waning messages during computation - suppression possible?

gbettega · November 13, 2024, 10:18

Hello forum,
is there the possibility to suppress this kind of message

Warning: y+ < 5 in 117 points, for which the wall model is not active.
Warning: Computation of wall coefficients (y+) did not converge in 1 points.

using a proper setting in the .cfg file?
Thank you very much
Giovanni

joshkellyjak · November 13, 2024, 10:28

Yes, set COMM_LEVEL= MINIMAL, this will supress the wall function output. It may also suppress other useful information so be careful.

gbettega · November 14, 2024, 01:37

Thank you. It works. World record in fast answer discipline. Buona giornata

Marcus021 · January 6, 2025, 16:44

Hello, I'm piggybacking off this because I noticed that when setting COMM_LEVEL = MINIMAL and using mpirun the evolution of the residuals and the entire flow changes. This does not happen when using directly SU2_CFD.

In my case I was using wall functions and FGMRES was consistently failing orthogonalization after 200 or so iterations, while by enabling this option my simulation continues further with seemingly no hassle.

I have also tested this behaviour in a slightly modified version of the RAE2822 test case, where by suppressing warnings the simulation converges much faster.

I'm wondering if this behaviour is intended since perhaps this affects the way parallel computation operates or if there's something else going on that I'm missing? My cfg files were exactly the same aside from this option being present or absent. Thank you very much for anybody that's willing to shed some light on this.

bigfootedrockmidget · January 7, 2025, 01:57

Sounds like we switch off too many things when COMM_LEVEL is not FULL. Can you create a ticket on github? If you can reproduce this with one of the standard testcases that we have, it will be easier to reproduce this.

https://github.com/su2code/SU2/issues

Marcus021 · January 7, 2025, 10:13

Thank you for the reply, I have created the ticket, hopefully the info I have provided is enough to reproduce the issue and look into this.
I was wondering if I could follow up with another odd behaviour I noticed when using wall functions, I am not sure if this is a bug or if it is something caused by a fault of my own. Shall I make a new post on the forum or is it ok to write here?

bigfootedrockmidget · January 7, 2025, 11:04

if the issue is unrelated to the warning messages, it is better to create a new issue

November 13, 2024, 10:18	waning messages during computation - suppression possible?	#1
gbettega Member Giovanni Bettega Join Date: Sep 2014 Posts: 42 Rep Power: 12	Hello forum, is there the possibility to suppress this kind of message Warning: y+ < 5 in 117 points, for which the wall model is not active. Warning: Computation of wall coefficients (y+) did not converge in 1 points. using a proper setting in the .cfg file? Thank you very much Giovanni

November 13, 2024, 10:28		#2
joshkellyjak Member Josh Kelly Join Date: Dec 2018 Posts: 57 Rep Power: 8	Yes, set COMM_LEVEL= MINIMAL, this will supress the wall function output. It may also suppress other useful information so be careful. bigfootedrockmidget, giovanni.medici and gbettega like this.

January 7, 2025, 11:04		#7
bigfootedrockmidget Senior Member bigfoot Join Date: Dec 2011 Location: Netherlands Posts: 772 Rep Power: 21	if the issue is unrelated to the warning messages, it is better to create a new issue Marcus021 likes this.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
PhD thesis with funding - Modelling of Fire Suppression in Heritage Buildings	AxelCable	CFD Freelancers	1	June 30, 2020 14:05
pimpleDyMFoam computation randomly stops	babapeti	OpenFOAM Running, Solving & CFD	5	January 24, 2018 05:28
problem in the CFX12.1 parallel computation	BalanceChen	ANSYS	2	July 7, 2011 10:26
Discussion on the computation of the multi-stage axial compressor	BalanceChen	CFX	5	January 5, 2011 02:23
Cell face values computation un unstructured grids	Sergio Rossi	Main CFD Forum	2	May 28, 2006 10:04

November 14, 2024, 01:37		#3
gbettega Member Giovanni Bettega Join Date: Sep 2014 Posts: 42 Rep Power: 12	Thank you. It works. World record in fast answer discipline. Buona giornata

January 6, 2025, 16:44		#4
Marcus021 New Member Join Date: Sep 2024 Posts: 6 Rep Power: 2	Hello, I'm piggybacking off this because I noticed that when setting COMM_LEVEL = MINIMAL and using mpirun the evolution of the residuals and the entire flow changes. This does not happen when using directly SU2_CFD. In my case I was using wall functions and FGMRES was consistently failing orthogonalization after 200 or so iterations, while by enabling this option my simulation continues further with seemingly no hassle. I have also tested this behaviour in a slightly modified version of the RAE2822 test case, where by suppressing warnings the simulation converges much faster. I'm wondering if this behaviour is intended since perhaps this affects the way parallel computation operates or if there's something else going on that I'm missing? My cfg files were exactly the same aside from this option being present or absent. Thank you very much for anybody that's willing to shed some light on this.

January 7, 2025, 01:57		#5
bigfootedrockmidget Senior Member bigfoot Join Date: Dec 2011 Location: Netherlands Posts: 772 Rep Power: 21	Sounds like we switch off too many things when COMM_LEVEL is not FULL. Can you create a ticket on github? If you can reproduce this with one of the standard testcases that we have, it will be easier to reproduce this. https://github.com/su2code/SU2/issues

January 7, 2025, 10:13		#6
Marcus021 New Member Join Date: Sep 2024 Posts: 6 Rep Power: 2	Thank you for the reply, I have created the ticket, hopefully the info I have provided is enough to reproduce the issue and look into this. I was wondering if I could follow up with another odd behaviour I noticed when using wall functions, I am not sure if this is a bug or if it is something caused by a fault of my own. Shall I make a new post on the forum or is it ok to write here?