Home

Context Navigation

source: palm/trunk/SOURCE/poismg.f90 @ 1433

Last change on this file since 1433 was 1354, checked in by heinze, 11 years ago
last commit documented
Property svn:keywords set to `Id`
File size: 64.7 KB

Rev	Line
[1]	1	SUBROUTINE poismg( r )
	2
[1036]	3	!--------------------------------------------------------------------------------!
	4	! This file is part of PALM.
	5	!
	6	! PALM is free software: you can redistribute it and/or modify it under the terms
	7	! of the GNU General Public License as published by the Free Software Foundation,
	8	! either version 3 of the License, or (at your option) any later version.
	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	!
[1310]	17	! Copyright 1997-2014 Leibniz Universitaet Hannover
[1036]	18	!--------------------------------------------------------------------------------!
	19	!
[1]	20	! Attention: Loop unrolling and cache optimization in SOR-Red/Black method
[707]	21	! still does not give the expected speedup! Further work required.
[1]	22	!
[257]	23	! Current revisions:
[1]	24	! -----------------
[1354]	25	!
	26	!
[1321]	27	! Former revisions:
	28	! -----------------
	29	! $Id: poismg.f90 1354 2014-04-08 15:22:57Z keck $
	30	!
[1354]	31	! 1353 2014-04-08 15:21:23Z heinze
	32	! REAL constants provided with KIND-attribute
	33	!
[1323]	34	! 1322 2014-03-20 16:38:49Z raasch
	35	! REAL constants defined as wp-kind
	36	!
[1321]	37	! 1320 2014-03-20 08:40:49Z raasch
[1320]	38	! ONLY-attribute added to USE-statements,
	39	! kind-parameters added to all INTEGER and REAL declaration statements,
	40	! kinds are defined in new module kinds,
	41	! old module precision_kind is removed,
	42	! revision history before 2012 removed,
	43	! comment fields (!:) to be used for variable explanations added to
	44	! all variable declaration statements
[708]	45	!
[1319]	46	! 1318 2014-03-17 13:35:16Z raasch
	47	! module interfaces removed
	48	!
[1160]	49	! 1159 2013-05-21 11:58:22Z fricke
	50	! bc_lr/ns_dirneu/neudir removed
	51	!
[1093]	52	! 1092 2013-02-02 11:24:22Z raasch
	53	! unused variables removed
	54	!
[1057]	55	! 1056 2012-11-16 15:28:04Z raasch
	56	! Bugfix: all ghost points have to be used for allocating p3
	57	! arrays p2, f2, and f2_l changed from allocatable to automatic
	58	!
[1037]	59	! 1036 2012-10-22 13:43:42Z raasch
	60	! code put under GPL (PALM 3.9)
	61	!
[997]	62	! 996 2012-09-07 10:41:47Z raasch
	63	! little reformatting
	64	!
[979]	65	! 978 2012-08-09 08:28:32Z fricke
	66	! bc_lr/ns_dirneu/neudir added
	67	!
[881]	68	! 880 2012-04-13 06:28:59Z raasch
	69	! Bugfix: preprocessor statements for parallel execution added
	70	!
[779]	71	! 778 2011-11-07 14:18:25Z fricke
	72	! Allocation of p3 changes when multigrid is used and the collected field on PE0
	73	! has more grid points than the subdomain of an PE.
	74	!
[708]	75	! 707 2011-03-29 11:39:40Z raasch
[707]	76	! p_loc is used instead of p in the main routine (poismg).
	77	! On coarse grid levels, gathered data are identically processed on all PEs
	78	! (before, on PE0 only), so that the subsequent scattering of data is not
	79	! neccessary any more.
	80	! bc_lr/ns replaced by bc_lr/ns_cyc/dirrad/raddir
	81	! Bugfix: bottom (nzb) and top (nzt+1) boundary conditions set in routines
	82	! resid and restrict. They were missed before which may have led to
	83	! unpredictable results.
[1]	84	!
[668]	85	! 667 2010-12-23 12:06:00Z suehring/gryschka
	86	! Calls of exchange_horiz are modified.
	87	!
[623]	88	! 622 2010-12-10 08:08:13Z raasch
	89	! optional barriers included in order to speed up collective operations
	90	!
[392]	91	! 257 2009-03-11 15:17:42Z heinze
	92	! Output of messages replaced by message handling routine.
	93	!
[198]	94	! 181 2008-07-30 07:07:47Z raasch
	95	! Bugfix: grid_level+1 has to be used in restrict for flags-array
	96	!
[139]	97	! 114 2007-10-10 00:03:15Z raasch
	98	! Boundary conditions at walls are implicitly set using flag arrays. Only
	99	! Neumann BC is allowed. Upper walls are still not realized.
	100	! Bottom and top BCs for array f_mg in restrict removed because boundary
	101	! values are not needed (right hand side of SOR iteration).
	102	!
[77]	103	! 75 2007-03-22 09:54:05Z raasch
	104	! 2nd+3rd argument removed from exchange horiz
	105	!
[3]	106	! RCS Log replace by Id keyword, revision history cleaned up
	107	!
[1]	108	! Revision 1.6 2005/03/26 20:55:54 raasch
	109	! Implementation of non-cyclic (Neumann) horizontal boundary conditions,
	110	! routine prolong simplified (one call of exchange_horiz spared)
	111	!
	112	! Revision 1.1 2001/07/20 13:10:51 raasch
	113	! Initial revision
	114	!
	115	!
	116	! Description:
	117	! ------------
	118	! Solves the Poisson equation for the perturbation pressure with a multigrid
	119	! V- or W-Cycle scheme.
	120	!
	121	! This multigrid method was originally developed for PALM by Joerg Uhlenbrock,
	122	! September 2000 - July 2001.
	123	!------------------------------------------------------------------------------!
	124
[1320]	125	USE arrays_3d, &
	126	ONLY: d, p_loc
	127
	128	USE control_parameters, &
	129	ONLY: gathered_size, grid_level, grid_level_count, &
	130	maximum_grid_level, message_string, mgcycles, mg_cycles, &
	131	mg_switch_to_pe0_level, residual_limit, subdomain_size
	132
	133	USE cpulog, &
	134	ONLY: cpu_log, log_point_s
	135
	136	USE indices, &
	137	ONLY: nxl, nxlg, nxl_mg, nxr, nxrg, nxr_mg, nys, nysg, nys_mg, nyn, &
	138	nyng, nyn_mg, nzb, nzt, nzt_mg
	139
	140	USE kinds
	141
[1]	142	USE pegrid
	143
	144	IMPLICIT NONE
	145
[1320]	146	REAL(wp) :: maxerror !:
	147	REAL(wp) :: maximum_mgcycles !:
	148	REAL(wp) :: residual_norm !:
[1]	149
[1320]	150	REAL(wp), DIMENSION(nzb:nzt+1,nys-1:nyn+1,nxl-1:nxr+1) :: r !:
[1]	151
[1320]	152	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: p3 !:
[1]	153
	154
	155	CALL cpu_log( log_point_s(29), 'poismg', 'start' )
	156	!
	157	!-- Initialize arrays and variables used in this subroutine
	158
[879]	159	!-- If the number of grid points of the gathered grid, which is collected
	160	!-- on PE0, is larger than the number of grid points of an PE, than array
	161	!-- p3 will be enlarged.
[778]	162	IF ( gathered_size > subdomain_size ) THEN
[879]	163	ALLOCATE( p3(nzb:nzt_mg(mg_switch_to_pe0_level)+1,nys_mg( &
	164	mg_switch_to_pe0_level)-1:nyn_mg(mg_switch_to_pe0_level)+1,&
	165	nxl_mg(mg_switch_to_pe0_level)-1:nxr_mg( &
[778]	166	mg_switch_to_pe0_level)+1) )
	167	ELSE
[1056]	168	ALLOCATE ( p3(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
[778]	169	ENDIF
[1056]	170
[1353]	171	p3 = 0.0_wp
[879]	172
[1]	173	!
[707]	174	!-- Ghost boundaries have to be added to divergence array.
	175	!-- Exchange routine needs to know the grid level!
	176	grid_level = maximum_grid_level
[667]	177	CALL exchange_horiz( d, 1)
[1]	178	d(nzb,:,:) = d(nzb+1,:,:)
	179
	180	!
	181	!-- Initiation of the multigrid scheme. Does n cycles until the
	182	!-- residual is smaller than the given limit. The accuracy of the solution
	183	!-- of the poisson equation will increase with the number of cycles.
	184	!-- If the number of cycles is preset by the user, this number will be
	185	!-- carried out regardless of the accuracy.
[707]	186	grid_level_count = 0
	187	mgcycles = 0
[1]	188	IF ( mg_cycles == -1 ) THEN
	189	maximum_mgcycles = 0
[1353]	190	residual_norm = 1.0_wp
[1]	191	ELSE
	192	maximum_mgcycles = mg_cycles
[1353]	193	residual_norm = 0.0_wp
[1]	194	ENDIF
	195
	196	DO WHILE ( residual_norm > residual_limit .OR. &
	197	mgcycles < maximum_mgcycles )
[778]	198
	199	CALL next_mg_level( d, p_loc, p3, r)
[1]	200
	201	!
	202	!-- Calculate the residual if the user has not preset the number of
	203	!-- cycles to be performed
	204	IF ( maximum_mgcycles == 0 ) THEN
[707]	205	CALL resid( d, p_loc, r )
[1]	206	maxerror = SUM( r(nzb+1:nzt,nys:nyn,nxl:nxr)**2 )
[778]	207
[1]	208	#if defined( __parallel )
[622]	209	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[778]	210	CALL MPI_ALLREDUCE( maxerror, residual_norm, 1, MPI_REAL, MPI_SUM, &
[1]	211	comm2d, ierr)
	212	#else
[778]	213	residual_norm = maxerror
[1]	214	#endif
	215	residual_norm = SQRT( residual_norm )
	216	ENDIF
	217
	218	mgcycles = mgcycles + 1
	219
	220	!
	221	!-- If the user has not limited the number of cycles, stop the run in case
	222	!-- of insufficient convergence
	223	IF ( mgcycles > 1000 .AND. mg_cycles == -1 ) THEN
[257]	224	message_string = 'no sufficient convergence within 1000 cycles'
	225	CALL message( 'poismg', 'PA0283', 1, 2, 0, 6, 0 )
[1]	226	ENDIF
	227
	228	ENDDO
	229
	230	DEALLOCATE( p3 )
	231
[707]	232	!
	233	!-- Unset the grid level. Variable is used to determine the MPI datatypes for
	234	!-- ghost point exchange
	235	grid_level = 0
	236
[1]	237	CALL cpu_log( log_point_s(29), 'poismg', 'stop' )
	238
	239	END SUBROUTINE poismg
	240
	241
	242
	243	SUBROUTINE resid( f_mg, p_mg, r )
	244
	245	!------------------------------------------------------------------------------!
	246	! Description:
	247	! ------------
	248	! Computes the residual of the perturbation pressure.
	249	!------------------------------------------------------------------------------!
	250
[1320]	251	USE arrays_3d, &
	252	ONLY: f1_mg, f2_mg, f3_mg
[1]	253
[1320]	254	USE control_parameters, &
	255	ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, inflow_l, &
	256	inflow_n, inflow_r, inflow_s, outflow_l, outflow_n, outflow_r, &
	257	outflow_s
	258
	259	USE grid_variables, &
	260	ONLY: ddx2_mg, ddy2_mg
	261
	262	USE indices, &
	263	ONLY: flags, wall_flags_1, wall_flags_2, wall_flags_3, wall_flags_4, &
	264	wall_flags_5, wall_flags_6, wall_flags_7, wall_flags_8, &
	265	wall_flags_9, wall_flags_10, nxl_mg, nxr_mg, nys_mg, nyn_mg, &
	266	nzb, nzt_mg
	267
	268	USE kinds
	269
[1]	270	IMPLICIT NONE
	271
[1320]	272	INTEGER(iwp) :: i
	273	INTEGER(iwp) :: j
	274	INTEGER(iwp) :: k
	275	INTEGER(iwp) :: l
[1]	276
[1320]	277	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
[1]	278	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
[1320]	279	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f_mg !:
	280	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
	281	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	282	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p_mg !:
	283	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
	284	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	285	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: r !:
[1]	286
	287	!
	288	!-- Calculate the residual
	289	l = grid_level
	290
[114]	291	!
	292	!-- Choose flag array of this level
	293	SELECT CASE ( l )
	294	CASE ( 1 )
	295	flags => wall_flags_1
	296	CASE ( 2 )
	297	flags => wall_flags_2
	298	CASE ( 3 )
	299	flags => wall_flags_3
	300	CASE ( 4 )
	301	flags => wall_flags_4
	302	CASE ( 5 )
	303	flags => wall_flags_5
	304	CASE ( 6 )
	305	flags => wall_flags_6
	306	CASE ( 7 )
	307	flags => wall_flags_7
	308	CASE ( 8 )
	309	flags => wall_flags_8
	310	CASE ( 9 )
	311	flags => wall_flags_9
	312	CASE ( 10 )
	313	flags => wall_flags_10
	314	END SELECT
	315
[1]	316	!$OMP PARALLEL PRIVATE (i,j,k)
	317	!$OMP DO
	318	DO i = nxl_mg(l), nxr_mg(l)
	319	DO j = nys_mg(l), nyn_mg(l)
	320	DO k = nzb+1, nzt_mg(l)
[114]	321	r(k,j,i) = f_mg(k,j,i) &
	322	- ddx2_mg(l) * &
	323	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	324	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	325	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	326	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	327	- ddy2_mg(l) * &
	328	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	329	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	330	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	331	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	332	- f2_mg(k,l) * p_mg(k+1,j,i) &
	333	- f3_mg(k,l) * &
	334	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	335	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
[1]	336	+ f1_mg(k,l) * p_mg(k,j,i)
[114]	337	!
	338	!-- Residual within topography should be zero
[1353]	339	r(k,j,i) = r(k,j,i) * ( 1.0_wp - IBITS( flags(k,j,i), 6, 1 ) )
[1]	340	ENDDO
	341	ENDDO
	342	ENDDO
	343	!$OMP END PARALLEL
	344
	345	!
	346	!-- Horizontal boundary conditions
[667]	347	CALL exchange_horiz( r, 1)
[1]	348
[707]	349	IF ( .NOT. bc_lr_cyc ) THEN
[1]	350	IF ( inflow_l .OR. outflow_l ) r(:,:,nxl_mg(l)-1) = r(:,:,nxl_mg(l))
	351	IF ( inflow_r .OR. outflow_r ) r(:,:,nxr_mg(l)+1) = r(:,:,nxr_mg(l))
	352	ENDIF
	353
[707]	354	IF ( .NOT. bc_ns_cyc ) THEN
[1]	355	IF ( inflow_n .OR. outflow_n ) r(:,nyn_mg(l)+1,:) = r(:,nyn_mg(l),:)
	356	IF ( inflow_s .OR. outflow_s ) r(:,nys_mg(l)-1,:) = r(:,nys_mg(l),:)
	357	ENDIF
	358
	359	!
[707]	360	!-- Boundary conditions at bottom and top of the domain.
	361	!-- These points are not handled by the above loop. Points may be within
	362	!-- buildings, but that doesn't matter.
	363	IF ( ibc_p_b == 1 ) THEN
	364	r(nzb,:,: ) = r(nzb+1,:,:)
	365	ELSE
[1353]	366	r(nzb,:,: ) = 0.0_wp
[707]	367	ENDIF
	368
[1]	369	IF ( ibc_p_t == 1 ) THEN
	370	r(nzt_mg(l)+1,:,: ) = r(nzt_mg(l),:,:)
	371	ELSE
[1353]	372	r(nzt_mg(l)+1,:,: ) = 0.0_wp
[1]	373	ENDIF
	374
	375
	376	END SUBROUTINE resid
	377
	378
	379
	380	SUBROUTINE restrict( f_mg, r )
	381
	382	!------------------------------------------------------------------------------!
	383	! Description:
	384	! ------------
	385	! Interpolates the residual on the next coarser grid with "full weighting"
	386	! scheme
	387	!------------------------------------------------------------------------------!
	388
[1320]	389	USE control_parameters, &
	390	ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, inflow_l, &
	391	inflow_n, inflow_r, inflow_s, outflow_l, outflow_n, outflow_r, &
	392	outflow_s
[1]	393
[1320]	394	USE indices, &
	395	ONLY: flags, wall_flags_1, wall_flags_2, wall_flags_3, wall_flags_4, &
	396	wall_flags_5, wall_flags_6, wall_flags_7, wall_flags_8, &
	397	wall_flags_9, wall_flags_10, nxl_mg, nxr_mg, nys_mg, nyn_mg, &
	398	nzb, nzt_mg
	399
	400	USE kinds
	401
[1]	402	IMPLICIT NONE
	403
[1320]	404	INTEGER(iwp) :: i !:
	405	INTEGER(iwp) :: ic !:
	406	INTEGER(iwp) :: j !:
	407	INTEGER(iwp) :: jc !:
	408	INTEGER(iwp) :: k !:
	409	INTEGER(iwp) :: kc !:
	410	INTEGER(iwp) :: l !:
[1]	411
[1320]	412	REAL(wp) :: rkjim !:
	413	REAL(wp) :: rkjip !:
	414	REAL(wp) :: rkjmi !:
	415	REAL(wp) :: rkjmim !:
	416	REAL(wp) :: rkjmip !:
	417	REAL(wp) :: rkjpi !:
	418	REAL(wp) :: rkjpim !:
	419	REAL(wp) :: rkjpip !:
	420	REAL(wp) :: rkmji !:
	421	REAL(wp) :: rkmjim !:
	422	REAL(wp) :: rkmjip !:
	423	REAL(wp) :: rkmjmi !:
	424	REAL(wp) :: rkmjmim !:
	425	REAL(wp) :: rkmjmip !:
	426	REAL(wp) :: rkmjpi !:
	427	REAL(wp) :: rkmjpim !:
	428	REAL(wp) :: rkmjpip !:
[114]	429
[1320]	430	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
	431	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	432	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f_mg !:
[1]	433
[1320]	434	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level+1)+1, &
	435	nys_mg(grid_level+1)-1:nyn_mg(grid_level+1)+1, &
	436	nxl_mg(grid_level+1)-1:nxr_mg(grid_level+1)+1) :: r !:
[1]	437
	438	!
	439	!-- Interpolate the residual
	440	l = grid_level
	441
[114]	442	!
	443	!-- Choose flag array of the upper level
[181]	444	SELECT CASE ( l+1 )
[114]	445	CASE ( 1 )
	446	flags => wall_flags_1
	447	CASE ( 2 )
	448	flags => wall_flags_2
	449	CASE ( 3 )
	450	flags => wall_flags_3
	451	CASE ( 4 )
	452	flags => wall_flags_4
	453	CASE ( 5 )
	454	flags => wall_flags_5
	455	CASE ( 6 )
	456	flags => wall_flags_6
	457	CASE ( 7 )
	458	flags => wall_flags_7
	459	CASE ( 8 )
	460	flags => wall_flags_8
	461	CASE ( 9 )
	462	flags => wall_flags_9
	463	CASE ( 10 )
	464	flags => wall_flags_10
	465	END SELECT
	466
[1]	467	!$OMP PARALLEL PRIVATE (i,j,k,ic,jc,kc)
	468	!$OMP DO
	469	DO ic = nxl_mg(l), nxr_mg(l)
	470	i = 2*ic
	471	DO jc = nys_mg(l), nyn_mg(l)
	472	j = 2*jc
	473	DO kc = nzb+1, nzt_mg(l)
	474	k = 2*kc-1
[114]	475	!
	476	!-- Use implicit Neumann BCs if the respective gridpoint is inside
	477	!-- the building
	478	rkjim = r(k,j,i-1) + IBITS( flags(k,j,i-1), 6, 1 ) * &
	479	( r(k,j,i) - r(k,j,i-1) )
	480	rkjip = r(k,j,i+1) + IBITS( flags(k,j,i+1), 6, 1 ) * &
	481	( r(k,j,i) - r(k,j,i+1) )
	482	rkjpi = r(k,j+1,i) + IBITS( flags(k,j+1,i), 6, 1 ) * &
	483	( r(k,j,i) - r(k,j+1,i) )
	484	rkjmi = r(k,j-1,i) + IBITS( flags(k,j-1,i), 6, 1 ) * &
	485	( r(k,j,i) - r(k,j-1,i) )
	486	rkjmim = r(k,j-1,i-1) + IBITS( flags(k,j-1,i-1), 6, 1 ) * &
	487	( r(k,j,i) - r(k,j-1,i-1) )
	488	rkjpim = r(k,j+1,i-1) + IBITS( flags(k,j+1,i-1), 6, 1 ) * &
	489	( r(k,j,i) - r(k,j+1,i-1) )
	490	rkjmip = r(k,j-1,i+1) + IBITS( flags(k,j-1,i+1), 6, 1 ) * &
	491	( r(k,j,i) - r(k,j-1,i+1) )
	492	rkjpip = r(k,j+1,i+1) + IBITS( flags(k,j+1,i+1), 6, 1 ) * &
	493	( r(k,j,i) - r(k,j+1,i+1) )
	494	rkmji = r(k-1,j,i) + IBITS( flags(k-1,j,i), 6, 1 ) * &
	495	( r(k,j,i) - r(k-1,j,i) )
	496	rkmjim = r(k-1,j,i-1) + IBITS( flags(k-1,j,i-1), 6, 1 ) * &
	497	( r(k,j,i) - r(k-1,j,i-1) )
	498	rkmjip = r(k-1,j,i+1) + IBITS( flags(k-1,j,i+1), 6, 1 ) * &
	499	( r(k,j,i) - r(k-1,j,i+1) )
	500	rkmjpi = r(k-1,j+1,i) + IBITS( flags(k-1,j+1,i), 6, 1 ) * &
	501	( r(k,j,i) - r(k-1,j+1,i) )
	502	rkmjmi = r(k-1,j-1,i) + IBITS( flags(k-1,j-1,i), 6, 1 ) * &
	503	( r(k,j,i) - r(k-1,j-1,i) )
	504	rkmjmim = r(k-1,j-1,i-1) + IBITS( flags(k-1,j-1,i-1), 6, 1 ) * &
	505	( r(k,j,i) - r(k-1,j-1,i-1) )
	506	rkmjpim = r(k-1,j+1,i-1) + IBITS( flags(k-1,j+1,i-1), 6, 1 ) * &
	507	( r(k,j,i) - r(k-1,j+1,i-1) )
	508	rkmjmip = r(k-1,j-1,i+1) + IBITS( flags(k-1,j-1,i+1), 6, 1 ) * &
	509	( r(k,j,i) - r(k-1,j-1,i+1) )
	510	rkmjpip = r(k-1,j+1,i+1) + IBITS( flags(k-1,j+1,i+1), 6, 1 ) * &
	511	( r(k,j,i) - r(k-1,j+1,i+1) )
	512
[1353]	513	f_mg(kc,jc,ic) = 1.0_wp / 64.0_wp * ( &
	514	8.0_wp * r(k,j,i) &
	515	+ 4.0_wp * ( rkjim + rkjip + &
	516	rkjpi + rkjmi ) &
	517	+ 2.0_wp * ( rkjmim + rkjpim + &
	518	rkjmip + rkjpip ) &
	519	+ 4.0_wp * rkmji &
	520	+ 2.0_wp * ( rkmjim + rkmjim + &
	521	rkmjpi + rkmjmi ) &
	522	+ ( rkmjmim + rkmjpim + &
	523	rkmjmip + rkmjpip ) &
	524	+ 4.0_wp * r(k+1,j,i) &
	525	+ 2.0_wp * ( r(k+1,j,i-1) + r(k+1,j,i+1) + &
	526	r(k+1,j+1,i) + r(k+1,j-1,i) ) &
	527	+ ( r(k+1,j-1,i-1) + r(k+1,j+1,i-1) + &
	528	r(k+1,j-1,i+1) + r(k+1,j+1,i+1) ) &
	529	)
[114]	530
[1353]	531	! f_mg(kc,jc,ic) = 1.0_wp / 64.0_wp * ( &
	532	! 8.0_wp * r(k,j,i) &
	533	! + 4.0_wp * ( r(k,j,i-1) + r(k,j,i+1) + &
	534	! r(k,j+1,i) + r(k,j-1,i) ) &
	535	! + 2.0_wp * ( r(k,j-1,i-1) + r(k,j+1,i-1) + &
	536	! r(k,j-1,i+1) + r(k,j+1,i+1) ) &
	537	! + 4.0_wp * r(k-1,j,i) &
	538	! + 2.0_wp * ( r(k-1,j,i-1) + r(k-1,j,i+1) + &
	539	! r(k-1,j+1,i) + r(k-1,j-1,i) ) &
	540	! + ( r(k-1,j-1,i-1) + r(k-1,j+1,i-1) + &
	541	! r(k-1,j-1,i+1) + r(k-1,j+1,i+1) ) &
	542	! + 4.0_wp * r(k+1,j,i) &
	543	! + 2.0_wp * ( r(k+1,j,i-1) + r(k+1,j,i+1) + &
	544	! r(k+1,j+1,i) + r(k+1,j-1,i) ) &
	545	! + ( r(k+1,j-1,i-1) + r(k+1,j+1,i-1) + &
	546	! r(k+1,j-1,i+1) + r(k+1,j+1,i+1) ) &
	547	! )
[1]	548	ENDDO
	549	ENDDO
	550	ENDDO
	551	!$OMP END PARALLEL
	552
	553	!
	554	!-- Horizontal boundary conditions
[667]	555	CALL exchange_horiz( f_mg, 1)
[1]	556
[707]	557	IF ( .NOT. bc_lr_cyc ) THEN
[1]	558	IF (inflow_l .OR. outflow_l) f_mg(:,:,nxl_mg(l)-1) = f_mg(:,:,nxl_mg(l))
	559	IF (inflow_r .OR. outflow_r) f_mg(:,:,nxr_mg(l)+1) = f_mg(:,:,nxr_mg(l))
	560	ENDIF
	561
[707]	562	IF ( .NOT. bc_ns_cyc ) THEN
[1]	563	IF (inflow_n .OR. outflow_n) f_mg(:,nyn_mg(l)+1,:) = f_mg(:,nyn_mg(l),:)
	564	IF (inflow_s .OR. outflow_s) f_mg(:,nys_mg(l)-1,:) = f_mg(:,nys_mg(l),:)
	565	ENDIF
	566
	567	!
[707]	568	!-- Boundary conditions at bottom and top of the domain.
	569	!-- These points are not handled by the above loop. Points may be within
	570	!-- buildings, but that doesn't matter.
	571	IF ( ibc_p_b == 1 ) THEN
	572	f_mg(nzb,:,: ) = f_mg(nzb+1,:,:)
	573	ELSE
[1353]	574	f_mg(nzb,:,: ) = 0.0_wp
[707]	575	ENDIF
[1]	576
[707]	577	IF ( ibc_p_t == 1 ) THEN
	578	f_mg(nzt_mg(l)+1,:,: ) = f_mg(nzt_mg(l),:,:)
	579	ELSE
[1353]	580	f_mg(nzt_mg(l)+1,:,: ) = 0.0_wp
[707]	581	ENDIF
[1]	582
[707]	583
[1]	584	END SUBROUTINE restrict
	585
	586
	587
	588	SUBROUTINE prolong( p, temp )
	589
	590	!------------------------------------------------------------------------------!
	591	! Description:
	592	! ------------
	593	! Interpolates the correction of the perturbation pressure
	594	! to the next finer grid.
	595	!------------------------------------------------------------------------------!
	596
[1320]	597	USE control_parameters, &
	598	ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, inflow_l, &
	599	inflow_n, inflow_r, inflow_s, outflow_l, outflow_n, outflow_r, &
	600	outflow_s
[1]	601
[1320]	602	USE indices, &
	603	ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg
	604
	605	USE kinds
	606
[1]	607	IMPLICIT NONE
	608
[1320]	609	INTEGER(iwp) :: i !:
	610	INTEGER(iwp) :: j !:
	611	INTEGER(iwp) :: k !:
	612	INTEGER(iwp) :: l !:
[1]	613
[1320]	614	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, &
	615	nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, &
	616	nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1 ) :: p !:
[1]	617
[1320]	618	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
	619	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	620	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: temp !:
[1]	621
	622
	623	!
	624	!-- First, store elements of the coarser grid on the next finer grid
	625	l = grid_level
	626
	627	!$OMP PARALLEL PRIVATE (i,j,k)
	628	!$OMP DO
	629	DO i = nxl_mg(l-1), nxr_mg(l-1)
	630	DO j = nys_mg(l-1), nyn_mg(l-1)
	631	!CDIR NODEP
	632	DO k = nzb+1, nzt_mg(l-1)
	633	!
	634	!-- Points of the coarse grid are directly stored on the next finer
	635	!-- grid
	636	temp(2k-1,2j,2*i) = p(k,j,i)
	637	!
	638	!-- Points between two coarse-grid points
[1353]	639	temp(2k-1,2j,2i+1) = 0.5_wp ( p(k,j,i) + p(k,j,i+1) )
	640	temp(2k-1,2j+1,2i) = 0.5_wp ( p(k,j,i) + p(k,j+1,i) )
	641	temp(2k,2j,2i) = 0.5_wp ( p(k,j,i) + p(k+1,j,i) )
[1]	642	!
	643	!-- Points in the center of the planes stretched by four points
	644	!-- of the coarse grid cube
[1353]	645	temp(2k-1,2j+1,2i+1) = 0.25_wp ( p(k,j,i) + p(k,j,i+1) + &
	646	p(k,j+1,i) + p(k,j+1,i+1) )
	647	temp(2k,2j,2i+1) = 0.25_wp ( p(k,j,i) + p(k,j,i+1) + &
	648	p(k+1,j,i) + p(k+1,j,i+1) )
	649	temp(2k,2j+1,2i) = 0.25_wp ( p(k,j,i) + p(k,j+1,i) + &
	650	p(k+1,j,i) + p(k+1,j+1,i) )
[1]	651	!
	652	!-- Points in the middle of coarse grid cube
[1353]	653	temp(2k,2j+1,2i+1) = 0.125_wp ( p(k,j,i) + p(k,j,i+1) + &
	654	p(k,j+1,i) + p(k,j+1,i+1) + &
	655	p(k+1,j,i) + p(k+1,j,i+1) + &
	656	p(k+1,j+1,i) + p(k+1,j+1,i+1) )
[1]	657	ENDDO
	658	ENDDO
	659	ENDDO
	660	!$OMP END PARALLEL
	661
	662	!
	663	!-- Horizontal boundary conditions
[667]	664	CALL exchange_horiz( temp, 1)
[1]	665
[707]	666	IF ( .NOT. bc_lr_cyc ) THEN
[1]	667	IF (inflow_l .OR. outflow_l) temp(:,:,nxl_mg(l)-1) = temp(:,:,nxl_mg(l))
	668	IF (inflow_r .OR. outflow_r) temp(:,:,nxr_mg(l)+1) = temp(:,:,nxr_mg(l))
	669	ENDIF
	670
[707]	671	IF ( .NOT. bc_ns_cyc ) THEN
[1]	672	IF (inflow_n .OR. outflow_n) temp(:,nyn_mg(l)+1,:) = temp(:,nyn_mg(l),:)
	673	IF (inflow_s .OR. outflow_s) temp(:,nys_mg(l)-1,:) = temp(:,nys_mg(l),:)
	674	ENDIF
	675
	676	!
	677	!-- Bottom and top boundary conditions
	678	IF ( ibc_p_b == 1 ) THEN
	679	temp(nzb,:,: ) = temp(nzb+1,:,:)
	680	ELSE
[1353]	681	temp(nzb,:,: ) = 0.0_wp
[1]	682	ENDIF
	683
	684	IF ( ibc_p_t == 1 ) THEN
	685	temp(nzt_mg(l)+1,:,: ) = temp(nzt_mg(l),:,:)
	686	ELSE
[1353]	687	temp(nzt_mg(l)+1,:,: ) = 0.0_wp
[1]	688	ENDIF
	689
	690
	691	END SUBROUTINE prolong
	692
	693
	694	SUBROUTINE redblack( f_mg, p_mg )
	695
	696	!------------------------------------------------------------------------------!
	697	! Description:
	698	! ------------
	699	! Relaxation method for the multigrid scheme. A Gauss-Seidel iteration with
	700	! 3D-Red-Black decomposition (GS-RB) is used.
	701	!------------------------------------------------------------------------------!
	702
[1320]	703	USE arrays_3d, &
	704	ONLY: f1_mg, f2_mg, f3_mg
[1]	705
[1320]	706	USE control_parameters, &
	707	ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, inflow_l, &
	708	inflow_n, inflow_r, inflow_s, ngsrb, outflow_l, outflow_n, &
	709	outflow_r, outflow_s
	710
	711	USE cpulog, &
	712	ONLY: cpu_log, log_point_s
	713
	714	USE grid_variables, &
	715	ONLY: ddx2_mg, ddy2_mg
	716
	717	USE indices, &
	718	ONLY: flags, wall_flags_1, wall_flags_2, wall_flags_3, wall_flags_4, &
	719	wall_flags_5, wall_flags_6, wall_flags_7, wall_flags_8, &
	720	wall_flags_9, wall_flags_10, nxl_mg, nxr_mg, nys_mg, nyn_mg, &
	721	nzb, nzt_mg
	722
	723	USE kinds
	724
[1]	725	IMPLICIT NONE
	726
[1320]	727	INTEGER(iwp) :: color !:
	728	INTEGER(iwp) :: i !:
	729	INTEGER(iwp) :: ic !:
	730	INTEGER(iwp) :: j !:
	731	INTEGER(iwp) :: jc !:
	732	INTEGER(iwp) :: jj !:
	733	INTEGER(iwp) :: k !:
	734	INTEGER(iwp) :: l !:
	735	INTEGER(iwp) :: n !:
[1]	736
[1320]	737	LOGICAL :: unroll !:
[1]	738
[1320]	739	REAL(wp) :: wall_left !:
	740	REAL(wp) :: wall_north !:
	741	REAL(wp) :: wall_right !:
	742	REAL(wp) :: wall_south !:
	743	REAL(wp) :: wall_total !:
	744	REAL(wp) :: wall_top !:
[114]	745
[1320]	746	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
	747	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	748	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f_mg !:
	749	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
	750	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	751	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p_mg !:
[1]	752
	753	l = grid_level
	754
[114]	755	!
