[3] | 1 | MODULE lib_mpp |
---|
[13] | 2 | !!====================================================================== |
---|
| 3 | !! *** MODULE lib_mpp *** |
---|
[1344] | 4 | !! Ocean numerics: massively parallel processing library |
---|
[13] | 5 | !!===================================================================== |
---|
[1344] | 6 | !! History : OPA ! 1994 (M. Guyon, J. Escobar, M. Imbard) Original code |
---|
| 7 | !! 7.0 ! 1997 (A.M. Treguier) SHMEM additions |
---|
| 8 | !! 8.0 ! 1998 (M. Imbard, J. Escobar, L. Colombet ) SHMEM and MPI |
---|
| 9 | !! ! 1998 (J.M. Molines) Open boundary conditions |
---|
| 10 | !! NEMO 1.0 ! 2003 (J.-M. Molines, G. Madec) F90, free form |
---|
| 11 | !! ! 2003 (J.M. Molines) add mpp_ini_north(_3d,_2d) |
---|
| 12 | !! - ! 2004 (R. Bourdalle Badie) isend option in mpi |
---|
| 13 | !! ! 2004 (J.M. Molines) minloc, maxloc |
---|
| 14 | !! - ! 2005 (G. Madec, S. Masson) npolj=5,6 F-point & ice cases |
---|
| 15 | !! - ! 2005 (R. Redler) Replacement of MPI_COMM_WORLD except for MPI_Abort |
---|
| 16 | !! - ! 2005 (R. Benshila, G. Madec) add extra halo case |
---|
| 17 | !! - ! 2008 (R. Benshila) add mpp_ini_ice |
---|
| 18 | !! 3.2 ! 2009 (R. Benshila) SHMEM suppression, north fold in lbc_nfd |
---|
[1345] | 19 | !! 3.2 ! 2009 (O. Marti) add mpp_ini_znl |
---|
[2715] | 20 | !! 4.0 ! 2011 (G. Madec) move ctl_ routines from in_out_manager |
---|
[13] | 21 | !!---------------------------------------------------------------------- |
---|
[2715] | 22 | |
---|
| 23 | !!---------------------------------------------------------------------- |
---|
| 24 | !! ctl_stop : update momentum and tracer Kz from a tke scheme |
---|
| 25 | !! ctl_warn : initialization, namelist read, and parameters control |
---|
| 26 | !! ctl_opn : Open file and check if required file is available. |
---|
| 27 | !! get_unit : give the index of an unused logical unit |
---|
| 28 | !!---------------------------------------------------------------------- |
---|
[1344] | 29 | #if defined key_mpp_mpi |
---|
[13] | 30 | !!---------------------------------------------------------------------- |
---|
[1344] | 31 | !! 'key_mpp_mpi' MPI massively parallel processing library |
---|
| 32 | !!---------------------------------------------------------------------- |
---|
[2715] | 33 | !! lib_mpp_alloc : allocate mpp arrays |
---|
| 34 | !! mynode : indentify the processor unit |
---|
| 35 | !! mpp_lnk : interface (defined in lbclnk) for message passing of 2d or 3d arrays (mpp_lnk_2d, mpp_lnk_3d) |
---|
[473] | 36 | !! mpp_lnk_3d_gather : Message passing manadgement for two 3D arrays |
---|
[2715] | 37 | !! mpp_lnk_e : interface (defined in lbclnk) for message passing of 2d array with extra halo (mpp_lnk_2d_e) |
---|
| 38 | !! mpprecv : |
---|
| 39 | !! mppsend : SUBROUTINE mpp_ini_znl |
---|
| 40 | !! mppscatter : |
---|
| 41 | !! mppgather : |
---|
| 42 | !! mpp_min : generic interface for mppmin_int , mppmin_a_int , mppmin_real, mppmin_a_real |
---|
| 43 | !! mpp_max : generic interface for mppmax_int , mppmax_a_int , mppmax_real, mppmax_a_real |
---|
| 44 | !! mpp_sum : generic interface for mppsum_int , mppsum_a_int , mppsum_real, mppsum_a_real |
---|
| 45 | !! mpp_minloc : |
---|
| 46 | !! mpp_maxloc : |
---|
| 47 | !! mppsync : |
---|
| 48 | !! mppstop : |
---|
| 49 | !! mppobc : variant of mpp_lnk for open boundary condition |
---|
[1344] | 50 | !! mpp_ini_north : initialisation of north fold |
---|
| 51 | !! mpp_lbc_north : north fold processors gathering |
---|
| 52 | !! mpp_lbc_north_e : variant of mpp_lbc_north for extra outer halo |
---|
[13] | 53 | !!---------------------------------------------------------------------- |
---|
[2715] | 54 | USE dom_oce ! ocean space and time domain |
---|
| 55 | USE lbcnfd ! north fold treatment |
---|
| 56 | USE in_out_manager ! I/O manager |
---|
[3] | 57 | |
---|
[13] | 58 | IMPLICIT NONE |
---|
[415] | 59 | PRIVATE |
---|
[1344] | 60 | |
---|
[2715] | 61 | PUBLIC ctl_stop, ctl_warn, get_unit, ctl_opn |
---|
[1344] | 62 | PUBLIC mynode, mppstop, mppsync, mpp_comm_free |
---|
| 63 | PUBLIC mpp_ini_north, mpp_lbc_north, mpp_lbc_north_e |
---|
| 64 | PUBLIC mpp_min, mpp_max, mpp_sum, mpp_minloc, mpp_maxloc |
---|
| 65 | PUBLIC mpp_lnk_3d, mpp_lnk_3d_gather, mpp_lnk_2d, mpp_lnk_2d_e |
---|
[3294] | 66 | PUBLIC mppscatter, mppgather |
---|
[1528] | 67 | PUBLIC mppobc, mpp_ini_ice, mpp_ini_znl |
---|
[2715] | 68 | PUBLIC mppsize |
---|
[3609] | 69 | PUBLIC mppsend, mpprecv ! needed by ICB routines |
---|
[2715] | 70 | PUBLIC lib_mpp_alloc ! Called in nemogcm.F90 |
---|
[415] | 71 | |
---|
[13] | 72 | !! * Interfaces |
---|
| 73 | !! define generic interface for these routine as they are called sometimes |
---|
[1344] | 74 | !! with scalar arguments instead of array arguments, which causes problems |
---|
| 75 | !! for the compilation on AIX system as well as NEC and SGI. Ok on COMPACQ |
---|
[13] | 76 | INTERFACE mpp_min |
---|
| 77 | MODULE PROCEDURE mppmin_a_int, mppmin_int, mppmin_a_real, mppmin_real |
---|
| 78 | END INTERFACE |
---|
| 79 | INTERFACE mpp_max |
---|
[681] | 80 | MODULE PROCEDURE mppmax_a_int, mppmax_int, mppmax_a_real, mppmax_real |
---|
[13] | 81 | END INTERFACE |
---|
| 82 | INTERFACE mpp_sum |
---|
[1976] | 83 | MODULE PROCEDURE mppsum_a_int, mppsum_int, mppsum_a_real, mppsum_real, & |
---|
| 84 | mppsum_realdd, mppsum_a_realdd |
---|
[13] | 85 | END INTERFACE |
---|
| 86 | INTERFACE mpp_lbc_north |
---|
| 87 | MODULE PROCEDURE mpp_lbc_north_3d, mpp_lbc_north_2d |
---|
| 88 | END INTERFACE |
---|
[1344] | 89 | INTERFACE mpp_minloc |
---|
| 90 | MODULE PROCEDURE mpp_minloc2d ,mpp_minloc3d |
---|
| 91 | END INTERFACE |
---|
| 92 | INTERFACE mpp_maxloc |
---|
| 93 | MODULE PROCEDURE mpp_maxloc2d ,mpp_maxloc3d |
---|
| 94 | END INTERFACE |
---|
[1976] | 95 | |
---|
[51] | 96 | !! ========================= !! |
---|
| 97 | !! MPI variable definition !! |
---|
| 98 | !! ========================= !! |
---|
[1629] | 99 | !$AGRIF_DO_NOT_TREAT |
---|
[2004] | 100 | INCLUDE 'mpif.h' |
---|
[1629] | 101 | !$AGRIF_END_DO_NOT_TREAT |
---|
[1344] | 102 | |
---|
| 103 | LOGICAL, PUBLIC, PARAMETER :: lk_mpp = .TRUE. !: mpp flag |
---|
[3] | 104 | |
---|
[1344] | 105 | INTEGER, PARAMETER :: nprocmax = 2**10 ! maximun dimension (required to be a power of 2) |
---|
| 106 | |
---|
| 107 | INTEGER :: mppsize ! number of process |
---|
| 108 | INTEGER :: mpprank ! process number [ 0 - size-1 ] |
---|
[2363] | 109 | !$AGRIF_DO_NOT_TREAT |
---|
[2249] | 110 | INTEGER, PUBLIC :: mpi_comm_opa ! opa local communicator |
---|
[2363] | 111 | !$AGRIF_END_DO_NOT_TREAT |
---|
[3] | 112 | |
---|
[2480] | 113 | INTEGER :: MPI_SUMDD |
---|
[1976] | 114 | |
---|
[869] | 115 | ! variables used in case of sea-ice |
---|
[3625] | 116 | INTEGER, PUBLIC :: ncomm_ice !: communicator made by the processors with sea-ice (public so that it can be freed in limthd) |
---|
| 117 | INTEGER :: ngrp_iworld ! group ID for the world processors (for rheology) |
---|
[1345] | 118 | INTEGER :: ngrp_ice ! group ID for the ice processors (for rheology) |
---|
| 119 | INTEGER :: ndim_rank_ice ! number of 'ice' processors |
---|
| 120 | INTEGER :: n_ice_root ! number (in the comm_ice) of proc 0 in the ice comm |
---|
[2715] | 121 | INTEGER, DIMENSION(:), ALLOCATABLE, SAVE :: nrank_ice ! dimension ndim_rank_ice |
---|
[1345] | 122 | |
---|
| 123 | ! variables used for zonal integration |
---|
| 124 | INTEGER, PUBLIC :: ncomm_znl !: communicator made by the processors on the same zonal average |
---|
| 125 | LOGICAL, PUBLIC :: l_znl_root ! True on the 'left'most processor on the same row |
---|
| 126 | INTEGER :: ngrp_znl ! group ID for the znl processors |
---|
| 127 | INTEGER :: ndim_rank_znl ! number of processors on the same zonal average |
---|
[2715] | 128 | INTEGER, DIMENSION(:), ALLOCATABLE, SAVE :: nrank_znl ! dimension ndim_rank_znl, number of the procs into the same znl domain |
---|
[1344] | 129 | |
---|
| 130 | ! North fold condition in mpp_mpi with jpni > 1 |
---|
| 131 | INTEGER :: ngrp_world ! group ID for the world processors |
---|
[1345] | 132 | INTEGER :: ngrp_opa ! group ID for the opa processors |
---|
[1344] | 133 | INTEGER :: ngrp_north ! group ID for the northern processors (to be fold) |
---|
| 134 | INTEGER :: ncomm_north ! communicator made by the processors belonging to ngrp_north |
---|
| 135 | INTEGER :: ndim_rank_north ! number of 'sea' processor in the northern line (can be /= jpni !) |
---|
| 136 | INTEGER :: njmppmax ! value of njmpp for the processors of the northern line |
---|
| 137 | INTEGER :: north_root ! number (in the comm_opa) of proc 0 in the northern comm |
---|
[2715] | 138 | INTEGER, DIMENSION(:), ALLOCATABLE, SAVE :: nrank_north ! dimension ndim_rank_north |
---|
[3] | 139 | |
---|
[1344] | 140 | ! Type of send : standard, buffered, immediate |
---|
[3609] | 141 | CHARACTER(len=1) :: cn_mpi_send = 'S' ! type od mpi send/recieve (S=standard, B=bsend, I=isend) |
---|
| 142 | LOGICAL , PUBLIC :: l_isend = .FALSE. ! isend use indicator (T if cn_mpi_send='I') |
---|
| 143 | INTEGER :: nn_buffer = 0 ! size of the buffer in case of mpi_bsend |
---|
[1344] | 144 | |
---|
[2715] | 145 | REAL(wp), DIMENSION(:), ALLOCATABLE, SAVE :: tampon ! buffer in case of bsend |
---|
[3] | 146 | |
---|
[1344] | 147 | ! message passing arrays |
---|
[2715] | 148 | REAL(wp), DIMENSION(:,:,:,:,:), ALLOCATABLE, SAVE :: t4ns, t4sn ! 2 x 3d for north-south & south-north |
---|
| 149 | REAL(wp), DIMENSION(:,:,:,:,:), ALLOCATABLE, SAVE :: t4ew, t4we ! 2 x 3d for east-west & west-east |
---|
| 150 | REAL(wp), DIMENSION(:,:,:,:,:), ALLOCATABLE, SAVE :: t4p1, t4p2 ! 2 x 3d for north fold |
---|
| 151 | REAL(wp), DIMENSION(:,:,:,:) , ALLOCATABLE, SAVE :: t3ns, t3sn ! 3d for north-south & south-north |
---|
| 152 | REAL(wp), DIMENSION(:,:,:,:) , ALLOCATABLE, SAVE :: t3ew, t3we ! 3d for east-west & west-east |
---|
| 153 | REAL(wp), DIMENSION(:,:,:,:) , ALLOCATABLE, SAVE :: t3p1, t3p2 ! 3d for north fold |
---|
| 154 | REAL(wp), DIMENSION(:,:,:) , ALLOCATABLE, SAVE :: t2ns, t2sn ! 2d for north-south & south-north |
---|
| 155 | REAL(wp), DIMENSION(:,:,:) , ALLOCATABLE, SAVE :: t2ew, t2we ! 2d for east-west & west-east |
---|
| 156 | REAL(wp), DIMENSION(:,:,:) , ALLOCATABLE, SAVE :: t2p1, t2p2 ! 2d for north fold |
---|
| 157 | |
---|
| 158 | ! Arrays used in mpp_lbc_north_3d() |
---|
| 159 | REAL(wp), DIMENSION(:,:,:) , ALLOCATABLE, SAVE :: ztab, znorthloc |
---|
| 160 | REAL(wp), DIMENSION(:,:,:,:), ALLOCATABLE, SAVE :: znorthgloio |
---|
[3294] | 161 | REAL(wp), DIMENSION(:,:,:) , ALLOCATABLE, SAVE :: zfoldwk ! Workspace for message transfers avoiding mpi_allgather |
---|
[2715] | 162 | |
---|
| 163 | ! Arrays used in mpp_lbc_north_2d() |
---|
| 164 | REAL(wp), DIMENSION(:,:) , ALLOCATABLE, SAVE :: ztab_2d, znorthloc_2d |
---|
| 165 | REAL(wp), DIMENSION(:,:,:), ALLOCATABLE, SAVE :: znorthgloio_2d |
---|
[3294] | 166 | REAL(wp), DIMENSION(:,:) , ALLOCATABLE, SAVE :: zfoldwk_2d ! Workspace for message transfers avoiding mpi_allgather |
---|
[2715] | 167 | |
---|
| 168 | ! Arrays used in mpp_lbc_north_e() |
---|
| 169 | REAL(wp), DIMENSION(:,:) , ALLOCATABLE, SAVE :: ztab_e, znorthloc_e |
---|
| 170 | REAL(wp), DIMENSION(:,:,:), ALLOCATABLE, SAVE :: znorthgloio_e |
---|
| 171 | |
---|
[3294] | 172 | ! North fold arrays used to minimise the use of allgather operations. Set in nemo_northcomms (nemogcm) so need to be public |
---|
| 173 | INTEGER, PUBLIC, PARAMETER :: jpmaxngh = 8 ! Assumed maximum number of active neighbours |
---|
| 174 | INTEGER, PUBLIC, PARAMETER :: jptyps = 5 ! Number of different neighbour lists to be used for northfold exchanges |
---|
| 175 | INTEGER, PUBLIC, DIMENSION (jpmaxngh,jptyps) :: isendto |
---|
| 176 | INTEGER, PUBLIC, DIMENSION (jptyps) :: nsndto |
---|
| 177 | LOGICAL, PUBLIC :: ln_nnogather = .FALSE. ! namelist control of northfold comms |
---|
| 178 | LOGICAL, PUBLIC :: l_north_nogather = .FALSE. ! internal control of northfold comms |
---|
| 179 | INTEGER, PUBLIC :: ityp |
---|
[51] | 180 | !!---------------------------------------------------------------------- |
---|
[2287] | 181 | !! NEMO/OPA 3.3 , NEMO Consortium (2010) |
---|
[888] | 182 | !! $Id$ |
---|
[2715] | 183 | !! Software governed by the CeCILL licence (NEMOGCM/NEMO_CeCILL.txt) |
---|
[1344] | 184 | !!---------------------------------------------------------------------- |
---|
[3] | 185 | CONTAINS |
---|
| 186 | |
---|
[2715] | 187 | INTEGER FUNCTION lib_mpp_alloc( kumout ) |
---|
[51] | 188 | !!---------------------------------------------------------------------- |
---|
[2715] | 189 | !! *** routine lib_mpp_alloc *** |
---|
| 190 | !!---------------------------------------------------------------------- |
---|
| 191 | INTEGER, INTENT(in) :: kumout ! ocean.output logical unit |
---|
| 192 | !!---------------------------------------------------------------------- |
---|
| 193 | ! |
---|
| 194 | ALLOCATE( t4ns(jpi,jprecj,jpk,2,2) , t4sn(jpi,jprecj,jpk,2,2) , & |
---|
| 195 | & t4ew(jpj,jpreci,jpk,2,2) , t4we(jpj,jpreci,jpk,2,2) , & |
---|
| 196 | & t4p1(jpi,jprecj,jpk,2,2) , t4p2(jpi,jprecj,jpk,2,2) , & |
---|
| 197 | & t3ns(jpi,jprecj,jpk,2) , t3sn(jpi,jprecj,jpk,2) , & |
---|
| 198 | & t3ew(jpj,jpreci,jpk,2) , t3we(jpj,jpreci,jpk,2) , & |
---|
| 199 | & t3p1(jpi,jprecj,jpk,2) , t3p2(jpi,jprecj,jpk,2) , & |
---|
| 200 | & t2ns(jpi,jprecj ,2) , t2sn(jpi,jprecj ,2) , & |
---|
| 201 | & t2ew(jpj,jpreci ,2) , t2we(jpj,jpreci ,2) , & |
---|
| 202 | & t2p1(jpi,jprecj ,2) , t2p2(jpi,jprecj ,2) , & |
---|
| 203 | ! |
---|
| 204 | & ztab(jpiglo,4,jpk) , znorthloc(jpi,4,jpk) , znorthgloio(jpi,4,jpk,jpni) , & |
---|
[3294] | 205 | & zfoldwk(jpi,4,jpk) , & |
---|
[2715] | 206 | ! |
---|
| 207 | & ztab_2d(jpiglo,4) , znorthloc_2d(jpi,4) , znorthgloio_2d(jpi,4,jpni) , & |
---|
[3294] | 208 | & zfoldwk_2d(jpi,4) , & |
---|
[2715] | 209 | ! |
---|
| 210 | & ztab_e(jpiglo,4+2*jpr2dj) , znorthloc_e(jpi,4+2*jpr2dj) , znorthgloio_e(jpi,4+2*jpr2dj,jpni) , & |
---|
| 211 | ! |
---|
| 212 | & STAT=lib_mpp_alloc ) |
---|
| 213 | ! |
---|
| 214 | IF( lib_mpp_alloc /= 0 ) THEN |
---|
| 215 | WRITE(kumout,cform_war) |
---|
| 216 | WRITE(kumout,*) 'lib_mpp_alloc : failed to allocate arrays' |
---|
| 217 | ENDIF |
---|
| 218 | ! |
---|
| 219 | END FUNCTION lib_mpp_alloc |
---|
| 220 | |
---|
| 221 | |
---|
| 222 | FUNCTION mynode( ldtxt, kumnam, kstop, localComm ) |
---|
| 223 | !!---------------------------------------------------------------------- |
---|
[51] | 224 | !! *** routine mynode *** |
---|
| 225 | !! |
---|
| 226 | !! ** Purpose : Find processor unit |
---|
| 227 | !!---------------------------------------------------------------------- |
---|
[1579] | 228 | CHARACTER(len=*),DIMENSION(:), INTENT( out) :: ldtxt |
---|
[2715] | 229 | INTEGER , INTENT(in ) :: kumnam ! namelist logical unit |
---|
| 230 | INTEGER , INTENT(inout) :: kstop ! stop indicator |
---|
[1579] | 231 | INTEGER, OPTIONAL , INTENT(in ) :: localComm |
---|
[2715] | 232 | ! |
---|
[2481] | 233 | INTEGER :: mynode, ierr, code, ji, ii |
---|
[532] | 234 | LOGICAL :: mpi_was_called |
---|
[2715] | 235 | ! |
---|
[3294] | 236 | NAMELIST/nammpp/ cn_mpi_send, nn_buffer, jpni, jpnj, jpnij, ln_nnogather |
---|
[51] | 237 | !!---------------------------------------------------------------------- |
---|
[1344] | 238 | ! |
---|
[2481] | 239 | ii = 1 |
---|
| 240 | WRITE(ldtxt(ii),*) ; ii = ii + 1 |
---|
| 241 | WRITE(ldtxt(ii),*) 'mynode : mpi initialisation' ; ii = ii + 1 |
---|
| 242 | WRITE(ldtxt(ii),*) '~~~~~~ ' ; ii = ii + 1 |
---|
[1344] | 243 | ! |
---|
[2715] | 244 | jpni = -1; jpnj = -1; jpnij = -1 |
---|
| 245 | REWIND( kumnam ) ! Namelist namrun : parameters of the run |
---|
| 246 | READ ( kumnam, nammpp ) |
---|
[1344] | 247 | ! ! control print |
---|
[2481] | 248 | WRITE(ldtxt(ii),*) ' Namelist nammpp' ; ii = ii + 1 |
---|
| 249 | WRITE(ldtxt(ii),*) ' mpi send type cn_mpi_send = ', cn_mpi_send ; ii = ii + 1 |
---|
| 250 | WRITE(ldtxt(ii),*) ' size in bytes of exported buffer nn_buffer = ', nn_buffer ; ii = ii + 1 |
---|
[300] | 251 | |
---|
[2731] | 252 | #if defined key_agrif |
---|
| 253 | IF( .NOT. Agrif_Root() ) THEN |
---|
| 254 | jpni = Agrif_Parent(jpni ) |
---|
| 255 | jpnj = Agrif_Parent(jpnj ) |
---|
| 256 | jpnij = Agrif_Parent(jpnij) |
---|
| 257 | ENDIF |
---|
| 258 | #endif |
---|
| 259 | |
---|
[2715] | 260 | IF(jpnij < 1)THEN |
---|
| 261 | ! If jpnij is not specified in namelist then we calculate it - this |
---|
| 262 | ! means there will be no land cutting out. |
---|
| 263 | jpnij = jpni * jpnj |
---|
| 264 | END IF |
---|
| 265 | |
---|
| 266 | IF( (jpni < 1) .OR. (jpnj < 1) )THEN |
---|
| 267 | WRITE(ldtxt(ii),*) ' jpni, jpnj and jpnij will be calculated automatically'; ii = ii + 1 |
---|
| 268 | ELSE |
---|
| 269 | WRITE(ldtxt(ii),*) ' processor grid extent in i jpni = ',jpni; ii = ii + 1 |
---|
| 270 | WRITE(ldtxt(ii),*) ' processor grid extent in j jpnj = ',jpnj; ii = ii + 1 |
---|
| 271 | WRITE(ldtxt(ii),*) ' number of local domains jpnij = ',jpnij; ii = ii +1 |
---|
| 272 | END IF |
---|
| 273 | |
---|
[3294] | 274 | WRITE(ldtxt(ii),*) ' avoid use of mpi_allgather at the north fold ln_nnogather = ', ln_nnogather ; ii = ii + 1 |
---|
| 275 | |
---|
[2480] | 276 | CALL mpi_initialized ( mpi_was_called, code ) |
---|
| 277 | IF( code /= MPI_SUCCESS ) THEN |
---|
[2481] | 278 | DO ji = 1, SIZE(ldtxt) |
---|
| 279 | IF( TRIM(ldtxt(ji)) /= '' ) WRITE(*,*) ldtxt(ji) ! control print of mynode |
---|
| 280 | END DO |
---|
[2480] | 281 | WRITE(*, cform_err) |
---|
| 282 | WRITE(*, *) 'lib_mpp: Error in routine mpi_initialized' |
---|
| 283 | CALL mpi_abort( mpi_comm_world, code, ierr ) |
---|
| 284 | ENDIF |
---|
[415] | 285 | |
---|
[2480] | 286 | IF( mpi_was_called ) THEN |
---|
| 287 | ! |
---|
| 288 | SELECT CASE ( cn_mpi_send ) |
---|
| 289 | CASE ( 'S' ) ! Standard mpi send (blocking) |
---|
[2481] | 290 | WRITE(ldtxt(ii),*) ' Standard blocking mpi send (send)' ; ii = ii + 1 |
---|
[2480] | 291 | CASE ( 'B' ) ! Buffer mpi send (blocking) |
---|
[2481] | 292 | WRITE(ldtxt(ii),*) ' Buffer blocking mpi send (bsend)' ; ii = ii + 1 |
---|
| 293 | IF( Agrif_Root() ) CALL mpi_init_opa( ldtxt, ii, ierr ) |
---|
[2480] | 294 | CASE ( 'I' ) ! Immediate mpi send (non-blocking send) |
---|
[2481] | 295 | WRITE(ldtxt(ii),*) ' Immediate non-blocking send (isend)' ; ii = ii + 1 |
---|
[2480] | 296 | l_isend = .TRUE. |
---|
| 297 | CASE DEFAULT |
---|
[2481] | 298 | WRITE(ldtxt(ii),cform_err) ; ii = ii + 1 |
---|
| 299 | WRITE(ldtxt(ii),*) ' bad value for cn_mpi_send = ', cn_mpi_send ; ii = ii + 1 |
---|
[2715] | 300 | kstop = kstop + 1 |
---|
[2480] | 301 | END SELECT |
---|
| 302 | ELSE IF ( PRESENT(localComm) .and. .not. mpi_was_called ) THEN |
---|
[2481] | 303 | WRITE(ldtxt(ii),*) ' lib_mpp: You cannot provide a local communicator ' ; ii = ii + 1 |
---|
| 304 | WRITE(ldtxt(ii),*) ' without calling MPI_Init before ! ' ; ii = ii + 1 |
---|
[2715] | 305 | kstop = kstop + 1 |
---|
[532] | 306 | ELSE |
---|
[1601] | 307 | SELECT CASE ( cn_mpi_send ) |
---|
[524] | 308 | CASE ( 'S' ) ! Standard mpi send (blocking) |
---|
[2481] | 309 | WRITE(ldtxt(ii),*) ' Standard blocking mpi send (send)' ; ii = ii + 1 |
---|
[2480] | 310 | CALL mpi_init( ierr ) |
---|
[524] | 311 | CASE ( 'B' ) ! Buffer mpi send (blocking) |
---|
[2481] | 312 | WRITE(ldtxt(ii),*) ' Buffer blocking mpi send (bsend)' ; ii = ii + 1 |
---|
| 313 | IF( Agrif_Root() ) CALL mpi_init_opa( ldtxt, ii, ierr ) |
---|
[524] | 314 | CASE ( 'I' ) ! Immediate mpi send (non-blocking send) |
---|
[2481] | 315 | WRITE(ldtxt(ii),*) ' Immediate non-blocking send (isend)' ; ii = ii + 1 |
---|
[524] | 316 | l_isend = .TRUE. |
---|
[2480] | 317 | CALL mpi_init( ierr ) |
---|
[524] | 318 | CASE DEFAULT |
---|
[2481] | 319 | WRITE(ldtxt(ii),cform_err) ; ii = ii + 1 |
---|
| 320 | WRITE(ldtxt(ii),*) ' bad value for cn_mpi_send = ', cn_mpi_send ; ii = ii + 1 |
---|
[2715] | 321 | kstop = kstop + 1 |
---|
[524] | 322 | END SELECT |
---|
[2480] | 323 | ! |
---|
[415] | 324 | ENDIF |
---|
[570] | 325 | |
---|
[2480] | 326 | IF( PRESENT(localComm) ) THEN |
---|
| 327 | IF( Agrif_Root() ) THEN |
---|
| 328 | mpi_comm_opa = localComm |
---|
| 329 | ENDIF |
---|
| 330 | ELSE |
---|
| 331 | CALL mpi_comm_dup( mpi_comm_world, mpi_comm_opa, code) |
---|
| 332 | IF( code /= MPI_SUCCESS ) THEN |
---|
[2481] | 333 | DO ji = 1, SIZE(ldtxt) |
---|
| 334 | IF( TRIM(ldtxt(ji)) /= '' ) WRITE(*,*) ldtxt(ji) ! control print of mynode |
---|
| 335 | END DO |
---|
[2480] | 336 | WRITE(*, cform_err) |
---|
| 337 | WRITE(*, *) ' lib_mpp: Error in routine mpi_comm_dup' |
---|
| 338 | CALL mpi_abort( mpi_comm_world, code, ierr ) |
---|
| 339 | ENDIF |
---|
| 340 | ENDIF |
---|
| 341 | |
---|
[1344] | 342 | CALL mpi_comm_rank( mpi_comm_opa, mpprank, ierr ) |
---|
| 343 | CALL mpi_comm_size( mpi_comm_opa, mppsize, ierr ) |
---|
[629] | 344 | mynode = mpprank |
---|
[2480] | 345 | ! |
---|
[1976] | 346 | CALL MPI_OP_CREATE(DDPDD_MPI, .TRUE., MPI_SUMDD, ierr) |
---|
| 347 | ! |
---|
[51] | 348 | END FUNCTION mynode |
---|
[3] | 349 | |
---|
| 350 | |
---|
[888] | 351 | SUBROUTINE mpp_lnk_3d( ptab, cd_type, psgn, cd_mpp, pval ) |
---|
[51] | 352 | !!---------------------------------------------------------------------- |
---|
| 353 | !! *** routine mpp_lnk_3d *** |
---|
| 354 | !! |
---|
| 355 | !! ** Purpose : Message passing manadgement |
---|
| 356 | !! |
---|
| 357 | !! ** Method : Use mppsend and mpprecv function for passing mask |
---|
| 358 | !! between processors following neighboring subdomains. |
---|
| 359 | !! domain parameters |
---|
| 360 | !! nlci : first dimension of the local subdomain |
---|
| 361 | !! nlcj : second dimension of the local subdomain |
---|
| 362 | !! nbondi : mark for "east-west local boundary" |
---|
| 363 | !! nbondj : mark for "north-south local boundary" |
---|
| 364 | !! noea : number for local neighboring processors |
---|
| 365 | !! nowe : number for local neighboring processors |
---|
| 366 | !! noso : number for local neighboring processors |
---|
| 367 | !! nono : number for local neighboring processors |
---|
| 368 | !! |
---|
| 369 | !! ** Action : ptab with update value at its periphery |
---|
| 370 | !! |
---|
| 371 | !!---------------------------------------------------------------------- |
---|
[1344] | 372 | REAL(wp), DIMENSION(jpi,jpj,jpk), INTENT(inout) :: ptab ! 3D array on which the boundary condition is applied |
---|
| 373 | CHARACTER(len=1) , INTENT(in ) :: cd_type ! define the nature of ptab array grid-points |
---|
| 374 | ! ! = T , U , V , F , W points |
---|
| 375 | REAL(wp) , INTENT(in ) :: psgn ! =-1 the sign change across the north fold boundary |
---|
| 376 | ! ! = 1. , the sign is kept |
---|
| 377 | CHARACTER(len=3), OPTIONAL , INTENT(in ) :: cd_mpp ! fill the overlap area only |
---|
| 378 | REAL(wp) , OPTIONAL , INTENT(in ) :: pval ! background value (used at closed boundaries) |
---|
| 379 | !! |
---|
[1718] | 380 | INTEGER :: ji, jj, jk, jl ! dummy loop indices |
---|
[1344] | 381 | INTEGER :: imigr, iihom, ijhom ! temporary integers |
---|
| 382 | INTEGER :: ml_req1, ml_req2, ml_err ! for key_mpi_isend |
---|
[888] | 383 | REAL(wp) :: zland |
---|
[1344] | 384 | INTEGER, DIMENSION(MPI_STATUS_SIZE) :: ml_stat ! for key_mpi_isend |
---|
[51] | 385 | !!---------------------------------------------------------------------- |
---|
[3] | 386 | |
---|
[1344] | 387 | IF( PRESENT( pval ) ) THEN ; zland = pval ! set land value |
---|
| 388 | ELSE ; zland = 0.e0 ! zero by default |
---|
| 389 | ENDIF |
---|
| 390 | |
---|
[51] | 391 | ! 1. standard boundary treatment |
---|
| 392 | ! ------------------------------ |
---|
[1718] | 393 | IF( PRESENT( cd_mpp ) ) THEN ! only fill added line/raw with existing values |
---|
[1344] | 394 | ! |
---|
[1718] | 395 | ! WARNING ptab is defined only between nld and nle |
---|
| 396 | DO jk = 1, jpk |
---|
| 397 | DO jj = nlcj+1, jpj ! added line(s) (inner only) |
---|
| 398 | ptab(nldi :nlei , jj ,jk) = ptab(nldi:nlei, nlej,jk) |
---|
| 399 | ptab(1 :nldi-1, jj ,jk) = ptab(nldi , nlej,jk) |
---|
| 400 | ptab(nlei+1:nlci , jj ,jk) = ptab( nlei, nlej,jk) |
---|
| 401 | END DO |
---|
| 402 | DO ji = nlci+1, jpi ! added column(s) (full) |
---|
| 403 | ptab(ji ,nldj :nlej ,jk) = ptab( nlei,nldj:nlej,jk) |
---|
| 404 | ptab(ji ,1 :nldj-1,jk) = ptab( nlei,nldj ,jk) |
---|
| 405 | ptab(ji ,nlej+1:jpj ,jk) = ptab( nlei, nlej,jk) |
---|
| 406 | END DO |
---|
[610] | 407 | END DO |
---|
[1344] | 408 | ! |
---|
| 409 | ELSE ! standard close or cyclic treatment |
---|
| 410 | ! |
---|
| 411 | ! ! East-West boundaries |
---|
| 412 | ! !* Cyclic east-west |
---|
| 413 | IF( nbondi == 2 .AND. (nperio == 1 .OR. nperio == 4 .OR. nperio == 6) ) THEN |
---|
[473] | 414 | ptab( 1 ,:,:) = ptab(jpim1,:,:) |
---|
| 415 | ptab(jpi,:,:) = ptab( 2 ,:,:) |
---|
[1344] | 416 | ELSE !* closed |
---|
| 417 | IF( .NOT. cd_type == 'F' ) ptab( 1 :jpreci,:,:) = zland ! south except F-point |
---|
| 418 | ptab(nlci-jpreci+1:jpi ,:,:) = zland ! north |
---|
[473] | 419 | ENDIF |
---|
[1344] | 420 | ! ! North-South boundaries (always closed) |
---|
| 421 | IF( .NOT. cd_type == 'F' ) ptab(:, 1 :jprecj,:) = zland ! south except F-point |
---|
| 422 | ptab(:,nlcj-jprecj+1:jpj ,:) = zland ! north |
---|
| 423 | ! |
---|
[51] | 424 | ENDIF |
---|
[3] | 425 | |
---|
[51] | 426 | ! 2. East and west directions exchange |
---|
| 427 | ! ------------------------------------ |
---|
[1344] | 428 | ! we play with the neigbours AND the row number because of the periodicity |
---|
| 429 | ! |
---|
| 430 | SELECT CASE ( nbondi ) ! Read Dirichlet lateral conditions |
---|
| 431 | CASE ( -1, 0, 1 ) ! all exept 2 (i.e. close case) |
---|
[51] | 432 | iihom = nlci-nreci |
---|
| 433 | DO jl = 1, jpreci |
---|
| 434 | t3ew(:,jl,:,1) = ptab(jpreci+jl,:,:) |
---|
| 435 | t3we(:,jl,:,1) = ptab(iihom +jl,:,:) |
---|
| 436 | END DO |
---|
[1345] | 437 | END SELECT |
---|
[1344] | 438 | ! |
---|
| 439 | ! ! Migrations |
---|
[51] | 440 | imigr = jpreci * jpj * jpk |
---|
[1344] | 441 | ! |
---|
[51] | 442 | SELECT CASE ( nbondi ) |
---|
| 443 | CASE ( -1 ) |
---|
[181] | 444 | CALL mppsend( 2, t3we(1,1,1,1), imigr, noea, ml_req1 ) |
---|
[3294] | 445 | CALL mpprecv( 1, t3ew(1,1,1,2), imigr, noea ) |
---|
[300] | 446 | IF(l_isend) CALL mpi_wait(ml_req1, ml_stat, ml_err) |
---|
[51] | 447 | CASE ( 0 ) |
---|
[181] | 448 | CALL mppsend( 1, t3ew(1,1,1,1), imigr, nowe, ml_req1 ) |
---|
| 449 | CALL mppsend( 2, t3we(1,1,1,1), imigr, noea, ml_req2 ) |
---|
[3294] | 450 | CALL mpprecv( 1, t3ew(1,1,1,2), imigr, noea ) |
---|
| 451 | CALL mpprecv( 2, t3we(1,1,1,2), imigr, nowe ) |
---|
[300] | 452 | IF(l_isend) CALL mpi_wait(ml_req1, ml_stat, ml_err) |
---|
| 453 | IF(l_isend) CALL mpi_wait(ml_req2, ml_stat, ml_err) |
---|
[51] | 454 | CASE ( 1 ) |
---|
[181] | 455 | CALL mppsend( 1, t3ew(1,1,1,1), imigr, nowe, ml_req1 ) |
---|
[3294] | 456 | CALL mpprecv( 2, t3we(1,1,1,2), imigr, nowe ) |
---|
[300] | 457 | IF(l_isend) CALL mpi_wait(ml_req1, ml_stat, ml_err) |
---|
[51] | 458 | END SELECT |
---|
[1344] | 459 | ! |
---|
| 460 | ! ! Write Dirichlet lateral conditions |
---|
[51] | 461 | iihom = nlci-jpreci |
---|
[1344] | 462 | ! |
---|
[51] | 463 | SELECT CASE ( nbondi ) |
---|
| 464 | CASE ( -1 ) |
---|
| 465 | DO jl = 1, jpreci |
---|
| 466 | ptab(iihom+jl,:,:) = t3ew(:,jl,:,2) |
---|
| 467 | END DO |
---|
| 468 | CASE ( 0 ) |
---|
| 469 | DO jl = 1, jpreci |
---|
| 470 | ptab(jl ,:,:) = t3we(:,jl,:,2) |
---|
| 471 | ptab(iihom+jl,:,:) = t3ew(:,jl,:,2) |
---|
| 472 | END DO |
---|
| 473 | CASE ( 1 ) |
---|
| 474 | DO jl = 1, jpreci |
---|
| 475 | ptab(jl ,:,:) = t3we(:,jl,:,2) |
---|
| 476 | END DO |
---|
| 477 | END SELECT |
---|
[3] | 478 | |
---|
| 479 | |
---|
[51] | 480 | ! 3. North and south directions |
---|
| 481 | ! ----------------------------- |
---|
[1344] | 482 | ! always closed : we play only with the neigbours |
---|
| 483 | ! |
---|
| 484 | IF( nbondj /= 2 ) THEN ! Read Dirichlet lateral conditions |
---|
[51] | 485 | ijhom = nlcj-nrecj |
---|
| 486 | DO jl = 1, jprecj |
---|
| 487 | t3sn(:,jl,:,1) = ptab(:,ijhom +jl,:) |
---|
| 488 | t3ns(:,jl,:,1) = ptab(:,jprecj+jl,:) |
---|
| 489 | END DO |
---|
| 490 | ENDIF |
---|
[1344] | 491 | ! |
---|
| 492 | ! ! Migrations |
---|
[51] | 493 | imigr = jprecj * jpi * jpk |
---|
[1344] | 494 | ! |
---|
[51] | 495 | SELECT CASE ( nbondj ) |
---|
| 496 | CASE ( -1 ) |
---|
[181] | 497 | CALL mppsend( 4, t3sn(1,1,1,1), imigr, nono, ml_req1 ) |
---|
[3294] | 498 | CALL mpprecv( 3, t3ns(1,1,1,2), imigr, nono ) |
---|
[300] | 499 | IF(l_isend) CALL mpi_wait(ml_req1, ml_stat, ml_err) |
---|
[51] | 500 | CASE ( 0 ) |
---|
[181] | 501 | CALL mppsend( 3, t3ns(1,1,1,1), imigr, noso, ml_req1 ) |
---|
| 502 | CALL mppsend( 4, t3sn(1,1,1,1), imigr, nono, ml_req2 ) |
---|
[3294] | 503 | CALL mpprecv( 3, t3ns(1,1,1,2), imigr, nono ) |
---|
| 504 | CALL mpprecv( 4, t3sn(1,1,1,2), imigr, noso ) |
---|
[300] | 505 | IF(l_isend) CALL mpi_wait(ml_req1, ml_stat, ml_err) |
---|
| 506 | IF(l_isend) CALL mpi_wait(ml_req2, ml_stat, ml_err) |
---|
[51] | 507 | CASE ( 1 ) |
---|
[181] | 508 | CALL mppsend( 3, t3ns(1,1,1,1), imigr, noso, ml_req1 ) |
---|
[3294] | 509 | CALL mpprecv( 4, t3sn(1,1,1,2), imigr, noso ) |
---|
[300] | 510 | IF(l_isend) CALL mpi_wait(ml_req1, ml_stat, ml_err) |
---|
[51] | 511 | END SELECT |
---|
[1344] | 512 | ! |
---|
| 513 | ! ! Write Dirichlet lateral conditions |
---|
[51] | 514 | ijhom = nlcj-jprecj |
---|
[1344] | 515 | ! |
---|
[51] | 516 | SELECT CASE ( nbondj ) |
---|
| 517 | CASE ( -1 ) |
---|
| 518 | DO jl = 1, jprecj |
---|
| 519 | ptab(:,ijhom+jl,:) = t3ns(:,jl,:,2) |
---|
| 520 | END DO |
---|
| 521 | CASE ( 0 ) |
---|
| 522 | DO jl = 1, jprecj |
---|
| 523 | ptab(:,jl ,:) = t3sn(:,jl,:,2) |
---|
| 524 | ptab(:,ijhom+jl,:) = t3ns(:,jl,:,2) |
---|
| 525 | END DO |
---|
| 526 | CASE ( 1 ) |
---|
| 527 | DO jl = 1, jprecj |
---|
| 528 | ptab(:,jl,:) = t3sn(:,jl,:,2) |
---|
| 529 | END DO |
---|
| 530 | END SELECT |
---|
[3] | 531 | |
---|
| 532 | |
---|
[51] | 533 | ! 4. north fold treatment |
---|
| 534 | ! ----------------------- |
---|
[1344] | 535 | ! |
---|
| 536 | IF( npolj /= 0 .AND. .NOT. PRESENT(cd_mpp) ) THEN |
---|
| 537 | ! |
---|
| 538 | SELECT CASE ( jpni ) |
---|
| 539 | CASE ( 1 ) ; CALL lbc_nfd ( ptab, cd_type, psgn ) ! only 1 northern proc, no mpp |
---|
| 540 | CASE DEFAULT ; CALL mpp_lbc_north( ptab, cd_type, psgn ) ! for all northern procs. |
---|
| 541 | END SELECT |
---|
| 542 | ! |
---|
[473] | 543 | ENDIF |
---|
[1344] | 544 | ! |
---|
[51] | 545 | END SUBROUTINE mpp_lnk_3d |
---|
[3] | 546 | |
---|
| 547 | |
---|
[888] | 548 | SUBROUTINE mpp_lnk_2d( pt2d, cd_type, psgn, cd_mpp, pval ) |
---|
[51] | 549 | !!---------------------------------------------------------------------- |
---|
| 550 | !! *** routine mpp_lnk_2d *** |
---|
| 551 | !! |
---|
| 552 | !! ** Purpose : Message passing manadgement for 2d array |
---|
| 553 | !! |
---|
| 554 | !! ** Method : Use mppsend and mpprecv function for passing mask |
---|
| 555 | !! between processors following neighboring subdomains. |
---|
| 556 | !! domain parameters |
---|
| 557 | !! nlci : first dimension of the local subdomain |
---|
| 558 | !! nlcj : second dimension of the local subdomain |
---|
| 559 | !! nbondi : mark for "east-west local boundary" |
---|
| 560 | !! nbondj : mark for "north-south local boundary" |
---|
| 561 | !! noea : number for local neighboring processors |
---|
| 562 | !! nowe : number for local neighboring processors |
---|
| 563 | !! noso : number for local neighboring processors |
---|
| 564 | !! nono : number for local neighboring processors |
---|
| 565 | !! |
---|
| 566 | !!---------------------------------------------------------------------- |
---|
[1344] | 567 | REAL(wp), DIMENSION(jpi,jpj), INTENT(inout) :: pt2d ! 2D array on which the boundary condition is applied |
---|
| 568 | CHARACTER(len=1) , INTENT(in ) :: cd_type ! define the nature of ptab array grid-points |
---|
| 569 | ! ! = T , U , V , F , W and I points |
---|
| 570 | REAL(wp) , INTENT(in ) :: psgn ! =-1 the sign change across the north fold boundary |
---|
| 571 | ! ! = 1. , the sign is kept |
---|
| 572 | CHARACTER(len=3), OPTIONAL , INTENT(in ) :: cd_mpp ! fill the overlap area only |
---|
| 573 | REAL(wp) , OPTIONAL , INTENT(in ) :: pval ! background value (used at closed boundaries) |
---|
| 574 | !! |
---|
| 575 | INTEGER :: ji, jj, jl ! dummy loop indices |
---|
| 576 | INTEGER :: imigr, iihom, ijhom ! temporary integers |
---|
| 577 | INTEGER :: ml_req1, ml_req2, ml_err ! for key_mpi_isend |
---|
[888] | 578 | REAL(wp) :: zland |
---|
[1344] | 579 | INTEGER, DIMENSION(MPI_STATUS_SIZE) :: ml_stat ! for key_mpi_isend |
---|
[51] | 580 | !!---------------------------------------------------------------------- |
---|
[3] | 581 | |
---|
[1344] | 582 | IF( PRESENT( pval ) ) THEN ; zland = pval ! set land value |
---|
| 583 | ELSE ; zland = 0.e0 ! zero by default |
---|
[888] | 584 | ENDIF |
---|
| 585 | |
---|
[51] | 586 | ! 1. standard boundary treatment |
---|
| 587 | ! ------------------------------ |
---|
[1344] | 588 | ! |
---|
[1718] | 589 | IF( PRESENT( cd_mpp ) ) THEN ! only fill added line/raw with existing values |
---|
[1344] | 590 | ! |
---|
[1718] | 591 | ! WARNING pt2d is defined only between nld and nle |
---|
| 592 | DO jj = nlcj+1, jpj ! added line(s) (inner only) |
---|
| 593 | pt2d(nldi :nlei , jj ) = pt2d(nldi:nlei, nlej) |
---|
| 594 | pt2d(1 :nldi-1, jj ) = pt2d(nldi , nlej) |
---|
| 595 | pt2d(nlei+1:nlci , jj ) = pt2d( nlei, nlej) |
---|
[610] | 596 | END DO |
---|
[1718] | 597 | DO ji = nlci+1, jpi ! added column(s) (full) |
---|
| 598 | pt2d(ji ,nldj :nlej ) = pt2d( nlei,nldj:nlej) |
---|
| 599 | pt2d(ji ,1 :nldj-1) = pt2d( nlei,nldj ) |
---|
| 600 | pt2d(ji ,nlej+1:jpj ) = pt2d( nlei, nlej) |
---|
[1344] | 601 | END DO |
---|
| 602 | ! |
---|
| 603 | ELSE ! standard close or cyclic treatment |
---|
| 604 | ! |
---|
| 605 | ! ! East-West boundaries |
---|
| 606 | IF( nbondi == 2 .AND. & ! Cyclic east-west |
---|
[473] | 607 | & (nperio == 1 .OR. nperio == 4 .OR. nperio == 6) ) THEN |
---|
[1344] | 608 | pt2d( 1 ,:) = pt2d(jpim1,:) ! west |
---|
| 609 | pt2d(jpi,:) = pt2d( 2 ,:) ! east |
---|
| 610 | ELSE ! closed |
---|
| 611 | IF( .NOT. cd_type == 'F' ) pt2d( 1 :jpreci,:) = zland ! south except F-point |
---|
| 612 | pt2d(nlci-jpreci+1:jpi ,:) = zland ! north |
---|
[473] | 613 | ENDIF |
---|
[1344] | 614 | ! ! North-South boundaries (always closed) |
---|
| 615 | IF( .NOT. cd_type == 'F' ) pt2d(:, 1 :jprecj) = zland !south except F-point |
---|
| 616 | pt2d(:,nlcj-jprecj+1:jpj ) = zland ! north |
---|
| 617 | ! |
---|
[51] | 618 | ENDIF |
---|
[3] | 619 | |
---|
[1344] | 620 | ! 2. East and west directions exchange |
---|
| 621 | ! ------------------------------------ |
---|
| 622 | ! we play with the neigbours AND the row number because of the periodicity |
---|
| 623 | ! |
---|
| 624 | SELECT CASE ( nbondi ) ! Read Dirichlet lateral conditions |
---|
| 625 | CASE ( -1, 0, 1 ) ! all exept 2 (i.e. close case) |
---|
[51] | 626 | iihom = nlci-nreci |
---|
| 627 | DO jl = 1, jpreci |
---|
| 628 | t2ew(:,jl,1) = pt2d(jpreci+jl,:) |
---|
| 629 | t2we(:,jl,1) = pt2d(iihom +jl,:) |
---|
| 630 | END DO |
---|
| 631 | END SELECT |
---|
[1344] | 632 | ! |
---|
| 633 | ! ! Migrations |
---|
[51] | 634 | imigr = jpreci * jpj |
---|
[1344] | 635 | ! |
---|
[51] | 636 | SELECT CASE ( nbondi ) |
---|
| 637 | CASE ( -1 ) |
---|
[181] | 638 | CALL mppsend( 2, t2we(1,1,1), imigr, noea, ml_req1 ) |
---|
[3294] | 639 | CALL mpprecv( 1, t2ew(1,1,2), imigr, noea ) |
---|
[300] | 640 | IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err) |
---|
[51] | 641 | CASE ( 0 ) |
---|
[181] | 642 | CALL mppsend( 1, t2ew(1,1,1), imigr, nowe, ml_req1 ) |
---|
| 643 | CALL mppsend( 2, t2we(1,1,1), imigr, noea, ml_req2 ) |
---|
[3294] | 644 | CALL mpprecv( 1, t2ew(1,1,2), imigr, noea ) |
---|
| 645 | CALL mpprecv( 2, t2we(1,1,2), imigr, nowe ) |
---|
[300] | 646 | IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err) |
---|
| 647 | IF(l_isend) CALL mpi_wait(ml_req2,ml_stat,ml_err) |
---|
[51] | 648 | CASE ( 1 ) |
---|
[181] | 649 | CALL mppsend( 1, t2ew(1,1,1), imigr, nowe, ml_req1 ) |
---|
[3294] | 650 | CALL mpprecv( 2, t2we(1,1,2), imigr, nowe ) |
---|
[300] | 651 | IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err) |
---|
[51] | 652 | END SELECT |
---|
[1344] | 653 | ! |
---|
| 654 | ! ! Write Dirichlet lateral conditions |
---|
[51] | 655 | iihom = nlci - jpreci |
---|
[1344] | 656 | ! |
---|
[51] | 657 | SELECT CASE ( nbondi ) |
---|
| 658 | CASE ( -1 ) |
---|
| 659 | DO jl = 1, jpreci |
---|
| 660 | pt2d(iihom+jl,:) = t2ew(:,jl,2) |
---|
| 661 | END DO |
---|
| 662 | CASE ( 0 ) |
---|
| 663 | DO jl = 1, jpreci |
---|
| 664 | pt2d(jl ,:) = t2we(:,jl,2) |
---|
| 665 | pt2d(iihom+jl,:) = t2ew(:,jl,2) |
---|
| 666 | END DO |
---|
| 667 | CASE ( 1 ) |
---|
| 668 | DO jl = 1, jpreci |
---|
| 669 | pt2d(jl ,:) = t2we(:,jl,2) |
---|
| 670 | END DO |
---|
| 671 | END SELECT |
---|
[3] | 672 | |
---|
| 673 | |
---|
[51] | 674 | ! 3. North and south directions |
---|
| 675 | ! ----------------------------- |
---|
[1344] | 676 | ! always closed : we play only with the neigbours |
---|
| 677 | ! |
---|
| 678 | IF( nbondj /= 2 ) THEN ! Read Dirichlet lateral conditions |
---|
[51] | 679 | ijhom = nlcj-nrecj |
---|
| 680 | DO jl = 1, jprecj |
---|
| 681 | t2sn(:,jl,1) = pt2d(:,ijhom +jl) |
---|
| 682 | t2ns(:,jl,1) = pt2d(:,jprecj+jl) |
---|
| 683 | END DO |
---|
| 684 | ENDIF |
---|
[1344] | 685 | ! |
---|
| 686 | ! ! Migrations |
---|
[51] | 687 | imigr = jprecj * jpi |
---|
[1344] | 688 | ! |
---|
[51] | 689 | SELECT CASE ( nbondj ) |
---|
| 690 | CASE ( -1 ) |
---|
[181] | 691 | CALL mppsend( 4, t2sn(1,1,1), imigr, nono, ml_req1 ) |
---|
[3294] | 692 | CALL mpprecv( 3, t2ns(1,1,2), imigr, nono ) |
---|
[300] | 693 | IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err) |
---|
[51] | 694 | CASE ( 0 ) |
---|
