摘要
为了充分利用现有的多核CPU计算资源,提出一种基于OpenMP框架的快速并行分层算法,并对其性能进行讨论.该算法利用模型自然分组特征建立拓扑关系分组,使得模型拓扑数据结构的建立时间缩短;在此基础上,采用基于OpenMP的多线程计算,将拓扑结构的建立过程和求取层片轮廓的过程并行化计算,可以取得接近CPU核数的加速比,因此分层时间明显降低.对于复杂三维模型的超大STL文件进行分层处理,实例计算表明文中算法是一种高效且易于实现的方法.
To reduce the slicing time by utilizing multi-core CPU, a parallel slicing algorithm is proposed in this paper, and the properties of the algorithm is analyzed. Natural grouping feature in the model is used to build to-pological information groups. On the basis of this, an OpenMP-based parallel computing method is adopted to speedup construction of topological information and slicing process. A speedup close to the number of CPU cores was achieved in the test, so less slicing time was spent. The example test result shows that the algorithm proposed in this paper provides an efficient and achievable way to deal with large STL file.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2015年第4期747-753,共7页
Journal of Computer-Aided Design & Computer Graphics
