[Paraview] seg fault paraview 3.0.2

[kmorel]

The problem is that when you first load in the file and the image is first
displayed, the surface is already extracted to display that.  That extra
large surface will stick around in memory afterwards.

In this case, it would be OK to run clean to grid, write out the result (if
you can get that far), and then load in the resolved grid.  If that is
working for you fine, but anything larger will probably fail.  A better
solution is to write out the data with proper connectivity in the first
place.  If possible, use cell data to model the discontinuous nature rather
than point data.

-Ken


On 11/6/07 1:29 PM, "Sreejith Pulloor" <sreejithpk at gmail.com> wrote:

> Thank you very much, I forgot to mention that, it is a discontinuous
> finite  (non-conforming) field  data. so what I  normally do
> is
> -> first do a clean to grid filter. so that the ''double-d' points
> will be gone.
> -> then I do the slice filter on the cleaned-grid filter. and there i
> get the segfault
> 
> Do you think that is also possible.
> 
> thanks,
> Sreejith
> 
> On Nov 6, 2007 8:34 PM, Moreland, Kenneth <kmorel at sandia.gov> wrote:
>> Sreejith,
>> 
>> I figured out the problem.  Your data is written inefficiently.  Each
>> cell points to its own set of points.  Instead, cells should be sharing
>> points.  Because no cells are sharing points, ParaView thinks that every
>> face of every cell is an external face.  The first thing that ParaView
>> does is to extract all the external faces to render them, and this
>> bloats the memory significantly.  After rendering the memory consumption
>> is near the 2 GB limit for 32 bit applications, and ParaView is crashing
>> because it cannot allocate memory.
>> 
>> Change your program to write out data so that cells share points and
>> everything should work fine.
>> 
>> -Ken
>> 
>>> -----Original Message-----
>>> From: Sreejith Pulloor [mailto:sreejithpk at gmail.com]
>>> Sent: Tuesday, November 06, 2007 2:28 AM
>>> To: Moreland, Kenneth
>> 
>>> Subject: Re: [Paraview] seg fault paraview 3.0.2
>>> 
>>> Dear Sir,
>>> 
>>> Here is a vtk xml unstructured grid binary  appended raw format file.
>>> I can really manage the file with paraview 2.6.2 version. But
>>> unfortunately wiht paraview 3.0.1 and 3.0.2 stable version (linux),
>>> the slice and clip things are not working. sometimes it exit with
>>> segfault , sometimes not but cannot see the slice or clip. that
>>> produced.
>>> what could be the problem. please find the file here:
>>> 
>>> http://hal.iwr.uni-heidelberg.de/~sreejith/flow3d.vtu   (rather big
>> file
>>> 186 MB)
>>> 
>>> please note, i dont have this problem with paraview 2.6.2. working
>>> great with that.
>>> 
>>> thank you
>>> Sreejith
>>> 
>>> On Sep 26, 2007 4:45 PM, Moreland, Kenneth <kmorel at sandia.gov> wrote:
>>>> I cannot replicate your problem.  Can you give some more details on
>> what
>>>> you are doing?
>>>> 
>>>> Note that when you apply the clip filter, it also tetrahedralizes
>> your
>>>> data and converts it into an unstructured grid.  This can
>> dramatically
>>>> increase the memory footprint (possibly by 2 orders of magnitude if
>> the
>>>> original data only had one array on it).  If you are applying a clip
>> to
>>>> a 3D array of over a million points total, you may simply be running
>> out
>>>> of memory.
>>>> 
>>>> Note that this is not an issue for the slice filter.
>>>> 
>>>> -Ken
>>>> 
>>>> 
>>>>> -----Original Message-----
>>>>> From: paraview-bounces+kmorel=sandia.gov at paraview.org
>>>> [mailto:paraview-
>>>>> bounces+kmorel=sandia.gov at paraview.org] On Behalf Of Sreejith
>> Pulloor
>>>>> Sent: Wednesday, September 26, 2007 9:29 AM
>>>>> To: paraview at paraview.org
>>>>> Subject: [Paraview] seg fault paraview 3.0.2
>>>>> 
>>>>> Hello,
>>>>> very often when I try to clip or slice a 3d rectangular geometry
>> data,
>>>>> the paraview (3.0.2 stable unix version) get exited.
>>>>> /bin/paraviewstable: line 2: 17123 Segmentation fault
>>>>> $HOME/paraview-3.0.2-stable/bin/paraview
>>>>> any help
>>>>> thank you
>>>>> _______________________________________________
>>>>> ParaView mailing list
>>>>> ParaView at paraview.org
>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>> 
>>>> 
>>>> 
>> 
>> 
>> 
>