| 超算环境下的并行切图性能分析 |
| |
| 引用本文： | 裴斐,杨耀环,刘怡菲.超算环境下的并行切图性能分析[J].测绘与空间地理信息,2017,40(Z1). |
| |
| 作者姓名： | 裴斐 杨耀环 刘怡菲 |
| |
| 作者单位： | 1. 北京超图软件股份有限公司郑州分公司,河南 郑州,450001;2. 郑州市数字城市办公室,河南 郑州,450007 |
| |
| 摘 要： | 在传统GIS二维地图缓存生成过程中,一般用普通的PC机来生成地图缓存.普通PC机的内存一般不大于8 G,CPU主频一般不大于2.2 GHz,且核数不大于8核,硬盘一般为机械硬盘,容量不大于2 T.受限于硬件性能的瓶颈,切图效率较低,不能够分析出上述各硬件因素中制约二维地图缓存性能的关键性因素.相比之下,超算环境有着丰富的资源,可将上述任一资源延伸到极致,因此,本文借助超算环境,研究分别增大内存、CPU主频、硬盘空间等因素后对地图缓存效率的影响.基于此,本研究在郑州大学超级计算机中心专门进行了比较测试.郑州大学超级计算机中心拥有两种类型的CPU,每个主机拥有128 G内存,CPU分为普通CPU和KNL CPU,分别拥有32核和64核,所有主机共享802 T的磁盘阵列,可以认为对于二维地图缓存来说硬件资源是无限的.本文利用郑州大学超级计算机中心的理想环境分析出了在内存和CPU等因素都不受限的情况下,影响二维地图缓存效率的关键因素.
|
| 关 键 词： | 地图缓存 切图性能 超算 |
An Analysis of the Efficiency of Parallel Map Cache in Supercomputer Environment |
| |
| PEI Fei,YANG Yao-huan,LIU Yi-fei.An Analysis of the Efficiency of Parallel Map Cache in Supercomputer Environment[J].Geomatics & Spatial Information Technology,2017,40(Z1). |
| |
| Authors: | PEI Fei YANG Yao-huan LIU Yi-fei |
| |
| Abstract: | When processing map cache by traditional method, we often use the normal PC to cache the map. The normal PC has at most 8G memories, and the CPU rate is less than 2. 2GHz, and the CPU's core number is less than 8. As the efficiency is limited to the hardware bottleneck, issues related with effect factor of map cache are hard to be analyzed. Compared with PC, supercomputer en-vironment has abundant hardware resources. Any of these resources can be extended beyond the limit. We analyzed the influence fac-tor by increasing each hardware condition including memory size, CPU rate, and hard disk space, and by changing the hard disk types using Zhengzhou University Supercomputer Center. The Center has two type of CPUs of normal CPU and KNL CPU with 32kerneland 64Kernel, respectively. For each computer, it has two CPU and each CPU has 128G memory. All computers share the 802T disk ar-ray. Such hardware condition is sufficient for two-dimensional map cache. We analyzed the key issues of map cache without the limit of CPU and memory by using the perfect environment of the Zhengzhou University Supercomputer Center. |
| |
| Keywords: | map cache the efficiency of map caching supercomputer |
| 本文献已被 万方数据 等数据库收录！ |
|
