What is the best proper value of second per iteration?

dlalstnf · April 6, 2023, 10:25

Hello everyone!

I'm novice of running CFD.
So I don't know proper seconds per iteration.

I know that it depends on problem.

But, generally, Do you have any guideline about this?

In my case, 5 sec / iteration using 124 core..
Is it proper or not?

Thank you!

bigfootedrockmidget · April 6, 2023, 12:54

A case I have has 1M points and the RANS solver with SA runs 5 s/iteration on 1 core.
So if it is scaling entirely linearly, then a 124M points case will also run 5s/iteration on 124 cores.

First thing to check is if your code was compiled with optimization on (debugging off, optimization flag -O2) and if you get a proper linear time reduction in nr of cores when using mpi.

April 6, 2023, 10:25	What is the best proper value of second per iteration?	#1
dlalstnf New Member MINSUL LEE Join Date: Apr 2023 Posts: 8 Rep Power: 3	Hello everyone! I'm novice of running CFD. So I don't know proper seconds per iteration. I know that it depends on problem. But, generally, Do you have any guideline about this? In my case, 5 sec / iteration using 124 core.. Is it proper or not? Thank you!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
pimpleFoam residual increase in final iteration	AhmedHafez	OpenFOAM Running, Solving & CFD	0	December 1, 2022 03:03
Slow calculation time on CFD server	killian153	Hardware	4	July 27, 2021 07:28
Free surface issues with interDyMFoam for hydroturbine	oumnion	OpenFOAM Running, Solving & CFD	0	October 6, 2017 14:05
Hardware-Configuration for Fluent HPC-Pack (8x)	JohHaas	Hardware	9	March 3, 2015 13:25
Parallel runs slower with MTU=9000 than MTU=1500	Javier Larrondo	FLUENT	0	October 28, 2007 22:30

April 6, 2023, 12:54		#2
bigfootedrockmidget Senior Member bigfoot Join Date: Dec 2011 Location: Netherlands Posts: 585 Rep Power: 17	A case I have has 1M points and the RANS solver with SA runs 5 s/iteration on 1 core. So if it is scaling entirely linearly, then a 124M points case will also run 5s/iteration on 124 cores. First thing to check is if your code was compiled with optimization on (debugging off, optimization flag -O2) and if you get a proper linear time reduction in nr of cores when using mpi.