| View Issue Details [ Jump to Notes ] | [ Print ] | ||||||||
| ID | Project | Category | View Status | Date Submitted | Last Update | ||||
| 0015397 | ParaView | (No Category) | public | 2015-03-26 17:27 | 2015-12-15 19:25 | ||||
| Reporter | Alan Scott | ||||||||
| Assigned To | Dan Lipsa | ||||||||
| Priority | normal | Severity | minor | Reproducibility | have not tried | ||||
| Status | closed | Resolution | fixed | ||||||
| Platform | OS | OS Version | |||||||
| Product Version | 4.3 | ||||||||
| Target Version | 4.4 | Fixed in Version | 5.0 | ||||||
| Summary | 0015397: Resample dataset doesn't work multi block | ||||||||
| Description | Resample datatset doesn't work correctly when the source is multiple block. * Linux, 4.3.1, local server. * Open disk_out_ref.exo. All vars on. Apply. * Open can.exo. All vars on. Apply. * Resample with dataset. Make disk_out_ref.exo the Input, can.exo the Source. Apply. Minor complaint - the visibility of the two input files doesn't go off. This is confusing. Please fix it. Major issue - Lots of warnings and errors, and nothing is displayed. Error is as follows: Warning: In /projects/viz/paraview/src/ParaView/4.3.1a/ParaView4/VTK/Common/DataModel/vtkDataSet.cxx, line 421 vtkUnstructuredGrid (0x8f9d550): Point array vtkValidPointMask with 1 components, has 7157 tuples but there are only 800 points ERROR: In /projects/viz/paraview/src/ParaView/4.3.1a/ParaView4/VTK/Common/DataModel/vtkDataSet.cxx, line 413 vtkUnstructuredGrid (0x3bc3fa0): Point array vtkValidPointMask with 1 components, only has 3364 tuples but there are 6724 points | ||||||||
| Tags | No tags attached. | ||||||||
| Project | Sandia | ||||||||
| Topic Name | |||||||||
| Type | incorrect functionality | ||||||||
| Attached Files | |||||||||
| Relationships | |
| Relationships |
| Notes | |
|
(0034327) Alan Scott (manager) 2015-03-26 17:30 |
Although not acceptable as a fix, a hack that gets around this bug is to run the merge blocks filter on the Source. |
|
(0034377) Alan Scott (manager) 2015-04-29 18:32 |
The hack listed here does not work for my user - he has a few dozen blocks, different materials, and needs to resample onto those blocks. Thus, I am marking this for PV 4.4. Also, I have an excellent dataset that shows this issue. Kitware may have it if necessary. Alan - Test using representative dataset. Dataset is on Windows, ..../resampleDatasetNew/ donor.g and recipient.g. |
|
(0034725) Dan Lipsa (developer) 2015-07-09 15:10 |
I don't get the warning with the latest master but I do get the error. A way to simplify this test is to use wavelet instead of disk_out_ref.exo. The same error is printed, so indeed the problem is when the source is multiblock. |
|
(0034747) Dan Lipsa (developer) 2015-07-10 15:37 edited on: 2015-07-14 14:16 |
=============== VTK ========================== commit f905347019816d21b212eb0a36286aa93d2eb8df Author: Dan Lipsa <dan.lipsa@kitware.com> Date: Fri Jul 10 14:26:05 2015 -0400 BUG: vtkProbeFilter stored state which did not work for multi-block data. For multi-block data, the pipeline calls vtkProbeFilter for each block of the multi-block mesh. vtkProbeFilter used the same array as vtkValidPointMask which resulted in all blocks writing to the same array. ================ ParaView =================== commit 71487c396af9bcc1a1e951583ec9e197515ae8b8 Author: Dan Lipsa <dan.lipsa@kitware.com> Date: Fri Jul 10 15:32:13 2015 -0400 BUG 15397: Hide input representations which showing ResampleWithDataset. |
|
(0035020) Dan Lipsa (developer) 2015-08-18 16:47 edited on: 2015-08-28 13:25 |
