文献类型：期刊文献

英文题名：GPU-based parallel simulation of silicon anisotropic etching

作者：Li, Jianhua[1]; Wang, Yan[2]; Chen, Jingyuan[1]; Yan, Li[1]

机构：[1] Department of Computer Science and Engineering, East China University of Science and Technology, Shanghai 200237, China; [2] Multiscale Systems Engineering Research Group, Woodruff School of Mechanical Engineering, Georgia Institute of Technology, Atlanta, GA, 30332, United States

年份：2012

卷号：2

期号：PARTS A AND B

起止页码：819

外文期刊名：Proceedings of the ASME Design Engineering Technical Conference

收录：EI(收录号：20134016799506)

语种：英文

外文关键词：Anisotropy - Application programming interfaces (API) - Computer graphics - Program processors - Silicon

摘要：Silicon anisotropic etching simulation, based on geometric model or cellular automata (CA) model, is highly timeconsuming. In this paper, we propose two parallelization methods for the simulation of the silicon anisotropic etching process with CA models on graphics processing units (GPUs). One is the direct parallelization of the serial CA algorithm, and the other is to use a spatial parallelization strategy where each crystal unit cell is allocated to a thread in GPU. The proposed simulation methods are implemented with the Compute Unified Device Architecture (CUDA) application programming interface. Several computational experiments are taken to analyze the efficiency of the methods. Copyright ? 2012 by ASME.