Eaton had problem with ncdata stability with ne30np4

brian-eaton commented 5 years ago

I went back to check this test and I see that I didn't use the latest ncdata file that you had provided. So I'm rerunning with that file when cheyenne comes back. The test was with FW2000climo on grid ne30_ne30_mg17 with --pecount set to 216x1.

FYI I copied that file to a standard location and renamed:

/glade/p/cgd/amp/pel/inic/waccm.i.spinup.nc copied to: /glade/p/cesmdata/cseg/inputdata/atm/waccm/ic/ waccm-tsmlt-mam4_2000_ne30np4_L70_0002-01-01_c181221.nc'

brian-eaton commented 5 years ago

I have rerun this test with the spunup IC file and am getting the following failure:

NEGATIVE LAYER THICKNESS DIAGNOSTICS:

(lon,lat) = 71.17 40.44

k,dp_star_moist,u,v,T 1 0.00016256 172.04 -49.12 530.20 ... ERROR: negative moist layer thickness. timestep or remap time too large Image PC Routine Line Source
cesm.exe 000000000B2884BD Unknown Unknown Unknown cesm.exe 000000000A5781A1 shr_abort_modmp 114 shr_abort_mod.F90 cesm.exe 000000000A578022 shr_abort_modmp 61 shr_abort_mod.F90 cesm.exe 0000000004BF8780 prim_advection_mo 1053 prim_advection_mod.F90 cesm.exe 0000000001242171 prim_driver_mod_m 282 prim_driver_mod.F90 cesm.exe 0000000000A8D063 dyn_comp_mp_dyn_r 986 dyn_comp.F90

PeterHjortLauritzen commented 5 years ago

Got it ... fixed with se_nsplit=12 change for ne30_ne30

PeterHjortLauritzen / opt-se-cslam

Eaton had problem with ncdata stability with ne30np4 #33