[181] | 695 | CALL mppsend( 3, t2ns(1,1,1), imigr, noso, ml_req1 ) |
---|
| 696 | CALL mppsend( 4, t2sn(1,1,1), imigr, nono, ml_req2 ) |
---|
[3294] | 697 | CALL mpprecv( 3, t2ns(1,1,2), imigr, nono ) |
---|
| 698 | CALL mpprecv( 4, t2sn(1,1,2), imigr, noso ) |
---|
[300] | 699 | IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err) |
---|
| 700 | IF(l_isend) CALL mpi_wait(ml_req2,ml_stat,ml_err) |
---|
[51] | 701 | CASE ( 1 ) |
---|
[181] | 702 | CALL mppsend( 3, t2ns(1,1,1), imigr, noso, ml_req1 ) |
---|
[3294] | 703 | CALL mpprecv( 4, t2sn(1,1,2), imigr, noso ) |
---|
[300] | 704 | IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err) |
---|
[51] | 705 | END SELECT |
---|
[1344] | 706 | ! |
---|
| 707 | ! ! Write Dirichlet lateral conditions |
---|
[51] | 708 | ijhom = nlcj - jprecj |
---|
[1344] | 709 | ! |
---|
[51] | 710 | SELECT CASE ( nbondj ) |
---|
| 711 | CASE ( -1 ) |
---|
| 712 | DO jl = 1, jprecj |
---|
| 713 | pt2d(:,ijhom+jl) = t2ns(:,jl,2) |
---|
| 714 | END DO |
---|
| 715 | CASE ( 0 ) |
---|
| 716 | DO jl = 1, jprecj |
---|
| 717 | pt2d(:,jl ) = t2sn(:,jl,2) |
---|
| 718 | pt2d(:,ijhom+jl) = t2ns(:,jl,2) |
---|
| 719 | END DO |
---|
| 720 | CASE ( 1 ) |
---|
| 721 | DO jl = 1, jprecj |
---|
| 722 | pt2d(:,jl ) = t2sn(:,jl,2) |
---|
| 723 | END DO |
---|
[1344] | 724 | END SELECT |
---|
[3] | 725 | |
---|
[1344] | 726 | |
---|
[51] | 727 | ! 4. north fold treatment |
---|
| 728 | ! ----------------------- |
---|
[1344] | 729 | ! |
---|
| 730 | IF( npolj /= 0 .AND. .NOT. PRESENT(cd_mpp) ) THEN |
---|
| 731 | ! |
---|
| 732 | SELECT CASE ( jpni ) |
---|
| 733 | CASE ( 1 ) ; CALL lbc_nfd ( pt2d, cd_type, psgn ) ! only 1 northern proc, no mpp |
---|
| 734 | CASE DEFAULT ; CALL mpp_lbc_north( pt2d, cd_type, psgn ) ! for all northern procs. |
---|
| 735 | END SELECT |
---|
| 736 | ! |
---|
[473] | 737 | ENDIF |
---|
[1344] | 738 | ! |
---|
[51] | 739 | END SUBROUTINE mpp_lnk_2d |
---|
[3] | 740 | |
---|
| 741 | |
---|
[473] | 742 | SUBROUTINE mpp_lnk_3d_gather( ptab1, cd_type1, ptab2, cd_type2, psgn ) |
---|
| 743 | !!---------------------------------------------------------------------- |
---|
| 744 | !! *** routine mpp_lnk_3d_gather *** |
---|
| 745 | !! |
---|
| 746 | !! ** Purpose : Message passing manadgement for two 3D arrays |
---|
| 747 | !! |
---|
| 748 | !! ** Method : Use mppsend and mpprecv function for passing mask |
---|
| 749 | !! between processors following neighboring subdomains. |
---|
| 750 | !! domain parameters |
---|
| 751 | !! nlci : first dimension of the local subdomain |
---|
| 752 | !! nlcj : second dimension of the local subdomain |
---|
| 753 | !! nbondi : mark for "east-west local boundary" |
---|
| 754 | !! nbondj : mark for "north-south local boundary" |
---|
| 755 | !! noea : number for local neighboring processors |
---|
| 756 | !! nowe : number for local neighboring processors |
---|
| 757 | !! noso : number for local neighboring processors |
---|
| 758 | !! nono : number for local neighboring processors |
---|
| 759 | !! |
---|
| 760 | !! ** Action : ptab1 and ptab2 with update value at its periphery |
---|
| 761 | !! |
---|
| 762 | !!---------------------------------------------------------------------- |
---|
[1344] | 763 | REAL(wp), DIMENSION(jpi,jpj,jpk), INTENT(inout) :: ptab1 ! first and second 3D array on which |
---|
| 764 | REAL(wp), DIMENSION(jpi,jpj,jpk), INTENT(inout) :: ptab2 ! the boundary condition is applied |
---|
| 765 | CHARACTER(len=1) , INTENT(in ) :: cd_type1 ! nature of ptab1 and ptab2 arrays |
---|
| 766 | CHARACTER(len=1) , INTENT(in ) :: cd_type2 ! i.e. grid-points = T , U , V , F or W points |
---|
| 767 | REAL(wp) , INTENT(in ) :: psgn ! =-1 the sign change across the north fold boundary |
---|
| 768 | !! ! = 1. , the sign is kept |
---|
| 769 | INTEGER :: jl ! dummy loop indices |
---|
| 770 | INTEGER :: imigr, iihom, ijhom ! temporary integers |
---|
| 771 | INTEGER :: ml_req1, ml_req2, ml_err ! for key_mpi_isend |
---|
| 772 | INTEGER, DIMENSION(MPI_STATUS_SIZE) :: ml_stat ! for key_mpi_isend |
---|
[473] | 773 | !!---------------------------------------------------------------------- |
---|
| 774 | |
---|
| 775 | ! 1. standard boundary treatment |
---|
| 776 | ! ------------------------------ |
---|
[1344] | 777 | ! ! East-West boundaries |
---|
| 778 | ! !* Cyclic east-west |
---|
| 779 | IF( nbondi == 2 .AND. (nperio == 1 .OR. nperio == 4 .OR. nperio == 6) ) THEN |
---|
[473] | 780 | ptab1( 1 ,:,:) = ptab1(jpim1,:,:) |
---|
| 781 | ptab1(jpi,:,:) = ptab1( 2 ,:,:) |
---|
| 782 | ptab2( 1 ,:,:) = ptab2(jpim1,:,:) |
---|
| 783 | ptab2(jpi,:,:) = ptab2( 2 ,:,:) |
---|
[1344] | 784 | ELSE !* closed |
---|
| 785 | IF( .NOT. cd_type1 == 'F' ) ptab1( 1 :jpreci,:,:) = 0.e0 ! south except at F-point |
---|
| 786 | IF( .NOT. cd_type2 == 'F' ) ptab2( 1 :jpreci,:,:) = 0.e0 |
---|
| 787 | ptab1(nlci-jpreci+1:jpi ,:,:) = 0.e0 ! north |
---|
| 788 | ptab2(nlci-jpreci+1:jpi ,:,:) = 0.e0 |
---|
[473] | 789 | ENDIF |
---|
| 790 | |
---|
[1344] | 791 | |
---|
| 792 | ! ! North-South boundaries |
---|
| 793 | IF( .NOT. cd_type1 == 'F' ) ptab1(:, 1 :jprecj,:) = 0.e0 ! south except at F-point |
---|
| 794 | IF( .NOT. cd_type2 == 'F' ) ptab2(:, 1 :jprecj,:) = 0.e0 |
---|
| 795 | ptab1(:,nlcj-jprecj+1:jpj ,:) = 0.e0 ! north |
---|
| 796 | ptab2(:,nlcj-jprecj+1:jpj ,:) = 0.e0 |
---|
[473] | 797 | |
---|
| 798 | |
---|
| 799 | ! 2. East and west directions exchange |
---|
| 800 | ! ------------------------------------ |
---|
[1344] | 801 | ! we play with the neigbours AND the row number because of the periodicity |
---|
| 802 | ! |
---|
| 803 | SELECT CASE ( nbondi ) ! Read Dirichlet lateral conditions |
---|
| 804 | CASE ( -1, 0, 1 ) ! all exept 2 (i.e. close case) |
---|
[473] | 805 | iihom = nlci-nreci |
---|
| 806 | DO jl = 1, jpreci |
---|
| 807 | t4ew(:,jl,:,1,1) = ptab1(jpreci+jl,:,:) |
---|
| 808 | t4we(:,jl,:,1,1) = ptab1(iihom +jl,:,:) |
---|
| 809 | t4ew(:,jl,:,2,1) = ptab2(jpreci+jl,:,:) |
---|
| 810 | t4we(:,jl,:,2,1) = ptab2(iihom +jl,:,:) |
---|
| 811 | END DO |
---|
| 812 | END SELECT |
---|
[1344] | 813 | ! |
---|
| 814 | ! ! Migrations |
---|
[473] | 815 | imigr = jpreci * jpj * jpk *2 |
---|
[1344] | 816 | ! |
---|
[473] | 817 | SELECT CASE ( nbondi ) |
---|
| 818 | CASE ( -1 ) |
---|
| 819 | CALL mppsend( 2, t4we(1,1,1,1,1), imigr, noea, ml_req1 ) |
---|
[3294] | 820 | CALL mpprecv( 1, t4ew(1,1,1,1,2), imigr, noea ) |
---|
[473] | 821 | IF(l_isend) CALL mpi_wait(ml_req1, ml_stat, ml_err) |
---|
| 822 | CASE ( 0 ) |
---|
| 823 | CALL mppsend( 1, t4ew(1,1,1,1,1), imigr, nowe, ml_req1 ) |
---|
| 824 | CALL mppsend( 2, t4we(1,1,1,1,1), imigr, noea, ml_req2 ) |
---|
[3294] | 825 | CALL mpprecv( 1, t4ew(1,1,1,1,2), imigr, noea ) |
---|
| 826 | CALL mpprecv( 2, t4we(1,1,1,1,2), imigr, nowe ) |
---|
[473] | 827 | IF(l_isend) CALL mpi_wait(ml_req1, ml_stat, ml_err) |
---|
| 828 | IF(l_isend) CALL mpi_wait(ml_req2, ml_stat, ml_err) |
---|
| 829 | CASE ( 1 ) |
---|
| 830 | CALL mppsend( 1, t4ew(1,1,1,1,1), imigr, nowe, ml_req1 ) |
---|
[3294] | 831 | CALL mpprecv( 2, t4we(1,1,1,1,2), imigr, nowe ) |
---|
[473] | 832 | IF(l_isend) CALL mpi_wait(ml_req1, ml_stat, ml_err) |
---|
| 833 | END SELECT |
---|
[1344] | 834 | ! |
---|
| 835 | ! ! Write Dirichlet lateral conditions |
---|
| 836 | iihom = nlci - jpreci |
---|
| 837 | ! |
---|
[473] | 838 | SELECT CASE ( nbondi ) |
---|
| 839 | CASE ( -1 ) |
---|
| 840 | DO jl = 1, jpreci |
---|
| 841 | ptab1(iihom+jl,:,:) = t4ew(:,jl,:,1,2) |
---|
| 842 | ptab2(iihom+jl,:,:) = t4ew(:,jl,:,2,2) |
---|
| 843 | END DO |
---|
| 844 | CASE ( 0 ) |
---|
| 845 | DO jl = 1, jpreci |
---|
| 846 | ptab1(jl ,:,:) = t4we(:,jl,:,1,2) |
---|
| 847 | ptab1(iihom+jl,:,:) = t4ew(:,jl,:,1,2) |
---|
| 848 | ptab2(jl ,:,:) = t4we(:,jl,:,2,2) |
---|
| 849 | ptab2(iihom+jl,:,:) = t4ew(:,jl,:,2,2) |
---|
| 850 | END DO |
---|
| 851 | CASE ( 1 ) |
---|
| 852 | DO jl = 1, jpreci |
---|
| 853 | ptab1(jl ,:,:) = t4we(:,jl,:,1,2) |
---|
| 854 | ptab2(jl ,:,:) = t4we(:,jl,:,2,2) |
---|
| 855 | END DO |
---|
| 856 | END SELECT |
---|
| 857 | |
---|
| 858 | |
---|
| 859 | ! 3. North and south directions |
---|
| 860 | ! ----------------------------- |
---|
[1344] | 861 | ! always closed : we play only with the neigbours |
---|
| 862 | ! |
---|
| 863 | IF( nbondj /= 2 ) THEN ! Read Dirichlet lateral conditions |
---|
| 864 | ijhom = nlcj - nrecj |
---|
[473] | 865 | DO jl = 1, jprecj |
---|
| 866 | t4sn(:,jl,:,1,1) = ptab1(:,ijhom +jl,:) |
---|
| 867 | t4ns(:,jl,:,1,1) = ptab1(:,jprecj+jl,:) |
---|
| 868 | t4sn(:,jl,:,2,1) = ptab2(:,ijhom +jl,:) |
---|
| 869 | t4ns(:,jl,:,2,1) = ptab2(:,jprecj+jl,:) |
---|
| 870 | END DO |
---|
| 871 | ENDIF |
---|
[1344] | 872 | ! |
---|
| 873 | ! ! Migrations |
---|
[473] | 874 | imigr = jprecj * jpi * jpk * 2 |
---|
[1344] | 875 | ! |
---|
[473] | 876 | SELECT CASE ( nbondj ) |
---|
| 877 | CASE ( -1 ) |
---|
| 878 | CALL mppsend( 4, t4sn(1,1,1,1,1), imigr, nono, ml_req1 ) |
---|
[3294] | 879 | CALL mpprecv( 3, t4ns(1,1,1,1,2), imigr, nono ) |
---|
[473] | 880 | IF(l_isend) CALL mpi_wait(ml_req1, ml_stat, ml_err) |
---|
| 881 | CASE ( 0 ) |
---|
| 882 | CALL mppsend( 3, t4ns(1,1,1,1,1), imigr, noso, ml_req1 ) |
---|
| 883 | CALL mppsend( 4, t4sn(1,1,1,1,1), imigr, nono, ml_req2 ) |
---|
[3294] | 884 | CALL mpprecv( 3, t4ns(1,1,1,1,2), imigr, nono ) |
---|
| 885 | CALL mpprecv( 4, t4sn(1,1,1,1,2), imigr, noso ) |
---|
[473] | 886 | IF(l_isend) CALL mpi_wait(ml_req1, ml_stat, ml_err) |
---|
| 887 | IF(l_isend) CALL mpi_wait(ml_req2, ml_stat, ml_err) |
---|
| 888 | CASE ( 1 ) |
---|
| 889 | CALL mppsend( 3, t4ns(1,1,1,1,1), imigr, noso, ml_req1 ) |
---|
[3294] | 890 | CALL mpprecv( 4, t4sn(1,1,1,1,2), imigr, noso ) |
---|
[473] | 891 | IF(l_isend) CALL mpi_wait(ml_req1, ml_stat, ml_err) |
---|
| 892 | END SELECT |
---|
[1344] | 893 | ! |
---|
| 894 | ! ! Write Dirichlet lateral conditions |
---|
| 895 | ijhom = nlcj - jprecj |
---|
| 896 | ! |
---|
[473] | 897 | SELECT CASE ( nbondj ) |
---|
| 898 | CASE ( -1 ) |
---|
| 899 | DO jl = 1, jprecj |
---|
| 900 | ptab1(:,ijhom+jl,:) = t4ns(:,jl,:,1,2) |
---|
| 901 | ptab2(:,ijhom+jl,:) = t4ns(:,jl,:,2,2) |
---|
| 902 | END DO |
---|
| 903 | CASE ( 0 ) |
---|
| 904 | DO jl = 1, jprecj |
---|
| 905 | ptab1(:,jl ,:) = t4sn(:,jl,:,1,2) |
---|
| 906 | ptab1(:,ijhom+jl,:) = t4ns(:,jl,:,1,2) |
---|
| 907 | ptab2(:,jl ,:) = t4sn(:,jl,:,2,2) |
---|
| 908 | ptab2(:,ijhom+jl,:) = t4ns(:,jl,:,2,2) |
---|
| 909 | END DO |
---|
| 910 | CASE ( 1 ) |
---|
| 911 | DO jl = 1, jprecj |
---|
| 912 | ptab1(:,jl,:) = t4sn(:,jl,:,1,2) |
---|
| 913 | ptab2(:,jl,:) = t4sn(:,jl,:,2,2) |
---|
| 914 | END DO |
---|
| 915 | END SELECT |
---|
| 916 | |
---|
| 917 | |
---|
| 918 | ! 4. north fold treatment |
---|
| 919 | ! ----------------------- |
---|
[1344] | 920 | IF( npolj /= 0 ) THEN |
---|
| 921 | ! |
---|
| 922 | SELECT CASE ( jpni ) |
---|
| 923 | CASE ( 1 ) |
---|
| 924 | CALL lbc_nfd ( ptab1, cd_type1, psgn ) ! only for northern procs. |
---|
| 925 | CALL lbc_nfd ( ptab2, cd_type2, psgn ) |
---|
| 926 | CASE DEFAULT |
---|
| 927 | CALL mpp_lbc_north( ptab1, cd_type1, psgn ) ! for all northern procs. |
---|
| 928 | CALL mpp_lbc_north (ptab2, cd_type2, psgn) |
---|
| 929 | END SELECT |
---|
| 930 | ! |
---|
| 931 | ENDIF |
---|
| 932 | ! |
---|
[473] | 933 | END SUBROUTINE mpp_lnk_3d_gather |
---|
| 934 | |
---|
| 935 | |
---|
[3609] | 936 | SUBROUTINE mpp_lnk_2d_e( pt2d, cd_type, psgn, jpri, jprj ) |
---|
[311] | 937 | !!---------------------------------------------------------------------- |
---|
| 938 | !! *** routine mpp_lnk_2d_e *** |
---|
| 939 | !! |
---|
| 940 | !! ** Purpose : Message passing manadgement for 2d array (with halo) |
---|
| 941 | !! |
---|
| 942 | !! ** Method : Use mppsend and mpprecv function for passing mask |
---|
| 943 | !! between processors following neighboring subdomains. |
---|
| 944 | !! domain parameters |
---|
| 945 | !! nlci : first dimension of the local subdomain |
---|
| 946 | !! nlcj : second dimension of the local subdomain |
---|
[3609] | 947 | !! jpri : number of rows for extra outer halo |
---|
| 948 | !! jprj : number of columns for extra outer halo |
---|
[311] | 949 | !! nbondi : mark for "east-west local boundary" |
---|
| 950 | !! nbondj : mark for "north-south local boundary" |
---|
| 951 | !! noea : number for local neighboring processors |
---|
| 952 | !! nowe : number for local neighboring processors |
---|
| 953 | !! noso : number for local neighboring processors |
---|
| 954 | !! nono : number for local neighboring processors |
---|
| 955 | !! |
---|
| 956 | !!---------------------------------------------------------------------- |
---|
[3609] | 957 | INTEGER , INTENT(in ) :: jpri |
---|
| 958 | INTEGER , INTENT(in ) :: jprj |
---|
| 959 | REAL(wp), DIMENSION(1-jpri:jpi+jpri,1-jprj:jpj+jprj), INTENT(inout) :: pt2d ! 2D array with extra halo |
---|
| 960 | CHARACTER(len=1) , INTENT(in ) :: cd_type ! nature of ptab array grid-points |
---|
| 961 | ! ! = T , U , V , F , W and I points |
---|
| 962 | REAL(wp) , INTENT(in ) :: psgn ! =-1 the sign change across the |
---|
| 963 | !! ! north boundary, = 1. otherwise |
---|
[1344] | 964 | INTEGER :: jl ! dummy loop indices |
---|
| 965 | INTEGER :: imigr, iihom, ijhom ! temporary integers |
---|
| 966 | INTEGER :: ipreci, iprecj ! temporary integers |
---|
| 967 | INTEGER :: ml_req1, ml_req2, ml_err ! for key_mpi_isend |
---|
| 968 | INTEGER, DIMENSION(MPI_STATUS_SIZE) :: ml_stat ! for key_mpi_isend |
---|
[3609] | 969 | !! |
---|
| 970 | REAL(wp), DIMENSION(1-jpri:jpi+jpri,jprecj+jprj,2) :: r2dns |
---|
| 971 | REAL(wp), DIMENSION(1-jpri:jpi+jpri,jprecj+jprj,2) :: r2dsn |
---|
| 972 | REAL(wp), DIMENSION(1-jprj:jpj+jprj,jpreci+jpri,2) :: r2dwe |
---|
| 973 | REAL(wp), DIMENSION(1-jprj:jpj+jprj,jpreci+jpri,2) :: r2dew |
---|
[1344] | 974 | !!---------------------------------------------------------------------- |
---|
[311] | 975 | |
---|
[3609] | 976 | ipreci = jpreci + jpri ! take into account outer extra 2D overlap area |
---|
| 977 | iprecj = jprecj + jprj |
---|
[311] | 978 | |
---|
| 979 | |
---|
| 980 | ! 1. standard boundary treatment |
---|
| 981 | ! ------------------------------ |
---|
[1344] | 982 | ! Order matters Here !!!! |
---|
| 983 | ! |
---|
| 984 | ! !* North-South boundaries (always colsed) |
---|
[3609] | 985 | IF( .NOT. cd_type == 'F' ) pt2d(:, 1-jprj : jprecj ) = 0.e0 ! south except at F-point |
---|
| 986 | pt2d(:,nlcj-jprecj+1:jpj+jprj) = 0.e0 ! north |
---|
[1344] | 987 | |
---|
| 988 | ! ! East-West boundaries |
---|
| 989 | ! !* Cyclic east-west |
---|
| 990 | IF( nbondi == 2 .AND. (nperio == 1 .OR. nperio == 4 .OR. nperio == 6) ) THEN |
---|
[3609] | 991 | pt2d(1-jpri: 1 ,:) = pt2d(jpim1-jpri: jpim1 ,:) ! east |
---|
| 992 | pt2d( jpi :jpi+jpri,:) = pt2d( 2 :2+jpri,:) ! west |
---|
[1344] | 993 | ! |
---|
| 994 | ELSE !* closed |
---|
[3609] | 995 | IF( .NOT. cd_type == 'F' ) pt2d( 1-jpri :jpreci ,:) = 0.e0 ! south except at F-point |
---|
| 996 | pt2d(nlci-jpreci+1:jpi+jpri,:) = 0.e0 ! north |
---|
[1344] | 997 | ENDIF |
---|
| 998 | ! |
---|
[311] | 999 | |
---|
[1344] | 1000 | ! north fold treatment |
---|
| 1001 | ! ----------------------- |
---|
| 1002 | IF( npolj /= 0 ) THEN |
---|
| 1003 | ! |
---|
| 1004 | SELECT CASE ( jpni ) |
---|
[3609] | 1005 | CASE ( 1 ) ; CALL lbc_nfd ( pt2d(1:jpi,1:jpj+jprj), cd_type, psgn, pr2dj=jprj ) |
---|
[1344] | 1006 | CASE DEFAULT ; CALL mpp_lbc_north_e( pt2d , cd_type, psgn ) |
---|
| 1007 | END SELECT |
---|
| 1008 | ! |
---|
[311] | 1009 | ENDIF |
---|
| 1010 | |
---|
[1344] | 1011 | ! 2. East and west directions exchange |
---|
| 1012 | ! ------------------------------------ |
---|
| 1013 | ! we play with the neigbours AND the row number because of the periodicity |
---|
| 1014 | ! |
---|
| 1015 | SELECT CASE ( nbondi ) ! Read Dirichlet lateral conditions |
---|
| 1016 | CASE ( -1, 0, 1 ) ! all exept 2 (i.e. close case) |
---|
[3609] | 1017 | iihom = nlci-nreci-jpri |
---|
[311] | 1018 | DO jl = 1, ipreci |
---|
[3609] | 1019 | r2dew(:,jl,1) = pt2d(jpreci+jl,:) |
---|
| 1020 | r2dwe(:,jl,1) = pt2d(iihom +jl,:) |
---|
[311] | 1021 | END DO |
---|
| 1022 | END SELECT |
---|
[1344] | 1023 | ! |
---|
| 1024 | ! ! Migrations |
---|
[3609] | 1025 | imigr = ipreci * ( jpj + 2*jprj) |
---|
[1344] | 1026 | ! |
---|
[311] | 1027 | SELECT CASE ( nbondi ) |
---|
| 1028 | CASE ( -1 ) |
---|
[3609] | 1029 | CALL mppsend( 2, r2dwe(1-jprj,1,1), imigr, noea, ml_req1 ) |
---|
| 1030 | CALL mpprecv( 1, r2dew(1-jprj,1,2), imigr, noea ) |
---|
[311] | 1031 | IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err) |
---|
| 1032 | CASE ( 0 ) |
---|
[3609] | 1033 | CALL mppsend( 1, r2dew(1-jprj,1,1), imigr, nowe, ml_req1 ) |
---|
| 1034 | CALL mppsend( 2, r2dwe(1-jprj,1,1), imigr, noea, ml_req2 ) |
---|
| 1035 | CALL mpprecv( 1, r2dew(1-jprj,1,2), imigr, noea ) |
---|
| 1036 | CALL mpprecv( 2, r2dwe(1-jprj,1,2), imigr, nowe ) |
---|
[311] | 1037 | IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err) |
---|
| 1038 | IF(l_isend) CALL mpi_wait(ml_req2,ml_stat,ml_err) |
---|
| 1039 | CASE ( 1 ) |
---|
[3609] | 1040 | CALL mppsend( 1, r2dew(1-jprj,1,1), imigr, nowe, ml_req1 ) |
---|
| 1041 | CALL mpprecv( 2, r2dwe(1-jprj,1,2), imigr, nowe ) |
---|
[311] | 1042 | IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err) |
---|
| 1043 | END SELECT |
---|
[1344] | 1044 | ! |
---|
| 1045 | ! ! Write Dirichlet lateral conditions |
---|
[311] | 1046 | iihom = nlci - jpreci |
---|
[1344] | 1047 | ! |
---|
[311] | 1048 | SELECT CASE ( nbondi ) |
---|
| 1049 | CASE ( -1 ) |
---|
| 1050 | DO jl = 1, ipreci |
---|
[3609] | 1051 | pt2d(iihom+jl,:) = r2dew(:,jl,2) |
---|
[311] | 1052 | END DO |
---|
| 1053 | CASE ( 0 ) |
---|
| 1054 | DO jl = 1, ipreci |
---|
[3609] | 1055 | pt2d(jl-jpri,:) = r2dwe(:,jl,2) |
---|
| 1056 | pt2d( iihom+jl,:) = r2dew(:,jl,2) |
---|
[311] | 1057 | END DO |
---|
| 1058 | CASE ( 1 ) |
---|
| 1059 | DO jl = 1, ipreci |
---|
[3609] | 1060 | pt2d(jl-jpri,:) = r2dwe(:,jl,2) |
---|
[311] | 1061 | END DO |
---|
| 1062 | END SELECT |
---|
| 1063 | |
---|
| 1064 | |
---|
| 1065 | ! 3. North and south directions |
---|
| 1066 | ! ----------------------------- |
---|
[1344] | 1067 | ! always closed : we play only with the neigbours |
---|
| 1068 | ! |
---|
| 1069 | IF( nbondj /= 2 ) THEN ! Read Dirichlet lateral conditions |
---|
[3609] | 1070 | ijhom = nlcj-nrecj-jprj |
---|
[311] | 1071 | DO jl = 1, iprecj |
---|
[3609] | 1072 | r2dsn(:,jl,1) = pt2d(:,ijhom +jl) |
---|
| 1073 | r2dns(:,jl,1) = pt2d(:,jprecj+jl) |
---|
[311] | 1074 | END DO |
---|
| 1075 | ENDIF |
---|
[1344] | 1076 | ! |
---|
| 1077 | ! ! Migrations |
---|
[3609] | 1078 | imigr = iprecj * ( jpi + 2*jpri ) |
---|
[1344] | 1079 | ! |
---|
[311] | 1080 | SELECT CASE ( nbondj ) |
---|
| 1081 | CASE ( -1 ) |
---|
[3609] | 1082 | CALL mppsend( 4, r2dsn(1-jpri,1,1), imigr, nono, ml_req1 ) |
---|
| 1083 | CALL mpprecv( 3, r2dns(1-jpri,1,2), imigr, nono ) |
---|
[311] | 1084 | IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err) |
---|
| 1085 | CASE ( 0 ) |
---|
[3609] | 1086 | CALL mppsend( 3, r2dns(1-jpri,1,1), imigr, noso, ml_req1 ) |
---|
| 1087 | CALL mppsend( 4, r2dsn(1-jpri,1,1), imigr, nono, ml_req2 ) |
---|
| 1088 | CALL mpprecv( 3, r2dns(1-jpri,1,2), imigr, nono ) |
---|
| 1089 | CALL mpprecv( 4, r2dsn(1-jpri,1,2), imigr, noso ) |
---|
[311] | 1090 | IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err) |
---|
| 1091 | IF(l_isend) CALL mpi_wait(ml_req2,ml_stat,ml_err) |
---|
| 1092 | CASE ( 1 ) |
---|
[3609] | 1093 | CALL mppsend( 3, r2dns(1-jpri,1,1), imigr, noso, ml_req1 ) |
---|
| 1094 | CALL mpprecv( 4, r2dsn(1-jpri,1,2), imigr, noso ) |
---|
[311] | 1095 | IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err) |
---|
| 1096 | END SELECT |
---|
[1344] | 1097 | ! |
---|
| 1098 | ! ! Write Dirichlet lateral conditions |
---|
[311] | 1099 | ijhom = nlcj - jprecj |
---|
[1344] | 1100 | ! |
---|
[311] | 1101 | SELECT CASE ( nbondj ) |
---|
| 1102 | CASE ( -1 ) |
---|
| 1103 | DO jl = 1, iprecj |
---|
[3609] | 1104 | pt2d(:,ijhom+jl) = r2dns(:,jl,2) |
---|
[311] | 1105 | END DO |
---|
| 1106 | CASE ( 0 ) |
---|
| 1107 | DO jl = 1, iprecj |
---|
[3609] | 1108 | pt2d(:,jl-jprj) = r2dsn(:,jl,2) |
---|
| 1109 | pt2d(:,ijhom+jl ) = r2dns(:,jl,2) |
---|
[311] | 1110 | END DO |
---|
| 1111 | CASE ( 1 ) |
---|
| 1112 | DO jl = 1, iprecj |
---|
[3609] | 1113 | pt2d(:,jl-jprj) = r2dsn(:,jl,2) |
---|
[311] | 1114 | END DO |
---|
[1344] | 1115 | END SELECT |
---|
[311] | 1116 | |
---|
| 1117 | END SUBROUTINE mpp_lnk_2d_e |
---|
| 1118 | |
---|
| 1119 | |
---|
[1344] | 1120 | SUBROUTINE mppsend( ktyp, pmess, kbytes, kdest, md_req ) |
---|
[51] | 1121 | !!---------------------------------------------------------------------- |
---|
| 1122 | !! *** routine mppsend *** |
---|
| 1123 | !! |
---|
| 1124 | !! ** Purpose : Send messag passing array |
---|
| 1125 | !! |
---|
| 1126 | !!---------------------------------------------------------------------- |
---|
[1344] | 1127 | REAL(wp), INTENT(inout) :: pmess(*) ! array of real |
---|
| 1128 | INTEGER , INTENT(in ) :: kbytes ! size of the array pmess |
---|
| 1129 | INTEGER , INTENT(in ) :: kdest ! receive process number |
---|
| 1130 | INTEGER , INTENT(in ) :: ktyp ! tag of the message |
---|
| 1131 | INTEGER , INTENT(in ) :: md_req ! argument for isend |
---|
| 1132 | !! |
---|
| 1133 | INTEGER :: iflag |
---|
[51] | 1134 | !!---------------------------------------------------------------------- |
---|
[1344] | 1135 | ! |
---|
[1601] | 1136 | SELECT CASE ( cn_mpi_send ) |
---|
[300] | 1137 | CASE ( 'S' ) ! Standard mpi send (blocking) |
---|
[1344] | 1138 | CALL mpi_send ( pmess, kbytes, mpi_double_precision, kdest , ktyp, mpi_comm_opa , iflag ) |
---|
[300] | 1139 | CASE ( 'B' ) ! Buffer mpi send (blocking) |
---|
[1344] | 1140 | CALL mpi_bsend( pmess, kbytes, mpi_double_precision, kdest , ktyp, mpi_comm_opa , iflag ) |
---|
[300] | 1141 | CASE ( 'I' ) ! Immediate mpi send (non-blocking send) |
---|
[1344] | 1142 | ! be carefull, one more argument here : the mpi request identifier.. |
---|
| 1143 | CALL mpi_isend( pmess, kbytes, mpi_double_precision, kdest , ktyp, mpi_comm_opa, md_req, iflag ) |
---|
[300] | 1144 | END SELECT |
---|
[1344] | 1145 | ! |
---|
[51] | 1146 | END SUBROUTINE mppsend |
---|
[3] | 1147 | |
---|
| 1148 | |
---|
[3294] | 1149 | SUBROUTINE mpprecv( ktyp, pmess, kbytes, ksource ) |
---|
[51] | 1150 | !!---------------------------------------------------------------------- |
---|
| 1151 | !! *** routine mpprecv *** |
---|
| 1152 | !! |
---|
| 1153 | !! ** Purpose : Receive messag passing array |
---|
| 1154 | !! |
---|
| 1155 | !!---------------------------------------------------------------------- |
---|
[1344] | 1156 | REAL(wp), INTENT(inout) :: pmess(*) ! array of real |
---|
| 1157 | INTEGER , INTENT(in ) :: kbytes ! suze of the array pmess |
---|
| 1158 | INTEGER , INTENT(in ) :: ktyp ! Tag of the recevied message |
---|
[3294] | 1159 | INTEGER, OPTIONAL, INTENT(in) :: ksource ! source process number |
---|
[1344] | 1160 | !! |
---|
[51] | 1161 | INTEGER :: istatus(mpi_status_size) |
---|
| 1162 | INTEGER :: iflag |
---|
[3294] | 1163 | INTEGER :: use_source |
---|
[1344] | 1164 | !!---------------------------------------------------------------------- |
---|
| 1165 | ! |
---|
[3294] | 1166 | |
---|
| 1167 | ! If a specific process number has been passed to the receive call, |
---|
| 1168 | ! use that one. Default is to use mpi_any_source |
---|
| 1169 | use_source=mpi_any_source |
---|
| 1170 | if(present(ksource)) then |
---|
| 1171 | use_source=ksource |
---|
| 1172 | end if |
---|
| 1173 | |
---|
| 1174 | CALL mpi_recv( pmess, kbytes, mpi_double_precision, use_source, ktyp, mpi_comm_opa, istatus, iflag ) |
---|
[1344] | 1175 | ! |
---|
[51] | 1176 | END SUBROUTINE mpprecv |
---|
[3] | 1177 | |
---|
| 1178 | |
---|
[51] | 1179 | SUBROUTINE mppgather( ptab, kp, pio ) |
---|
| 1180 | !!---------------------------------------------------------------------- |
---|
| 1181 | !! *** routine mppgather *** |
---|
| 1182 | !! |
---|
| 1183 | !! ** Purpose : Transfert between a local subdomain array and a work |
---|
| 1184 | !! array which is distributed following the vertical level. |
---|
| 1185 | !! |
---|
[1344] | 1186 | !!---------------------------------------------------------------------- |
---|
| 1187 | REAL(wp), DIMENSION(jpi,jpj), INTENT(in ) :: ptab ! subdomain input array |
---|
| 1188 | INTEGER , INTENT(in ) :: kp ! record length |
---|
| 1189 | REAL(wp), DIMENSION(jpi,jpj,jpnij), INTENT( out) :: pio ! subdomain input array |
---|
[51] | 1190 | !! |
---|
[1344] | 1191 | INTEGER :: itaille, ierror ! temporary integer |
---|
[51] | 1192 | !!--------------------------------------------------------------------- |
---|
[1344] | 1193 | ! |
---|
| 1194 | itaille = jpi * jpj |
---|
| 1195 | CALL mpi_gather( ptab, itaille, mpi_double_precision, pio, itaille , & |
---|
[532] | 1196 | & mpi_double_precision, kp , mpi_comm_opa, ierror ) |
---|
[1344] | 1197 | ! |
---|
[51] | 1198 | END SUBROUTINE mppgather |
---|
[3] | 1199 | |
---|
| 1200 | |
---|
[51] | 1201 | SUBROUTINE mppscatter( pio, kp, ptab ) |
---|
| 1202 | !!---------------------------------------------------------------------- |
---|
| 1203 | !! *** routine mppscatter *** |
---|
| 1204 | !! |
---|
| 1205 | !! ** Purpose : Transfert between awork array which is distributed |
---|
| 1206 | !! following the vertical level and the local subdomain array. |
---|
| 1207 | !! |
---|
| 1208 | !!---------------------------------------------------------------------- |
---|
| 1209 | REAL(wp), DIMENSION(jpi,jpj,jpnij) :: pio ! output array |
---|
| 1210 | INTEGER :: kp ! Tag (not used with MPI |
---|
| 1211 | REAL(wp), DIMENSION(jpi,jpj) :: ptab ! subdomain array input |
---|
[1344] | 1212 | !! |
---|
| 1213 | INTEGER :: itaille, ierror ! temporary integer |
---|
[51] | 1214 | !!--------------------------------------------------------------------- |
---|
[1344] | 1215 | ! |
---|
[51] | 1216 | itaille=jpi*jpj |
---|
[1344] | 1217 | ! |
---|
| 1218 | CALL mpi_scatter( pio, itaille, mpi_double_precision, ptab, itaille , & |
---|
| 1219 | & mpi_double_precision, kp , mpi_comm_opa, ierror ) |
---|
| 1220 | ! |
---|
[51] | 1221 | END SUBROUTINE mppscatter |
---|
[3] | 1222 | |
---|
| 1223 | |
---|
[869] | 1224 | SUBROUTINE mppmax_a_int( ktab, kdim, kcom ) |
---|
[681] | 1225 | !!---------------------------------------------------------------------- |
---|
| 1226 | !! *** routine mppmax_a_int *** |
---|
| 1227 | !! |
---|
| 1228 | !! ** Purpose : Find maximum value in an integer layout array |
---|
| 1229 | !! |
---|
| 1230 | !!---------------------------------------------------------------------- |
---|
[1344] | 1231 | INTEGER , INTENT(in ) :: kdim ! size of array |
---|
| 1232 | INTEGER , INTENT(inout), DIMENSION(kdim) :: ktab ! input array |
---|
| 1233 | INTEGER , INTENT(in ), OPTIONAL :: kcom ! |
---|
| 1234 | !! |
---|
| 1235 | INTEGER :: ierror, localcomm ! temporary integer |
---|
[681] | 1236 | INTEGER, DIMENSION(kdim) :: iwork |
---|
[1344] | 1237 | !!---------------------------------------------------------------------- |
---|
| 1238 | ! |
---|
[869] | 1239 | localcomm = mpi_comm_opa |
---|
[1344] | 1240 | IF( PRESENT(kcom) ) localcomm = kcom |
---|
| 1241 | ! |
---|
| 1242 | CALL mpi_allreduce( ktab, iwork, kdim, mpi_integer, mpi_max, localcomm, ierror ) |
---|
| 1243 | ! |
---|
[681] | 1244 | ktab(:) = iwork(:) |
---|
[1344] | 1245 | ! |
---|
[681] | 1246 | END SUBROUTINE mppmax_a_int |
---|
| 1247 | |
---|
| 1248 | |
---|
[869] | 1249 | SUBROUTINE mppmax_int( ktab, kcom ) |
---|
[681] | 1250 | !!---------------------------------------------------------------------- |
---|
| 1251 | !! *** routine mppmax_int *** |
---|
| 1252 | !! |
---|
[1344] | 1253 | !! ** Purpose : Find maximum value in an integer layout array |
---|
[681] | 1254 | !! |
---|
| 1255 | !!---------------------------------------------------------------------- |
---|
[1344] | 1256 | INTEGER, INTENT(inout) :: ktab ! ??? |
---|
| 1257 | INTEGER, INTENT(in ), OPTIONAL :: kcom ! ??? |
---|
| 1258 | !! |
---|
| 1259 | INTEGER :: ierror, iwork, localcomm ! temporary integer |
---|
| 1260 | !!---------------------------------------------------------------------- |
---|
| 1261 | ! |
---|
[869] | 1262 | localcomm = mpi_comm_opa |
---|
[1344] | 1263 | IF( PRESENT(kcom) ) localcomm = kcom |
---|
| 1264 | ! |
---|
| 1265 | CALL mpi_allreduce( ktab, iwork, 1, mpi_integer, mpi_max, localcomm, ierror) |
---|
| 1266 | ! |
---|
[681] | 1267 | ktab = iwork |
---|
[1344] | 1268 | ! |
---|
[681] | 1269 | END SUBROUTINE mppmax_int |
---|
| 1270 | |
---|
| 1271 | |
---|
[869] | 1272 | SUBROUTINE mppmin_a_int( ktab, kdim, kcom ) |
---|
[51] | 1273 | !!---------------------------------------------------------------------- |
---|
| 1274 | !! *** routine mppmin_a_int *** |
---|
| 1275 | !! |
---|
| 1276 | !! ** Purpose : Find minimum value in an integer layout array |
---|
| 1277 | !! |
---|
| 1278 | !!---------------------------------------------------------------------- |
---|
| 1279 | INTEGER , INTENT( in ) :: kdim ! size of array |
---|
| 1280 | INTEGER , INTENT(inout), DIMENSION(kdim) :: ktab ! input array |
---|
[888] | 1281 | INTEGER , INTENT( in ), OPTIONAL :: kcom ! input array |
---|
[1344] | 1282 | !! |
---|
| 1283 | INTEGER :: ierror, localcomm ! temporary integer |
---|
[51] | 1284 | INTEGER, DIMENSION(kdim) :: iwork |
---|
[1344] | 1285 | !!---------------------------------------------------------------------- |
---|
| 1286 | ! |
---|
[869] | 1287 | localcomm = mpi_comm_opa |
---|
[1344] | 1288 | IF( PRESENT(kcom) ) localcomm = kcom |
---|
| 1289 | ! |
---|
| 1290 | CALL mpi_allreduce( ktab, iwork, kdim, mpi_integer, mpi_min, localcomm, ierror ) |
---|
| 1291 | ! |
---|
[51] | 1292 | ktab(:) = iwork(:) |
---|
[1344] | 1293 | ! |
---|
[51] | 1294 | END SUBROUTINE mppmin_a_int |
---|
[3] | 1295 | |
---|
[13] | 1296 | |
---|
[1345] | 1297 | SUBROUTINE mppmin_int( ktab, kcom ) |
---|
[51] | 1298 | !!---------------------------------------------------------------------- |
---|
| 1299 | !! *** routine mppmin_int *** |
---|
| 1300 | !! |
---|
[1344] | 1301 | !! ** Purpose : Find minimum value in an integer layout array |
---|
[51] | 1302 | !! |
---|
| 1303 | !!---------------------------------------------------------------------- |
---|
| 1304 | INTEGER, INTENT(inout) :: ktab ! ??? |
---|
[1345] | 1305 | INTEGER , INTENT( in ), OPTIONAL :: kcom ! input array |
---|
[1344] | 1306 | !! |
---|
[1345] | 1307 | INTEGER :: ierror, iwork, localcomm |
---|
[1344] | 1308 | !!---------------------------------------------------------------------- |
---|
| 1309 | ! |
---|
[1345] | 1310 | localcomm = mpi_comm_opa |
---|
| 1311 | IF( PRESENT(kcom) ) localcomm = kcom |
---|
[1344] | 1312 | ! |
---|
[1345] | 1313 | CALL mpi_allreduce( ktab, iwork, 1, mpi_integer, mpi_min, localcomm, ierror ) |
---|
| 1314 | ! |
---|
[51] | 1315 | ktab = iwork |
---|
[1344] | 1316 | ! |
---|
[51] | 1317 | END SUBROUTINE mppmin_int |
---|
[3] | 1318 | |
---|
[13] | 1319 | |
---|
[51] | 1320 | SUBROUTINE mppsum_a_int( ktab, kdim ) |
---|
| 1321 | !!---------------------------------------------------------------------- |
---|
| 1322 | !! *** routine mppsum_a_int *** |
---|
| 1323 | !! |
---|
[1344] | 1324 | !! ** Purpose : Global integer sum, 1D array case |
---|
[51] | 1325 | !! |
---|
| 1326 | !!---------------------------------------------------------------------- |
---|
[1344] | 1327 | INTEGER, INTENT(in ) :: kdim ! ??? |
---|
[51] | 1328 | INTEGER, INTENT(inout), DIMENSION (kdim) :: ktab ! ??? |
---|
[1344] | 1329 | !! |
---|
[51] | 1330 | INTEGER :: ierror |
---|
| 1331 | INTEGER, DIMENSION (kdim) :: iwork |
---|
[1344] | 1332 | !!---------------------------------------------------------------------- |
---|
| 1333 | ! |
---|
| 1334 | CALL mpi_allreduce( ktab, iwork, kdim, mpi_integer, mpi_sum, mpi_comm_opa, ierror ) |
---|
| 1335 | ! |
---|
[51] | 1336 | ktab(:) = iwork(:) |
---|
[1344] | 1337 | ! |
---|
[51] | 1338 | END SUBROUTINE mppsum_a_int |
---|
[3] | 1339 | |
---|
[13] | 1340 | |
---|
[1344] | 1341 | SUBROUTINE mppsum_int( ktab ) |
---|
| 1342 | !!---------------------------------------------------------------------- |
---|
| 1343 | !! *** routine mppsum_int *** |
---|
| 1344 | !! |
---|
| 1345 | !! ** Purpose : Global integer sum |
---|
| 1346 | !! |
---|
| 1347 | !!---------------------------------------------------------------------- |
---|
| 1348 | INTEGER, INTENT(inout) :: ktab |
---|
| 1349 | !! |
---|
| 1350 | INTEGER :: ierror, iwork |
---|
| 1351 | !!---------------------------------------------------------------------- |
---|
| 1352 | ! |
---|
| 1353 | CALL mpi_allreduce( ktab, iwork, 1, mpi_integer, mpi_sum, mpi_comm_opa, ierror ) |
---|
| 1354 | ! |
---|
| 1355 | ktab = iwork |
---|
| 1356 | ! |
---|
| 1357 | END SUBROUTINE mppsum_int |
---|
[3] | 1358 | |
---|
[13] | 1359 | |
---|
[1344] | 1360 | SUBROUTINE mppmax_a_real( ptab, kdim, kcom ) |
---|
| 1361 | !!---------------------------------------------------------------------- |
---|
| 1362 | !! *** routine mppmax_a_real *** |
---|
| 1363 | !! |
---|
| 1364 | !! ** Purpose : Maximum |
---|
| 1365 | !! |
---|
| 1366 | !!---------------------------------------------------------------------- |
---|
| 1367 | INTEGER , INTENT(in ) :: kdim |
---|
| 1368 | REAL(wp), INTENT(inout), DIMENSION(kdim) :: ptab |
---|
| 1369 | INTEGER , INTENT(in ), OPTIONAL :: kcom |
---|
| 1370 | !! |
---|
| 1371 | INTEGER :: ierror, localcomm |
---|
| 1372 | REAL(wp), DIMENSION(kdim) :: zwork |
---|
| 1373 | !!---------------------------------------------------------------------- |
---|
| 1374 | ! |
---|
| 1375 | localcomm = mpi_comm_opa |
---|
| 1376 | IF( PRESENT(kcom) ) localcomm = kcom |
---|
| 1377 | ! |
---|
| 1378 | CALL mpi_allreduce( ptab, zwork, kdim, mpi_double_precision, mpi_max, localcomm, ierror ) |
---|
| 1379 | ptab(:) = zwork(:) |
---|
| 1380 | ! |
---|
| 1381 | END SUBROUTINE mppmax_a_real |
---|
[3] | 1382 | |
---|
| 1383 | |
---|
[1344] | 1384 | SUBROUTINE mppmax_real( ptab, kcom ) |
---|
| 1385 | !!---------------------------------------------------------------------- |
---|
| 1386 | !! *** routine mppmax_real *** |
---|
| 1387 | !! |
---|
| 1388 | !! ** Purpose : Maximum |
---|
| 1389 | !! |
---|
| 1390 | !!---------------------------------------------------------------------- |
---|
| 1391 | REAL(wp), INTENT(inout) :: ptab ! ??? |
---|
| 1392 | INTEGER , INTENT(in ), OPTIONAL :: kcom ! ??? |
---|
| 1393 | !! |
---|
| 1394 | INTEGER :: ierror, localcomm |
---|
| 1395 | REAL(wp) :: zwork |
---|
| 1396 | !!---------------------------------------------------------------------- |
---|
| 1397 | ! |
---|
| 1398 | localcomm = mpi_comm_opa |
---|
| 1399 | IF( PRESENT(kcom) ) localcomm = kcom |
---|
| 1400 | ! |
---|
| 1401 | CALL mpi_allreduce( ptab, zwork, 1, mpi_double_precision, mpi_max, localcomm, ierror ) |
---|
| 1402 | ptab = zwork |
---|
| 1403 | ! |
---|
| 1404 | END SUBROUTINE mppmax_real |
---|
[13] | 1405 | |
---|
[3] | 1406 | |
---|
[1344] | 1407 | SUBROUTINE mppmin_a_real( ptab, kdim, kcom ) |
---|
| 1408 | !!---------------------------------------------------------------------- |
---|
| 1409 | !! *** routine mppmin_a_real *** |
---|
| 1410 | !! |
---|
| 1411 | !! ** Purpose : Minimum of REAL, array case |
---|
| 1412 | !! |
---|
| 1413 | !!----------------------------------------------------------------------- |
---|
| 1414 | INTEGER , INTENT(in ) :: kdim |
---|
| 1415 | REAL(wp), INTENT(inout), DIMENSION(kdim) :: ptab |
---|
| 1416 | INTEGER , INTENT(in ), OPTIONAL :: kcom |
---|
| 1417 | !! |
---|
| 1418 | INTEGER :: ierror, localcomm |
---|
| 1419 | REAL(wp), DIMENSION(kdim) :: zwork |
---|
| 1420 | !!----------------------------------------------------------------------- |
---|
| 1421 | ! |
---|
| 1422 | localcomm = mpi_comm_opa |
---|
| 1423 | IF( PRESENT(kcom) ) localcomm = kcom |
---|
| 1424 | ! |
---|
| 1425 | CALL mpi_allreduce( ptab, zwork, kdim, mpi_double_precision, mpi_min, localcomm, ierror ) |
---|
| 1426 | ptab(:) = zwork(:) |
---|
| 1427 | ! |
---|
| 1428 | END SUBROUTINE mppmin_a_real |
---|
[3] | 1429 | |
---|
| 1430 | |
---|
[1344] | 1431 | SUBROUTINE mppmin_real( ptab, kcom ) |
---|
| 1432 | !!---------------------------------------------------------------------- |
---|
| 1433 | !! *** routine mppmin_real *** |
---|
| 1434 | !! |
---|
| 1435 | !! ** Purpose : minimum of REAL, scalar case |
---|
| 1436 | !! |
---|
| 1437 | !!----------------------------------------------------------------------- |
---|
| 1438 | REAL(wp), INTENT(inout) :: ptab ! |
---|
| 1439 | INTEGER , INTENT(in ), OPTIONAL :: kcom |
---|
| 1440 | !! |
---|
| 1441 | INTEGER :: ierror |
---|
| 1442 | REAL(wp) :: zwork |
---|
| 1443 | INTEGER :: localcomm |
---|
| 1444 | !!----------------------------------------------------------------------- |
---|
| 1445 | ! |
---|
| 1446 | localcomm = mpi_comm_opa |
---|
| 1447 | IF( PRESENT(kcom) ) localcomm = kcom |
---|
| 1448 | ! |
---|
| 1449 | CALL mpi_allreduce( ptab, zwork, 1, mpi_double_precision, mpi_min, localcomm, ierror ) |
---|
| 1450 | ptab = zwork |
---|
| 1451 | ! |
---|
| 1452 | END SUBROUTINE mppmin_real |
---|
[13] | 1453 | |
---|
[3] | 1454 | |
---|
[1344] | 1455 | SUBROUTINE mppsum_a_real( ptab, kdim, kcom ) |
---|
| 1456 | !!---------------------------------------------------------------------- |
---|
| 1457 | !! *** routine mppsum_a_real *** |
---|
| 1458 | !! |
---|
| 1459 | !! ** Purpose : global sum, REAL ARRAY argument case |
---|
| 1460 | !! |
---|
| 1461 | !!----------------------------------------------------------------------- |
---|
| 1462 | INTEGER , INTENT( in ) :: kdim ! size of ptab |
---|
| 1463 | REAL(wp), DIMENSION(kdim), INTENT( inout ) :: ptab ! input array |
---|
| 1464 | INTEGER , INTENT( in ), OPTIONAL :: kcom |
---|
| 1465 | !! |
---|
| 1466 | INTEGER :: ierror ! temporary integer |
---|
| 1467 | INTEGER :: localcomm |
---|
| 1468 | REAL(wp), DIMENSION(kdim) :: zwork ! temporary workspace |
---|
| 1469 | !!----------------------------------------------------------------------- |
---|
| 1470 | ! |
---|
| 1471 | localcomm = mpi_comm_opa |
---|
| 1472 | IF( PRESENT(kcom) ) localcomm = kcom |
---|
| 1473 | ! |
---|
| 1474 | CALL mpi_allreduce( ptab, zwork, kdim, mpi_double_precision, mpi_sum, localcomm, ierror ) |
---|
| 1475 | ptab(:) = zwork(:) |
---|
| 1476 | ! |
---|
| 1477 | END SUBROUTINE mppsum_a_real |
---|
[869] | 1478 | |
---|
[3] | 1479 | |
---|
[1344] | 1480 | SUBROUTINE mppsum_real( ptab, kcom ) |
---|
| 1481 | !!---------------------------------------------------------------------- |
---|
| 1482 | !! *** routine mppsum_real *** |
---|
| 1483 | !! |
---|
| 1484 | !! ** Purpose : global sum, SCALAR argument case |
---|
| 1485 | !! |
---|
| 1486 | !!----------------------------------------------------------------------- |
---|
| 1487 | REAL(wp), INTENT(inout) :: ptab ! input scalar |
---|
| 1488 | INTEGER , INTENT(in ), OPTIONAL :: kcom |
---|
| 1489 | !! |
---|
| 1490 | INTEGER :: ierror, localcomm |
---|
| 1491 | REAL(wp) :: zwork |
---|
| 1492 | !!----------------------------------------------------------------------- |
---|
| 1493 | ! |
---|
| 1494 | localcomm = mpi_comm_opa |
---|
| 1495 | IF( PRESENT(kcom) ) localcomm = kcom |
---|
| 1496 | ! |
---|
| 1497 | CALL mpi_allreduce( ptab, zwork, 1, mpi_double_precision, mpi_sum, localcomm, ierror ) |
---|
| 1498 | ptab = zwork |
---|
| 1499 | ! |
---|
| 1500 | END SUBROUTINE mppsum_real |
---|
[3] | 1501 | |
---|
[1976] | 1502 | SUBROUTINE mppsum_realdd( ytab, kcom ) |
---|
| 1503 | !!---------------------------------------------------------------------- |
---|
| 1504 | !! *** routine mppsum_realdd *** |
---|
| 1505 | !! |
---|
| 1506 | !! ** Purpose : global sum in Massively Parallel Processing |
---|
| 1507 | !! SCALAR argument case for double-double precision |
---|
| 1508 | !! |
---|
| 1509 | !!----------------------------------------------------------------------- |
---|
| 1510 | COMPLEX(wp), INTENT(inout) :: ytab ! input scalar |
---|
| 1511 | INTEGER , INTENT( in ), OPTIONAL :: kcom |
---|
[3] | 1512 | |
---|
[1976] | 1513 | !! * Local variables (MPI version) |
---|
| 1514 | INTEGER :: ierror |
---|
| 1515 | INTEGER :: localcomm |
---|
| 1516 | COMPLEX(wp) :: zwork |
---|
| 1517 | |
---|
| 1518 | localcomm = mpi_comm_opa |
---|
| 1519 | IF( PRESENT(kcom) ) localcomm = kcom |
---|
| 1520 | |
---|
| 1521 | ! reduce local sums into global sum |
---|
| 1522 | CALL MPI_ALLREDUCE (ytab, zwork, 1, MPI_DOUBLE_COMPLEX, & |
---|
| 1523 | MPI_SUMDD,localcomm,ierror) |
---|
| 1524 | ytab = zwork |
---|
| 1525 | |
---|
| 1526 | END SUBROUTINE mppsum_realdd |
---|
| 1527 | |
---|
| 1528 | |
---|
| 1529 | SUBROUTINE mppsum_a_realdd( ytab, kdim, kcom ) |
---|
| 1530 | !!---------------------------------------------------------------------- |
---|
| 1531 | !! *** routine mppsum_a_realdd *** |
---|
| 1532 | !! |
---|
| 1533 | !! ** Purpose : global sum in Massively Parallel Processing |
---|
| 1534 | !! COMPLEX ARRAY case for double-double precision |
---|
| 1535 | !! |
---|
| 1536 | !!----------------------------------------------------------------------- |
---|
| 1537 | INTEGER , INTENT( in ) :: kdim ! size of ytab |
---|
| 1538 | COMPLEX(wp), DIMENSION(kdim), INTENT( inout ) :: ytab ! input array |
---|
| 1539 | INTEGER , INTENT( in ), OPTIONAL :: kcom |
---|
| 1540 | |
---|
| 1541 | !! * Local variables (MPI version) |
---|
| 1542 | INTEGER :: ierror ! temporary integer |
---|
| 1543 | INTEGER :: localcomm |
---|
| 1544 | COMPLEX(wp), DIMENSION(kdim) :: zwork ! temporary workspace |
---|
| 1545 | |
---|
| 1546 | localcomm = mpi_comm_opa |
---|
| 1547 | IF( PRESENT(kcom) ) localcomm = kcom |
---|
| 1548 | |
---|
| 1549 | CALL MPI_ALLREDUCE (ytab, zwork, kdim, MPI_DOUBLE_COMPLEX, & |
---|
| 1550 | MPI_SUMDD,localcomm,ierror) |
---|
| 1551 | ytab(:) = zwork(:) |
---|
| 1552 | |
---|
| 1553 | END SUBROUTINE mppsum_a_realdd |
---|
| 1554 | |
---|
[1344] | 1555 | SUBROUTINE mpp_minloc2d( ptab, pmask, pmin, ki,kj ) |
---|
| 1556 | !!------------------------------------------------------------------------ |
---|
| 1557 | !! *** routine mpp_minloc *** |
---|
| 1558 | !! |
---|
| 1559 | !! ** Purpose : Compute the global minimum of an array ptab |
---|
| 1560 | !! and also give its global position |
---|
| 1561 | !! |
---|
| 1562 | !! ** Method : Use MPI_ALLREDUCE with MPI_MINLOC |
---|
| 1563 | !! |
---|
| 1564 | !!-------------------------------------------------------------------------- |
---|
| 1565 | REAL(wp), DIMENSION (jpi,jpj), INTENT(in ) :: ptab ! Local 2D array |
---|
| 1566 | REAL(wp), DIMENSION (jpi,jpj), INTENT(in ) :: pmask ! Local mask |
---|
| 1567 | REAL(wp) , INTENT( out) :: pmin ! Global minimum of ptab |
---|
| 1568 | INTEGER , INTENT( out) :: ki, kj ! index of minimum in global frame |
---|
| 1569 | !! |
---|
| 1570 | INTEGER , DIMENSION(2) :: ilocs |
---|
| 1571 | INTEGER :: ierror |
---|
| 1572 | REAL(wp) :: zmin ! local minimum |
---|
| 1573 | REAL(wp), DIMENSION(2,1) :: zain, zaout |
---|
| 1574 | !!----------------------------------------------------------------------- |
---|
| 1575 | ! |
---|
| 1576 | zmin = MINVAL( ptab(:,:) , mask= pmask == 1.e0 ) |
---|
| 1577 | ilocs = MINLOC( ptab(:,:) , mask= pmask == 1.e0 ) |
---|
| 1578 | ! |
---|
| 1579 | ki = ilocs(1) + nimpp - 1 |
---|
| 1580 | kj = ilocs(2) + njmpp - 1 |
---|
| 1581 | ! |
---|
| 1582 | zain(1,:)=zmin |
---|
| 1583 | zain(2,:)=ki+10000.*kj |
---|
| 1584 | ! |
---|
| 1585 | CALL MPI_ALLREDUCE( zain,zaout, 1, MPI_2DOUBLE_PRECISION,MPI_MINLOC,MPI_COMM_OPA,ierror) |
---|
| 1586 | ! |
---|
| 1587 | pmin = zaout(1,1) |
---|
| 1588 | kj = INT(zaout(2,1)/10000.) |
---|
| 1589 | ki = INT(zaout(2,1) - 10000.*kj ) |
---|
| 1590 | ! |
---|
| 1591 | END SUBROUTINE mpp_minloc2d |
---|
[13] | 1592 | |
---|
[3] | 1593 | |
---|
[1344] | 1594 | SUBROUTINE mpp_minloc3d( ptab, pmask, pmin, ki, kj ,kk) |
---|
| 1595 | !!------------------------------------------------------------------------ |
---|
| 1596 | !! *** routine mpp_minloc *** |
---|
| 1597 | !! |
---|
| 1598 | !! ** Purpose : Compute the global minimum of an array ptab |
---|
| 1599 | !! and also give its global position |
---|
| 1600 | !! |
---|
| 1601 | !! ** Method : Use MPI_ALLREDUCE with MPI_MINLOC |
---|
| 1602 | !! |
---|
| 1603 | !!-------------------------------------------------------------------------- |
---|
| 1604 | REAL(wp), DIMENSION (jpi,jpj,jpk), INTENT(in ) :: ptab ! Local 2D array |
---|
| 1605 | REAL(wp), DIMENSION (jpi,jpj,jpk), INTENT(in ) :: pmask ! Local mask |
---|
| 1606 | REAL(wp) , INTENT( out) :: pmin ! Global minimum of ptab |
---|
| 1607 | INTEGER , INTENT( out) :: ki, kj, kk ! index of minimum in global frame |
---|
| 1608 | !! |
---|
| 1609 | INTEGER :: ierror |
---|
| 1610 | REAL(wp) :: zmin ! local minimum |
---|
| 1611 | INTEGER , DIMENSION(3) :: ilocs |
---|
| 1612 | REAL(wp), DIMENSION(2,1) :: zain, zaout |
---|
| 1613 | !!----------------------------------------------------------------------- |
---|
| 1614 | ! |
---|
| 1615 | zmin = MINVAL( ptab(:,:,:) , mask= pmask == 1.e0 ) |
---|
| 1616 | ilocs = MINLOC( ptab(:,:,:) , mask= pmask == 1.e0 ) |
---|
| 1617 | ! |
---|
| 1618 | ki = ilocs(1) + nimpp - 1 |
---|
| 1619 | kj = ilocs(2) + njmpp - 1 |
---|
| 1620 | kk = ilocs(3) |
---|
| 1621 | ! |
---|
| 1622 | zain(1,:)=zmin |
---|
| 1623 | zain(2,:)=ki+10000.*kj+100000000.*kk |
---|
| 1624 | ! |
---|
| 1625 | CALL MPI_ALLREDUCE( zain,zaout, 1, MPI_2DOUBLE_PRECISION,MPI_MINLOC,MPI_COMM_OPA,ierror) |
---|
| 1626 | ! |
---|
| 1627 | pmin = zaout(1,1) |
---|
| 1628 | kk = INT( zaout(2,1) / 100000000. ) |
---|
| 1629 | kj = INT( zaout(2,1) - kk * 100000000. ) / 10000 |
---|
| 1630 | ki = INT( zaout(2,1) - kk * 100000000. -kj * 10000. ) |
---|
| 1631 | ! |
---|
| 1632 | END SUBROUTINE mpp_minloc3d |
---|
[13] | 1633 | |
---|
[3] | 1634 | |
---|
[1344] | 1635 | SUBROUTINE mpp_maxloc2d( ptab, pmask, pmax, ki, kj ) |
---|
| 1636 | !!------------------------------------------------------------------------ |
---|
| 1637 | !! *** routine mpp_maxloc *** |
---|
| 1638 | !! |
---|
| 1639 | !! ** Purpose : Compute the global maximum of an array ptab |
---|
| 1640 | !! and also give its global position |
---|
| 1641 | !! |
---|
| 1642 | !! ** Method : Use MPI_ALLREDUCE with MPI_MINLOC |
---|
| 1643 | !! |
---|
| 1644 | !!-------------------------------------------------------------------------- |
---|
| 1645 | REAL(wp), DIMENSION (jpi,jpj), INTENT(in ) :: ptab ! Local 2D array |
---|
| 1646 | REAL(wp), DIMENSION (jpi,jpj), INTENT(in ) :: pmask ! Local mask |
---|
| 1647 | REAL(wp) , INTENT( out) :: pmax ! Global maximum of ptab |
---|
| 1648 | INTEGER , INTENT( out) :: ki, kj ! index of maximum in global frame |
---|
| 1649 | !! |
---|
| 1650 | INTEGER :: ierror |
---|
| 1651 | INTEGER, DIMENSION (2) :: ilocs |
---|
| 1652 | REAL(wp) :: zmax ! local maximum |
---|
| 1653 | REAL(wp), DIMENSION(2,1) :: zain, zaout |
---|
| 1654 | !!----------------------------------------------------------------------- |
---|
| 1655 | ! |
---|
| 1656 | zmax = MAXVAL( ptab(:,:) , mask= pmask == 1.e0 ) |
---|
| 1657 | ilocs = MAXLOC( ptab(:,:) , mask= pmask == 1.e0 ) |
---|
| 1658 | ! |
---|
| 1659 | ki = ilocs(1) + nimpp - 1 |
---|
| 1660 | kj = ilocs(2) + njmpp - 1 |
---|
| 1661 | ! |
---|
| 1662 | zain(1,:) = zmax |
---|
| 1663 | zain(2,:) = ki + 10000. * kj |
---|
| 1664 | ! |
---|
| 1665 | CALL MPI_ALLREDUCE( zain,zaout, 1, MPI_2DOUBLE_PRECISION,MPI_MAXLOC,MPI_COMM_OPA,ierror) |
---|
| 1666 | ! |
---|
| 1667 | pmax = zaout(1,1) |
---|
| 1668 | kj = INT( zaout(2,1) / 10000. ) |
---|
| 1669 | ki = INT( zaout(2,1) - 10000.* kj ) |
---|
| 1670 | ! |
---|
| 1671 | END SUBROUTINE mpp_maxloc2d |
---|
[3] | 1672 | |
---|
[13] | 1673 | |
---|
[1344] | 1674 | SUBROUTINE mpp_maxloc3d( ptab, pmask, pmax, ki, kj, kk ) |
---|
| 1675 | !!------------------------------------------------------------------------ |
---|
| 1676 | !! *** routine mpp_maxloc *** |
---|
| 1677 | !! |
---|
| 1678 | !! ** Purpose : Compute the global maximum of an array ptab |
---|
| 1679 | !! and also give its global position |
---|
| 1680 | !! |
---|
| 1681 | !! ** Method : Use MPI_ALLREDUCE with MPI_MINLOC |
---|
| 1682 | !! |
---|
| 1683 | !!-------------------------------------------------------------------------- |
---|
| 1684 | REAL(wp), DIMENSION (jpi,jpj,jpk), INTENT(in ) :: ptab ! Local 2D array |
---|
| 1685 | REAL(wp), DIMENSION (jpi,jpj,jpk), INTENT(in ) :: pmask ! Local mask |
---|
| 1686 | REAL(wp) , INTENT( out) :: pmax ! Global maximum of ptab |
---|
| 1687 | INTEGER , INTENT( out) :: ki, kj, kk ! index of maximum in global frame |
---|
| 1688 | !! |
---|
| 1689 | REAL(wp) :: zmax ! local maximum |
---|
| 1690 | REAL(wp), DIMENSION(2,1) :: zain, zaout |
---|
| 1691 | INTEGER , DIMENSION(3) :: ilocs |
---|
| 1692 | INTEGER :: ierror |
---|
| 1693 | !!----------------------------------------------------------------------- |
---|
| 1694 | ! |
---|
| 1695 | zmax = MAXVAL( ptab(:,:,:) , mask= pmask == 1.e0 ) |
---|
| 1696 | ilocs = MAXLOC( ptab(:,:,:) , mask= pmask == 1.e0 ) |
---|
| 1697 | ! |
---|
| 1698 | ki = ilocs(1) + nimpp - 1 |
---|
| 1699 | kj = ilocs(2) + njmpp - 1 |
---|
| 1700 | kk = ilocs(3) |
---|
| 1701 | ! |
---|
| 1702 | zain(1,:)=zmax |
---|
| 1703 | zain(2,:)=ki+10000.*kj+100000000.*kk |
---|
| 1704 | ! |
---|
| 1705 | CALL MPI_ALLREDUCE( zain,zaout, 1, MPI_2DOUBLE_PRECISION,MPI_MAXLOC,MPI_COMM_OPA,ierror) |
---|
| 1706 | ! |
---|
| 1707 | pmax = zaout(1,1) |
---|
| 1708 | kk = INT( zaout(2,1) / 100000000. ) |
---|
| 1709 | kj = INT( zaout(2,1) - kk * 100000000. ) / 10000 |
---|
| 1710 | ki = INT( zaout(2,1) - kk * 100000000. -kj * 10000. ) |
---|
| 1711 | ! |
---|
| 1712 | END SUBROUTINE mpp_maxloc3d |
---|
[3] | 1713 | |
---|
[869] | 1714 | |
---|
[1344] | 1715 | SUBROUTINE mppsync() |
---|
| 1716 | !!---------------------------------------------------------------------- |
---|
| 1717 | !! *** routine mppsync *** |
---|
| 1718 | !! |
---|
| 1719 | !! ** Purpose : Massively parallel processors, synchroneous |
---|
| 1720 | !! |
---|
| 1721 | !!----------------------------------------------------------------------- |
---|
| 1722 | INTEGER :: ierror |
---|
| 1723 | !!----------------------------------------------------------------------- |
---|
| 1724 | ! |
---|
| 1725 | CALL mpi_barrier( mpi_comm_opa, ierror ) |
---|
| 1726 | ! |
---|
| 1727 | END SUBROUTINE mppsync |
---|
[3] | 1728 | |
---|
| 1729 | |
---|
[1344] | 1730 | SUBROUTINE mppstop |
---|
| 1731 | !!---------------------------------------------------------------------- |
---|
| 1732 | !! *** routine mppstop *** |
---|
| 1733 | !! |
---|
[3294] | 1734 | !! ** purpose : Stop massively parallel processors method |
---|
[1344] | 1735 | !! |
---|
| 1736 | !!---------------------------------------------------------------------- |
---|
| 1737 | INTEGER :: info |
---|
| 1738 | !!---------------------------------------------------------------------- |
---|
| 1739 | ! |
---|
| 1740 | CALL mppsync |
---|
| 1741 | CALL mpi_finalize( info ) |
---|
| 1742 | ! |
---|
| 1743 | END SUBROUTINE mppstop |
---|
[3] | 1744 | |
---|
| 1745 | |
---|
[2715] | 1746 | SUBROUTINE mppobc( ptab, kd1, kd2, kl, kk, ktype, kij , kumout) |
---|
[1344] | 1747 | !!---------------------------------------------------------------------- |
---|
| 1748 | !! *** routine mppobc *** |
---|
| 1749 | !! |
---|
| 1750 | !! ** Purpose : Message passing manadgement for open boundary |
---|
| 1751 | !! conditions array |
---|
| 1752 | !! |
---|
| 1753 | !! ** Method : Use mppsend and mpprecv function for passing mask |
---|
| 1754 | !! between processors following neighboring subdomains. |
---|
| 1755 | !! domain parameters |
---|
| 1756 | !! nlci : first dimension of the local subdomain |
---|
| 1757 | !! nlcj : second dimension of the local subdomain |
---|
| 1758 | !! nbondi : mark for "east-west local boundary" |
---|
| 1759 | !! nbondj : mark for "north-south local boundary" |
---|
| 1760 | !! noea : number for local neighboring processors |
---|
| 1761 | !! nowe : number for local neighboring processors |
---|
| 1762 | !! noso : number for local neighboring processors |
---|
| 1763 | !! nono : number for local neighboring processors |
---|
| 1764 | !! |
---|
| 1765 | !!---------------------------------------------------------------------- |
---|
[3294] | 1766 | USE wrk_nemo ! Memory allocation |
---|
[2715] | 1767 | ! |
---|
[1344] | 1768 | INTEGER , INTENT(in ) :: kd1, kd2 ! starting and ending indices |
---|
| 1769 | INTEGER , INTENT(in ) :: kl ! index of open boundary |
---|
| 1770 | INTEGER , INTENT(in ) :: kk ! vertical dimension |
---|
| 1771 | INTEGER , INTENT(in ) :: ktype ! define north/south or east/west cdt |
---|
| 1772 | ! ! = 1 north/south ; = 2 east/west |
---|
| 1773 | INTEGER , INTENT(in ) :: kij ! horizontal dimension |
---|
[2715] | 1774 | INTEGER , INTENT(in ) :: kumout ! ocean.output logical unit |
---|
[1344] | 1775 | REAL(wp), INTENT(inout), DIMENSION(kij,kk) :: ptab ! variable array |
---|
[2715] | 1776 | ! |
---|
| 1777 | INTEGER :: ji, jj, jk, jl ! dummy loop indices |
---|
| 1778 | INTEGER :: iipt0, iipt1, ilpt1 ! local integers |
---|
| 1779 | INTEGER :: ijpt0, ijpt1 ! - - |
---|
| 1780 | INTEGER :: imigr, iihom, ijhom ! - - |
---|
[1344] | 1781 | INTEGER :: ml_req1, ml_req2, ml_err ! for key_mpi_isend |
---|
| 1782 | INTEGER :: ml_stat(MPI_STATUS_SIZE) ! for key_mpi_isend |
---|
[3294] | 1783 | REAL(wp), POINTER, DIMENSION(:,:) :: ztab ! temporary workspace |
---|
[1344] | 1784 | !!---------------------------------------------------------------------- |
---|
[3] | 1785 | |
---|
[3294] | 1786 | CALL wrk_alloc( jpi,jpj, ztab ) |
---|
[2715] | 1787 | |
---|
[1344] | 1788 | ! boundary condition initialization |
---|
| 1789 | ! --------------------------------- |
---|
| 1790 | ztab(:,:) = 0.e0 |
---|
| 1791 | ! |
---|
| 1792 | IF( ktype==1 ) THEN ! north/south boundaries |
---|
| 1793 | iipt0 = MAX( 1, MIN(kd1 - nimpp+1, nlci ) ) |
---|
| 1794 | iipt1 = MAX( 0, MIN(kd2 - nimpp+1, nlci - 1 ) ) |
---|
| 1795 | ilpt1 = MAX( 1, MIN(kd2 - nimpp+1, nlci ) ) |
---|
| 1796 | ijpt0 = MAX( 1, MIN(kl - njmpp+1, nlcj ) ) |
---|
| 1797 | ijpt1 = MAX( 0, MIN(kl - njmpp+1, nlcj - 1 ) ) |
---|
| 1798 | ELSEIF( ktype==2 ) THEN ! east/west boundaries |
---|
| 1799 | iipt0 = MAX( 1, MIN(kl - nimpp+1, nlci ) ) |
---|
| 1800 | iipt1 = MAX( 0, MIN(kl - nimpp+1, nlci - 1 ) ) |
---|
| 1801 | ijpt0 = MAX( 1, MIN(kd1 - njmpp+1, nlcj ) ) |
---|
| 1802 | ijpt1 = MAX( 0, MIN(kd2 - njmpp+1, nlcj - 1 ) ) |
---|
| 1803 | ilpt1 = MAX( 1, MIN(kd2 - njmpp+1, nlcj ) ) |
---|
| 1804 | ELSE |
---|
[2715] | 1805 | WRITE(kumout, cform_err) |
---|
| 1806 | WRITE(kumout,*) 'mppobc : bad ktype' |
---|
| 1807 | CALL mppstop |
---|
[1344] | 1808 | ENDIF |
---|
| 1809 | |
---|
| 1810 | ! Communication level by level |
---|
| 1811 | ! ---------------------------- |
---|
| 1812 | !!gm Remark : this is very time consumming!!! |
---|
| 1813 | ! ! ------------------------ ! |
---|
| 1814 | DO jk = 1, kk ! Loop over the levels ! |
---|
| 1815 | ! ! ------------------------ ! |
---|
| 1816 | ! |
---|
| 1817 | IF( ktype == 1 ) THEN ! north/south boundaries |
---|
| 1818 | DO jj = ijpt0, ijpt1 |
---|
| 1819 | DO ji = iipt0, iipt1 |
---|
| 1820 | ztab(ji,jj) = ptab(ji,jk) |
---|
| 1821 | END DO |
---|
| 1822 | END DO |
---|
| 1823 | ELSEIF( ktype == 2 ) THEN ! east/west boundaries |
---|
| 1824 | DO jj = ijpt0, ijpt1 |
---|
| 1825 | DO ji = iipt0, iipt1 |
---|
| 1826 | ztab(ji,jj) = ptab(jj,jk) |
---|
| 1827 | END DO |
---|
| 1828 | END DO |
---|
| 1829 | ENDIF |
---|
[13] | 1830 | |
---|
[3] | 1831 | |
---|
[1344] | 1832 | ! 1. East and west directions |
---|
| 1833 | ! --------------------------- |
---|
| 1834 | ! |
---|
| 1835 | IF( nbondi /= 2 ) THEN ! Read Dirichlet lateral conditions |
---|
| 1836 | iihom = nlci-nreci |
---|
| 1837 | DO jl = 1, jpreci |
---|
| 1838 | t2ew(:,jl,1) = ztab(jpreci+jl,:) |
---|
| 1839 | t2we(:,jl,1) = ztab(iihom +jl,:) |
---|
| 1840 | END DO |
---|
| 1841 | ENDIF |
---|
| 1842 | ! |
---|
| 1843 | ! ! Migrations |
---|
| 1844 | imigr=jpreci*jpj |
---|
| 1845 | ! |
---|
| 1846 | IF( nbondi == -1 ) THEN |
---|
| 1847 | CALL mppsend( 2, t2we(1,1,1), imigr, noea, ml_req1 ) |
---|
[3294] | 1848 | CALL mpprecv( 1, t2ew(1,1,2), imigr, noea ) |
---|
[1344] | 1849 | IF(l_isend) CALL mpi_wait( ml_req1, ml_stat, ml_err ) |
---|
| 1850 | ELSEIF( nbondi == 0 ) THEN |
---|
| 1851 | CALL mppsend( 1, t2ew(1,1,1), imigr, nowe, ml_req1 ) |
---|
| 1852 | CALL mppsend( 2, t2we(1,1,1), imigr, noea, ml_req2 ) |
---|
[3294] | 1853 | CALL mpprecv( 1, t2ew(1,1,2), imigr, noea ) |
---|
| 1854 | CALL mpprecv( 2, t2we(1,1,2), imigr, nowe ) |
---|
[1344] | 1855 | IF(l_isend) CALL mpi_wait( ml_req1, ml_stat, ml_err ) |
---|
| 1856 | IF(l_isend) CALL mpi_wait( ml_req2, ml_stat, ml_err ) |
---|
| 1857 | ELSEIF( nbondi == 1 ) THEN |
---|
| 1858 | CALL mppsend( 1, t2ew(1,1,1), imigr, nowe, ml_req1 ) |
---|
