Hello,
With Cheyenne's retirement around the corner, I decided to move my work with iCESM onto my institution's cluster, where my team has a working port of the model. I am trying to test my configuration by running a branch of my Cheyenne case on the new machine. I created a new case based on the original case's compset and resolution, and I imported the original case's user namelists and sourcemods, along with all the necessary input files, and my most recent restart files from my work on Cheyenne. However, when the model begins running, POP produces this error:
From my previous experiences, POP fails like this after the first day of running when it is passed bad boundary conditions. In fact, with the debugging output turned on, I can see that several physical variables ("Si_ifrac", "Foxx_tau", "Foxx_lat", "Foxx_sen", "Foxx_lwu", "Fioi_mel", and "Fioi_sal") send NaN values to the coupler, so it looks like this is the case. However, I have no idea what could be bad about my configuration. I did not change anything in the restart, and everything in my namelists, except for paths and directories relevant to my new workspace, is identical to what I used for my working case on Cheyenne. Does anyone know what could be going wrong?
With Cheyenne's retirement around the corner, I decided to move my work with iCESM onto my institution's cluster, where my team has a working port of the model. I am trying to test my configuration by running a branch of my Cheyenne case on the new machine. I created a new case based on the original case's compset and resolution, and I imported the original case's user namelists and sourcemods, along with all the necessary input files, and my most recent restart files from my work on Cheyenne. However, when the model begins running, POP produces this error:
Code:
POP Exiting...
POP_SolversChronGear: solver not converged
POP_SolverRun: error in ChronGear
POP_BarotropicDriver: error in solver
Step: error in barotropicFrom my previous experiences, POP fails like this after the first day of running when it is passed bad boundary conditions. In fact, with the debugging output turned on, I can see that several physical variables ("Si_ifrac", "Foxx_tau", "Foxx_lat", "Foxx_sen", "Foxx_lwu", "Fioi_mel", and "Fioi_sal") send NaN values to the coupler, so it looks like this is the case. However, I have no idea what could be bad about my configuration. I did not change anything in the restart, and everything in my namelists, except for paths and directories relevant to my new workspace, is identical to what I used for my working case on Cheyenne. Does anyone know what could be going wrong?