WRF Stop Working With No Error

[dia]

hi guys

i have problem when i run wrf.exe let me give you brief summary
i use WRF V4.1.3
i use 20 nodes
1600 core
so no problem with the hardware

the same name list works with mp 8 and mp 1
but when change it to mp 6 and mp 2 don't work

wrfbdy_d01 show all times

the problem wrf stop randomly sometime goes 10 days and some 16 with no error

namelist + rsl.error from 0000 to 0009 attached
Im using CFSv2 data
can you guys help
thanks

 

[Ming Chen]

Your namelist.inputy looks fine except that the time step of 60 is way too small for dx=20km.

There is no error message in any of the rsl files you posted. Would you please double check all the rsl files and log file for this run? If the model crashed, there must be some error messages output in the rsl files or log file. Such information is important for us to figure pout what is wrong.

 

[dia]

Your namelist.inputy looks fine except that the time step of 60 is way too small for dx=20km.

There is no error message in any of the rsl files you posted. Would you please double check all the rsl files and log file for this run? If the model crashed, there must be some error messages output in the rsl files or log file. Such information is important for us to figure pout what is wrong.

Thanks for replying
There is no error in all 1599 rsl file
As I mentioned the problem only appears when i use mp 6 and mp 2 , when i use mp8 and mp1 with the same exact namelist and it works, what do you think the problem is or what suspicion do you have it might give any clue, and there is no sing of crashing

 

[Ming Chen]

If this case crashed, there must be something wrong and the error message may appear in RSL files.
Please double check all the files. Without any helpful information, it is hard for us to figure out what is wrong.

 

[dia]

If this case crashed, there must be something wrong and the error message may appear in RSL files.
Please double check all the files. Without any helpful information, it is hard for us to figure out what is wrong.

ok i did the run with 1 node 10 core so there will be 10 rsl file i think i found the error but i dont know the reason

please check rsl.error.0002

thanks again

 

[Ming Chen]

There are many CFL violations in your rsl file, which indicates that this case is numerically unstable.

I notice that you have deleted this thread, ---- does this indicate that you have solved your issue? If so, we would appreciate that you post your solutions instead of deleting the thread. This Forum is supposed for users to share information and discuss possible problems. Your solution might be helpful for others who have the same issue.

 

[dia]

There are many CFL violations in your rsl file, which indicates that this case is numerically unstable.

I notice that you have deleted this thread, ---- does this indicate that you have solved your issue? If so, we would appreciate that you post your solutions instead of deleting the thread. This Forum is supposed for users to share information and discuss possible problems. Your solution might be helpful for others who have the same issue.

i just change
w_damping = 0,

to

w_damping = 1,