| 3 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
| 4 |
============================== |
============================== |
| 5 |
|
|
| 6 |
|
ecco_c44_e22 |
| 7 |
|
o Finally, the merges with the Scripps cost function terms: |
| 8 |
|
Various additions, modifications, quality checks. |
| 9 |
|
o Plus: Included switch ALLOW_AUTODIFF_MONITOR to handle |
| 10 |
|
monitoring of adjoint state. |
| 11 |
|
o Plus: removed comm_stats, ini_communication_patterns from diff. list |
| 12 |
|
(no more need to "keep" MPI headers for adjoint generation) |
| 13 |
|
o Plus: exf changes: |
| 14 |
|
o added run-off routine |
| 15 |
|
o additional storing needed for ustress,vstess |
| 16 |
|
because of state-dependent quality check |
| 17 |
|
o added missing initialisations |
| 18 |
|
o Plus: genmake with MAKEDEPEND variable was still missing in ecco-branch |
| 19 |
|
NB: - verification/natl_box_adjoint/ 240 steps run |
| 20 |
|
* looks ok under Linux (twain.lcs.mit.edu) |
| 21 |
|
* looks ok on MIT SUN's (fjord.mit.edu) |
| 22 |
|
* NaN in reverse mode on Scripps SUN's at iter=214 |
| 23 |
|
* NaN in reverse mode on IBM SP3 (BlueHorizon) at iter=204 |
| 24 |
|
- 1 year forward int. global 2x2 deg. produces similar cost functions |
| 25 |
|
* all cost terms of new code are slightly smaller (0.5 to 3%). |
| 26 |
|
|
| 27 |
|
ecco_c44_e21 |
| 28 |
|
o reversed order of ctrl/cost initialisation |
| 29 |
|
o gmredi_slope_psi.F: Inconsistent types in MIN/MAX expressions |
| 30 |
|
o modified to enable ctrl field swapping |
| 31 |
|
o bug fix for global sum of ncvarlength in ctrl_init |
| 32 |
|
o no more call to ctrl_unpack from ctrl_init |
| 33 |
|
o no more write of ctrl vector in ctrl_pack |
| 34 |
|
however, init. of control vector from ctrl_unpack |
| 35 |
|
o various bug fixes & consistent nobcs handling in |
| 36 |
|
ctrl_pack/unpack routines (by G. Gebbie) |
| 37 |
|
o optim package: Added obcs control part for lsopt I/O |
| 38 |
|
(by G. Gebbie) |
| 39 |
|
o missing call to obcs_diags |
| 40 |
|
o moved calls in packages_init_fixed to packages_readparms |
| 41 |
|
because some quantities are needed earlier. |
| 42 |
|
|
| 43 |
|
ecco_c44_e20 |
| 44 |
|
o before major merges from the Scripps setup |
| 45 |
|
here are two small updates: |
| 46 |
|
- field swapping for surface flux controls enabled |
| 47 |
|
(remains t.b.d. for atmos. state plus obcs) |
| 48 |
|
- bug fix in active_file.F for xz-/yz- routines |
| 49 |
|
- added missing xz-/yz- routines to active_file_ad.F |
| 50 |
|
|
| 51 |
|
ecco_c44_e19 |
| 52 |
|
o merged necessary patches: |
| 53 |
|
- AD-related changes for GMRedi |
| 54 |
|
- fixes i KPP (delZ -> drF) |
| 55 |
|
- hook to OBCS songe layer code in external_forcing |
| 56 |
|
- errorMessageUnit non-zero in eeboot.F |
| 57 |
|
- included GMREDI, EXACT_CONSERV to natl_box_adjoint verif. |
| 58 |
|
|
| 59 |
|
ecco_c44_e18 |
| 60 |
|
o sponge layer code for obcs was missing |
| 61 |
|
- added hook in external_forcing.F |
| 62 |
|
- added storing for TAF |
| 63 |
|
o new routine cost_init_barfiles to initialise |
| 64 |
|
adjoint fields adtbar, adsbar, adpsbar |
| 65 |
|
(needed in conjunction with corrected ladinit for active I/O) |
| 66 |
|
o cleaned up and corrected order in packages_readparms, ecco_readparms |
| 67 |
|
o corrected and simplified weights for obcs controls |
| 68 |
|
- weight fields simplified |
| 69 |
|
- bug fix in ctrl_pack |
| 70 |
|
- no more write of control vector in ctrl_pack |
| 71 |
|
o i/j index bug fixes in cost_ssh_mean |
| 72 |
|
|
| 73 |
|
ecco_c44_e17 |
| 74 |
|
o bug fixes in cost package |
| 75 |
|
* wrong identifier for active I/O of monthly mean psbar |
| 76 |
|
(cost_averagesfields, cost_ssh) xx_psbar_dummy |
| 77 |
|
* cleaned all identifier for active I/O of monthly fields |
| 78 |
|
* wrong ladinit in various cost routines |
| 79 |
|
o merge with interval (chkpt44c_post - chkpt44d_post) |
| 80 |
|
* added missing EXCLUDE_MONITOR flags |
| 81 |
|
* changed "e" to "_d" in gmredi_slope_limit, gmredi_slope_psi |
| 82 |
|
(incompatible typ in MIN/MAX expressions caused problems |
| 83 |
|
on IBM SP3) |
| 84 |
|
* in genmake added variable MAKEDEPEND |
| 85 |
|
plus resetting for case SunOS |
| 86 |
|
added timer_stats.c routine for IBM SP3 |
| 87 |
|
* removed variables in dynamics |
| 88 |
|
|
| 89 |
|
ecco_c44_e16 |
| 90 |
|
o new ECCO checkpoint, the first since ecco_c43_e15 |
| 91 |
|
which is branched from MITgcm. |
| 92 |
|
All changes were merged in between (see ecco-branch-mod? tags). |
| 93 |
|
o last changes: |
| 94 |
|
* assuming path for TAF and TAMC are known (e.g. in $path) |
| 95 |
|
* genmake: |
| 96 |
|
- added tools/xmakedepend to circumvene shortcoming of |
| 97 |
|
SUN's makedepend (size too small) |
| 98 |
|
- invoking ../tools/xmakedepend instead of makedepend |
| 99 |
|
- changed compiler flags for case SunOS |
| 100 |
|
|
| 101 |
|
ecco-branch-mod5 |
| 102 |
|
o mod's to reproduce open boundary run |
| 103 |
|
o enable field swapping for OB's and climatology |
| 104 |
|
o update store directives for OB control |
| 105 |
|
o added option ALLOW_OBCS_PRESCRIBE |
| 106 |
|
o call of getobcs now out of obcs_calc (param. list modified) |
| 107 |
|
|
| 108 |
|
ecco-branch-mod4 |
| 109 |
|
o this is the second part of changes necessary to reconcile |
| 110 |
|
the natl_box_adjoint verification (with and without grdchk). |
| 111 |
|
>>>!!! CHANGES BETWEEN THE TAGS ecco-branch-mod3 !!!<<< |
| 112 |
|
>>>!!! AND ecco-branch-mod4 !!!<<< |
| 113 |
|
>>>!!! SHOULD BE MERGED INTO THE MAIN TRUNK !!!<<< |
| 114 |
|
- genmake: removed $S64 overwrite for case SunOS |
| 115 |
|
- pkg/exf: update and corrections for field swapping and obcs |
| 116 |
|
- pkg/ecco: parameter lists for the_model_main, the_main_loop |
| 117 |
|
harmonized between ECCO and MITgcm |
| 118 |
|
- pkg/autodiff: added flow directives for obcs, mdsio_gl_slice |
| 119 |
|
updated checkpointing_lev... lists for obcs |
| 120 |
|
- model/src: minor changes in forward_step, plot_field |
| 121 |
|
- pkg/mdsio: added mdsio_gl_slice |
| 122 |
|
|
| 123 |
|
ecco-branch-mod3 |
| 124 |
|
o this is the first part of changes necessary to reconcile |
| 125 |
|
natl_box_adjoint verification (with and without grdchk). |
| 126 |
|
This set of changes is relevant for ECCO only. |
| 127 |
|
The next set/tag of changes (ecco-branch-mod4) |
| 128 |
|
should be merged into the main trunk. |
| 129 |
|
- update some options |
| 130 |
|
- parameter lists of ecco_the_model_main and ecco_the_main_loop |
| 131 |
|
differed from those of the MITgcm counterparts. |
| 132 |
|
|
| 133 |
|
ecco-branch-mod2 |
| 134 |
|
o update ecco-branch-mod1 to checkpoint44 |
| 135 |
|
o exclude from updates: |
| 136 |
|
- verification/ |
| 137 |
|
- adjoint/ |
| 138 |
|
- pkg/cost |
| 139 |
|
- pkg/ctrl |
| 140 |
|
- pkg/ecco |
| 141 |
|
- pkg/obcs |
| 142 |
|
- model/src/the_model_main.F |
| 143 |
|
- model/src/the_main_loop.F |
| 144 |
|
|
| 145 |
|
checkpoint44 |
| 146 |
|
o apply Shapiro filter on T&S before calc_PhiHyd (stagger time stepping) |
| 147 |
|
o fix bug in mom_vi_del2uv |
| 148 |
|
o added some stuff for AD |
| 149 |
|
|
| 150 |
|
ecco-branch-mod1 |
| 151 |
|
o based on c43 |
| 152 |
|
o replaced packages cost, ctrl, ecco, obcs |
| 153 |
|
o removed MITgcm verif.'s, added ECCO verif.'s |
| 154 |
|
o added lsopt, optim for line search optimization. |
| 155 |
|
|
| 156 |
checkpoint43 |
checkpoint43 |
| 157 |
o added two new advection tests |
o added two new advection tests |
| 158 |
o added global 4x4 degree ocean test |
o added global 4x4 degree ocean test |
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch