[visit-developers] Fwd: Question regarding invariant metadata
funkysidd at gmail.com
Mon Oct 6 18:55:28 EDT 2008
This is my own plugin. Let me the check the points below and I would let you
On Mon, Oct 6, 2008 at 3:42 PM, Mark Miller <miller86 at llnl.gov> wrote:
> Hi Siddharth,
> Hmm. Is this your own plugin that you developed or one that comes with
> VisIt distro?
> Reason I ask is that I find it hard to explain a failure of VisIt to
> call GetMesh()/GetVar(). The ONLY case where VisIt will not call these
> methods is if it thinks it already has what it needs 'cached'. In that
> case, it won't attempt to re-read objects it already has in cache.
> However, the cache is keyed, among other things, on timestep number. So,
> it won't by itself make a mistake and think something at an earlier
> timestep is the proper cached result for something at a later timestep.
> However, I can see how a new plugin developer could easily make this
> mistake if s/he was manipulating the cache directly from within the
> plugin. Do you know if that is the case?
> By telling VisIt that you DO NOT have INvariant metadata, it means VisIt
> will re-populate database metadata on every change in timestep and you
> appear to be observer that behavior is happening as expected.
> Another thing that could be happening (I am really just guessing), is
> that maybe the calls to PopulateDatabaseMetadata are failing to
> advertise the existance of the desired mesh and/or variable. So that
> when VisIt attempts to plot them, it says 'hey, that object doesn't
> exist' and never attempts to GetMesh/GetVar from the plugin. The more I
> think about it, the more this sounds like the the likely cause of your
> problems. So, examine what later calls to PopulateDatabaseMetaData are
> returning and confirm they are returning what you expect.
> On Mon, 2008-10-06 at 15:07 -0600, Siddharth Shankar wrote:
> > Hi,
> > Sent this mail a while ago. Any pointers on this.
> > Thanks,
> > Siddharth
> > ---------- Forwarded message ----------
> > From: Siddharth Shankar <funkysidd at gmail.com>
> > Date: Thu, Oct 2, 2008 at 4:03 PM
> > Subject: Question regarding invariant metadata
> > To: VisIt Developers <visit-developers at email.ornl.gov>
> > Hello,
> > I have a scenario here, mainly with AMR datasets, that the no of
> > patches change or may not change with every time step. As a result I
> > have set the functions HasInvariantMetaData() and HasInvariantSIL()
> > both to return false, so that this info is read everytime in
> > PopulateDatabaseMetaData(), while setting up the mesh.
> > However when I click on the play button on the gui to iterate through
> > the time steps, everything works fine if the no. of patches across
> > the timesteps doesn't vary. However, if this is not the case, the
> > program goes to PopulateDatabaseMetaData() (which is desired), but
> > then never enters GetMesh()/ GetVar() and a warning is flashed. As a
> > result no visualization is generated and the program just keeps on
> > skipping through the time steps. As mentioned earlier, this only
> > happens when the no. of patches across the time steps varies, and
> > everything works fine if the no. of patches/ time step remains
> > invariant.
> > Is is that I am missing some implementation. Do let me know if
> > somethng is no clear.
> > -Siddharth
> Mark C. Miller, Lawrence Livermore National Laboratory
> email: mailto:miller86 at llnl.gov
> (M/T/W) (925)-423-5901 (!!LLNL BUSINESS ONLY!!)
> (Th/F) (530)-753-8511 (!!LLNL BUSINESS ONLY!!)
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the visit-developers