my case in CFX runs in single mode but in parallel it does not

gaza · September 16, 2014, 11:47

Hi all,
I am doing 3D mixer simulation with radiation. I have two domains (stationary and rotating). I use transient rotor stator interface. My question is why CFX runs on one processor but in the parallel mode (e.g. on 4 processors) solver crashes?
If anybody can help?
best regards
pBlasiak

ghorrocks · September 16, 2014, 18:52

It is unusual for a run which is OK on single processor to crash multiprocessor. The CFX documentation has some comment on this, you should read it. It suggests your simulation is numerically unstable as small changes can cause numerical instability. So the usual fixes for numerical instability apply: Improve mesh quality, smaller time steps, double precision solver and a better initial condition.

September 16, 2014, 11:47	my case in CFX runs in single mode but in parallel it does not	#1
gaza Senior Member Przemek Join Date: Jun 2011 Posts: 249 Rep Power: 15	Hi all, I am doing 3D mixer simulation with radiation. I have two domains (stationary and rotating). I use transient rotor stator interface. My question is why CFX runs on one processor but in the parallel mode (e.g. on 4 processors) solver crashes? If anybody can help? best regards pBlasiak

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Parallel Moving Mesh Bug for Multi-patch Case	albcem	OpenFOAM Bugs	17	April 28, 2013 23:44
Parallel runs with sonicDyMFoam crashes (works fine with sonicFoam)	jnilsson	OpenFOAM Running, Solving & CFD	0	March 9, 2012 06:45
CFX Parallel Setup on windows 7 x64	SlicedBread	CFX	1	November 14, 2011 17:06
Grid Check Fails in Parallel Processing Mode	askance	Main CFD Forum	0	October 20, 2010 10:11
parallel mode - small problem?	co2	FLUENT	2	June 1, 2004 23:47

September 16, 2014, 18:52		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,719 Rep Power: 143	It is unusual for a run which is OK on single processor to crash multiprocessor. The CFX documentation has some comment on this, you should read it. It suggests your simulation is numerically unstable as small changes can cause numerical instability. So the usual fixes for numerical instability apply: Improve mesh quality, smaller time steps, double precision solver and a better initial condition.