A crash happens for a simpler problem: Resample with Dataset with wavelet for attribute values and blow.vtk (unstructured grid) distributed through D3 for probe positions. The problem is the vtkPProbeFilter assumes the whole geometry dataset (which determines positions to probe) is available on all nodes. We fix the crash and print an error message. commit c0f2e89244022c28209cf28a7c46fceea7f33979 Author: Dan Lipsa <dan.lipsa@kitware.com> Date: Fri Aug 28 13:12:25 2015 -0400 BUG 0015397: vtkPProbeFilter crash when geometry dataset is distributed. If running in parallel, vtkPProbeFilter assumes that the whole geometry dataset (which determines positions to probe) is available on all nodes. If that is not the case a crash happened. This commit fixes the crash and prints an error message if different geometry datsets are available on different nodes. |
|
(0035123) Dan Lipsa (developer) 2015-09-02 14:12 |
A known limitation of vtkPProbeFilter is that it expects the whole geometry dataset (which determines positions to probe) to be available on all nodes. If that is not the case, possible options are 1. the filter will do the broadcast - time consuming and possible wasteful 2. the user is provided with a filter that does this. Further discussion is required to determine use case. |
|
(0035155) Utkarsh Ayachit (administrator) 2015-09-03 09:40 |
A fix for the crash has been merged in `master`. https://gitlab.kitware.com/paraview/paraview/merge_requests/338 [^] |
|
(0035543) Utkarsh Ayachit (administrator) 2015-12-11 15:03 |
Note the current fix in only a partial fix. More work is needed to totally fix the filter to work with composite datasets. |
|
(0035558) Alan Scott (manager) 2015-12-15 19:25 |
I'm going to mark this as closed. I tested it, and it worked as I would have expected. If there still are issues, we can open another bug. Tested 5.0.0-RC2, Linux, remote server. |
| Notes |
| Issue History | |||
| Date Modified | Username | Field | Change |
| 2015-03-26 17:27 | Alan Scott | New Issue | |
| 2015-03-26 17:30 | Alan Scott | Note Added: 0034327 | |
| 2015-04-29 18:32 | Alan Scott | Note Added: 0034377 | |
| 2015-04-29 18:32 | Alan Scott | Target Version | => 4.4 |
| 2015-06-07 14:13 | Utkarsh Ayachit | Status | backlog => todo |
| 2015-07-08 16:30 | Utkarsh Ayachit | Status | todo => backlog |
| 2015-07-08 16:36 | Utkarsh Ayachit | Status | backlog => todo |
| 2015-07-09 13:49 | Dan Lipsa | Assigned To | => Dan Lipsa |
| 2015-07-09 13:50 | Dan Lipsa | Status | todo => active development |
| 2015-07-09 15:10 | Dan Lipsa | Note Added: 0034725 | |
| 2015-07-10 15:37 | Dan Lipsa | Note Added: 0034747 | |
| 2015-07-14 14:16 | Dan Lipsa | Note Edited: 0034747 | |
| 2015-08-18 16:47 | Dan Lipsa | Note Added: 0035020 | |
| 2015-08-28 13:00 | Dan Lipsa | Note Edited: 0035020 | |
| 2015-08-28 13:25 | Dan Lipsa | Note Edited: 0035020 | |
| 2015-09-02 14:12 | Dan Lipsa | Note Added: 0035123 | |
| 2015-09-02 14:35 | Dan Lipsa | Status | active development => gatekeeper review |
| 2015-09-03 09:40 | Utkarsh Ayachit | Note Added: 0035155 | |
| 2015-12-11 15:02 | Utkarsh Ayachit | Status | gatekeeper review => customer review |
| 2015-12-11 15:02 | Utkarsh Ayachit | Resolution | open => fixed |
| 2015-12-11 15:02 | Utkarsh Ayachit | Fixed in Version | => 5.0 |
| 2015-12-11 15:03 | Utkarsh Ayachit | Note Added: 0035543 | |
| 2015-12-15 19:25 | Alan Scott | Note Added: 0035558 | |
| 2015-12-15 19:25 | Alan Scott | Status | customer review => closed |
| Issue History |
| Copyright © 2000 - 2018 MantisBT Team |