Login
Research at the Institute of Data Analysis and Visualization
Jump to:
Query-Driven Visualization of Time-Varying Adaptive Mesh

Luke Gosink, John C. Anderson, E. Wes Bethel, and Ken Joy


Abstract

Image

We present a new approach that enables query-driven analysis and multitemporal visualization of time-varying AMR data. Previously, such analysis and visualization efforts were hindered by the dynamic temporal and spatial properties of AMR grid hierarchies. We present a two-step method for compositing and synchronizing AMR data from a series of timesteps. We first generate a composite template from the AMR grid hierarchies of these timesteps; the composite template preserves the finest level of grid cell refinement from each grid hierarchy. We then synchronize each timestep's grid hierarchy to the composite template. This approach enables our method to process queries on a common AMR grid hierarchy. Using this data structure, we move the work of query processing to the GPU to realize the benefit of greatly accelerated QDV analysis. On the GPU side, we integrate our new method with a GPU-based query engine, called the Bin-Hash index.

Our method facilitates query-driven analysis of time-varying AMR data, and generates two types of time-dependent visualization: temporally sequential and temporally concurrent. In temporally sequential visualizations, features from each timestep are analyzed and visualized individually in sequential frames as an animation. Comparatively, in temporally concurrent visualizations, a single multitemporal image conveys how queries characterizing important features evolve over time. In temporally sequential visualizations, our GPU-based QDV engine enables accelerated analysis; users can process queries over multiple time steps and view the results in real-time as an animation.

Introduction

Computational simulation has become an essential and powerful tool impacting a diverse group of scientific disciplines such as engineering, biology, and medicine. Detailed simulations that model time-dependent, continuous physical phenomena, along with analysis and visualization tools that address the temporal aspects of these simulations, are essential to generate new understanding and insight into many domain-specific problems.

In scientific simulations, the immense size and sheer complexity of data generated from highly-detailed numerical methods has popularized the use of adaptive mesh refinement (AMR) strategies. In numerical simulations, AMR-based techniques adaptively refine the domain space of a simulation, both spatially and temporally, into a hierarchy of nested, sequentially refined grids. Though these strategies are computationally efficient and provide significant storage benefits, the dynamic aspects of the grid hierarchies pose significant challenges for visualization methods

In this work, we address the challenges of using a query-driven visualization (QDV) approach to visualize time-varying AMR data. We present a two-step method for compositing and synchronizing AMR data from a series of timesteps. We first generate a composite template from the AMR grid hierarchies of these timesteps; the composite template preserves the finest level of grid cell refinement from each grid hierarchy. We then synchronize each timestep's grid hierarchy to the composite template. This approach enables our method to process queries on a common AMR grid hierarchy. Using this data structure, we move the work of query processing to the GPU to realize the benefit of greatly accelerated QDV analysis. On the GPU side, we integrate our new method with a GPU-based query engine, called the Bin-Hash index (the project page for the Bin-Hash index may be found here).

Contribution

The main contributions of this work include:

  • We develop a new framework for doing QDV processing and visualization of time-varying AMR data. The core of this method is based upon a synchronization strategy that addresses the disparities in spatial refinement that exist between any series of timesteps in an AMR-based simulation.
  • We demonstrate the first GPU-based QDV approach that utilizes a GPU-based indexing strategy to accelerate query processing, efficiently utilize GPU memory, and accelerate QDV methods.

Method

The following images depict our two step process that facilitates query-driven visualization of time-varying AMR data.

The figure at left illustrates the sequential process of compositing the AMR grid hierarchies of two selected timesteps. The process begins by filling the composite template with all grid cells, from both timesteps, of the finest level of refinement. In each subsequent pass, our procedure adds grid cells of the next level of lesser refinement to the template - conditioned on the basis that a more finely refined grid cell has not already been placed at that position. Finally, we add grid cells of the coarsest level of refinement to the template.

The figure at right depicts the sequential process of synchronizing the grid hierarchy of a given timestep with a composite template. At each level of synchronization, grid cells conditionally refine themselves by one additional level according to whether or not they are synchronized with the composite template. In this example, synchronization is complete for the grid hierarchy in the second level of synchronization.

Results:

This series of images, selected from 48 timesteps, compares query results from non-sychronized (top row), and sychronized (bottom row) AMR grids of the Hurricane Isabel dataset. The query used on each timestep consists of two parts; we query for regions of low pressure $(-200 \leq \emph{pressure} \leq 20)$ OR regions of high pressure $(500 \leq \emph{pressure} \leq 1000)$.

This multitemporal image depicts summary statistic information gathered from queries processed over 48 timesteps of the Hurricane Isabel WRF Model dataset. In this multitemporal image, hurricane direction and velocity information are conveyed by querying for regions of low-pressure.

Publications