ncdio_pio.F90.in at line 2

[MrIgnition]

Hi all,
I submitted my case which uses compset I2000Clm50SpGs. Got an error pointing to /clm5.0/src/main/ncdio_pio.F90.in at line 2. Any idea?

 

[oleson]

When you are posting to the forums because you want help with an issue, please include as much of the following information as possible.

Always:

• First look through the CIME troubleshooting guide to see if any of the suggestions there solve your problem. (If you think of ways we could improve the troubleshooting guide, please give us feedback.)
• Document the version of the code are you using: the output from running git describe from the top level of the CESM clone, and also the output from running ./manage_externals/checkout_externals --status --verbose
• If you have made any changes to files in the source tree (code, xml files, etc.), please make those changes available (or at least describe them in your help request)
• Describe every step you took, starting with the create_newcase command and including any changes you made to xml files (via xmlchange or hand edits), user_nl files, or any other changes in your case. (Note that the create_newcase command can be found in README.case, and any changes made via xmlchange can be found in CaseStatus.) It is best if you try to reproduce the problem first using your own instructions, to make sure that your instructions are complete and accurate.

If this is a port to a new machine:

• Include your config_compilers.xml, config_machines.xml, and config_batch.xml files
• Document the compiler version you are using on this machine

If the error occurs in the model run:

• Attach all log files from the run (cpl.log, cesm.log and all component log files)

 

[MrIgnition]

Thank you for this guide. I am actually reading it for the first time. As advised, I went through each component log and I actually got more information from my lnd.log file. It mentions ERROR in vsize after opening the surface file. Cant specify what vsize means, but I think I should include my preview_run in case the vsize is referring to virtual memory. I have attached the lnd.log and preview run files for your perusal.
Thanks once again for the guide.

 

[oleson]

It would be helpful to attach all of the log files, particularly the cesm.log....

 

[MrIgnition]

Ok. thanks, they are hereby attached.

 

[oleson]

What do you get when you do this:

Document the version of the code are you using: the output from running git describe from the top level of the CESM clone, and also the output from running ./manage_externals --status --verbose

 

[oleson]

Also, can you post your "lnd_in"?

 

[MrIgnition]

my release is release-clm5.0.28-1-gd5f3086
my manage_externals as found in clm5.0 directory is a directory so i couldn't run ./manage_externals --status --verbose.
I am working on getting the lnd_in.
Thanks.

 

[MrIgnition]

Here is lnd.in

 

[MrIgnition]

Here is lnd.in

Also doing git describe from my clm5.0 directory shows:
heads/release-clm5.0

 

[MrIgnition]

Also doing git describe from my clm5.0 directory shows:
heads/release-clm5.0

Externals.cfg

 

[oleson]

I'm not sure, but it looks like the GLACIER_REGION field on your surface dataset might be inconsistent with the domain size.
Seems like it should be dimensioned as (lsmlat=2160, lsmlon=4320), but might be something different.
Note that you can find a traceback of where in the code the error is occurring by looking in the cesm.log:

ERROR:
ERROR in /p/project/cjicg41/bamidele/clm5.0/src/main/ncdio_pio.F90.in at line 2388
Image PC Routine Line Source
cesm.exe 00000000012515FA Unknown Unknown Unknown
cesm.exe 0000000000D9BEBE shr_abort_mod_mp_ 114 shr_abort_mod.F90
cesm.exe 00000000005FF38C ncdio_pio_mp_ncd_ 2388 ncdio_pio.F90.in
cesm.exe 00000000005FA890 ncdio_pio_mp_ncd_ 1619 ncdio_pio.F90.in
cesm.exe 000000000056F822 glcbehaviormod_mp 659 glcBehaviorMod.F90
cesm.exe 0000000000573D9D glcbehaviormod_mp 217 glcBehaviorMod.F90
cesm.exe 000000000051B77C clm_initializemod 150 clm_initializeMod.F90
cesm.exe 0000000000502BF5 lnd_comp_mct_mp_l 199 lnd_comp_mct.F90
cesm.exe 0000000000435CD9 component_mod_mp_ 267 component_mod.F90
cesm.exe 0000000000424CF4 cime_comp_mod_mp_ 1237 cime_comp_mod.F90
cesm.exe 0000000000432E59 MAIN__ 114 cime_driver.F90

 

[MrIgnition]

Ok. I got more light on this. The code snippet causing the eror:
vsize = fullsize / gsize
if (mod(fullsize,gsize) /= 0) then
write(iulog,*) subname,' ERROR in vsize ',fullsize,gsize,vsize
call shr_sys_abort(errMsg(sourcefile, __LINE__))
endif
fullsize is the total number of grid cells in my domaon 260*260 hence 67600
gsize is "global size of clmlevel"
and the machine is looking for the remainder of fullsize/gsize which in my case is not applicable because my fullsize is smaller than my gzise. What does this infer? Am I running clm5 at a horizontal resolution that is too coarse? or my selected domain is too small for clm?
Or better still, can I reduce the gsize ("global size of clm level") ?

 

[sacks]

Is it possible that you copied GLACIER_REGION from another surface dataset onto this one? As @oleson mentioned, you should make sure that GLACIER_REGION has the same dimensions as other fields for this domain - in this case, 260x260. If this is a run over Nigeria, then you probably don't care about glaciers, and so you can set this field to 0 everywhere.

 

[MrIgnition]

Ok. I'll try that. Not sure how to do that though. I'll get back on the outcome. Thanks, I'm grateful.

 

[MrIgnition]

Is it possible that you copied GLACIER_REGION from another surface dataset onto this one? As @oleson mentioned, you should make sure that GLACIER_REGION has the same dimensions as other fields for this domain - in this case, 260x260. If this is a run over Nigeria, then you probably don't care about glaciers, and so you can set this field to 0 everywhere.

Kindly take a look at this ncview images. My GLACIER_REGION is already 260x260 and its all set to value 0. A look at the last instructions that caused the error:
vsize = fullsize / gsize
if (mod(fullsize,gsize) /= 0) then
write(iulog,*) subname,' ERROR in vsize ',fullsize,gsize,vsize
call shr_sys_abort(errMsg(sourcefile, __LINE__))
endif
especially the second line in this code snippet already shows what may not allow my simulation go through. my fullsize is 67600 while gsize is 9331200. Meanwhile the code is looking for the remainder of fullsize/gsize to be 0. This makes it impossible to gave a valid value for if (mod(fullsize,gsize) /= 0) then ... so my question is, how do i pass through this? or is this issue spectacular to me? Please help me.

 

[sacks]

Sorry, I got my earlier suggestion backwards; I think @oleson's suggestion was more correct. Reading through this again, and looking more closely at the code, it looks like GLACIER_REGION is the first field read from the surface dataset, so this is just an indication that your surface dataset is inconsistent with your domain file. From your lnd.log file, it looks like your domain file (domain.lnd.Nigeria005.191223.nc) has dimensions 4320*2160, but it sounds like your intent is to run over a domain that is 260x260. Can you double-check the consistency of the dimensions on your different input files - especially your domain and surface datasets? If you can't figure this out yourself, can you please do an ncdump -h of each of the .nc files you are pointing to in lnd_in and attach those netcdf headers?

 

[MrIgnition]

Ok. I will do this immediately.

 

[MrIgnition]

Thanks, I have uploaded the files as advised. ncdump -h does not work on my supercomputers so i have used ncks -M. I believe it does the magic as well.
Thanks once again.

 

[sacks]

Thanks. Indeed, there is an inconsistency in dimension sizes between fsurdat and fatmlndfrc. You need to ensure that the domain sizes are the same for both of these files.