New strategies for developing GPU accelerated disk-based discontinuous deformation analysis for large-scale modeling
文献类型：期刊论文
| 作者 | Liu, Feng3; Chen, Zediao3; Xia, Kaiwen1,3; Xu, Dongdong4; Yang, Yongtao2 |
| 刊名 | INTERNATIONAL JOURNAL FOR NUMERICAL AND ANALYTICAL METHODS IN GEOMECHANICS
 |
| 出版日期 | 2023-01-29 |
| 卷号 | -期号:-页码:- |
| ISSN号 | 0363-9061 |
| 关键词 | contact detection discontinuous deformation analysis disk-based DDA graphics processing unit (GPU) large-scale simulation parallel computing |
| 英文摘要 | The major obstacle for the application of discontinuous deformation analysis (DDA) in engineering problems is the high computational cost and poor efficiency. In this paper, the main algorithms of disk-based DDA are redesigned and implemented on graphics processing unit (GPU) to improve its performance. First, a contact pair-wise scheme is proposed to assemble the stiffness matrix on GPU. Second, a buffer strategy and a GPU version of grid-based contact detection algorithm are developed to improve the efficiency of contact detection. Third, for solving the simultaneous equations, two iterative methods are considered along with the direct solver method. The parallel performances of proposed strategies are tested and compared with the CPU counterparts. The results show that the maximum speedup ratio is 14 for the assembly of the stiffness matrix and 215 for contact detection. The speedup ratio for solving simultaneous equations depends on several factors, and the preconditioned conjugate gradients method (pcg) is suggested. Finally, the effectiveness and performance of the proposed GPU accelerated disk-based DDA is further demonstrated by several examples, one of which consisted of over 500,000 particles. The results show that the proposed method can achieve a satisfactory speedup ratio, and is ready for large-scale problems. |
| 学科主题 | Engineering ; Materials Science ; Mechanics |
| 语种 | 英语 |
| 出版者 | WILEY |
| WOS记录号 | WOS:000919193600001 |
| 源URL | [http://119.78.100.198/handle/2S6PX9GI/35451]  |
| 专题 | 中科院武汉岩土力学所 |
| 作者单位 | 1.Department of Civil & Mineral Engineering, University of Toronto, Toronto, Ontario, Canada 2.State Key Laboratory of Geomechanics and Geotechnical Engineering, Institute of Rock and Soil Mechanics, Chinese Academy of Sciences, Wuhan, China 3.State Key Laboratory of Hydraulic Engineering Simulation and Safety, School of Civil Engineering, Tianjin University, Tianjin, China 4.Key Laboratory of Geotechnical Mechanics and Engineering, Ministry of Water Resources, Yangtze River Scientific Research Institute, Wuhan, China |
| 推荐引用方式 GB/T 7714 | Liu, Feng,Chen, Zediao,Xia, Kaiwen,et al. New strategies for developing GPU accelerated disk-based discontinuous deformation analysis for large-scale modeling[J]. INTERNATIONAL JOURNAL FOR NUMERICAL AND ANALYTICAL METHODS IN GEOMECHANICS,2023,-(-):-. |
| APA | Liu, Feng,Chen, Zediao,Xia, Kaiwen,Xu, Dongdong,&Yang, Yongtao.(2023).New strategies for developing GPU accelerated disk-based discontinuous deformation analysis for large-scale modeling.INTERNATIONAL JOURNAL FOR NUMERICAL AND ANALYTICAL METHODS IN GEOMECHANICS,-(-),-. |
| MLA | Liu, Feng,et al."New strategies for developing GPU accelerated disk-based discontinuous deformation analysis for large-scale modeling".INTERNATIONAL JOURNAL FOR NUMERICAL AND ANALYTICAL METHODS IN GEOMECHANICS -.-(2023):-. |
入库方式： OAI收割
来源：武汉岩土力学研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。