New URL for NEMO forge! http://forge.nemo-ocean.eu

Since March 2022 along with NEMO 4.2 release, the code development moved to a self-hosted GitLab.
This present forge is now archived and remained online for history.

lib_mpp.F90 in NEMO/branches/2018/dev_r10164_HPC09_ESIWACE_PREP_MERGE/src/OCE/LBC – NEMO

Context Navigation

source: NEMO/branches/2018/dev_r10164_HPC09_ESIWACE_PREP_MERGE/src/OCE/LBC/lib_mpp.F90 @ 10407

Last change on this file since 10407 was 10407, checked in by smasson, 5 years ago
dev_r10164_HPC09_ESIWACE_PREP_MERGE: send back error code, see #2181
Property svn:keywords set to `Id`
File size: 88.0 KB

Rev	Line
[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
[9019]	10	!! NEMO 1.0 ! 2003 (J.M. Molines, G. Madec) F90, free form
[1344]	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
[3764]	19	!! 3.2 ! 2009 (O. Marti) add mpp_ini_znl
[2715]	20	!! 4.0 ! 2011 (G. Madec) move ctl_ routines from in_out_manager
[9019]	21	!! 3.5 ! 2012 (S.Mocavero, I. Epicoco) Add mpp_lnk_bdy_3d/2d routines to optimize the BDY comm.
	22	!! 3.5 ! 2013 (C. Ethe, G. Madec) message passing arrays as local variables
[6140]	23	!! 3.5 ! 2013 (S.Mocavero, I.Epicoco - CMCC) north fold optimizations
[9019]	24	!! 3.6 ! 2015 (O. Tintó and M. Castrillo - BSC) Added '_multiple' case for 2D lbc and max
	25	!! 4.0 ! 2017 (G. Madec) automatique allocation of array argument (use any 3rd dimension)
	26	!! - ! 2017 (G. Madec) create generic.h90 files to generate all lbc and north fold routines
[13]	27	!!----------------------------------------------------------------------
[2715]	28
	29	!!----------------------------------------------------------------------
[6140]	30	!! ctl_stop : update momentum and tracer Kz from a tke scheme
	31	!! ctl_warn : initialization, namelist read, and parameters control
	32	!! ctl_opn : Open file and check if required file is available.
	33	!! ctl_nam : Prints informations when an error occurs while reading a namelist
	34	!! get_unit : give the index of an unused logical unit
[2715]	35	!!----------------------------------------------------------------------
[3764]	36	#if defined key_mpp_mpi
[13]	37	!!----------------------------------------------------------------------
[1344]	38	!! 'key_mpp_mpi' MPI massively parallel processing library
	39	!!----------------------------------------------------------------------
[2715]	40	!! lib_mpp_alloc : allocate mpp arrays
	41	!! mynode : indentify the processor unit
	42	!! mpp_lnk : interface (defined in lbclnk) for message passing of 2d or 3d arrays (mpp_lnk_2d, mpp_lnk_3d)
[4990]	43	!! mpp_lnk_icb : interface for message passing of 2d arrays with extra halo for icebergs (mpp_lnk_2d_icb)
[6140]	44	!! mpprecv :
[9019]	45	!! mppsend :
[2715]	46	!! mppscatter :
	47	!! mppgather :
	48	!! mpp_min : generic interface for mppmin_int , mppmin_a_int , mppmin_real, mppmin_a_real
	49	!! mpp_max : generic interface for mppmax_int , mppmax_a_int , mppmax_real, mppmax_a_real
	50	!! mpp_sum : generic interface for mppsum_int , mppsum_a_int , mppsum_real, mppsum_a_real
	51	!! mpp_minloc :
	52	!! mpp_maxloc :
	53	!! mppsync :
	54	!! mppstop :
[1344]	55	!! mpp_ini_north : initialisation of north fold
[9019]	56	!! mpp_lbc_north_icb : alternative to mpp_nfd for extra outer halo with icebergs
[13]	57	!!----------------------------------------------------------------------
[3764]	58	USE dom_oce ! ocean space and time domain
[2715]	59	USE lbcnfd ! north fold treatment
	60	USE in_out_manager ! I/O manager
[3]	61
[13]	62	IMPLICIT NONE
[415]	63	PRIVATE
[9019]	64
	65	INTERFACE mpp_nfd
[10368]	66	MODULE PROCEDURE mpp_nfd_2d , mpp_nfd_3d , mpp_nfd_4d
[9019]	67	MODULE PROCEDURE mpp_nfd_2d_ptr, mpp_nfd_3d_ptr, mpp_nfd_4d_ptr
	68	END INTERFACE
	69
	70	! Interface associated to the mpp_lnk_... routines is defined in lbclnk
[10368]	71	PUBLIC mpp_lnk_2d , mpp_lnk_3d , mpp_lnk_4d
[9019]	72	PUBLIC mpp_lnk_2d_ptr, mpp_lnk_3d_ptr, mpp_lnk_4d_ptr
	73	!
	74	!!gm this should be useless
	75	PUBLIC mpp_nfd_2d , mpp_nfd_3d , mpp_nfd_4d
	76	PUBLIC mpp_nfd_2d_ptr, mpp_nfd_3d_ptr, mpp_nfd_4d_ptr
	77	!!gm end
	78	!
[4147]	79	PUBLIC ctl_stop, ctl_warn, get_unit, ctl_opn, ctl_nam
[1344]	80	PUBLIC mynode, mppstop, mppsync, mpp_comm_free
[9019]	81	PUBLIC mpp_ini_north
	82	PUBLIC mpp_lnk_2d_icb
	83	PUBLIC mpp_lbc_north_icb
[1344]	84	PUBLIC mpp_min, mpp_max, mpp_sum, mpp_minloc, mpp_maxloc
[10397]	85	PUBLIC mpp_delay_max, mpp_delay_sum, mpp_delay_rcv
[3294]	86	PUBLIC mppscatter, mppgather
[10180]	87	PUBLIC mpp_ini_znl
[3764]	88	PUBLIC mppsend, mpprecv ! needed by TAM and ICB routines
[9890]	89	PUBLIC mpp_lnk_bdy_2d, mpp_lnk_bdy_3d, mpp_lnk_bdy_4d
[5429]	90
[13]	91	!! * Interfaces
	92	!! define generic interface for these routine as they are called sometimes
[1344]	93	!! with scalar arguments instead of array arguments, which causes problems
	94	!! for the compilation on AIX system as well as NEC and SGI. Ok on COMPACQ
[13]	95	INTERFACE mpp_min
	96	MODULE PROCEDURE mppmin_a_int, mppmin_int, mppmin_a_real, mppmin_real
	97	END INTERFACE
	98	INTERFACE mpp_max
[681]	99	MODULE PROCEDURE mppmax_a_int, mppmax_int, mppmax_a_real, mppmax_real
[13]	100	END INTERFACE
	101	INTERFACE mpp_sum
[6140]	102	MODULE PROCEDURE mppsum_a_int, mppsum_int, mppsum_a_real, mppsum_real, &
[9019]	103	& mppsum_realdd, mppsum_a_realdd
[13]	104	END INTERFACE
[1344]	105	INTERFACE mpp_minloc
	106	MODULE PROCEDURE mpp_minloc2d ,mpp_minloc3d
	107	END INTERFACE
	108	INTERFACE mpp_maxloc
	109	MODULE PROCEDURE mpp_maxloc2d ,mpp_maxloc3d
	110	END INTERFACE
[6490]	111
[51]	112	!! ========================= !!
	113	!! MPI variable definition !!
	114	!! ========================= !!
[1629]	115	!$AGRIF_DO_NOT_TREAT
[2004]	116	INCLUDE 'mpif.h'
[1629]	117	!$AGRIF_END_DO_NOT_TREAT
[3764]	118
[1344]	119	LOGICAL, PUBLIC, PARAMETER :: lk_mpp = .TRUE. !: mpp flag
[3]	120
[1344]	121	INTEGER, PARAMETER :: nprocmax = 2**10 ! maximun dimension (required to be a power of 2)
[3764]	122
[10330]	123	INTEGER, PUBLIC :: mppsize ! number of process
	124	INTEGER, PUBLIC :: mpprank ! process number [ 0 - size-1 ]
[2363]	125	!$AGRIF_DO_NOT_TREAT
[9570]	126	INTEGER, PUBLIC :: mpi_comm_oce ! opa local communicator
[2363]	127	!$AGRIF_END_DO_NOT_TREAT
[3]	128
[2480]	129	INTEGER :: MPI_SUMDD
[1976]	130
[1345]	131	! variables used for zonal integration
	132	INTEGER, PUBLIC :: ncomm_znl !: communicator made by the processors on the same zonal average
[9019]	133	LOGICAL, PUBLIC :: l_znl_root !: True on the 'left'most processor on the same row
	134	INTEGER :: ngrp_znl ! group ID for the znl processors
	135	INTEGER :: ndim_rank_znl ! number of processors on the same zonal average
[2715]	136	INTEGER, DIMENSION(:), ALLOCATABLE, SAVE :: nrank_znl ! dimension ndim_rank_znl, number of the procs into the same znl domain
[3]	137
[3764]	138	! North fold condition in mpp_mpi with jpni > 1 (PUBLIC for TAM)
[9019]	139	INTEGER, PUBLIC :: ngrp_world !: group ID for the world processors
	140	INTEGER, PUBLIC :: ngrp_opa !: group ID for the opa processors
	141	INTEGER, PUBLIC :: ngrp_north !: group ID for the northern processors (to be fold)
	142	INTEGER, PUBLIC :: ncomm_north !: communicator made by the processors belonging to ngrp_north
	143	INTEGER, PUBLIC :: ndim_rank_north !: number of 'sea' processor in the northern line (can be /= jpni !)
	144	INTEGER, PUBLIC :: njmppmax !: value of njmpp for the processors of the northern line
	145	INTEGER, PUBLIC :: north_root !: number (in the comm_opa) of proc 0 in the northern comm
	146	INTEGER, PUBLIC, DIMENSION(:), ALLOCATABLE, SAVE :: nrank_north !: dimension ndim_rank_north
[3764]	147
[1344]	148	! Type of send : standard, buffered, immediate
[9019]	149	CHARACTER(len=1), PUBLIC :: cn_mpi_send !: type od mpi send/recieve (S=standard, B=bsend, I=isend)
	150	LOGICAL , PUBLIC :: l_isend = .FALSE. !: isend use indicator (T if cn_mpi_send='I')
	151	INTEGER , PUBLIC :: nn_buffer !: size of the buffer in case of mpi_bsend
[3764]	152
[10170]	153	! Communications summary report
[10297]	154	CHARACTER(len=128), DIMENSION(:), ALLOCATABLE :: crname_lbc !: names of lbc_lnk calling routines
	155	CHARACTER(len=128), DIMENSION(:), ALLOCATABLE :: crname_glb !: names of global comm calling routines
[10170]	156	INTEGER, PUBLIC :: ncom_stp = 0 !: copy of time step # istp
[10297]	157	INTEGER, PUBLIC :: ncom_fsbc = 1 !: copy of sbc time step # nn_fsbc
	158	INTEGER, PUBLIC :: ncom_dttrc = 1 !: copy of top time step # nn_dttrc
[10314]	159	INTEGER, PUBLIC :: ncom_freq !: frequency of comm diagnostic
[10170]	160	INTEGER, PUBLIC , DIMENSION(:,:), ALLOCATABLE :: ncomm_sequence !: size of communicated arrays (halos)
[10314]	161	INTEGER, PARAMETER, PUBLIC :: ncom_rec_max = 2000 !: max number of communication record
[10297]	162	INTEGER, PUBLIC :: n_sequence_lbc = 0 !: # of communicated arraysvia lbc
	163	INTEGER, PUBLIC :: n_sequence_glb = 0 !: # of global communications
	164	INTEGER, PUBLIC :: numcom = -1 !: logical unit for communicaton report
[10386]	165	LOGICAL, PUBLIC :: l_full_nf_update = .TRUE. !: logical for a full (2lines) update of bc at North fold report
[10397]	166	INTEGER, PARAMETER, PUBLIC :: nbdelay = 2 !: number of delayed operations
	167	!: name (used as id) of allreduce-delayed operations
	168	CHARACTER(len=32), DIMENSION(nbdelay), PUBLIC :: c_delaylist = (/ 'cflice', 'fwb' /)
	169	!: component name where the allreduce-delayed operation is performed
	170	CHARACTER(len=3), DIMENSION(nbdelay), PUBLIC :: c_delaycpnt = (/ 'ICE' , 'OCE' /)
	171	TYPE, PUBLIC :: DELAYARR
	172	REAL( wp), POINTER, DIMENSION(:) :: z1d => NULL()
	173	COMPLEX(wp), POINTER, DIMENSION(:) :: y1d => NULL()
	174	END TYPE DELAYARR
	175	TYPE( DELAYARR ), DIMENSION(nbdelay), PUBLIC :: todelay
	176	INTEGER, DIMENSION(nbdelay), PUBLIC :: ndelayid = -1 !: mpi request id of the delayed operations
[10297]	177
[10172]	178	! timing summary report
[10300]	179	REAL(wp), DIMENSION(2), PUBLIC :: waiting_time = 0._wp
	180	REAL(wp) , PUBLIC :: compute_time = 0._wp, elapsed_time = 0._wp
[10172]	181
[9019]	182	REAL(wp), DIMENSION(:), ALLOCATABLE, SAVE :: tampon ! buffer in case of bsend
[3]	183
[9019]	184	LOGICAL, PUBLIC :: ln_nnogather !: namelist control of northfold comms
	185	LOGICAL, PUBLIC :: l_north_nogather = .FALSE. !: internal control of northfold comms
	186
[51]	187	!!----------------------------------------------------------------------
[9598]	188	!! NEMO/OCE 4.0 , NEMO Consortium (2018)
[888]	189	!! $Id$
[10068]	190	!! Software governed by the CeCILL license (see ./LICENSE)
[1344]	191	!!----------------------------------------------------------------------
[3]	192	CONTAINS
	193
[9019]	194	FUNCTION mynode( ldtxt, ldname, kumnam_ref, kumnam_cfg, kumond, kstop, localComm )
[2715]	195	!!----------------------------------------------------------------------
[51]	196	!! * routine mynode *
[3764]	197	!!
[51]	198	!! ** Purpose : Find processor unit
	199	!!----------------------------------------------------------------------
[6140]	200	CHARACTER(len=*),DIMENSION(:), INTENT( out) :: ldtxt !
