摘要
针对三维地形数据生成时间过长的问题,提出了一种基于集群和GPGPU技术的HPC(High Performance Computing)高性能计算方法,能够有效整合集群中多台计算机的CPU和GPU的处理能力.在三维地形数据生成过程中,该方法将地形数据进行二次细化,分别分配给GPU的每个任务管线,大幅度提高运算吞吐量,缩短三维地形数据的生成时间.
In order to solve the problem of huge time cost in the generating process of 3D terrain data ,this paper presents a HPC(High Performance Computing) method based on GPGPU and computer cluster ,which can integrate the CPUs and GPUs of the computer cluster together to provide huge computing ability .When the HPC method is used to generate 3D terrain data ,the terrain is divided into pieces with two levels of details first ,and then assigned to GPU pipelines .In this way ,the computing throughput is greatly improved and plenty of time can be saved .
出处
《微电子学与计算机》
CSCD
北大核心
2014年第10期1-4,共4页
Microelectronics & Computer
基金
921载人航天领域预先研究项目(030202)
辽宁省教育厅科研项目(L2011238)
