大规模数值模拟数据的多分辨组织

摘要

Large-scale simulation data challenges the ability of visualization, and our research is to reorganize large-scale data into multiresolution representation and improve the interactive performance of visualization. Considering the data block size produced by large-scale simulation mismatches the need of visualization, this paper introduces an algorithm of I/O opti-mized multiresolution data organization, and it works on the two storage schemes: multi-level and hierarchy. This technique has been tested in practice data and obtains good interactive visualization performance.%大规模数值模拟数据对可视化分析提出了挑战,本文研究如何将大规模数据组织成多分辨的形式以提高大规模数据可视化的交互性能.鉴于大规模数值模拟剖分的数据块粒度与可视化分析优化的I/O粒度不匹配,本文提出了I/O优化的多分辨数据组织算法,并实现了多层存储与多级存储两种存储方案.使用实际的大规模数据进行了测试并取得了很好的效果.