摘要
在大规模科学计算中,核外计算的应用越来越广泛,而核外计算时间优化的关键是I／O执行 时间的长短．通过对I/O优化方法的研究,实现了数据筛选、数据预取、数据重用三种I／O优化策略．实 验结果表明这三种优化策略均能隐藏I／O延迟,优化核外计算的系统性能．
The application of out-of-core computing is more and more broad in the large scientific computing. The key of time optimization in the out-of-core computing due to the time of I/O performs. This paper mainly studies the optimizing methods I/O optimizations which including data sieving, data prefetching and data reuse. Several sets of experimentations showed that the I/O performance is efficiently improved and enhance the holistic performance of out-of-core computing by using these optimization methods.
出处
《哈尔滨商业大学学报（自然科学版）》
CAS
2005年第5期600-603,共4页
Journal of Harbin University of Commerce：Natural Sciences Edition
基金
863计划(2001AA147010)十五国防预研(41316.1.1)项目.
关键词
核外计算
数据筛选
数据预取
数据重用
out-of-core computing
filter data
data prefetching
data reuse
