高级检索
刘德儿, 熊证, 沈敬伟, 朱晓璠, 徐昕. 完全拓扑关系结构D-TIN多核并行构建[J]. 计算机辅助设计与图形学学报, 2016, 28(6): 950-959.
引用本文: 刘德儿, 熊证, 沈敬伟, 朱晓璠, 徐昕. 完全拓扑关系结构D-TIN多核并行构建[J]. 计算机辅助设计与图形学学报, 2016, 28(6): 950-959.
Liu De'er, Xiong Zheng, Shen Jingwei, Zhu Xiaofan, Xu Xin. Creation of Full Topological D-TIN Based on Multi-core Parallel[J]. Journal of Computer-Aided Design & Computer Graphics, 2016, 28(6): 950-959.
Citation: Liu De'er, Xiong Zheng, Shen Jingwei, Zhu Xiaofan, Xu Xin. Creation of Full Topological D-TIN Based on Multi-core Parallel[J]. Journal of Computer-Aided Design & Computer Graphics, 2016, 28(6): 950-959.

完全拓扑关系结构D-TIN多核并行构建

Creation of Full Topological D-TIN Based on Multi-core Parallel

  • 摘要: 为了充分利用普通PC机多核CPU计算资源实现大数据量离散点集矢量地形快速构建,提出具有完全拓扑关系结构D-TIN的多核并行生成算法.首先基于四方边缘结构设计具有完全拓扑关系的D-TIN数据结构及系列函数接口,用于操作该数据结构;然后顾及线程间数据负载平衡和空间均质性,采用坐标排序法对离散点集进行均等虚拟划分;再基于离散点集凸包"生成—消融—生成"周期性特征和分治算法思想设计D-TIN串行构建算法;最后以其作为D-TIN并行构建算法的内核,Open MP作为多核并行编程模型,实现了单机多核环境下的D-TIN并行构建.实验结果表明,当数据量较大时,该算法效率明显高于串行算法,且两者的计算结果具有完全一致性;同时,参与计算线程数与总耗时呈负相关.

     

    Abstract: To take full advantage of the multi-core CPU to quickly create the vector terrain based on a large number of disperse points, this paper proposed a multi-core parallel algorithm to create the full topological D-TIN. Firstly, the quad-edge structure was chose to design the D-TIN data structure, and a series of function interfaces were developed to manipulate it; and given consideration to the data load balancing within threads and well-distributed in spatial, the method of coordinate sorting was used to divide the discrete points set into equal parts in logical; then, the serial construction algorithm of D-TIN was developed based on the periodicity of "creation-melting-creation" of the convex hull of the disperse point set and the divide-and-conquer algorithm; lastly, it was taken as the kernel of D-TIN parallel construction algorithm and Open MP as the model of multi-core parallel programming, then the parallel construction of D-TIN was achieved based on the stand-alone multi-core CPU. The experimental results show that the efficiency of this algorithm is significantly higher than that of the serial algorithm when the data is lager, and the D-TINs cre-ated by them are completely consistent; and there is negative correlation between the number of threads used and the total elapsed time.

     

/

返回文章
返回