[visit-users] Volimage plugin
Jose A. Abell M.
jaabell at ucdavis.edu
Thu Jul 23 19:41:47 EDT 2015
I fixed it! I used my other plugin as a template and changed the line(s)
that declare(s) the metadata by one call to
I think this is what you meant too.
The mesh and data now display nicely within VisIt. I'm going to add some
other features to the plugin (change it into a MTMD plugin so I can
visualize a series of files generated by the simulation, also add
visualization of displacement fields instead of just the components). If
you're interested I can contribute the new plugin to VisIt.
I'll treat it as a new plugin since the old format, I discovered, is being
used by some legacy users.
Maybe call it 'volimage2' and version it at 2.0.
Thanks for all the help!
Computational Geomechanics Group
Dept. of Civil and Environmental Engineering
On Thu, Jul 23, 2015 at 11:52 AM, Miller, Mark C. <miller86 at llnl.gov> wrote:
> Well, I was thinking the reason for the error had to do with an
> inconsistency between what the mdserver sees in the file and what the
> engine sees in the file.
> We used to print the result of metadata object after call to PopMD on
> *both* mdserver and engine vlogs. However, it would appear we no longer
> print it in engine vlogs and so I cannot know for sure if that is going on.
> I looked at your code and doen't see anything that would suggest different
> code paths for mdserver and engine (though that is possible using either
> conditional compiliation directives or some runtime calls to test if a code
> block should execute only on mdserver or only on engine).
> Bottom line, at this point I am confused as you.
> Its failing deeply inside VisIt's attempt to setup the database based on
> metadata it got. Its failing to associate mesh object and variable names.
> the metadata on the mdserver looks fine. I don't see any problems with
> it. On the engine, it *should* be using the same metadata to initialize
> some internal structures and that is failing.
> I am wondering if we're failing to set some field in the avtMeshMetaData
> object that is causing issues. Maybe you can try using one of the
> avtMeshMetaData constructors instead of just new avtMeshMetaData. For
> examples, see some other plugins that use them like Silo plugin.
> From: "Jose A. Abell M." <jaabell at ucdavis.edu>
> Reply-To: VisIt software users community <visit-users at elist.ornl.gov>
> Date: Thursday, July 23, 2015 10:01 AM
> To: VisIt software users community <visit-users at elist.ornl.gov>
> Subject: Re: [visit-users] Volimage plugin
> Thanksfor the feedback. I've been looking into what you suggested.
> Indeed the .so files were outdated. So I erased everything and recompiled.
> The issue still comes up, but now I get "Populatedatabase begins" in both
> mdserver and the engine_ser files. Here are the logs.
> Maybe I'm missing some overload somewhere? I've atached the current
> source I'm working with (only the ones I've messed around with.
> And BTW, I'm compiling on visit2.8.1
> José Abell
> PhD Candidate
> Computational Geomechanics Group
> Dept. of Civil and Environmental Engineering
> UC Davis
> On Wed, Jul 22, 2015 at 4:58 PM, Miller, Mark C. <miller86 at llnl.gov>
>> I am a bit confused by what I am seeing (or not seeing) in your logs.
>> If you look at volimage's PopulateDatabaseMetaData, it has this debug5
>> statement in it right at the top. . .
>> avtvolimageFileFormat::PopulateDatabaseMetaData(avtDatabaseMetaData *md)
>> debug5 << "Populatedatabase begins" << endl;
>> And, when I look in the mdserver . 5 . vlog file, I see the message
>> "Populatedatabase begins"
>> However, I do not see this message in engine_ser . 5 . vlog.
>> Both the mdserver and the engine *should* wind up executing
>> So, its puzzling to me why we're not seeing that message in the engine
>> logs. And, that is the crux of the problem.
>> The engine doesn't know about any objects in the file. From the logs,
>> it looks like the engine has 'succesfully' opened the file with the
>> volimage plugin. But, it looks like it didn't actually execute anything in
>> the plugin to find out about the objects there.
>> When you did 'make' for the plugin, did it indeed make all the .so's
>> for mdserver and for engine? (look in <build-dir>/plugins/databases)
>> Is the timestamp on these files today?
>> From: "Jose A. Abell M." <jaabell at ucdavis.edu>
>> Reply-To: VisIt software users community <visit-users at elist.ornl.gov>
>> Date: Wednesday, July 22, 2015 4:03 PM
>> To: VisIt software users community <visit-users at elist.ornl.gov>
>> Subject: Re: [visit-users] Volimage plugin
>> On Wed, Jul 22, 2015 at 3:33 PM, Miller, Mark C. <miller86 at llnl.gov>
>>> If all that has changed is the file extension, that is an easy thing to
>>> deal with.
>> So apparently the format has changed since the plugin was implemented
>> in 2011. I'm trying to reimplement the plugin so that it reads the new
>> format. I can open the file, but when I try visualizing the mesh
>> (Rectilinear grid) I get a complaint from mdserver.
>> "Was not able to match up griddy with any of the existing top sets."
>> Here "griddy" is a new name I gave the mesh, which I changed to confirm
>> that the message was related to this name. I've attached the results from
>> running with the "-debug 5" option.
>> The format is pretty simple, I think I can implement it and send it over
>> to you guys. But I'm lost with this error.
>> José Abell
>> PhD Candidate
>> Computational Geomechanics Group
>> Dept. of Civil and Environmental Engineering
>> UC Davis
>> VisIt Users Wiki: hxxp://visitusers.org/
>> Frequently Asked Questions for VisIt: hxxp://visit.llnl.gov/FAQ.html
>> To Unsubscribe: send a blank email to
>> visit-users-unsubscribe at elist.ornl.gov
>> More Options: hxxps://elist.ornl.gov/mailman/listinfo/visit-users
> VisIt Users Wiki: hxxp://visitusers.org/
> Frequently Asked Questions for VisIt: hxxp://visit.llnl.gov/FAQ.html
> To Unsubscribe: send a blank email to
> visit-users-unsubscribe at elist.ornl.gov
> More Options: hxxps://elist.ornl.gov/mailman/listinfo/visit-users
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the visit-users