PC集群环境下地图切片的并行计算方法  

Parallel computing of map tile under PC cluster

作　　者：杨轶 

机构地区：[1]云南省测绘工程院,昆明650033

出　　处：《测绘科学》2014年第3期120-123,共4页Science of Surveying and Mapping

摘　　要：WebGIS的地图服务普遍采用地图数据切片的发布策略,切片数量随着地图渲染精细程度的加大以几何级数增长,以单机或单服务器的计算方式已经无法满足地图切片生成与更新的效率需求,而构建高性能计算机集群或采用高端分布式环境增加了计算的成本和复杂度。本文以相同条件进行PC集群环境下地图切片的并行计算实验与分析,随着计算节点的加入,并行计算执行时间相对于串行计算缩短了2-6倍。实验结果表明,PC集群环境下进行地图切片的并行计算,是一种成本相对较低、具有伸缩性和提升效率的解决方法。WebGIS map service commonly adopts map tile release strategy, the amount of map tile in- creases in geometric progression with the fine degree growing of map rendering, stand-alone or single- server computing mode has been unable to meet the efficiency requirements of map tile generating and up- dating, but building high-performance computer cluster or high-end distributed environment could increase the cost and complexity of computing. In this paper, the same conditions were used for experiment and a- nalysis of parallel computing of map tile under PC cluster, the execution time by parallel computing was shortened 2-6 times relative to serial computing with the addition of computing nodes. Experimental results indicated that the method could be relative low-cost, scalability, higher efficiency in parallel computing of map tile under PC cluster environment.

关 键 词：计算机集群 地图切片 并行计算 MPI 

分 类 号：P208[天文地球—地图制图学与地理信息工程]