MantisBT - ParaView
View Issue Details
IDProjectCategoryView StatusDate SubmittedLast Update
0015377ParaView(No Category)public2015-03-12 10:572015-09-06 12:17
ReporterUtkarsh Ayachit 
Assigned ToUtkarsh Ayachit 
PrioritynormalSeverityminorReproducibilityhave not tried
StatusclosedResolutionfixed 
PlatformOSOS Version
Product Version 
Target Version4.4Fixed in Version4.4 
ProjectSandia
Topic Name
Typeincorrect functionality
Summary0015377: Surface selection broken for AMR datasets
DescriptionSteps to reproduce:

+ Create HierarchicalFractal, Apply
+ Extract Surface, Apply
+ Select surface cells -- they are often incorrect.

The issue seems to be the fact that the CompositeID returned by the selection is incorrect.
Steps To Reproduce
Additional Information
TagsNo tags attached.
Relationships
related to 0015437closed Utkarsh Ayachit Multipiece dataset selection tools select data on the first piece only 
Attached Files
Issue History
Date ModifiedUsernameFieldChange
2015-03-12 10:57Utkarsh AyachitNew Issue
2015-03-12 10:57Utkarsh AyachitTarget Version => 4.4
2015-06-22 09:13Utkarsh AyachitRelationship addedrelated to 0015437
2015-06-22 09:13Utkarsh AyachitNote Added: 0034591
2015-06-22 09:13Utkarsh AyachitStatusbacklog => customer review
2015-06-22 09:13Utkarsh AyachitFixed in Version => git-master
2015-06-22 09:13Utkarsh AyachitResolutionopen => fixed
2015-06-22 09:13Utkarsh AyachitAssigned To => Utkarsh Ayachit
2015-06-25 21:53Alan ScottNote Added: 0034615
2015-07-09 20:58Alan ScottNote Added: 0034730
2015-07-09 20:58Alan ScottStatuscustomer review => todo
2015-07-14 12:49Alan ScottNote Added: 0034774
2015-07-14 12:49Alan ScottStatustodo => closed
2015-09-06 12:17Utkarsh AyachitFixed in Versiongit-master => 4.4

Notes
(0034591)
Utkarsh Ayachit   
2015-06-22 09:13   
This was fixed by BUG 0015437.
(0034615)
Alan Scott   
2015-06-25 21:53   
Note - Following Utkarsh's fix, this appears incorrect parallel server.
(0034730)
Alan Scott   
2015-07-09 20:58   
See e-mail notes sent to Utkarsh.
(0034774)
Alan Scott   
2015-07-14 12:49   
There are two issues here - is the AMR selection fixed, and is the Hierarchical Fractal working correctly in parallel.

We believe there is a problem with the Hierarchical fractal code in parallel. This isn't this bug.

We believe that the AMR selection is fixed (which is this bug).

From an e-mail from Utkarsh:

So the issue is weird. It so happens that the Hierarchical Fractals produces incorrect image extents on the last rank when running in parallel. Consequently, when ExtractSurface it applied, it produces the lowest res grid to cover all partitions on the last rank (see attached image).

When selecting, the last process wins and hence it selects the lower-res cell from the rank #3 (I was running 4 procs here, but same is true for 2 procs).

If you don't apply "Extract Surface" but directly switch to "Surface" representation on the Fractal itself and then select it doesn't run into this issue.


Closing the bug.