Re: [Paraview] vtkCompositeDataSet - memory ?

2009-02-23 Thread Jean Favre
Andy Bauer wrote: > ...They should also have the same vtkPointData as a result so that > data is not duplicated. > In my own experience, it is often the case that different Data arrays can be defined on the 3D blocks and on the 2D boundary surfaces. I just wanted to point out that ParaView is smar

Re: [Paraview] vtkCompositeDataSet - memory ?

2009-02-21 Thread Jacques Papper
Thank you, This is what I coded up (using setpoints for each vtkUnstructuredGrid with the same vtkPoints). I was then just surprised to see that the number of points in ParaView was multiplied by my number of blocks and therefore believed the memory used was multiplied. Is there any way of making

Re: [Paraview] vtkCompositeDataSet - memory ?

2009-02-20 Thread Andy Bauer
If all of the grids that are used in the vtkMultiBlockDataSet derive from vtkPointSet, you can use the same vtkPoints for each of the grids. They should also have the same vtkPointData as a result so that data is not duplicated. Andy On Fri, Feb 20, 2009 at 2:23 PM, Jacques Papper wrote: > Hi

Re: [Paraview] vtkCompositeDataSet - memory ?

2009-02-20 Thread Moreland, Kenneth
You can share the points by simply creating one vtkPoints object and then passing those points to all the unstructured grids in your data object. Then all the blocks will be sharing that data. Note, however, that ParaView will not actually check that. Thus it will probably incorrectly count t

[Paraview] vtkCompositeDataSet - memory ?

2009-02-20 Thread Jacques Papper
Hi all, I'm writing a reader for a dataset which has boundary faces (tri and quads) as well as volume cells (tets etc..). In would like to use a CompositeDataSet as a container, so that I can easily extract boundary faces which are groupesd in named zones (for ex: upper_surf lower_surf etc...).