long iteration time

BakedAlmonds · July 19, 2016, 11:27

Hi, I am running a simulation with 12 million cells with 8 processors and 8gb ram. It is taking 20 mins ++ per iteration.

Is this normal? If so then wont it take weeks before convergence?

Or does it means that something is wrong with my model or case setup? I am only using k-e turb model.

schuyler · July 19, 2016, 14:47

Doesn't really surprise me. But I could be wrong. That's 1.5 million cells per core. Personally when I run parallel cases I am usually using around 20000 cells per core or even smaller sometimes.

For ram you could turn on your system monitor while it is running. You will see if your available ram is being maxed out.

Just my opinion on it. Let me know if you figure something out! I would be interested to know.

BakedAlmonds · July 19, 2016, 15:02

HEY! thanks for the reply. I figured out what I did wrong.

I set the iteration for each field calculation to too high a value by making the tolerance in fvsolution insanely low. Before, it runs 1000 iteration to calculate each parameter but not it mostly does <5

schuyler · July 19, 2016, 15:04

Awesome! Glad you figured it out

July 19, 2016, 11:27	long iteration time	#1
BakedAlmonds Member Almond Wong Join Date: May 2016 Posts: 68 Rep Power: 10	Hi, I am running a simulation with 12 million cells with 8 processors and 8gb ram. It is taking 20 mins ++ per iteration. Is this normal? If so then wont it take weeks before convergence? Or does it means that something is wrong with my model or case setup? I am only using k-e turb model.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
High Courant Number @ icoFoam	Artex85	OpenFOAM Running, Solving & CFD	11	February 16, 2017 14:40
Moving mesh	Niklas Wikstrom (Wikstrom)	OpenFOAM Running, Solving & CFD	122	June 15, 2014 07:20
Help for the small implementation in turbulence model	shipman	OpenFOAM Programming & Development	25	March 19, 2014 11:08
dynamic Mesh is faster than MRF????	sharonyue	OpenFOAM Running, Solving & CFD	14	August 26, 2013 08:47
Missing math.h header	Travis	FLUENT	4	January 15, 2009 12:48

July 19, 2016, 14:47		#2
schuyler Member W. Schuyler Hinman Join Date: Apr 2013 Location: Calgary, Alberta, Canada Posts: 38 Rep Power: 13	Doesn't really surprise me. But I could be wrong. That's 1.5 million cells per core. Personally when I run parallel cases I am usually using around 20000 cells per core or even smaller sometimes. For ram you could turn on your system monitor while it is running. You will see if your available ram is being maxed out. Just my opinion on it. Let me know if you figure something out! I would be interested to know.

July 19, 2016, 15:02		#3
BakedAlmonds Member Almond Wong Join Date: May 2016 Posts: 68 Rep Power: 10	HEY! thanks for the reply. I figured out what I did wrong. I set the iteration for each field calculation to too high a value by making the tolerance in fvsolution insanely low. Before, it runs 1000 iteration to calculate each parameter but not it mostly does <5

July 19, 2016, 15:04		#4
schuyler Member W. Schuyler Hinman Join Date: Apr 2013 Location: Calgary, Alberta, Canada Posts: 38 Rep Power: 13	Awesome! Glad you figured it out