search

CICE-Consortium / Icepack

Development repository for sea-ice column physics
Other
25 stars 131 forks source link

closing/flushing nu_diag at end #401

Open cmbitz opened 1 year ago

cmbitz commented 1 year ago

When I compiled with netcdf, I had to add a close(nu_diag) or flush(nu_diag) at the end of the program icedrv, otherwise I don't get the last output written to the log file, including the ICEPACK COMPLETED statement, and the stdio would echo "ICEPACK run did not complete".

I tried making it a call icepack_warnings_flush(nu_diag) but it didn't work

I'm using my macbook pro with Monterey OSX and gnu compiler. I didn't have this problem until I compiled with netcdf (didn't matter whether I wrote the history file though). With this change all is well getting it to run. Thanks, Cecilia

eclare108213 commented 1 year ago

HI CC, I also have been having this problem and hadn't fixed it. Which OSX are you running? I'm having trouble with Monterey 12.4 on my regular laptop, using the conda environment. The loaner laptop last week did not have this issue, running OSX Big Sur 11.6.2. If the close statement is enough to fix it, let's do that...

phil-blain commented 1 year ago

It's actually a bug in the OS, the close/flush statement should not be necessary in theory: https://bb.cgd.ucar.edu/cesm/threads/icepack-macos-monterey.6907/post-42716.

I agree that if we can work around it by adding a close statement, let's do it.

cmbitz commented 1 year ago

I'm running OSX Monterey 12.6. Thanks for entertaining the fix for me, though it is the OS's problem.

eclare108213 commented 1 year ago

There is still a problem with the grid-cell diagnostic files not being completely written before the program ends, so keeping this issue open for now.