Changes between Version 4 and Version 5 of 2020WP/TOP-05_Ethe_Agrif

Timestamp:

2020-11-27T16:44:33+01:00 (4 years ago)

Author:

cetlod

Comment:

--

2020WP/TOP-05_Ethe_Agrif

@@ -45 +45 @@
 == Preview 
 
-{{{#!box width=50em info
-[[Include(wiki:Developers/DevProcess#preview_)]]
-}}}
+The methodology is just to correct appropriate routines to make NEMO-TOP compiles and runs successfully when using ARIF
 
-''...''
+All the modfied routines are listed here : https://forge.ipsl.jussieu.fr/nemo/changeset?sfp_email=&sfph_mail=&reponame=&new=13373%40NEMO%2Fbranches%2F2020%2Fdev_r13333_TOP-05_Ethe_Agrif%2Fsrc&old=13372%40NEMO%2Fbranches%2F2020%2Fdev_r13333_TOP-05_Ethe_Agrif%2Fsrc
+
+Most of them have minor changes. 
+
+The real bottleneck came from to the fact that AGRIF doesn't like structure as an argument of a subroutine and so the compilation fails unless we add, as it is done now, a directive $AGRIF_DO_NOT_TREAT in trcdta.F90. But doing that implies we can't use this routine to initialize the tracers using on-the-fly interpolation.
+
+
+As suggested Tomas Lovato an alternative is to pass as an argument only the jl index to select data from the structure within trcdta. It works but not very optimal since with the "CALL fld_read( kt, 1, sf_trcdta )" we always read the whole structure ( all the initialized tracers ) even if we need only to get the tracer with jl index. Also change in trcdmp.F90
+
+The dev branch has been successfully phased with trunk at revision r13787
 
 == Tests
 
-{{{#!box width=50em info
-[[Include(wiki:Developers/DevProcess#tests)]]
-}}}
+The run done is described in ticket https://forge.ipsl.jussieu.fr/nemo/ticket/2508
+
 
 ''...''