source: trunk/LMDZ.COMMON/libf/dyn3dpar/grad_p.F

Last change on this file was 1019, checked in by emillour, 11 years ago

Common dynamics; keep up with updates (seq and ) in LMDZ5 (up tio rev 1845):

  • General stuff:
  • makelmdz_fcm: add options -j # (compile using # threads) and -full, and to keep up

with Earth model, possibility to compile with various versions of orchidee

  • bld.cfg: adaptations to enable compiling using multiple threads
  • build_gcm: adaptations to enable compiling using multiple threads
  • makelmdz: keep up with Earth model: possibility to compile with various versions of orchidee + cosmetic changes + library directory name change
  • bibio:
  • wxios.F90 : Added for possible future use of XIOS library
  • filtrez:
  • mkl_dft_type.f90 & mkl_dfti.f90 : MKL (for MKL FFT) interface definitions
  • filtreg_mod : limit use of FFT to parallel mode
  • mod_filtre_fft.F90 & mod_filtre_fft_lov.F90 : swich to use parallel_lmdz
  • dyn3d:
  • abort_gcm.F : add things for xios
  • advtrac.F90 : minor change in CFL outputs
  • ce0l.F90 : indicesol.h is now module indice_sol_mod
  • comvert.h : cosmetic change on comments
  • gcm.F : add xios and use module indice_sol_mod (for INCA)
  • inigeom.F : move two computations outside loop
  • dyn3dpar:
  • parallel.F90 => parallel_lmdz.F90 : and change all the "use parallel" into "use parallel_lmdz" in all files in dyn3dpar
  • comvert.h : cosmetic change on comments
  • gcm.F : add xios and use module indice_sol_mod (for INCA)
  • leapfrog_p.F : add xios + correction for times in Newtonian case
  • ce0l.F90 : indicesol.h is now module indice_sol_mod
  • inigeom.F : move two computations outside loop

EM
File size: 1.3 KB
Line 
1      SUBROUTINE  grad_p(klevel, pg,pgx,pgy )
2c
3c      P. Le Van
4c
5c    ******************************************************************
6c     .. calcul des composantes covariantes en x et y du gradient de g
7c
8c    ******************************************************************
9c             pg        est un   argument  d'entree pour le s-prog
10c       pgx  et  pgy    sont des arguments de sortie pour le s-prog
11c
12      USE parallel_lmdz
13      IMPLICIT NONE
14c
15#include "dimensions.h"
16#include "paramet.h"
17      INTEGER klevel
18      REAL  pg( ip1jmp1,klevel )
19      REAL pgx( ip1jmp1,klevel ) , pgy( ip1jm,klevel )
20      INTEGER  l,ij
21      INTEGER :: ijb,ije,jjb,jje
22c
23c
24c$OMP DO SCHEDULE(STATIC,OMP_CHUNK)
25      DO 6 l = 1,klevel
26c
27      ijb=ij_begin
28      ije=ij_end
29      DO 2  ij = ijb, ije - 1
30        pgx( ij,l ) = pg( ij +1,l ) - pg( ij,l )
31   2  CONTINUE
32c
33c    .... correction pour  pgx(ip1,j,l)  ....
34c    ...    pgx(iip1,j,l)= pgx(1,j,l)  ....
35CDIR$ IVDEP
36      DO 3  ij = ijb+iip1-1, ije, iip1
37        pgx( ij,l ) = pgx( ij -iim,l )
38   3  CONTINUE
39c
40      ijb=ij_begin-iip1
41      ije=ij_end
42      if (pole_nord) ijb=ij_begin
43      if (pole_sud)  ije=ij_end-iip1
44     
45      DO 4 ij = ijb,ije
46        pgy( ij,l ) = pg( ij,l ) - pg( ij +iip1,l )
47   4  CONTINUE
48c
49   6  CONTINUE
50c$OMP END DO NOWAIT
51
52      RETURN
53      END
Note: See TracBrowser for help on using the repository browser.