|
[Sponsors] |
[snappyHexMesh] snappyHexMesh - parallel processing problems |
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
March 5, 2014, 00:55 |
snappyHexMesh - parallel processing problems
|
#1 |
New Member
J. Brindley
Join Date: Mar 2014
Posts: 11
Rep Power: 12 |
When I run snappy regularly everything is fine. As soon as I use decomposePar, the resulting mesh is the same as that of blockMesh. Is this a simple fix?
|
|
March 5, 2014, 05:52 |
|
#2 |
Senior Member
Lieven
Join Date: Dec 2011
Location: Leuven, Belgium
Posts: 299
Rep Power: 22 |
You first need to copy the mesh in the last time directory created by snappy to the constant/polyMesh folder. Or alternatively run snappy with the '-overwrite' option.
Cheers, L |
|
March 5, 2014, 10:35 |
|
#3 |
New Member
J. Brindley
Join Date: Mar 2014
Posts: 11
Rep Power: 12 |
after running reconstructPar -latestTime, no new folders are being made? All I have is 8 processor folders and the 0, constant, and systems folder. I feel like a new time folder should be created correct?
|
|
March 5, 2014, 10:49 |
|
#4 |
Member
Julian Langowski
Join Date: May 2011
Location: Bremen, Germany
Posts: 91
Rep Power: 15 |
Dear B,
if you use -overwrite, no new folders (1,2,3) are created. Instead 0 is overwritten. Best regards Julian
__________________
πάντα ῥεῖ - Heraclitus |
|
March 5, 2014, 10:56 |
|
#5 |
New Member
J. Brindley
Join Date: Mar 2014
Posts: 11
Rep Power: 12 |
when I run snappy with -overwrite I cannot get recontructPar to work?
|
|
March 5, 2014, 11:21 |
|
#6 |
New Member
J. Brindley
Join Date: Mar 2014
Posts: 11
Rep Power: 12 |
blockMesh
decontructPar -force foamJob -parallel -screen snappyHexMesh -overwrite reconstructPar -zeroTime ---> gives error: Size of maps does not correspont to size of mesh for processor 0 reconstructPar -latestTime ---> gives error: No times selected when I run without overwrite, two new times folders are created within each processor (one for mesh, one for snap), but reconstructPar does not create new times folders in the directory. when I run with overwrite, no new times folders are created within each processor, and I cannot use reconstructPar. again, if I just use snappyHexMesh, everything works out perfectly. Could this have to do with the way decomposePar is splitting up the mesh? the piece I am adding is not evenly distributed. |
|
March 5, 2014, 11:38 |
|
#7 |
Member
Julian Langowski
Join Date: May 2011
Location: Bremen, Germany
Posts: 91
Rep Power: 15 |
Try
Code:
reconstructParMesh -noZero -constant -mergeTol 1e-6 Best regards Julian
__________________
πάντα ῥεῖ - Heraclitus |
|
March 5, 2014, 11:57 |
|
#8 |
New Member
J. Brindley
Join Date: Mar 2014
Posts: 11
Rep Power: 12 |
RULI you're the best! Worked like a charm.
|
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
[snappyHexMesh] Parallel processing SnappyHexMesh | mullet904 | OpenFOAM Meshing & Mesh Conversion | 6 | September 25, 2016 14:30 |
Explicitly filtered LES | saeedi | Main CFD Forum | 16 | October 14, 2015 12:58 |
multiphase solver - parallel processing - GAMG | thibault_pringuey | OpenFOAM Programming & Development | 2 | August 27, 2013 23:03 |
OF Parallel Processing with Core i7 - How to Handle Hyperthreading | dancfd | OpenFOAM Running, Solving & CFD | 5 | March 5, 2012 22:43 |
Cyclic patches and parallel postprocessing problems | askjak | OpenFOAM Bugs | 18 | October 27, 2010 04:35 |