Histogram-Based I/O Optimization for Visualizing Large-Scale Data
|Title||Histogram-Based I/O Optimization for Visualizing Large-Scale Data|
|Publication Type||Conference Proceedings|
|Year of Publication||2009|
|Authors||Hong, Y, Peterka, T, Shen, HW|
|Conference Name||SuperComputing 2009 (SC 2009)|
|Conference Location||Portland, OR|
We present an I/O optimization method for parallel volume ren- dering based on visibility and spatial locality. The combined met- ric is used to organize the file layout of the dataset on a paral- lel file system. This reduces the number of small, noncontiguous I/O operations and improves load balance among I/O servers. The net result is reduced I/O time. Since large-scale visualization is data-intensive, overall visualization performance improves using this method. This paper explains the preprocessing of data blocks to compute feature vectors and the storage organization based on them. Run-time performance is analyzed with a variety of transfer functions, view directions, system scales, and datasets. Our re- sults show significant performance gains over file layouts based on space-filling curves.