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

Real.exe corrupted without error information

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.

lacstorm

New member
Hi everyone. I'm using WRF3.5.1.
When I run real.exe by "mpirun -np 20 real.exe" for three nested domains.
The progress will stop after the first 2 domains are processed.
I got
Code:
===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   RANK xx PID xxxx RUNNING at xxx pc
=   KILLED BY SIGNAL: 9 (Killed)
===================================================================================
The the rsl.out.0000 file did not show any error information and is given in the attachment.
The namelist is also given in the attachments since I always get a hint like "spam message, contains contatct."

However, when I use only one core by "./real.exe".
The real.exe could be ran successfully.
I'm very confused about this.
And actually I need a larger domain for
Code:
e_we = 361,  385,  421 
e_sn = 361, 373, 301.

And the real.exe will be killed even if I use one core with this larger domain.
When I type "free -m", the free memory is 61491, the total memory is 63911.
My disk place is not full and the enviroment WRFIO_NCD_LARGE_FILE_SUPPORT=1 and ulimit -s unlimited are set.

View attachment namelist.input
View attachment rsl.error.0000.one_core.txt
View attachment rsl.out.0000.20cores.txt
 
It's hard to say why this is happening, but the real program really does not need many processors to run because it's such a short/simple program. I almost always only use 1 processor for real. It's the wrf program that depends on multiple processors. If you are able to run real without problems with 1 processor, then perhaps you can move on to wrf.exe to see if you're able to run with the number of processors that you need.
 
kwerner said:
It's hard to say why this is happening, but the real program really does not need many processors to run because it's such a short/simple program. I almost always only use 1 processor for real. It's the wrf program that depends on multiple processors. If you are able to run real without problems with 1 processor, then perhaps you can move on to wrf.exe to see if you're able to run with the number of processors that you need.
Hi, kwerner, thanks for your reply.
Actually, I need a larger domain as
Code:
e_we = 361,  385,  421 
e_sn = 361, 373, 301.
And I can't run real.exe successfully even with only one core for this larger domain.
 
Hi,
Can you send your configure.wrf file (found in the WRF/ directory), along with all of the rsl* files when your real.exe run fails with multiple processors? You can package them altogether in a single *.TAR file. Thanks!
 
Top