摘要
为减少空间降水插值的计算时间,以MPI并行接口为技术手段,采用数据划分建模方法,实现改进Kriging算法的并行算法。在Linux操作系统上搭建并行计算环境,试验数据表明,该并行算法能有效节省计算时间并具有良好的加速比、并行效率和扩展性。为Kriging插值算法的并行化实现和应用提供有意义的参考。
In order to improve the calculation speed for interpolation of Spatial Precipitation（SP）,using MPI parallel interface library and modeling approach using data partitioning,achieve parallel Kriging algorithm for interpolation of SP.The parallel computing environment is based on the Linux operating system.Experimental results show that the parallel algorithm can effectively save computing time and has a good Speedup,parallel efficiency and scalability.It also provides a meaningful reference for the parallel implementation and application of Kriging interpolation algorithm.
出处
《计算技术与自动化》
2011年第1期71-74,共4页
Computing Technology and Automation
基金
国家气象局的"精细化农业区划试点"项目
湖南省科技计划重大专项资助(2008FJ1006)
