Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

Segmentation fault due to physics?

This post was from a previous version of the WRF&MPAS-A Support Forum. New replies have been disabled and if you have follow up questions related to this post, then please start a new thread from the forum home page.

janac4

Member
Hello,
I'm trying to execute a wrf simulation in Panama. I wanted to use two domains (one using PBL physics and another with LES conditions), but I got a segmentation fault error.
I also tried to compile the model with only one domain but it cracks anyway and I don't know why.
I tried reducing the time_step, changing the value of epssm, using smooth_cg_topo = true but it always cracks. I also wanted to use the w_damping variable but when running real.exe:

------ ERROR while reading namelist dynamics ------
Maybe here?: w_damping = 1.,
Maybe here?: mix_isotropic = 1

Is there any inconsistency in the physics or dynamics? I wanted to use the CONUS parameterization to avoid problems but I'm not able to make it work.
I attach the namelists I used in case they are helpful.
Thanks
Jana.
 

Attachments

  • namelist.input
    5.1 KB · Views: 34
  • namelist.wps
    1.3 KB · Views: 35
Jana,
The error you're seeing is simply a namelist issue. I believe the problem may be that you have "w_damping = 1." There shouldn't be a "." after the one because it expects an integer.
 
I found this issue after posting the question, but the simulation cracks anyway so there must be something else wrong with the phyisics.
 
I checked the post you mentioned but It's not very helpful. The real.exe program has no problem when running, even the wrf.exe starts without a problem, but in the middle of the simulation it cracks.
I have used another namelist with the CONUS phyisics and changed values from that and it has worked. I still don't know if there where inconsistencies in the first namelis.input.
Thanks.
 
Can you send the most recent namelist.input file you're using, along with the runtime error files (e.g., rsl.error.* If sending rsl files, please package them all together in a single *.TAR file). Thanks!
 
Hi kwerner,
Here you have the last namelist.input and rsl.error files that are giving me problems.
Jana.
 

Attachments

  • error.rar
    14 KB · Views: 28
Hi Jana,
Unfortunately I'm unable to open a .rar file. Can you send it again as a .tar? Thanks!
 
Hello kwerner,
I'm not able to change the format from .rar to .tar, I hope you can open the attached files.
I was able to make the simulation work after changing the PBL option. The only one that has worked is YSU scheme, are there any incompatibilities when using another PBL scheme?
Jana.
 

Attachments

  • namelist.input
    5 KB · Views: 34
  • rsl.error.0000.txt
    76.9 KB · Views: 25
  • rsl.out.0000.txt
    75.9 KB · Views: 24
Jana,

Looking at your namelist.input, I can see that you're using mp_physics = 10. I recall from other case, that this option seems to be unstable. Please, take a look here:
https://forum.mmm.ucar.edu/phpBB3/viewtopic.php?f=58&t=10273

Can you try to change that and see then how your case is going on?
Bye,
Ivan
 
I agree with Ivan. I would try a different MP scheme. The Morrison scheme has been known to be tricky. It's also really hard to tell what modifications helped since your new namelist is so drastically different than your original one, so I can't narrow down the issue.

When I pointed you to the other post, I was simply advising you to follow the method I instructed the user to use to narrow down the namelist setting that caused the problem - not the specific issue they were having (running real.exe).
 
Thanks for your answers. I finally made it work after changing the PBL scheme to YSU. I didn't think the problem could be in the mp_physics option as I had tried it in other simulations and it worked fine. I'm doing a comparison between different schemes (5 in total) and in two of them I had to force YSU scheme to be used as the other PBL options made it crush.
Thanks for everything!
 
Top