摘要
主要介绍了百万网格点规模的精细油藏数值模拟在国产高性能并行计算机与微机机群系统上的应用情况 .针对若干组来自于国内油田的百万网格点实际数据 ,给出了在多种国产并行机环境下的运行结果 ,并作了分析与评价 .在此基础上 ,讨论并行油藏数值模拟软件高效实现过程中遇到的关键技术 。
This paper reports some application instances of a finery petroleum reservoir simulation parallel software on some native high performance computers and PC cluster systems. This software was developed by Institute of Software and mainly used for solving petroleum description data more than one million grid cells. By making use of parallel computers from National High Performance Computing Environment, we have accomplished several sets of mega-griding scale simulation projects from domestic oil field. Beyond of the analysis and evaluation of the reported running results, some sticking points of the implementation process of the petroleum reservoir simulation parallel software are emphasized. Generic technical bottlenecks in the course of parallelism process of large-scale numerical software are also involved and discussed in the paper.
出处
《计算机研究与发展》
EI
CSCD
北大核心
2002年第8期973-980,共8页
Journal of Computer Research and Development
基金
国家"八六三"高技术研究发展计划基金 ( 86 3-30 6 -ZD11-0 3-1
86 3-30 6 -ZT0 6 -0 2 -1)
国家重点基础研究发展规划基金( G19990 32 80 3)
国家自然科学基金 ( 6 98830 0 6 )资助
国家高性能计算环境基金
国家基金委 NCF C超级计算机机时的资助
关键词
油藏
大型应用软件
数值模拟软件
并行模拟
高性能计算机
parallelism, petroleum reservoir simulation, large-scale application software, high performance computing environment, native parallel computers
