Optimizing Data Loading And Memory Allocation For Out-of-Core Simplification

The algorithm of Out-of-Core simplification is one of the research hotspots in Computer Graphics in resent years for its importance in many practical applications. The research related to this thesis is conducted on the optimization of Out-of-Core simplification.A survey of Out-of-Core simplification is given in the second chapter. The algorithms are classified according to their realization methods, then an introduction is given particularly to each class that is classified again according to the difference of their algorithms, and some typical algorithms are briefly described along with their merits and shortcomings analyzed.Current algorithms have improved the efficiency of Out-of-Core simplification, but it is too simple to deal with data reading and memory allocation in the algorithms. The simplification of Out-of-Core model must be handled by reading data in batches and using local operations. Obviously, both data reading and main memory allocation have much influence on simplification efficiency. With regard to the investigation on this topic, we put forward novel measures in this paper for dynamically optimizing data reading and memory allocation. By the approach, while the algorithm is simplifying the data at the beginning, it also examines the data block size and memory allocation pattern that may influence the efficiency of the simplification. This allows the computers with different configurations to be able to adaptively get their own optimized data sizes and memory allocation patterns when processing different Out-of-Core models, thus accelerating the simplification process of the remaining data. Experimental results show that our measures can considerably improve Out-of-Core simplification efficiency.