Using the public release App with 127 vertical levels

The planned implementation of GFS v16 uses 127 vertical levels. What would be involved in running the public release with that configuration? I assume that chgres_cube would have to output more levels, that the model namelist would have to change, and that possibly also the computational resources would need to be different.

I assume that these changes would need to be made in the ufs-weather-model first and then the app would need to make the required changes in the buildlib and buildnml scripts. 

The ufs-weather-model itself can already run with 127 L as long as the namelist is set accordingly. 

However, I don't know what would be involved in chgres_cube outputting ICs for 127L.

Hello,

I did a test run for 127 levels. By including the below two options in user_nl_ufsatm, the 127 levels IC files are successfully generated from chgres from CIME workflow

levp = 128. (for chgres)

 npz = 127 (for ufs)

The ufs.exe crashes at the case.run step, indicating error of "out of memory". I see that tasks and tasks per node for case.chgres and case.gfs_post tasks can be changed in env_workflow.xml.

Can you suggest to me where I can change the tasks, nodes, tasks per node for case.run step?

Thanks!