| 大规模时变可视化数据的预取加速研究 |
| |
| 引用本文： | 曹立强,夏芳,沈卫超. 大规模时变可视化数据的预取加速研究[J]. 华中科技大学学报(自然科学版), 2011, 39(Z1): 76-78 |
| |
| 作者姓名： | 曹立强 夏芳 沈卫超 |
| |
| 作者单位： | 北京应用物理与计算数学研究所,北京,100094 |
| |
| 基金项目： | 国家高技术研究发展计划资助项目(2009AA01A134) |
| |
| 摘 要： | 针对读取数据这一制约绘制性能的主要因素,从时变数据场的相邻帧相似性入手,利用前一帧图像的绘图区域预测下一帧图像的绘图区域,进而转换成I/O参数,实现连续绘制中的数据预取.并在多帧图像绘制过程中形成载人、处理、绘制三级流水线,提高可视化的速度.微机测试表明,由于图像相邻帧相似度的不同,时变可视化数据预取将可视化性能提高17.2％～47.5％.
|
| 关 键 词： | 可视化 知识发现 数据预取 I/O性能优化 绘图流水线 |
Speeding up visualization of large-scale time-varying data by prefetching |
| |
| Cao Liqiang,Xia Fang,Shen Weichao. Speeding up visualization of large-scale time-varying data by prefetching[J]. JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE, 2011, 39(Z1): 76-78 |
| |
| Authors: | Cao Liqiang Xia Fang Shen Weichao |
| |
| Affiliation: | Cao Liqiang Xia Fang Shen Weichao(Institute of Applied Physics and Computational Mathematics,Beijing 100094,China) |
| |
| Abstract: | As loading is a main constraint factor in visualization,considering the similarity of adjacent frames,data loading pattern in the current can be used to predict the loading pattern of the following step.With this opinion,data is read into memory of next step after data loading in current step by a prefetching thread.Then,visualization of next step can be accelerated by reading data from memory directly.Tests show that due to the difference of data sets,visual data prefetching improves the visual performance... |
| |
| Keywords: | visualization knowledge discovery fata prefetching I/O optimization visualization pipeline |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