[3294] | 1859 | CALL mpprecv( 2, t2we(1,1,2), imigr, nowe ) |
---|
[1344] | 1860 | IF(l_isend) CALL mpi_wait( ml_req1, ml_stat, ml_err ) |
---|
| 1861 | ENDIF |
---|
| 1862 | ! |
---|
| 1863 | ! ! Write Dirichlet lateral conditions |
---|
| 1864 | iihom = nlci-jpreci |
---|
| 1865 | ! |
---|
| 1866 | IF( nbondi == 0 .OR. nbondi == 1 ) THEN |
---|
| 1867 | DO jl = 1, jpreci |
---|
| 1868 | ztab(jl,:) = t2we(:,jl,2) |
---|
| 1869 | END DO |
---|
| 1870 | ENDIF |
---|
| 1871 | IF( nbondi == -1 .OR. nbondi == 0 ) THEN |
---|
| 1872 | DO jl = 1, jpreci |
---|
| 1873 | ztab(iihom+jl,:) = t2ew(:,jl,2) |
---|
| 1874 | END DO |
---|
| 1875 | ENDIF |
---|
[3] | 1876 | |
---|
| 1877 | |
---|
[1344] | 1878 | ! 2. North and south directions |
---|
| 1879 | ! ----------------------------- |
---|
| 1880 | ! |
---|
| 1881 | IF( nbondj /= 2 ) THEN ! Read Dirichlet lateral conditions |
---|
| 1882 | ijhom = nlcj-nrecj |
---|
| 1883 | DO jl = 1, jprecj |
---|
| 1884 | t2sn(:,jl,1) = ztab(:,ijhom +jl) |
---|
| 1885 | t2ns(:,jl,1) = ztab(:,jprecj+jl) |
---|
| 1886 | END DO |
---|
| 1887 | ENDIF |
---|
| 1888 | ! |
---|
| 1889 | ! ! Migrations |
---|
| 1890 | imigr = jprecj * jpi |
---|
| 1891 | ! |
---|
| 1892 | IF( nbondj == -1 ) THEN |
---|
| 1893 | CALL mppsend( 4, t2sn(1,1,1), imigr, nono, ml_req1 ) |
---|
[3294] | 1894 | CALL mpprecv( 3, t2ns(1,1,2), imigr, nono ) |
---|
[1344] | 1895 | IF(l_isend) CALL mpi_wait( ml_req1, ml_stat, ml_err ) |
---|
| 1896 | ELSEIF( nbondj == 0 ) THEN |
---|
| 1897 | CALL mppsend( 3, t2ns(1,1,1), imigr, noso, ml_req1 ) |
---|
| 1898 | CALL mppsend( 4, t2sn(1,1,1), imigr, nono, ml_req2 ) |
---|
[3294] | 1899 | CALL mpprecv( 3, t2ns(1,1,2), imigr, nono ) |
---|
| 1900 | CALL mpprecv( 4, t2sn(1,1,2), imigr, noso ) |
---|
[1344] | 1901 | IF( l_isend ) CALL mpi_wait( ml_req1, ml_stat, ml_err ) |
---|
| 1902 | IF( l_isend ) CALL mpi_wait( ml_req2, ml_stat, ml_err ) |
---|
| 1903 | ELSEIF( nbondj == 1 ) THEN |
---|
| 1904 | CALL mppsend( 3, t2ns(1,1,1), imigr, noso, ml_req1 ) |
---|
[3294] | 1905 | CALL mpprecv( 4, t2sn(1,1,2), imigr, noso) |
---|
[1344] | 1906 | IF( l_isend ) CALL mpi_wait( ml_req1, ml_stat, ml_err ) |
---|
| 1907 | ENDIF |
---|
| 1908 | ! |
---|
| 1909 | ! ! Write Dirichlet lateral conditions |
---|
| 1910 | ijhom = nlcj - jprecj |
---|
| 1911 | IF( nbondj == 0 .OR. nbondj == 1 ) THEN |
---|
| 1912 | DO jl = 1, jprecj |
---|
| 1913 | ztab(:,jl) = t2sn(:,jl,2) |
---|
| 1914 | END DO |
---|
| 1915 | ENDIF |
---|
| 1916 | IF( nbondj == 0 .OR. nbondj == -1 ) THEN |
---|
| 1917 | DO jl = 1, jprecj |
---|
| 1918 | ztab(:,ijhom+jl) = t2ns(:,jl,2) |
---|
| 1919 | END DO |
---|
| 1920 | ENDIF |
---|
| 1921 | IF( ktype==1 .AND. kd1 <= jpi+nimpp-1 .AND. nimpp <= kd2 ) THEN |
---|
| 1922 | DO jj = ijpt0, ijpt1 ! north/south boundaries |
---|
| 1923 | DO ji = iipt0,ilpt1 |
---|
| 1924 | ptab(ji,jk) = ztab(ji,jj) |
---|
| 1925 | END DO |
---|
| 1926 | END DO |
---|
| 1927 | ELSEIF( ktype==2 .AND. kd1 <= jpj+njmpp-1 .AND. njmpp <= kd2 ) THEN |
---|
| 1928 | DO jj = ijpt0, ilpt1 ! east/west boundaries |
---|
| 1929 | DO ji = iipt0,iipt1 |
---|
| 1930 | ptab(jj,jk) = ztab(ji,jj) |
---|
| 1931 | END DO |
---|
| 1932 | END DO |
---|
| 1933 | ENDIF |
---|
| 1934 | ! |
---|
| 1935 | END DO |
---|
| 1936 | ! |
---|
[3294] | 1937 | CALL wrk_dealloc( jpi,jpj, ztab ) |
---|
[2715] | 1938 | ! |
---|
[1344] | 1939 | END SUBROUTINE mppobc |
---|
| 1940 | |
---|
[13] | 1941 | |
---|
[1344] | 1942 | SUBROUTINE mpp_comm_free( kcom ) |
---|
| 1943 | !!---------------------------------------------------------------------- |
---|
| 1944 | !!---------------------------------------------------------------------- |
---|
| 1945 | INTEGER, INTENT(in) :: kcom |
---|
| 1946 | !! |
---|
| 1947 | INTEGER :: ierr |
---|
| 1948 | !!---------------------------------------------------------------------- |
---|
| 1949 | ! |
---|
| 1950 | CALL MPI_COMM_FREE(kcom, ierr) |
---|
| 1951 | ! |
---|
| 1952 | END SUBROUTINE mpp_comm_free |
---|
[3] | 1953 | |
---|
[869] | 1954 | |
---|
[2715] | 1955 | SUBROUTINE mpp_ini_ice( pindic, kumout ) |
---|
[1344] | 1956 | !!---------------------------------------------------------------------- |
---|
| 1957 | !! *** routine mpp_ini_ice *** |
---|
| 1958 | !! |
---|
| 1959 | !! ** Purpose : Initialize special communicator for ice areas |
---|
| 1960 | !! condition together with global variables needed in the ddmpp folding |
---|
| 1961 | !! |
---|
| 1962 | !! ** Method : - Look for ice processors in ice routines |
---|
| 1963 | !! - Put their number in nrank_ice |
---|
| 1964 | !! - Create groups for the world processors and the ice processors |
---|
| 1965 | !! - Create a communicator for ice processors |
---|
| 1966 | !! |
---|
| 1967 | !! ** output |
---|
| 1968 | !! njmppmax = njmpp for northern procs |
---|
| 1969 | !! ndim_rank_ice = number of processors with ice |
---|
| 1970 | !! nrank_ice (ndim_rank_ice) = ice processors |
---|
[3625] | 1971 | !! ngrp_iworld = group ID for the world processors |
---|
[1344] | 1972 | !! ngrp_ice = group ID for the ice processors |
---|
| 1973 | !! ncomm_ice = communicator for the ice procs. |
---|
| 1974 | !! n_ice_root = number (in the world) of proc 0 in the ice comm. |
---|
| 1975 | !! |
---|
| 1976 | !!---------------------------------------------------------------------- |
---|
[2715] | 1977 | INTEGER, INTENT(in) :: pindic |
---|
| 1978 | INTEGER, INTENT(in) :: kumout ! ocean.output logical unit |
---|
[1344] | 1979 | !! |
---|
| 1980 | INTEGER :: jjproc |
---|
[2715] | 1981 | INTEGER :: ii, ierr |
---|
| 1982 | INTEGER, ALLOCATABLE, DIMENSION(:) :: kice |
---|
| 1983 | INTEGER, ALLOCATABLE, DIMENSION(:) :: zwork |
---|
[1344] | 1984 | !!---------------------------------------------------------------------- |
---|
| 1985 | ! |
---|
[2715] | 1986 | ! Since this is just an init routine and these arrays are of length jpnij |
---|
| 1987 | ! then don't use wrk_nemo module - just allocate and deallocate. |
---|
| 1988 | ALLOCATE( kice(jpnij), zwork(jpnij), STAT=ierr ) |
---|
| 1989 | IF( ierr /= 0 ) THEN |
---|
| 1990 | WRITE(kumout, cform_err) |
---|
| 1991 | WRITE(kumout,*) 'mpp_ini_ice : failed to allocate 2, 1D arrays (jpnij in length)' |
---|
| 1992 | CALL mppstop |
---|
| 1993 | ENDIF |
---|
| 1994 | |
---|
[1344] | 1995 | ! Look for how many procs with sea-ice |
---|
| 1996 | ! |
---|
| 1997 | kice = 0 |
---|
| 1998 | DO jjproc = 1, jpnij |
---|
| 1999 | IF( jjproc == narea .AND. pindic .GT. 0 ) kice(jjproc) = 1 |
---|
| 2000 | END DO |
---|
| 2001 | ! |
---|
| 2002 | zwork = 0 |
---|
| 2003 | CALL MPI_ALLREDUCE( kice, zwork, jpnij, mpi_integer, mpi_sum, mpi_comm_opa, ierr ) |
---|
| 2004 | ndim_rank_ice = SUM( zwork ) |
---|
[3] | 2005 | |
---|
[1344] | 2006 | ! Allocate the right size to nrank_north |
---|
[1441] | 2007 | IF( ALLOCATED ( nrank_ice ) ) DEALLOCATE( nrank_ice ) |
---|
[1344] | 2008 | ALLOCATE( nrank_ice(ndim_rank_ice) ) |
---|
| 2009 | ! |
---|
| 2010 | ii = 0 |
---|
| 2011 | nrank_ice = 0 |
---|
| 2012 | DO jjproc = 1, jpnij |
---|
| 2013 | IF( zwork(jjproc) == 1) THEN |
---|
| 2014 | ii = ii + 1 |
---|
| 2015 | nrank_ice(ii) = jjproc -1 |
---|
| 2016 | ENDIF |
---|
| 2017 | END DO |
---|
[1208] | 2018 | |
---|
[1344] | 2019 | ! Create the world group |
---|
[3625] | 2020 | CALL MPI_COMM_GROUP( mpi_comm_opa, ngrp_iworld, ierr ) |
---|
[869] | 2021 | |
---|
[1344] | 2022 | ! Create the ice group from the world group |
---|
[3625] | 2023 | CALL MPI_GROUP_INCL( ngrp_iworld, ndim_rank_ice, nrank_ice, ngrp_ice, ierr ) |
---|
[869] | 2024 | |
---|
[1344] | 2025 | ! Create the ice communicator , ie the pool of procs with sea-ice |
---|
| 2026 | CALL MPI_COMM_CREATE( mpi_comm_opa, ngrp_ice, ncomm_ice, ierr ) |
---|
[869] | 2027 | |
---|
[1344] | 2028 | ! Find proc number in the world of proc 0 in the north |
---|
| 2029 | ! The following line seems to be useless, we just comment & keep it as reminder |
---|
[3625] | 2030 | ! CALL MPI_GROUP_TRANSLATE_RANKS(ngrp_ice,1,0,ngrp_iworld,n_ice_root,ierr) |
---|
[1344] | 2031 | ! |
---|
[3625] | 2032 | CALL MPI_GROUP_FREE(ngrp_ice, ierr) |
---|
| 2033 | CALL MPI_GROUP_FREE(ngrp_iworld, ierr) |
---|
| 2034 | |
---|
[2715] | 2035 | DEALLOCATE(kice, zwork) |
---|
| 2036 | ! |
---|
[1344] | 2037 | END SUBROUTINE mpp_ini_ice |
---|
[869] | 2038 | |
---|
| 2039 | |
---|
[2715] | 2040 | SUBROUTINE mpp_ini_znl( kumout ) |
---|
[1345] | 2041 | !!---------------------------------------------------------------------- |
---|
| 2042 | !! *** routine mpp_ini_znl *** |
---|
| 2043 | !! |
---|
| 2044 | !! ** Purpose : Initialize special communicator for computing zonal sum |
---|
| 2045 | !! |
---|
| 2046 | !! ** Method : - Look for processors in the same row |
---|
| 2047 | !! - Put their number in nrank_znl |
---|
| 2048 | !! - Create group for the znl processors |
---|
| 2049 | !! - Create a communicator for znl processors |
---|
| 2050 | !! - Determine if processor should write znl files |
---|
| 2051 | !! |
---|
| 2052 | !! ** output |
---|
| 2053 | !! ndim_rank_znl = number of processors on the same row |
---|
| 2054 | !! ngrp_znl = group ID for the znl processors |
---|
| 2055 | !! ncomm_znl = communicator for the ice procs. |
---|
| 2056 | !! n_znl_root = number (in the world) of proc 0 in the ice comm. |
---|
| 2057 | !! |
---|
| 2058 | !!---------------------------------------------------------------------- |
---|
[2715] | 2059 | INTEGER, INTENT(in) :: kumout ! ocean.output logical units |
---|
[1345] | 2060 | ! |
---|
[2715] | 2061 | INTEGER :: jproc ! dummy loop integer |
---|
| 2062 | INTEGER :: ierr, ii ! local integer |
---|
| 2063 | INTEGER, ALLOCATABLE, DIMENSION(:) :: kwork |
---|
| 2064 | !!---------------------------------------------------------------------- |
---|
[1345] | 2065 | !-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - ngrp_world : ', ngrp_world |
---|
| 2066 | !-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - mpi_comm_world : ', mpi_comm_world |
---|
| 2067 | !-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - mpi_comm_opa : ', mpi_comm_opa |
---|
| 2068 | ! |
---|
[2715] | 2069 | ALLOCATE( kwork(jpnij), STAT=ierr ) |
---|
| 2070 | IF( ierr /= 0 ) THEN |
---|
| 2071 | WRITE(kumout, cform_err) |
---|
| 2072 | WRITE(kumout,*) 'mpp_ini_znl : failed to allocate 1D array of length jpnij' |
---|
| 2073 | CALL mppstop |
---|
| 2074 | ENDIF |
---|
| 2075 | |
---|
| 2076 | IF( jpnj == 1 ) THEN |
---|
[1345] | 2077 | ngrp_znl = ngrp_world |
---|
| 2078 | ncomm_znl = mpi_comm_opa |
---|
| 2079 | ELSE |
---|
| 2080 | ! |
---|
| 2081 | CALL MPI_ALLGATHER ( njmpp, 1, mpi_integer, kwork, 1, mpi_integer, mpi_comm_opa, ierr ) |
---|
| 2082 | !-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - kwork pour njmpp : ', kwork |
---|
| 2083 | !-$$ CALL flush(numout) |
---|
| 2084 | ! |
---|
| 2085 | ! Count number of processors on the same row |
---|
| 2086 | ndim_rank_znl = 0 |
---|
| 2087 | DO jproc=1,jpnij |
---|
| 2088 | IF ( kwork(jproc) == njmpp ) THEN |
---|
| 2089 | ndim_rank_znl = ndim_rank_znl + 1 |
---|
| 2090 | ENDIF |
---|
| 2091 | END DO |
---|
| 2092 | !-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - ndim_rank_znl : ', ndim_rank_znl |
---|
| 2093 | !-$$ CALL flush(numout) |
---|
| 2094 | ! Allocate the right size to nrank_znl |
---|
[1441] | 2095 | IF (ALLOCATED (nrank_znl)) DEALLOCATE(nrank_znl) |
---|
[1345] | 2096 | ALLOCATE(nrank_znl(ndim_rank_znl)) |
---|
| 2097 | ii = 0 |
---|
| 2098 | nrank_znl (:) = 0 |
---|
| 2099 | DO jproc=1,jpnij |
---|
| 2100 | IF ( kwork(jproc) == njmpp) THEN |
---|
| 2101 | ii = ii + 1 |
---|
| 2102 | nrank_znl(ii) = jproc -1 |
---|
| 2103 | ENDIF |
---|
| 2104 | END DO |
---|
| 2105 | !-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - nrank_znl : ', nrank_znl |
---|
| 2106 | !-$$ CALL flush(numout) |
---|
| 2107 | |
---|
| 2108 | ! Create the opa group |
---|
| 2109 | CALL MPI_COMM_GROUP(mpi_comm_opa,ngrp_opa,ierr) |
---|
| 2110 | !-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - ngrp_opa : ', ngrp_opa |
---|
| 2111 | !-$$ CALL flush(numout) |
---|
| 2112 | |
---|
| 2113 | ! Create the znl group from the opa group |
---|
| 2114 | CALL MPI_GROUP_INCL ( ngrp_opa, ndim_rank_znl, nrank_znl, ngrp_znl, ierr ) |
---|
| 2115 | !-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - ngrp_znl ', ngrp_znl |
---|
| 2116 | !-$$ CALL flush(numout) |
---|
| 2117 | |
---|
| 2118 | ! Create the znl communicator from the opa communicator, ie the pool of procs in the same row |
---|
| 2119 | CALL MPI_COMM_CREATE ( mpi_comm_opa, ngrp_znl, ncomm_znl, ierr ) |
---|
| 2120 | !-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - ncomm_znl ', ncomm_znl |
---|
| 2121 | !-$$ CALL flush(numout) |
---|
| 2122 | ! |
---|
| 2123 | END IF |
---|
| 2124 | |
---|
| 2125 | ! Determines if processor if the first (starting from i=1) on the row |
---|
| 2126 | IF ( jpni == 1 ) THEN |
---|
| 2127 | l_znl_root = .TRUE. |
---|
| 2128 | ELSE |
---|
| 2129 | l_znl_root = .FALSE. |
---|
| 2130 | kwork (1) = nimpp |
---|
| 2131 | CALL mpp_min ( kwork(1), kcom = ncomm_znl) |
---|
| 2132 | IF ( nimpp == kwork(1)) l_znl_root = .TRUE. |
---|
| 2133 | END IF |
---|
| 2134 | |
---|
[2715] | 2135 | DEALLOCATE(kwork) |
---|
| 2136 | |
---|
[1345] | 2137 | END SUBROUTINE mpp_ini_znl |
---|
| 2138 | |
---|
| 2139 | |
---|
[1344] | 2140 | SUBROUTINE mpp_ini_north |
---|
| 2141 | !!---------------------------------------------------------------------- |
---|
| 2142 | !! *** routine mpp_ini_north *** |
---|
| 2143 | !! |
---|
| 2144 | !! ** Purpose : Initialize special communicator for north folding |
---|
| 2145 | !! condition together with global variables needed in the mpp folding |
---|
| 2146 | !! |
---|
| 2147 | !! ** Method : - Look for northern processors |
---|
| 2148 | !! - Put their number in nrank_north |
---|
| 2149 | !! - Create groups for the world processors and the north processors |
---|
| 2150 | !! - Create a communicator for northern processors |
---|
| 2151 | !! |
---|
| 2152 | !! ** output |
---|
| 2153 | !! njmppmax = njmpp for northern procs |
---|
| 2154 | !! ndim_rank_north = number of processors in the northern line |
---|
| 2155 | !! nrank_north (ndim_rank_north) = number of the northern procs. |
---|
| 2156 | !! ngrp_world = group ID for the world processors |
---|
| 2157 | !! ngrp_north = group ID for the northern processors |
---|
| 2158 | !! ncomm_north = communicator for the northern procs. |
---|
| 2159 | !! north_root = number (in the world) of proc 0 in the northern comm. |
---|
| 2160 | !! |
---|
| 2161 | !!---------------------------------------------------------------------- |
---|
| 2162 | INTEGER :: ierr |
---|
| 2163 | INTEGER :: jjproc |
---|
| 2164 | INTEGER :: ii, ji |
---|
| 2165 | !!---------------------------------------------------------------------- |
---|
| 2166 | ! |
---|
| 2167 | njmppmax = MAXVAL( njmppt ) |
---|
| 2168 | ! |
---|
| 2169 | ! Look for how many procs on the northern boundary |
---|
| 2170 | ndim_rank_north = 0 |
---|
| 2171 | DO jjproc = 1, jpnij |
---|
| 2172 | IF( njmppt(jjproc) == njmppmax ) ndim_rank_north = ndim_rank_north + 1 |
---|
| 2173 | END DO |
---|
| 2174 | ! |
---|
| 2175 | ! Allocate the right size to nrank_north |
---|
[1441] | 2176 | IF (ALLOCATED (nrank_north)) DEALLOCATE(nrank_north) |
---|
[1344] | 2177 | ALLOCATE( nrank_north(ndim_rank_north) ) |
---|
[869] | 2178 | |
---|
[1344] | 2179 | ! Fill the nrank_north array with proc. number of northern procs. |
---|
| 2180 | ! Note : the rank start at 0 in MPI |
---|
| 2181 | ii = 0 |
---|
| 2182 | DO ji = 1, jpnij |
---|
| 2183 | IF ( njmppt(ji) == njmppmax ) THEN |
---|
| 2184 | ii=ii+1 |
---|
| 2185 | nrank_north(ii)=ji-1 |
---|
| 2186 | END IF |
---|
| 2187 | END DO |
---|
| 2188 | ! |
---|
| 2189 | ! create the world group |
---|
| 2190 | CALL MPI_COMM_GROUP( mpi_comm_opa, ngrp_world, ierr ) |
---|
| 2191 | ! |
---|
| 2192 | ! Create the North group from the world group |
---|
| 2193 | CALL MPI_GROUP_INCL( ngrp_world, ndim_rank_north, nrank_north, ngrp_north, ierr ) |
---|
| 2194 | ! |
---|
| 2195 | ! Create the North communicator , ie the pool of procs in the north group |
---|
| 2196 | CALL MPI_COMM_CREATE( mpi_comm_opa, ngrp_north, ncomm_north, ierr ) |
---|
| 2197 | ! |
---|
| 2198 | END SUBROUTINE mpp_ini_north |
---|
[869] | 2199 | |
---|
| 2200 | |
---|
[1344] | 2201 | SUBROUTINE mpp_lbc_north_3d( pt3d, cd_type, psgn ) |
---|
[51] | 2202 | !!--------------------------------------------------------------------- |
---|
| 2203 | !! *** routine mpp_lbc_north_3d *** |
---|
| 2204 | !! |
---|
[1344] | 2205 | !! ** Purpose : Ensure proper north fold horizontal bondary condition |
---|
| 2206 | !! in mpp configuration in case of jpn1 > 1 |
---|
[51] | 2207 | !! |
---|
[1344] | 2208 | !! ** Method : North fold condition and mpp with more than one proc |
---|
| 2209 | !! in i-direction require a specific treatment. We gather |
---|
| 2210 | !! the 4 northern lines of the global domain on 1 processor |
---|
| 2211 | !! and apply lbc north-fold on this sub array. Then we |
---|
| 2212 | !! scatter the north fold array back to the processors. |
---|
[51] | 2213 | !! |
---|
| 2214 | !!---------------------------------------------------------------------- |
---|
[1344] | 2215 | REAL(wp), DIMENSION(jpi,jpj,jpk), INTENT(inout) :: pt3d ! 3D array on which the b.c. is applied |
---|
| 2216 | CHARACTER(len=1) , INTENT(in ) :: cd_type ! nature of pt3d grid-points |
---|
| 2217 | ! ! = T , U , V , F or W gridpoints |
---|
| 2218 | REAL(wp) , INTENT(in ) :: psgn ! = -1. the sign change across the north fold |
---|
| 2219 | !! ! = 1. , the sign is kept |
---|
| 2220 | INTEGER :: ji, jj, jr |
---|
| 2221 | INTEGER :: ierr, itaille, ildi, ilei, iilb |
---|
| 2222 | INTEGER :: ijpj, ijpjm1, ij, iproc |
---|
[3294] | 2223 | INTEGER, DIMENSION (jpmaxngh) :: ml_req_nf ! for mpi_isend when avoiding mpi_allgather |
---|
| 2224 | INTEGER :: ml_err ! for mpi_isend when avoiding mpi_allgather |
---|
| 2225 | INTEGER, DIMENSION(MPI_STATUS_SIZE) :: ml_stat ! for mpi_isend when avoiding mpi_allgather |
---|
[51] | 2226 | !!---------------------------------------------------------------------- |
---|
[1344] | 2227 | ! |
---|
| 2228 | ijpj = 4 |
---|
[3294] | 2229 | ityp = -1 |
---|
[1344] | 2230 | ijpjm1 = 3 |
---|
[2480] | 2231 | ztab(:,:,:) = 0.e0 |
---|
[1344] | 2232 | ! |
---|
| 2233 | DO jj = nlcj - ijpj +1, nlcj ! put in znorthloc the last 4 jlines of pt3d |
---|
| 2234 | ij = jj - nlcj + ijpj |
---|
| 2235 | znorthloc(:,ij,:) = pt3d(:,jj,:) |
---|
| 2236 | END DO |
---|
| 2237 | ! |
---|
| 2238 | ! ! Build in procs of ncomm_north the znorthgloio |
---|
| 2239 | itaille = jpi * jpk * ijpj |
---|
[3294] | 2240 | IF ( l_north_nogather ) THEN |
---|
| 2241 | ! |
---|
| 2242 | ! Avoid the use of mpi_allgather by exchanging only with the processes already identified |
---|
| 2243 | ! (in nemo_northcomms) as being involved in this process' northern boundary exchange |
---|
| 2244 | ! |
---|
| 2245 | DO jj = nlcj-ijpj+1, nlcj ! First put local values into the global array |
