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 while running init_atmosphere_model for ever

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.

DaShao

New member
Hi all,

I got the following error while running init_atmosphere_model for generating static.nc.
Program received signal SIGSEGV: Segmentation fault occur.
And it is not useful for "ulimit -s unlimited".
My MPAS version is 6.1.
Would you please help me?Or can you give me a namelist.init_atmosphere file and streams.init_atmosphere for static?
Thank you very much.

DaShao
 
For which mesh are you trying to process the static fields? How much memory do you have available on the machine/node where you are running? Can you attach your namelist.init_atmosphere and log.init_atmosphere.* files?
 
mgduda said:
For which mesh are you trying to process the static fields? How much memory do you have available on the machine/node where you are running? Can you attach your namelist.init_atmosphere and log.init_atmosphere.* files?

My mesh is 30-km mesh (655362 horizontal grid cells).
My machine has 247G memory.
The namelist.init_atmosphere and streams.init_atmosphere are following.
Thank you.View attachment static_streams_init_atmosphere.txtView attachment static_namelist_init_atmosphere.txtView attachment log.init_atmosphere.0000.out.txt
 
Thanks very much for attaching these files -- all of your settings look reasonable to me, and 247 GB of memory should be more than sufficient to process the static fields for the 30-km mesh. I did notice that the standard log file header doesn't appear in your log.init_atmosphere.0000.out file; is the file you've attached the entire log file? For example, I would expect to have seen something like the following at the top of the log file:
Code:
----------------------------------------------------------------------
Beginning MPAS-init_atmosphere Output Log File for task       0 of       1
    Opened at 2019/05/02 11:29:40
----------------------------------------------------------------------

Which compiler and compiler version are you using? I'm not sure whether it will give any additional information given that the code is crashing so early in the static interpolation process, but would you be willing to 'make clean CORE=init_atmosphere' and recompile the code with 'DEBUG=true' set in your make command?

Lastly, are you using the 'bash' shell? You mentioned that 'ulimit -s unlimited' doesn't help, and I'd just like to confirm that you're not using, e.g., tcsh.
 
mgduda said:
Thanks very much for attaching these files -- all of your settings look reasonable to me, and 247 GB of memory should be more than sufficient to process the static fields for the 30-km mesh. I did notice that the standard log file header doesn't appear in your log.init_atmosphere.0000.out file; is the file you've attached the entire log file? For example, I would expect to have seen something like the following at the top of the log file:
Code:
----------------------------------------------------------------------
Beginning MPAS-init_atmosphere Output Log File for task       0 of       1
    Opened at 2019/05/02 11:29:40
----------------------------------------------------------------------

Which compiler and compiler version are you using? I'm not sure whether it will give any additional information given that the code is crashing so early in the static interpolation process, but would you be willing to 'make clean CORE=init_atmosphere' and recompile the code with 'DEBUG=true' set in your make command?

Lastly, are you using the 'bash' shell? You mentioned that 'ulimit -s unlimited' doesn't help, and I'd just like to confirm that you're not using, e.g., tcsh.

Oh,no.I neglected the top of the log file at the last as you show.But there is nothing at the end of the log file beside.
My compiler is ifort version 18.0.5 and I use 'bash' shell.
The Mpas did not become well after recompiling the 'init_atmosphere'.
Thank you.
 
You mentioned that recompiling did not help. But, after running 'make clean CORE=init_atmosphere' and 'make ifort CORE=atmosphere DEBUG=true' (possibly adding USE_PIO2=true depending on your PIO library version), does the executable create any log.init_atmosphere.0000.err file, or are there any messages printed to the terminal screen?
 
Top