Home

Context Navigation

source: palm/trunk/SOURCE/exchange_horiz.f90 @ 2709

Last change on this file since 2709 was 2696, checked in by kanani, 7 years ago
Merge of branch palm4u into trunk
Property svn:keywords set to `Id`
File size: 13.4 KB

Rev	Line
[1682]	1	!> @file exchange_horiz.f90
[1320]	2	!------------------------------------------------------------------------------!
[2696]	3	! This file is part of the PALM model system.
[1036]	4	!
[2000]	5	! PALM is free software: you can redistribute it and/or modify it under the
	6	! terms of the GNU General Public License as published by the Free Software
	7	! Foundation, either version 3 of the License, or (at your option) any later
	8	! version.
[1036]	9	!
	10	! PALM is distributed in the hope that it will be useful, but WITHOUT ANY
	11	! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
	12	! A PARTICULAR PURPOSE. See the GNU General Public License for more details.
	13	!
	14	! You should have received a copy of the GNU General Public License along with
	15	! PALM. If not, see <http://www.gnu.org/licenses/>.
	16	!
[2101]	17	! Copyright 1997-2017 Leibniz Universitaet Hannover
[1320]	18	!------------------------------------------------------------------------------!
[1036]	19	!
[484]	20	! Current revisions:
[1]	21	! -----------------
[1678]	22	!
[2119]	23	!
[1321]	24	! Former revisions:
	25	! -----------------
	26	! $Id: exchange_horiz.f90 2696 2017-12-14 17:12:51Z suehring $
[2696]	27	! 3D-Integer exchange on multigrid level (MS)
	28	!
	29	! 2298 2017-06-29 09:28:18Z raasch
[2298]	30	! sendrecv_in_background related parts removed
	31	!
	32	! 2119 2017-01-17 16:51:50Z raasch
[1321]	33	!
[2119]	34	! 2118 2017-01-17 16:38:49Z raasch
	35	! OpenACC directives and related code removed
	36	!
[2001]	37	! 2000 2016-08-20 18:09:15Z knoop
	38	! Forced header and separation lines into 80 columns
	39	!
[1805]	40	! 1804 2016-04-05 16:30:18Z maronga
	41	! Removed code for parameter file check (__check)
	42	!
[1683]	43	! 1682 2015-10-07 23:56:08Z knoop
	44	! Code annotations made doxygen readable
	45	!
[1678]	46	! 1677 2015-10-02 13:25:23Z boeske
	47	! Added new routine for exchange of three-dimensional integer arrays
	48	!
[1570]	49	! 1569 2015-03-12 07:54:38Z raasch
	50	! bugfix in background communication part
	51	!
[1349]	52	! 1348 2014-03-27 18:01:03Z raasch
[1569]	53	! bugfix: on_device added to ONLY-list
[1349]	54	!
[1345]	55	! 1344 2014-03-26 17:33:09Z kanani
	56	! Added missing parameters to ONLY-attribute
	57	!
[1321]	58	! 1320 2014-03-20 08:40:49Z raasch
[1320]	59	! ONLY-attribute added to USE-statements,
	60	! kind-parameters added to all INTEGER and REAL declaration statements,
	61	! kinds are defined in new module kinds,
	62	! revision history before 2012 removed,
	63	! comment fields (!:) to be used for variable explanations added to
	64	! all variable declaration statements
[668]	65	!
[1258]	66	! 1257 2013-11-08 15:18:40Z raasch
	67	! openacc loop and loop vector clauses removed
	68	!
[1132]	69	! 1128 2013-04-12 06:19:32Z raasch
	70	! modifications for asynchronous transfer,
	71	! local variables req, wait_stat are global now, and have been moved to module
	72	! pegrid
	73	!
[1114]	74	! 1113 2013-03-10 02:48:14Z raasch
	75	! GPU-porting for single-core (1PE) mode
	76	!
[1037]	77	! 1036 2012-10-22 13:43:42Z raasch
	78	! code put under GPL (PALM 3.9)
	79	!
[842]	80	! 841 2012-02-28 12:29:49Z maronga
	81	! Excluded routine from compilation of namelist_file_check
	82	!
[1]	83	! Revision 1.1 1997/07/24 11:13:29 raasch
	84	! Initial revision
	85	!
	86	!
	87	! Description:
	88	! ------------
[1682]	89	!> Exchange of lateral boundary values (parallel computers) and cyclic
	90	!> lateral boundary conditions, respectively.
[1]	91	!------------------------------------------------------------------------------!
[1682]	92	SUBROUTINE exchange_horiz( ar, nbgp_local)
	93
[1]	94
[1320]	95	USE control_parameters, &
[1344]	96	ONLY: bc_lr, bc_lr_cyc, bc_ns, bc_ns_cyc, grid_level, &
[2118]	97	mg_switch_to_pe0, synchronous_exchange
[1320]	98
	99	USE cpulog, &
	100	ONLY: cpu_log, log_point_s
	101
	102	USE indices, &
	103	ONLY: nxl, nxr, nyn, nys, nzb, nzt
	104
	105	USE kinds
	106
[1]	107	USE pegrid
	108
	109	IMPLICIT NONE
	110
[841]	111
[1682]	112	INTEGER(iwp) :: i !<
	113	INTEGER(iwp) :: j !<
	114	INTEGER(iwp) :: k !<
	115	INTEGER(iwp) :: nbgp_local !<
[1320]	116
	117	REAL(wp), DIMENSION(nzb:nzt+1,nys-nbgp_local:nyn+nbgp_local, &
[1682]	118	nxl-nbgp_local:nxr+nbgp_local) :: ar !<
[1320]	119
[841]	120
[1]	121	CALL cpu_log( log_point_s(2), 'exchange_horiz', 'start' )
	122
	123	#if defined( __parallel )
	124
	125	!
[1128]	126	!-- Exchange in x-direction of lateral boundaries
[1]	127	IF ( pdims(1) == 1 .OR. mg_switch_to_pe0 ) THEN
	128	!
	129	!-- One-dimensional decomposition along y, boundary values can be exchanged
	130	!-- within the PE memory
[707]	131	IF ( bc_lr_cyc ) THEN
[667]	132	ar(:,:,nxl-nbgp_local:nxl-1) = ar(:,:,nxr-nbgp_local+1:nxr)
	133	ar(:,:,nxr+1:nxr+nbgp_local) = ar(:,:,nxl:nxl+nbgp_local-1)
[1]	134	ENDIF
	135
	136	ELSE
[75]	137
[683]	138	IF ( synchronous_exchange ) THEN
[1]	139	!
[683]	140	!-- Send left boundary, receive right one (synchronous)
	141	CALL MPI_SENDRECV( &
[707]	142	ar(nzb,nys-nbgp_local,nxl), 1, type_yz(grid_level), pleft, 0, &
	143	ar(nzb,nys-nbgp_local,nxr+1), 1, type_yz(grid_level), pright, 0, &
	144	comm2d, status, ierr )
[1]	145	!
[683]	146	!-- Send right boundary, receive left one (synchronous)
[1320]	147	CALL MPI_SENDRECV( ar(nzb,nys-nbgp_local,nxr+1-nbgp_local), 1, &
	148	type_yz(grid_level), pright, 1, &
	149	ar(nzb,nys-nbgp_local,nxl-nbgp_local), 1, &
	150	type_yz(grid_level), pleft, 1, &
[707]	151	comm2d, status, ierr )
[667]	152
[683]	153	ELSE
[667]	154
[683]	155	!
[2298]	156	!-- Asynchroneous exchange
[1128]	157	IF ( send_receive == 'lr' .OR. send_receive == 'al' ) THEN
	158
[2298]	159	req(1:4) = 0
	160	req_count = 0
[683]	161	!
