4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
5 |
============================== |
============================== |
6 |
|
|
7 |
|
o pkg/exf: |
8 |
|
- interpolation with #undef EXF_INTERP_USE_DYNALLOC: move buffer size |
9 |
|
definition outside EXF_PARAM.h in new header file: EXF_INTERP_SIZE.h; |
10 |
|
use fixed size buffer (passed to S/R EXF_INTERP & EXF_INTERP_UV) to |
11 |
|
avoid any dynamic allocation ; |
12 |
|
- S/R EXF_CHECK: add new check & stop when file-name is defined but not |
13 |
|
read-in ; add warning when file is read-in but not used within pkg/exf ; |
14 |
|
+ check for buffer-size (if #undef EXF_INTERP_USE_DYNALLOC). |
15 |
|
- update exf_summary.F (print {FLD}RepCycle). |
16 |
|
o verification/global_with_exf: |
17 |
|
- switch on exf interpolation in AD test (just like the forward test) |
18 |
|
o pkg/mom_vecinv: |
19 |
|
- add storage directives to avoid recomputations within mom_vecinv.F |
20 |
|
which were generating seg. fault with gfortran -devel |
21 |
|
o pkg/ctrl, pkg/smooth: |
22 |
|
- initialize array halos which were generating seg. fault with |
23 |
|
gfortran -devel in adjoint mode |
24 |
|
o pkg/exf: |
25 |
|
- for each input field, enable to specify a field specific repeat-cycle |
26 |
|
({FLD}RepCycle) which defaults to pkg/exf main "repeatPeriod". |
27 |
|
o verification/lab_sea: |
28 |
|
- update AD results of primary test and noseaicedyn test after switching |
29 |
|
to new version of TAF (5.3.13 vs 5.3.2 before); agreement with gradient |
30 |
|
check stays about the same. |
31 |
|
o pkg/exf: |
32 |
|
- fix bug in exf_zenithangle.F for cases select_ZenAlbedo > 0 with useCAL=T. |
33 |
|
- exf_check: MAX_LAT_INC need only be "loc_nlat-1", not equal to loc_nlat |
34 |
|
o pkg/bling, USE_EXFCO2 code: |
35 |
|
- add parameter "apco2RepCycle" (set to pkg/exf "repeatPeriod" by default); |
36 |
|
- update calls to EXF S/R (use new version: EXF_INIT_FLD & EXF_SET_FLD). |
37 |
|
o pkg/exf: |
38 |
|
- rename S/R EXF_INIT to EXF_INIT_VARIA ; |
39 |
|
- add arguments: fldName (for debug msg) and fldRepeatCycle (set to |
40 |
|
repeatPeriod) to S/R EXF_SET_UV and to S/R EXF_SET_GEN (renamed EXF_SET_FLD) |
41 |
|
and update calls in pkg/exf calling S/R (exf_getffields.F & exf_getclim.F) ; |
42 |
|
- add argument: fldName (for debug msg) to S/R EXF_INIT_GEN (renamed |
43 |
|
EXF_INIT_FLD) and update calls in exf_init_varia.F |
44 |
|
- For now, for backward compatibility, keep a simplified version of S/Rs |
45 |
|
EXF_SET_GEN & EXF_INIT_GEN in file exf_set_gen.F |
46 |
|
- add and pass new arguments: fldName (for debug msg) and fldRepeatCycle |
47 |
|
to S/R EXF_GetFFieldRec ; update calls in pkg/obcs/obcs_exf_load.F |
48 |
|
o tools/genmake2: |
49 |
|
- replace "bc -l" (only used once) by "expr" (already used). |
50 |
|
o verification/global_ocean.cs32x15: |
51 |
|
- turn off pkg/cal (useCAL=F) in secondary FWD exp. "icedyn". |
52 |
|
Note: exf setting (data.exf) can be compared to other exf secondary exp |
53 |
|
"seaice" (kept with useCAL=T) which has same time dependent forcing. |
54 |
|
|
55 |
|
checkpoint66d (2017/02/13) |
56 |
|
o pkg/exf: |
57 |
|
