The Time Slice Selection Bake-off

Loading...
Thumbnail Image

Date

2022-10-04

Authors

Kawakami, Yuya

Journal Title

Journal ISSN

Volume Title

Publisher

University of Oregon

Abstract

As the size of data from scientific simulations grows, the ability to identify key time steps in a simulation has emerged as a key challenge. In response, a number of time slice selection methods and algorithms have been proposed. However, no past work has performed a comparative analysis of selection methods as well as their evaluation metrics. This thesis presents results from quantitative and qualitative study of selection methods and evaluation metrics to fill thisgap. Our work has three major thrusts. First, we identify similarities and dissimilarities between different time slice slice selection algorithms. Second, we evaluate conditions under which these methods may fail. Third, we also perform a comparative study with evaluation metrics to investigate how selection methods perform under the set of evaluation metrics in literature. In all, this thesis aims to understand the space of time slice selection methods, to inform future research in this area.

Description

Keywords

High-Performance Computing, In Situ Processing, Scientific Visualization

Citation