RedistributeDataSet filter froze for CONVERGECFD Reader

I have a CONVERGECFD data set, I am loading it in ParaView Version 5.8.1 mpi version. When I am using RedistributeDataSet filter, it froze after progressing to 25%.

In the attachment, you will find the data.

Any suggestions?

attachment update: I am not able to upload the data file. its size is 267 MB. Is there any way I can upload that kind of big file?

Sure, use google drive or wetransfer.

Hi @mwestphal: Can you check if you are able to access the data?

Data Link:

Yes, I have access.

I am able to reproduce your issue. This would require further investigation to understand what is going on.

Thanks @mwestphal.

Should I submit an issue at GitLab? ( So the developer can track the issue.

Yes please. Ideally it would be nice to understand what is special about your dataset and makes ParaView crashes.

Was this ever reported? I’ve been looking at other issues related to this filter, I’d like to make sure any other known issues are addressed too.

From what I can tell, it’s the vtkCellCenters filter that’s used internally that’s taking an inordinate amount of time when computing cell-centers for polyhedral cells.

It doesn’t appear to have been reported in the issue tracker. Since the reader and filter are in VTK, I reported an issue on the VTK tracker:

@cory.quammen: Thanks for reporting.

@utkarsh.ayachit: I have found other issues when using filters with CONVERGE CFD data set:

  1. Tetrahedralize: When I use Tetrahedralize filters, it progresses up to 60 %, then without showing any error message it crashes.

  2. Extract Surface: When I use extract surface filters, the surface is choppy, not smooth. See the image below


It’s a little hard to tell from the image, but my guess is that the surface is not smooth because no surface normals are available. You can try to use the Generate Surface Normals filter to make the surface appear smoother.

thanks @cory.quammen.

I have tried using Generate Surface Normal, doesn’t improve. As you can see from the attached image below. I agree it’s hard to tell from the image. I was comparing Extract Surface image with Extract Block image. You can see extract block is much smoother. Version 5.8.1 Extract block has some issues when used a python script.

Image using Generate Surface Normal:

Image using Extract Block: