Spherical Space Grid Of Distributed Parallel Technology Research And Implementation

In recent years, with the cloud of distributed computing, virtual reality, computer graphics, mobile visualization development, the use of remote sensing real-time access to air and terrain elevation data, the use of distributed parallel computing real-time generation of high levels of detail spherical grid data, in the a variety of fixed and mobile terminals browsing and querying the entire spherical space mesh of geographic service information is also becoming more common. But the whole spherical space, terrain and image data are massive, and these spherical subdivision mesh treatment is extremely complex and time-consuming process, it is difficult in ordinary stand-alone workstation,Therefore, a distributed parallel computing to deal with sphere mesh is a good solution, especially the cloud-based distributed computing, is a spherical mesh real-time subdivision processing a research focus.In this paper, fully understand the spherical mesh generation based on the study and implementation of the spherical mesh, distributed parallel processing methods and technology principles. This paper has done the main work of three parts. First examine the sphere mesh-based engine design methods and implementation techniques. The specific approach is based on the Earth's latitude and longitude of four equal portions the sphere of the initial split of the four largest diamond plots Land, each Land was divided into under the resolution of N x N has a different resolution level of detail of the grid tiles Patches. Each Patch consists of two isosceles triangle formed, known as Binary Triangle Tree, as the rendering of each frame split basis. Then describes math foundation of spherical apatial mesh and the sources of terrain elevation data to generate and read methods. Spherical mesh data structure and data organization, as well as the spherical surface of the grid subdivision process.Focuses on the dynamics of spherical mesh generation and visual field cut.The second part describes the basic parallel computing concepts and principles, using the parallel library spherical space grid implementation of the parallel partition approach, which focuses on the basic features of OpenMP and MPI parallel program development and the model. For distributed network data packet communication and the process of reconciliation package. The use of modular and object-oriented approach, the overall serial, parallel design of the core computing part of the model, the sphere grid, parallel processing is divided into five modules encapsulate the coordination of processing to be completed, and the last describes the whole process of workflow execution process and code implementation, while the experiment proved the feasibility of this theory.Finally, introduce the principal of distribute parallel processing and technological methold, which highlights the Winsock network data communication program design method. Windows service design methods and to start Windows services for registration and participation in parallel distributed Split service node; and finally introduced a spherical mesh of specific distributed parallel subdivision technology architecture and processing process, and gives a specific experiment. Compared to OpenMP and MPI parallel library, Winsock-based distributed parallel sphere grid more flexible and intuitive display of spherical parallel grid subdivision process, experimental verification of the spherical grid distributed parallel feasibility.