Dissertation > Excellent graduate degree dissertation topics show

Research of Parallel FDTD Algorithm Based on LAN

Author: LiHaoYu
Tutor: FuJiaHui
School: Harbin Institute of Technology
Course: Electromagnetic Field and Microwave Technology
Keywords: Finite difference time domain method Computational electromagnetics High-performance computing Concurrent Technologies Parallel computer cluster system
CLC: TP393.1
Type: Master's thesis
Year: 2011
Downloads: 42
Quote: 0
Read: Download Dissertation

Abstract


Finite difference time domain (FDTD) method has been widely used in computational electromagnetics. FDTD algorithm formula is simple and intuitive, suitable for solving all kinds of electromagnetic engineering problems and visual computing implementations, the calculation process is recursive in the FDTD grid for solving electromagnetic field values, to deal with large-scale electromagnetic simulation problems to ensuring the accuracy and stability conditions, FDTD grid increased, so that the huge demand for computer memory and solution time is too long, the computer can not run this calculation process, which has become a major bottleneck of the FDTD applications. Improve the performance of computer hardware and fleet platform, high-performance parallel computing applications more widely used. In this paper, parallel LAN technology combined with the FDTD algorithm, a region of space subdivision Policy discrete FDTD region, the FDTD algorithm itself has a good parallel. Call MPI message completion of the transfer function of the discrete computational domain boundary electromagnetic field data communication, the use of data collection technology to complete the host output of each node after the completion of discrete data, so that the the complete FDTD areas each node in parallel cluster system of parallel computing to solve, and Parallel process, test parallel computing performance. In this paper, the hardware of microwave laboratory conditions, to build a parallel computer cluster is divided into parallel hardware platform, software platform, programming mode three parts. Hardware based on existing established laboratory router, LAN environment and personal computer node. The software platform includes two aspects of the strategy based on the MPI message passing parallel software operating platform and programming software development platform. Parallel programming mode selection optimization of the master-slave programming model to write the calculation program code, write and maintain two-part program code, which is more suitable for the existing hardware structure, at the same time makes the calculation procedure of the stability and scalability enhancements, more suitable for parallel computing in the late . This paper presents a dynamic the split parallel regional strategy, the DAMA parallel each node in the system to take the concept of computational. Traditional optimization such as the combination of two parallel parallel particle size or the use of OpenMP and MPI not fundamentally efficient use of resources within the system. The dynamic triangulation strategy is calculated for each node in the test performance according returned to the host normalized performance index distribution calculation amount. FDTD procedures performed in the parallel system prior to test the performance every time according to the normalized index computation assignment. Another in this research is the The parallel FDTD technical data structure pointer instead of an array. Traditional parameters array. Intuitive array definition, a simple statement, and the difficulty of programming, there are two drawbacks: declare an array space and the initialization process waste too many resources and determined from the array is often calculated data overflow. The pointer can compensate for this drawback, the correct flexibility in the use of the pointer, so that the the parallel FDTD calculation procedure more simple, compact and highly efficient.

Related Dissertations

  1. Research of Finite Element Method on GPU,O241.82
  2. Applications of Vector Finite Element Method to Simulate Logging While Drilling Resistivity Tools,P631.83
  3. The Research on Characteristics of Reconfigurable Optical Add/Drop Multiplexer Based on Photonic Crystals,TN929.1
  4. Overload virtual machine performance improvement under MPI communication method,TP302
  5. Based on GPU / CPU multi- level parallel CFD Optimization Strategy,V221
  6. Based on multi-core SMP cluster environment parallel ray tracing simulation of satellite imagery and Implementation,TP391.41
  7. Time Domain Analysis of Basic Eletromagnetic Characteristics for Left-handed Materials,O441.4
  8. Research on Absorbing Boundary Condition and Its Application on Electromagnetic Scattering,O441.4
  9. FDTD Analysis on Influence of Micro-Nano Surface on Near-Field Light Intensity Distribution in KDP Crystal,TG665
  10. Research of Multicast Algorithm on A Novel Interconnection Network PGLH,TP393.02
  11. High - performance computing systems based on particle simulation of the GPU,TP338
  12. The Research of Optimization Technologies for the File System of High Performance Computing Cluster,TP316
  13. The Research of High-performance Computing Course Based on Case-task Driven Teaching Method,TP338-4
  14. Research and Implementation of the Registration Center in the Enterprise Grid Computing System,TP393.18
  15. Study and Design of Tianjin University’s Grid System,TP393.02
  16. Based on the two-dimensional metamaterial FDTD simulation and design of electromagnetic devices,TN011
  17. Fabrication and Band Gap Research of the Two-Dimensional Silicon Photonic Crystals,O734
  18. A Study of GPU-based FDTD Method and Its Application in LED Performance Optimization,TN312.8
  19. GPU Based Spot Noise Parallel Algorithm for 2D Vector Field Visualization,TP391.41
  20. The Study on Characterisce of Multi-wavegength Phtonic Crsytal Wavelength Division Multiplexers Based Multimode Interference Thoery,TN929.11
  21. The Research on Characteristics of Optical-control Two-wavelength Terahertz Modulator Based on Photonic Crystal,TN761

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Computer network > Local area network (LAN), metropolitan - area network (MAN)
© 2012 www.DissertationTopic.Net  Mobile