[1128]	162	!-- Send left boundary, receive right one (asynchronous)
	163	CALL MPI_ISEND( ar(nzb,nys-nbgp_local,nxl), 1, type_yz(grid_level), &
	164	pleft, req_count, comm2d, req(req_count+1), ierr )
	165	CALL MPI_IRECV( ar(nzb,nys-nbgp_local,nxr+1), 1, type_yz(grid_level), &
	166	pright, req_count, comm2d, req(req_count+2), ierr )
	167	!
	168	!-- Send right boundary, receive left one (asynchronous)
[1320]	169	CALL MPI_ISEND( ar(nzb,nys-nbgp_local,nxr+1-nbgp_local), 1, &
	170	type_yz(grid_level), pright, req_count+1, comm2d, &
[1128]	171	req(req_count+3), ierr )
[1320]	172	CALL MPI_IRECV( ar(nzb,nys-nbgp_local,nxl-nbgp_local), 1, &
	173	type_yz(grid_level), pleft, req_count+1, comm2d, &
[1128]	174	req(req_count+4), ierr )
[667]	175
[2298]	176	CALL MPI_WAITALL( 4, req, wait_stat, ierr )
[75]	177
[1128]	178	ENDIF
	179
[683]	180	ENDIF
	181
[1]	182	ENDIF
	183
	184
	185	IF ( pdims(2) == 1 .OR. mg_switch_to_pe0 ) THEN
	186	!
	187	!-- One-dimensional decomposition along x, boundary values can be exchanged
	188	!-- within the PE memory
[707]	189	IF ( bc_ns_cyc ) THEN
[667]	190	ar(:,nys-nbgp_local:nys-1,:) = ar(:,nyn-nbgp_local+1:nyn,:)
	191	ar(:,nyn+1:nyn+nbgp_local,:) = ar(:,nys:nys+nbgp_local-1,:)
[1]	192	ENDIF
	193
	194	ELSE
	195
[683]	196	IF ( synchronous_exchange ) THEN
[1]	197	!
[683]	198	!-- Send front boundary, receive rear one (synchronous)
	199	CALL MPI_SENDRECV( &
[707]	200	ar(nzb,nys,nxl-nbgp_local), 1, type_xz(grid_level), psouth, 0, &
	201	ar(nzb,nyn+1,nxl-nbgp_local), 1, type_xz(grid_level), pnorth, 0, &
	202	comm2d, status, ierr )
[683]	203	!
	204	!-- Send rear boundary, receive front one (synchronous)
[1320]	205	CALL MPI_SENDRECV( ar(nzb,nyn-nbgp_local+1,nxl-nbgp_local), 1, &
	206	type_xz(grid_level), pnorth, 1, &
	207	ar(nzb,nys-nbgp_local,nxl-nbgp_local), 1, &
	208	type_xz(grid_level), psouth, 1, &
[707]	209	comm2d, status, ierr )
[667]	210
[683]	211	ELSE
	212
[1]	213	!
[2298]	214	!-- Asynchroneous exchange
[1569]	215	IF ( send_receive == 'ns' .OR. send_receive == 'al' ) THEN
[1128]	216
[2298]	217	req(1:4) = 0
	218	req_count = 0
[1128]	219
[683]	220	!
[1128]	221	!-- Send front boundary, receive rear one (asynchronous)
	222	CALL MPI_ISEND( ar(nzb,nys,nxl-nbgp_local), 1, type_xz(grid_level), &
	223	psouth, req_count, comm2d, req(req_count+1), ierr )
	224	CALL MPI_IRECV( ar(nzb,nyn+1,nxl-nbgp_local), 1, type_xz(grid_level), &
	225	pnorth, req_count, comm2d, req(req_count+2), ierr )
	226	!
	227	!-- Send rear boundary, receive front one (asynchronous)
[1320]	228	CALL MPI_ISEND( ar(nzb,nyn-nbgp_local+1,nxl-nbgp_local), 1, &
	229	type_xz(grid_level), pnorth, req_count+1, comm2d, &
[1128]	230	req(req_count+3), ierr )
[1320]	231	CALL MPI_IRECV( ar(nzb,nys-nbgp_local,nxl-nbgp_local), 1, &
	232	type_xz(grid_level), psouth, req_count+1, comm2d, &
[1128]	233	req(req_count+4), ierr )
[75]	234
[2298]	235	CALL MPI_WAITALL( 4, req, wait_stat, ierr )
[683]	236
[1128]	237	ENDIF
	238
[683]	239	ENDIF
	240
[1]	241	ENDIF
	242
	243	#else
	244
	245	!
[1113]	246	!-- Lateral boundary conditions in the non-parallel case.
	247	!-- Case dependent, because in GPU mode still not all arrays are on device. This
	248	!-- workaround has to be removed later. Also, since PGI compiler 12.5 has problems
	249	!-- with array syntax, explicit loops are used.
[1]	250	IF ( bc_lr == 'cyclic' ) THEN
[2118]	251	ar(:,:,nxl-nbgp_local:nxl-1) = ar(:,:,nxr-nbgp_local+1:nxr)
	252	ar(:,:,nxr+1:nxr+nbgp_local) = ar(:,:,nxl:nxl+nbgp_local-1)
[1]	253	ENDIF
	254
	255	IF ( bc_ns == 'cyclic' ) THEN
[2118]	256	ar(:,nys-nbgp_local:nys-1,:) = ar(:,nyn-nbgp_local+1:nyn,:)
	257	ar(:,nyn+1:nyn+nbgp_local,:) = ar(:,nys:nys+nbgp_local-1,:)
[1]	258	ENDIF
	259
	260	#endif
	261	CALL cpu_log( log_point_s(2), 'exchange_horiz', 'stop' )
	262
	263	END SUBROUTINE exchange_horiz
[1677]	264
	265
[1682]	266	!------------------------------------------------------------------------------!
	267	! Description:
	268	! ------------
	269	!> @todo Missing subroutine description.
	270	!------------------------------------------------------------------------------!
[2696]	271	SUBROUTINE exchange_horiz_int( ar, nys_l, nyn_l, nxl_l, nxr_l, nzt_l, nbgp_local)
[1677]	272
	273	USE control_parameters, &
[2696]	274	ONLY: bc_lr, bc_lr_cyc, bc_ns, bc_ns_cyc, grid_level
[1677]	275
	276	USE indices, &
	277	ONLY: nxl, nxr, nyn, nys, nzb, nzt
	278
	279	USE kinds
	280
	281	USE pegrid
	282
	283	IMPLICIT NONE
	284
[2696]	285	INTEGER(iwp) :: nxl_l !< local index bound at current grid level, left side
	286	INTEGER(iwp) :: nxr_l !< local index bound at current grid level, right side
	287	INTEGER(iwp) :: nyn_l !< local index bound at current grid level, north side
	288	INTEGER(iwp) :: nys_l !< local index bound at current grid level, south side
	289	INTEGER(iwp) :: nzt_l !< local index bound at current grid level, top
[1682]	290	INTEGER(iwp) :: nbgp_local !< number of ghost points
[1677]	291
[2696]	292	INTEGER(iwp), DIMENSION(nzb:nzt_l+1,nys_l-nbgp_local:nyn_l+nbgp_local, &
	293	nxl_l-nbgp_local:nxr_l+nbgp_local) :: ar !< treated array
[1677]	294
	295
	296	#if defined( __parallel )
	297	IF ( pdims(1) == 1 ) THEN
	298	!
	299	!-- One-dimensional decomposition along y, boundary values can be exchanged
	300	!-- within the PE memory
	301	IF ( bc_lr_cyc ) THEN
[2696]	302	ar(:,:,nxl_l-nbgp_local:nxl_l-1) = ar(:,:,nxr_l-nbgp_local+1:nxr_l)
	303	ar(:,:,nxr_l+1:nxr_l+nbgp_local) = ar(:,:,nxl_l:nxl_l+nbgp_local-1)
[1677]	304	ENDIF
	305	ELSE
	306	!
	307	!-- Send left boundary, receive right one (synchronous)
[2696]	308	CALL MPI_SENDRECV( &
	309	ar(nzb,nys_l-nbgp_local,nxl_l), 1, type_yz_int(grid_level), pleft, 0,&
	310	ar(nzb,nys_l-nbgp_local,nxr_l+1), 1, type_yz_int(grid_level), pright, 0,&
[1677]	311	comm2d, status, ierr )
	312	!
	313	!-- Send right boundary, receive left one (synchronous)
[2696]	314	CALL MPI_SENDRECV( &
	315	ar(nzb,nys_l-nbgp_local,nxr_l+1-nbgp_local), 1, type_yz_int(grid_level),&
	316	pright, 1, &
	317	ar(nzb,nys_l-nbgp_local,nxl_l-nbgp_local), 1, type_yz_int(grid_level),&
	318	pleft, 1, &
[1677]	319	comm2d, status, ierr )
	320	ENDIF
	321
	322
	323	IF ( pdims(2) == 1 ) THEN
	324	!
	325	!-- One-dimensional decomposition along x, boundary values can be exchanged
	326	!-- within the PE memory
	327	IF ( bc_ns_cyc ) THEN
[2696]	328	ar(:,nys_l-nbgp_local:nys_l-1,:) = ar(:,nyn_l-nbgp_local+1:nyn_l,:)
	329	ar(:,nyn_l+1:nyn_l+nbgp_local,:) = ar(:,nys_l:nys_l+nbgp_local-1,:)
[1677]	330	ENDIF
	331
	332	ELSE
	333
	334	!
	335	!-- Send front boundary, receive rear one (synchronous)
[2696]	336	CALL MPI_SENDRECV( &
	337	ar(nzb,nys_l,nxl_l-nbgp_local), 1, type_xz_int(grid_level), psouth, 0,&
	338	ar(nzb,nyn_l+1,nxl_l-nbgp_local), 1, type_xz_int(grid_level), pnorth, 0,&
[1677]	339	comm2d, status, ierr )
	340	!
	341	!-- Send rear boundary, receive front one (synchronous)
[2696]	342	CALL MPI_SENDRECV( ar(nzb,nyn_l-nbgp_local+1,nxl_l-nbgp_local), 1, &
	343	type_xz_int(grid_level), pnorth, 1, &
	344	ar(nzb,nys_l-nbgp_local,nxl_l-nbgp_local), 1, &
	345	type_xz_int(grid_level), psouth, 1, &
[1677]	346	comm2d, status, ierr )
	347
	348	ENDIF
	349
	350	#else
	351
	352	IF ( bc_lr == 'cyclic' ) THEN
[2696]	353	ar(:,:,nxl_l-nbgp_local:nxl_l-1) = ar(:,:,nxr_l-nbgp_local+1:nxr_l)
	354	ar(:,:,nxr_l+1:nxr_l+nbgp_local) = ar(:,:,nxl_l:nxl_l+nbgp_local-1)
[1677]	355	ENDIF
	356
	357	IF ( bc_ns == 'cyclic' ) THEN
[2696]	358	ar(:,nys_l-nbgp_local:nys_l-1,:) = ar(:,nyn_l-nbgp_local+1:nyn_l,:)
	359	ar(:,nyn_l+1:nyn_l+nbgp_local,:) = ar(:,nys_l:nys_l+nbgp_local-1,:)
[1677]	360	ENDIF
	361
	362	#endif
	363
	364
[1933]	365	END SUBROUTINE exchange_horiz_int

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |