MPAS-8.0.1 - Floating overflow - error (72)

[makinde]

Dear all,

I am currently facing an issue while running the atmosphere_model component of the new MPAS-8.0.1. Approximately two weeks ago, our technical support team assisted me in successfully compiling MPAS-8.0.01 using Intel compilers and other necessary requirements, including parallel I/O.

However, my attempts to run the atmosphere_model at a 60 km uniform resolution have consistently resulted in the same error: “forrtl: error (72): floating overflow.” Despite compiling the model in debug mode, I have not been able to extract useful information regarding the root cause of this error.


More detail:
I have successfully executed the init_atmosphere step, generating all the initial conditions (init.nc, static.nc, and sfc_update.nc) without encountering any error messages. The issue arises during the model integration phase when running the atmosphere_model using the outputs from the initialization step. In an effort to troubleshoot, I have experimented with modifying certain parameters in my namelist.atmosphere. For instance, I adjusted the config_dt from the usual 300 seconds (used with the 60 km mesh) to 360 seconds and even tested with 150 seconds, but unfortunately, the error persists.

Interestingly, the atmosphere_model ran successfully when I tested it with the supercell idealized case but will not run real-data. I have attached all the relevant namelist and stream files that I used for both the init_atmosphere_model and atmosphere_model.

Could you kindly provide any insights or guidance on resolving this issue? Your assistance would be greatly appreciated.
Thank you in advance for your time and expertise.



Best regards,
Makinde

 

[gdicker]

Hello @makinde,

Apologies that there hasn't been a reply before now. However, I am still lacking some details that would help me troubleshoot this.

You indicated that you would attach some files but I don't see any. Could you please try attaching them again? Along with the namelist or streams files, could you please include any log.atmosphere.0000.out, log.atmosphere.0000.err, and the (stdout, stderr) output from running the model? These would help to give context to your attempts (especially the line number for the forrtl: error(72) ...).

 

[makinde]

Hello @makinde,

Apologies that there hasn't been a reply before now. However, I am still lacking some details that would help me troubleshoot this.

You indicated that you would attach some files but I don't see any. Could you please try attaching them again? Along with the namelist or streams files, could you please include any log.atmosphere.0000.out, log.atmosphere.0000.err, and the (stdout, stderr) output from running the model? These would help to give context to your attempts (especially the line number for the forrtl: error(72) ...).

Thank you @gdicker for the response and sorry for the missing attachment. Please find attached the namelist files, streams file and one of the log.atmosphere.0000.out.

Though I have re-compiled the model (MPAS-8.0.1) several times again, I had luck with a working model when I allowed MPAS to use the default SMIOL instead of PIO. Previously, we had compiled MPAS-7.x.x with PIO and were all working fine. However, I am still unsure what caused the "Floating overflow - error (72)".