- remove setting of useCAL=T when using pkg/exf (useEXF=T); add warning |
58 |
|
early on (in packages_boot.F) to stress this change in pkg setting. |
59 |
|
- In (most) verification exp. that use pkg/exf, add useCAL=.TRUE. to data.pkg |
60 |
|
o model/src: |
61 |
|
- move call to ecco_phys to end of time step; this may induce minor |
62 |
|
cost function changes by shifting time averages by one time step |
63 |
|
for some variables; this revision resulted in changed adjoint results |
64 |
|
in MITgcm_contrib/verification_other/global_oce_cs32 |
65 |
|
o pkg/exf: |
66 |
|
- Change EXF_READ_EVAP to only activate code within #ifdef ALLOW_ATM_TEMP ; |
67 |
|
this prevents to read evap when ALLOW_ATM_TEMP is undef and other fields, |
68 |
|
e.g. precip, are not available. |
69 |
|
o pkg/flt: |
70 |
|
- flt_init_varia.F: use pickupSuff if specified |
71 |
|
- add code (within #ifdef DEVEL_FLT_EXCH2) and test case as a first, |
72 |
|
intermediate, step towards completing pkg/exch2 support within pkg/flt; |
73 |
|
ifdef DEVEL_FLT_EXCH2 then: |
74 |
|
- flt_init_fixed.F: comment out stop |
75 |
|
- flt_exch2.F: conserve ipart, jpart (grid coordinates) |
76 |
|
- exch2_recv_get_vec.F: transform ipart, jpart (grid coordinates) |
77 |
|
- flt_mapping.F: attempt to fix longitude diagnostic near date-line |
78 |
|
- flt_init_varia.F: hack initialization (for testing purposes w. LLC90) |
79 |
|
o tools/genmake2 & pkg/pkg_depend: |
80 |
|
- define new type of pkg dependency rule ("=pkg_x"), less strict than |
81 |
|
the "+" rule, that add the dependent pkg (pkg_x) unless it's disabled |
82 |
|
(-pkg_x in packages.conf). |
83 |
|
- change pkg/exf dependency rule from "exf +cal" to "exf =cal"; |
84 |
|
- add "+cal" to dependency list of pkg/ecco (needed to compile). |
85 |
|
- simplify FWD experiment "offline_exf_seaice" by removing pkg/cal. |
86 |
|
o utils/matlab: |
87 |
|
- adding ocean_basin routines for defining basin masks. |
88 |
|
o pkg/exf: |
89 |
|
- new S/R EXF_GETFFIELD_START replacing EXF_GETFIELD_START, but with more |
90 |
|
arguments, allowing to check (and stop) if parameter setting of |
91 |
|
{FLD}startdate1 & date2 and of {FLD}StartTime are consistent with |
92 |
|
useCAL= T or F. For now, kept a simplified version of EXF_GETFIELD_START |
93 |
|
for backward compatibility. |
94 |
|
- change accordingly calls in exf_init_fixed.F + default {FLD}StartTime in |
95 |
|
exf_readparms.F ; apply same changes to pkg/icefront (for SGRunOff field). |
96 |
|
o pkg/bling (USE_EXFCO2 code): |
97 |
|
- fix missing setting of apco2StartTime according to apco2startdate1 & date2 |
98 |
|
by adding a call to S/R EXF_GETFFIELD_START. |
99 |
|
o pkg/exf: |
100 |
|
- allows to compile without pkg/cal (but disabled by current dependency |
101 |
|
rules, i.e., default "pkg/pkg_depend"). Minimal changes to enable to |
102 |
|
run with useCAL=F ; this involves setting directly {FLD}StartTime instead |
103 |
|
of {FLD}startdate1 & date2 ; consistency check not yet coded. |
104 |
|
o verification/global_ocean.cs32x15: |
105 |
|
- fix all data.exf for a shift of 1 day in all field startdate1 so that it now |
106 |
|
corresponds to middle of 1rst month; update outputs (2 fwd + 3 adm + 3 tlm). |
107 |
|
o pkg/exf: |
108 |
|
- allow to use EXF_ZENITHANGLE without pkg/cal + simplify (saving some |
109 |
|
calculations); change FWD experiment global_ocean.cs32x15 |
110 |
|
(#define ALLOW_ZENITHANGLE) to test/check that this code compiles. |
111 |
|
o pkg/exf: |
112 |
|
- In preparation for addition to NAMELIST, rename internal (_RL) parameter |
113 |
|
{FIELD}startdate to {FIELD}StartTime |
114 |
|
Note: new name also fit better the content, i.e, a start time in second. |
115 |
|
o eesupp/src: |
116 |
|
- add 1 argument to S/R EEBOOT & EEBOOT_MINIMAL for interface with ESMF |
117 |
|
o pkg/exf: |
118 |
|
- add option ALLOW_READ_TURBFLUXES (and corresponding set of parameters) |
119 |
|
to allow to read-in Sensible & Latent Heat flux ; |
120 |
|
- simplify pkg/exf main options: |
121 |
|
1) move small piece of code (case ALLOW_ATM_TEMP undef) out of |
122 |
|
exf_bulkformulae.F into exf_wind.F so that exf_bulkformulae.F is |
123 |
|
compiled only when both ALLOW_ATM_TEMP & ALLOW_BULKFORMULAE are defined; |
124 |
|
2) make call to EXF_RADIATION and to EXF_WIND independent of CPP option |
125 |
|
ALLOW_BULKFORMULAE which now only applies to S/R EXF_BULKFORMULAE ; |
126 |
|
3) improve options documentation in EXF_OPTIONS.h. |
127 |
|
o pkg/ecco: |
128 |
|
- cost_gencost_customize.F,ecco.h,ecco_phys.F,ecco_readparms.F, and |
129 |
|
ecco_summary.F: add passive tracer cost functions to gencost; |
130 |
|
m_ptracer and m_boxmean_ptracer |
131 |
|
o pkg/ctrl: |
132 |
|
- ctrl_map_forcing.F: pass myIter and myTime to ctrl_map_forcing as input |
133 |
|
parameters; move myIter .EQ. nIter0 inside ctrl_map_forcing; add to list |
134 |
|
of implemented gentim2d contol parameters: xx_qnet, xx_empmr, xx_qsw, |
135 |
|
xx_sst, xx_sss, xx_pload, xx_saltflux, xx_fu, xx_fv, xx_fe, xx_fn |
136 |
|
o model/src: |
137 |
|
- forward_step.F: pass myIter and myTime to ctrl_map_forcing as input |
138 |
|
parameters; move myIter .EQ. nIter0 inside ctrl_map_forcing |
139 |
|
o pkg/exf: |
140 |
|
- exf_getffields.F: add to list of implemented gentim2d contol parameters: |
141 |
|
xx_lwflux, xx_swflux, xx_evap, xx_apressure, xx_areamask |
142 |
|
|
143 |
|
checkpoint66c (2017/01/19) |
144 |
|
o verification: |
145 |
|
- add new option "-src" to testreport to only produce source code (*.f), |
146 |
|
including taf output src code (if using '-adm'/'-tlm' option). |
147 |
|
o pkg/exf: |
148 |
|
- add saltflx :: Net upward salt flux in psu.kg/m^2/s |
149 |
|
o pkg/diagnostics: |
150 |
|
- add run-time variable diagMdsDir to specify a subdirectory for |
151 |
|
diagnostic output when mds is being used and mdsioLocalDir is not. |
152 |
|
o pkg/mnc: |
153 |
|
- mnc_utils.F: edit stop message |
154 |
|
o pkg/autodiff: |
155 |
|
- autodiff_ini_model_io.F: PRINT_MESSAGE, clean-up comments, add _*_MASTER, |
156 |
|
add HAVE_SYSTEM bracket around system call to mkdir |
157 |
|
o pkg/profiles: |
158 |
|
- profiles_findunit.F, profiles_init_fixed.F, profiles_interp.F, |
159 |
|
profiles_readparms.F: call ALL_PROC_DIE before 'stop' |
160 |
|
- profiles_ini_io.F: use PRINT_MESSAGE, clean-up comments, add _*_MASTER, |
