4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
5 |
============================== |
============================== |
6 |
|
|
7 |
|
o adjustment.cs-32x32x1: changed to oceanic test with 1 continent |
8 |
|
+ use 48 tiles (16x8): this will allow to test "blank tiles". |
9 |
|
o start to test MPI + multi-threaded (change testreport to skip tests |
10 |
|
where number of tiles do not match). |
11 |
|
o genmake2: add option "-omp" to turn on Open-MP ; |
12 |
|
passed to testreport (if using -mth) ; merge few (*+mth) optfiles. |
13 |
|
o pkg/exch2 (+ utils/exch2): ignore blank tiles (instead of removing them) |
14 |
|
and keep the original tile numbering. Note: compatible with old |
15 |
|
W2_EXCH2_TOPOLOGY.h & w2_e2setup.F topology files. |
16 |
|
|
17 |
|
checkpoint61l |
18 |
|
o Overlaps had been forgotten in calculating ijk keys : fixed |
19 |
|
o fix cyrus makedepend for gcc 4.4 |
20 |
|
o experiment fizhi-cs-aqualev20: |
21 |
|
switching to new version of S/R GETPWHERE (#define TRY_NEW_GETPWHERE) |
22 |
|
and generate new output. Note: new output is much closer to the old |
23 |
|
one (v1.8) generated with pgf77 -O2 on Mar 14 2007 than it is from |
24 |
|
the previous one (v1.9) generated with pgf77 -O0 on Jun 13 2007. |
25 |
|
o pkg/fizhi: |
26 |
|
- change to avoid indices going deliberately over array-bounds |
27 |
|
- remove unnecessary MPI stuff |
28 |
|
- perpetual spring equinox with "#define FIZHI_USE_FIXED_DAY" |
29 |
|
- new version of S/R GETPWHERE available with #define TRY_NEW_GETPWHERE: |
30 |
|
does not go over array-bounds and therefore returns reliable output. |
31 |
|
- in fizhi_driver.F & fizhi_turb.F, fix "get_alarm" calls: |
32 |
|
this S/R has 4 output arguments, which are used inside "get_alarm"; |
33 |
|
when it is called with the same variable "ndum" in place of several |
34 |
|
output arguments, this produces weird side effect inside "get_alarm" |
35 |
|
(e.g., datein=dates(n) is overwritten by timein=times(n) since both |
36 |
|
datein & timein, resp. 2nd and 3rd arg., are pointed to the same "ndum" |
37 |
|
address) and produces error. |
38 |
|
o lab_sea experiment: change from 2x1 tiles to 2x2 tiles and regenerate output |
39 |
|
o pkg/seaice: fix bug in new version of S/R SEAICE_LSR.F: |
40 |
|
update lab_sea.lsr & lab_sea.salt_plume results |
41 |
|
o pkg/seaice: |
42 |
|
Major overhaul of the LSOR solver: change discretization from original |
43 |
|
finite differences to finite volume. The old routine is still in the |
44 |
|
file seaice_lsr.F and can be turned on with the CPP-flag |
45 |
|
SEAICE_OLD_AND_BAD_DISCRETIZATION |
46 |
|
Along with this change, the discretization of the strain rates is changed. |
47 |
|
The same flag turns on the old discretization in seaice_calc_strainrates.F |
48 |
|
All verification experiments with seaice are affected (except for |
49 |
|
seaice_exf_offline); global_ocean.cs32x15/input.icedyn recovers the old |
50 |
|
results by setting the above flag. |
51 |
|
Further changes involve changes in defaults: |
52 |
|
- ALLOW_FLOODING is defined and turned on by default |
53 |
|
- SEAICEadvSnow = .true. is now the default |
54 |
|
- SEAICEadvSalt = .true. is now the default |
55 |
|
- SEAICEadvAge = .true. is now the default |
56 |
|
- SEAICE_clipVelocities = .false. is now the default |
57 |
|
(as per J. Zhang's recommendation) |
58 |
|
- B-grid, and thus not tested: SEAICE_TEST_ICE_STRESS_1/EXPLICIT_SSH_SLOPE |
59 |
|
is defined, SEAICE_TEST_ICE_STRESS_1 is renamed into SEAICE_BICE_STRESS |
60 |
|
- seaice_growth: replace computation of UG by a simple copy from wspeed |
61 |
|
|
62 |
|
checkpoint61k |
63 |
|
o pkg/seaice: a few bug fixes mostly concerning the metric terms: |
64 |
|
seaice_lsr: fixed a few metric terms, |
65 |
|
affects lab_sea.lsr, lab_sea.salt_plume, seaice_obcs |
66 |
|
add more debugging diagnostics |
67 |
|
seaice_evp: removed superfluous many metric terms of FV discretization |
68 |
|
of stress divergence |
69 |
|
affects lab_sea, lab_sea.hb87 |
70 |
|
seaice_ocean_stress: removed superfluous many metric terms of FV discretization |
71 |
|
of stress divergence for non-evp solver |
72 |
|
(this part is not tested anywhere) |
73 |
|
seaice_calc_strainrates: fixed a bug in no-slip boundary conditions, |
74 |
|
they were effectively only half slip |
75 |
|
affects lab_sea.hb87 |
76 |
|
updated all relevant verification experiments (except lab_sea//output_adm.txt |
77 |
|
which was left over, but fails at level 7 & 6) |
78 |
|
|
79 |
|
checkpoint61j |
80 |
|
o pkg/seaice: turn a few hard wired parameters into run time parameters |
81 |
|
o pkg/seaice/exf/monitor: turn netcdf coordinate variable T into model |
82 |
|
time in seconds as promised by the usints in seaice_monitor and exf_monitor |
83 |
|
o pkg/mnc: fix the coordinate variables for multi-tile curvi-linear grids |
84 |
|
that are not cubed-sphere grids |
85 |
|
o pkg/seaice: fix a bug for the rare case of SEAICE_VECTORIZE_LSR |
86 |
|
o pkg/ctrl: |
87 |
|
fix bug in ctrl_depth_ini.F (EXCH_XYZ was applied to 2-D field xx_r_low) |
88 |
|
o non-hydrostatic (calc_gw.F): |
89 |
|
fix viscA4 on CS-grid + fix diagnostics: Wm_Diss & Wm_Advec |
90 |
|
update output of exp. global_ocean.cs32x15.viscA4 |
91 |
|
o remove backward compatibility for thetaMin,phiMin |
92 |
|
(replaced by xgOrigin,ygOrigin on Jan 27, 2009) |
93 |
|
o pkg/exf: fix a serious bug in exf_set_uv.F: do a rotation of the |
94 |
|
interpolated wind field not only for curvilineargrid=.true., but also |
95 |
|
for a rotated grid (rotateGrid=.true.) |
96 |
o pkg/exf: add capability to bypass interpolation for runoff |
o pkg/exf: add capability to bypass interpolation for runoff |
97 |
by setting yet another CPP-flag |
by setting yet another CPP-flag |
98 |
|
o eesupp/src, B-grid exchange: |
99 |
|
add simple (6 tiles, non MPI) CS exch + standard interface S/R. |
100 |
o autodiff |
o autodiff |
101 |
new CPP option ALLOW_TAMC_SINGLEPREC_COMLEV |
new CPP option ALLOW_TAMC_SINGLEPREC_COMLEV |
102 |
allows to switch to single-prec common blocks; |
allows to switch to single-prec common blocks; |
112 |
Add flag to reduce all tapelev I/O to single-prec. |
Add flag to reduce all tapelev I/O to single-prec. |
113 |
In data.ctrl set |
In data.ctrl set |
114 |
doSinglePrecTapelev = .TRUE. |
doSinglePrecTapelev = .TRUE. |
115 |
o model/src/ini_curvilinear_grid.F: check in hack for SX8 compiler |
o model/src/ini_curvilinear_grid.F, new grid-file format without EXCH2: |
116 |
to make sure that recip_dxV and recip_dyU are never zero, until we |
check in hack for SX8 compiler (for dxV, dyU); remove it afterwards. |
|
have a proper exchange for them. Most compilers do not seem |
|
|
to mind divisions by zero in the overlaps, as long as the results are |
|
|
not used anywhere, but sxf90 does. |
|
117 |
o pkg/shelfice: fix bug with SHELFICEconserve flag, prepare for |
o pkg/shelfice: fix bug with SHELFICEconserve flag, prepare for |
118 |
real fresh water flux (not yet) |
real fresh water flux (not yet) |
119 |
o pkg/mdsio & rw: |
o pkg/mdsio & rw: |