	756	!-- Choose flag array of this level
	757	SELECT CASE ( l )
	758	CASE ( 1 )
	759	flags => wall_flags_1
	760	CASE ( 2 )
	761	flags => wall_flags_2
	762	CASE ( 3 )
	763	flags => wall_flags_3
	764	CASE ( 4 )
	765	flags => wall_flags_4
	766	CASE ( 5 )
	767	flags => wall_flags_5
	768	CASE ( 6 )
	769	flags => wall_flags_6
	770	CASE ( 7 )
	771	flags => wall_flags_7
	772	CASE ( 8 )
	773	flags => wall_flags_8
	774	CASE ( 9 )
	775	flags => wall_flags_9
	776	CASE ( 10 )
	777	flags => wall_flags_10
	778	END SELECT
	779
[1]	780	unroll = ( MOD( nyn_mg(l)-nys_mg(l)+1, 4 ) == 0 .AND. &
	781	MOD( nxr_mg(l)-nxl_mg(l)+1, 2 ) == 0 )
	782
	783	DO n = 1, ngsrb
	784
[1320]	785	DO color = 1, 2
[1]	786
	787	IF ( .NOT. unroll ) THEN
[778]	788
[1]	789	CALL cpu_log( log_point_s(36), 'redblack_no_unroll', 'start' )
	790
	791	!
	792	!-- Without unrolling of loops, no cache optimization
	793	DO i = nxl_mg(l), nxr_mg(l), 2
[1320]	794	DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2
[1]	795	DO k = nzb+1, nzt_mg(l), 2
[1353]	796	! p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	797	! ddx2_mg(l) * ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) &
	798	! + ddy2_mg(l) * ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) &
	799	! + f2_mg(k,l) * p_mg(k+1,j,i) &
	800	! + f3_mg(k,l) * p_mg(k-1,j,i) - f_mg(k,j,i) &
	801	! )
	802
[1353]	803	p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	804	ddx2_mg(l) * &
	805	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	806	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	807	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	808	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	809	+ ddy2_mg(l) * &
	810	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	811	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	812	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	813	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	814	+ f2_mg(k,l) * p_mg(k+1,j,i) &
	815	+ f3_mg(k,l) * &
	816	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	817	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
	818	- f_mg(k,j,i) )
[1]	819	ENDDO
	820	ENDDO
	821	ENDDO
	822
	823	DO i = nxl_mg(l)+1, nxr_mg(l), 2
[1320]	824	DO j = nys_mg(l) + (color-1), nyn_mg(l), 2
[1]	825	DO k = nzb+1, nzt_mg(l), 2
[1353]	826	p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	827	ddx2_mg(l) * &
	828	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	829	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	830	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	831	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	832	+ ddy2_mg(l) * &
	833	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	834	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	835	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	836	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	837	+ f2_mg(k,l) * p_mg(k+1,j,i) &
	838	+ f3_mg(k,l) * &
	839	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	840	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
	841	- f_mg(k,j,i) )
[1]	842	ENDDO
	843	ENDDO
	844	ENDDO
	845
	846	DO i = nxl_mg(l), nxr_mg(l), 2
[1320]	847	DO j = nys_mg(l) + (color-1), nyn_mg(l), 2
[1]	848	DO k = nzb+2, nzt_mg(l), 2
[1353]	849	p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	850	ddx2_mg(l) * &
	851	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	852	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	853	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	854	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	855	+ ddy2_mg(l) * &
	856	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	857	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	858	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	859	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	860	+ f2_mg(k,l) * p_mg(k+1,j,i) &
	861	+ f3_mg(k,l) * &
	862	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	863	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
	864	- f_mg(k,j,i) )
[1]	865	ENDDO
	866	ENDDO
	867	ENDDO
	868
	869	DO i = nxl_mg(l)+1, nxr_mg(l), 2
[1320]	870	DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2
[1]	871	DO k = nzb+2, nzt_mg(l), 2
[1353]	872	p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	873	ddx2_mg(l) * &
	874	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	875	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	876	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	877	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	878	+ ddy2_mg(l) * &
	879	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	880	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	881	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	882	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	883	+ f2_mg(k,l) * p_mg(k+1,j,i) &
	884	+ f3_mg(k,l) * &
	885	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	886	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
	887	- f_mg(k,j,i) )
[1]	888	ENDDO
	889	ENDDO
	890	ENDDO
	891	CALL cpu_log( log_point_s(36), 'redblack_no_unroll', 'stop' )
	892
	893	ELSE
	894
	895	!
	896	!-- Loop unrolling along y, only one i loop for better cache use
	897	CALL cpu_log( log_point_s(38), 'redblack_unroll', 'start' )
	898	DO ic = nxl_mg(l), nxr_mg(l), 2
	899	DO jc = nys_mg(l), nyn_mg(l), 4
	900	i = ic
[1320]	901	jj = jc+2-color
[1]	902	DO k = nzb+1, nzt_mg(l), 2
	903	j = jj
[1353]	904	p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	905	ddx2_mg(l) * &
	906	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	907	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	908	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	909	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	910	+ ddy2_mg(l) * &
	911	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	912	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	913	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	914	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	915	+ f2_mg(k,l) * p_mg(k+1,j,i) &
	916	+ f3_mg(k,l) * &
	917	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	918	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
	919	- f_mg(k,j,i) )
[1]	920	j = jj+2
[1353]	921	p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	922	ddx2_mg(l) * &
	923	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	924	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	925	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	926	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	927	+ ddy2_mg(l) * &
	928	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	929	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	930	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	931	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	932	+ f2_mg(k,l) * p_mg(k+1,j,i) &
	933	+ f3_mg(k,l) * &
	934	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	935	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
	936	- f_mg(k,j,i) )
[1]	937	ENDDO
	938
	939	i = ic+1
[1320]	940	jj = jc+color-1
[1]	941	DO k = nzb+1, nzt_mg(l), 2
	942	j =jj
[1353]	943	p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	944	ddx2_mg(l) * &
	945	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	946	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	947	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	948	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	949	+ ddy2_mg(l) * &
	950	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	951	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	952	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	953	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	954	+ f2_mg(k,l) * p_mg(k+1,j,i) &
	955	+ f3_mg(k,l) * &
	956	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	957	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
	958	- f_mg(k,j,i) )
[1]	959	j = jj+2
[1353]	960	p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	961	ddx2_mg(l) * &
	962	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	963	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	964	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	965	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	966	+ ddy2_mg(l) * &
	967	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	968	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	969	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	970	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	971	+ f2_mg(k,l) * p_mg(k+1,j,i) &
	972	+ f3_mg(k,l) * &
	973	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	974	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
	975	- f_mg(k,j,i) )
[1]	976	ENDDO
	977
	978	i = ic
[1320]	979	jj = jc+color-1
[1]	980	DO k = nzb+2, nzt_mg(l), 2
	981	j =jj
[1353]	982	p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	983	ddx2_mg(l) * &
	984	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	985	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	986	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	987	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	988	+ ddy2_mg(l) * &
	989	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	990	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	991	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	992	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	993	+ f2_mg(k,l) * p_mg(k+1,j,i) &
	994	+ f3_mg(k,l) * &
	995	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	996	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
	997	- f_mg(k,j,i) )
[1]	998	j = jj+2
[1353]	999	p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	1000	ddx2_mg(l) * &
	1001	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	1002	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	1003	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	1004	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	1005	+ ddy2_mg(l) * &
	1006	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	1007	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	1008	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	1009	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	1010	+ f2_mg(k,l) * p_mg(k+1,j,i) &
	1011	+ f3_mg(k,l) * &
	1012	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	1013	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
	1014	- f_mg(k,j,i) )
[1]	1015	ENDDO
	1016
	1017	i = ic+1
[1320]	1018	jj = jc+2-color
[1]	1019	DO k = nzb+2, nzt_mg(l), 2
	1020	j =jj
[1353]	1021	p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	1022	ddx2_mg(l) * &
	1023	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	1024	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	1025	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	1026	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	1027	+ ddy2_mg(l) * &
	1028	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	1029	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	1030	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	1031	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	1032	+ f2_mg(k,l) * p_mg(k+1,j,i) &
	1033	+ f3_mg(k,l) * &
	1034	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	1035	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
	1036	- f_mg(k,j,i) )
[1]	1037	j = jj+2
[1353]	1038	p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( &
[114]	1039	ddx2_mg(l) * &
	1040	( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &
	1041	( p_mg(k,j,i) - p_mg(k,j,i+1) ) + &
	1042	p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * &
	1043	( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) &
	1044	+ ddy2_mg(l) * &
	1045	( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * &
	1046	( p_mg(k,j,i) - p_mg(k,j+1,i) ) + &
	1047	p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * &
	1048	( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) &
	1049	+ f2_mg(k,l) * p_mg(k+1,j,i) &
	1050	+ f3_mg(k,l) * &
	1051	( p_mg(k-1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * &
	1052	( p_mg(k,j,i) - p_mg(k-1,j,i) ) ) &
	1053	- f_mg(k,j,i) )
[1]	1054	ENDDO
	1055
	1056	ENDDO
	1057	ENDDO
	1058	CALL cpu_log( log_point_s(38), 'redblack_unroll', 'stop' )
	1059
	1060	ENDIF
	1061
	1062	!