	201	CHARACTER(len=*) , INTENT(in ) :: ldname !
	202	INTEGER , INTENT(in ) :: kumnam_ref ! logical unit for reference namelist
	203	INTEGER , INTENT(in ) :: kumnam_cfg ! logical unit for configuration namelist
	204	INTEGER , INTENT(inout) :: kumond ! logical unit for namelist output
	205	INTEGER , INTENT(inout) :: kstop ! stop indicator
	206	INTEGER , OPTIONAL , INTENT(in ) :: localComm !
[2715]	207	!
[4147]	208	INTEGER :: mynode, ierr, code, ji, ii, ios
[532]	209	LOGICAL :: mpi_was_called
[2715]	210	!
[3294]	211	NAMELIST/nammpp/ cn_mpi_send, nn_buffer, jpni, jpnj, jpnij, ln_nnogather
[51]	212	!!----------------------------------------------------------------------
[1344]	213	!
[2481]	214	ii = 1
[6140]	215	WRITE(ldtxt(ii),*) ; ii = ii + 1
	216	WRITE(ldtxt(ii),*) 'mynode : mpi initialisation' ; ii = ii + 1
	217	WRITE(ldtxt(ii),*) '~~~~~~ ' ; ii = ii + 1
[1344]	218	!
[4147]	219	REWIND( kumnam_ref ) ! Namelist nammpp in reference namelist: mpi variables
	220	READ ( kumnam_ref, nammpp, IOSTAT = ios, ERR = 901)
[9168]	221	901 IF( ios /= 0 ) CALL ctl_nam ( ios , 'nammpp in reference namelist', lwp )
[9019]	222	!
[4147]	223	REWIND( kumnam_cfg ) ! Namelist nammpp in configuration namelist: mpi variables
	224	READ ( kumnam_cfg, nammpp, IOSTAT = ios, ERR = 902 )
[9168]	225	902 IF( ios > 0 ) CALL ctl_nam ( ios , 'nammpp in configuration namelist', lwp )
[9019]	226	!
[1344]	227	! ! control print
[6140]	228	WRITE(ldtxt(ii),*) ' Namelist nammpp' ; ii = ii + 1
	229	WRITE(ldtxt(ii),*) ' mpi send type cn_mpi_send = ', cn_mpi_send ; ii = ii + 1
	230	WRITE(ldtxt(ii),*) ' size exported buffer nn_buffer = ', nn_buffer,' bytes'; ii = ii + 1
[9019]	231	!
	232	IF( jpni < 1 .OR. jpnj < 1 ) THEN
[6140]	233	WRITE(ldtxt(ii),*) ' jpni, jpnj and jpnij will be calculated automatically' ; ii = ii + 1
[2715]	234	ELSE
[6140]	235	WRITE(ldtxt(ii),*) ' processor grid extent in i jpni = ',jpni ; ii = ii + 1
	236	WRITE(ldtxt(ii),*) ' processor grid extent in j jpnj = ',jpnj ; ii = ii + 1
[9019]	237	ENDIF
[2715]	238
[3294]	239	WRITE(ldtxt(ii),*) ' avoid use of mpi_allgather at the north fold ln_nnogather = ', ln_nnogather ; ii = ii + 1
	240
[2480]	241	CALL mpi_initialized ( mpi_was_called, code )
	242	IF( code /= MPI_SUCCESS ) THEN
[3764]	243	DO ji = 1, SIZE(ldtxt)
[2481]	244	IF( TRIM(ldtxt(ji)) /= '' ) WRITE(,) ldtxt(ji) ! control print of mynode
[3764]	245	END DO
[2480]	246	WRITE(*, cform_err)
	247	WRITE(, ) 'lib_mpp: Error in routine mpi_initialized'
	248	CALL mpi_abort( mpi_comm_world, code, ierr )
	249	ENDIF
[415]	250
[2480]	251	IF( mpi_was_called ) THEN
	252	!
	253	SELECT CASE ( cn_mpi_send )
	254	CASE ( 'S' ) ! Standard mpi send (blocking)
[6140]	255	WRITE(ldtxt(ii),*) ' Standard blocking mpi send (send)' ; ii = ii + 1
[2480]	256	CASE ( 'B' ) ! Buffer mpi send (blocking)
[6140]	257	WRITE(ldtxt(ii),*) ' Buffer blocking mpi send (bsend)' ; ii = ii + 1
[9570]	258	IF( Agrif_Root() ) CALL mpi_init_oce( ldtxt, ii, ierr )
[2480]	259	CASE ( 'I' ) ! Immediate mpi send (non-blocking send)
[6140]	260	WRITE(ldtxt(ii),*) ' Immediate non-blocking send (isend)' ; ii = ii + 1
[2480]	261	l_isend = .TRUE.
	262	CASE DEFAULT
[6140]	263	WRITE(ldtxt(ii),cform_err) ; ii = ii + 1
	264	WRITE(ldtxt(ii),*) ' bad value for cn_mpi_send = ', cn_mpi_send ; ii = ii + 1
[2715]	265	kstop = kstop + 1
[2480]	266	END SELECT
[9019]	267	!
	268	ELSEIF ( PRESENT(localComm) .AND. .NOT. mpi_was_called ) THEN
[6140]	269	WRITE(ldtxt(ii),*) ' lib_mpp: You cannot provide a local communicator ' ; ii = ii + 1
	270	WRITE(ldtxt(ii),*) ' without calling MPI_Init before ! ' ; ii = ii + 1
[2715]	271	kstop = kstop + 1
[532]	272	ELSE
[1601]	273	SELECT CASE ( cn_mpi_send )
[524]	274	CASE ( 'S' ) ! Standard mpi send (blocking)
[6140]	275	WRITE(ldtxt(ii),*) ' Standard blocking mpi send (send)' ; ii = ii + 1
[2480]	276	CALL mpi_init( ierr )
[524]	277	CASE ( 'B' ) ! Buffer mpi send (blocking)
[6140]	278	WRITE(ldtxt(ii),*) ' Buffer blocking mpi send (bsend)' ; ii = ii + 1
[9570]	279	IF( Agrif_Root() ) CALL mpi_init_oce( ldtxt, ii, ierr )
[524]	280	CASE ( 'I' ) ! Immediate mpi send (non-blocking send)
[6140]	281	WRITE(ldtxt(ii),*) ' Immediate non-blocking send (isend)' ; ii = ii + 1
[524]	282	l_isend = .TRUE.
[2480]	283	CALL mpi_init( ierr )
[524]	284	CASE DEFAULT
[6140]	285	WRITE(ldtxt(ii),cform_err) ; ii = ii + 1
	286	WRITE(ldtxt(ii),*) ' bad value for cn_mpi_send = ', cn_mpi_send ; ii = ii + 1
[2715]	287	kstop = kstop + 1
[524]	288	END SELECT
[2480]	289	!
[415]	290	ENDIF
[570]	291
[3764]	292	IF( PRESENT(localComm) ) THEN
[2480]	293	IF( Agrif_Root() ) THEN
[9570]	294	mpi_comm_oce = localComm
[2480]	295	ENDIF
	296	ELSE
[9570]	297	CALL mpi_comm_dup( mpi_comm_world, mpi_comm_oce, code)
[2480]	298	IF( code /= MPI_SUCCESS ) THEN
[3764]	299	DO ji = 1, SIZE(ldtxt)
[2481]	300	IF( TRIM(ldtxt(ji)) /= '' ) WRITE(,) ldtxt(ji) ! control print of mynode
	301	END DO
[2480]	302	WRITE(*, cform_err)
	303	WRITE(, ) ' lib_mpp: Error in routine mpi_comm_dup'
	304	CALL mpi_abort( mpi_comm_world, code, ierr )
	305	ENDIF
[3764]	306	ENDIF
[2480]	307
[5656]	308	#if defined key_agrif
[9019]	309	IF( Agrif_Root() ) THEN
[9570]	310	CALL Agrif_MPI_Init(mpi_comm_oce)
[5656]	311	ELSE
[9570]	312	CALL Agrif_MPI_set_grid_comm(mpi_comm_oce)
[5656]	313	ENDIF
	314	#endif
	315
[9570]	316	CALL mpi_comm_rank( mpi_comm_oce, mpprank, ierr )
	317	CALL mpi_comm_size( mpi_comm_oce, mppsize, ierr )
[629]	318	mynode = mpprank
[4624]	319
	320	IF( mynode == 0 ) THEN
[5407]	321	CALL ctl_opn( kumond, TRIM(ldname), 'UNKNOWN', 'FORMATTED', 'SEQUENTIAL', -1, 6, .FALSE. , 1 )
	322	WRITE(kumond, nammpp)
[4624]	323	ENDIF
[3764]	324	!
[1976]	325	CALL MPI_OP_CREATE(DDPDD_MPI, .TRUE., MPI_SUMDD, ierr)
	326	!
[51]	327	END FUNCTION mynode
[3]	328
[9019]	329	!!----------------------------------------------------------------------
	330	!! * routine mpp_lnk_(2,3,4)d *
	331	!!
	332	!! * Argument : dummy argument use in mpp_lnk_... routines
	333	!! ptab : array or pointer of arrays on which the boundary condition is applied
	334	!! cd_nat : nature of array grid-points
	335	!! psgn : sign used across the north fold boundary
	336	!! kfld : optional, number of pt3d arrays
	337	!! cd_mpp : optional, fill the overlap area only
	338	!! pval : optional, background value (used at closed boundaries)
	339	!!----------------------------------------------------------------------
	340	!
	341	! !== 2D array and array of 2D pointer ==!
	342	!
	343	# define DIM_2d
	344	# define ROUTINE_LNK mpp_lnk_2d
	345	# include "mpp_lnk_generic.h90"
	346	# undef ROUTINE_LNK
	347	# define MULTI
	348	# define ROUTINE_LNK mpp_lnk_2d_ptr
	349	# include "mpp_lnk_generic.h90"
	350	# undef ROUTINE_LNK
	351	# undef MULTI
	352	# undef DIM_2d
	353	!
	354	! !== 3D array and array of 3D pointer ==!
	355	!
	356	# define DIM_3d
	357	# define ROUTINE_LNK mpp_lnk_3d
	358	# include "mpp_lnk_generic.h90"
	359	# undef ROUTINE_LNK
	360	# define MULTI
	361	# define ROUTINE_LNK mpp_lnk_3d_ptr
	362	# include "mpp_lnk_generic.h90"
	363	# undef ROUTINE_LNK
	364	# undef MULTI
	365	# undef DIM_3d
	366	!
	367	! !== 4D array and array of 4D pointer ==!
	368	!
	369	# define DIM_4d
	370	# define ROUTINE_LNK mpp_lnk_4d
	371	# include "mpp_lnk_generic.h90"
	372	# undef ROUTINE_LNK
	373	# define MULTI
	374	# define ROUTINE_LNK mpp_lnk_4d_ptr
	375	# include "mpp_lnk_generic.h90"
	376	# undef ROUTINE_LNK
	377	# undef MULTI
	378	# undef DIM_4d
[6140]	379
[9019]	380	!!----------------------------------------------------------------------
	381	!! * routine mpp_nfd_(2,3,4)d *
	382	!!
	383	!! * Argument : dummy argument use in mpp_nfd_... routines
	384	!! ptab : array or pointer of arrays on which the boundary condition is applied
	385	!! cd_nat : nature of array grid-points
	386	!! psgn : sign used across the north fold boundary
	387	!! kfld : optional, number of pt3d arrays
	388	!! cd_mpp : optional, fill the overlap area only
	389	!! pval : optional, background value (used at closed boundaries)
	390	!!----------------------------------------------------------------------
	391	!
	392	! !== 2D array and array of 2D pointer ==!
	393	!
	394	# define DIM_2d
	395	# define ROUTINE_NFD mpp_nfd_2d
	396	# include "mpp_nfd_generic.h90"
	397	# undef ROUTINE_NFD
	398	# define MULTI
	399	# define ROUTINE_NFD mpp_nfd_2d_ptr
	400	# include "mpp_nfd_generic.h90"
	401	# undef ROUTINE_NFD
	402	# undef MULTI
	403	# undef DIM_2d
	404	!
	405	! !== 3D array and array of 3D pointer ==!
	406	!
	407	# define DIM_3d
	408	# define ROUTINE_NFD mpp_nfd_3d
	409	# include "mpp_nfd_generic.h90"
	410	# undef ROUTINE_NFD
	411	# define MULTI
	412	# define ROUTINE_NFD mpp_nfd_3d_ptr
	413	# include "mpp_nfd_generic.h90"
	414	# undef ROUTINE_NFD
	415	# undef MULTI
	416	# undef DIM_3d
	417	!
	418	! !== 4D array and array of 4D pointer ==!
	419	!
	420	# define DIM_4d
	421	# define ROUTINE_NFD mpp_nfd_4d
	422	# include "mpp_nfd_generic.h90"
	423	# undef ROUTINE_NFD
	424	# define MULTI
	425	# define ROUTINE_NFD mpp_nfd_4d_ptr
	426	# include "mpp_nfd_generic.h90"
	427	# undef ROUTINE_NFD
	428	# undef MULTI
	429	# undef DIM_4d
[3]	430
[1344]	431
[9019]	432	!!----------------------------------------------------------------------
	433	!! * routine mpp_lnk_bdy_(2,3,4)d *
	434	!!
	435	!! * Argument : dummy argument use in mpp_lnk_... routines
	436	!! ptab : array or pointer of arrays on which the boundary condition is applied
	437	!! cd_nat : nature of array grid-points
	438	!! psgn : sign used across the north fold boundary
	439	!! kb_bdy : BDY boundary set
	440	!! kfld : optional, number of pt3d arrays
	441	!!----------------------------------------------------------------------
	442	!
	443	! !== 2D array and array of 2D pointer ==!
	444	!
	445	# define DIM_2d
	446	# define ROUTINE_BDY mpp_lnk_bdy_2d
	447	# include "mpp_bdy_generic.h90"
	448	# undef ROUTINE_BDY
	449	# undef DIM_2d
	450	!
	451	! !== 3D array and array of 3D pointer ==!
	452	!
	453	# define DIM_3d
	454	# define ROUTINE_BDY mpp_lnk_bdy_3d
	455	# include "mpp_bdy_generic.h90"
	456	# undef ROUTINE_BDY
	457	# undef DIM_3d
	458	!
	459	! !== 4D array and array of 4D pointer ==!
	460	!
[9890]	461	# define DIM_4d
	462	# define ROUTINE_BDY mpp_lnk_bdy_4d
	463	# include "mpp_bdy_generic.h90"
	464	# undef ROUTINE_BDY
	465	# undef DIM_4d
[3]	466
[9019]	467	!!----------------------------------------------------------------------
	468	!!
	469	!! load_array & mpp_lnk_2d_9 à generaliser a 3D et 4D
[5429]	470
	471
[9019]	472	!! mpp_lnk_sum_2d et 3D ====>>>>>> à virer du code !!!!