---|
| 2246 | ij = jj - nlcj + ijpj |
---|
| 2247 | DO ji = 1, nlci |
---|
| 2248 | ztab(ji+nimpp-1,ij,:) = pt3d(ji,jj,:) |
---|
[1344] | 2249 | END DO |
---|
| 2250 | END DO |
---|
[3294] | 2251 | |
---|
| 2252 | ! |
---|
| 2253 | ! Set the exchange type in order to access the correct list of active neighbours |
---|
| 2254 | ! |
---|
| 2255 | SELECT CASE ( cd_type ) |
---|
| 2256 | CASE ( 'T' , 'W' ) |
---|
| 2257 | ityp = 1 |
---|
| 2258 | CASE ( 'U' ) |
---|
| 2259 | ityp = 2 |
---|
| 2260 | CASE ( 'V' ) |
---|
| 2261 | ityp = 3 |
---|
| 2262 | CASE ( 'F' ) |
---|
| 2263 | ityp = 4 |
---|
| 2264 | CASE ( 'I' ) |
---|
| 2265 | ityp = 5 |
---|
| 2266 | CASE DEFAULT |
---|
| 2267 | ityp = -1 ! Set a default value for unsupported types which |
---|
| 2268 | ! will cause a fallback to the mpi_allgather method |
---|
| 2269 | END SELECT |
---|
| 2270 | IF ( ityp .gt. 0 ) THEN |
---|
| 2271 | |
---|
| 2272 | DO jr = 1,nsndto(ityp) |
---|
| 2273 | CALL mppsend(5, znorthloc, itaille, isendto(jr,ityp), ml_req_nf(jr) ) |
---|
| 2274 | END DO |
---|
| 2275 | DO jr = 1,nsndto(ityp) |
---|
| 2276 | CALL mpprecv(5, zfoldwk, itaille, isendto(jr,ityp)) |
---|
| 2277 | iproc = isendto(jr,ityp) + 1 |
---|
| 2278 | ildi = nldit (iproc) |
---|
| 2279 | ilei = nleit (iproc) |
---|
| 2280 | iilb = nimppt(iproc) |
---|
| 2281 | DO jj = 1, ijpj |
---|
| 2282 | DO ji = ildi, ilei |
---|
| 2283 | ztab(ji+iilb-1,jj,:) = zfoldwk(ji,jj,:) |
---|
| 2284 | END DO |
---|
| 2285 | END DO |
---|
| 2286 | END DO |
---|
| 2287 | IF (l_isend) THEN |
---|
| 2288 | DO jr = 1,nsndto(ityp) |
---|
| 2289 | CALL mpi_wait(ml_req_nf(jr), ml_stat, ml_err) |
---|
| 2290 | END DO |
---|
| 2291 | ENDIF |
---|
| 2292 | |
---|
| 2293 | ENDIF |
---|
| 2294 | |
---|
| 2295 | ENDIF |
---|
| 2296 | |
---|
| 2297 | IF ( ityp .lt. 0 ) THEN |
---|
| 2298 | CALL MPI_ALLGATHER( znorthloc , itaille, MPI_DOUBLE_PRECISION, & |
---|
| 2299 | & znorthgloio, itaille, MPI_DOUBLE_PRECISION, ncomm_north, ierr ) |
---|
| 2300 | ! |
---|
| 2301 | DO jr = 1, ndim_rank_north ! recover the global north array |
---|
| 2302 | iproc = nrank_north(jr) + 1 |
---|
| 2303 | ildi = nldit (iproc) |
---|
| 2304 | ilei = nleit (iproc) |
---|
| 2305 | iilb = nimppt(iproc) |
---|
| 2306 | DO jj = 1, ijpj |
---|
| 2307 | DO ji = ildi, ilei |
---|
| 2308 | ztab(ji+iilb-1,jj,:) = znorthgloio(ji,jj,:,jr) |
---|
| 2309 | END DO |
---|
| 2310 | END DO |
---|
| 2311 | END DO |
---|
| 2312 | ENDIF |
---|
[1344] | 2313 | ! |
---|
[3294] | 2314 | ! The ztab array has been either: |
---|
| 2315 | ! a. Fully populated by the mpi_allgather operation or |
---|
| 2316 | ! b. Had the active points for this domain and northern neighbours populated |
---|
| 2317 | ! by peer to peer exchanges |
---|
| 2318 | ! Either way the array may be folded by lbc_nfd and the result for the span of |
---|
| 2319 | ! this domain will be identical. |
---|
| 2320 | ! |
---|
[1344] | 2321 | CALL lbc_nfd( ztab, cd_type, psgn ) ! North fold boundary condition |
---|
| 2322 | ! |
---|
| 2323 | DO jj = nlcj-ijpj+1, nlcj ! Scatter back to pt3d |
---|
| 2324 | ij = jj - nlcj + ijpj |
---|
| 2325 | DO ji= 1, nlci |
---|
| 2326 | pt3d(ji,jj,:) = ztab(ji+nimpp-1,ij,:) |
---|
| 2327 | END DO |
---|
| 2328 | END DO |
---|
| 2329 | ! |
---|
| 2330 | END SUBROUTINE mpp_lbc_north_3d |
---|
[3] | 2331 | |
---|
| 2332 | |
---|
[1344] | 2333 | SUBROUTINE mpp_lbc_north_2d( pt2d, cd_type, psgn) |
---|
| 2334 | !!--------------------------------------------------------------------- |
---|
| 2335 | !! *** routine mpp_lbc_north_2d *** |
---|
| 2336 | !! |
---|
| 2337 | !! ** Purpose : Ensure proper north fold horizontal bondary condition |
---|
| 2338 | !! in mpp configuration in case of jpn1 > 1 (for 2d array ) |
---|
| 2339 | !! |
---|
| 2340 | !! ** Method : North fold condition and mpp with more than one proc |
---|
| 2341 | !! in i-direction require a specific treatment. We gather |
---|
| 2342 | !! the 4 northern lines of the global domain on 1 processor |
---|
| 2343 | !! and apply lbc north-fold on this sub array. Then we |
---|
| 2344 | !! scatter the north fold array back to the processors. |
---|
| 2345 | !! |
---|
| 2346 | !!---------------------------------------------------------------------- |
---|
| 2347 | REAL(wp), DIMENSION(jpi,jpj), INTENT(inout) :: pt2d ! 3D array on which the b.c. is applied |
---|
| 2348 | CHARACTER(len=1) , INTENT(in ) :: cd_type ! nature of pt3d grid-points |
---|
| 2349 | ! ! = T , U , V , F or W gridpoints |
---|
| 2350 | REAL(wp) , INTENT(in ) :: psgn ! = -1. the sign change across the north fold |
---|
| 2351 | !! ! = 1. , the sign is kept |
---|
| 2352 | INTEGER :: ji, jj, jr |
---|
| 2353 | INTEGER :: ierr, itaille, ildi, ilei, iilb |
---|
| 2354 | INTEGER :: ijpj, ijpjm1, ij, iproc |
---|
[3294] | 2355 | INTEGER, DIMENSION (jpmaxngh) :: ml_req_nf ! for mpi_isend when avoiding mpi_allgather |
---|
| 2356 | INTEGER :: ml_err ! for mpi_isend when avoiding mpi_allgather |
---|
| 2357 | INTEGER, DIMENSION(MPI_STATUS_SIZE):: ml_stat ! for mpi_isend when avoiding mpi_allgather |
---|
[1344] | 2358 | !!---------------------------------------------------------------------- |
---|
| 2359 | ! |
---|
| 2360 | ijpj = 4 |
---|
[3294] | 2361 | ityp = -1 |
---|
[1344] | 2362 | ijpjm1 = 3 |
---|
[2715] | 2363 | ztab_2d(:,:) = 0.e0 |
---|
[1344] | 2364 | ! |
---|
[2715] | 2365 | DO jj = nlcj-ijpj+1, nlcj ! put in znorthloc_2d the last 4 jlines of pt2d |
---|
[1344] | 2366 | ij = jj - nlcj + ijpj |
---|
[2715] | 2367 | znorthloc_2d(:,ij) = pt2d(:,jj) |
---|
[1344] | 2368 | END DO |
---|
[3] | 2369 | |
---|
[2715] | 2370 | ! ! Build in procs of ncomm_north the znorthgloio_2d |
---|
[1344] | 2371 | itaille = jpi * ijpj |
---|
[3294] | 2372 | IF ( l_north_nogather ) THEN |
---|
| 2373 | ! |
---|
| 2374 | ! Avoid the use of mpi_allgather by exchanging only with the processes already identified |
---|
| 2375 | ! (in nemo_northcomms) as being involved in this process' northern boundary exchange |
---|
| 2376 | ! |
---|
| 2377 | DO jj = nlcj-ijpj+1, nlcj ! First put local values into the global array |
---|
| 2378 | ij = jj - nlcj + ijpj |
---|
| 2379 | DO ji = 1, nlci |
---|
| 2380 | ztab_2d(ji+nimpp-1,ij) = pt2d(ji,jj) |
---|
[13] | 2381 | END DO |
---|
| 2382 | END DO |
---|
[3294] | 2383 | |
---|
| 2384 | ! |
---|
| 2385 | ! Set the exchange type in order to access the correct list of active neighbours |
---|
| 2386 | ! |
---|
| 2387 | SELECT CASE ( cd_type ) |
---|
| 2388 | CASE ( 'T' , 'W' ) |
---|
| 2389 | ityp = 1 |
---|
| 2390 | CASE ( 'U' ) |
---|
| 2391 | ityp = 2 |
---|
| 2392 | CASE ( 'V' ) |
---|
| 2393 | ityp = 3 |
---|
| 2394 | CASE ( 'F' ) |
---|
| 2395 | ityp = 4 |
---|
| 2396 | CASE ( 'I' ) |
---|
| 2397 | ityp = 5 |
---|
| 2398 | CASE DEFAULT |
---|
| 2399 | ityp = -1 ! Set a default value for unsupported types which |
---|
| 2400 | ! will cause a fallback to the mpi_allgather method |
---|
| 2401 | END SELECT |
---|
| 2402 | |
---|
| 2403 | IF ( ityp .gt. 0 ) THEN |
---|
| 2404 | |
---|
| 2405 | DO jr = 1,nsndto(ityp) |
---|
| 2406 | CALL mppsend(5, znorthloc_2d, itaille, isendto(jr,ityp), ml_req_nf(jr) ) |
---|
| 2407 | END DO |
---|
| 2408 | DO jr = 1,nsndto(ityp) |
---|
| 2409 | CALL mpprecv(5, zfoldwk_2d, itaille, isendto(jr,ityp)) |
---|
| 2410 | iproc = isendto(jr,ityp) + 1 |
---|
| 2411 | ildi = nldit (iproc) |
---|
| 2412 | ilei = nleit (iproc) |
---|
| 2413 | iilb = nimppt(iproc) |
---|
| 2414 | DO jj = 1, ijpj |
---|
| 2415 | DO ji = ildi, ilei |
---|
| 2416 | ztab_2d(ji+iilb-1,jj) = zfoldwk_2d(ji,jj) |
---|
| 2417 | END DO |
---|
| 2418 | END DO |
---|
| 2419 | END DO |
---|
| 2420 | IF (l_isend) THEN |
---|
| 2421 | DO jr = 1,nsndto(ityp) |
---|
| 2422 | CALL mpi_wait(ml_req_nf(jr), ml_stat, ml_err) |
---|
| 2423 | END DO |
---|
| 2424 | ENDIF |
---|
| 2425 | |
---|
| 2426 | ENDIF |
---|
| 2427 | |
---|
| 2428 | ENDIF |
---|
| 2429 | |
---|
| 2430 | IF ( ityp .lt. 0 ) THEN |
---|
| 2431 | CALL MPI_ALLGATHER( znorthloc_2d , itaille, MPI_DOUBLE_PRECISION, & |
---|
| 2432 | & znorthgloio_2d, itaille, MPI_DOUBLE_PRECISION, ncomm_north, ierr ) |
---|
| 2433 | ! |
---|
| 2434 | DO jr = 1, ndim_rank_north ! recover the global north array |
---|
| 2435 | iproc = nrank_north(jr) + 1 |
---|
| 2436 | ildi = nldit (iproc) |
---|
| 2437 | ilei = nleit (iproc) |
---|
| 2438 | iilb = nimppt(iproc) |
---|
| 2439 | DO jj = 1, ijpj |
---|
| 2440 | DO ji = ildi, ilei |
---|
| 2441 | ztab_2d(ji+iilb-1,jj) = znorthgloio_2d(ji,jj,jr) |
---|
| 2442 | END DO |
---|
| 2443 | END DO |
---|
| 2444 | END DO |
---|
| 2445 | ENDIF |
---|
[1344] | 2446 | ! |
---|
[3294] | 2447 | ! The ztab array has been either: |
---|
| 2448 | ! a. Fully populated by the mpi_allgather operation or |
---|
| 2449 | ! b. Had the active points for this domain and northern neighbours populated |
---|
| 2450 | ! by peer to peer exchanges |
---|
| 2451 | ! Either way the array may be folded by lbc_nfd and the result for the span of |
---|
| 2452 | ! this domain will be identical. |
---|
| 2453 | ! |
---|
[2715] | 2454 | CALL lbc_nfd( ztab_2d, cd_type, psgn ) ! North fold boundary condition |
---|
[1344] | 2455 | ! |
---|
| 2456 | ! |
---|
| 2457 | DO jj = nlcj-ijpj+1, nlcj ! Scatter back to pt2d |
---|
[13] | 2458 | ij = jj - nlcj + ijpj |
---|
[1344] | 2459 | DO ji = 1, nlci |
---|
[2715] | 2460 | pt2d(ji,jj) = ztab_2d(ji+nimpp-1,ij) |
---|
[1344] | 2461 | END DO |
---|
[13] | 2462 | END DO |
---|
[1344] | 2463 | ! |
---|
[13] | 2464 | END SUBROUTINE mpp_lbc_north_2d |
---|
[3] | 2465 | |
---|
| 2466 | |
---|
[1344] | 2467 | SUBROUTINE mpp_lbc_north_e( pt2d, cd_type, psgn) |
---|
| 2468 | !!--------------------------------------------------------------------- |
---|
| 2469 | !! *** routine mpp_lbc_north_2d *** |
---|
| 2470 | !! |
---|
| 2471 | !! ** Purpose : Ensure proper north fold horizontal bondary condition |
---|
| 2472 | !! in mpp configuration in case of jpn1 > 1 and for 2d |
---|
| 2473 | !! array with outer extra halo |
---|
| 2474 | !! |
---|
| 2475 | !! ** Method : North fold condition and mpp with more than one proc |
---|
| 2476 | !! in i-direction require a specific treatment. We gather |
---|
| 2477 | !! the 4+2*jpr2dj northern lines of the global domain on 1 |
---|
| 2478 | !! processor and apply lbc north-fold on this sub array. |
---|
| 2479 | !! Then we scatter the north fold array back to the processors. |
---|
| 2480 | !! |
---|
| 2481 | !!---------------------------------------------------------------------- |
---|
| 2482 | REAL(wp), DIMENSION(1-jpr2di:jpi+jpr2di,1-jpr2dj:jpj+jpr2dj), INTENT(inout) :: pt2d ! 2D array with extra halo |
---|
| 2483 | CHARACTER(len=1) , INTENT(in ) :: cd_type ! nature of pt3d grid-points |
---|
| 2484 | ! ! = T , U , V , F or W -points |
---|
| 2485 | REAL(wp) , INTENT(in ) :: psgn ! = -1. the sign change across the |
---|
| 2486 | !! ! north fold, = 1. otherwise |
---|
| 2487 | INTEGER :: ji, jj, jr |
---|
| 2488 | INTEGER :: ierr, itaille, ildi, ilei, iilb |
---|
| 2489 | INTEGER :: ijpj, ij, iproc |
---|
| 2490 | !!---------------------------------------------------------------------- |
---|
| 2491 | ! |
---|
| 2492 | ijpj=4 |
---|
[2715] | 2493 | ztab_e(:,:) = 0.e0 |
---|
[311] | 2494 | |
---|
[1344] | 2495 | ij=0 |
---|
[2715] | 2496 | ! put in znorthloc_e the last 4 jlines of pt2d |
---|
[1344] | 2497 | DO jj = nlcj - ijpj + 1 - jpr2dj, nlcj +jpr2dj |
---|
| 2498 | ij = ij + 1 |
---|
| 2499 | DO ji = 1, jpi |
---|
[2715] | 2500 | znorthloc_e(ji,ij)=pt2d(ji,jj) |
---|
[1344] | 2501 | END DO |
---|
| 2502 | END DO |
---|
| 2503 | ! |
---|
| 2504 | itaille = jpi * ( ijpj + 2 * jpr2dj ) |
---|
[2715] | 2505 | CALL MPI_ALLGATHER( znorthloc_e(1,1) , itaille, MPI_DOUBLE_PRECISION, & |
---|
| 2506 | & znorthgloio_e(1,1,1), itaille, MPI_DOUBLE_PRECISION, ncomm_north, ierr ) |
---|
[1344] | 2507 | ! |
---|
| 2508 | DO jr = 1, ndim_rank_north ! recover the global north array |
---|
| 2509 | iproc = nrank_north(jr) + 1 |
---|
| 2510 | ildi = nldit (iproc) |
---|
| 2511 | ilei = nleit (iproc) |
---|
| 2512 | iilb = nimppt(iproc) |
---|
| 2513 | DO jj = 1, ijpj+2*jpr2dj |
---|
| 2514 | DO ji = ildi, ilei |
---|
[2715] | 2515 | ztab_e(ji+iilb-1,jj) = znorthgloio_e(ji,jj,jr) |
---|
[311] | 2516 | END DO |
---|
[1344] | 2517 | END DO |
---|
| 2518 | END DO |
---|
[311] | 2519 | |
---|
| 2520 | |
---|
[1344] | 2521 | ! 2. North-Fold boundary conditions |
---|
| 2522 | ! ---------------------------------- |
---|
[2715] | 2523 | CALL lbc_nfd( ztab_e(:,:), cd_type, psgn, pr2dj = jpr2dj ) |
---|
[311] | 2524 | |
---|
[1344] | 2525 | ij = jpr2dj |
---|
| 2526 | !! Scatter back to pt2d |
---|
| 2527 | DO jj = nlcj - ijpj + 1 , nlcj +jpr2dj |
---|
| 2528 | ij = ij +1 |
---|
| 2529 | DO ji= 1, nlci |
---|
[2715] | 2530 | pt2d(ji,jj) = ztab_e(ji+nimpp-1,ij) |
---|
[311] | 2531 | END DO |
---|
| 2532 | END DO |
---|
[1344] | 2533 | ! |
---|
[311] | 2534 | END SUBROUTINE mpp_lbc_north_e |
---|
| 2535 | |
---|
[389] | 2536 | |
---|
[2481] | 2537 | SUBROUTINE mpi_init_opa( ldtxt, ksft, code ) |
---|
[1344] | 2538 | !!--------------------------------------------------------------------- |
---|
| 2539 | !! *** routine mpp_init.opa *** |
---|
| 2540 | !! |
---|
| 2541 | !! ** Purpose :: export and attach a MPI buffer for bsend |
---|
| 2542 | !! |
---|
| 2543 | !! ** Method :: define buffer size in namelist, if 0 no buffer attachment |
---|
| 2544 | !! but classical mpi_init |
---|
| 2545 | !! |
---|
| 2546 | !! History :: 01/11 :: IDRIS initial version for IBM only |
---|
| 2547 | !! 08/04 :: R. Benshila, generalisation |
---|
| 2548 | !!--------------------------------------------------------------------- |
---|
[2481] | 2549 | CHARACTER(len=*),DIMENSION(:), INTENT( out) :: ldtxt |
---|
| 2550 | INTEGER , INTENT(inout) :: ksft |
---|
| 2551 | INTEGER , INTENT( out) :: code |
---|
| 2552 | INTEGER :: ierr, ji |
---|
| 2553 | LOGICAL :: mpi_was_called |
---|
[1344] | 2554 | !!--------------------------------------------------------------------- |
---|
| 2555 | ! |
---|
| 2556 | CALL mpi_initialized( mpi_was_called, code ) ! MPI initialization |
---|
[532] | 2557 | IF ( code /= MPI_SUCCESS ) THEN |
---|
[2481] | 2558 | DO ji = 1, SIZE(ldtxt) |
---|
| 2559 | IF( TRIM(ldtxt(ji)) /= '' ) WRITE(*,*) ldtxt(ji) ! control print of mynode |
---|
| 2560 | END DO |
---|
| 2561 | WRITE(*, cform_err) |
---|
| 2562 | WRITE(*, *) ' lib_mpp: Error in routine mpi_initialized' |
---|
[1344] | 2563 | CALL mpi_abort( mpi_comm_world, code, ierr ) |
---|
[532] | 2564 | ENDIF |
---|
[1344] | 2565 | ! |
---|
| 2566 | IF( .NOT. mpi_was_called ) THEN |
---|
| 2567 | CALL mpi_init( code ) |
---|
| 2568 | CALL mpi_comm_dup( mpi_comm_world, mpi_comm_opa, code ) |
---|
[532] | 2569 | IF ( code /= MPI_SUCCESS ) THEN |
---|
[2481] | 2570 | DO ji = 1, SIZE(ldtxt) |
---|
| 2571 | IF( TRIM(ldtxt(ji)) /= '' ) WRITE(*,*) ldtxt(ji) ! control print of mynode |
---|
| 2572 | END DO |
---|
| 2573 | WRITE(*, cform_err) |
---|
| 2574 | WRITE(*, *) ' lib_mpp: Error in routine mpi_comm_dup' |
---|
[532] | 2575 | CALL mpi_abort( mpi_comm_world, code, ierr ) |
---|
| 2576 | ENDIF |
---|
| 2577 | ENDIF |
---|
[1344] | 2578 | ! |
---|
[897] | 2579 | IF( nn_buffer > 0 ) THEN |
---|
[2481] | 2580 | WRITE(ldtxt(ksft),*) 'mpi_bsend, buffer allocation of : ', nn_buffer ; ksft = ksft + 1 |
---|
[897] | 2581 | ! Buffer allocation and attachment |
---|
[2481] | 2582 | ALLOCATE( tampon(nn_buffer), stat = ierr ) |
---|
[2715] | 2583 | IF( ierr /= 0 ) THEN |
---|
[2481] | 2584 | DO ji = 1, SIZE(ldtxt) |
---|
| 2585 | IF( TRIM(ldtxt(ji)) /= '' ) WRITE(*,*) ldtxt(ji) ! control print of mynode |
---|
| 2586 | END DO |
---|
| 2587 | WRITE(*, cform_err) |
---|
| 2588 | WRITE(*, *) ' lib_mpp: Error in ALLOCATE', ierr |
---|
| 2589 | CALL mpi_abort( mpi_comm_world, code, ierr ) |
---|
| 2590 | END IF |
---|
| 2591 | CALL mpi_buffer_attach( tampon, nn_buffer, code ) |
---|
[897] | 2592 | ENDIF |
---|
[1344] | 2593 | ! |
---|
[13] | 2594 | END SUBROUTINE mpi_init_opa |
---|
[3] | 2595 | |
---|
[1976] | 2596 | SUBROUTINE DDPDD_MPI (ydda, yddb, ilen, itype) |
---|
| 2597 | !!--------------------------------------------------------------------- |
---|
| 2598 | !! Routine DDPDD_MPI: used by reduction operator MPI_SUMDD |
---|
| 2599 | !! |
---|
| 2600 | !! Modification of original codes written by David H. Bailey |
---|
| 2601 | !! This subroutine computes yddb(i) = ydda(i)+yddb(i) |
---|
| 2602 | !!--------------------------------------------------------------------- |
---|
| 2603 | INTEGER, INTENT(in) :: ilen, itype |
---|
| 2604 | COMPLEX(wp), DIMENSION(ilen), INTENT(in) :: ydda |
---|
| 2605 | COMPLEX(wp), DIMENSION(ilen), INTENT(inout) :: yddb |
---|
| 2606 | ! |
---|
| 2607 | REAL(wp) :: zerr, zt1, zt2 ! local work variables |
---|
| 2608 | INTEGER :: ji, ztmp ! local scalar |
---|
| 2609 | |
---|
| 2610 | ztmp = itype ! avoid compilation warning |
---|
| 2611 | |
---|
| 2612 | DO ji=1,ilen |
---|
| 2613 | ! Compute ydda + yddb using Knuth's trick. |
---|
| 2614 | zt1 = real(ydda(ji)) + real(yddb(ji)) |
---|
| 2615 | zerr = zt1 - real(ydda(ji)) |
---|
| 2616 | zt2 = ((real(yddb(ji)) - zerr) + (real(ydda(ji)) - (zt1 - zerr))) & |
---|
| 2617 | + aimag(ydda(ji)) + aimag(yddb(ji)) |
---|
| 2618 | |
---|
| 2619 | ! The result is zt1 + zt2, after normalization. |
---|
| 2620 | yddb(ji) = cmplx ( zt1 + zt2, zt2 - ((zt1 + zt2) - zt1),wp ) |
---|
| 2621 | END DO |
---|
| 2622 | |
---|
| 2623 | END SUBROUTINE DDPDD_MPI |
---|
| 2624 | |
---|
[13] | 2625 | #else |
---|
| 2626 | !!---------------------------------------------------------------------- |
---|
| 2627 | !! Default case: Dummy module share memory computing |
---|
| 2628 | !!---------------------------------------------------------------------- |
---|
[2715] | 2629 | USE in_out_manager |
---|
[1976] | 2630 | |
---|
[13] | 2631 | INTERFACE mpp_sum |
---|
[3294] | 2632 | MODULE PROCEDURE mpp_sum_a2s, mpp_sum_as, mpp_sum_ai, mpp_sum_s, mpp_sum_i, mppsum_realdd, mppsum_a_realdd |
---|
[13] | 2633 | END INTERFACE |
---|
| 2634 | INTERFACE mpp_max |
---|
[681] | 2635 | MODULE PROCEDURE mppmax_a_int, mppmax_int, mppmax_a_real, mppmax_real |
---|
[13] | 2636 | END INTERFACE |
---|
| 2637 | INTERFACE mpp_min |
---|
| 2638 | MODULE PROCEDURE mppmin_a_int, mppmin_int, mppmin_a_real, mppmin_real |
---|
| 2639 | END INTERFACE |
---|
| 2640 | INTERFACE mppobc |
---|
| 2641 | MODULE PROCEDURE mppobc_1d, mppobc_2d, mppobc_3d, mppobc_4d |
---|
| 2642 | END INTERFACE |
---|
[1344] | 2643 | INTERFACE mpp_minloc |
---|
| 2644 | MODULE PROCEDURE mpp_minloc2d ,mpp_minloc3d |
---|
| 2645 | END INTERFACE |
---|
| 2646 | INTERFACE mpp_maxloc |
---|
| 2647 | MODULE PROCEDURE mpp_maxloc2d ,mpp_maxloc3d |
---|
| 2648 | END INTERFACE |
---|
[3] | 2649 | |
---|
[13] | 2650 | LOGICAL, PUBLIC, PARAMETER :: lk_mpp = .FALSE. !: mpp flag |
---|
[3294] | 2651 | LOGICAL, PUBLIC :: ln_nnogather = .FALSE. !: namelist control of northfold comms (needed here in case "key_mpp_mpi" is not used) |
---|
[869] | 2652 | INTEGER :: ncomm_ice |
---|
