[MITgcm-devel] TAF TLM initialisation problem

[Patrick Heimbach]

Hi Ralf,

here's how you can compare the two codes:
Check out both versions, i.e.
(1st works, 2nd doesn't)
cvs co -r checkpoint52c_post -d c52ok MITgcm
cvs co -r checkpoint52d_post -d c52notlm MITgcm

# Then, for both versions, do following
# (assume you are in base directory c52* )
#
cd verification/carbon/build/
../../../tools/genmake2 -mods ../code_ad
make depend
make ftlall
#
#Then, run the code by:
#
ln -s ../input_ad/* .
./mitgcmuv_ftl >&! output.txt
#
# Then grep for 4 gradient checks
grep grad-res output.txt

You'll see that while in one case (c52ok) all
for look good, in the other case (c52motlm)
only the first is ok.

In both cases the relevant code is ftl_taf_output.f
You'll see that the c52ok case has the correct
initialisation of g_etan (as part of S/R g_ini_psurf)
whereas in the c52notlm case it is missing.

You can also diff both ftl_ codes,
they don't look very different
(apart from the missing S/R g_ini_psurf)

Thanks
-Patrick



Quoting Patrick Heimbach <heimbach at MIT.EDU>:

> Hi Ralf,
> 
> I've noticed that for our newer versions of
> the code, TAF seems to falsely ignore some
> array initialisations.
> The initialisations of the original variables
> is seen by TAF, but e.g. for
> initialisation of etaN in S/R ini_psurf
> TAF is not generating a S/R g_ini_psurf
> which should have initialised g_etan.
> 
> I have two setups to compare on sea,
> let me know  whether you want to view
> them directly on sea or whether I should
> tar them up for you.
> 
> Thanks
> -p.
> 
> _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
> Patrick Heimbach     Massachusetts Institute of Technology
> FON: +1/617/253-5259                    EAPS, Room 54-1518
> FAX: +1/617/253-4464               77 Massachusetts Avenue
> mailto:heimbach at mit.edu                 Cambridge MA 02139
> http://www.mit.edu/~heimbach/                          USA
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-devel
> 


_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
Patrick Heimbach     Massachusetts Institute of Technology
FON: +1/617/253-5259                    EAPS, Room 54-1518
FAX: +1/617/253-4464               77 Massachusetts Avenue
mailto:heimbach at mit.edu                 Cambridge MA 02139
http://www.mit.edu/~heimbach/                          USA