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

WRF compilation not finding modules

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.

Hello,
I'm trying to compile WRF with the GNU compilers (v9.3.0) to check any potential improvements vs. GNUv7. However, and unlike for the latter, the compilation cannot find some modules (e.g. module_dm.mod). I'm unsure why of the differences for compilers of the same family.
Thanks
 

Attachments

  • configure.wrf
    20.2 KB · Views: 63
  • output_compilation.txt
    183.3 KB · Views: 66
Hi,
It's hard to say what may be causing that. I just did a test with version GNU V9.1.0 (we don't have access to 9.3 yet), and WRFV4.1.5 and had no troubles. What version of WRF were you compiling? Can you send the full compile log? If you don't have it, issue a 'clean -a' and then reconfigure and recompile, then make sure to send the standard error and output to a log file, using a syntax like:
./comple em_real >& log.compile

Also, if you're interested, someone else recently did a comparison test of GNU V9 (not sure of the exact version of 9) and Intel. It's not exactly what you're doing, but thought you might like to see their results:
https://forum.mmm.ucar.edu/phpBB3/viewtopic.php?f=40&t=8953&p=16214#p16214
 
Hi @kwerner,
As usually, I forgot to write the output to a file. It's now attached. It might be that WRF detects that there are two sets of GNU compilers (the one with the OS plus GNUv9.3.0 installed at /opt), and this causes the installation of module_dm.mod to fail. Unless you have any suggestion on how to tell the WRF installer to somehow overcome this situation, the only solution that I can think of is to replace the older GNU with the newer one while installing it.
I saw the post that you mention (one of my objectives is to compare different configurations). I still haven't gotten to the Intel compiler but will try as soon as other obligations permit.
Thanks.
P.S. Forgot to mention that the version is 4.1 (not sure exactly which release)
 

Attachments

  • log.compile
    832.2 KB · Views: 68
Thanks for sending that. Is the code that you're compiling the pristine "out of the box" code that we provide, or have there been modifications to the code prior to your compiling?
 
Hi,

I am wondering whether this problem has been resolved? I met similar situation.

We cloned the version from your github site. We managed to compile and run a series run with single domain.

Then we did a ./clean all and tried to re configure it with 'dm' and for nested domain. When we compile em_real, we got following error message:

=============================================================
module_comm_nesting_dm.F:30:0: fatal error: REGISTRY_COMM_NESTING_DM_subs.inc: No such file or directory
#include "REGISTRY_COMM_NESTING_DM_subs.inc"
^
compilation terminated.
===================================================================

Those inc files should be generated automatically? I saw this file in the 'configure.wrf' in the session # compile without OMP?

We are using GNU fortran and gcc compiler.

Any clue?

Many thanks,

Qian
 
The error message indicates that the file REGISTRY_COMM_NESTING_DM_subs.inc doesn't exist. However, this file should be automatically generated during the compiling process. Failing to generate this file often indicates something wrong in the environmental settings or libraries required for building WRF.
ONe suggestion: please build WRF in dumper mode, and check your environmental settings to make sure mpi library is installed correctly and located at the right place (e.g., /usr/local/bin)
 
Top