LIMA Refactoring #61
LIMA Refactoring #61
Conversation
| REAL, DIMENSION(MERGE(D%NIJT,0,OMFCONV), & | ||
| MERGE(D%NKT,0,OMFCONV)), INTENT(IN) :: PMFCONV ! | ||
| REAL, DIMENSION(D%NIJT, D%NKT), INTENT(IN) :: PPABST ! Absolute Pressure at t | ||
| REAL, DIMENSION(D%NIJT, D%NKT), INTENT(IN) :: PPABSTT ! Absolute Pressure at t+dt |
There was a problem hiding this comment.
This is why I said "code still uses 3 time steps whereas the third time step has been removed from other parametrisations". I haven't checked all the code to see if there are others, perhaps it's the last variable at t+dt.
| END WHERE | ||
| WHERE(ZT(:)<=263.15) | ||
| ZLBDS(:) = MAX(MIN(XLBDAS_MAX, 10**(6.226-0.0106*ZT(:))),XLBDAS_MIN) | ||
| ZLBDS(:) = MAX(MIN(LIMAC%XLBDAS_MAX, 10**(6.226-0.0106*ZT(:))),LIMAC%XLBDAS_MIN) | ||
| END WHERE | ||
| END WHERE |
There was a problem hiding this comment.
Nested where are dangerous and do not necessarily do what is expected on some compilers.
| @@ -175,7 +180,7 @@ SUBROUTINE GETDATA_LIMA_ADJUST (NPROMA, NGPBLKS, NFLEVG, KRR, KSV, & | |||
| READ (IFILE) KLON, KDUM, KLEV | |||
| READ (IFILE) KRR, KSV | |||
| READ (IFILE) CCONDENS, CLAMBDA3 | |||
| READ (IFILE) OSUBG_COND, OSIGMAS, PTSTEP, PSIGQSAT | |||
| READ (IFILE) OSUBG_COND, OSIGMAS, PTSTEP | |||
There was a problem hiding this comment.
To avoid having to generate the files again, we should keep the same reading, even if it means doing nothing with this variable. Alternatively, it can be used to fill the ZSIGQSAT_B array.
There was a problem hiding this comment.
There seem to be some forgotten prints.
The refversion variable cannot be hard-coded, it would break other usages.
There was a problem hiding this comment.
Why those modifications?
Solves several items from issue #57
bit-repro until commit f1f0bc7 (compared to commit dc0be8a). lima_adjust results changed when switching PSIGQSAT to an array.
commit f1f0bc7 tested in Meso-NH -> OK (Quentin)
Remaining items to be addressed :