Aggregate dataset doesn't appear to work
The Aggregate Dataset filter appears to not function properly. Here is how to replicate:
- 5.8.0-RC2, Linux, remote server - 64 ranks.
- Open g1s1-10fps.e.16.*. All vars on. Apply.
Note - we now have 64 ranks and pvserver processes, with 16 files, in 16 processes.
- Paint by vtkProcessId.
You will see that vtkProcessId goes from 0 to 15. This makes sense.
- Merge blocks. Apply.
ProcessId still goes from 0 to 15.
- Aggregate Dataset. Set number of target processes to 4.
- Paint by vtkProcessId. Rescale to Data Range.
Everything is now running on 1 process! This is a bug.
- File/ Save Data. Save as Exodus data. Take defaults. OK.
We now have 64 files, all of which are small except rank 7 which is really big. I assume we should only have 4 files, all about the same size. This is a bug.
Edited by W. Alan Scott