161 |
|
add HAVE_SYSTEM bracket around system call to mkdir |
162 |
|
o pkg/smooth: |
163 |
|
- smooth_filtervar2d.F, smooth_filtervar3d.F: call ALL_PROC_DIE before 'stop' |
164 |
|
- smooth_init2d.F, smooth_init3d.F: move print statements to readparms |
165 |
|
- smooth_readparms.F: add cleaner print statements from smooth_init?d.F |
166 |
|
o pkg/exf: |
167 |
|
- check: count errors and stop cleanly at the end (call to ALL_PROC_DIE); |
168 |
|
issue error message when climsst_interpMethod.GT.0 and climsst_lon_inc > 500 |
169 |
|
generalize tests on interpMethod, nlat, and lon_inc to all exf variables. |
170 |
|
- summary: skip detailed field report if not used (empty file-name). |
171 |
|
o tools/xmakedepend: |
172 |
|
- In case line editor "ed" is missing, switch to "sed" command to clean-up |
173 |
|
Makefile from previous dependencies (anything after "# DO NOT DELETE" line). |
174 |
|
o pkg/exf: |
175 |
|
- change definition of "hflux" input file to always include Short-Wave; |
176 |
|
this is now consistent with hflux computation (if using bulk-formulae) |
177 |
|
and hflux diagnostic & documentation (in EXF_FIELDS.h). |
178 |
|
|
179 |
|
checkpoint66b (2016/12/01) |
180 |
|
o model & momentum pkgs: |
181 |
|
- implement fully implicit bottom friction combined with implicit surf. |
182 |
|
pressure, for now, only in hydrostic formulation: |
183 |
|
+ to use: compile with #define ALLOW_SOLVE4_PS_AND_DRAG in CPP_OPTIONS.h |
184 |
|
and run with selectImplicitDrag=2 ; |
185 |
|
+ currently missing diagnostics of bottom drag and vert. viscosity tendency; |
186 |
|
- partial implicit time-stepping (bot. friction + vert. visc first, then |
187 |
|
surf. press) corresponding to selectImplicitDrag=1 should be applicable |
188 |
|
to NH formulation but not available until we solve for surf. pressure |
189 |
|
increment (instead of full etaN). |
190 |
|
- change exp. global_ocean.cs32x15.thsice to test implicit bottom drag |
191 |
|
with implicit vertical viscosity. |
192 |
o pkg/seaice: add linear remapping scheme (Lipscomb 2001) to ITD model |
o pkg/seaice: add linear remapping scheme (Lipscomb 2001) to ITD model |
193 |
- this is a first version that includes many regularisations that may |
- this is a first version that includes many regularisations that may |
194 |
not all be necessary, to be sorted out, |
not all be necessary, to be sorted out, |
195 |
- turned on by runtime parameter SEAICEuseLinRemapITD (true by default) |
- turned on by runtime parameter SEAICEuseLinRemapITD (true by default) |
196 |
- change one verification experiment in seaice_itd to use and test new code |
- change one verification experiment in seaice_itd to use and test new code |
197 |
- seaice_growth: catch a case where the argument of log can be zero |
- seaice_growth: catch a case where the argument of log can be zero |
198 |
or smaller (important for code stability) |
or smaller (important for code stability) |
199 |
o model/src: |
o model/src: |
200 |
- modify SOLVE_DIAGONAL_LOWMEMORY version of solve_tridiagonal.F |
- modify SOLVE_DIAGONAL_LOWMEMORY version of solve_tridiagonal.F |