Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Add new unit tests sumchk and bcstchk and update tests (#606)
* Update testing - Set default debug_model_step=0 (was 99999999) - Add debug_model_[i,j,iblk,task] to define the debug_model diagnostic point in local grid index space. If this point is not set and debug_model is turned on, it will use lonpnt(1),latpnt(1). - Rename forcing_diag namelist/variable to debug_forcing to be more consistent with other "debug_" namelist variables - Rename the local variable forcing_debug in ice_forcing.F90 to local_debug to avoid confusion with global varaible debug_forcing. - Add namelist variable optics_file. Was hardwired in ice_forcing_bgc.F90 - Update optics file variable name to read, still hardwired in model. - Update setting of nbtrcr_sw and allocation of trcrn_sw. nbtrcr_sw was not set in icepack after it was computed and trcrn_sw was allocated before nbtrcr_sw was computed. This impacts the dedd_algae implementation which still isn't working. - move default distribution_wgt_file for gx1 to set_nml.gx1 - update test suite, add testing of debug_model_[i,j,iblk,task], add addtional testing of maskhalo - update documentation * add sumchk unit test to test global reduction methods * - add bcstchk unit test - update ice_broadcast to sync up serial and mpi versions - add get_rank to ice_communicate.F90 - add global_[min/max]val_scalar_int_nodist method to ice_global_reductions.F90 - add tripole output in ice_blocks.F90 with debug_blocks - update set_nml.tx1 to set ns_boundary_type to 'tripole', was 'open' * update lsum16 to revert to double precision if NO_R16 is set * sync up serial ice_global_reductions.F90 * - add optics_file_fieldname namelist - add grid_type and ns_boundary_type tripole check - update sumchk unit test to check both Nface and center points. these are treated differently for tripole grids. - update documentation of unit tests
- Loading branch information