[5429]	473
[9019]	474
	475	!!----------------------------------------------------------------------
[5429]	476
	477
[888]	478
[1344]	479	SUBROUTINE mppsend( ktyp, pmess, kbytes, kdest, md_req )
[51]	480	!!----------------------------------------------------------------------
	481	!! * routine mppsend *
[3764]	482	!!
[51]	483	!! ** Purpose : Send messag passing array
	484	!!
	485	!!----------------------------------------------------------------------
[1344]	486	REAL(wp), INTENT(inout) :: pmess(*) ! array of real
	487	INTEGER , INTENT(in ) :: kbytes ! size of the array pmess
	488	INTEGER , INTENT(in ) :: kdest ! receive process number
	489	INTEGER , INTENT(in ) :: ktyp ! tag of the message
	490	INTEGER , INTENT(in ) :: md_req ! argument for isend
	491	!!
	492	INTEGER :: iflag
[51]	493	!!----------------------------------------------------------------------
[1344]	494	!
[1601]	495	SELECT CASE ( cn_mpi_send )
[300]	496	CASE ( 'S' ) ! Standard mpi send (blocking)
[9570]	497	CALL mpi_send ( pmess, kbytes, mpi_double_precision, kdest , ktyp, mpi_comm_oce , iflag )
[300]	498	CASE ( 'B' ) ! Buffer mpi send (blocking)
[9570]	499	CALL mpi_bsend( pmess, kbytes, mpi_double_precision, kdest , ktyp, mpi_comm_oce , iflag )
[300]	500	CASE ( 'I' ) ! Immediate mpi send (non-blocking send)
[1344]	501	! be carefull, one more argument here : the mpi request identifier..
[9570]	502	CALL mpi_isend( pmess, kbytes, mpi_double_precision, kdest , ktyp, mpi_comm_oce, md_req, iflag )
[300]	503	END SELECT
[1344]	504	!
[51]	505	END SUBROUTINE mppsend
[3]	506
	507
[3294]	508	SUBROUTINE mpprecv( ktyp, pmess, kbytes, ksource )
[51]	509	!!----------------------------------------------------------------------
	510	!! * routine mpprecv *
	511	!!
	512	!! ** Purpose : Receive messag passing array
	513	!!
	514	!!----------------------------------------------------------------------
[1344]	515	REAL(wp), INTENT(inout) :: pmess(*) ! array of real
	516	INTEGER , INTENT(in ) :: kbytes ! suze of the array pmess
	517	INTEGER , INTENT(in ) :: ktyp ! Tag of the recevied message
[3764]	518	INTEGER, OPTIONAL, INTENT(in) :: ksource ! source process number
[1344]	519	!!
[51]	520	INTEGER :: istatus(mpi_status_size)
	521	INTEGER :: iflag
[3294]	522	INTEGER :: use_source
[1344]	523	!!----------------------------------------------------------------------
	524	!
[3764]	525	! If a specific process number has been passed to the receive call,
[3294]	526	! use that one. Default is to use mpi_any_source
[6140]	527	use_source = mpi_any_source
	528	IF( PRESENT(ksource) ) use_source = ksource
	529	!
[9570]	530	CALL mpi_recv( pmess, kbytes, mpi_double_precision, use_source, ktyp, mpi_comm_oce, istatus, iflag )
[1344]	531	!
[51]	532	END SUBROUTINE mpprecv
[3]	533
	534
[51]	535	SUBROUTINE mppgather( ptab, kp, pio )
	536	!!----------------------------------------------------------------------
	537	!! * routine mppgather *
[3764]	538	!!
	539	!! ** Purpose : Transfert between a local subdomain array and a work
[51]	540	!! array which is distributed following the vertical level.
	541	!!
[1344]	542	!!----------------------------------------------------------------------
[6140]	543	REAL(wp), DIMENSION(jpi,jpj) , INTENT(in ) :: ptab ! subdomain input array
	544	INTEGER , INTENT(in ) :: kp ! record length
[1344]	545	REAL(wp), DIMENSION(jpi,jpj,jpnij), INTENT( out) :: pio ! subdomain input array
[51]	546	!!
[1344]	547	INTEGER :: itaille, ierror ! temporary integer
[51]	548	!!---------------------------------------------------------------------
[1344]	549	!
	550	itaille = jpi * jpj
	551	CALL mpi_gather( ptab, itaille, mpi_double_precision, pio, itaille , &
[9570]	552	& mpi_double_precision, kp , mpi_comm_oce, ierror )
[1344]	553	!
[51]	554	END SUBROUTINE mppgather
[3]	555
	556
[51]	557	SUBROUTINE mppscatter( pio, kp, ptab )
	558	!!----------------------------------------------------------------------
	559	!! * routine mppscatter *
	560	!!
[3764]	561	!! ** Purpose : Transfert between awork array which is distributed
[51]	562	!! following the vertical level and the local subdomain array.
	563	!!
	564	!!----------------------------------------------------------------------
[6140]	565	REAL(wp), DIMENSION(jpi,jpj,jpnij) :: pio ! output array
	566	INTEGER :: kp ! Tag (not used with MPI
	567	REAL(wp), DIMENSION(jpi,jpj) :: ptab ! subdomain array input
[1344]	568	!!
	569	INTEGER :: itaille, ierror ! temporary integer
[51]	570	!!---------------------------------------------------------------------
[1344]	571	!
[6140]	572	itaille = jpi * jpj
[1344]	573	!
	574	CALL mpi_scatter( pio, itaille, mpi_double_precision, ptab, itaille , &
[9570]	575	& mpi_double_precision, kp , mpi_comm_oce, ierror )
[1344]	576	!
[51]	577	END SUBROUTINE mppscatter
[3]	578
[10397]	579
	580	SUBROUTINE mpp_delay_sum( cdname, cdelay, y_in, pout, ldlast, kcom )
	581	!!----------------------------------------------------------------------
	582	!! * routine mpp_delay_sum *
[10359]	583	!!
[10397]	584	!! ** Purpose : performed delayed mpp_sum, the result is received on next call
[10359]	585	!!
[10292]	586	!!----------------------------------------------------------------------
[10359]	587	CHARACTER(len=*), INTENT(in ) :: cdname ! name of the calling subroutine
	588	CHARACTER(len=*), INTENT(in ) :: cdelay ! name (used as id) of the delayed operation
[10397]	589	COMPLEX(wp), INTENT(in ), DIMENSION(:) :: y_in
	590	REAL(wp), INTENT( out), DIMENSION(:) :: pout
[10359]	591	LOGICAL, INTENT(in ) :: ldlast ! true if this is the last time we call this routine
	592	INTEGER, INTENT(in ), OPTIONAL :: kcom
	593	!!
[10397]	594	INTEGER :: ji, isz
[10359]	595	INTEGER :: idvar
[10397]	596	INTEGER :: ierr, ilocalcomm
	597	COMPLEX(wp), ALLOCATABLE, DIMENSION(:) :: ytmp
[10359]	598	!!----------------------------------------------------------------------
[10292]	599	ilocalcomm = mpi_comm_oce
[10359]	600	IF( PRESENT(kcom) ) ilocalcomm = kcom
	601
[10397]	602	isz = SIZE(y_in)
	603
[10359]	604	IF( narea == 1 .AND. numcom == -1 ) CALL mpp_report( cdname, ld_glb = .TRUE. )
	605
	606	idvar = -1
	607	DO ji = 1, nbdelay
[10397]	608	IF( TRIM(cdelay) == TRIM(c_delaylist(ji)) ) idvar = ji
[10359]	609	END DO
[10397]	610	IF ( idvar == -1 ) CALL ctl_stop( 'STOP',' mpp_delay_sum : please add a new delayed exchange for '//TRIM(cdname) )
	611
	612	IF ( ndelayid(idvar) == 0 ) THEN ! first call with restart: %z1d defined in iom_delay_rst
	613	! --------------------------
	614	IF ( SIZE(todelay(idvar)%z1d) /= isz ) THEN ! Check dimension coherence
	615	IF(lwp) WRITE(numout,*) ' WARNING: the nb of delayed variables in restart file is not the model one'
	616	DEALLOCATE(todelay(idvar)%z1d)
	617	ndelayid(idvar) = -1 ! do as if we had no restart
	618	ELSE
	619	ALLOCATE(todelay(idvar)%y1d(isz))
	620	todelay(idvar)%y1d(:) = CMPLX(todelay(idvar)%z1d(:), 0., wp) ! create %y1d, complex variable needed by mpi_sumdd
	621	END IF
	622	ENDIF
	623
	624	IF( ndelayid(idvar) == -1 ) THEN ! first call without restart: define %y1d and %z1d from y_in with blocking allreduce
	625	! --------------------------
	626	ALLOCATE(todelay(idvar)%z1d(isz), todelay(idvar)%y1d(isz)) ! allocate also %z1d as used for the restart
	627	CALL mpi_allreduce( y_in(:), todelay(idvar)%y1d(:), isz, MPI_DOUBLE_COMPLEX, mpi_sumdd, ilocalcomm, ierr ) ! get %y1d
	628	todelay(idvar)%z1d(:) = REAL(todelay(idvar)%y1d(:), wp) ! define %z1d from %y1d
	629	ENDIF
	630
	631	IF( ndelayid(idvar) > 0 ) CALL mpp_delay_rcv( idvar ) ! make sure %z1d is received
	632
	633	! send back pout from todelay(idvar)%z1d defined at previous call
	634	pout(:) = todelay(idvar)%z1d(:)
	635
	636	! send y_in into todelay(idvar)%y1d with a non-blocking communication
	637	CALL mpi_iallreduce( y_in(:), todelay(idvar)%y1d(:), isz, MPI_DOUBLE_COMPLEX, mpi_sumdd, ilocalcomm, ndelayid(idvar), ierr )
	638
	639	END SUBROUTINE mpp_delay_sum
	640
	641
	642	SUBROUTINE mpp_delay_max( cdname, cdelay, p_in, pout, ldlast, kcom )
	643	!!----------------------------------------------------------------------
	644	!! * routine mpp_delay_max *
	645	!!
	646	!! ** Purpose : performed delayed mpp_max, the result is received on next call
	647	!!
	648	!!----------------------------------------------------------------------
	649	CHARACTER(len=*), INTENT(in ) :: cdname ! name of the calling subroutine
	650	CHARACTER(len=*), INTENT(in ) :: cdelay ! name (used as id) of the delayed operation
	651	REAL(wp), INTENT(in ), DIMENSION(:) :: p_in !
	652	REAL(wp), INTENT( out), DIMENSION(:) :: pout !
	653	LOGICAL, INTENT(in ) :: ldlast ! true if this is the last time we call this routine
	654	INTEGER, INTENT(in ), OPTIONAL :: kcom
	655	!!
	656	INTEGER :: ji, isz
	657	INTEGER :: idvar
	658	INTEGER :: ierr, ilocalcomm
	659	!!----------------------------------------------------------------------
	660	ilocalcomm = mpi_comm_oce
	661	IF( PRESENT(kcom) ) ilocalcomm = kcom
	662
	663	isz = SIZE(p_in)
	664
	665	IF( narea == 1 .AND. numcom == -1 ) CALL mpp_report( cdname, ld_glb = .TRUE. )
	666
	667	idvar = -1
	668	DO ji = 1, nbdelay
	669	IF( TRIM(cdelay) == TRIM(c_delaylist(ji)) ) idvar = ji
	670	END DO
[10359]	671	IF ( idvar == -1 ) CALL ctl_stop( 'STOP',' mpp_delay_max : please add a new delayed exchange for '//TRIM(cdname) )
	672
[10397]	673	IF ( ndelayid(idvar) == 0 ) THEN ! first call with restart: %z1d defined in iom_delay_rst
	674	! --------------------------
	675	IF ( SIZE(todelay(idvar)%z1d) /= isz ) THEN ! Check dimension coherence
	676	IF(lwp) WRITE(numout,*) ' WARNING: the nb of delayed variables in restart file is not the model one'
	677	DEALLOCATE(todelay(idvar)%z1d)
	678	ndelayid(idvar) = -1 ! do as if we had no restart
	679	END IF
[10292]	680	ENDIF
[681]	681
[10397]	682	IF( ndelayid(idvar) == -1 ) THEN ! first call without restart: define %z1d from p_in with a blocking allreduce
	683	! --------------------------
	684	ALLOCATE(todelay(idvar)%z1d(isz))
	685	CALL mpi_allreduce( p_in(:), todelay(idvar)%z1d(:), isz, MPI_DOUBLE_PRECISION, mpi_max, ilocalcomm, ierr ) ! get %z1d
[10359]	686	ENDIF
	687
[10397]	688	IF( ndelayid(idvar) > 0 ) CALL mpp_delay_rcv( idvar ) ! make sure %z1d is received
	689
	690	! send back pout from todelay(idvar)%z1d defined at previous call
	691	pout(:) = todelay(idvar)%z1d(:)
	692
	693	! send p_in into todelay(idvar)%z1d with a non-blocking communication
	694	CALL mpi_iallreduce( p_in(:), todelay(idvar)%z1d(:), isz, MPI_DOUBLE_PRECISION, mpi_max, ilocalcomm, ndelayid(idvar), ierr )
	695
[10359]	696	END SUBROUTINE mpp_delay_max
[10397]	697
[10297]	698
[10397]	699	SUBROUTINE mpp_delay_rcv( kid )
	700	!!----------------------------------------------------------------------
	701	!! * routine mpp_delay_rcv *
	702	!!
	703	!! ** Purpose : force barrier for delayed mpp (needed for restart)
	704	!!
	705	!!----------------------------------------------------------------------
	706	INTEGER,INTENT(in ) :: kid
	707	INTEGER :: ierr
	708	!!----------------------------------------------------------------------
	709	IF( ndelayid(kid) /= -2 ) THEN
	710	IF( ln_timing ) CALL tic_tac( .TRUE., ld_global = .TRUE.)
	711	CALL mpi_wait( ndelayid(kid), MPI_STATUS_IGNORE, ierr ) ! make sure todelay(kid) is received
	712	IF( ln_timing ) CALL tic_tac(.FALSE., ld_global = .TRUE.)
	713	IF( ASSOCIATED(todelay(kid)%y1d) ) todelay(kid)%z1d(:) = REAL(todelay(kid)%y1d(:), wp) ! define %z1d from %y1d
	714	ndelayid(kid) = -2 ! add flag to know that mpi_wait was already called on kid
	715	ENDIF
	716	END SUBROUTINE mpp_delay_rcv
	717
	718
[9019]	719	!!----------------------------------------------------------------------
[10297]	720	!! * mppmax_a_int, mppmax_int, mppmax_a_real, mppmax_real *
	721	!!
	722	!!----------------------------------------------------------------------
	723	!!
	724	# define OPERATION_MAX
	725	# define INTEGER_TYPE
	726	# define DIM_0d
	727	# define ROUTINE_ALLREDUCE mppmax_int
	728	# include "mpp_allreduce_generic.h90"
	729	# undef ROUTINE_ALLREDUCE
	730	# undef DIM_0d
	731	# define DIM_1d
	732	# define ROUTINE_ALLREDUCE mppmax_a_int
	733	# include "mpp_allreduce_generic.h90"
	734	# undef ROUTINE_ALLREDUCE
	735	# undef DIM_1d
	736	# undef INTEGER_TYPE
	737	!
	738	# define REAL_TYPE
	739	# define DIM_0d
	740	# define ROUTINE_ALLREDUCE mppmax_real
	741	# include "mpp_allreduce_generic.h90"
	742	# undef ROUTINE_ALLREDUCE
	743	# undef DIM_0d
	744	# define DIM_1d
	745	# define ROUTINE_ALLREDUCE mppmax_a_real
	746	# include "mpp_allreduce_generic.h90"
	747	# undef ROUTINE_ALLREDUCE
	748	# undef DIM_1d
	749	# undef REAL_TYPE
	750	# undef OPERATION_MAX
	751	!!----------------------------------------------------------------------
[9019]	752	!! * mppmin_a_int, mppmin_int, mppmin_a_real, mppmin_real *
	753	!!
	754	!!----------------------------------------------------------------------
	755	!!
[10297]	756	# define OPERATION_MIN
	757	# define INTEGER_TYPE
	758	# define DIM_0d
	759	# define ROUTINE_ALLREDUCE mppmin_int
	760	# include "mpp_allreduce_generic.h90"
	761	# undef ROUTINE_ALLREDUCE
	762	# undef DIM_0d
	763	# define DIM_1d
	764	# define ROUTINE_ALLREDUCE mppmin_a_int
	765	# include "mpp_allreduce_generic.h90"
	766	# undef ROUTINE_ALLREDUCE
	767	# undef DIM_1d
	768	# undef INTEGER_TYPE
	769	!
	770	# define REAL_TYPE
	771	# define DIM_0d
	772	# define ROUTINE_ALLREDUCE mppmin_real
	773	# include "mpp_allreduce_generic.h90"
	774	# undef ROUTINE_ALLREDUCE
	775	# undef DIM_0d
	776	# define DIM_1d
	777	# define ROUTINE_ALLREDUCE mppmin_a_real
	778	# include "mpp_allreduce_generic.h90"
	779	# undef ROUTINE_ALLREDUCE
	780	# undef DIM_1d
	781	# undef REAL_TYPE
	782	# undef OPERATION_MIN
[3]	783
[9019]	784	!!----------------------------------------------------------------------
	785	!! * mppsum_a_int, mppsum_int, mppsum_a_real, mppsum_real *
	786	!!
	787	!! Global sum of 1D array or a variable (integer, real or complex)
	788	!!----------------------------------------------------------------------
	789	!!
[10297]	790	# define OPERATION_SUM
	791	# define INTEGER_TYPE
	792	# define DIM_0d
	793	# define ROUTINE_ALLREDUCE mppsum_int
	794	# include "mpp_allreduce_generic.h90"
	795	# undef ROUTINE_ALLREDUCE
	796	# undef DIM_0d
	797	# define DIM_1d
	798	# define ROUTINE_ALLREDUCE mppsum_a_int
	799	# include "mpp_allreduce_generic.h90"
	800	# undef ROUTINE_ALLREDUCE
	801	# undef DIM_1d
	802	# undef INTEGER_TYPE
	803	!
	804	# define REAL_TYPE
	805	# define DIM_0d
	806	# define ROUTINE_ALLREDUCE mppsum_real
	807	# include "mpp_allreduce_generic.h90"
	808	# undef ROUTINE_ALLREDUCE
	809	# undef DIM_0d
	810	# define DIM_1d
	811	# define ROUTINE_ALLREDUCE mppsum_a_real
	812	# include "mpp_allreduce_generic.h90"
	813	# undef ROUTINE_ALLREDUCE
	814	# undef DIM_1d
	815	# undef REAL_TYPE
	816	# undef OPERATION_SUM
[3764]	817
[10297]	818	# define OPERATION_SUM_DD
	819	# define COMPLEX_TYPE
	820	# define DIM_0d
	821	# define ROUTINE_ALLREDUCE mppsum_realdd
	822	# include "mpp_allreduce_generic.h90"
	823	# undef ROUTINE_ALLREDUCE
	824	# undef DIM_0d
	825	# define DIM_1d
	826	# define ROUTINE_ALLREDUCE mppsum_a_realdd
	827	# include "mpp_allreduce_generic.h90"
	828	# undef ROUTINE_ALLREDUCE
	829	# undef DIM_1d
	830	# undef COMPLEX_TYPE
	831	# undef OPERATION_SUM_DD
	832
[10314]	833	!!----------------------------------------------------------------------
	834	!! *** mpp_minloc2d, mpp_minloc3d, mpp_maxloc2d, mpp_maxloc3d
	835	!!
	836	!!----------------------------------------------------------------------
	837	!!
	838	# define OPERATION_MINLOC
	839	# define DIM_2d
	840	# define ROUTINE_LOC mpp_minloc2d
	841	# include "mpp_loc_generic.h90"
	842	# undef ROUTINE_LOC
	843	# undef DIM_2d
	844	# define DIM_3d
	845	# define ROUTINE_LOC mpp_minloc3d
	846	# include "mpp_loc_generic.h90"
	847	# undef ROUTINE_LOC
	848	# undef DIM_3d
	849	# undef OPERATION_MINLOC
[6140]	850
[10314]	851	# define OPERATION_MAXLOC
	852	# define DIM_2d
	853	# define ROUTINE_LOC mpp_maxloc2d
	854	# include "mpp_loc_generic.h90"
	855	# undef ROUTINE_LOC
	856	# undef DIM_2d
	857	# define DIM_3d
	858	# define ROUTINE_LOC mpp_maxloc3d
	859	# include "mpp_loc_generic.h90"
	860	# undef ROUTINE_LOC
	861	# undef DIM_3d
	862	# undef OPERATION_MAXLOC
[9019]	863
[1344]	864	SUBROUTINE mppsync()
	865	!!----------------------------------------------------------------------
	866	!! * routine mppsync *
[3764]	867	!!
[1344]	868	!! ** Purpose : Massively parallel processors, synchroneous
	869	!!
	870	!!-----------------------------------------------------------------------
	871	INTEGER :: ierror
	872	!!-----------------------------------------------------------------------
	873	!
[9570]	874	CALL mpi_barrier( mpi_comm_oce, ierror )
[1344]	875	!
	876	END SUBROUTINE mppsync
[3]	877
	878
[10358]	879	SUBROUTINE mppstop( ldfinal, ld_force_abort )
[1344]	880	!!----------------------------------------------------------------------
	881	!! * routine mppstop *
[3764]	882	!!
[3294]	883	!! ** purpose : Stop massively parallel processors method
[1344]	884	!!
	885	!!----------------------------------------------------------------------
[10330]	886	LOGICAL, OPTIONAL, INTENT(in) :: ldfinal ! source process number
[10358]	887	LOGICAL, OPTIONAL, INTENT(in) :: ld_force_abort ! source process number
	888	LOGICAL :: llfinal, ll_force_abort
[1344]	889	INTEGER :: info
	890	!!----------------------------------------------------------------------
[10330]	891	llfinal = .FALSE.
	892	IF( PRESENT(ldfinal) ) llfinal = ldfinal
[10358]	893	ll_force_abort = .FALSE.
	894	IF( PRESENT(ld_force_abort) ) ll_force_abort = ld_force_abort
	895	!
	896	IF(ll_force_abort) THEN
	897	CALL mpi_abort( MPI_COMM_WORLD )
	898	ELSE
	899	CALL mppsync
	900	CALL mpi_finalize( info )
	901	ENDIF
[10330]	902	IF( .NOT. llfinal ) STOP 123456
[1344]	903	!
	904	END SUBROUTINE mppstop
[3]	905
	906
[1344]	907	SUBROUTINE mpp_comm_free( kcom )
	908	!!----------------------------------------------------------------------
	909	INTEGER, INTENT(in) :: kcom
	910	!!
	911	INTEGER :: ierr
	912	!!----------------------------------------------------------------------
	913	!
	914	CALL MPI_COMM_FREE(kcom, ierr)
	915	!
	916	END SUBROUTINE mpp_comm_free
[3]	917
[869]	918
[2715]	919	SUBROUTINE mpp_ini_znl( kumout )
[1345]	920	!!----------------------------------------------------------------------
	921	!! * routine mpp_ini_znl *
	922	!!
	923	!! ** Purpose : Initialize special communicator for computing zonal sum
	924	!!
	925	!! ** Method : - Look for processors in the same row
	926	!! - Put their number in nrank_znl
	927	!! - Create group for the znl processors
	928	!! - Create a communicator for znl processors
	929	!! - Determine if processor should write znl files
	930	!!
	931	!! ** output
	932	!! ndim_rank_znl = number of processors on the same row
	933	!! ngrp_znl = group ID for the znl processors
	934	!! ncomm_znl = communicator for the ice procs.
	935	!! n_znl_root = number (in the world) of proc 0 in the ice comm.
	936	!!
	937	!!----------------------------------------------------------------------
[2715]	938	INTEGER, INTENT(in) :: kumout ! ocean.output logical units
[1345]	939	!
[2715]	940	INTEGER :: jproc ! dummy loop integer
	941	INTEGER :: ierr, ii ! local integer
	942	INTEGER, ALLOCATABLE, DIMENSION(:) :: kwork
	943	!!----------------------------------------------------------------------
[1345]	944	!-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - ngrp_world : ', ngrp_world
	945	!-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - mpi_comm_world : ', mpi_comm_world
[9570]	946	!-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - mpi_comm_oce : ', mpi_comm_oce
[1345]	947	!
[2715]	948	ALLOCATE( kwork(jpnij), STAT=ierr )
	949	IF( ierr /= 0 ) THEN
	950	WRITE(kumout, cform_err)
	951	WRITE(kumout,*) 'mpp_ini_znl : failed to allocate 1D array of length jpnij'
	952	CALL mppstop
	953	ENDIF
	954
	955	IF( jpnj == 1 ) THEN
[1345]	956	ngrp_znl = ngrp_world
[9570]	957	ncomm_znl = mpi_comm_oce
[1345]	958	ELSE
	959	!
[9570]	960	CALL MPI_ALLGATHER ( njmpp, 1, mpi_integer, kwork, 1, mpi_integer, mpi_comm_oce, ierr )
[1345]	961	!-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - kwork pour njmpp : ', kwork
	962	!-$$ CALL flush(numout)
	963	!
	964	! Count number of processors on the same row
	965	ndim_rank_znl = 0
	966	DO jproc=1,jpnij
	967	IF ( kwork(jproc) == njmpp ) THEN
	968	ndim_rank_znl = ndim_rank_znl + 1
	969	ENDIF
	970	END DO
	971	!-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - ndim_rank_znl : ', ndim_rank_znl
	972	!-$$ CALL flush(numout)
	973	! Allocate the right size to nrank_znl
[1441]	974	IF (ALLOCATED (nrank_znl)) DEALLOCATE(nrank_znl)
[1345]	975	ALLOCATE(nrank_znl(ndim_rank_znl))
[3764]	976	ii = 0
[1345]	977	nrank_znl (:) = 0
	978	DO jproc=1,jpnij
	979	IF ( kwork(jproc) == njmpp) THEN
	980	ii = ii + 1
[3764]	981	nrank_znl(ii) = jproc -1
[1345]	982	ENDIF
	983	END DO
	984	!-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - nrank_znl : ', nrank_znl
	985	!-$$ CALL flush(numout)
	986
	987	! Create the opa group
[9570]	988	CALL MPI_COMM_GROUP(mpi_comm_oce,ngrp_opa,ierr)
[1345]	989	!-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - ngrp_opa : ', ngrp_opa
	990	!-$$ CALL flush(numout)
	991
	992	! Create the znl group from the opa group
	993	CALL MPI_GROUP_INCL ( ngrp_opa, ndim_rank_znl, nrank_znl, ngrp_znl, ierr )
	994	!-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - ngrp_znl ', ngrp_znl
	995	!-$$ CALL flush(numout)
	996
	997	! Create the znl communicator from the opa communicator, ie the pool of procs in the same row
[9570]	998	CALL MPI_COMM_CREATE ( mpi_comm_oce, ngrp_znl, ncomm_znl, ierr )
[1345]	999	!-$$ WRITE (numout,*) 'mpp_ini_znl ', nproc, ' - ncomm_znl ', ncomm_znl
	1000	!-$$ CALL flush(numout)
	1001	!
	1002	END IF
	1003
	1004	! Determines if processor if the first (starting from i=1) on the row
[3764]	1005	IF ( jpni == 1 ) THEN
[1345]	1006	l_znl_root = .TRUE.
	1007	ELSE
	1008	l_znl_root = .FALSE.
	1009	kwork (1) = nimpp
[10297]	1010	CALL mpp_min ( 'lib_mpp', kwork(1), kcom = ncomm_znl)
[1345]	1011	IF ( nimpp == kwork(1)) l_znl_root = .TRUE.
	1012	END IF
	1013
[2715]	1014	DEALLOCATE(kwork)
	1015
[1345]	1016	END SUBROUTINE mpp_ini_znl
	1017
	1018
[1344]	1019	SUBROUTINE mpp_ini_north
	1020	!!----------------------------------------------------------------------
	1021	!! * routine mpp_ini_north *
	1022	!!
[3764]	1023	!! ** Purpose : Initialize special communicator for north folding
[1344]	1024	!! condition together with global variables needed in the mpp folding
	1025	!!
	1026	!! ** Method : - Look for northern processors
	1027	!! - Put their number in nrank_north
	1028	!! - Create groups for the world processors and the north processors
	1029	!! - Create a communicator for northern processors
	1030	!!
	1031	!! ** output
	1032	!! njmppmax = njmpp for northern procs
	1033	!! ndim_rank_north = number of processors in the northern line
	1034	!! nrank_north (ndim_rank_north) = number of the northern procs.
	1035	!! ngrp_world = group ID for the world processors
	1036	!! ngrp_north = group ID for the northern processors
	1037	!! ncomm_north = communicator for the northern procs.
	1038	!! north_root = number (in the world) of proc 0 in the northern comm.
	1039	!!
	1040	!!----------------------------------------------------------------------
	1041	INTEGER :: ierr
	1042	INTEGER :: jjproc
	1043	INTEGER :: ii, ji
	1044	!!----------------------------------------------------------------------
	1045	!
	1046	njmppmax = MAXVAL( njmppt )
	1047	!
	1048	! Look for how many procs on the northern boundary
	1049	ndim_rank_north = 0
	1050	DO jjproc = 1, jpnij
	1051	IF( njmppt(jjproc) == njmppmax ) ndim_rank_north = ndim_rank_north + 1
	1052	END DO
	1053	!
	1054	! Allocate the right size to nrank_north
[1441]	1055	IF (ALLOCATED (nrank_north)) DEALLOCATE(nrank_north)
[1344]	1056	ALLOCATE( nrank_north(ndim_rank_north) )
[869]	1057
[1344]	1058	! Fill the nrank_north array with proc. number of northern procs.
	1059	! Note : the rank start at 0 in MPI
	1060	ii = 0
	1061	DO ji = 1, jpnij
	1062	IF ( njmppt(ji) == njmppmax ) THEN
	1063	ii=ii+1
	1064	nrank_north(ii)=ji-1
	1065	END IF
	1066	END DO
	1067	!
	1068	! create the world group
[9570]	1069	CALL MPI_COMM_GROUP( mpi_comm_oce, ngrp_world, ierr )
[1344]	1070	!
	1071	! Create the North group from the world group
	1072	CALL MPI_GROUP_INCL( ngrp_world, ndim_rank_north, nrank_north, ngrp_north, ierr )
	1073	!
	1074	! Create the North communicator , ie the pool of procs in the north group
[9570]	1075	CALL MPI_COMM_CREATE( mpi_comm_oce, ngrp_north, ncomm_north, ierr )
[1344]	1076	!
	1077	END SUBROUTINE mpp_ini_north
[869]	1078
	1079
[9570]	1080	SUBROUTINE mpi_init_oce( ldtxt, ksft, code )
[1344]	1081	!!---------------------------------------------------------------------
	1082	!! * routine mpp_init.opa *
	1083	!!
	1084	!! ** Purpose :: export and attach a MPI buffer for bsend
	1085	!!
	1086	!! ** Method :: define buffer size in namelist, if 0 no buffer attachment
	1087	!! but classical mpi_init
[3764]	1088	!!
	1089	!! History :: 01/11 :: IDRIS initial version for IBM only
[1344]	1090	!! 08/04 :: R. Benshila, generalisation
	1091	!!---------------------------------------------------------------------
[3764]	1092	CHARACTER(len=*),DIMENSION(:), INTENT( out) :: ldtxt
[2481]	1093	INTEGER , INTENT(inout) :: ksft
	1094	INTEGER , INTENT( out) :: code
	1095	INTEGER :: ierr, ji
	1096	LOGICAL :: mpi_was_called
[1344]	1097	!!---------------------------------------------------------------------
	1098	!
	1099	CALL mpi_initialized( mpi_was_called, code ) ! MPI initialization
[532]	1100	IF ( code /= MPI_SUCCESS ) THEN
[3764]	1101	DO ji = 1, SIZE(ldtxt)
[2481]	1102	IF( TRIM(ldtxt(ji)) /= '' ) WRITE(,) ldtxt(ji) ! control print of mynode
[3764]	1103	END DO
[2481]	1104	WRITE(*, cform_err)
	1105	WRITE(, ) ' lib_mpp: Error in routine mpi_initialized'
[1344]	1106	CALL mpi_abort( mpi_comm_world, code, ierr )
[532]	1107	ENDIF
[1344]	1108	!
	1109	IF( .NOT. mpi_was_called ) THEN
	1110	CALL mpi_init( code )
[9570]	1111	CALL mpi_comm_dup( mpi_comm_world, mpi_comm_oce, code )
[532]	1112	IF ( code /= MPI_SUCCESS ) THEN
[3764]	1113	DO ji = 1, SIZE(ldtxt)
[2481]	1114	IF( TRIM(ldtxt(ji)) /= '' ) WRITE(,) ldtxt(ji) ! control print of mynode
	1115	END DO
	1116	WRITE(*, cform_err)
	1117	WRITE(, ) ' lib_mpp: Error in routine mpi_comm_dup'
[532]	1118	CALL mpi_abort( mpi_comm_world, code, ierr )
	1119	ENDIF
	1120	ENDIF
[1344]	1121	!
[897]	1122	IF( nn_buffer > 0 ) THEN
[2481]	1123	WRITE(ldtxt(ksft),*) 'mpi_bsend, buffer allocation of : ', nn_buffer ; ksft = ksft + 1
[897]	1124	! Buffer allocation and attachment
[2481]	1125	ALLOCATE( tampon(nn_buffer), stat = ierr )
[3764]	1126	IF( ierr /= 0 ) THEN
	1127	DO ji = 1, SIZE(ldtxt)
[2481]	1128	IF( TRIM(ldtxt(ji)) /= '' ) WRITE(,) ldtxt(ji) ! control print of mynode
	1129	END DO
	1130	WRITE(*, cform_err)
	1131	WRITE(, ) ' lib_mpp: Error in ALLOCATE', ierr
	1132	CALL mpi_abort( mpi_comm_world, code, ierr )
	1133	END IF
	1134	CALL mpi_buffer_attach( tampon, nn_buffer, code )
[897]	1135	ENDIF
[1344]	1136	!
[9570]	1137	END SUBROUTINE mpi_init_oce
[3]	1138
[9019]	1139
	1140	SUBROUTINE DDPDD_MPI( ydda, yddb, ilen, itype )
[1976]	1141	!!---------------------------------------------------------------------
	1142	!! Routine DDPDD_MPI: used by reduction operator MPI_SUMDD
	1143	!!
	1144	!! Modification of original codes written by David H. Bailey
	1145	!! This subroutine computes yddb(i) = ydda(i)+yddb(i)
	1146	!!---------------------------------------------------------------------
[9019]	1147	INTEGER , INTENT(in) :: ilen, itype
	1148	COMPLEX(wp), DIMENSION(ilen), INTENT(in) :: ydda
	1149	COMPLEX(wp), DIMENSION(ilen), INTENT(inout) :: yddb
[1976]	1150	!
	1151	REAL(wp) :: zerr, zt1, zt2 ! local work variables
[9019]	1152	INTEGER :: ji, ztmp ! local scalar
	1153	!!---------------------------------------------------------------------
	1154	!
[1976]	1155	ztmp = itype ! avoid compilation warning
[9019]	1156	!
[1976]	1157	DO ji=1,ilen
	1158	! Compute ydda + yddb using Knuth's trick.
	1159	zt1 = real(ydda(ji)) + real(yddb(ji))
	1160	zerr = zt1 - real(ydda(ji))
	1161	zt2 = ((real(yddb(ji)) - zerr) + (real(ydda(ji)) - (zt1 - zerr))) &
	1162	+ aimag(ydda(ji)) + aimag(yddb(ji))
	1163
	1164	! The result is zt1 + zt2, after normalization.
	1165	yddb(ji) = cmplx ( zt1 + zt2, zt2 - ((zt1 + zt2) - zt1),wp )
	1166	END DO
[9019]	1167	!
[1976]	1168	END SUBROUTINE DDPDD_MPI
	1169
[6140]	1170
[9019]	1171	SUBROUTINE mpp_lbc_north_icb( pt2d, cd_type, psgn, kextj)
[4990]	1172	!!---------------------------------------------------------------------
	1173	!! * routine mpp_lbc_north_icb *
	1174	!!
	1175	!! ** Purpose : Ensure proper north fold horizontal bondary condition
	1176	!! in mpp configuration in case of jpn1 > 1 and for 2d
	1177	!! array with outer extra halo
	1178	!!
	1179	!! ** Method : North fold condition and mpp with more than one proc
	1180	!! in i-direction require a specific treatment. We gather
[9019]	1181	!! the 4+kextj northern lines of the global domain on 1
[4990]	1182	!! processor and apply lbc north-fold on this sub array.
	1183	!! Then we scatter the north fold array back to the processors.
[9019]	1184	!! This routine accounts for an extra halo with icebergs
	1185	!! and assumes ghost rows and columns have been suppressed.
[4990]	1186	!!
	1187	!!----------------------------------------------------------------------
	1188	REAL(wp), DIMENSION(:,:), INTENT(inout) :: pt2d ! 2D array with extra halo
	1189	CHARACTER(len=1) , INTENT(in ) :: cd_type ! nature of pt3d grid-points
	1190	! ! = T , U , V , F or W -points
	1191	REAL(wp) , INTENT(in ) :: psgn ! = -1. the sign change across the
	1192	!! ! north fold, = 1. otherwise
[9019]	1193	INTEGER , INTENT(in ) :: kextj ! Extra halo width at north fold
[6140]	1194	!
[4990]	1195	INTEGER :: ji, jj, jr
	1196	INTEGER :: ierr, itaille, ildi, ilei, iilb
[9019]	1197	INTEGER :: ipj, ij, iproc
[4990]	1198	!
	1199	REAL(wp), DIMENSION(:,:) , ALLOCATABLE :: ztab_e, znorthloc_e
	1200	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: znorthgloio_e
	1201	!!----------------------------------------------------------------------
	1202	!
[9019]	1203	ipj=4
[9467]	1204	ALLOCATE( ztab_e(jpiglo, 1-kextj:ipj+kextj) , &
	1205	& znorthloc_e(jpimax, 1-kextj:ipj+kextj) , &
	1206	& znorthgloio_e(jpimax, 1-kextj:ipj+kextj,jpni) )
[4990]	1207	!
[9019]	1208	ztab_e(:,:) = 0._wp
	1209	znorthloc_e(:,:) = 0._wp
[6140]	1210	!
[9467]	1211	ij = 1 - kextj
	1212	! put the last ipj+2*kextj lines of pt2d into znorthloc_e
	1213	DO jj = jpj - ipj + 1 - kextj , jpj + kextj
	1214	znorthloc_e(1:jpi,ij)=pt2d(1:jpi,jj)
[4990]	1215	ij = ij + 1
	1216	END DO
	1217	!
[9467]	1218	itaille = jpimax * ( ipj + 2*kextj )
[10314]	1219	!
	1220	IF( ln_timing ) CALL tic_tac(.TRUE.)
[9467]	1221	CALL MPI_ALLGATHER( znorthloc_e(1,1-kextj) , itaille, MPI_DOUBLE_PRECISION, &
	1222	& znorthgloio_e(1,1-kextj,1), itaille, MPI_DOUBLE_PRECISION, &
	1223	& ncomm_north, ierr )
[4990]	1224	!
[10314]	1225	IF( ln_timing ) CALL tic_tac(.FALSE.)
	1226	!
[4990]	1227	DO jr = 1, ndim_rank_north ! recover the global north array
	1228	iproc = nrank_north(jr) + 1
	1229	ildi = nldit (iproc)
	1230	ilei = nleit (iproc)
	1231	iilb = nimppt(iproc)
[9467]	1232	DO jj = 1-kextj, ipj+kextj
[4990]	1233	DO ji = ildi, ilei
	1234	ztab_e(ji+iilb-1,jj) = znorthgloio_e(ji,jj,jr)
	1235	END DO
	1236	END DO
	1237	END DO
	1238
	1239	! 2. North-Fold boundary conditions
	1240	! ----------------------------------
[9467]	1241	CALL lbc_nfd( ztab_e(:,1-kextj:ipj+kextj), cd_type, psgn, kextj )
[4990]	1242
[9467]	1243	ij = 1 - kextj
[4990]	1244	!! Scatter back to pt2d
[9467]	1245	DO jj = jpj - ipj + 1 - kextj , jpj + kextj
[9019]	1246	DO ji= 1, jpi
[4990]	1247	pt2d(ji,jj) = ztab_e(ji+nimpp-1,ij)
	1248	END DO
[9467]	1249	ij = ij +1
[4990]	1250	END DO
	1251	!
	1252	DEALLOCATE( ztab_e, znorthloc_e, znorthgloio_e )
	1253	!
	1254	END SUBROUTINE mpp_lbc_north_icb
	1255
[6140]	1256
[10314]	1257	SUBROUTINE mpp_lnk_2d_icb( cdname, pt2d, cd_type, psgn, kexti, kextj )
[4990]	1258	!!----------------------------------------------------------------------
	1259	!! * routine mpp_lnk_2d_icb *
	1260	!!
[9019]	1261	!! ** Purpose : Message passing management for 2d array (with extra halo for icebergs)
	1262	!! This routine receives a (1-kexti:jpi+kexti,1-kexti:jpj+kextj)
	1263	!! array (usually (0:jpi+1, 0:jpj+1)) from lbc_lnk_icb calls.
[4990]	1264	!!
	1265	!! ** Method : Use mppsend and mpprecv function for passing mask
	1266	!! between processors following neighboring subdomains.
	1267	!! domain parameters
[9019]	1268	!! jpi : first dimension of the local subdomain
	1269	!! jpj : second dimension of the local subdomain
	1270	!! kexti : number of columns for extra outer halo
	1271	!! kextj : number of rows for extra outer halo
[4990]	1272	!! nbondi : mark for "east-west local boundary"
	1273	!! nbondj : mark for "north-south local boundary"
	1274	!! noea : number for local neighboring processors
	1275	!! nowe : number for local neighboring processors
	1276	!! noso : number for local neighboring processors
	1277	!! nono : number for local neighboring processors
	1278	!!----------------------------------------------------------------------
[10314]	1279	CHARACTER(len=*) , INTENT(in ) :: cdname ! name of the calling subroutine
[9019]	1280	REAL(wp), DIMENSION(1-kexti:jpi+kexti,1-kextj:jpj+kextj), INTENT(inout) :: pt2d ! 2D array with extra halo
	1281	CHARACTER(len=1) , INTENT(in ) :: cd_type ! nature of ptab array grid-points
	1282	REAL(wp) , INTENT(in ) :: psgn ! sign used across the north fold
	1283	INTEGER , INTENT(in ) :: kexti ! extra i-halo width
	1284	INTEGER , INTENT(in ) :: kextj ! extra j-halo width
	1285	!
[4990]	1286	INTEGER :: jl ! dummy loop indices
[9019]	1287	INTEGER :: imigr, iihom, ijhom ! local integers
	1288	INTEGER :: ipreci, iprecj ! - -
[4990]	1289	INTEGER :: ml_req1, ml_req2, ml_err ! for key_mpi_isend
	1290	INTEGER, DIMENSION(MPI_STATUS_SIZE) :: ml_stat ! for key_mpi_isend
	1291	!!
[9019]	1292	REAL(wp), DIMENSION(1-kexti:jpi+kexti,nn_hls+kextj,2) :: r2dns, r2dsn
	1293	REAL(wp), DIMENSION(1-kextj:jpj+kextj,nn_hls+kexti,2) :: r2dwe, r2dew
[4990]	1294	!!----------------------------------------------------------------------
	1295
[9019]	1296	ipreci = nn_hls + kexti ! take into account outer extra 2D overlap area
	1297	iprecj = nn_hls + kextj
[4990]	1298
[10314]	1299	IF( narea == 1 .AND. numcom == -1 ) CALL mpp_report( cdname, 1, 1, 1, ld_lbc = .TRUE. )
[4990]	1300
	1301	! 1. standard boundary treatment
	1302	! ------------------------------
	1303	! Order matters Here !!!!
	1304	!
	1305	! ! East-West boundaries
	1306	! !* Cyclic east-west
[9667]	1307	IF( l_Iperio ) THEN
	1308	pt2d(1-kexti: 1 ,:) = pt2d(jpim1-kexti: jpim1 ,:) ! east
	1309	pt2d( jpi :jpi+kexti,:) = pt2d( 2 :2+kexti,:) ! west
[4990]	1310	!
	1311	ELSE !* closed
[9667]	1312	IF( .NOT. cd_type == 'F' ) pt2d( 1-kexti :nn_hls ,:) = 0._wp ! east except at F-point
	1313	pt2d(jpi-nn_hls+1:jpi+kexti,:) = 0._wp ! west
[4990]	1314	ENDIF
[9667]	1315	! ! North-South boundaries
	1316	IF( l_Jperio ) THEN !* cyclic (only with no mpp j-split)
	1317	pt2d(:,1-kextj: 1 ) = pt2d(:,jpjm1-kextj: jpjm1) ! north
	1318	pt2d(:, jpj :jpj+kextj) = pt2d(:, 2 :2+kextj) ! south
	1319	ELSE !* closed
	1320	IF( .NOT. cd_type == 'F' ) pt2d(:, 1-kextj :nn_hls ) = 0._wp ! north except at F-point
	1321	pt2d(:,jpj-nn_hls+1:jpj+kextj) = 0._wp ! south
	1322	ENDIF
[4990]	1323	!
	1324
	1325	! north fold treatment
	1326	! -----------------------
	1327	IF( npolj /= 0 ) THEN
	1328	!
	1329	SELECT CASE ( jpni )
[9019]	1330	CASE ( 1 ) ; CALL lbc_nfd ( pt2d(1:jpi,1:jpj+kextj), cd_type, psgn, kextj )
	1331	CASE DEFAULT ; CALL mpp_lbc_north_icb( pt2d(1:jpi,1:jpj+kextj), cd_type, psgn, kextj )
[4990]	1332	END SELECT
	1333	!
	1334	ENDIF
	1335
	1336	! 2. East and west directions exchange
	1337	! ------------------------------------
	1338	! we play with the neigbours AND the row number because of the periodicity
	1339	!
	1340	SELECT CASE ( nbondi ) ! Read Dirichlet lateral conditions
	1341	CASE ( -1, 0, 1 ) ! all exept 2 (i.e. close case)
[9019]	1342	iihom = jpi-nreci-kexti
[4990]	1343	DO jl = 1, ipreci
[9019]	1344	r2dew(:,jl,1) = pt2d(nn_hls+jl,:)
[4990]	1345	r2dwe(:,jl,1) = pt2d(iihom +jl,:)
	1346	END DO
	1347	END SELECT
	1348	!
	1349	! ! Migrations
[9019]	1350	imigr = ipreci * ( jpj + 2*kextj )
[4990]	1351	!
[10314]	1352	IF( ln_timing ) CALL tic_tac(.TRUE.)
	1353	!
[4990]	1354	SELECT CASE ( nbondi )
	1355	CASE ( -1 )
[9019]	1356	CALL mppsend( 2, r2dwe(1-kextj,1,1), imigr, noea, ml_req1 )
	1357	CALL mpprecv( 1, r2dew(1-kextj,1,2), imigr, noea )
[4990]	1358	IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err)
	1359	CASE ( 0 )
[9019]	1360	CALL mppsend( 1, r2dew(1-kextj,1,1), imigr, nowe, ml_req1 )
	1361	CALL mppsend( 2, r2dwe(1-kextj,1,1), imigr, noea, ml_req2 )
	1362	CALL mpprecv( 1, r2dew(1-kextj,1,2), imigr, noea )
	1363	CALL mpprecv( 2, r2dwe(1-kextj,1,2), imigr, nowe )
[4990]	1364	IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err)
	1365	IF(l_isend) CALL mpi_wait(ml_req2,ml_stat,ml_err)
	1366	CASE ( 1 )
[9019]	1367	CALL mppsend( 1, r2dew(1-kextj,1,1), imigr, nowe, ml_req1 )
	1368	CALL mpprecv( 2, r2dwe(1-kextj,1,2), imigr, nowe )
[4990]	1369	IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err)
	1370	END SELECT
	1371	!
[10314]	1372	IF( ln_timing ) CALL tic_tac(.FALSE.)
	1373	!
[4990]	1374	! ! Write Dirichlet lateral conditions
[9019]	1375	iihom = jpi - nn_hls
[4990]	1376	!
	1377	SELECT CASE ( nbondi )
	1378	CASE ( -1 )
	1379	DO jl = 1, ipreci
	1380	pt2d(iihom+jl,:) = r2dew(:,jl,2)
	1381	END DO
	1382	CASE ( 0 )
	1383	DO jl = 1, ipreci
[9019]	1384	pt2d(jl-kexti,:) = r2dwe(:,jl,2)
	1385	pt2d(iihom+jl,:) = r2dew(:,jl,2)
[4990]	1386	END DO
	1387	CASE ( 1 )
	1388	DO jl = 1, ipreci
[9019]	1389	pt2d(jl-kexti,:) = r2dwe(:,jl,2)
[4990]	1390	END DO
	1391	END SELECT
	1392
	1393
	1394	! 3. North and south directions
	1395	! -----------------------------
	1396	! always closed : we play only with the neigbours
	1397	!
	1398	IF( nbondj /= 2 ) THEN ! Read Dirichlet lateral conditions
[9019]	1399	ijhom = jpj-nrecj-kextj
[4990]	1400	DO jl = 1, iprecj
	1401	r2dsn(:,jl,1) = pt2d(:,ijhom +jl)
[9019]	1402	r2dns(:,jl,1) = pt2d(:,nn_hls+jl)
[4990]	1403	END DO
	1404	ENDIF
	1405	!
	1406	! ! Migrations
[9019]	1407	imigr = iprecj * ( jpi + 2*kexti )
[4990]	1408	!
[10314]	1409	IF( ln_timing ) CALL tic_tac(.TRUE.)
	1410	!
[4990]	1411	SELECT CASE ( nbondj )
	1412	CASE ( -1 )
[9019]	1413	CALL mppsend( 4, r2dsn(1-kexti,1,1), imigr, nono, ml_req1 )
	1414	CALL mpprecv( 3, r2dns(1-kexti,1,2), imigr, nono )
[4990]	1415	IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err)
	1416	CASE ( 0 )
[9019]	1417	CALL mppsend( 3, r2dns(1-kexti,1,1), imigr, noso, ml_req1 )
	1418	CALL mppsend( 4, r2dsn(1-kexti,1,1), imigr, nono, ml_req2 )
	1419	CALL mpprecv( 3, r2dns(1-kexti,1,2), imigr, nono )
	1420	CALL mpprecv( 4, r2dsn(1-kexti,1,2), imigr, noso )
[4990]	1421	IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err)
	1422	IF(l_isend) CALL mpi_wait(ml_req2,ml_stat,ml_err)
	1423	CASE ( 1 )
[9019]	1424	CALL mppsend( 3, r2dns(1-kexti,1,1), imigr, noso, ml_req1 )
	1425	CALL mpprecv( 4, r2dsn(1-kexti,1,2), imigr, noso )
[4990]	1426	IF(l_isend) CALL mpi_wait(ml_req1,ml_stat,ml_err)
	1427	END SELECT
	1428	!
[10314]	1429	IF( ln_timing ) CALL tic_tac(.FALSE.)
	1430	!
[4990]	1431	! ! Write Dirichlet lateral conditions
[9019]	1432	ijhom = jpj - nn_hls
[4990]	1433	!
	1434	SELECT CASE ( nbondj )
	1435	CASE ( -1 )
	1436	DO jl = 1, iprecj
	1437	pt2d(:,ijhom+jl) = r2dns(:,jl,2)
	1438	END DO
	1439	CASE ( 0 )
	1440	DO jl = 1, iprecj
[9019]	1441	pt2d(:,jl-kextj) = r2dsn(:,jl,2)
	1442	pt2d(:,ijhom+jl) = r2dns(:,jl,2)
[4990]	1443	END DO
	1444	CASE ( 1 )
	1445	DO jl = 1, iprecj
[9019]	1446	pt2d(:,jl-kextj) = r2dsn(:,jl,2)
[4990]	1447	END DO
	1448	END SELECT
[9019]	1449	!
[4990]	1450	END SUBROUTINE mpp_lnk_2d_icb
[10172]	1451
[10314]	1452
	1453	SUBROUTINE mpp_report( cdname, kpk, kpl, kpf, ld_lbc, ld_glb )
	1454	!!----------------------------------------------------------------------
	1455	!! * routine mpp_report *
	1456	!!
	1457	!! ** Purpose : report use of mpp routines per time-setp
	1458	!!
	1459	!!----------------------------------------------------------------------
	1460	CHARACTER(len=*), INTENT(in ) :: cdname ! name of the calling subroutine
	1461	INTEGER , OPTIONAL, INTENT(in ) :: kpk, kpl, kpf
	1462	LOGICAL , OPTIONAL, INTENT(in ) :: ld_lbc, ld_glb
	1463	!!
	1464	LOGICAL :: ll_lbc, ll_glb
	1465	INTEGER :: ji, jj, jk, jh, jf ! dummy loop indices
	1466	!!----------------------------------------------------------------------
	1467	!
	1468	ll_lbc = .FALSE.
	1469	IF( PRESENT(ld_lbc) ) ll_lbc = ld_lbc
	1470	ll_glb = .FALSE.
	1471	IF( PRESENT(ld_glb) ) ll_glb = ld_glb
	1472	!
	1473	! find the smallest common frequency: default = frequency product, if multiple, choose the larger of the 2 frequency
[10397]	1474	IF( ncom_dttrc /= 1 ) CALL ctl_stop( 'STOP', 'mpp_report, ncom_dttrc /= 1 not coded...' )
	1475	ncom_freq = ncom_fsbc
[10314]	1476	!
	1477	IF ( ncom_stp == nit000+ncom_freq ) THEN ! avoid to count extra communications in potential initializations at nit000
	1478	IF( ll_lbc ) THEN
	1479	IF( .NOT. ALLOCATED(ncomm_sequence) ) ALLOCATE( ncomm_sequence(ncom_rec_max,2) )
	1480	IF( .NOT. ALLOCATED( crname_lbc) ) ALLOCATE( crname_lbc(ncom_rec_max ) )
	1481	n_sequence_lbc = n_sequence_lbc + 1
[10397]	1482	IF( n_sequence_lbc > ncom_rec_max ) CALL ctl_stop( 'STOP', 'lib_mpp, increase ncom_rec_max' ) ! deadlock
[10314]	1483	crname_lbc(n_sequence_lbc) = cdname ! keep the name of the calling routine
	1484	ncomm_sequence(n_sequence_lbc,1) = kpk*kpl ! size of 3rd and 4th dimensions
	1485	ncomm_sequence(n_sequence_lbc,2) = kpf ! number of arrays to be treated (multi)
	1486	ENDIF
	1487	IF( ll_glb ) THEN
	1488	IF( .NOT. ALLOCATED(crname_glb) ) ALLOCATE( crname_glb(ncom_rec_max) )
	1489	n_sequence_glb = n_sequence_glb + 1
[10397]	1490	IF( n_sequence_glb > ncom_rec_max ) CALL ctl_stop( 'STOP', 'lib_mpp, increase ncom_rec_max' ) ! deadlock
[10314]	1491	crname_glb(n_sequence_glb) = cdname ! keep the name of the calling routine
	1492	ENDIF
	1493	ELSE IF ( ncom_stp == nit000+2*ncom_freq ) THEN
	1494	CALL ctl_opn( numcom, 'communication_report.txt', 'REPLACE', 'FORMATTED', 'SEQUENTIAL', -1, numout, .FALSE., narea )
	1495	WRITE(numcom,*) ' '
	1496	WRITE(numcom,*) ' ------------------------------------------------------------'
	1497	WRITE(numcom,*) ' Communication pattern report (second oce+sbc+top time step):'
	1498	WRITE(numcom,*) ' ------------------------------------------------------------'
	1499	WRITE(numcom,*) ' '
	1500	WRITE(numcom,'(A,I4)') ' Exchanged halos : ', n_sequence_lbc
	1501	jj = 0; jk = 0; jf = 0; jh = 0
	1502	DO ji = 1, n_sequence_lbc
	1503	IF ( ncomm_sequence(ji,1) .GT. 1 ) jk = jk + 1
	1504	IF ( ncomm_sequence(ji,2) .GT. 1 ) jf = jf + 1
	1505	IF ( ncomm_sequence(ji,1) .GT. 1 .AND. ncomm_sequence(ji,2) .GT. 1 ) jj = jj + 1
	1506	jh = MAX (jh, ncomm_sequence(ji,1)*ncomm_sequence(ji,2))
	1507	END DO
	1508	WRITE(numcom,'(A,I3)') ' 3D Exchanged halos : ', jk
	1509	WRITE(numcom,'(A,I3)') ' Multi arrays exchanged halos : ', jf
	1510	WRITE(numcom,'(A,I3)') ' from which 3D : ', jj
	1511	WRITE(numcom,'(A,I10)') ' Array max size : ', jhjpijpj
	1512	WRITE(numcom,*) ' '
	1513	WRITE(numcom,*) ' lbc_lnk called'
	1514	jj = 1
	1515	DO ji = 2, n_sequence_lbc
	1516	IF( crname_lbc(ji-1) /= crname_lbc(ji) ) THEN
	1517	WRITE(numcom,'(A, I4, A, A)') ' - ', jj,' times by subroutine ', TRIM(crname_lbc(ji-1))
	1518	jj = 0
	1519	END IF
	1520	jj = jj + 1
	1521	END DO
	1522	WRITE(numcom,'(A, I4, A, A)') ' - ', jj,' times by subroutine ', TRIM(crname_lbc(n_sequence_lbc))
	1523	WRITE(numcom,*) ' '
	1524	IF ( n_sequence_glb > 0 ) THEN
	1525	WRITE(numcom,'(A,I4)') ' Global communications : ', n_sequence_glb
	1526	jj = 1
	1527	DO ji = 2, n_sequence_glb
	1528	IF( crname_glb(ji-1) /= crname_glb(ji) ) THEN
	1529	WRITE(numcom,'(A, I4, A, A)') ' - ', jj,' times by subroutine ', TRIM(crname_glb(ji-1))
	1530	jj = 0
	1531	END IF
	1532	jj = jj + 1
	1533	END DO
	1534	WRITE(numcom,'(A, I4, A, A)') ' - ', jj,' times by subroutine ', TRIM(crname_glb(n_sequence_glb))
	1535	DEALLOCATE(crname_glb)
	1536	ELSE
	1537	WRITE(numcom,*) ' No MPI global communication '
	1538	ENDIF
	1539	WRITE(numcom,*) ' '
	1540	WRITE(numcom,*) ' -----------------------------------------------'
	1541	WRITE(numcom,*) ' '
	1542	DEALLOCATE(ncomm_sequence)
	1543	DEALLOCATE(crname_lbc)
	1544	ENDIF
	1545	END SUBROUTINE mpp_report
	1546
[6140]	1547
[10300]	1548	SUBROUTINE tic_tac (ld_tic, ld_global)
[10172]	1549
[10300]	1550	LOGICAL, INTENT(IN) :: ld_tic
	1551	LOGICAL, OPTIONAL, INTENT(IN) :: ld_global
	1552	REAL(wp), DIMENSION(2), SAVE :: tic_wt
	1553	REAL(wp), SAVE :: tic_ct = 0._wp
	1554	INTEGER :: ii
[10172]	1555
	1556	IF( ncom_stp <= nit000 ) RETURN
	1557	IF( ncom_stp == nitend ) RETURN
[10300]	1558	ii = 1
	1559	IF( PRESENT( ld_global ) ) THEN
	1560	IF( ld_global ) ii = 2
	1561	END IF
[10172]	1562
[10300]	1563	IF ( ld_tic ) THEN
	1564	tic_wt(ii) = MPI_Wtime() ! start count tic->tac (waiting time)
[10172]	1565	IF ( tic_ct > 0.0_wp ) compute_time = compute_time + MPI_Wtime() - tic_ct ! cumulate count tac->tic
	1566	ELSE
[10300]	1567	waiting_time(ii) = waiting_time(ii) + MPI_Wtime() - tic_wt(ii) ! cumulate count tic->tac
[10172]	1568	tic_ct = MPI_Wtime() ! start count tac->tic (waiting time)
	1569	ENDIF
	1570
	1571	END SUBROUTINE tic_tac
	1572
	1573
[13]	1574	#else
	1575	!!----------------------------------------------------------------------
	1576	!! Default case: Dummy module share memory computing
	1577	!!----------------------------------------------------------------------
[2715]	1578	USE in_out_manager
[1976]	1579
[13]	1580	INTERFACE mpp_sum
[10314]	1581	MODULE PROCEDURE mppsum_int, mppsum_a_int, mppsum_real, mppsum_a_real, mppsum_realdd, mppsum_a_realdd
[13]	1582	END INTERFACE
	1583	INTERFACE mpp_max
[681]	1584	MODULE PROCEDURE mppmax_a_int, mppmax_int, mppmax_a_real, mppmax_real
[13]	1585	END INTERFACE
	1586	INTERFACE mpp_min
	1587	MODULE PROCEDURE mppmin_a_int, mppmin_int, mppmin_a_real, mppmin_real
	1588	END INTERFACE
[1344]	1589	INTERFACE mpp_minloc
	1590	MODULE PROCEDURE mpp_minloc2d ,mpp_minloc3d
	1591	END INTERFACE
	1592	INTERFACE mpp_maxloc
	1593	MODULE PROCEDURE mpp_maxloc2d ,mpp_maxloc3d
	1594	END INTERFACE
[3]	1595
[13]	1596	LOGICAL, PUBLIC, PARAMETER :: lk_mpp = .FALSE. !: mpp flag
[4147]	1597	LOGICAL, PUBLIC :: ln_nnogather !: namelist control of northfold comms (needed here in case "key_mpp_mpi" is not used)
[9570]	1598	INTEGER, PUBLIC :: mpi_comm_oce ! opa local communicator
[10359]	1599
	1600	INTEGER, PARAMETER, PUBLIC :: nbdelay = 0 ! make sure we don't enter loops: DO ji = 1, nbdelay
	1601	CHARACTER(len=32), DIMENSION(1), PUBLIC :: c_delaylist = 'empty'
[10397]	1602	CHARACTER(len=32), DIMENSION(1), PUBLIC :: c_delaycpnt = 'empty'
	1603	TYPE :: DELAYARR
	1604	REAL( wp), POINTER, DIMENSION(:) :: z1d => NULL()
	1605	COMPLEX(wp), POINTER, DIMENSION(:) :: y1d => NULL()
	1606	END TYPE DELAYARR
	1607	TYPE( DELAYARR ), DIMENSION(1), PUBLIC :: todelay
[10359]	1608	INTEGER, PUBLIC, DIMENSION(1) :: ndelayid = -1
[2715]	1609	!!----------------------------------------------------------------------
[13]	1610	CONTAINS
[3]	1611
[2715]	1612	INTEGER FUNCTION lib_mpp_alloc(kumout) ! Dummy function
	1613	INTEGER, INTENT(in) :: kumout
	1614	lib_mpp_alloc = 0
	1615	END FUNCTION lib_mpp_alloc
	1616
[5407]	1617	FUNCTION mynode( ldtxt, ldname, kumnam_ref, knumnam_cfg, kumond , kstop, localComm ) RESULT (function_value)
[1579]	1618	INTEGER, OPTIONAL , INTENT(in ) :: localComm
[3764]	1619	CHARACTER(len=*),DIMENSION(:) :: ldtxt
[5407]	1620	CHARACTER(len=*) :: ldname
[4314]	1621	INTEGER :: kumnam_ref, knumnam_cfg , kumond , kstop
[9570]	1622	IF( PRESENT( localComm ) ) mpi_comm_oce = localComm
[5412]	1623	function_value = 0
[1579]	1624	IF( .FALSE. ) ldtxt(:) = 'never done'
[5407]	1625	CALL ctl_opn( kumond, TRIM(ldname), 'UNKNOWN', 'FORMATTED', 'SEQUENTIAL', -1, 6, .FALSE. , 1 )
[13]	1626	END FUNCTION mynode
[3]	1627
[13]	1628	SUBROUTINE mppsync ! Dummy routine
	1629	END SUBROUTINE mppsync
[3]	1630
[10314]	1631	!!----------------------------------------------------------------------
	1632	!! * mppmax_a_int, mppmax_int, mppmax_a_real, mppmax_real *
	1633	!!
	1634	!!----------------------------------------------------------------------
	1635	!!
	1636	# define OPERATION_MAX
	1637	# define INTEGER_TYPE
	1638	# define DIM_0d
	1639	# define ROUTINE_ALLREDUCE mppmax_int
	1640	# include "mpp_allreduce_generic.h90"
	1641	# undef ROUTINE_ALLREDUCE
	1642	# undef DIM_0d
	1643	# define DIM_1d
	1644	# define ROUTINE_ALLREDUCE mppmax_a_int
	1645	# include "mpp_allreduce_generic.h90"
	1646	# undef ROUTINE_ALLREDUCE
	1647	# undef DIM_1d
	1648	# undef INTEGER_TYPE
	1649	!
	1650	# define REAL_TYPE
	1651	# define DIM_0d
	1652	# define ROUTINE_ALLREDUCE mppmax_real
	1653	# include "mpp_allreduce_generic.h90"
	1654	# undef ROUTINE_ALLREDUCE
	1655	# undef DIM_0d
	1656	# define DIM_1d
	1657	# define ROUTINE_ALLREDUCE mppmax_a_real
	1658	# include "mpp_allreduce_generic.h90"
	1659	# undef ROUTINE_ALLREDUCE
	1660	# undef DIM_1d
	1661	# undef REAL_TYPE
	1662	# undef OPERATION_MAX
	1663	!!----------------------------------------------------------------------
	1664	!! * mppmin_a_int, mppmin_int, mppmin_a_real, mppmin_real *
	1665	!!
	1666	!!----------------------------------------------------------------------
	1667	!!
	1668	# define OPERATION_MIN
	1669	# define INTEGER_TYPE
	1670	# define DIM_0d
	1671	# define ROUTINE_ALLREDUCE mppmin_int
	1672	# include "mpp_allreduce_generic.h90"
	1673	# undef ROUTINE_ALLREDUCE
	1674	# undef DIM_0d
	1675	# define DIM_1d
	1676	# define ROUTINE_ALLREDUCE mppmin_a_int
	1677	# include "mpp_allreduce_generic.h90"
	1678	# undef ROUTINE_ALLREDUCE
	1679	# undef DIM_1d
	1680	# undef INTEGER_TYPE
	1681	!
	1682	# define REAL_TYPE
	1683	# define DIM_0d
	1684	# define ROUTINE_ALLREDUCE mppmin_real
	1685	# include "mpp_allreduce_generic.h90"
	1686	# undef ROUTINE_ALLREDUCE
	1687	# undef DIM_0d
	1688	# define DIM_1d
	1689	# define ROUTINE_ALLREDUCE mppmin_a_real
	1690	# include "mpp_allreduce_generic.h90"
	1691	# undef ROUTINE_ALLREDUCE
	1692	# undef DIM_1d
	1693	# undef REAL_TYPE
	1694	# undef OPERATION_MIN
[3]	1695
[10314]	1696	!!----------------------------------------------------------------------
	1697	!! * mppsum_a_int, mppsum_int, mppsum_a_real, mppsum_real *
	1698	!!
	1699	!! Global sum of 1D array or a variable (integer, real or complex)
	1700	!!----------------------------------------------------------------------
	1701	!!
	1702	# define OPERATION_SUM
	1703	# define INTEGER_TYPE
	1704	# define DIM_0d
	1705	# define ROUTINE_ALLREDUCE mppsum_int
	1706	# include "mpp_allreduce_generic.h90"
	1707	# undef ROUTINE_ALLREDUCE
	1708	# undef DIM_0d
	1709	# define DIM_1d
	1710	# define ROUTINE_ALLREDUCE mppsum_a_int
	1711	# include "mpp_allreduce_generic.h90"
	1712	# undef ROUTINE_ALLREDUCE
	1713	# undef DIM_1d
	1714	# undef INTEGER_TYPE
	1715	!
	1716	# define REAL_TYPE
	1717	# define DIM_0d
	1718	# define ROUTINE_ALLREDUCE mppsum_real
	1719	# include "mpp_allreduce_generic.h90"
	1720	# undef ROUTINE_ALLREDUCE
	1721	# undef DIM_0d
	1722	# define DIM_1d
	1723	# define ROUTINE_ALLREDUCE mppsum_a_real
	1724	# include "mpp_allreduce_generic.h90"
	1725	# undef ROUTINE_ALLREDUCE
	1726	# undef DIM_1d
	1727	# undef REAL_TYPE
	1728	# undef OPERATION_SUM
[3]	1729
[10314]	1730	# define OPERATION_SUM_DD
	1731	# define COMPLEX_TYPE
	1732	# define DIM_0d
	1733	# define ROUTINE_ALLREDUCE mppsum_realdd
	1734	# include "mpp_allreduce_generic.h90"
	1735	# undef ROUTINE_ALLREDUCE
	1736	# undef DIM_0d
	1737	# define DIM_1d
	1738	# define ROUTINE_ALLREDUCE mppsum_a_realdd
	1739	# include "mpp_allreduce_generic.h90"
	1740	# undef ROUTINE_ALLREDUCE
	1741	# undef DIM_1d
	1742	# undef COMPLEX_TYPE
	1743	# undef OPERATION_SUM_DD
[3]	1744
[10314]	1745	!!----------------------------------------------------------------------
	1746	!! *** mpp_minloc2d, mpp_minloc3d, mpp_maxloc2d, mpp_maxloc3d
	1747	!!
	1748	!!----------------------------------------------------------------------
	1749	!!
	1750	# define OPERATION_MINLOC
	1751	# define DIM_2d
	1752	# define ROUTINE_LOC mpp_minloc2d
	1753	# include "mpp_loc_generic.h90"
	1754	# undef ROUTINE_LOC
	1755	# undef DIM_2d
	1756	# define DIM_3d
	1757	# define ROUTINE_LOC mpp_minloc3d
	1758	# include "mpp_loc_generic.h90"
	1759	# undef ROUTINE_LOC
	1760	# undef DIM_3d
	1761	# undef OPERATION_MINLOC
[2480]	1762
[10314]	1763	# define OPERATION_MAXLOC
	1764	# define DIM_2d
	1765	# define ROUTINE_LOC mpp_maxloc2d
	1766	# include "mpp_loc_generic.h90"
	1767	# undef ROUTINE_LOC
	1768	# undef DIM_2d
	1769	# define DIM_3d
	1770	# define ROUTINE_LOC mpp_maxloc3d
	1771	# include "mpp_loc_generic.h90"
	1772	# undef ROUTINE_LOC
	1773	# undef DIM_3d
	1774	# undef OPERATION_MAXLOC
[13]	1775
[10397]	1776	SUBROUTINE mpp_delay_sum( cdname, cdelay, y_in, pout, ldlast, kcom )
[10359]	1777	CHARACTER(len=*), INTENT(in ) :: cdname ! name of the calling subroutine
	1778	CHARACTER(len=*), INTENT(in ) :: cdelay ! name (used as id) of the delayed operation
[10397]	1779	COMPLEX(wp), INTENT(in ), DIMENSION(:) :: y_in
	1780	REAL(wp), INTENT( out), DIMENSION(:) :: pout
[10359]	1781	LOGICAL, INTENT(in ) :: ldlast ! true if this is the last time we call this routine
	1782	INTEGER, INTENT(in ), OPTIONAL :: kcom
[10397]	1783	!
	1784	pout(:) = REAL(y_in(:), wp)
	1785	END SUBROUTINE mpp_delay_sum
	1786
	1787	SUBROUTINE mpp_delay_max( cdname, cdelay, p_in, pout, ldlast, kcom )
	1788	CHARACTER(len=*), INTENT(in ) :: cdname ! name of the calling subroutine
	1789	CHARACTER(len=*), INTENT(in ) :: cdelay ! name (used as id) of the delayed operation
	1790	REAL(wp), INTENT(in ), DIMENSION(:) :: p_in
	1791	REAL(wp), INTENT( out), DIMENSION(:) :: pout
	1792	LOGICAL, INTENT(in ) :: ldlast ! true if this is the last time we call this routine
	1793	INTEGER, INTENT(in ), OPTIONAL :: kcom
	1794	!
