发射”高速风影响地球气候和导航系统[J].大连:Space Travel,2009(24):105.
[3]余英.太阳活动对卫星通信的影响[J].广播与电视技术,2000,(12):120-123.
[4]涂传诒.太阳风源于日冕的漏斗状磁场结构[J].中国基础科学,2006,(02).
[5]欧阳自远等.绕月探测工程科学应用研究进展[J]中国科学院院刊，2009，24（5）：530-536.
[6]刘国栋，陈乐寿。大地电磁测深研究[M].地震出版社，1984（06）.
[7] MarschEetal.Wave heating and acceleration of solar wind by cyclotron resonance.J.Geophys.Res.,1982,87:5030-5036.
[8] HollwegJV, etal. Possibleevidencefor coronalAlfvenwaves.J.Geophys. Res.,1982,57:1
[9] Isenberg PA.Resonantaccelerationandheatingofsolarwindions:Anisotropyanddispersion.JGeophys.Res.,1984,89:6613-6621.
[10]吴其芬,李桦著.磁流体动力学[M].北京:国防科技大学出版社，2007.
[11] J. W. Dungey.Interplanetary Magnetic Field and the Auroral Zones.Phys. Rev.1961:6,47–48.
[12] Axford, W.L., and C. O. Hines, A unifying theory of high-latitude geophysical phenomenaand geomagnetic storms, Can. J. phys.,39,1961:1433.
[13] Ming Xue, CAPS/Univ. of Oklahoma, Norman, OK; and F. Kong, D. Weber, K. W. Thomas,Y. Wang, K. Brewster, K. K. Droegemeier, J. S. Kain, S. J. Weiss, D. R. Bright, M. S. Wandishin,M. C. Coniglio, and J. Du. CAPS Realtime Storm-scale Ensemble and High-resolution Forecastsas Part of the NOAA Hazardous Weather Testbed2007Spring Experiment.22nd Conference onWeather Analysis and Forecasting/18th Conference on Numerical Weather Prediction.
[14]张子平,马庆林.帕克与太阳风的发现.天文爱好者.2005(2):8-10.
[15]束传政,王玲华,涂传诒.回旋波共振加热加速的多重离子流体模型.空间科学学报,2000,(03):201-207.
[16]彭忠,陈耀.流管几何对低速太阳风alpha粒子丰度的影响.中国科学技术大学学报,2007,(08):852-858.
[17]敖先志,沈迹,涂传诒.太阳风质子各向异性速度分布的产生机制.中国科学A辑,2002,(10),940-944.
[18] David A. Bader.面向千万亿次计算[M].都志辉,等,译.北京：清华大学出版社,2008:59-72.
[19] P. Hut. Dense stellar systems as laboratiories for fundamental physics. In ArXiv Astrophysicse-prints,2006.
[20] John D Evans, Robort R Kessler. A communication-ordered task graph allocation algorithm.IEEE transaction on Computers, March1981, C(3):207-214.
[21]崔雪冰,张延红,李国徽.基于通用计算的GPU-CPU协作计算模式研究微.电子学与计算机,2009,(08),30-33.
[22]王栋栋,庄雷. CPU与GPU并行计算的火焰模拟.计算机应用,2009,(06),1702-1706+1710.
[23]周建,刘鹏,梅优良等.基于微处理器核的AAC解码软硬件协同优化.浙江大学学报(工学版),2007,(08),1271-1276.
[24]许丽华,刘淼. MPI并行编程环境的研究.现代计算机,2003,(04),28-31.
[25]冯学尚,向长青,钟鼎坤等.三维太阳风结构的Ulysses观测和MHD模拟的比较研究.科学通报,2005,(08),820-826.
[26]刘小龙,赵华,C.T.Russell等.上游太阳风中高能电子向同步轨道区的传输.空间科学学报,2003,(03),174-180.
[27]石勇,魏奉思,冯学尚等.1997年1月事件行星际扰动在三维背景太阳风中传播的数值模拟.地球物理学报,2001,(03),303-310.
[1]冯恒强，吴德金，赵寄昆.通过建立自然坐标系确定磁云边界方法的探讨[J].天文学报，2006，47(4):371-386.
[2]徐龙道.物理学词典[M].1版.北京:科学出版社,2007:9-9.
[3]钟江帆.大学物理(上册)[M].北京:高等教育出版社,2005:77-78.
[4]解恩泽.爱因斯坦统一场论思想的命运[J/OL].中国青年科技,1999,3:28-30.
[5]何圣静.物理定律的形成与发展[M].北京:测绘出版,1988:55-81.
[6]复旦大学《物理学》编写组.物理学.1版.北京:高等教育出版社,1985:374-377.
[7]张三慧.大学基础物理学(上册)[M].1版.北京:清华大学出版社,2003:206-207.
[8]王晓鸥.物理学概论[M].1版.上海:同济大学出版社,2007:103.
[9]卢新平.简明普通物理学[M].1版.上海:同济大学出版社,2007:131.
[10] B.M.亚沃尔斯基(Яворский,В.М.), A.A.杰特拉无(Детлаф,А.А.).物理学手册[M].宓鼎樑,等,译.1版.北京:科学出版社,1986:146-147.
[11]张桂兰，聂东壁山，夏玉珍.对理想气体分子按平均动能分布的讨论[J].临沂师范学院学报，2003，25(6):32-34.
[12]徐劳立.关于能量均分定理的讨论[J].物理与工程，2005,15(6):61-62.
[13]西南中南九所高等院校.大学物理学[M].成都:四川科学技术出版社,1987:385-391.
[14]钟江帆.大学物理(下册)[M].高等教育出版社.2005:253-258.
[15]马根源,王松立,金庆华.物理学[M].1版.天津:南开大学出版社,1993:448-463.
[16]程守洙,江之永.普通物理学[M].5版.北京:高等教育出版社,1997:295-298.
[17] Gavryuseva, E., Godoli, G., Structure and rotation of the large scale solar magnetic fieldobserved at the Wilcox Solar Observatory. Elsevier Ltd,2006(31):68-76.
[18]徐兵.地球和太阳的磁场从哪里来[J].大科技科学之谜,2009,10:12-13.
[19]林元章.太阳活动起源研究(Ⅱ):太阳发电机理论[J].天文学进展,2002,(01)23:32.
[20]陈宏芳.原子物理学[M].1版.北京:科学出版社,2006:1-32.
[21]王永昌.近代物理学[M].1版.北京:高等教育出版社,2006:169-170.
[22]杨桂林,江兴方,柯善哲.近代物理.1版.北京:科学出版社,2006:22-23.
[23]陈强顺，冯承天.电荷守恒定律可确定麦克斯韦方程组的数学形式[J].同济大学学报，1989,17(3):395-400.
[24] P.J.OUSEPH.技术物理学[M].黄伟民,褚孟似译.北京：高等教育出版社,1992:597-598.
[25]陈丹.太阳的日冕观测和磁场观测[J].太空探索,2010(6):62-65.
[26] WengYing, WengZi-Hua. Heavy ions acceleration of solar wind in electromagnetic andgravitational fields. Progress in Electromagnetics Research Symposium, PIERS2011Marrakesh,2011:1780-1784.
[1]杨立,齐建玲,曲风娟等.计算机控制与仿真技术[M].1版.北京:中国水利水电出版社,2006:14.
[2]王正林,王胜开,陈国顺等.MATLAB/Simulink与控制系统仿真[M].2版.北京:电子工业出版社,2008.7:7.
[3]雷永林.仿真模型重用理论、方法与异构集成技术研究[D].长沙:国防科学技术大学,2006.
[4]马元元,孙志挥,张南翔.增量关联规则在大型火力发电厂实时控制中的应用[J].工业控制计算机,2000,(01)14-15.
[5]高薇,张强,孙军.通过超实时仿真验证轨控策略方法研究[J].航天控制,2008,26(2):71-73,78.
[6]林源.一种基于虚拟仪器的局域网导弹测控仿真培训系统[J].软件导刊,2009(03):79-81.
[7]许喆.某外场水下试验平台的视景仿真研究[D].西安:西北工业大学,2007:3.
[8]詹剑锋.挑战下一个大家伙写在2011超级计算大会之后[J].程序员2012(1):14-15.
[9] Alpert, D.,Avnon, D..Architecture of the Pentium microprocessor.Micro, IEEE,2002,13:11–21.
[10] Behrouz A.Forouzan, Firouz Mosharraf.计算机科学导论[M].刘艺,瞿高峰,译.2版.北京:机械工业出版社,2009:83.
[11] Katz, R.Reduced instruction set computers[C]. Solid-State Circuits Conference. Digest ofTechnical Papers.1986:218-219.
[12] Schlansker, M.S.,Rau, B.R., Hewlett-Packard Labs.EPIC: Explicitly Parallel InstructionComputing[J]. Computer,2000(33):37-45.
[13] Eggers, S.J.,Emer, J.S.,Leby, H.M., et al. Simultaneous multithreading: a platform fornext-generation processors. Micro, IEEE,1997(17):12-19.
[14] Nayfeh, B.A., Olukotun, K.,Stanford Univ., CA. A single-chip multiprocessor[J].Computer,1997(30):79-85.
[15]武剑锋,朱家维.对称多处理机系统[J].微型机与应用,1992(9):21-23.
[16] Haas, Peter W., Resch, Michael M., High performance computing in science andengineering[C].17th International Symposium on Computer Architecture and High PerformanceComputing, SBAC-PAD2005:10-18.
[17] Chen, Guo-Liang, Sun, Guang-Zhong, Zhang, Yun-Quan, et al.Study on parallel computing.Journal of Computer Science and Technology.Springer New York LLC,2006:665-673.
[18] AdlerM.,GibbonsP.B., MatiasY.; RamachandranV., Modeling Parallel Bandwidth: Localversus Global Restrictions. Springer-Verlag New York Inc,1999:381-404.
[19] Boghosian, Bruce M., Coveney, Peter V.Introduction of grid computing application projectsat the NASA earth science technology office[C]. IEEE Computer Society,2006:289-298.
[20] Kalagiakos, Panagiotis, Karampelas, Panagiotis. Cloud computing learning[C].20115thInternational Conference on Application of Information and Communication Technologies, AICT2011:1-4.
[21] Anthony T. Velte Toby J. Velte Robert Elsenpeter.云计算实践指南[M].周庆辉,陈宗斌,译.1版.北京:机械工业出版社,2010:50-57.
[22] ChohanNavraj, BunchChris, Pang, Sydney, et al.. AppScale: Scalable and open AppEngineapplication development and deployment[C].1st International Conference on Cloud Computing,CloudComp2009:57-70.
[23] Jackson, Keith R., RamakrishnanLavanya, MurikiKrishna. Performance analysis of highperformance computing applications on the Amazon Web Services cloud[C].2nd IEEEInternational Conference on Cloud Computing Technology and Science, CloudCom2010:159-168.
[24] Saha, Proshanta, El-Ghazawi, Tarek. Applications of heterogeneous computing inhardware/software co-scheduling[C].2007IEEE/ACS International Conference on ComputerSystems and Applications, AICCSA2007:942-949.
[25]许桢.关于CPU+GPU异构计算的研究与分析[J].科技信息,2010,(17):613-614.
[26]赵彦.基于GPU技术的增强现实系统[D].成都:电子科技大学,2009.
[27] Mark J.Kilgard, Randima Fernando. Cg教程:可编程实时图形权威指南[M].洪伟,译.北京:人民邮电出版社,2004:1-30.
[28] St-Laurent Sebastien, Sebastien St-Laurent. The COMPLETE Effect and HLSL Guide[M].Washington:Paradoxal Press,2005:3-22.
[29]李俊山,罗蓉,赵方舟.数据库原理及应用[M].1版.北京:清华大学出版社,2009:270.284.
[30] E. F. Codd. A Relational Model of Data for Large Shared Data Banks. Communications ofthe ACM,1970,13(6):377-387.
[31] Ozsu, M. Tamer, Patrick Valduriez.Principles of Distributed Database Systems.3th ed.NewYork:Springer,2011.
[32] Afshar, M.,Bates, J.,Bierman, G., Moody, K..A new general purpose parallel databasesystem.Parallel Architectures, Algorithms, and Networks,1997.(I-SPAN '97) Proceedings., ThirdInternational Symposium on,1997:2-8.
[33] W.H. Inmon. Building Data Warehouse.2th ed. Wiley,2005.
[34] E.F.Codd， Providing OLAP (On-Line Analytical Processing) to User-Analysts. PCWorld,1993,9.
[35] Jiawei Han. Data Mining:Concepts and Techniques.2th ed. Diane Cerra,2006.
[36] Charles Petzold.3D Programming for Windows[M].1th ed. Washington: MicrosoftPress,2007:3.
[37]耿肇英. Direct3D和XNA游戏开发基础（C#语言版）[M].1版.北京:清华大学出版社,2009.
[38] HPC选型必备六招[J].微电脑世界,2010,(07):90-91.
[39]吴炜煜.面向对象分析设计与编程OOA/OOD/OOP/AOP.2版.北京:清华大学出版社,2011:10-17.
[40]郝培锋,崔建江,肖文栋.计算机仿真技术[M],沈阳:东北大学出版社,2003.7:110-111.
[41] Christian Nagel, Bill Evjen, Jay Glynn. Professional C#4.0and.Net4[M].UK:Wrox,2010:1-18.
[42] Charles Petzold, Programming Windows Phone7[M]. United States of America:MicrosoftPress,2010:19.
[43] David A. Bader.面向千万亿次计算[M].都志辉,等译.北京：清华大学出版社,2008:70.
[1]陈雄华,涂传滨.精通JBuilder2005[M].1版.北京:电子工业出版社,2005:606.
[2] Stefan Zerbst.3D Game Engine Programming,Thomson Coursetechnology PTR,June,2004:3-39.
[3] Wang, Hechuang; Miao, Fang; Guo, Xirong; Zhang, Peng; Leng, Xiaopeng. DesigningArchitecture in Practice of the3D Engine for Geo-Browser.The International Conference onMultimedia Technology (ICMT2010)2010. Accession number:20111013721815ISBN:978-1-4244-7872-9:1427-1430.
[4] Nancy A Lynch. Distributed Algorithms. San Francisco: Morgan KaufmannPublishers,1996:2-5.
[5]陈明.分布式计算应用模型[M].北京:科学出版社,2009:10-49.
[6] J.Barnes and P. Hut A hierarchical O(NlogN) force-calculation algorithm. Nature,324:446-449, December1986.
[7]Floyd S, Fall K. Promoting the use of end-to-end congestion control in the internet. IEEE/ACMTranscation on Networking.1999,7(4):458-472.
[8]Padhye J, Firoiu V, Towsley D. Et al..Modeling TCP throughput: a simple model and itsempirical validation. In: Oran D,ed.. Proceeding of the SIGCOMM. Vancouver: ACM Press,1998:303-314.
[9]董科军,阎保平.科学数据环境下基于域的存储网格模型[J].微电子学与计算机,2007.24(6):158-161.
[10] Kempler Steven J.. Earth science remote sensing data and services and informationtechnology at the NASA/GES DISC
[11]王燕,何小明,臧海佳,等.美国卫星科学数据处理存档系统S4PA综述[J].遥感信息,2010(2):109-113.
[12]马高峰,郑勇,郭睿.月球参考系的建立[A].中国空间科学学会空间探测专业委员会第十八次学术会议论文集（上册）[C],2005:56-62.
[13]Christian Nagel,etc. Wrox Professional C#2008[M]. Wiley Publishing， Inc.2008:865-876.
[14]谢剑斌,李沛秦,刘通,程江华.一种基于内容分割的数字地图插值新算法[J].计算机工程与科学,2008.30(2):55-57,138
[15] Aaron Reed. Learning XNA4.0.1th,ed. O’Reilly Media,2010:299:321.
[16] Gastón C. Hillar.3D Game Development with Microsoft Silverlight3Beginner's Guide.1thed. Packt Publishing Ltd,2009:255-273.
[17] Whaler K. Does the whole of the Earth’s core convect [J] Nature,1980,287:528-529.
[18]李泳泉，刘建忠，欧阳自远，郑永春，李春来。月球磁场与月球演化[J]。地球物理学进展，2005，20（4）：1003-1008.
[19]王亶文.地磁场模型研究[J].国际地震动态,2001,(04).
[20]安振昌.中国地区地磁场的球冠谐和分析[J].地球物理学报,1993,(06).
[21]安振昌.地磁场模型的计算和评述[J].地球科学进展，1993（04）.
[22] Willianms D J,Mead G D.Night side magnetospheric configuration as obtained fromtrapped electrons at1100kilometers[J].J.Geophys.Res.,1965,70:3017-3029.
[23] Alekseev I I,Shabansky V P. A model of a magnetic field in the Geomagnetosphere[J].Planet.Space Sci.,1972,20:117-133.
[24] Voigt G H. A three dimensional analytical magnetosphericmodel with definedmagnetopause [J].Geophys.J.,1972,38:319-331.
[25] Choe J Y, Beard D B. The compressed geomagnetic field as a function of dipoletilt.Planet[J].Space Sci.,1974,22:595-608.
[26] Choe J Y,Beard D B.The near earth magnetic field of themagnetotail current[J].Planet.Space Sci.,1974,22:609-615.
[27] Walker R J. Quantitative modeling of planetary magnetospheric magnetic fields.In:Quantitative modeling of magnetospheric processes,ed[J].Olson W P. AGU,WashingtonDC,1979,7-34.
[28] Olson W P,Pfitzer K A. A dynamic model of the magnetospheric magnetic and electric fieldsfor July29,1977[J].J.Geophys. Res.,1982,87:5943-5948.
[29] Tsyganenko N A. Numerical models of quiet and disturbedgeomagnetic field in the cislunarpart of the magnetosphere[J].Ann.Geophys.,1981,37:381-391.
[30] Voigt G H. A mathematical magnetospheric field model with independent physicalparameters [J].Planet.Space Sci.,1981,29:1-20.
[31] Taylor H E, Hones E W. Adiabatic motion of auroral particles in a model ofthe electric andmagnetic fields surrounding the earth[J].J.Geophys. Res.,1965,70:3605-3628.
[32] Willis D M, Pratt R T.A quantitative model of the geomagnetic tail [J].J.Atmos.Terr.Phys.,1972,34:1955-1976.
[33]刘国栋,陈乐寿.大地电磁测深研究[M].地震出版社,1984(06).
[2] Cranmer S R et al. Spectroscopic constraints on models of ions-cyclotron resonance heating inthe polar solar corona and high-speed Solar wind. Astrophys.J.,1999,518-L937.
[3]束传政,王玲华,涂传诒.回旋波共振加热加速的多重离子流体模型.空间科学学报,2000,(03):201-207.
[4]冶萍,张靖周.求解N-S方程SIMPLE算法的半交错网格研究.南京航空航天大学学报.2004,36(1):11-15.
[5]阚圣哲,陈国兴,陈磊.基于Abaqus软件的并行计算集群平台构建与优化方法[J].防灾减灾工程学报,2009,29(06):644-651
[6]段孝国.分布式计算技术介绍[J].电脑知识与技术,2011.7(22):5463-5465.
[7] Behrouz A. Forouzan,Firouz Mosharraf. Foundations of Computer Science[M].2th ed.London:Gaynor Redvers-Mutton,2008:4-5.
[8]刘久星,孙永强. PRAM,BSP和LogP并行模型之间的关系及其比较.小型微型计算机系统,1999,20(11):824-827.
[9]罗达强.探析Windows Azure Platform微软云计算平台.硅谷,2010(6):9-10.
[10] Grama,M.,等.并行计算导论[M].张武,等,译.2版.北京:机械工业出版社,2004:1-6.
[11] David A. Bader.面向千万亿次计算[M].都志辉,等译.北京：清华大学出版社,2008:59-72.
[12] P. Hut. Dense stellar systems as laboratiories for fundamental physics. In ArXiv Astrophysicse-prints,2006.
[13] John D Evans, Robort R Kessler. A communication-ordered task graph allocation algorithm.IEEE transaction on Computers, March1981, C(3):207-214.
[14]ollweg, Joseph V., Heating and acceleration of the solar wind in coronal holes: Cyclotronresonances. Advances in Space Research, v30, n3, p469,2002.
[15]Li, Xing. Acceleration of the fast solar wind through minor ions. European Space Agency,(Special Publication) ESA SP, n547, p369-374.
[16]Marsch, Eckart. Status of knowledge after helios, ulysses and soho of the microstate of thecoronal and solar-wind plasma. European Space Agency,(Special Publication) ESA SP, n SP-641,October2006.
[17]Chouliaras.Customization of an embedded RISC CPU with SIMD extensions for videoencoding: A case study.Integration, the VLSI Journal, v41, n1, p135-152, January2008.
[18] Mohanty, Saraju P..GPU-CPU multi-core for real-time signal processing. Digest ofTechnical Papers-IEEE International Conference on Consumer Electronics,2009.
[19]Smelyanskiy, Mikhail (Intel Corporation); Holmes, David; Chhugani, Jatin; etc..Mappinghigh-fidelity volume rendering for medical imaging to CPU, GPU and many-core architectures.IEEE Transactions on Visualization and Computer Graphics, v15, n6, p1563-1570.
[20] Teodoro, George. Coordinating the use of GPU and CPU for improving performance ofcompute intensive applications. Proceedings-IEEE International Conference on ClusterComputing, ICCC,2009,2009IEEE International Conference on Cluster Computing andWorkshops, CLUSTER '09Language: English
[21]涂传诒.太阳风起源和形成的研究进展.中国科学A辑,2000,(S1),5-8.
[22] P. Hut. Dense stellar systems as laboratiories for fundamental physics. In ArXiv Astrophysicse-prints,2006.
[23]石勇,魏奉思,冯学尚等.1997年1月事件行星际扰动在三维背景太阳风中传播的数值模拟.地球物理学报,2001,(03),303-310。
[24]敖先志,沈迹,涂传诒.太阳风质子各向异性速度分布的产生机制.中国科学A辑,2002,(10),940-944.
[25]刘小龙,赵华,C.T.Russell等.上游太阳风中高能电子向同步轨道区的传输.空间科学学报,2003,(03),174-180.
[26]薛金涛,王敬芳,钱昕等.单流体二维太阳风结构的磁流体力学模拟.空间科学学报,2003,(01),1-6.
[27]叶占银,魏奉思,王赤.两冕流间CME事件数值模拟的改进.空间科学学报,2004,(01),1-7.
[28]冯学尚,向长青,钟鼎坤等.三维太阳风结构的Ulysses观测和MHD模拟的比较研究.科学通报,2005,(08),820-826.
[29]彭忠.太阳风-磁层-电离层耦合的全球MHD数组模拟研究[D].合肥:中国科学技术大学,2009:2-14.
[30]许丽华,刘淼. MPI并行编程环境的研究.现代计算机,2003,(04):28-31.
[31]崔雪冰,张延红,李国徽.基于通用计算的GPU-CPU协作计算模式研究微.电子学与计算机,2009,(08):30-33.
[32]王栋栋,庄雷. CPU与GPU并行计算的火焰模拟.计算机应用,2009,(06):1702-1706,1710
[1] Horne, Richard B., Wave acceleration of electrons in the Van Allen radiation belts. NaturePublishing Group,2005,437:227-230.
[2]吴锤结.太阳“发射”高速风影响地球气候和导航系统[J].大连:Space Travel,2009(24):105.
[3]余英.太阳活动对卫星通信的影响[J].广播与电视技术,2000,(12):120-123.
[4]涂传诒.太阳风源于日冕的漏斗状磁场结构[J].中国基础科学,2006,(02).
[5]欧阳自远等.绕月探测工程科学应用研究进展[J]中国科学院院刊，2009，24（5）：530-536.
[6]刘国栋，陈乐寿。大地电磁测深研究[M].地震出版社，1984（06）.
[7] MarschEetal.Wave heating and acceleration of solar wind by cyclotron resonance.J.Geophys.Res.,1982,87:5030-5036.
[8] HollwegJV, etal. Possibleevidencefor coronalAlfvenwaves.J.Geophys. Res.,1982,57:1
[9] Isenberg PA.Resonantaccelerationandheatingofsolarwindions:Anisotropyanddispersion.JGeophys.Res.,1984,89:6613-6621.
[10]吴其芬,李桦著.磁流体动力学[M].北京:国防科技大学出版社，2007.
[11] J. W. Dungey.Interplanetary Magnetic Field and the Auroral Zones.Phys. Rev.1961:6,47–48.
[12] Axford, W.L., and C. O. Hines, A unifying theory of high-latitude geophysical phenomenaand geomagnetic storms, Can. J. phys.,39,1961:1433.
[13] Ming Xue, CAPS/Univ. of Oklahoma, Norman, OK; and F. Kong, D. Weber, K. W. Thomas,Y. Wang, K. Brewster, K. K. Droegemeier, J. S. Kain, S. J. Weiss, D. R. Bright, M. S. Wandishin,M. C. Coniglio, and J. Du. CAPS Realtime Storm-scale Ensemble and High-resolution Forecastsas Part of the NOAA Hazardous Weather Testbed2007Spring Experiment.22nd Conference onWeather Analysis and Forecasting/18th Conference on Numerical Weather Prediction.
[14]张子平,马庆林.帕克与太阳风的发现.天文爱好者.2005(2):8-10.
[15]束传政,王玲华,涂传诒.回旋波共振加热加速的多重离子流体模型.空间科学学报,2000,(03):201-207.
[16]彭忠,陈耀.流管几何对低速太阳风alpha粒子丰度的影响.中国科学技术大学学报,2007,(08):852-858.
[17]敖先志,沈迹,涂传诒.太阳风质子各向异性速度分布的产生机制.中国科学A辑,2002,(10),940-944.
[18] David A. Bader.面向千万亿次计算[M].都志辉,等,译.北京：清华大学出版社,2008:59-72.
[19] P. Hut. Dense stellar systems as laboratiories for fundamental physics. In ArXiv Astrophysicse-prints,2006.
[20] John D Evans, Robort R Kessler. A communication-ordered task graph allocation algorithm.IEEE transaction on Computers, March1981, C(3):207-214.
[21]崔雪冰,张延红,李国徽.基于通用计算的GPU-CPU协作计算模式研究微.电子学与计算机,2009,(08),30-33.
[22]王栋栋,庄雷. CPU与GPU并行计算的火焰模拟.计算机应用,2009,(06),1702-1706+1710.
[23]周建,刘鹏,梅优良等.基于微处理器核的AAC解码软硬件协同优化.浙江大学学报(工学版),2007,(08),1271-1276.
[24]许丽华,刘淼. MPI并行编程环境的研究.现代计算机,2003,(04),28-31.
[25]冯学尚,向长青,钟鼎坤等.三维太阳风结构的Ulysses观测和MHD模拟的比较研究.科学通报,2005,(08),820-826.
[26]刘小龙,赵华,C.T.Russell等.上游太阳风中高能电子向同步轨道区的传输.空间科学学报,2003,(03),174-180.
[27]石勇,魏奉思,冯学尚等.1997年1月事件行星际扰动在三维背景太阳风中传播的数值模拟.地球物理学报,2001,(03),303-310.
[1]冯恒强，吴德金，赵寄昆.通过建立自然坐标系确定磁云边界方法的探讨[J].天文学报，2006，47(4):371-386.
[2]徐龙道.物理学词典[M].1版.北京:科学出版社,2007:9-9.
[3]钟江帆.大学物理(上册)[M].北京:高等教育出版社,2005:77-78.
[4]解恩泽.爱因斯坦统一场论思想的命运[J/OL].中国青年科技,1999,3:28-30.
[5]何圣静.物理定律的形成与发展[M].北京:测绘出版,1988:55-81.
[6]复旦大学《物理学》编写组.物理学.1版.北京:高等教育出版社,1985:374-377.
[7]张三慧.大学基础物理学(上册)[M].1版.北京:清华大学出版社,2003:206-207.
[8]王晓鸥.物理学概论[M].1版.上海:同济大学出版社,2007:103.
[9]卢新平.简明普通物理学[M].1版.上海:同济大学出版社,2007:131.
[10] B.M.亚沃尔斯基(Яворский,В.М.), A.A.杰特拉无(Детлаф,А.А.).物理学手册[M].宓鼎樑,等,译.1版.北京:科学出版社,1986:146-147.
[11]张桂兰，聂东壁山，夏玉珍.对理想气体分子按平均动能分布的讨论[J].临沂师范学院学报，2003，25(6):32-34.
[12]徐劳立.关于能量均分定理的讨论[J].物理与工程，2005,15(6):61-62.
[13]西南中南九所高等院校.大学物理学[M].成都:四川科学技术出版社,1987:385-391.
[14]钟江帆.大学物理(下册)[M].高等教育出版社.2005:253-258.
[15]马根源,王松立,金庆华.物理学[M].1版.天津:南开大学出版社,1993:448-463.
[16]程守洙,江之永.普通物理学[M].5版.北京:高等教育出版社,1997:295-298.
[17] Gavryuseva, E., Godoli, G., Structure and rotation of the large scale solar magnetic fieldobserved at the Wilcox Solar Observatory. Elsevier Ltd,2006(31):68-76.
[18]徐兵.地球和太阳的磁场从哪里来[J].大科技科学之谜,2009,10:12-13.
[19]林元章.太阳活动起源研究(Ⅱ):太阳发电机理论[J].天文学进展,2002,(01)23:32.
[20]陈宏芳.原子物理学[M].1版.北京:科学出版社,2006:1-32.
[21]王永昌.近代物理学[M].1版.北京:高等教育出版社,2006:169-170.
[22]杨桂林,江兴方,柯善哲.近代物理.1版.北京:科学出版社,2006:22-23.
[23]陈强顺，冯承天.电荷守恒定律可确定麦克斯韦方程组的数学形式[J].同济大学学报，1989,17(3):395-400.
[24] P.J.OUSEPH.技术物理学[M].黄伟民,褚孟似译.北京：高等教育出版社,1992:597-598.
[25]陈丹.太阳的日冕观测和磁场观测[J].太空探索,2010(6):62-65.
[26] WengYing, WengZi-Hua. Heavy ions acceleration of solar wind in electromagnetic andgravitational fields. Progress in Electromagnetics Research Symposium, PIERS2011Marrakesh,2011:1780-1784.
[1]杨立,齐建玲,曲风娟等.计算机控制与仿真技术[M].1版.北京:中国水利水电出版社,2006:14.
[2]王正林,王胜开,陈国顺等.MATLAB/Simulink与控制系统仿真[M].2版.北京:电子工业出版社,2008.7:7.
[3]雷永林.仿真模型重用理论、方法与异构集成技术研究[D].长沙:国防科学技术大学,2006.
[4]马元元,孙志挥,张南翔.增量关联规则在大型火力发电厂实时控制中的应用[J].工业控制计算机,2000,(01)14-15.
[5]高薇,张强,孙军.通过超实时仿真验证轨控策略方法研究[J].航天控制,2008,26(2):71-73,78.
[6]林源.一种基于虚拟仪器的局域网导弹测控仿真培训系统[J].软件导刊,2009(03):79-81.
[7]许喆.某外场水下试验平台的视景仿真研究[D].西安:西北工业大学,2007:3.
[8]詹剑锋.挑战下一个大家伙写在2011超级计算大会之后[J].程序员2012(1):14-15.
[9] Alpert, D.,Avnon, D..Architecture of the Pentium microprocessor.Micro, IEEE,2002,13:11–21.
[10] Behrouz A.Forouzan, Firouz Mosharraf.计算机科学导论[M].刘艺,瞿高峰,译.2版.北京:机械工业出版社,2009:83.
[11] Katz, R.Reduced instruction set computers[C]. Solid-State Circuits Conference. Digest ofTechnical Papers.1986:218-219.
[12] Schlansker, M.S.,Rau, B.R., Hewlett-Packard Labs.EPIC: Explicitly Parallel InstructionComputing[J]. Computer,2000(33):37-45.
[13] Eggers, S.J.,Emer, J.S.,Leby, H.M., et al. Simultaneous multithreading: a platform fornext-generation processors. Micro, IEEE,1997(17):12-19.
[14] Nayfeh, B.A., Olukotun, K.,Stanford Univ., CA. A single-chip multiprocessor[J].Computer,1997(30):79-85.
[15]武剑锋,朱家维.对称多处理机系统[J].微型机与应用,1992(9):21-23.
[16] Haas, Peter W., Resch, Michael M., High performance computing in science andengineering[C].17th International Symposium on Computer Architecture and High PerformanceComputing, SBAC-PAD2005:10-18.
[17] Chen, Guo-Liang, Sun, Guang-Zhong, Zhang, Yun-Quan, et al.Study on parallel computing.Journal of Computer Science and Technology.Springer New York LLC,2006:665-673.
[18] AdlerM.,GibbonsP.B., MatiasY.; RamachandranV., Modeling Parallel Bandwidth: Localversus Global Restrictions. Springer-Verlag New York Inc,1999:381-404.
[19] Boghosian, Bruce M., Coveney, Peter V.Introduction of grid computing application projectsat the NASA earth science technology office[C]. IEEE Computer Society,2006:289-298.
[20] Kalagiakos, Panagiotis, Karampelas, Panagiotis. Cloud computing learning[C].20115thInternational Conference on Application of Information and Communication Technologies, AICT2011:1-4.
[21] Anthony T. Velte Toby J. Velte Robert Elsenpeter.云计算实践指南[M].周庆辉,陈宗斌,译.1版.北京:机械工业出版社,2010:50-57.
[22] ChohanNavraj, BunchChris, Pang, Sydney, et al.. AppScale: Scalable and open AppEngineapplication development and deployment[C].1st International Conference on Cloud Computing,CloudComp2009:57-70.
[23] Jackson, Keith R., RamakrishnanLavanya, MurikiKrishna. Performance analysis of highperformance computing applications on the Amazon Web Services cloud[C].2nd IEEEInternational Conference on Cloud Computing Technology and Science, CloudCom2010:159-168.
[24] Saha, Proshanta, El-Ghazawi, Tarek. Applications of heterogeneous computing inhardware/software co-scheduling[C].2007IEEE/ACS International Conference on ComputerSystems and Applications, AICCSA2007:942-949.
[25]许桢.关于CPU+GPU异构计算的研究与分析[J].科技信息,2010,(17):613-614.
[26]赵彦.基于GPU技术的增强现实系统[D].成都:电子科技大学,2009.
[27] Mark J.Kilgard, Randima Fernando. Cg教程:可编程实时图形权威指南[M].洪伟,译.北京:人民邮电出版社,2004:1-30.
[28] St-Laurent Sebastien, Sebastien St-Laurent. The COMPLETE Effect and HLSL Guide[M].Washington:Paradoxal Press,2005:3-22.
[29]李俊山,罗蓉,赵方舟.数据库原理及应用[M].1版.北京:清华大学出版社,2009:270.284.
[30] E. F. Codd. A Relational Model of Data for Large Shared Data Banks. Communications ofthe ACM,1970,13(6):377-387.
[31] Ozsu, M. Tamer, Patrick Valduriez.Principles of Distributed Database Systems.3th ed.NewYork:Springer,2011.
[32] Afshar, M.,Bates, J.,Bierman, G., Moody, K..A new general purpose parallel databasesystem.Parallel Architectures, Algorithms, and Networks,1997.(I-SPAN '97) Proceedings., ThirdInternational Symposium on,1997:2-8.
[33] W.H. Inmon. Building Data Warehouse.2th ed. Wiley,2005.
[34] E.F.Codd， Providing OLAP (On-Line Analytical Processing) to User-Analysts. PCWorld,1993,9.
[35] Jiawei Han. Data Mining:Concepts and Techniques.2th ed. Diane Cerra,2006.
[36] Charles Petzold.3D Programming for Windows[M].1th ed. Washington: MicrosoftPress,2007:3.
[37]耿肇英. Direct3D和XNA游戏开发基础（C#语言版）[M].1版.北京:清华大学出版社,2009.
[38] HPC选型必备六招[J].微电脑世界,2010,(07):90-91.
[39]吴炜煜.面向对象分析设计与编程OOA/OOD/OOP/AOP.2版.北京:清华大学出版社,2011:10-17.
[40]郝培锋,崔建江,肖文栋.计算机仿真技术[M],沈阳:东北大学出版社,2003.7:110-111.
[41] Christian Nagel, Bill Evjen, Jay Glynn. Professional C#4.0and.Net4[M].UK:Wrox,2010:1-18.
[42] Charles Petzold, Programming Windows Phone7[M]. United States of America:MicrosoftPress,2010:19.
[43] David A. Bader.面向千万亿次计算[M].都志辉,等译.北京：清华大学出版社,2008:70.
[1]陈雄华,涂传滨.精通JBuilder2005[M].1版.北京:电子工业出版社,2005:606.
[2] Stefan Zerbst.3D Game Engine Programming,Thomson Coursetechnology PTR,June,2004:3-39.
[3] Wang, Hechuang; Miao, Fang; Guo, Xirong; Zhang, Peng; Leng, Xiaopeng. DesigningArchitecture in Practice of the3D Engine for Geo-Browser.The International Conference onMultimedia Technology (ICMT2010)2010. Accession number:20111013721815ISBN:978-1-4244-7872-9:1427-1430.
[4] Nancy A Lynch. Distributed Algorithms. San Francisco: Morgan KaufmannPublishers,1996:2-5.
[5]陈明.分布式计算应用模型[M].北京:科学出版社,2009:10-49.
[6] J.Barnes and P. Hut A hierarchical O(NlogN) force-calculation algorithm. Nature,324:446-449, December1986.
[7]Floyd S, Fall K. Promoting the use of end-to-end congestion control in the internet. IEEE/ACMTranscation on Networking.1999,7(4):458-472.
[8]Padhye J, Firoiu V, Towsley D. Et al..Modeling TCP throughput: a simple model and itsempirical validation. In: Oran D,ed.. Proceeding of the SIGCOMM. Vancouver: ACM Press,1998:303-314.
[9]董科军,阎保平.科学数据环境下基于域的存储网格模型[J].微电子学与计算机,2007.24(6):158-161.
[10] Kempler Steven J.. Earth science remote sensing data and services and informationtechnology at the NASA/GES DISC
[11]王燕,何小明,臧海佳,等.美国卫星科学数据处理存档系统S4PA综述[J].遥感信息,2010(2):109-113.
[12]马高峰,郑勇,郭睿.月球参考系的建立[A].中国空间科学学会空间探测专业委员会第十八次学术会议论文集（上册）[C],2005:56-62.
[13]Christian Nagel,etc. Wrox Professional C#2008[M]. Wiley Publishing， Inc.2008:865-876.
[14]谢剑斌,李沛秦,刘通,程江华.一种基于内容分割的数字地图插值新算法[J].计算机工程与科学,2008.30(2):55-57,138
[15] Aaron Reed. Learning XNA4.0.1th,ed. O’Reilly Media,2010:299:321.
[16] Gastón C. Hillar.3D Game Development with Microsoft Silverlight3Beginner's Guide.1thed. Packt Publishing Ltd,2009:255-273.
[17] Whaler K. Does the whole of the Earth’s core convect [J] Nature,1980,287:528-529.
[18]李泳泉，刘建忠，欧阳自远，郑永春，李春来。月球磁场与月球演化[J]。地球物理学进展，2005，20（4）：1003-1008.
[19]王亶文.地磁场模型研究[J].国际地震动态,2001,(04).
[20]安振昌.中国地区地磁场的球冠谐和分析[J].地球物理学报,1993,(06).
[21]安振昌.地磁场模型的计算和评述[J].地球科学进展，1993（04）.
[22] Willianms D J,Mead G D.Night side magnetospheric configuration as obtained fromtrapped electrons at1100kilometers[J].J.Geophys.Res.,1965,70:3017-3029.
[23] Alekseev I I,Shabansky V P. A model of a magnetic field in the Geomagnetosphere[J].Planet.Space Sci.,1972,20:117-133.
[24] Voigt G H. A three dimensional analytical magnetosphericmodel with definedmagnetopause [J].Geophys.J.,1972,38:319-331.
[25] Choe J Y, Beard D B. The compressed geomagnetic field as a function of dipoletilt.Planet[J].Space Sci.,1974,22:595-608.
[26] Choe J Y,Beard D B.The near earth magnetic field of themagnetotail current[J].Planet.Space Sci.,1974,22:609-615.
[27] Walker R J. Quantitative modeling of planetary magnetospheric magnetic fields.In:Quantitative modeling of magnetospheric processes,ed[J].Olson W P. AGU,WashingtonDC,1979,7-34.
[28] Olson W P,Pfitzer K A. A dynamic model of the magnetospheric magnetic and electric fieldsfor July29,1977[J].J.Geophys. Res.,1982,87:5943-5948.
[29] Tsyganenko N A. Numerical models of quiet and disturbedgeomagnetic field in the cislunarpart of the magnetosphere[J].Ann.Geophys.,1981,37:381-391.
[30] Voigt G H. A mathematical magnetospheric field model with independent physicalparameters [J].Planet.Space Sci.,1981,29:1-20.
[31] Taylor H E, Hones E W. Adiabatic motion of auroral particles in a model ofthe electric andmagnetic fields surrounding the earth[J].J.Geophys. Res.,1965,70:3605-3628.
[32] Willis D M, Pratt R T.A quantitative model of the geomagnetic tail [J].J.Atmos.Terr.Phys.,1972,34:1955-1976.
[33]刘国栋,陈乐寿.大地电磁测深研究[M].地震出版社,1984(06).