Hi WT, You may be running into an issue that can come up with the OpenGL2 renderer in ParaView 5. Please see section 2.2.3 in the GeantToVTK plugin page: https://cdcvs.fnal.gov/redmine/projects/geanttovtk/wiki/Wiki
I have quoted that section here for convenience: [begin quote] The output of the GeantToVTK plugin is a vtkMultiBlockDataSet, which is a hierarchical structure of vtkPolyData objects. ParaView version 5 has a new renderer using OpenGL2 that is extremely fast and is used when a vtkMultiBlockDataSet is homogeneous. A vtkMultiBlockDataSet is homogeneous when all of the constituentvtkPolyData objects within have the same list of attributes. That is if one vtkPolyData object has a Normals vector, then all of the objects should. Or none at all should. AllvtkPolyData objects need to have the same list of attributes (e.g. scalar point data). The vtkMultiBlockDataSet from GeantToVTK is heterogeneous. Some vtkPolyData have Normal vectors, others don't. Some have particular attributes, like color and density; others don't. This situation occurs because the vtkMultiBlockDataSet follows the Geant4 visualization output. Some Geant objects have particular attributes and others don't. The consequence is that a slow generic renderer is used. For a complicated geometry, the generic renderer could be so slow as to make ParaView unusable. Note that this situation only affects ParaView v5. ParaView v4 uses an OpenGL1 renderer. The fast OpenGL2 renderer can be many times faster though, and so it is useful to upgrade to ParaView v5. The Normalize a multi-block dataset filter makes a heterogeneous vtkMultiBlockDataSet homogeneous by adding missing attributes to the vtkPolyData with appropriate defaults. The resulting vtkMultiBlockDataSet should trigger the fast renderer. This fast renderer is anywhere from 20x - 200x faster than the generic renderer. [end quote] Would you mind trying this "Normalize a multi-block dataset" filter and see if that improves the animation speed? Thanks, Cory On Fri, Mar 4, 2016 at 5:43 AM, wturner <[email protected]> wrote: > Hi Cory, > > for the geometry VTM (attached) looks to be ASCII and the VTP is base64 > (random attached, but different VTP's have different lengths of base64 > stuff). > for the events, they are stored as .heprep (again, attached) then loaded in > using a GeantToVTK plugin to a Multi-block Dataset. > > I didnt build from source, just using the binaries on SL6. > > Thanks, > WT > > > > On 03-03-2016 22:13, Cory Quammen wrote: >> >> WT, >> >> That's a decent number of cells and points, but not super big. Can you >> provide details on the file format in which you are storing the data? >> VTK legacy file in ASCII encoding would be slow to load, for example. >> >> Did you build ParaView in debug mode, by change? That may also explain >> the slow data loading. >> >> Is there any sample data you can provide? >> >> Thanks, >> Cory >> >> On Thu, Mar 3, 2016 at 4:05 PM, wturner <[email protected]> wrote: >>> >>> Hi Cory, >>> >>> Yes, I've had this problem in previous versions. I am currently using >>> Paraview v5 and previously used 4.3.1. >>> >>> Geometry is 1347798 Cells, 3909201 Points and takes up 3.6e+02 MB in >>> memory. >>> Events are around 3039 Cells, 3040 Points and 0.19 MB memory >>> >>> Not sure if it's important but it takes a similar amount of time if I'm >>> running locally on my 4core desktop, or connected to a pvserver using a >>> 980Ti for rendering. >>> >>> Thanks, >>> WT >>> >>> >>> On 26-02-2016 19:45, Cory Quammen wrote: >>>> >>>> >>>> Hi WT, >>>> >>>> Is this a new problem or is this the first time you have tried this. >>>> If this is a new problem, in which version of ParaView did it work >>>> quickly and in which version of ParaView did it start? If this is your >>>> first time trying this, how large are your geometries you are trying >>>> to display? >>>> >>>> Thanks, >>>> Cory >>>> >>>> On Fri, Feb 26, 2016 at 12:49 PM, wturner <[email protected]> >>>> wrote: >>>> >>>>> Hi all, >>>>> >>>>> This should be an easy one. I've got GEANT4 exported geometry and >>>>> events. I've got a plugin to convert from GEANT to VTK I cut down >>>>> the geometry to what I need, save it as VTM and use this in the >>>>> future. If I load the HEPREP.zip events without the geometry I can >>>>> press play and flip though event by event quickly, ~10ms per event.. >>>>> But if I try and flip through events with geometry it takes >>>>> ~10-20seconds per event. >>>>> >>>>> I've tried right clicking on the geom and selecting "ignore time", >>>>> but no luck. Can't find anything in the documentation. >>>>> >>>>> Thanks, >>>>> WT >>>>> _______________________________________________ >>>>> Powered by www.kitware.com [1] >>>>> >>>>> Visit other Kitware open-source projects at >>>>> http://www.kitware.com/opensource/opensource.html [2] >>>>> >>>>> Please keep messages on-topic and check the ParaView Wiki at: >>>>> http://paraview.org/Wiki/ParaView [3] >>>>> >>>>> Search the list archives at: http://markmail.org/search/?q=ParaView >>>>> [4] >>>>> >>>>> Follow this link to subscribe/unsubscribe: >>>>> http://public.kitware.com/mailman/listinfo/paraview [5] >>>> >>>> >>>> >>>> -- >>>> >>>> Cory Quammen >>>> R&D Engineer >>>> Kitware, Inc. >>>> >>>> Links: >>>> ------ >>>> [1] http://www.kitware.com >>>> [2] http://www.kitware.com/opensource/opensource.html >>>> [3] http://paraview.org/Wiki/ParaView >>>> [4] http://markmail.org/search/?q=ParaView >>>> [5] http://public.kitware.com/mailman/listinfo/paraview >>> >>> >>> > -- Cory Quammen R&D Engineer Kitware, Inc. _______________________________________________ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView Search the list archives at: http://markmail.org/search/?q=ParaView Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
