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

Not able to modify domain decomposition

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,
To avoid small subdomains, I am using nproc_x
Code:
 &domains
 time_step                           = 150,
 ...
 nproc_x                             = 16
 /
However, WRF(v4.3) seems to be ignoring it and still performing the domain decomposition as 8x16 instead of 16x8 as expected:
Code:
   Minimum decomposed computational patch size, either x-dir or y-dir, is 10 grid cells.
  e_we =   165, nproc_x =    8, with cell width in x-direction =   20
  e_sn =   131, nproc_y =   16, with cell width in y-direction =    8
  --- ERROR: Reduce the MPI rank count, or redistribute the tasks.
Thanks
 
Hi,
Can you try also adding "nproc_y = 8" to the namelist and see if that makes a difference?
 
When I try that, I'm getting the error
Code:
 module_io_quilt_old.F        2931 T
-------------- FATAL CALLED ---------------
FATAL CALLED FROM FILE:  <stdin>  LINE:    5711
 check comm_start, nest_pes_x, nest_pes_y settings in namelist for comm            1
-------------------------------------------
 
Hi,
Can you attach your full namelist.input file, as well as your rsl.out.0000 and rsl.error.0000 files so that I can take a look at those? Thanks!
 
Hello Kelly,
Due to confidentiality matters, I cannot share the files. The only piece of information that I can share is that the size of the parent domain (the one producing the error) is 165 x 131. As far as the rsl files, they looked like usual but with the error (when both nproc_x & nproc_y were specified) previously posted. At the end, we are not pursuing to change the domain decomposition from 8x16 to 16x8. However, custom domain decomposition is still an important feature so understanding what might have caused the error of interest.
Thanks,
Arturo
 
Arturo,
Unfortunately it may be difficult to understand the problem without those files, but I can understand your privacy concerns. If you want to do some troubleshooting, perhaps you can attempt to add some print statements in the code to track down the issue. This particular error you are seeing (check comm_start, nest_pes_x, nest_pes_y settings in namelist for comm) occurs in the WRF/external/RSL_LITE/module_dm.F file.

I'm glad you've been able to move forward, though. Thank you for letting us know.
 
Top