/[MITgcm]/MITgcm/doc/tag-index
ViewVC logotype

Diff of /MITgcm/doc/tag-index

Parent Directory Parent Directory | Revision Log Revision Log | View Revision Graph Revision Graph | View Patch Patch

revision 1.1287 by mlosch, Mon Aug 2 10:34:13 2010 UTC revision 1.1321 by dimitri, Wed Oct 27 05:48:33 2010 UTC
# Line 4  $Name$ Line 4  $Name$
4      Notes on tags used in MITgcmUV      Notes on tags used in MITgcmUV
5      ==============================      ==============================
6    
7    o pkg/obcs:
8      - writing sections with globalFiles=T does not work: switch to tiled-files
9        when writing sections for OBCS (change pkg/rw/write_rec.F accordingly).
10      - fix index bug (OB_Ie -> OB_Iw_) in obcs_apply_uvice.F
11      - apply OBCS_SEAICE_COMPUTE_UVICE to both uice and vice at all obcs
12    
13    checkpoint62m (2010/10/21)
14    o pkg/seaice: vectorize seaice_solve4temp.F
15      - move iterations loops outside of i/j-loops
16      - requires 2D arrays for TB and effConduct
17      - add 2D iceOrNot to replace HICE_ACTUAL.GT.0, as HICE_ACTUAL is
18        modified in this routine
19      - a little bit of clean up, adjusting indentation, etc.
20      - replaced a few TEN**a by exp(a*log(10.)) in new code for better
21        performance and '**TWO' by '**2' (because TWO is a float),
22        replaced ONE, ZERO
23    o eesupp/src:
24       fix flow dir for exch1_z_RX_cube, exch1_uv_RX_cube & exch1_bg_RX_cube
25    o pkg/seaice/seaice_growth.F
26      Revision 1.89 of seaice_growth.F completes the overall october 2010 revision and
27      merging process. The merged routine has two branches, and one activates the 'evolution'
28      branch by setting SEAICE_GROWTH_LEGACY to undef (at the top of the routine for now).
29      In summary, here is what the '10/10 revision' did consist of
30        part 1: seaice_growth.F revisions 1.71 to 1.75
31          - Preliminary seaice_growth.F clean-up, keeping test results strictly unchanged.
32            The emphasis here was on improving readability (e.g. by trying to use meaningful
33            variable names) and modularity (e.g. by adding arrays to store heff etc. updates).
34        part 1.5:
35          - revising the lab_sea experiment to make it a more useful (i.e. stable
36            and representative of an actual model run). This was done by generating a
37            (winter) pickup to start from, switching back the main experiment dynamic
38            solver from evp to lsr, having one of the experiments use the free drift solution.
39        part 2: seaice_growth.F revisions 1.76 to 1.85
40          - Sorting out hard-coded constants that prevented heat and fresh water conservation.
41          - Sorting out the various fields units and sign (set early on and reset at the end)
42            to avoid numerous scalings/re-scalings. By v1.85 the basic unit for ice, heat and
43            fresh water stock is 'effective ice meters', and >0 implies more ice/snow.
44          - Using tendency terms arrays (d_HEFF* etc.) to alleviate remaining code locks,
45            and simplifying logical expressions to further improve readibility.
46        part 3: seaice_growth.F revisions 1.86 to 1.89
47          - Re-organize the blocks of code more logically.
48          - Merging the '_if.F' pieces that add something new.
49    o pkg/thsice:
50      - vectorize thsice_solve4temp and thsice_get_exf.
51        Problem: thsice_solve4temp has an unvectorizable statement that
52        might also be problematic for TAF
53      - thsice_get_exf: add the ALLOW_BULK_LARGEYEAGER04 CPP-flag. This
54        does not change the verifcation experiment, but will change all
55        experiments that have this flag undefined (not tested in verification);
56        a little reorganization to save a few 2D fields and recomputations
57      - 1st step of revisiting thsice adjoint (major recomp. sorted out)
58      - 2nd step:  fix some store dirs
59        o seems to work somewhat, i.e. using ifort but not g77
60        o likely cause are incomplete init. or recomputation
61        o start verification/global_ocean.cs32x15/input_ad.thsice
62    o pkg/obcs: add support for exch2
63      - the position of the boundary and prescribed values are specified using
64        a global domain with exch2 facets stacked
65        - in x for N,S boundaries (like W2_mapIO=-1)
66        - in y for E,W boundaries (so E,W boundaries do not overlap)
67    o pkg/mdsio: add support for writing global sections with exch2
68        file layout is as for obcs described above:
69        - for x-z sections, facets are stacked in x (like W2_mapIO=-1)
70        - for y-z sections, facets are stacked in y
71    o pkg/thsice:
72       - change a few routines for better vectorization: thsice_albedo.F,
73         thsice_step_temp.F, and in thsice_calc_thickn.F. Results are
74         not affected, but in particular the latter routine has changed
75         dramatically (not necessarily easier to read now).
76       - re-arrange calculation in THSICE_SOLVE4TEMP: move outside of i,j loops
77       iteration loop and call to S/R THSICE_GET_BULKF & THSICE_GET_EXF.
78    o verification/lab_sea:
79      - overhaul of the forward model experiments. The main motivation
80        for this revision is that lab_sea has proven quite fragile (when
81        changing the sea-ice code, and switching platforms) to a point
82        when failures were very hard to attribute to specific code problems.
83      - the suspected reason for this fragility is noise amplification by evp so,
84        to start with, we decided to replace evp with lsr in the main experiment.
85      - the input.lsr experiment thus became obsolete and was removed.
86      - it was replaced with the input.fd experiment that uses an
87        analytical "free drift" solution as the ice velocity field.
88      - Taken together, input (lsr), input.fd (free drift) and input.hb87 (evp)
89        will hopefully allow us to discriminate problems due to dynamic solvers
90        from problems due to the thermodynamic code. input.salt_plume (lsr) allows
91        us to test additional options, along with the salt plume package.
92      - Finally a pickup was generated (at the end of a one year run) to
93        make the tests more representative of an actual model run. Previously
94        the tests starting point was a 1m thickness of ice everywhere.
95    o pkg/seaice:
96      - introducing S\R seaice_freedrift.F that compute the
97        analytical solution to an approximate ice momentum
98        equation (i.e. the simple stress/coriolis balance).
99      - define SEAICE_ALLOW_FREEDRIFT to trigger the computation
100        of the free drift velocity field, and enable its use.
101        To actually use the free drift solution as the model ice
102        velocity field set SEAICEuseFREEDRIFT to .TRUE. in data.seaice.
103    o move "useOASIS" from PARAMS.h to EEPARAMS.h and read it from "eedata"
104    
105    checkpoint62l (2010/10/04)
106    o pkg/seaice:
107      -fix the USE_ORIGINAL_SBI version (for variable freezing point);
108      -use "celsius2K" in SEAICE_VARIABLE_FREEZING_POINT and in new
109       version (with USE_ORIGINAL_SBI undef);
110    o add hooks for new packages OASIS, the package will follow
111      (hooks exclude the seaice pkg for now)
112    o pkg/autodiff: ALLOW_AUTODIFF_WHTAPEIO
113      Option to use mdsio_read_whalos.F/mdsio_write_whalos.F,
114      rather than mdsio_readvector.F/mdsio_writevector.F, within
115      adread_adwrite.F to read/write adjoint-related tapes to disk.
116    o eesupp and pkg/mdsio: ALLOW_WHIO
117      Option to write/read 2D fields to files including tiles halos.
118      The main purpose is for adjoint related "tape I/O".
119      The secondary purpose is for debugging phases.
120      A word on those rather specifically intended purposes.
121      The code was meant to be stripped down to the bare minimum.
122      Typically, there is a unique and basic mapping to the global
123      buffer (tile by tile, then proc by proc, then level by level), and
124      byteswaps and meta-files are omitted. Indeed the (*_WH*) code is not
125      meant to generate user friendly outputs of various shapes and forms.
126      Rather it is intended, for the knowledgeabale developers,
127      to be largely independent of the many mdsio options, to allow
128      outputs of tiles+halos in debugging phases and in adjoint runs,
129      and to allow some flexibility in tuning (adoint) I/O depending on each
130      file system behavior. With regard to tuning, most of the potential knobs
131      (to leave files open, to allocate the I/O indep. of usesinglecpuio,
132      to specify directories, to specify the typical size of I/O calls, etc.)
133      are not yet included, and will be added if they prove useful.
134    o pkg/seaice:
135       fix uninitialised arrays in USE_ORIGINAL_SBI version of seaice_solve4temp.F
136    o pkg/seaice:
137       new S/R seaice_solve4temp.F: this is the merged version of Ian Fenty
138       code (seaice_budget_ice_if.F) which enable to recover old standard
139       code (seaice_budget_ice.F) when option USE_ORIGINAL_SBI is defined.
140       temporary: add "#define USE_ORIGINAL_SBI" at the top of seaice_solve4temp.F.
141    o model/src (rStar):
142       use simple average (not area weighted) of rStarFac at U,V point when using
143       vectorInvariant and selectKEscheme=1,3 (consistent with mom. vertical advection)
144       update output of exp. fizhi-cs-32x32x40 and fizhi-cs-aqualev20.
145    
146    checkpoint62k (2010/09/17)
147    o adjoint
148       * changed verification/global_ocean.cs32x15 ; adjoint now tests
149         Adams-Bashforth 3 + NLFS + RSTAR
150       * some more modifs required for ECCO-GODAE v4
151    o pkg/seaice
152       add SEAICE_AGE_VOL option to use volume instead of area for
153       integrating SEAICE_AGE
154    o model/src:
155       first check-in of sigma & hybrid-sigma coordinate code ; also includes
156       non-linear free-surface parts ; but pressure-gradient not yet coded.
157    o pkg/ptracers:
158       option for ptracer to convert Salt Relax into additional EmP.
159        tested in exp. global_ocean.90x40x15.dwnslp (update output)
160    
161    checkpoint62j (2010/08/25)
162    o pkg/ggl90:
163       other minor changes: remove 3-D temp array gTKE (+ add argument myIter)
164    o pkg/dic: car_flux.F, phos_flux.F
165      - reduce number of exp-calls in order to improve performance
166      - clean up a little (replace fields by their macros, etc., fix
167        typo in CPP-flag)
168    o pkg/ggl90: minor changes
169      - GGL90diffKrS was removed --> always use GGL90diffKr
170      - GGL90viscAr was removed --> replaced with GGL90viscArU, GGL90viscArV
171      - hack of mxlMaxFlag=2 --> ensure mixing between first and second level
172        (commented out for now)
173      - change in max/min operations to ensure that smoothing is ok
174      - smoothing of GGL90viscAr was moved to ggl90_calc.F (as done for GGL90diffKr)
175      - always use diffKrNrT as background profile (i.e. never use diffKr field)
176    o pkg/cfc: bug fix, AREA(i,j,1,bi,bj) -> AREA(i,j,bi,bj)
177  o pkg/dic:  o pkg/dic:
178      - model/src/impldiff.F: if TARGET_NEC_SX, extend loop ranges for better
179        vectorization
180    - small bug fix in insol for the unlikely case of a cartesian grid    - small bug fix in insol for the unlikely case of a cartesian grid
181    - towards vectorizing the dic package: bio_export.F, car_flux.F, phos_flux.F    - towards vectorizing the dic package: bio_export.F, car_flux.F, phos_flux.F
182      + now vectorize, which also removes adjoint recomputations;      + now vectorize, which also removes adjoint recomputations;
183      + replace integer maksp1 by _RL maskp1      + replace integer maksp1 by _RL maskp1
184        + replace a**x by exp(x*log(a)) in phos_flux.F for speed
185    
186  checkpoint62i (2010/07/21)  checkpoint62i (2010/07/21)
187  o pkg/flt:  o pkg/flt:
Legend:
Removed from v.1.1287  
changed lines
  Added in v.1.1321
  ViewVC Help
Powered by ViewVC 1.1.22