|November 10, 2009, 11:34||
Run in parallel: bug or what?
Join Date: Mar 2009
Location: Polonghera, Cuneo, Italy
Posts: 27Rep Power: 9
I've observed a strange behaviour when I tried to run in parallel a case (with multiple regions) decomposed in a "strange" way. For many reasons I have to decompose some region (with manual method) without using all processor.
if I set numberOfSubdomains = 3 (for many reasons I can't change this value) and I want to decompose my domain only using 2 CPU, it is sufficient for me that in the "manual file decomposition" appear only the number 0 and 1.
decomposePar works well and splits the domain (with only a WARNING). If I check the mesh decomposition with paraview all seems ok. But when I run
mpirun -np 3 icoFoam (for example) -parallel
the solver doesn't works...
Anyone can tell me why? Is this a bug or I can't decompose (never) a domain in this way?
...The best way to acquire new knowledge is to share it...
|Thread||Thread Starter||Forum||Replies||Last Post|
|Parallel Moving Mesh Bug for Multi-patch Case||albcem||OpenFOAM Bugs||17||April 28, 2013 23:44|
|Script to Run Parallel Jobs in Rocks Cluster||asaha||OpenFOAM Running, Solving & CFD||12||July 4, 2012 22:51|
|Working directory via command line||Luiz||CFX||4||March 6, 2011 21:02|
|Parallel Run on dynamically mounted partition||braennstroem||OpenFOAM Running, Solving & CFD||14||October 5, 2010 14:43|
|Parallel Moving Mesh Bug for Multi-patch Case||albcem||OpenFOAM||0||May 21, 2009 00:23|