The Parallel Drawing Of Large-scale Terrain Models Based On Multi-core Platform

In the graphics fields, such as virtual reality, three-dimensional games, military simulation, film special effects, etc., the application of multi-core platform has just begun. Real-time drawing and rendering large-scale three-dimensional scenes in the graphics are very important content of research. Therefore initialization drawing process of using large-scale mapping of forest terrain model has widely research value in this field. With the promoting chip manufacturers to the mainstream, multi-core processors have become increasingly popular at the same time, the design method of serial programming can not take full advantage of multicore resources of CPU, so multi-threaded program design has become an important problem of software development.Research work includes the following two aspects: First, multi-core platform with parallel programming ideas and drawing three-dimensional mountain forest scene roaming procedures, analysis a three-dimensional scene rendering a large number of mountain roaming instances of small entities of the third-party plug-ins PagedGeometry of OGRE engine. OGRE platform in the large-scale rendering of forest scenes initialization is designed to load a multi-core parallel fast rendering algorithm with the designed large number of entities Mesh, multi-core processors in the full use of individual processing capacity, enhanced realism and real-time systems; For fast rendering of large scale terrain adopt LOD levels detail model by the need to use mesh simplification algorithm to simplify the model in pre-processing stage. Analysis of large-scale terrain level detail model algorithm can be optimized in parallel place in the pretreatment, summed up a multi-core parallel model simplification algorithms based on meshing of quad-tree. Use of common parallel programming environment for OpenMP its threaded by comparing before and after optimization results show that after the parallel speedup and efficiency have improved significantly.Innovation is using the Intel Parallel Amplifier analyzer from a function point view to analysis the performance bottlenecks, and designing the multi-thread creation and synchronization,dynamic set number of threads,thread scheduling policy while OpenMP paralleling, so that performance is improved. Obtaining validation in the multi-core platform, improving the loading rate of the large-scale terrain model effectively, roaming the display of the terrain scene and improving the efficient of drawing, and proving that the excellent combine the graphics applications with multi-core parallel computing theory. The key is to set the shared variable development of data protection and load distribution between the processor core and balance to ensure that the number of processor cores is high loading and accurately complete the parallel task.