Home

Context Navigation

source: palm/trunk/SOURCE/exchange_horiz.f90 @ 2213

Last change on this file since 2213 was 2119, checked in by raasch, 8 years ago
last commit documented
Property svn:keywords set to `Id`
File size: 13.0 KB

Rev	Line
[1682]	1	!> @file exchange_horiz.f90
[1320]	2	!------------------------------------------------------------------------------!
[1036]	3	! This file is part of PALM.
	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 2119 2017-01-17 16:51:50Z kanani $
	27	!
[2119]	28	! 2118 2017-01-17 16:38:49Z raasch
	29	! OpenACC directives and related code removed
	30	!
[2001]	31	! 2000 2016-08-20 18:09:15Z knoop
	32	! Forced header and separation lines into 80 columns
	33	!
[1805]	34	! 1804 2016-04-05 16:30:18Z maronga
	35	! Removed code for parameter file check (__check)
	36	!
[1683]	37	! 1682 2015-10-07 23:56:08Z knoop
	38	! Code annotations made doxygen readable
	39	!
[1678]	40	! 1677 2015-10-02 13:25:23Z boeske
	41	! Added new routine for exchange of three-dimensional integer arrays
	42	!
[1570]	43	! 1569 2015-03-12 07:54:38Z raasch
	44	! bugfix in background communication part
	45	!
[1349]	46	! 1348 2014-03-27 18:01:03Z raasch
[1569]	47	! bugfix: on_device added to ONLY-list
[1349]	48	!
[1345]	49	! 1344 2014-03-26 17:33:09Z kanani
	50	! Added missing parameters to ONLY-attribute
	51	!
[1321]	52	! 1320 2014-03-20 08:40:49Z raasch
[1320]	53	! ONLY-attribute added to USE-statements,
	54	! kind-parameters added to all INTEGER and REAL declaration statements,
	55	! kinds are defined in new module kinds,
	56	! revision history before 2012 removed,
	57	! comment fields (!:) to be used for variable explanations added to
	58	! all variable declaration statements
[668]	59	!
[1258]	60	! 1257 2013-11-08 15:18:40Z raasch
	61	! openacc loop and loop vector clauses removed
	62	!
[1132]	63	! 1128 2013-04-12 06:19:32Z raasch
	64	! modifications for asynchronous transfer,
	65	! local variables req, wait_stat are global now, and have been moved to module
	66	! pegrid
	67	!
[1114]	68	! 1113 2013-03-10 02:48:14Z raasch
	69	! GPU-porting for single-core (1PE) mode
	70	!
[1037]	71	! 1036 2012-10-22 13:43:42Z raasch
	72	! code put under GPL (PALM 3.9)
	73	!
[842]	74	! 841 2012-02-28 12:29:49Z maronga
	75	! Excluded routine from compilation of namelist_file_check
	76	!
[1]	77	! Revision 1.1 1997/07/24 11:13:29 raasch
	78	! Initial revision
	79	!
	80	!
	81	! Description:
	82	! ------------
[1682]	83	!> Exchange of lateral boundary values (parallel computers) and cyclic
	84	!> lateral boundary conditions, respectively.
[1]	85	!------------------------------------------------------------------------------!
[1682]	86	SUBROUTINE exchange_horiz( ar, nbgp_local)
	87
[1]	88
[1320]	89	USE control_parameters, &
[1344]	90	ONLY: bc_lr, bc_lr_cyc, bc_ns, bc_ns_cyc, grid_level, &
[2118]	91	mg_switch_to_pe0, synchronous_exchange
[1320]	92
	93	USE cpulog, &
	94	ONLY: cpu_log, log_point_s
	95
	96	USE indices, &
	97	ONLY: nxl, nxr, nyn, nys, nzb, nzt
	98
	99	USE kinds
	100
[1]	101	USE pegrid
	102
	103	IMPLICIT NONE
	104
[841]	105
[1682]	106	INTEGER(iwp) :: i !<
	107	INTEGER(iwp) :: j !<
	108	INTEGER(iwp) :: k !<
	109	INTEGER(iwp) :: nbgp_local !<
[1320]	110
	111	REAL(wp), DIMENSION(nzb:nzt+1,nys-nbgp_local:nyn+nbgp_local, &
[1682]	112	nxl-nbgp_local:nxr+nbgp_local) :: ar !<
[1320]	113
[841]	114
[1]	115	CALL cpu_log( log_point_s(2), 'exchange_horiz', 'start' )
	116
	117	#if defined( __parallel )
	118
	119	!
[1128]	120	!-- Exchange in x-direction of lateral boundaries
[1]	121	IF ( pdims(1) == 1 .OR. mg_switch_to_pe0 ) THEN
	122	!
	123	!-- One-dimensional decomposition along y, boundary values can be exchanged
	124	!-- within the PE memory
[707]	125	IF ( bc_lr_cyc ) THEN
[667]	126	ar(:,:,nxl-nbgp_local:nxl-1) = ar(:,:,nxr-nbgp_local+1:nxr)
	127	ar(:,:,nxr+1:nxr+nbgp_local) = ar(:,:,nxl:nxl+nbgp_local-1)
[1]	128	ENDIF
	129
	130	ELSE
[75]	131
[683]	132	IF ( synchronous_exchange ) THEN
[1]	133	!
[683]	134	!-- Send left boundary, receive right one (synchronous)
	135	CALL MPI_SENDRECV( &
[707]	136	ar(nzb,nys-nbgp_local,nxl), 1, type_yz(grid_level), pleft, 0, &
	137	ar(nzb,nys-nbgp_local,nxr+1), 1, type_yz(grid_level), pright, 0, &
	138	comm2d, status, ierr )
[1]	139	!
[683]	140	!-- Send right boundary, receive left one (synchronous)
[1320]	141	CALL MPI_SENDRECV( ar(nzb,nys-nbgp_local,nxr+1-nbgp_local), 1, &
	142	type_yz(grid_level), pright, 1, &
	143	ar(nzb,nys-nbgp_local,nxl-nbgp_local), 1, &
	144	type_yz(grid_level), pleft, 1, &
[707]	145	comm2d, status, ierr )
[667]	146
[683]	147	ELSE
[667]	148
[683]	149	!
[1128]	150	!-- In case of background communication switched on, exchange is done
	151	!-- either along x or along y
	152	IF ( send_receive == 'lr' .OR. send_receive == 'al' ) THEN
	153
	154	IF ( .NOT. sendrecv_in_background ) THEN
	155	req(1:4) = 0
	156	req_count = 0
	157	ENDIF
[683]	158	!
[1128]	159	!-- Send left boundary, receive right one (asynchronous)
	160	CALL MPI_ISEND( ar(nzb,nys-nbgp_local,nxl), 1, type_yz(grid_level), &
	161	pleft, req_count, comm2d, req(req_count+1), ierr )
	162	CALL MPI_IRECV( ar(nzb,nys-nbgp_local,nxr+1), 1, type_yz(grid_level), &
	163	pright, req_count, comm2d, req(req_count+2), ierr )
	164	!
	165	!-- Send right boundary, receive left one (asynchronous)
[1320]	166	CALL MPI_ISEND( ar(nzb,nys-nbgp_local,nxr+1-nbgp_local), 1, &
	167	type_yz(grid_level), pright, req_count+1, comm2d, &
[1128]	168	req(req_count+3), ierr )
[1320]	169	CALL MPI_IRECV( ar(nzb,nys-nbgp_local,nxl-nbgp_local), 1, &
	170	type_yz(grid_level), pleft, req_count+1, comm2d, &
[1128]	171	req(req_count+4), ierr )
[667]	172
[1128]	173	IF ( .NOT. sendrecv_in_background ) THEN
	174	CALL MPI_WAITALL( 4, req, wait_stat, ierr )
	175	ELSE
	176	req_count = req_count + 4
	177	ENDIF
[75]	178
[1128]	179	ENDIF
	180
[683]	181	ENDIF
	182
[1]	183	ENDIF
	184
	185
	186	IF ( pdims(2) == 1 .OR. mg_switch_to_pe0 ) THEN
	187	!
	188	!-- One-dimensional decomposition along x, boundary values can be exchanged
	189	!-- within the PE memory
[707]	190	IF ( bc_ns_cyc ) THEN
[667]	191	ar(:,nys-nbgp_local:nys-1,:) = ar(:,nyn-nbgp_local+1:nyn,:)
	192	ar(:,nyn+1:nyn+nbgp_local,:) = ar(:,nys:nys+nbgp_local-1,:)
[1]	193	ENDIF
	194
	195	ELSE
	196
[683]	197	IF ( synchronous_exchange ) THEN
[1]	198	!
[683]	199	!-- Send front boundary, receive rear one (synchronous)
	200	CALL MPI_SENDRECV( &
[707]	201	ar(nzb,nys,nxl-nbgp_local), 1, type_xz(grid_level), psouth, 0, &
	202	ar(nzb,nyn+1,nxl-nbgp_local), 1, type_xz(grid_level), pnorth, 0, &
	203	comm2d, status, ierr )
[683]	204	!
	205	!-- Send rear boundary, receive front one (synchronous)
[1320]	206	CALL MPI_SENDRECV( ar(nzb,nyn-nbgp_local+1,nxl-nbgp_local), 1, &
	207	type_xz(grid_level), pnorth, 1, &
	208	ar(nzb,nys-nbgp_local,nxl-nbgp_local), 1, &
	209	type_xz(grid_level), psouth, 1, &
[707]	210	comm2d, status, ierr )
[667]	211
[683]	212	ELSE
	213
[1]	214	!
[1128]	215	!-- In case of background communication switched on, exchange is done
	216	!-- either along x or along y
[1569]	217	IF ( send_receive == 'ns' .OR. send_receive == 'al' ) THEN
[1128]	218
	219	IF ( .NOT. sendrecv_in_background ) THEN
	220	req(1:4) = 0
	221	req_count = 0
	222	ENDIF
	223
[683]	224	!
[1128]	225	!-- Send front boundary, receive rear one (asynchronous)
	226	CALL MPI_ISEND( ar(nzb,nys,nxl-nbgp_local), 1, type_xz(grid_level), &
	227	psouth, req_count, comm2d, req(req_count+1), ierr )
	228	CALL MPI_IRECV( ar(nzb,nyn+1,nxl-nbgp_local), 1, type_xz(grid_level), &
	229	pnorth, req_count, comm2d, req(req_count+2), ierr )
	230	!
	231	!-- Send rear boundary, receive front one (asynchronous)
[1320]	232	CALL MPI_ISEND( ar(nzb,nyn-nbgp_local+1,nxl-nbgp_local), 1, &
	233	type_xz(grid_level), pnorth, req_count+1, comm2d, &
[1128]	234	req(req_count+3), ierr )
[1320]	235	CALL MPI_IRECV( ar(nzb,nys-nbgp_local,nxl-nbgp_local), 1, &
	236	type_xz(grid_level), psouth, req_count+1, comm2d, &
[1128]	237	req(req_count+4), ierr )
[75]	238
[1128]	239	IF ( .NOT. sendrecv_in_background ) THEN
	240	CALL MPI_WAITALL( 4, req, wait_stat, ierr )
	241	ELSE
	242	req_count = req_count + 4
	243	ENDIF
[683]	244
[1128]	245	ENDIF
	246
[683]	247	ENDIF
	248
[1]	249	ENDIF
	250
	251	#else
	252
	253	!
[1113]	254	!-- Lateral boundary conditions in the non-parallel case.
	255	!-- Case dependent, because in GPU mode still not all arrays are on device. This
	256	!-- workaround has to be removed later. Also, since PGI compiler 12.5 has problems
	257	!-- with array syntax, explicit loops are used.
[1]	258	IF ( bc_lr == 'cyclic' ) THEN
[2118]	259	ar(:,:,nxl-nbgp_local:nxl-1) = ar(:,:,nxr-nbgp_local+1:nxr)
	260	ar(:,:,nxr+1:nxr+nbgp_local) = ar(:,:,nxl:nxl+nbgp_local-1)
[1]	261	ENDIF
	262
	263	IF ( bc_ns == 'cyclic' ) THEN
[2118]	264	ar(:,nys-nbgp_local:nys-1,:) = ar(:,nyn-nbgp_local+1:nyn,:)
	265	ar(:,nyn+1:nyn+nbgp_local,:) = ar(:,nys:nys+nbgp_local-1,:)
[1]	266	ENDIF
	267
	268	#endif
	269	CALL cpu_log( log_point_s(2), 'exchange_horiz', 'stop' )
	270
	271	END SUBROUTINE exchange_horiz
[1677]	272
	273
[1682]	274	!------------------------------------------------------------------------------!
	275	! Description:
	276	! ------------
	277	!> @todo Missing subroutine description.
	278	!------------------------------------------------------------------------------!
[1677]	279	SUBROUTINE exchange_horiz_int( ar, nbgp_local)
	280
	281	USE control_parameters, &
	282	ONLY: bc_lr, bc_lr_cyc, bc_ns, bc_ns_cyc
	283
	284	USE indices, &
	285	ONLY: nxl, nxr, nyn, nys, nzb, nzt
	286
	287	USE kinds
	288
	289	USE pegrid
	290
	291	IMPLICIT NONE
	292
	293
[1682]	294	INTEGER(iwp) :: nbgp_local !< number of ghost points
[1677]	295
	296	INTEGER(iwp), DIMENSION(nzb:nzt+1,nys-nbgp_local:nyn+nbgp_local, &
[1682]	297	nxl-nbgp_local:nxr+nbgp_local) :: ar !< treated array
[1677]	298
	299
	300	#if defined( __parallel )
	301	IF ( pdims(1) == 1 ) THEN
	302	!
	303	!-- One-dimensional decomposition along y, boundary values can be exchanged
	304	!-- within the PE memory
	305	IF ( bc_lr_cyc ) THEN
	306	ar(:,:,nxl-nbgp_local:nxl-1) = ar(:,:,nxr-nbgp_local+1:nxr)
	307	ar(:,:,nxr+1:nxr+nbgp_local) = ar(:,:,nxl:nxl+nbgp_local-1)
	308	ENDIF
	309	ELSE
	310	!
	311	!-- Send left boundary, receive right one (synchronous)
	312	CALL MPI_SENDRECV( &
	313	ar(nzb,nys-nbgp_local,nxl), 1, type_yz_int, pleft, 0, &
	314	ar(nzb,nys-nbgp_local,nxr+1), 1, type_yz_int, pright, 0, &
	315	comm2d, status, ierr )
	316	!
	317	!-- Send right boundary, receive left one (synchronous)
	318	CALL MPI_SENDRECV( &
	319	ar(nzb,nys-nbgp_local,nxr+1-nbgp_local), 1, type_yz_int, pright, 1, &
	320	ar(nzb,nys-nbgp_local,nxl-nbgp_local), 1, type_yz_int, pleft, 1, &
	321	comm2d, status, ierr )
	322	ENDIF
	323
	324
	325	IF ( pdims(2) == 1 ) THEN
	326	!
	327	!-- One-dimensional decomposition along x, boundary values can be exchanged
	328	!-- within the PE memory
	329	IF ( bc_ns_cyc ) THEN
	330	ar(:,nys-nbgp_local:nys-1,:) = ar(:,nyn-nbgp_local+1:nyn,:)
	331	ar(:,nyn+1:nyn+nbgp_local,:) = ar(:,nys:nys+nbgp_local-1,:)
	332	ENDIF
	333
	334	ELSE
	335
	336	!
	337	!-- Send front boundary, receive rear one (synchronous)
	338	CALL MPI_SENDRECV( &
	339	ar(nzb,nys,nxl-nbgp_local), 1, type_xz_int, psouth, 0, &
	340	ar(nzb,nyn+1,nxl-nbgp_local), 1, type_xz_int, pnorth, 0, &
	341	comm2d, status, ierr )
	342	!
	343	!-- Send rear boundary, receive front one (synchronous)
	344	CALL MPI_SENDRECV( ar(nzb,nyn-nbgp_local+1,nxl-nbgp_local), 1, &
	345	type_xz_int, pnorth, 1, &
	346	ar(nzb,nys-nbgp_local,nxl-nbgp_local), 1, &
	347	type_xz_int, psouth, 1, &
	348	comm2d, status, ierr )
	349
	350	ENDIF
	351
	352	#else
	353
	354	IF ( bc_lr == 'cyclic' ) THEN
	355	ar(:,:,nxl-nbgp_local:nxl-1) = ar(:,:,nxr-nbgp_local+1:nxr)
	356	ar(:,:,nxr+1:nxr+nbgp_local) = ar(:,:,nxl:nxl+nbgp_local-1)
	357	ENDIF
	358
	359	IF ( bc_ns == 'cyclic' ) THEN
	360	ar(:,nys-nbgp_local:nys-1,:) = ar(:,nyn-nbgp_local+1:nyn,:)
	361	ar(:,nyn+1:nyn+nbgp_local,:) = ar(:,nys:nys+nbgp_local-1,:)
	362	ENDIF
	363
	364	#endif
	365
	366
[1933]	367	END SUBROUTINE exchange_horiz_int

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |