Research And Implementation Of Cache Optimization Algorithm Based On 3D Mesh Model

In recent years, with the improvement of computer technology of graphics software and hardware and the growth of application requirements, the research and application of three-dimensional mesh models present a series of features, such as the rapid increase of model complexity, the complexity of scene object, the requirement of realistic rendering, the number of display resolution is increasing, which are far beyond the computing power provided by the existing processing hardware. Such large scale graphics processing pipeline has brought unprecedented challenges to the entire graphics processing pipeline. In order to solve the problem of large-scale rendering performance, the need for effective management of bandwidth requirements, the size of the control data set, the reduction of the volume of data, and the characteristics of the multi-level storage hierarchy are used to reduce the data access miss rate. This paper is focused on the research of layout optimization technology based on 3D mesh models, the main work is the following three parts:1. In this paper, several common mesh layout optimization algorithms are compared and classified. Based on the solving technology and describes the different ways of different levels can improve the cache access efficiency, at the same time, along with the model of interactive rendering efficiency and algorithm of layout optimization in different time. In the typical algorithm, Lin's algorithm in the average cache miss ratio is the lowest, and there is no specific requirements on the relationship between the topological structure, a detailed analysis of the cache access miss ratio, optimization model and time arrangement between the size and cache size.2. Mesh layout algorithm based on greedy optimization strategy. In view of the problem of low rendering performance of complex data set which is caused by the limited memory bandwidth and data access speed, this paper presents a mesh layout algorithm based on greedy optimization strategy, by rearranging triangle sequences to improve spatial locality and temporal locality. Firstly, divided the vertices into three categories, according to the improved cost function, we choose the vertex with the minimum cost as currently active vertex. Then, output all adjacent faces which not to be rendered, and press neighbors into buffer, the above steps are executed iteratively until adjacent faces of all vertices are rendered. Finally, we get the rearranged triangle sequence. Experimental results show that the algorithm provides a higher vertex cache hit ratio with less execution time, which can effectively solve the problem of data access speed lagging behind the processing speed of GPU seriously.3. Mesh layout optimization algorithm based on mesh segmentation. The algorithm first uses a mesh surface segmentation method of graph partitioning mesh into multiple sub meshes, this method is simple, and the segmentation speed. Then, for each sub mesh cache optimization. This algorithm can solve large-scale mesh model to join the whole memory rearrangement problem, and is suitable multiresolution representation for massive meshes based on clustered hierarchy, for each cluster of the sub mesh to achieve good triangle rearrangement and provide a theoretical basis.4. Design and implementation a mesh layout optimization platform. Design and implementation a massive meshes optimization system with display of the mesh, segmentation and optimization functions, and carry on the detailed test to the platform.