[2715] | 2653 | !!---------------------------------------------------------------------- |
---|
[13] | 2654 | CONTAINS |
---|
[3] | 2655 | |
---|
[2715] | 2656 | INTEGER FUNCTION lib_mpp_alloc(kumout) ! Dummy function |
---|
| 2657 | INTEGER, INTENT(in) :: kumout |
---|
| 2658 | lib_mpp_alloc = 0 |
---|
| 2659 | END FUNCTION lib_mpp_alloc |
---|
| 2660 | |
---|
| 2661 | FUNCTION mynode( ldtxt, kumnam, kstop, localComm ) RESULT (function_value) |
---|
[1579] | 2662 | INTEGER, OPTIONAL , INTENT(in ) :: localComm |
---|
[2715] | 2663 | CHARACTER(len=*),DIMENSION(:) :: ldtxt |
---|
| 2664 | INTEGER :: kumnam, kstop |
---|
[1559] | 2665 | IF( PRESENT( localComm ) .OR. .NOT.PRESENT( localComm ) ) function_value = 0 |
---|
[1579] | 2666 | IF( .FALSE. ) ldtxt(:) = 'never done' |
---|
[13] | 2667 | END FUNCTION mynode |
---|
[3] | 2668 | |
---|
[13] | 2669 | SUBROUTINE mppsync ! Dummy routine |
---|
| 2670 | END SUBROUTINE mppsync |
---|
[3] | 2671 | |
---|
[869] | 2672 | SUBROUTINE mpp_sum_as( parr, kdim, kcom ) ! Dummy routine |
---|
[13] | 2673 | REAL , DIMENSION(:) :: parr |
---|
| 2674 | INTEGER :: kdim |
---|
[869] | 2675 | INTEGER, OPTIONAL :: kcom |
---|
| 2676 | WRITE(*,*) 'mpp_sum_as: You should not have seen this print! error?', kdim, parr(1), kcom |
---|
[13] | 2677 | END SUBROUTINE mpp_sum_as |
---|
[3] | 2678 | |
---|
[869] | 2679 | SUBROUTINE mpp_sum_a2s( parr, kdim, kcom ) ! Dummy routine |
---|
[13] | 2680 | REAL , DIMENSION(:,:) :: parr |
---|
| 2681 | INTEGER :: kdim |
---|
[869] | 2682 | INTEGER, OPTIONAL :: kcom |
---|
| 2683 | WRITE(*,*) 'mpp_sum_a2s: You should not have seen this print! error?', kdim, parr(1,1), kcom |
---|
[13] | 2684 | END SUBROUTINE mpp_sum_a2s |
---|
[3] | 2685 | |
---|
[869] | 2686 | SUBROUTINE mpp_sum_ai( karr, kdim, kcom ) ! Dummy routine |
---|
[13] | 2687 | INTEGER, DIMENSION(:) :: karr |
---|
| 2688 | INTEGER :: kdim |
---|
[869] | 2689 | INTEGER, OPTIONAL :: kcom |
---|
| 2690 | WRITE(*,*) 'mpp_sum_ai: You should not have seen this print! error?', kdim, karr(1), kcom |
---|
[13] | 2691 | END SUBROUTINE mpp_sum_ai |
---|
[3] | 2692 | |
---|
[869] | 2693 | SUBROUTINE mpp_sum_s( psca, kcom ) ! Dummy routine |
---|
[13] | 2694 | REAL :: psca |
---|
[869] | 2695 | INTEGER, OPTIONAL :: kcom |
---|
| 2696 | WRITE(*,*) 'mpp_sum_s: You should not have seen this print! error?', psca, kcom |
---|
[13] | 2697 | END SUBROUTINE mpp_sum_s |
---|
[2480] | 2698 | |
---|
[869] | 2699 | SUBROUTINE mpp_sum_i( kint, kcom ) ! Dummy routine |
---|
[13] | 2700 | integer :: kint |
---|
[2480] | 2701 | INTEGER, OPTIONAL :: kcom |
---|
[869] | 2702 | WRITE(*,*) 'mpp_sum_i: You should not have seen this print! error?', kint, kcom |
---|
[13] | 2703 | END SUBROUTINE mpp_sum_i |
---|
| 2704 | |
---|
[3294] | 2705 | SUBROUTINE mppsum_realdd( ytab, kcom ) |
---|
| 2706 | COMPLEX(wp), INTENT(inout) :: ytab ! input scalar |
---|
| 2707 | INTEGER , INTENT( in ), OPTIONAL :: kcom |
---|
| 2708 | WRITE(*,*) 'mppsum_realdd: You should not have seen this print! error?', ytab |
---|
| 2709 | END SUBROUTINE mppsum_realdd |
---|
| 2710 | |
---|
| 2711 | SUBROUTINE mppsum_a_realdd( ytab, kdim, kcom ) |
---|
| 2712 | INTEGER , INTENT( in ) :: kdim ! size of ytab |
---|
| 2713 | COMPLEX(wp), DIMENSION(kdim), INTENT( inout ) :: ytab ! input array |
---|
| 2714 | INTEGER , INTENT( in ), OPTIONAL :: kcom |
---|
| 2715 | WRITE(*,*) 'mppsum_a_realdd: You should not have seen this print! error?', kdim, ytab(1), kcom |
---|
| 2716 | END SUBROUTINE mppsum_a_realdd |
---|
| 2717 | |
---|
[869] | 2718 | SUBROUTINE mppmax_a_real( parr, kdim, kcom ) |
---|
[13] | 2719 | REAL , DIMENSION(:) :: parr |
---|
| 2720 | INTEGER :: kdim |
---|
[869] | 2721 | INTEGER, OPTIONAL :: kcom |
---|
| 2722 | WRITE(*,*) 'mppmax_a_real: You should not have seen this print! error?', kdim, parr(1), kcom |
---|
[13] | 2723 | END SUBROUTINE mppmax_a_real |
---|
| 2724 | |
---|
[869] | 2725 | SUBROUTINE mppmax_real( psca, kcom ) |
---|
[13] | 2726 | REAL :: psca |
---|
[869] | 2727 | INTEGER, OPTIONAL :: kcom |
---|
| 2728 | WRITE(*,*) 'mppmax_real: You should not have seen this print! error?', psca, kcom |
---|
[13] | 2729 | END SUBROUTINE mppmax_real |
---|
| 2730 | |
---|
[869] | 2731 | SUBROUTINE mppmin_a_real( parr, kdim, kcom ) |
---|
[13] | 2732 | REAL , DIMENSION(:) :: parr |
---|
| 2733 | INTEGER :: kdim |
---|
[869] | 2734 | INTEGER, OPTIONAL :: kcom |
---|
| 2735 | WRITE(*,*) 'mppmin_a_real: You should not have seen this print! error?', kdim, parr(1), kcom |
---|
[13] | 2736 | END SUBROUTINE mppmin_a_real |
---|
| 2737 | |
---|
[869] | 2738 | SUBROUTINE mppmin_real( psca, kcom ) |
---|
[13] | 2739 | REAL :: psca |
---|
[869] | 2740 | INTEGER, OPTIONAL :: kcom |
---|
| 2741 | WRITE(*,*) 'mppmin_real: You should not have seen this print! error?', psca, kcom |
---|
[13] | 2742 | END SUBROUTINE mppmin_real |
---|
| 2743 | |
---|
[869] | 2744 | SUBROUTINE mppmax_a_int( karr, kdim ,kcom) |
---|
[681] | 2745 | INTEGER, DIMENSION(:) :: karr |
---|
| 2746 | INTEGER :: kdim |
---|
[869] | 2747 | INTEGER, OPTIONAL :: kcom |
---|
[888] | 2748 | WRITE(*,*) 'mppmax_a_int: You should not have seen this print! error?', kdim, karr(1), kcom |
---|
[681] | 2749 | END SUBROUTINE mppmax_a_int |
---|
| 2750 | |
---|
[869] | 2751 | SUBROUTINE mppmax_int( kint, kcom) |
---|
[681] | 2752 | INTEGER :: kint |
---|
[869] | 2753 | INTEGER, OPTIONAL :: kcom |
---|
| 2754 | WRITE(*,*) 'mppmax_int: You should not have seen this print! error?', kint, kcom |
---|
[681] | 2755 | END SUBROUTINE mppmax_int |
---|
| 2756 | |
---|
[869] | 2757 | SUBROUTINE mppmin_a_int( karr, kdim, kcom ) |
---|
[13] | 2758 | INTEGER, DIMENSION(:) :: karr |
---|
| 2759 | INTEGER :: kdim |
---|
[869] | 2760 | INTEGER, OPTIONAL :: kcom |
---|
| 2761 | WRITE(*,*) 'mppmin_a_int: You should not have seen this print! error?', kdim, karr(1), kcom |
---|
[13] | 2762 | END SUBROUTINE mppmin_a_int |
---|
| 2763 | |
---|
[869] | 2764 | SUBROUTINE mppmin_int( kint, kcom ) |
---|
[13] | 2765 | INTEGER :: kint |
---|
[869] | 2766 | INTEGER, OPTIONAL :: kcom |
---|
| 2767 | WRITE(*,*) 'mppmin_int: You should not have seen this print! error?', kint, kcom |
---|
[13] | 2768 | END SUBROUTINE mppmin_int |
---|
| 2769 | |
---|
[2715] | 2770 | SUBROUTINE mppobc_1d( parr, kd1, kd2, kl, kk, ktype, kij, knum ) |
---|
| 2771 | INTEGER :: kd1, kd2, kl , kk, ktype, kij, knum |
---|
[1344] | 2772 | REAL, DIMENSION(:) :: parr ! variable array |
---|
[2715] | 2773 | WRITE(*,*) 'mppobc: You should not have seen this print! error?', parr(1), kd1, kd2, kl, kk, ktype, kij, knum |
---|
[13] | 2774 | END SUBROUTINE mppobc_1d |
---|
| 2775 | |
---|
[2715] | 2776 | SUBROUTINE mppobc_2d( parr, kd1, kd2, kl, kk, ktype, kij, knum ) |
---|
| 2777 | INTEGER :: kd1, kd2, kl , kk, ktype, kij, knum |
---|
[1344] | 2778 | REAL, DIMENSION(:,:) :: parr ! variable array |
---|
[2715] | 2779 | WRITE(*,*) 'mppobc: You should not have seen this print! error?', parr(1,1), kd1, kd2, kl, kk, ktype, kij, knum |
---|
[13] | 2780 | END SUBROUTINE mppobc_2d |
---|
| 2781 | |
---|
[2715] | 2782 | SUBROUTINE mppobc_3d( parr, kd1, kd2, kl, kk, ktype, kij, knum ) |
---|
| 2783 | INTEGER :: kd1, kd2, kl , kk, ktype, kij, knum |
---|
[1344] | 2784 | REAL, DIMENSION(:,:,:) :: parr ! variable array |
---|
[2715] | 2785 | WRITE(*,*) 'mppobc: You should not have seen this print! error?', parr(1,1,1), kd1, kd2, kl, kk, ktype, kij, knum |
---|
[13] | 2786 | END SUBROUTINE mppobc_3d |
---|
| 2787 | |
---|
[2715] | 2788 | SUBROUTINE mppobc_4d( parr, kd1, kd2, kl, kk, ktype, kij, knum ) |
---|
| 2789 | INTEGER :: kd1, kd2, kl , kk, ktype, kij, knum |
---|
[1344] | 2790 | REAL, DIMENSION(:,:,:,:) :: parr ! variable array |
---|
[2715] | 2791 | WRITE(*,*) 'mppobc: You should not have seen this print! error?', parr(1,1,1,1), kd1, kd2, kl, kk, ktype, kij, knum |
---|
[13] | 2792 | END SUBROUTINE mppobc_4d |
---|
| 2793 | |
---|
[1344] | 2794 | SUBROUTINE mpp_minloc2d( ptab, pmask, pmin, ki, kj ) |
---|
[181] | 2795 | REAL :: pmin |
---|
| 2796 | REAL , DIMENSION (:,:) :: ptab, pmask |
---|
| 2797 | INTEGER :: ki, kj |
---|
[1528] | 2798 | WRITE(*,*) 'mpp_minloc2d: You should not have seen this print! error?', pmin, ki, kj, ptab(1,1), pmask(1,1) |
---|
[181] | 2799 | END SUBROUTINE mpp_minloc2d |
---|
| 2800 | |
---|
[1344] | 2801 | SUBROUTINE mpp_minloc3d( ptab, pmask, pmin, ki, kj, kk ) |
---|
[181] | 2802 | REAL :: pmin |
---|
| 2803 | REAL , DIMENSION (:,:,:) :: ptab, pmask |
---|
| 2804 | INTEGER :: ki, kj, kk |
---|
[1528] | 2805 | WRITE(*,*) 'mpp_minloc3d: You should not have seen this print! error?', pmin, ki, kj, kk, ptab(1,1,1), pmask(1,1,1) |
---|
[181] | 2806 | END SUBROUTINE mpp_minloc3d |
---|
| 2807 | |
---|
[1344] | 2808 | SUBROUTINE mpp_maxloc2d( ptab, pmask, pmax, ki, kj ) |
---|
[181] | 2809 | REAL :: pmax |
---|
| 2810 | REAL , DIMENSION (:,:) :: ptab, pmask |
---|
| 2811 | INTEGER :: ki, kj |
---|
[1528] | 2812 | WRITE(*,*) 'mpp_maxloc2d: You should not have seen this print! error?', pmax, ki, kj, ptab(1,1), pmask(1,1) |
---|
[181] | 2813 | END SUBROUTINE mpp_maxloc2d |
---|
| 2814 | |
---|
[1344] | 2815 | SUBROUTINE mpp_maxloc3d( ptab, pmask, pmax, ki, kj, kk ) |
---|
[181] | 2816 | REAL :: pmax |
---|
| 2817 | REAL , DIMENSION (:,:,:) :: ptab, pmask |
---|
| 2818 | INTEGER :: ki, kj, kk |
---|
[1528] | 2819 | WRITE(*,*) 'mpp_maxloc3d: You should not have seen this print! error?', pmax, ki, kj, kk, ptab(1,1,1), pmask(1,1,1) |
---|
[181] | 2820 | END SUBROUTINE mpp_maxloc3d |
---|
| 2821 | |
---|
[51] | 2822 | SUBROUTINE mppstop |
---|
[3294] | 2823 | WRITE(*,*) 'mppstop: You should not have seen this print if running in mpp mode! error?...' |
---|
| 2824 | WRITE(*,*) 'mppstop: ..otherwise this is a stop condition raised by ctl_stop in single processor mode' |
---|
| 2825 | STOP |
---|
[51] | 2826 | END SUBROUTINE mppstop |
---|
| 2827 | |
---|
[2715] | 2828 | SUBROUTINE mpp_ini_ice( kcom, knum ) |
---|
| 2829 | INTEGER :: kcom, knum |
---|
| 2830 | WRITE(*,*) 'mpp_ini_ice: You should not have seen this print! error?', kcom, knum |
---|
[888] | 2831 | END SUBROUTINE mpp_ini_ice |
---|
[869] | 2832 | |
---|
[2715] | 2833 | SUBROUTINE mpp_ini_znl( knum ) |
---|
| 2834 | INTEGER :: knum |
---|
| 2835 | WRITE(*,*) 'mpp_ini_znl: You should not have seen this print! error?', knum |
---|
[1345] | 2836 | END SUBROUTINE mpp_ini_znl |
---|
| 2837 | |
---|
[1344] | 2838 | SUBROUTINE mpp_comm_free( kcom ) |
---|
[869] | 2839 | INTEGER :: kcom |
---|
[1344] | 2840 | WRITE(*,*) 'mpp_comm_free: You should not have seen this print! error?', kcom |
---|
[869] | 2841 | END SUBROUTINE mpp_comm_free |
---|
[3] | 2842 | #endif |
---|
[2715] | 2843 | |
---|
[13] | 2844 | !!---------------------------------------------------------------------- |
---|
[2715] | 2845 | !! All cases: ctl_stop, ctl_warn, get_unit, ctl_opn routines |
---|
| 2846 | !!---------------------------------------------------------------------- |
---|
| 2847 | |
---|
| 2848 | SUBROUTINE ctl_stop( cd1, cd2, cd3, cd4, cd5 , & |
---|
| 2849 | & cd6, cd7, cd8, cd9, cd10 ) |
---|
| 2850 | !!---------------------------------------------------------------------- |
---|
| 2851 | !! *** ROUTINE stop_opa *** |
---|
| 2852 | !! |
---|
| 2853 | !! ** Purpose : print in ocean.outpput file a error message and |
---|
| 2854 | !! increment the error number (nstop) by one. |
---|
| 2855 | !!---------------------------------------------------------------------- |
---|
| 2856 | CHARACTER(len=*), INTENT(in), OPTIONAL :: cd1, cd2, cd3, cd4, cd5 |
---|
| 2857 | CHARACTER(len=*), INTENT(in), OPTIONAL :: cd6, cd7, cd8, cd9, cd10 |
---|
| 2858 | !!---------------------------------------------------------------------- |
---|
| 2859 | ! |
---|
| 2860 | nstop = nstop + 1 |
---|
| 2861 | IF(lwp) THEN |
---|
| 2862 | WRITE(numout,cform_err) |
---|
| 2863 | IF( PRESENT(cd1 ) ) WRITE(numout,*) cd1 |
---|
| 2864 | IF( PRESENT(cd2 ) ) WRITE(numout,*) cd2 |
---|
| 2865 | IF( PRESENT(cd3 ) ) WRITE(numout,*) cd3 |
---|
| 2866 | IF( PRESENT(cd4 ) ) WRITE(numout,*) cd4 |
---|
| 2867 | IF( PRESENT(cd5 ) ) WRITE(numout,*) cd5 |
---|
| 2868 | IF( PRESENT(cd6 ) ) WRITE(numout,*) cd6 |
---|
| 2869 | IF( PRESENT(cd7 ) ) WRITE(numout,*) cd7 |
---|
| 2870 | IF( PRESENT(cd8 ) ) WRITE(numout,*) cd8 |
---|
| 2871 | IF( PRESENT(cd9 ) ) WRITE(numout,*) cd9 |
---|
| 2872 | IF( PRESENT(cd10) ) WRITE(numout,*) cd10 |
---|
| 2873 | ENDIF |
---|
| 2874 | CALL FLUSH(numout ) |
---|
| 2875 | IF( numstp /= -1 ) CALL FLUSH(numstp ) |
---|
| 2876 | IF( numsol /= -1 ) CALL FLUSH(numsol ) |
---|
| 2877 | IF( numevo_ice /= -1 ) CALL FLUSH(numevo_ice) |
---|
| 2878 | ! |
---|
| 2879 | IF( cd1 == 'STOP' ) THEN |
---|
| 2880 | IF(lwp) WRITE(numout,*) 'huge E-R-R-O-R : immediate stop' |
---|
| 2881 | CALL mppstop() |
---|
| 2882 | ENDIF |
---|
| 2883 | ! |
---|
| 2884 | END SUBROUTINE ctl_stop |
---|
| 2885 | |
---|
| 2886 | |
---|
| 2887 | SUBROUTINE ctl_warn( cd1, cd2, cd3, cd4, cd5, & |
---|
| 2888 | & cd6, cd7, cd8, cd9, cd10 ) |
---|
| 2889 | !!---------------------------------------------------------------------- |
---|
| 2890 | !! *** ROUTINE stop_warn *** |
---|
| 2891 | !! |
---|
| 2892 | !! ** Purpose : print in ocean.outpput file a error message and |
---|
| 2893 | !! increment the warning number (nwarn) by one. |
---|
| 2894 | !!---------------------------------------------------------------------- |
---|
| 2895 | CHARACTER(len=*), INTENT(in), OPTIONAL :: cd1, cd2, cd3, cd4, cd5 |
---|
| 2896 | CHARACTER(len=*), INTENT(in), OPTIONAL :: cd6, cd7, cd8, cd9, cd10 |
---|
| 2897 | !!---------------------------------------------------------------------- |
---|
| 2898 | ! |
---|
| 2899 | nwarn = nwarn + 1 |
---|
| 2900 | IF(lwp) THEN |
---|
| 2901 | WRITE(numout,cform_war) |
---|
| 2902 | IF( PRESENT(cd1 ) ) WRITE(numout,*) cd1 |
---|
| 2903 | IF( PRESENT(cd2 ) ) WRITE(numout,*) cd2 |
---|
| 2904 | IF( PRESENT(cd3 ) ) WRITE(numout,*) cd3 |
---|
| 2905 | IF( PRESENT(cd4 ) ) WRITE(numout,*) cd4 |
---|
| 2906 | IF( PRESENT(cd5 ) ) WRITE(numout,*) cd5 |
---|
| 2907 | IF( PRESENT(cd6 ) ) WRITE(numout,*) cd6 |
---|
| 2908 | IF( PRESENT(cd7 ) ) WRITE(numout,*) cd7 |
---|
| 2909 | IF( PRESENT(cd8 ) ) WRITE(numout,*) cd8 |
---|
| 2910 | IF( PRESENT(cd9 ) ) WRITE(numout,*) cd9 |
---|
| 2911 | IF( PRESENT(cd10) ) WRITE(numout,*) cd10 |
---|
| 2912 | ENDIF |
---|
| 2913 | CALL FLUSH(numout) |
---|
| 2914 | ! |
---|
| 2915 | END SUBROUTINE ctl_warn |
---|
| 2916 | |
---|
| 2917 | |
---|
| 2918 | SUBROUTINE ctl_opn( knum, cdfile, cdstat, cdform, cdacce, klengh, kout, ldwp, karea ) |
---|
| 2919 | !!---------------------------------------------------------------------- |
---|
| 2920 | !! *** ROUTINE ctl_opn *** |
---|
| 2921 | !! |
---|
| 2922 | !! ** Purpose : Open file and check if required file is available. |
---|
| 2923 | !! |
---|
| 2924 | !! ** Method : Fortan open |
---|
| 2925 | !!---------------------------------------------------------------------- |
---|
| 2926 | INTEGER , INTENT( out) :: knum ! logical unit to open |
---|
| 2927 | CHARACTER(len=*) , INTENT(in ) :: cdfile ! file name to open |
---|
| 2928 | CHARACTER(len=*) , INTENT(in ) :: cdstat ! disposition specifier |
---|
| 2929 | CHARACTER(len=*) , INTENT(in ) :: cdform ! formatting specifier |
---|
| 2930 | CHARACTER(len=*) , INTENT(in ) :: cdacce ! access specifier |
---|
| 2931 | INTEGER , INTENT(in ) :: klengh ! record length |
---|
| 2932 | INTEGER , INTENT(in ) :: kout ! number of logical units for write |
---|
| 2933 | LOGICAL , INTENT(in ) :: ldwp ! boolean term for print |
---|
| 2934 | INTEGER, OPTIONAL, INTENT(in ) :: karea ! proc number |
---|
| 2935 | !! |
---|
| 2936 | CHARACTER(len=80) :: clfile |
---|
| 2937 | INTEGER :: iost |
---|
| 2938 | !!---------------------------------------------------------------------- |
---|
| 2939 | |
---|
| 2940 | ! adapt filename |
---|
| 2941 | ! ---------------- |
---|
| 2942 | clfile = TRIM(cdfile) |
---|
| 2943 | IF( PRESENT( karea ) ) THEN |
---|
| 2944 | IF( karea > 1 ) WRITE(clfile, "(a,'_',i4.4)") TRIM(clfile), karea-1 |
---|
| 2945 | ENDIF |
---|
| 2946 | #if defined key_agrif |
---|
| 2947 | IF( .NOT. Agrif_Root() ) clfile = TRIM(Agrif_CFixed())//'_'//TRIM(clfile) |
---|
| 2948 | knum=Agrif_Get_Unit() |
---|
| 2949 | #else |
---|
| 2950 | knum=get_unit() |
---|
| 2951 | #endif |
---|
| 2952 | |
---|
| 2953 | iost=0 |
---|
| 2954 | IF( cdacce(1:6) == 'DIRECT' ) THEN |
---|
| 2955 | OPEN( UNIT=knum, FILE=clfile, FORM=cdform, ACCESS=cdacce, STATUS=cdstat, RECL=klengh, ERR=100, IOSTAT=iost ) |
---|
| 2956 | ELSE |
---|
| 2957 | OPEN( UNIT=knum, FILE=clfile, FORM=cdform, ACCESS=cdacce, STATUS=cdstat , ERR=100, IOSTAT=iost ) |
---|
| 2958 | ENDIF |
---|
| 2959 | IF( iost == 0 ) THEN |
---|
| 2960 | IF(ldwp) THEN |
---|
| 2961 | WRITE(kout,*) ' file : ', clfile,' open ok' |
---|
| 2962 | WRITE(kout,*) ' unit = ', knum |
---|
| 2963 | WRITE(kout,*) ' status = ', cdstat |
---|
| 2964 | WRITE(kout,*) ' form = ', cdform |
---|
| 2965 | WRITE(kout,*) ' access = ', cdacce |
---|
| 2966 | WRITE(kout,*) |
---|
| 2967 | ENDIF |
---|
| 2968 | ENDIF |
---|
| 2969 | 100 CONTINUE |
---|
| 2970 | IF( iost /= 0 ) THEN |
---|
| 2971 | IF(ldwp) THEN |
---|
| 2972 | WRITE(kout,*) |
---|
| 2973 | WRITE(kout,*) ' ===>>>> : bad opening file: ', clfile |
---|
| 2974 | WRITE(kout,*) ' ======= === ' |
---|
| 2975 | WRITE(kout,*) ' unit = ', knum |
---|
| 2976 | WRITE(kout,*) ' status = ', cdstat |
---|
| 2977 | WRITE(kout,*) ' form = ', cdform |
---|
| 2978 | WRITE(kout,*) ' access = ', cdacce |
---|
| 2979 | WRITE(kout,*) ' iostat = ', iost |
---|
| 2980 | WRITE(kout,*) ' we stop. verify the file ' |
---|
| 2981 | WRITE(kout,*) |
---|
| 2982 | ENDIF |
---|
| 2983 | STOP 'ctl_opn bad opening' |
---|
| 2984 | ENDIF |
---|
| 2985 | |
---|
| 2986 | END SUBROUTINE ctl_opn |
---|
| 2987 | |
---|
| 2988 | |
---|
| 2989 | INTEGER FUNCTION get_unit() |
---|
| 2990 | !!---------------------------------------------------------------------- |
---|
| 2991 | !! *** FUNCTION get_unit *** |
---|
| 2992 | !! |
---|
| 2993 | !! ** Purpose : return the index of an unused logical unit |
---|
| 2994 | !!---------------------------------------------------------------------- |
---|
| 2995 | LOGICAL :: llopn |
---|
| 2996 | !!---------------------------------------------------------------------- |
---|
| 2997 | ! |
---|
| 2998 | get_unit = 15 ! choose a unit that is big enough then it is not already used in NEMO |
---|
| 2999 | llopn = .TRUE. |
---|
| 3000 | DO WHILE( (get_unit < 998) .AND. llopn ) |
---|
| 3001 | get_unit = get_unit + 1 |
---|
| 3002 | INQUIRE( unit = get_unit, opened = llopn ) |
---|
| 3003 | END DO |
---|
| 3004 | IF( (get_unit == 999) .AND. llopn ) THEN |
---|
| 3005 | CALL ctl_stop( 'get_unit: All logical units until 999 are used...' ) |
---|
| 3006 | get_unit = -1 |
---|
| 3007 | ENDIF |
---|
| 3008 | ! |
---|
| 3009 | END FUNCTION get_unit |
---|
| 3010 | |
---|
| 3011 | !!---------------------------------------------------------------------- |
---|
[3] | 3012 | END MODULE lib_mpp |
---|