	1063	!-- Horizontal boundary conditions
[667]	1064	CALL exchange_horiz( p_mg, 1 )
[1]	1065
[707]	1066	IF ( .NOT. bc_lr_cyc ) THEN
[1]	1067	IF ( inflow_l .OR. outflow_l ) THEN
	1068	p_mg(:,:,nxl_mg(l)-1) = p_mg(:,:,nxl_mg(l))
	1069	ENDIF
	1070	IF ( inflow_r .OR. outflow_r ) THEN
	1071	p_mg(:,:,nxr_mg(l)+1) = p_mg(:,:,nxr_mg(l))
	1072	ENDIF
	1073	ENDIF
	1074
[707]	1075	IF ( .NOT. bc_ns_cyc ) THEN
[1]	1076	IF ( inflow_n .OR. outflow_n ) THEN
	1077	p_mg(:,nyn_mg(l)+1,:) = p_mg(:,nyn_mg(l),:)
	1078	ENDIF
	1079	IF ( inflow_s .OR. outflow_s ) THEN
	1080	p_mg(:,nys_mg(l)-1,:) = p_mg(:,nys_mg(l),:)
	1081	ENDIF
	1082	ENDIF
	1083
	1084	!
	1085	!-- Bottom and top boundary conditions
	1086	IF ( ibc_p_b == 1 ) THEN
	1087	p_mg(nzb,:,: ) = p_mg(nzb+1,:,:)
	1088	ELSE
[1353]	1089	p_mg(nzb,:,: ) = 0.0_wp
[1]	1090	ENDIF
	1091
	1092	IF ( ibc_p_t == 1 ) THEN
	1093	p_mg(nzt_mg(l)+1,:,: ) = p_mg(nzt_mg(l),:,:)
	1094	ELSE
[1353]	1095	p_mg(nzt_mg(l)+1,:,: ) = 0.0_wp
[1]	1096	ENDIF
	1097
	1098	ENDDO
	1099
	1100	ENDDO
	1101
[114]	1102	!
	1103	!-- Set pressure within topography and at the topography surfaces
	1104	!$OMP PARALLEL PRIVATE (i,j,k,wall_left,wall_north,wall_right,wall_south,wall_top,wall_total)
	1105	!$OMP DO
	1106	DO i = nxl_mg(l), nxr_mg(l)
	1107	DO j = nys_mg(l), nyn_mg(l)
	1108	DO k = nzb, nzt_mg(l)
	1109	!
	1110	!-- First, set pressure inside topography to zero
[1353]	1111	p_mg(k,j,i) = p_mg(k,j,i) * ( 1.0_wp - IBITS( flags(k,j,i), 6, 1 ) )
[114]	1112	!
	1113	!-- Second, determine if the gridpoint inside topography is adjacent
	1114	!-- to a wall and set its value to a value given by the average of
	1115	!-- those values obtained from Neumann boundary condition
	1116	wall_left = IBITS( flags(k,j,i-1), 5, 1 )
	1117	wall_right = IBITS( flags(k,j,i+1), 4, 1 )
	1118	wall_south = IBITS( flags(k,j-1,i), 3, 1 )
	1119	wall_north = IBITS( flags(k,j+1,i), 2, 1 )
	1120	wall_top = IBITS( flags(k+1,j,i), 0, 1 )
	1121	wall_total = wall_left + wall_right + wall_south + wall_north + &
	1122	wall_top
[1]	1123
[1353]	1124	IF ( wall_total > 0.0_wp ) THEN
	1125	p_mg(k,j,i) = 1.0_wp / wall_total * &
	1126	( wall_left * p_mg(k,j,i-1) + &
	1127	wall_right * p_mg(k,j,i+1) + &
	1128	wall_south * p_mg(k,j-1,i) + &
	1129	wall_north * p_mg(k,j+1,i) + &
	1130	wall_top * p_mg(k+1,j,i) )
[114]	1131	ENDIF
	1132	ENDDO
	1133	ENDDO
	1134	ENDDO
[1056]	1135	!$OMP END PARALLEL
[114]	1136
	1137	!
	1138	!-- One more time horizontal boundary conditions
[667]	1139	CALL exchange_horiz( p_mg, 1)
[114]	1140
[778]	1141
[1]	1142	END SUBROUTINE redblack
	1143
	1144
	1145
	1146	SUBROUTINE mg_gather( f2, f2_sub )
	1147
[1320]	1148	USE control_parameters, &
	1149	ONLY: grid_level
	1150
	1151	USE cpulog, &
	1152	ONLY: cpu_log, log_point_s
	1153
	1154	USE indices, &
	1155	ONLY: mg_loc_ind, nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg
	1156
	1157	USE kinds
	1158
[1]	1159	USE pegrid
	1160
	1161	IMPLICIT NONE
	1162
[1320]	1163	INTEGER(iwp) :: i !:
	1164	INTEGER(iwp) :: il !:
	1165	INTEGER(iwp) :: ir !:
	1166	INTEGER(iwp) :: j !:
	1167	INTEGER(iwp) :: jn !:
	1168	INTEGER(iwp) :: js !:
	1169	INTEGER(iwp) :: k !:
	1170	INTEGER(iwp) :: nwords !:
[1]	1171
[1320]	1172	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
	1173	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	1174	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f2 !:
	1175	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
	1176	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	1177	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f2_l !:
[1]	1178
[1320]	1179	REAL(wp), DIMENSION(nzb:mg_loc_ind(5,myid)+1, &
	1180	mg_loc_ind(3,myid)-1:mg_loc_ind(4,myid)+1, &
	1181	mg_loc_ind(1,myid)-1:mg_loc_ind(2,myid)+1) :: f2_sub !:
[1]	1182
	1183
	1184	#if defined( __parallel )
	1185	CALL cpu_log( log_point_s(34), 'mg_gather', 'start' )
	1186
[1353]	1187	f2_l = 0.0_wp
[1]	1188
	1189	!
[707]	1190	!-- Store the local subdomain array on the total array
	1191	js = mg_loc_ind(3,myid)
	1192	IF ( south_border_pe ) js = js - 1
	1193	jn = mg_loc_ind(4,myid)
	1194	IF ( north_border_pe ) jn = jn + 1
	1195	il = mg_loc_ind(1,myid)
	1196	IF ( left_border_pe ) il = il - 1
	1197	ir = mg_loc_ind(2,myid)
	1198	IF ( right_border_pe ) ir = ir + 1
	1199	DO i = il, ir
	1200	DO j = js, jn
	1201	DO k = nzb, nzt_mg(grid_level)+1
	1202	f2_l(k,j,i) = f2_sub(k,j,i)
	1203	ENDDO
[1]	1204	ENDDO
[707]	1205	ENDDO
[1]	1206
	1207	!
[707]	1208	!-- Find out the number of array elements of the total array
	1209	nwords = SIZE( f2 )
[1]	1210
[707]	1211	!
	1212	!-- Gather subdomain data from all PEs
	1213	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
	1214	CALL MPI_ALLREDUCE( f2_l(nzb,nys_mg(grid_level)-1,nxl_mg(grid_level)-1), &
	1215	f2(nzb,nys_mg(grid_level)-1,nxl_mg(grid_level)-1), &
	1216	nwords, MPI_REAL, MPI_SUM, comm2d, ierr )
	1217
[1]	1218	CALL cpu_log( log_point_s(34), 'mg_gather', 'stop' )
	1219	#endif
	1220
	1221	END SUBROUTINE mg_gather
	1222
	1223
	1224
	1225	SUBROUTINE mg_scatter( p2, p2_sub )
	1226	!
	1227	!-- TODO: It may be possible to improve the speed of this routine by using
	1228	!-- non-blocking communication
	1229
[1320]	1230	USE control_parameters, &
	1231	ONLY: grid_level
	1232
	1233	USE cpulog, &
	1234	ONLY: cpu_log, log_point_s
	1235
	1236	USE indices, &
	1237	ONLY: mg_loc_ind, nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg
	1238
	1239	USE kinds
	1240
[1]	1241	USE pegrid
	1242
	1243	IMPLICIT NONE
	1244
[1320]	1245	INTEGER(iwp) :: nwords !:
[1]	1246
[1320]	1247	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, &
	1248	nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, &
	1249	nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1) :: p2 !:
[1]	1250
[1320]	1251	REAL(wp), DIMENSION(nzb:mg_loc_ind(5,myid)+1, &
	1252	mg_loc_ind(3,myid)-1:mg_loc_ind(4,myid)+1, &
	1253	mg_loc_ind(1,myid)-1:mg_loc_ind(2,myid)+1) :: p2_sub !:
[1]	1254
	1255	!
	1256	!-- Find out the number of array elements of the subdomain array
	1257	nwords = SIZE( p2_sub )
	1258
	1259	#if defined( __parallel )
	1260	CALL cpu_log( log_point_s(35), 'mg_scatter', 'start' )
	1261
[707]	1262	p2_sub = p2(:,mg_loc_ind(3,myid)-1:mg_loc_ind(4,myid)+1, &
	1263	mg_loc_ind(1,myid)-1:mg_loc_ind(2,myid)+1)
[1]	1264
	1265	CALL cpu_log( log_point_s(35), 'mg_scatter', 'stop' )
	1266	#endif
	1267
	1268	END SUBROUTINE mg_scatter
	1269
	1270
	1271
	1272	RECURSIVE SUBROUTINE next_mg_level( f_mg, p_mg, p3, r )
	1273
	1274	!------------------------------------------------------------------------------!
	1275	! Description:
	1276	! ------------
	1277	! This is where the multigrid technique takes place. V- and W- Cycle are
	1278	! implemented and steered by the parameter "gamma". Parameter "nue" determines
	1279	! the convergence of the multigrid iterative solution. There are nue times
	1280	! RB-GS iterations. It should be set to "1" or "2", considering the time effort
	1281	! one would like to invest. Last choice shows a very good converging factor,
	1282	! but leads to an increase in computing time.
	1283	!------------------------------------------------------------------------------!
	1284
[1320]	1285	USE control_parameters, &
	1286	ONLY: bc_lr_dirrad, bc_lr_raddir, bc_ns_dirrad, bc_ns_raddir, &
	1287	gamma_mg, grid_level, grid_level_count, ibc_p_b, ibc_p_t, &
	1288	inflow_l, inflow_n, inflow_r, inflow_s, maximum_grid_level, &
	1289	mg_switch_to_pe0_level, mg_switch_to_pe0, ngsrb, outflow_l, &
	1290	outflow_n, outflow_r, outflow_s
	1291
	1292
	1293	USE indices, &
	1294	ONLY: mg_loc_ind, nxl, nxl_mg, nxr, nxr_mg, nys, nys_mg, nyn, &
	1295	nyn_mg, nzb, nzt, nzt_mg
	1296
	1297	USE kinds
	1298
[1]	1299	USE pegrid
	1300
	1301	IMPLICIT NONE
	1302
[1320]	1303	INTEGER(iwp) :: i !:
	1304	INTEGER(iwp) :: j !:
	1305	INTEGER(iwp) :: k !:
	1306	INTEGER(iwp) :: nxl_mg_save !:
	1307	INTEGER(iwp) :: nxr_mg_save !:
	1308	INTEGER(iwp) :: nyn_mg_save !:
	1309	INTEGER(iwp) :: nys_mg_save !:
	1310	INTEGER(iwp) :: nzt_mg_save !:
[1]	1311
[1320]	1312	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
	1313	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	1314	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f_mg !:
	1315	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
	1316	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	1317	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p_mg !:
	1318	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
	1319	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	1320	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p3 !:
	1321	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, &
	1322	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	1323	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: r !:
[1]	1324
[1320]	1325	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, &
	1326	nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, &
	1327	nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1) :: f2 !:
	1328	REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, &
	1329	nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, &
	1330	nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1) :: p2 !:
[1]	1331
[1320]	1332	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: f2_sub !:
	1333	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: p2_sub !:
[778]	1334
[1]	1335	!
	1336	!-- Restriction to the coarsest grid
	1337	10 IF ( grid_level == 1 ) THEN
	1338
	1339	!
	1340	!-- Solution on the coarsest grid. Double the number of Gauss-Seidel
	1341	!-- iterations in order to get a more accurate solution.
	1342	ngsrb = 2 * ngsrb
[778]	1343
[1]	1344	CALL redblack( f_mg, p_mg )
[778]	1345
[1]	1346	ngsrb = ngsrb / 2
	1347
[778]	1348
[1]	1349	ELSEIF ( grid_level /= 1 ) THEN
	1350
	1351	grid_level_count(grid_level) = grid_level_count(grid_level) + 1
	1352
	1353	!
	1354	!-- Solution on the actual grid level
	1355	CALL redblack( f_mg, p_mg )
	1356
	1357	!
	1358	!-- Determination of the actual residual
	1359	CALL resid( f_mg, p_mg, r )
	1360
	1361	!
	1362	!-- Restriction of the residual (finer grid values!) to the next coarser
	1363	!-- grid. Therefore, the grid level has to be decremented now. nxl..nzt have
	1364	!-- to be set to the coarse grid values, because these variables are needed
	1365	!-- for the exchange of ghost points in routine exchange_horiz
	1366	grid_level = grid_level - 1
	1367	nxl = nxl_mg(grid_level)
[778]	1368	nys = nys_mg(grid_level)
[1]	1369	nxr = nxr_mg(grid_level)
	1370	nyn = nyn_mg(grid_level)
	1371	nzt = nzt_mg(grid_level)
	1372
	1373	IF ( grid_level == mg_switch_to_pe0_level ) THEN
[778]	1374
[1]	1375	!
	1376	!-- From this level on, calculations are done on PE0 only.
	1377	!-- First, carry out restriction on the subdomain.
	1378	!-- Therefore, indices of the level have to be changed to subdomain values
	1379	!-- in between (otherwise, the restrict routine would expect
	1380	!-- the gathered array)
[778]	1381
[1]	1382	nxl_mg_save = nxl_mg(grid_level)
	1383	nxr_mg_save = nxr_mg(grid_level)
	1384	nys_mg_save = nys_mg(grid_level)
	1385	nyn_mg_save = nyn_mg(grid_level)
	1386	nzt_mg_save = nzt_mg(grid_level)
	1387	nxl_mg(grid_level) = mg_loc_ind(1,myid)
	1388	nxr_mg(grid_level) = mg_loc_ind(2,myid)
	1389	nys_mg(grid_level) = mg_loc_ind(3,myid)
	1390	nyn_mg(grid_level) = mg_loc_ind(4,myid)
	1391	nzt_mg(grid_level) = mg_loc_ind(5,myid)
	1392	nxl = mg_loc_ind(1,myid)
	1393	nxr = mg_loc_ind(2,myid)
	1394	nys = mg_loc_ind(3,myid)
	1395	nyn = mg_loc_ind(4,myid)
	1396	nzt = mg_loc_ind(5,myid)
	1397
	1398	ALLOCATE( f2_sub(nzb:nzt_mg(grid_level)+1, &
	1399	nys_mg(grid_level)-1:nyn_mg(grid_level)+1, &
	1400	nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) )
	1401
	1402	CALL restrict( f2_sub, r )
	1403
	1404	!
	1405	!-- Restore the correct indices of this level
	1406	nxl_mg(grid_level) = nxl_mg_save
	1407	nxr_mg(grid_level) = nxr_mg_save
	1408	nys_mg(grid_level) = nys_mg_save
	1409	nyn_mg(grid_level) = nyn_mg_save
	1410	nzt_mg(grid_level) = nzt_mg_save
	1411	nxl = nxl_mg(grid_level)
	1412	nxr = nxr_mg(grid_level)
	1413	nys = nys_mg(grid_level)
	1414	nyn = nyn_mg(grid_level)
	1415	nzt = nzt_mg(grid_level)
	1416	!
	1417	!-- Gather all arrays from the subdomains on PE0
	1418	CALL mg_gather( f2, f2_sub )
	1419
	1420	!
	1421	!-- Set switch for routine exchange_horiz, that no ghostpoint exchange
	1422	!-- has to be carried out from now on
	1423	mg_switch_to_pe0 = .TRUE.
	1424
	1425	!
	1426	!-- In case of non-cyclic lateral boundary conditions, both in- and
[707]	1427	!-- outflow conditions have to be used on all PEs after the switch,
	1428	!-- because then they have the total domain.
[1159]	1429	IF ( bc_lr_dirrad ) THEN
[707]	1430	inflow_l = .TRUE.
	1431	inflow_r = .FALSE.
	1432	outflow_l = .FALSE.
	1433	outflow_r = .TRUE.
[1159]	1434	ELSEIF ( bc_lr_raddir ) THEN
[707]	1435	inflow_l = .FALSE.
	1436	inflow_r = .TRUE.
	1437	outflow_l = .TRUE.
	1438	outflow_r = .FALSE.
[1]	1439	ENDIF
	1440
[1159]	1441	IF ( bc_ns_dirrad ) THEN
[707]	1442	inflow_n = .TRUE.
	1443	inflow_s = .FALSE.
	1444	outflow_n = .FALSE.
	1445	outflow_s = .TRUE.
[1159]	1446	ELSEIF ( bc_ns_raddir ) THEN
[707]	1447	inflow_n = .FALSE.
	1448	inflow_s = .TRUE.
	1449	outflow_n = .TRUE.
	1450	outflow_s = .FALSE.
	1451	ENDIF
	1452
[1]	1453	DEALLOCATE( f2_sub )
	1454
	1455	ELSE
[1056]	1456
[1]	1457	CALL restrict( f2, r )
	1458
	1459	ENDIF
[707]	1460
[1353]	1461	p2 = 0.0_wp
[1]	1462
	1463	!
	1464	!-- Repeat the same procedure till the coarsest grid is reached
[707]	1465	CALL next_mg_level( f2, p2, p3, r )
[1]	1466
	1467	ENDIF
	1468
	1469	!
	1470	!-- Now follows the prolongation
	1471	IF ( grid_level >= 2 ) THEN
	1472
	1473	!
	1474	!-- Prolongation of the new residual. The values are transferred
	1475	!-- from the coarse to the next finer grid.
	1476	IF ( grid_level == mg_switch_to_pe0_level+1 ) THEN
[879]	1477
	1478	#if defined( __parallel )
[1]	1479	!
	1480	!-- At this level, the new residual first has to be scattered from
	1481	!-- PE0 to the other PEs
	1482	ALLOCATE( p2_sub(nzb:mg_loc_ind(5,myid)+1, &
	1483	mg_loc_ind(3,myid)-1:mg_loc_ind(4,myid)+1, &
	1484	mg_loc_ind(1,myid)-1:mg_loc_ind(2,myid)+1) )
	1485
	1486	CALL mg_scatter( p2, p2_sub )
	1487
	1488	!
	1489	!-- Therefore, indices of the previous level have to be changed to
	1490	!-- subdomain values in between (otherwise, the prolong routine would
	1491	!-- expect the gathered array)
	1492	nxl_mg_save = nxl_mg(grid_level-1)
	1493	nxr_mg_save = nxr_mg(grid_level-1)
	1494	nys_mg_save = nys_mg(grid_level-1)
	1495	nyn_mg_save = nyn_mg(grid_level-1)
	1496	nzt_mg_save = nzt_mg(grid_level-1)
	1497	nxl_mg(grid_level-1) = mg_loc_ind(1,myid)
	1498	nxr_mg(grid_level-1) = mg_loc_ind(2,myid)
	1499	nys_mg(grid_level-1) = mg_loc_ind(3,myid)
	1500	nyn_mg(grid_level-1) = mg_loc_ind(4,myid)
	1501	nzt_mg(grid_level-1) = mg_loc_ind(5,myid)
	1502
	1503	!
	1504	!-- Set switch for routine exchange_horiz, that ghostpoint exchange
	1505	!-- has to be carried again out from now on
	1506	mg_switch_to_pe0 = .FALSE.
	1507
	1508	!
[707]	1509	!-- For non-cyclic lateral boundary conditions, restore the
	1510	!-- in-/outflow conditions
	1511	inflow_l = .FALSE.; inflow_r = .FALSE.
	1512	inflow_n = .FALSE.; inflow_s = .FALSE.
	1513	outflow_l = .FALSE.; outflow_r = .FALSE.
	1514	outflow_n = .FALSE.; outflow_s = .FALSE.
	1515
	1516	IF ( pleft == MPI_PROC_NULL ) THEN
[1159]	1517	IF ( bc_lr_dirrad ) THEN
[707]	1518	inflow_l = .TRUE.
[1159]	1519	ELSEIF ( bc_lr_raddir ) THEN
[707]	1520	outflow_l = .TRUE.
[1]	1521	ENDIF
[707]	1522	ENDIF
	1523
	1524	IF ( pright == MPI_PROC_NULL ) THEN
[1159]	1525	IF ( bc_lr_dirrad ) THEN
[707]	1526	outflow_r = .TRUE.
[1159]	1527	ELSEIF ( bc_lr_raddir ) THEN
[707]	1528	inflow_r = .TRUE.
[1]	1529	ENDIF
	1530	ENDIF
	1531
[707]	1532	IF ( psouth == MPI_PROC_NULL ) THEN
[1159]	1533	IF ( bc_ns_dirrad ) THEN
[707]	1534	outflow_s = .TRUE.
[1159]	1535	ELSEIF ( bc_ns_raddir ) THEN
[707]	1536	inflow_s = .TRUE.
	1537	ENDIF
	1538	ENDIF
	1539
	1540	IF ( pnorth == MPI_PROC_NULL ) THEN
[1159]	1541	IF ( bc_ns_dirrad ) THEN
[707]	1542	inflow_n = .TRUE.
[1159]	1543	ELSEIF ( bc_ns_raddir ) THEN
[707]	1544	outflow_n = .TRUE.
	1545	ENDIF
	1546	ENDIF
	1547
[1]	1548	CALL prolong( p2_sub, p3 )
	1549
	1550	!
	1551	!-- Restore the correct indices of the previous level
	1552	nxl_mg(grid_level-1) = nxl_mg_save
	1553	nxr_mg(grid_level-1) = nxr_mg_save
	1554	nys_mg(grid_level-1) = nys_mg_save
	1555	nyn_mg(grid_level-1) = nyn_mg_save
	1556	nzt_mg(grid_level-1) = nzt_mg_save
	1557
	1558	DEALLOCATE( p2_sub )
[879]	1559	#endif
[1]	1560
	1561	ELSE
[879]	1562
[1]	1563	CALL prolong( p2, p3 )
	1564
	1565	ENDIF
	1566
	1567	!
	1568	!-- Computation of the new pressure correction. Therefore,
	1569	!-- values from prior grids are added up automatically stage by stage.
	1570	DO i = nxl_mg(grid_level)-1, nxr_mg(grid_level)+1
	1571	DO j = nys_mg(grid_level)-1, nyn_mg(grid_level)+1
	1572	DO k = nzb, nzt_mg(grid_level)+1
	1573	p_mg(k,j,i) = p_mg(k,j,i) + p3(k,j,i)
	1574	ENDDO
	1575	ENDDO
	1576	ENDDO
	1577
	1578	!
	1579	!-- Relaxation of the new solution
	1580	CALL redblack( f_mg, p_mg )
	1581
	1582	ENDIF
	1583
[778]	1584
[1]	1585	!
	1586	!-- The following few lines serve the steering of the multigrid scheme
	1587	IF ( grid_level == maximum_grid_level ) THEN
	1588
	1589	GOTO 20
	1590
	1591	ELSEIF ( grid_level /= maximum_grid_level .AND. grid_level /= 1 .AND. &
	1592	grid_level_count(grid_level) /= gamma_mg ) THEN
	1593
	1594	GOTO 10
	1595
	1596	ENDIF
	1597
	1598	!
	1599	!-- Reset counter for the next call of poismg
	1600	grid_level_count(grid_level) = 0
	1601
	1602	!
	1603	!-- Continue with the next finer level. nxl..nzt have to be
	1604	!-- set to the finer grid values, because these variables are needed for the
	1605	!-- exchange of ghost points in routine exchange_horiz
	1606	grid_level = grid_level + 1
	1607	nxl = nxl_mg(grid_level)
	1608	nxr = nxr_mg(grid_level)
	1609	nys = nys_mg(grid_level)
	1610	nyn = nyn_mg(grid_level)
	1611	nzt = nzt_mg(grid_level)
	1612
	1613	20 CONTINUE
	1614
	1615	END SUBROUTINE next_mg_level

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |