/[MITgcm]/MITgcm/verification/natl_box_adjoint/README
ViewVC logotype

Diff of /MITgcm/verification/natl_box_adjoint/README

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

revision 1.1.2.12 by dimitri, Sat Feb 22 01:15:52 2003 UTC revision 1.1.2.13 by dimitri, Sun Feb 23 01:42:41 2003 UTC
# Line 47  Comments: Line 47  Comments:
47  - 4 timesteps  - 4 timesteps
48    
49    
   
50  ####################  ####################
51  # Experiment no. 2 #  # Experiment no. 2 #
52  ####################  ####################
# Line 123  grep ph-grd ../verification/natl_box_adj Line 122  grep ph-grd ../verification/natl_box_adj
122    cp ../verification/natl_box_adjoint/input/PH_*.data .    cp ../verification/natl_box_adjoint/input/PH_*.data .
123    ./mitgcmuv >&! output.txt    ./mitgcmuv >&! output.txt
124    
 #I've changed the run setup:  
 #We now perform gradient checks for the first for elements of  
 #the air temperture part of the control vector (grdchkvarindex = 7).  
 #To verify the results of the gradient check, type  
   
125  grep ph-grd output.txt  grep ph-grd output.txt
126    
127  #A reference output is in:  #A reference output is in:
128  grep ph-grd ../verification/natl_box_adjoint/results_bulk/output.exp3  grep ph-grd ../verification/natl_box_adjoint/results_bulk/output.txt
129    
130  Comments:  Comments:
131  - Produces following error message during "make adtaf" compilation:  - Produces following error message during "make adtaf" compilation:
132      INTERNAL ERROR: cannot find var clone to ada2      INTERNAL ERROR: cannot find var clone to ada2
 - The gradient check results should be identical to those of Exp. 2,  
   but aren't.  Compare results_bulk/output.txt to output.exp3  
133    
134    
135  ####################  ####################
136  # Experiment no. 4 #  # Experiment no. 4 #
137  ####################  ####################
138    
139  #Same as exp. 3, but with useseaice = .true.  #Uses same executable as exp. 3, but with useseaice = .true.
140    
 #Compile code in bin/:  
   cd ../bin  
   rm -rf *.o *.f *.p rii_files  
   rm -rf *.F *.h *.c Makefile*  
   cp ../verification/natl_box_adjoint/code_seaice/*.h .  
   cp ../verification/natl_box_adjoint/code_seaice/.genmakerc .  
   cp ../verification/natl_box_adjoint/code_seaice/SEAICE_OPTIONS.h.FLUXES SEAICE_OPTIONS.h  
 #Configure and compile the code:  
   ../tools/genmake -makefile  
   make depend  
 #Generate the adjoint code:  
   cd ../adjoint  
   mv makefile makefile.old  
   cp ../verification/natl_box_adjoint/code_seaice/makefile .  
   make adtaf  
   make adchange  
   rm -f makefile  
   mv makefile.old makefile  
 #Go back and compile the code:  
   cd ../bin  
   make  
141  #To run:  #To run:
142    cd ../exe    cd ../exe
143    cp ../verification/natl_box_adjoint/input_bulk/* .    cp ../verification/natl_box_adjoint/input_bulk/* .
# Line 175  Comments: Line 146  Comments:
146    cp ../verification/natl_box_adjoint/input/PH_*.data .    cp ../verification/natl_box_adjoint/input/PH_*.data .
147    ./mitgcmuv >&! output.txt    ./mitgcmuv >&! output.txt
148    
 #I've changed the run setup:  
 #We now perform gradient checks for the first for elements of  
 #the air temperture part of the control vector (grdchkvarindex = 7).  
 #To verify the results of the gradient check, type  
   
149  grep ph-grd output.txt  grep ph-grd output.txt
150    
151  #A reference output is in:  #A reference output is in:
# Line 189  Comments: Line 155  Comments:
155  - Produces following error message during "make adtaf" compilation:  - Produces following error message during "make adtaf" compilation:
156      INTERNAL ERROR: cannot find var clone to ada2      INTERNAL ERROR: cannot find var clone to ada2
157  - The gradient check results are different from those of exp. 3,  - The gradient check results are different from those of exp. 3,
158    perhaps because of big initial pulse of freshwater?    because of big initial pulse of freshwater, corrected below?
159    
160    
161  ####################  ####################
# Line 198  Comments: Line 164  Comments:
164    
165  #Same as exp. 4, but with HEFF=0 initial condition  #Same as exp. 4, but with HEFF=0 initial condition
166    
 #Compile code in bin/:  
   cd ../bin  
   rm -rf *.o *.f *.p rii_files  
   rm -rf *.F *.h *.c Makefile*  
   cp ../verification/natl_box_adjoint/code_seaice/*.h .  
   cp ../verification/natl_box_adjoint/code_seaice/.genmakerc .  
   cp ../verification/natl_box_adjoint/code_seaice/SEAICE_OPTIONS.h.FLUXES SEAICE_OPTIONS.h  
 #Configure and compile the code:  
   ../tools/genmake -makefile  
   make depend  
 #Generate the adjoint code:  
   cd ../adjoint  
   mv makefile makefile.old  
   cp ../verification/natl_box_adjoint/code_seaice/makefile .  
   make adtaf  
   make adchange  
   rm -f makefile  
   mv makefile.old makefile  
 #Go back and compile the code:  
   cd ../bin  
   make  
167  #To run:  #To run:
168    cd ../exe    cd ../exe
169    cp ../verification/natl_box_adjoint/input_bulk/* .    cp ../verification/natl_box_adjoint/input_bulk/* .
170    cp ../verification/natl_box_adjoint/input_seaice/data.seaice.cgrid data.seaice    cp ../verification/natl_box_adjoint/input_seaice/data.seaice.heff data.seaice
171    cp ../verification/natl_box_adjoint/input_seaice/data.pkg .    cp ../verification/natl_box_adjoint/input_seaice/data.pkg .
172    cp ../verification/natl_box_adjoint/input/PH_*.data .    cp ../verification/natl_box_adjoint/input/PH_*.data .
173    ./mitgcmuv >&! output.txt    ./mitgcmuv >&! output.txt
# Line 230  Comments: Line 175  Comments:
175  grep ph-grd output.txt  grep ph-grd output.txt
176    
177  #A reference output is in:  #A reference output is in:
178  grep ph-grd ../verification/natl_box_adjoint/results_bulk/output.exp5  grep ph-grd ../verification/natl_box_adjoint/results_bulk/output.txt
179    
180  Comments:  Comments:
181  - Produces following error message during "make adtaf" compilation:  - Produces following error message during "make adtaf" compilation:
182      INTERNAL ERROR: cannot find var clone to ada2      INTERNAL ERROR: cannot find var clone to ada2
 - Gradient check results are closer to exp3 than were those of exp4.  
183    
184    
185  ####################  ####################
# Line 268  Comments: Line 212  Comments:
212  #To run:  #To run:
213    cd ../exe    cd ../exe
214    cp ../verification/natl_box_adjoint/input_bulk/* .    cp ../verification/natl_box_adjoint/input_bulk/* .
215    cp ../verification/natl_box_adjoint/input_seaice/data.seaice.cgrid data.seaice    cp ../verification/natl_box_adjoint/input_seaice/data.seaice.heff data.seaice
216    cp ../verification/natl_box_adjoint/input_seaice/data.pkg .    cp ../verification/natl_box_adjoint/input_seaice/data.pkg .
217    cp ../verification/natl_box_adjoint/input/PH_*.data .    cp ../verification/natl_box_adjoint/input/PH_*.data .
218    ./mitgcmuv >&! output.txt    ./mitgcmuv >&! output.txt
# Line 281  grep ph-grd ../verification/natl_box_adj Line 225  grep ph-grd ../verification/natl_box_adj
225  Comments:  Comments:
226  - Produces following error message during "make adtaf" compilation:  - Produces following error message during "make adtaf" compilation:
227      INTERNAL ERROR: cannot find var clone to ada2      INTERNAL ERROR: cannot find var clone to ada2
228  - The gradient check results should be identical than those of Exp. 2,  - The gradient check results should be comparable to that of Exps. 2, 3, and
229    but aren't.  Compare results_bulk/output.txt to output.exp3    5, but it is not.  Also there is a 3-4% discrepancy between ajoint model
230      and finite difference gradients.
231    
232    
233  ####################  ####################
Legend:
Removed from v.1.1.2.12  
changed lines
  Added in v.1.1.2.13
  ViewVC Help
Powered by ViewVC 1.1.22