Recent Advances in VisIt: AMR Streamlines and Query-Driven Visualization Page: 5 of 6
This article is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to Digital Library by the UNT Libraries Government Documents Department.
The following text was automatically extracted from the image on this page using optical character recognition software:
Recent Advances in VisIt: AMR Streamlines and Query-driven Visualization 5
enters a region in a patch that is also available at higher resolution. Instead,
a streamline is calculated based on the samples of a patch until it exits its
To remedy this problem, we are working on an implementation that detects
when streamlines enter a refined region (Deines et al. 2009). This algorithm ter-
minates streamline calculations in a lower resolution patch when the streamline
enters a region that is available at a higher resolution and restarts calculations
in the refining patch, see Fig. 2 (right). For the purpose of detecting this transi-
tion, we are considering two schemes. The first scheme does not use patch layout
information explicitly. Instead, it checks during streamline integration whether
the currently considered cell is superseded by a finer representation. If it is, it
uses the adaptive step size control of the integration scheme to reduce terminate
the streamline close to the level boundary and restarts calculation in a second
patch. Our second scheme explicitly uses the bounding boxes of levels to identify
when a streamline enters a refining level. It turns out that this approach leads
to more accurate results. For both schemes, we use ghost cells to interpolate
values at level boundaries. Currently, we assume that the simulation produces
meaningful ghost data. For a more detailed discussion, we refer the reader to
Deines et al. (2009).
3. Visualization of Large-scale Particle Simulations with FastBit
In last year's Astronum proceedings, we gave examples how the combination of
VisIt and FastBit facilitates data discovery in large scale particle data sets (Bethel
et al. 2009). In this prototype, the H5Part file reader (Adelmann et al. 2007),
uses FastBit (Wu, Otoo, & Shoshani 2006) to create an index for the depen-
dent variables in large scale particle simulations. The underlying idea is that
when considering billions of particles, not all of these particles are of interest.
Instead, one usually focuses only on a subset of these particles, e.g., based on
thresholding, and analyzes the behavior of these particles.
The use case in last year's paper (Bethel et al. 2009), for example, focused
on Laser-Wakefield particle acceleration. Here, particles of interest included
those particles with an momentum in x-direction larger than a given threshold
within a particular time step. Subsequently, one only considered these particles.
If a particle file in H5Part format contains an index, VisIt will now communicate
the settings of the threshold operator to the file reader via VisIt's contract based
scheme (Childs et al. 2005). The reader can utilize the FastBit index to load
only particles that satisfy the thresholding criteria. (While currently only the
H5Part file reader supports FastBit indices, VisIt has the infrastructure in place
to communicate thresholds to any custom file readers, so it is possible for users
to write their own file format and possibly utilize FastBit. We are considering
adding this functionality for particle data in Chombo files.)
Furthermore, VisIt supports now the concept of named selections. If one
selects particles in a given time step, e.g., via the threshold operator, it is now
possible to create a named selection comprising these particles. Using this named
selection it is possible to restrict consideration in other time steps to the same
set of particles as well. For example, one can use this functionality to compute
the traces of particles that exceed a certain momentum in a given time step.
Here’s what’s next.
This article can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Article.
Weber, Gunther; Ahern, Sean; Bethel, Wes; Borovikov, Sergey; Childs, Hank; Deines, Eduard et al. Recent Advances in VisIt: AMR Streamlines and Query-Driven Visualization, article, November 12, 2009; Berkeley, California. (https://digital.library.unt.edu/ark:/67531/metadc1013741/m1/5/: accessed March 23, 2019), University of North Texas Libraries, Digital Library, https://digital.library.unt.edu; crediting UNT Libraries Government Documents Department.