Using Visualization and Animation to Convey Motion in Experimental
Data
Chapter 5: Static Visualization
Static visualization refers to the process of visualizing the state
information of objects. This involves defining the objects under study
and a finite set of states of the objects,
classifying objects by their states, extracting state information from
the original data set, choosing a proper way to display the information,
and explaining the resulting displays.
Object and state modeling is the key design part of these visualizations.
It tends to be problem specific.
In the context of parallel program performance evaluation, the objects
are processors and the states refer to processor action such as
busy, overhead, and idle.
An object in the flame system can be
a visible, concrete concept such as a cell or a flame front in the image.
It can also be an abstract object such as angle, speed, etc.
State definitions should be mutually exclusively so that any object can
be classified in one and only one state at a time.
Extracting state information from the data
is not always an easy job. It depends on object and state modeling to
a great extent. A successful data extraction filters out all
the redundant or irrelevant information while keeping all the state
information needed for visualization.
This work adapts a parallel program visualization tool called
ParaGraph for displaying the information.
The next section introduces the ParaGraph visualization tool in detail.
The remainder of the chapter discusses the application of ParaGraph
to the flame problem.
The evaluation of the static visualization technique can be found
in Chapter 7.