sHM - reconstructPar
Hi all,
i run snappy with 6 cores. After it, i will reconstruct it again. But how?! Code:
tobi@tobi:~/OpenFOAM/tobi-1.7.x/run/teg$ reconstructPar -constant Any ideas. Tobi |
Why do you use -constant?
Have you tried it without? |
Hi,
yes i ve tried it with all tags. but i realized, that i need "reconstructParMesh" instead of reconstructPar after mesh generation. I am so stupid :( tobi |
Quote:
|
Quote:
hi Sandy,... if you split your mesh to solve sHM with more cores you have to reconstruct your mesh with the command Code:
reconstructMeshPar Code:
reconstructPar Hope it 's helpful. -> reconstructPar for solver -> reconstructParMesh for sHM I hope its correct - but i am certain sure about it. Tobi |
Hi Tobi, thank you for your reply. However, now I use lines step by step as follows:
1. blockMesh 2. snappyHexMesh 3. decomposePar 4. mpirun -np 8 interFoam -parallel ..... Could you tell me which one I should use, reconstructPar or reconstructParMesh? Waiting for your help again. Sandy |
Quote:
5. reconstructPar :) I added a bash file you can execute it with ./solve and change it - you should use your solver ;) http://www.file-upload.net/download-...solve.zip.html Tobi |
Hello Tobi,
maybe you can help me. I got the same error message: Quote:
I run these commands: setFields >> log.setFields decomposePar >> log.decomposePar mpirun -np 2 interDyMFoam -case /home/stephanie/Schreibtisch/FFW_E2/ -parallel >> log.interDyMFoam reconstructPar >> log.reconstructPar There seemed to be no problem until the command reconstructPar. There are both processor-folders with the timesteps inside. Might you have any idea, where I did a mistake? I would be very grateful for your help. Thank you and best regards, Stephie |
Hi Stephie,
with adaptive mesh refinement you have each time step a different mesh. Use 1. reconstructParMesh for the mesh for each time step 2. reconstructPar for the fields Best regards Tobias |
reconstructPar not reconstructing time 0
Hi Tobi and Stephie,
Hope you can help me out with something that has been keeping my mind busy, related with reconstructing the case after running in it parallel. The base case has the directories: 0.orig (with the respective fields), constant and system. Pretty standar. I run the simulation (PISO LES solver), decomposing the domain and reconstructing it, with the following Allrun script: Code:
#!/bin/sh reconstructPar (or reconstructParMesh, not sure which of both) does not reconstruct time 0 (initial values), but it does it perfectly with the following time step. I have already deleted the lines Code:
# Make dummy 0 directory Any thoughts? Thanks in advance for the time you might take reading my question. |
reconstructPar has the option -withZero, use it like this:
reconstructPar -withZero |
Thank you very much! I tried -time 0 before asking, but i didn't encounter (or thought) the right way to use it so i received nothing but error messages . Thank you again, i will try it out very soon.
|
Hi tobi! I've run my case in oepnFoam 7 with the order below :
runApplication surfaceFeatures runApplication blockMesh runApplication decomposePar -copyZero runParallel snappyHexMesh -overwrite -parallel runParallel setRefinedFields to recompose the case, i've trie reconstrucParmesh and reconstrucPar, event for each time step with a script : #!/bin/bash timeList=$(ls processor0/ | awk '$1!="constant"'); for line in $timeList do echo "Recontructing mesh for t = "$line" s"; reconstructParMesh -time $line > logTmp; echo "Reconstructing fields for t = "$line" s"; reconstructPar -time $line > logTmp; echo ""; done rm logTmp; but i have alaway the same error : processor0/0/polyMesh/pointProcAddressing not exist. Could you please give me a hint! |
All times are GMT -4. The time now is 01:51. |