	1795	pout(:) = p_in(:)
[10359]	1796	END SUBROUTINE mpp_delay_max
[10292]	1797
[10397]	1798	SUBROUTINE mpp_delay_rcv( kid )
	1799	INTEGER,INTENT(in ) :: kid
	1800	WRITE(,) 'mpp_delay_rcv: You should not have seen this print! error?', kid
	1801	END SUBROUTINE mpp_delay_rcv
	1802
[10358]	1803	SUBROUTINE mppstop( ldfinal, ld_force_abort )
	1804	LOGICAL, OPTIONAL, INTENT(in) :: ldfinal ! source process number
	1805	LOGICAL, OPTIONAL, INTENT(in) :: ld_force_abort ! source process number
[3799]	1806	STOP ! non MPP case, just stop the run
[51]	1807	END SUBROUTINE mppstop
	1808
[2715]	1809	SUBROUTINE mpp_ini_znl( knum )
	1810	INTEGER :: knum
	1811	WRITE(,) 'mpp_ini_znl: You should not have seen this print! error?', knum
[1345]	1812	END SUBROUTINE mpp_ini_znl
	1813
[1344]	1814	SUBROUTINE mpp_comm_free( kcom )
[869]	1815	INTEGER :: kcom
[1344]	1816	WRITE(,) 'mpp_comm_free: You should not have seen this print! error?', kcom
[869]	1817	END SUBROUTINE mpp_comm_free
[9019]	1818
[3]	1819	#endif
[2715]	1820
[13]	1821	!!----------------------------------------------------------------------
[4147]	1822	!! All cases: ctl_stop, ctl_warn, get_unit, ctl_opn, ctl_nam routines
[2715]	1823	!!----------------------------------------------------------------------
	1824
	1825	SUBROUTINE ctl_stop( cd1, cd2, cd3, cd4, cd5 , &
	1826	& cd6, cd7, cd8, cd9, cd10 )
	1827	!!----------------------------------------------------------------------
	1828	!! * ROUTINE stop_opa *
	1829	!!
[3764]	1830	!! ** Purpose : print in ocean.outpput file a error message and
[2715]	1831	!! increment the error number (nstop) by one.
	1832	!!----------------------------------------------------------------------
	1833	CHARACTER(len=*), INTENT(in), OPTIONAL :: cd1, cd2, cd3, cd4, cd5
	1834	CHARACTER(len=*), INTENT(in), OPTIONAL :: cd6, cd7, cd8, cd9, cd10
	1835	!!----------------------------------------------------------------------
	1836	!
[3764]	1837	nstop = nstop + 1
[2715]	1838	IF(lwp) THEN
	1839	WRITE(numout,cform_err)
[10330]	1840	IF( PRESENT(cd1 ) ) WRITE(numout,*) TRIM(cd1)
	1841	IF( PRESENT(cd2 ) ) WRITE(numout,*) TRIM(cd2)
	1842	IF( PRESENT(cd3 ) ) WRITE(numout,*) TRIM(cd3)
	1843	IF( PRESENT(cd4 ) ) WRITE(numout,*) TRIM(cd4)
	1844	IF( PRESENT(cd5 ) ) WRITE(numout,*) TRIM(cd5)
	1845	IF( PRESENT(cd6 ) ) WRITE(numout,*) TRIM(cd6)
	1846	IF( PRESENT(cd7 ) ) WRITE(numout,*) TRIM(cd7)
	1847	IF( PRESENT(cd8 ) ) WRITE(numout,*) TRIM(cd8)
	1848	IF( PRESENT(cd9 ) ) WRITE(numout,*) TRIM(cd9)
	1849	IF( PRESENT(cd10) ) WRITE(numout,*) TRIM(cd10)
[2715]	1850	ENDIF
	1851	CALL FLUSH(numout )
	1852	IF( numstp /= -1 ) CALL FLUSH(numstp )
[9019]	1853	IF( numrun /= -1 ) CALL FLUSH(numrun )
[2715]	1854	IF( numevo_ice /= -1 ) CALL FLUSH(numevo_ice)
	1855	!
	1856	IF( cd1 == 'STOP' ) THEN
	1857	IF(lwp) WRITE(numout,*) 'huge E-R-R-O-R : immediate stop'
[10407]	1858	CALL mppstop(ld_force_abort = .true.)
[2715]	1859	ENDIF
	1860	!
	1861	END SUBROUTINE ctl_stop
	1862
	1863
	1864	SUBROUTINE ctl_warn( cd1, cd2, cd3, cd4, cd5, &
	1865	& cd6, cd7, cd8, cd9, cd10 )
	1866	!!----------------------------------------------------------------------
	1867	!! * ROUTINE stop_warn *
	1868	!!
[3764]	1869	!! ** Purpose : print in ocean.outpput file a error message and
[2715]	1870	!! increment the warning number (nwarn) by one.
	1871	!!----------------------------------------------------------------------
	1872	CHARACTER(len=*), INTENT(in), OPTIONAL :: cd1, cd2, cd3, cd4, cd5
	1873	CHARACTER(len=*), INTENT(in), OPTIONAL :: cd6, cd7, cd8, cd9, cd10
	1874	!!----------------------------------------------------------------------
[3764]	1875	!
	1876	nwarn = nwarn + 1
[2715]	1877	IF(lwp) THEN
	1878	WRITE(numout,cform_war)
[10330]	1879	IF( PRESENT(cd1 ) ) WRITE(numout,*) TRIM(cd1)
	1880	IF( PRESENT(cd2 ) ) WRITE(numout,*) TRIM(cd2)
	1881	IF( PRESENT(cd3 ) ) WRITE(numout,*) TRIM(cd3)
	1882	IF( PRESENT(cd4 ) ) WRITE(numout,*) TRIM(cd4)
	1883	IF( PRESENT(cd5 ) ) WRITE(numout,*) TRIM(cd5)
	1884	IF( PRESENT(cd6 ) ) WRITE(numout,*) TRIM(cd6)
	1885	IF( PRESENT(cd7 ) ) WRITE(numout,*) TRIM(cd7)
	1886	IF( PRESENT(cd8 ) ) WRITE(numout,*) TRIM(cd8)
	1887	IF( PRESENT(cd9 ) ) WRITE(numout,*) TRIM(cd9)
	1888	IF( PRESENT(cd10) ) WRITE(numout,*) TRIM(cd10)
[2715]	1889	ENDIF
	1890	CALL FLUSH(numout)
	1891	!
	1892	END SUBROUTINE ctl_warn
	1893
	1894
	1895	SUBROUTINE ctl_opn( knum, cdfile, cdstat, cdform, cdacce, klengh, kout, ldwp, karea )
	1896	!!----------------------------------------------------------------------
	1897	!! * ROUTINE ctl_opn *
	1898	!!
	1899	!! ** Purpose : Open file and check if required file is available.
	1900	!!
	1901	!! ** Method : Fortan open
	1902	!!----------------------------------------------------------------------
	1903	INTEGER , INTENT( out) :: knum ! logical unit to open
	1904	CHARACTER(len=*) , INTENT(in ) :: cdfile ! file name to open
	1905	CHARACTER(len=*) , INTENT(in ) :: cdstat ! disposition specifier
	1906	CHARACTER(len=*) , INTENT(in ) :: cdform ! formatting specifier
	1907	CHARACTER(len=*) , INTENT(in ) :: cdacce ! access specifier
	1908	INTEGER , INTENT(in ) :: klengh ! record length
	1909	INTEGER , INTENT(in ) :: kout ! number of logical units for write
	1910	LOGICAL , INTENT(in ) :: ldwp ! boolean term for print
	1911	INTEGER, OPTIONAL, INTENT(in ) :: karea ! proc number
[5836]	1912	!
[2715]	1913	CHARACTER(len=80) :: clfile
	1914	INTEGER :: iost
	1915	!!----------------------------------------------------------------------
[5836]	1916	!
[2715]	1917	! adapt filename
	1918	! ----------------
	1919	clfile = TRIM(cdfile)
	1920	IF( PRESENT( karea ) ) THEN
	1921	IF( karea > 1 ) WRITE(clfile, "(a,'_',i4.4)") TRIM(clfile), karea-1
	1922	ENDIF
	1923	#if defined key_agrif
	1924	IF( .NOT. Agrif_Root() ) clfile = TRIM(Agrif_CFixed())//'_'//TRIM(clfile)
	1925	knum=Agrif_Get_Unit()
	1926	#else
	1927	knum=get_unit()
	1928	#endif
[10330]	1929	IF( TRIM(cdfile) == '/dev/null' ) clfile = TRIM(cdfile) ! force the use of /dev/null
[5836]	1930	!
[2715]	1931	iost=0
[10368]	1932	IF( cdacce(1:6) == 'DIRECT' ) THEN ! cdacce has always more than 6 characters
[10358]	1933	OPEN( UNIT=knum, FILE=clfile, FORM=cdform, ACCESS=cdacce, STATUS=cdstat, RECL=klengh , ERR=100, IOSTAT=iost )
[10368]	1934	ELSE IF( TRIM(cdstat) == 'APPEND' ) THEN ! cdstat can have less than 6 characters
[10358]	1935	OPEN( UNIT=knum, FILE=clfile, FORM=cdform, ACCESS=cdacce, STATUS='UNKNOWN', POSITION='APPEND', ERR=100, IOSTAT=iost )
[2715]	1936	ELSE
[10358]	1937	OPEN( UNIT=knum, FILE=clfile, FORM=cdform, ACCESS=cdacce, STATUS=cdstat , ERR=100, IOSTAT=iost )
[2715]	1938	ENDIF
[10358]	1939	IF( iost /= 0 .AND. TRIM(clfile) == '/dev/null' ) & ! for windows
	1940	& OPEN(UNIT=knum,FILE='NUL', FORM=cdform, ACCESS=cdacce, STATUS=cdstat , ERR=100, IOSTAT=iost )
[2715]	1941	IF( iost == 0 ) THEN
	1942	IF(ldwp) THEN
[10358]	1943	WRITE(kout,*) ' file : ', TRIM(clfile),' open ok'
[2715]	1944	WRITE(kout,*) ' unit = ', knum
	1945	WRITE(kout,*) ' status = ', cdstat
	1946	WRITE(kout,*) ' form = ', cdform
	1947	WRITE(kout,*) ' access = ', cdacce
	1948	WRITE(kout,*)
	1949	ENDIF
	1950	ENDIF
	1951	100 CONTINUE
	1952	IF( iost /= 0 ) THEN
	1953	IF(ldwp) THEN
	1954	WRITE(kout,*)
[10358]	1955	WRITE(kout,*) ' ===>>>> : bad opening file: ', TRIM(clfile)
[2715]	1956	WRITE(kout,*) ' ======= === '
	1957	WRITE(kout,*) ' unit = ', knum
	1958	WRITE(kout,*) ' status = ', cdstat
	1959	WRITE(kout,*) ' form = ', cdform
	1960	WRITE(kout,*) ' access = ', cdacce
	1961	WRITE(kout,*) ' iostat = ', iost
	1962	WRITE(kout,*) ' we stop. verify the file '
	1963	WRITE(kout,*)
[9438]	1964	ELSE !!! Force writing to make sure we get the information - at least once - in this violent STOP!!
	1965	WRITE(,)
[10358]	1966	WRITE(,) ' ===>>>> : bad opening file: ', TRIM(clfile)
[9438]	1967	WRITE(,) ' ======= === '
	1968	WRITE(,) ' unit = ', knum
	1969	WRITE(,) ' status = ', cdstat
	1970	WRITE(,) ' form = ', cdform
	1971	WRITE(,) ' access = ', cdacce
	1972	WRITE(,) ' iostat = ', iost
	1973	WRITE(,) ' we stop. verify the file '
	1974	WRITE(,)
[2715]	1975	ENDIF
[9019]	1976	CALL FLUSH( kout )
[2715]	1977	STOP 'ctl_opn bad opening'
	1978	ENDIF
[5836]	1979	!
[2715]	1980	END SUBROUTINE ctl_opn
	1981
[5836]	1982
[4147]	1983	SUBROUTINE ctl_nam ( kios, cdnam, ldwp )
	1984	!!----------------------------------------------------------------------
	1985	!! * ROUTINE ctl_nam *
	1986	!!
	1987	!! ** Purpose : Informations when error while reading a namelist
	1988	!!
	1989	!! ** Method : Fortan open
	1990	!!----------------------------------------------------------------------
[5836]	1991	INTEGER , INTENT(inout) :: kios ! IO status after reading the namelist
	1992	CHARACTER(len=*), INTENT(in ) :: cdnam ! group name of namelist for which error occurs
[7646]	1993	CHARACTER(len=5) :: clios ! string to convert iostat in character for print
[5836]	1994	LOGICAL , INTENT(in ) :: ldwp ! boolean term for print
[4147]	1995	!!----------------------------------------------------------------------
[5836]	1996	!
[7646]	1997	WRITE (clios, '(I5.0)') kios
[4147]	1998	IF( kios < 0 ) THEN
[5836]	1999	CALL ctl_warn( 'end of record or file while reading namelist ' &
	2000	& // TRIM(cdnam) // ' iostat = ' // TRIM(clios) )
[4147]	2001	ENDIF
[5836]	2002	!
[4147]	2003	IF( kios > 0 ) THEN
[5836]	2004	CALL ctl_stop( 'misspelled variable in namelist ' &
	2005	& // TRIM(cdnam) // ' iostat = ' // TRIM(clios) )
[4147]	2006	ENDIF
	2007	kios = 0
	2008	RETURN
[5836]	2009	!
[4147]	2010	END SUBROUTINE ctl_nam
	2011
[5836]	2012
[2715]	2013	INTEGER FUNCTION get_unit()
	2014	!!----------------------------------------------------------------------
	2015	!! * FUNCTION get_unit *
	2016	!!
	2017	!! ** Purpose : return the index of an unused logical unit
	2018	!!----------------------------------------------------------------------
[3764]	2019	LOGICAL :: llopn
[2715]	2020	!!----------------------------------------------------------------------
	2021	!
	2022	get_unit = 15 ! choose a unit that is big enough then it is not already used in NEMO
	2023	llopn = .TRUE.
	2024	DO WHILE( (get_unit < 998) .AND. llopn )
	2025	get_unit = get_unit + 1
	2026	INQUIRE( unit = get_unit, opened = llopn )
	2027	END DO
	2028	IF( (get_unit == 999) .AND. llopn ) THEN
	2029	CALL ctl_stop( 'get_unit: All logical units until 999 are used...' )
	2030	get_unit = -1
	2031	ENDIF
	2032	!
	2033	END FUNCTION get_unit
	2034
	2035	!!----------------------------------------------------------------------
[3]	2036	END MODULE lib_mpp

Note: See TracBrowser for help on using the repository browser.

Download in other formats: