Home

Context Navigation

source: palm/trunk/SOURCE/pres.f90 @ 1434

Last change on this file since 1434 was 1343, checked in by kanani, 11 years ago
last commit documented
Property svn:keywords set to `Id`
File size: 24.0 KB

Rev	Line
[1]	1	SUBROUTINE pres
	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	!
[484]	20	! Current revisions:
[1212]	21	! ------------------
[1343]	22	!
	23	!
[1321]	24	! Former revisions:
	25	! -----------------
	26	! $Id: pres.f90 1343 2014-03-26 17:07:58Z keck $
	27	!
[1343]	28	! 1342 2014-03-26 17:04:47Z kanani
	29	! REAL constants defined as wp-kind
	30	!
[1321]	31	! 1320 2014-03-20 08:40:49Z raasch
[1320]	32	! ONLY-attribute added to USE-statements,
	33	! kind-parameters added to all INTEGER and REAL declaration statements,
	34	! kinds are defined in new module kinds,
	35	! old module precision_kind is removed,
	36	! revision history before 2012 removed,
	37	! comment fields (!:) to be used for variable explanations added to
	38	! all variable declaration statements
[708]	39	!
[1319]	40	! 1318 2014-03-17 13:35:16Z raasch
	41	! module interfaces removed
	42	!
[1307]	43	! 1306 2014-03-13 14:30:59Z raasch
	44	! second argument removed from routine poisfft
	45	!
[1258]	46	! 1257 2013-11-08 15:18:40Z raasch
	47	! openacc loop and loop vector clauses removed, independent clauses added,
	48	! end parallel replaced by end parallel loop
	49	!
[1222]	50	! 1221 2013-09-10 08:59:13Z raasch
	51	! openACC porting of reduction operations, loops for calculating d are
	52	! using the rflags_s_inner multiply flag instead of the nzb_s_inner loop index
	53	!
[1213]	54	! 1212 2013-08-15 08:46:27Z raasch
	55	! call of poisfft_hybrid removed
	56	!
[1118]	57	! 1117 2013-03-27 11:15:36Z suehring
	58	! Bugfix in OpenMP parallelization.
	59	!
[1114]	60	! 1113 2013-03-10 02:48:14Z raasch
	61	! GPU-porting of several loops, some loops rearranged
	62	!
[1112]	63	! 1111 2013-03-08 23:54:10Z
	64	! openACC statements added,
	65	! ibc_p_b = 2 removed
	66	!
[1093]	67	! 1092 2013-02-02 11:24:22Z raasch
	68	! unused variables removed
	69	!
[1037]	70	! 1036 2012-10-22 13:43:42Z raasch
	71	! code put under GPL (PALM 3.9)
	72	!
[1004]	73	! 1003 2012-09-14 14:35:53Z raasch
	74	! adjustment of array tend for cases with unequal subdomain sizes removed
	75	!
[1]	76	! Revision 1.1 1997/07/24 11:24:44 raasch
	77	! Initial revision
	78	!
	79	!
	80	! Description:
	81	! ------------
	82	! Compute the divergence of the provisional velocity field. Solve the Poisson
	83	! equation for the perturbation pressure. Compute the final velocities using
	84	! this perturbation pressure. Compute the remaining divergence.
	85	!------------------------------------------------------------------------------!
	86
[1320]	87	USE arrays_3d, &
	88	ONLY: d, ddzu, ddzu_pres, ddzw, dzw, p, p_loc, tend, u, v, w
	89
	90	USE control_parameters, &
	91	ONLY: bc_lr_cyc, bc_ns_cyc, conserve_volume_flow, dt_3d, &
	92	gathered_size, ibc_p_b, ibc_p_t, intermediate_timestep_count, &
	93	mg_switch_to_pe0_level, on_device, outflow_l, outflow_n, &
	94	outflow_r, outflow_s, psolver, simulated_time, subdomain_size, &
	95	topography, volume_flow, volume_flow_area, volume_flow_initial
	96
	97	USE cpulog, &
	98	ONLY: cpu_log, log_point, log_point_s
	99
	100	USE grid_variables, &
	101	ONLY: ddx, ddy
	102
	103	USE indices, &
	104	ONLY: nbgp, ngp_2dh_outer, nx, nxl, nxlg, nxl_mg, nxr, nxrg, nxr_mg, &
	105	ny, nys, nysg, nys_mg, nyn, nyng, nyn_mg, nzb, nzb_s_inner, &
	106	nzb_u_inner, nzb_v_inner, nzb_w_inner, nzb_2d, nzt, nzt_mg, &
	107	rflags_s_inner
	108
	109	USE kinds
	110
[1]	111	USE pegrid
	112
[1320]	113	USE poisfft_mod, &
	114	ONLY: poisfft
	115
	116	USE statistics, &
	117	ONLY: statistic_regions, sums_divnew_l, sums_divold_l, weight_pres, &
	118	weight_substep
	119
[1]	120	IMPLICIT NONE
	121
[1320]	122	INTEGER(iwp) :: i !:
	123	INTEGER(iwp) :: j !:
	124	INTEGER(iwp) :: k !:
[1]	125
[1320]	126	REAL(wp) :: ddt_3d !:
	127	REAL(wp) :: localsum !:
	128	REAL(wp) :: threadsum !:
	129	REAL(wp) :: d_weight_pres !:
[1]	130
[1320]	131	REAL(wp), DIMENSION(1:2) :: volume_flow_l !:
	132	REAL(wp), DIMENSION(1:2) :: volume_flow_offset !:
	133	REAL(wp), DIMENSION(1:nzt) :: w_l !:
	134	REAL(wp), DIMENSION(1:nzt) :: w_l_l !:
[1]	135
	136
	137	CALL cpu_log( log_point(8), 'pres', 'start' )
	138
[85]	139
[1342]	140	ddt_3d = 1.0_wp / dt_3d
	141	d_weight_pres = 1.0_wp / weight_pres(intermediate_timestep_count)
[85]	142
[1]	143	!
[707]	144	!-- Multigrid method expects array d to have one ghost layer.
	145	!--
[1]	146	IF ( psolver == 'multigrid' ) THEN
[667]	147
[1]	148	DEALLOCATE( d )
[667]	149	ALLOCATE( d(nzb:nzt+1,nys-1:nyn+1,nxl-1:nxr+1) )
[707]	150
	151	!
	152	!-- Since p is later used to hold the weighted average of the substeps, it
	153	!-- cannot be used in the iterative solver. Therefore, its initial value is
	154	!-- stored on p_loc, which is then iteratively advanced in every substep.
	155	IF ( intermediate_timestep_count == 1 ) THEN
	156	DO i = nxl-1, nxr+1
	157	DO j = nys-1, nyn+1
	158	DO k = nzb, nzt+1
	159	p_loc(k,j,i) = p(k,j,i)
	160	ENDDO
	161	ENDDO
	162	ENDDO
[667]	163	ENDIF
	164
[707]	165	ELSEIF ( psolver == 'sor' .AND. intermediate_timestep_count == 1 ) THEN
	166
	167	!
	168	!-- Since p is later used to hold the weighted average of the substeps, it
	169	!-- cannot be used in the iterative solver. Therefore, its initial value is
	170	!-- stored on p_loc, which is then iteratively advanced in every substep.
	171	p_loc = p
	172
[1]	173	ENDIF
	174
	175	!
[75]	176	!-- Conserve the volume flow at the outflow in case of non-cyclic lateral
	177	!-- boundary conditions
[106]	178	!-- WARNING: so far, this conservation does not work at the left/south
	179	!-- boundary if the topography at the inflow differs from that at the
	180	!-- outflow! For this case, volume_flow_area needs adjustment!
	181	!
	182	!-- Left/right
[709]	183	IF ( conserve_volume_flow .AND. ( outflow_l .OR. outflow_r ) ) THEN
[680]	184
[1342]	185	volume_flow(1) = 0.0_wp
	186	volume_flow_l(1) = 0.0_wp
[106]	187
	188	IF ( outflow_l ) THEN
	189	i = 0
	190	ELSEIF ( outflow_r ) THEN
	191	i = nx+1
	192	ENDIF
	193
	194	DO j = nys, nyn
	195	!
	196	!-- Sum up the volume flow through the south/north boundary
[709]	197	DO k = nzb_2d(j,i)+1, nzt
[667]	198	volume_flow_l(1) = volume_flow_l(1) + u(k,j,i) * dzw(k)
[106]	199	ENDDO
	200	ENDDO
	201
	202	#if defined( __parallel )
[680]	203	IF ( collective_wait ) CALL MPI_BARRIER( comm1dy, ierr )
[106]	204	CALL MPI_ALLREDUCE( volume_flow_l(1), volume_flow(1), 1, MPI_REAL, &
	205	MPI_SUM, comm1dy, ierr )
	206	#else
	207	volume_flow = volume_flow_l
	208	#endif
[709]	209	volume_flow_offset(1) = ( volume_flow_initial(1) - volume_flow(1) ) &
[106]	210	/ volume_flow_area(1)
	211
[667]	212	DO j = nysg, nyng
[709]	213	DO k = nzb_2d(j,i)+1, nzt
[106]	214	u(k,j,i) = u(k,j,i) + volume_flow_offset(1)
	215	ENDDO
	216	ENDDO
	217
	218	ENDIF
	219
	220	!
	221	!-- South/north
[709]	222	IF ( conserve_volume_flow .AND. ( outflow_n .OR. outflow_s ) ) THEN
[106]	223
[1342]	224	volume_flow(2) = 0.0_wp
	225	volume_flow_l(2) = 0.0_wp
[75]	226
[106]	227	IF ( outflow_s ) THEN
	228	j = 0
	229	ELSEIF ( outflow_n ) THEN
[75]	230	j = ny+1
[106]	231	ENDIF
	232
	233	DO i = nxl, nxr
[75]	234	!
[106]	235	!-- Sum up the volume flow through the south/north boundary
[709]	236	DO k = nzb_2d(j,i)+1, nzt
[667]	237	volume_flow_l(2) = volume_flow_l(2) + v(k,j,i) * dzw(k)
[75]	238	ENDDO
[106]	239	ENDDO
	240
[75]	241	#if defined( __parallel )
[680]	242	IF ( collective_wait ) CALL MPI_BARRIER( comm1dx, ierr )
[75]	243	CALL MPI_ALLREDUCE( volume_flow_l(2), volume_flow(2), 1, MPI_REAL, &
	244	MPI_SUM, comm1dx, ierr )
	245	#else
	246	volume_flow = volume_flow_l
	247	#endif
	248	volume_flow_offset(2) = ( volume_flow_initial(2) - volume_flow(2) ) &
[106]	249	/ volume_flow_area(2)
[75]	250
[667]	251	DO i = nxlg, nxrg
[709]	252	DO k = nzb_v_inner(j,i)+1, nzt
[106]	253	v(k,j,i) = v(k,j,i) + volume_flow_offset(2)
[75]	254	ENDDO
[106]	255	ENDDO
[75]	256
	257	ENDIF
	258
[76]	259	!
	260	!-- Remove mean vertical velocity
	261	IF ( ibc_p_b == 1 .AND. ibc_p_t == 1 ) THEN
[1342]	262	IF ( simulated_time > 0.0_wp ) THEN ! otherwise nzb_w_inner not yet known
	263	w_l = 0.0_wp; w_l_l = 0.0_wp
[76]	264	DO i = nxl, nxr
	265	DO j = nys, nyn
	266	DO k = nzb_w_inner(j,i)+1, nzt
	267	w_l_l(k) = w_l_l(k) + w(k,j,i)
	268	ENDDO
	269	ENDDO
	270	ENDDO
	271	#if defined( __parallel )
[622]	272	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[709]	273	CALL MPI_ALLREDUCE( w_l_l(1), w_l(1), nzt, MPI_REAL, MPI_SUM, &
	274	comm2d, ierr )
[76]	275	#else
	276	w_l = w_l_l
	277	#endif
	278	DO k = 1, nzt
	279	w_l(k) = w_l(k) / ngp_2dh_outer(k,0)
	280	ENDDO
[667]	281	DO i = nxlg, nxrg
	282	DO j = nysg, nyng
[76]	283	DO k = nzb_w_inner(j,i)+1, nzt
	284	w(k,j,i) = w(k,j,i) - w_l(k)
	285	ENDDO
	286	ENDDO
	287	ENDDO
	288	ENDIF
	289	ENDIF
[75]	290
	291	!
[1]	292	!-- Compute the divergence of the provisional velocity field.
	293	CALL cpu_log( log_point_s(1), 'divergence', 'start' )
	294
	295	IF ( psolver == 'multigrid' ) THEN
	296	!$OMP PARALLEL DO SCHEDULE( STATIC )
	297	DO i = nxl-1, nxr+1
	298	DO j = nys-1, nyn+1
	299	DO k = nzb, nzt+1
[1342]	300	d(k,j,i) = 0.0_wp
[1]	301	ENDDO
	302	ENDDO
	303	ENDDO
	304	ELSE
	305	!$OMP PARALLEL DO SCHEDULE( STATIC )
[1111]	306	!$acc kernels present( d )
[1003]	307	DO i = nxl, nxr
	308	DO j = nys, nyn
	309	DO k = nzb+1, nzt
[1342]	310	d(k,j,i) = 0.0_wp
[1]	311	ENDDO
	312	ENDDO
	313	ENDDO
[1111]	314	!$acc end kernels
[1]	315	ENDIF
	316
[1342]	317	localsum = 0.0_wp
	318	threadsum = 0.0_wp
[1]	319
	320	#if defined( __ibm )
	321	!$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum)
	322	!$OMP DO SCHEDULE( STATIC )
	323	DO i = nxl, nxr
	324	DO j = nys, nyn
	325	DO k = nzb_s_inner(j,i)+1, nzt
[85]	326	d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
	327	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
[673]	328	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k) ) * ddt_3d &
	329	* d_weight_pres
[1]	330	ENDDO
	331	!
	332	!-- Compute possible PE-sum of divergences for flow_statistics
	333	DO k = nzb_s_inner(j,i)+1, nzt
	334	threadsum = threadsum + ABS( d(k,j,i) )
	335	ENDDO
	336
	337	ENDDO
	338	ENDDO
	339
[707]	340	localsum = localsum + threadsum * dt_3d * &
	341	weight_pres(intermediate_timestep_count)
[693]	342
[1]	343	!$OMP END PARALLEL
	344	#else
	345
[1111]	346	!$OMP PARALLEL PRIVATE (i,j,k)
	347	!$OMP DO SCHEDULE( STATIC )
[1221]	348	!$acc kernels present( d, ddzw, rflags_s_inner, u, v, w )
	349	!$acc loop collapse( 3 )
[1111]	350	DO i = nxl, nxr
	351	DO j = nys, nyn
	352	DO k = 1, nzt
[1221]	353	d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
	354	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
	355	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k) ) * ddt_3d &
	356	* d_weight_pres * rflags_s_inner(k,j,i)
[1]	357	ENDDO
	358	ENDDO
[1111]	359	ENDDO
	360	!$acc end kernels
	361	!$OMP END PARALLEL
[1]	362
	363	!
	364	!-- Compute possible PE-sum of divergences for flow_statistics
	365	!$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum)
	366	!$OMP DO SCHEDULE( STATIC )
[1221]	367	!$acc parallel loop collapse(3) present( d ) reduction(+:threadsum)
[1]	368	DO i = nxl, nxr
	369	DO j = nys, nyn
	370	DO k = nzb+1, nzt
	371	threadsum = threadsum + ABS( d(k,j,i) )
	372	ENDDO
	373	ENDDO
	374	ENDDO
[1257]	375	!$acc end parallel loop
[707]	376	localsum = localsum + threadsum * dt_3d * &
	377	weight_pres(intermediate_timestep_count)
[1]	378	!$OMP END PARALLEL
	379	#endif
	380
	381	!
	382	!-- For completeness, set the divergence sum of all statistic regions to those
	383	!-- of the total domain
	384	sums_divold_l(0:statistic_regions) = localsum
	385
	386	CALL cpu_log( log_point_s(1), 'divergence', 'stop' )
	387
	388	!
	389	!-- Compute the pressure perturbation solving the Poisson equation
	390	IF ( psolver(1:7) == 'poisfft' ) THEN
	391
	392	!
	393	!-- Solve Poisson equation via FFT and solution of tridiagonal matrices
	394	IF ( psolver == 'poisfft' ) THEN
[1212]	395
[1306]	396	CALL poisfft( d )
[1113]	397
[1]	398	ENDIF
	399
	400	!
	401	!-- Store computed perturbation pressure and set boundary condition in
	402	!-- z-direction
	403	!$OMP PARALLEL DO
[1113]	404	!$acc kernels present( d, tend )
[1]	405	DO i = nxl, nxr
	406	DO j = nys, nyn
	407	DO k = nzb+1, nzt
	408	tend(k,j,i) = d(k,j,i)
	409	ENDDO
	410	ENDDO
	411	ENDDO
[1113]	412	!$acc end kernels
[1]	413
	414	!
	415	!-- Bottom boundary:
	416	!-- This condition is only required for internal output. The pressure
	417	!-- gradient (dp(nzb+1)-dp(nzb))/dz is not used anywhere else.
	418	IF ( ibc_p_b == 1 ) THEN
	419	!
	420	!-- Neumann (dp/dz = 0)
	421	!$OMP PARALLEL DO
[1113]	422	!$acc kernels present( nzb_s_inner, tend )
[667]	423	DO i = nxlg, nxrg
	424	DO j = nysg, nyng
[1]	425	tend(nzb_s_inner(j,i),j,i) = tend(nzb_s_inner(j,i)+1,j,i)
	426	ENDDO
	427	ENDDO
[1113]	428	!$acc end kernels
[1]	429
	430	ELSE
	431	!
	432	!-- Dirichlet
	433	!$OMP PARALLEL DO
[1113]	434	!$acc kernels present( tend )
[667]	435	DO i = nxlg, nxrg
	436	DO j = nysg, nyng
[1342]	437	tend(nzb_s_inner(j,i),j,i) = 0.0_wp
[1]	438	ENDDO
	439	ENDDO
[1113]	440	!$acc end kernels
[1]	441
	442	ENDIF
	443
	444	!
	445	!-- Top boundary
	446	IF ( ibc_p_t == 1 ) THEN
	447	!
	448	!-- Neumann
	449	!$OMP PARALLEL DO
[1113]	450	!$acc kernels present( tend )
[667]	451	DO i = nxlg, nxrg
	452	DO j = nysg, nyng
[1]	453	tend(nzt+1,j,i) = tend(nzt,j,i)
	454	ENDDO
	455	ENDDO
[1113]	456	!$acc end kernels
[1]	457
	458	ELSE
	459	!
	460	!-- Dirichlet
	461	!$OMP PARALLEL DO
[1113]	462	!$acc kernels present( tend )
[667]	463	DO i = nxlg, nxrg
	464	DO j = nysg, nyng
[1342]	465	tend(nzt+1,j,i) = 0.0_wp
[1]	466	ENDDO
	467	ENDDO
[1113]	468	!$acc end kernels
[1]	469
	470	ENDIF
	471
	472	!
	473	!-- Exchange boundaries for p
[1113]	474	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
	475	on_device = .TRUE. ! to be removed after complete porting
	476	ELSE ! of ghost point exchange
	477	!$acc update host( tend )
	478	ENDIF
[667]	479	CALL exchange_horiz( tend, nbgp )
[1113]	480	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
	481	on_device = .FALSE. ! to be removed after complete porting
	482	ELSE ! of ghost point exchange
	483	!$acc update device( tend )
	484	ENDIF
[1]	485
	486	ELSEIF ( psolver == 'sor' ) THEN
	487
	488	!
	489	!-- Solve Poisson equation for perturbation pressure using SOR-Red/Black
	490	!-- scheme
[707]	491	CALL sor( d, ddzu_pres, ddzw, p_loc )
	492	tend = p_loc
[1]	493
	494	ELSEIF ( psolver == 'multigrid' ) THEN
	495
	496	!
	497	!-- Solve Poisson equation for perturbation pressure using Multigrid scheme,
[667]	498	!-- array tend is used to store the residuals, logical exchange_mg is used
	499	!-- to discern data exchange in multigrid ( 1 ghostpoint ) and normal grid
	500	!-- ( nbgp ghost points ).
[778]	501
	502	!-- If the number of grid points of the gathered grid, which is collected
	503	!-- on PE0, is larger than the number of grid points of an PE, than array
	504	!-- tend will be enlarged.
	505	IF ( gathered_size > subdomain_size ) THEN
	506	DEALLOCATE( tend )
	507	ALLOCATE( tend(nzb:nzt_mg(mg_switch_to_pe0_level)+1,nys_mg( &
	508	mg_switch_to_pe0_level)-1:nyn_mg(mg_switch_to_pe0_level)+1,&
	509	nxl_mg(mg_switch_to_pe0_level)-1:nxr_mg( &
	510	mg_switch_to_pe0_level)+1) )
	511	ENDIF
	512
[1]	513	CALL poismg( tend )
[707]	514
[778]	515	IF ( gathered_size > subdomain_size ) THEN
	516	DEALLOCATE( tend )
	517	ALLOCATE( tend(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	518	ENDIF
	519
[1]	520	!
	521	!-- Restore perturbation pressure on tend because this array is used
	522	!-- further below to correct the velocity fields
[707]	523	DO i = nxl-1, nxr+1
	524	DO j = nys-1, nyn+1
	525	DO k = nzb, nzt+1
	526	tend(k,j,i) = p_loc(k,j,i)
	527	ENDDO
	528	ENDDO
	529	ENDDO
[667]	530
[1]	531	ENDIF
	532
	533	!
[707]	534	!-- Store perturbation pressure on array p, used for pressure data output.
	535	!-- Ghost layers are added in the output routines (except sor-method: see below)
	536	IF ( intermediate_timestep_count == 1 ) THEN
	537	!$OMP PARALLEL PRIVATE (i,j,k)
	538	!$OMP DO
[1113]	539	!$acc kernels present( p, tend, weight_substep )
[1257]	540	!$acc loop independent
[707]	541	DO i = nxl-1, nxr+1
[1257]	542	!$acc loop independent
[707]	543	DO j = nys-1, nyn+1
[1257]	544	!$acc loop independent
[707]	545	DO k = nzb, nzt+1
	546	p(k,j,i) = tend(k,j,i) * &
	547	weight_substep(intermediate_timestep_count)
[673]	548	ENDDO
[1]	549	ENDDO
[707]	550	ENDDO
[1113]	551	!$acc end kernels
[707]	552	!$OMP END PARALLEL
	553
	554	ELSE
	555	!$OMP PARALLEL PRIVATE (i,j,k)
	556	!$OMP DO
[1113]	557	!$acc kernels present( p, tend, weight_substep )
[1257]	558	!$acc loop independent
[707]	559	DO i = nxl-1, nxr+1
[1257]	560	!$acc loop independent
[707]	561	DO j = nys-1, nyn+1
[1257]	562	!$acc loop independent
[707]	563	DO k = nzb, nzt+1
	564	p(k,j,i) = p(k,j,i) + tend(k,j,i) * &
	565	weight_substep(intermediate_timestep_count)
[673]	566	ENDDO
	567	ENDDO
[707]	568	ENDDO
[1113]	569	!$acc end kernels
[707]	570	!$OMP END PARALLEL
	571
	572	ENDIF
[673]	573
[707]	574	!
	575	!-- SOR-method needs ghost layers for the next timestep
	576	IF ( psolver == 'sor' ) CALL exchange_horiz( p, nbgp )
[682]	577
[1]	578	!
	579	!-- Correction of the provisional velocities with the current perturbation
	580	!-- pressure just computed
[709]	581	IF ( conserve_volume_flow .AND. ( bc_lr_cyc .OR. bc_ns_cyc ) ) THEN
[1342]	582	volume_flow_l(1) = 0.0_wp
	583	volume_flow_l(2) = 0.0_wp
[1]	584	ENDIF
[707]	585
[1]	586	!$OMP PARALLEL PRIVATE (i,j,k)
	587	!$OMP DO
[1113]	588	!$acc kernels present( ddzu, nzb_u_inner, nzb_v_inner, nzb_w_inner, tend, u, v, w, weight_pres )
[1257]	589	!$acc loop independent
[673]	590	DO i = nxl, nxr
[1257]	591	!$acc loop independent
[1]	592	DO j = nys, nyn
[1257]	593	!$acc loop independent
[1113]	594	DO k = 1, nzt
	595	IF ( k > nzb_w_inner(j,i) ) THEN
	596	w(k,j,i) = w(k,j,i) - dt_3d * &
	597	( tend(k+1,j,i) - tend(k,j,i) ) * ddzu(k+1) * &
	598	weight_pres(intermediate_timestep_count)
	599	ENDIF
[1]	600	ENDDO
[1257]	601	!$acc loop independent
[1113]	602	DO k = 1, nzt
	603	IF ( k > nzb_u_inner(j,i) ) THEN
	604	u(k,j,i) = u(k,j,i) - dt_3d * &
	605	( tend(k,j,i) - tend(k,j,i-1) ) * ddx * &
	606	weight_pres(intermediate_timestep_count)
	607	ENDIF
[1]	608	ENDDO
[1257]	609	!$acc loop independent
[1113]	610	DO k = 1, nzt
	611	IF ( k > nzb_v_inner(j,i) ) THEN
	612	v(k,j,i) = v(k,j,i) - dt_3d * &
	613	( tend(k,j,i) - tend(k,j-1,i) ) * ddy * &
	614	weight_pres(intermediate_timestep_count)
	615	ENDIF
[673]	616	ENDDO
[1]	617
	618	ENDDO
	619	ENDDO
[1113]	620	!$acc end kernels
[1]	621	!$OMP END PARALLEL
[1113]	622
	623	!
	624	!-- Sum up the volume flow through the right and north boundary
	625	IF ( conserve_volume_flow .AND. bc_lr_cyc .AND. bc_ns_cyc .AND. &
	626	nxr == nx ) THEN
	627
	628	!$OMP PARALLEL PRIVATE (j,k)
	629	!$OMP DO
	630	DO j = nys, nyn
	631	!$OMP CRITICAL
	632	DO k = nzb_2d(j,nx) + 1, nzt
	633	volume_flow_l(1) = volume_flow_l(1) + u(k,j,nx) * dzw(k)
	634	ENDDO
	635	!$OMP END CRITICAL
	636	ENDDO
	637	!$OMP END PARALLEL
	638
	639	ENDIF
	640
	641	IF ( conserve_volume_flow .AND. bc_ns_cyc .AND. bc_lr_cyc .AND. &
	642	nyn == ny ) THEN
	643
	644	!$OMP PARALLEL PRIVATE (i,k)
	645	!$OMP DO
	646	DO i = nxl, nxr
	647	!$OMP CRITICAL
	648	DO k = nzb_2d(ny,i) + 1, nzt
	649	volume_flow_l(2) = volume_flow_l(2) + v(k,ny,i) * dzw(k)
	650	ENDDO
	651	!$OMP END CRITICAL
	652	ENDDO
	653	!$OMP END PARALLEL
	654
	655	ENDIF
[673]	656
[1]	657	!
	658	!-- Conserve the volume flow
[707]	659	IF ( conserve_volume_flow .AND. ( bc_lr_cyc .AND. bc_ns_cyc ) ) THEN
[1]	660
	661	#if defined( __parallel )
[622]	662	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1]	663	CALL MPI_ALLREDUCE( volume_flow_l(1), volume_flow(1), 2, MPI_REAL, &
	664	MPI_SUM, comm2d, ierr )
	665	#else
	666	volume_flow = volume_flow_l
	667	#endif
	668
	669	volume_flow_offset = ( volume_flow_initial - volume_flow ) / &
	670	volume_flow_area
	671
	672	!$OMP PARALLEL PRIVATE (i,j,k)
	673	!$OMP DO
	674	DO i = nxl, nxr
	675	DO j = nys, nyn
[667]	676	DO k = nzb_u_inner(j,i) + 1, nzt
	677	u(k,j,i) = u(k,j,i) + volume_flow_offset(1)
[719]	678	ENDDO
	679	DO k = nzb_v_inner(j,i) + 1, nzt
[667]	680	v(k,j,i) = v(k,j,i) + volume_flow_offset(2)
	681	ENDDO
[1]	682	ENDDO
	683	ENDDO
[667]	684
[1]	685	!$OMP END PARALLEL
	686
	687	ENDIF
	688
	689	!
	690	!-- Exchange of boundaries for the velocities
[1113]	691	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
	692	on_device = .TRUE. ! to be removed after complete porting
	693	ELSE ! of ghost point exchange
	694	!$acc update host( u, v, w )
	695	ENDIF
[667]	696	CALL exchange_horiz( u, nbgp )
	697	CALL exchange_horiz( v, nbgp )
	698	CALL exchange_horiz( w, nbgp )
[1113]	699	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
	700	on_device = .FALSE. ! to be removed after complete porting
	701	ELSE ! of ghost point exchange
	702	!$acc update device( u, v, w )
	703	ENDIF
[1]	704
	705	!
	706	!-- Compute the divergence of the corrected velocity field,
	707	!-- a possible PE-sum is computed in flow_statistics
	708	CALL cpu_log( log_point_s(1), 'divergence', 'start' )
[1342]	709	sums_divnew_l = 0.0_wp
[1]	710
	711	!
	712	!-- d must be reset to zero because it can contain nonzero values below the
	713	!-- topography
[1342]	714	IF ( topography /= 'flat' ) d = 0.0_wp
[1]	715
[1342]	716	localsum = 0.0_wp
	717	threadsum = 0.0_wp
[1]	718
	719	!$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum)
	720	!$OMP DO SCHEDULE( STATIC )
	721	#if defined( __ibm )
	722	DO i = nxl, nxr
	723	DO j = nys, nyn
	724	DO k = nzb_s_inner(j,i)+1, nzt
	725	d(k,j,i) = ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
	726	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
	727	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k)
	728	ENDDO
	729	DO k = nzb+1, nzt
	730	threadsum = threadsum + ABS( d(k,j,i) )
	731	ENDDO
	732	ENDDO
	733	ENDDO
	734	#else
[1221]	735	!$acc kernels present( d, ddzw, rflags_s_inner, u, v, w )
	736	!$acc loop collapse( 3 )
[1]	737	DO i = nxl, nxr
	738	DO j = nys, nyn
[1113]	739	DO k = 1, nzt
[1221]	740	d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
	741	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
	742	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k) &
	743	) * rflags_s_inner(k,j,i)
[1113]	744	ENDDO
	745	ENDDO
	746	ENDDO
	747	!$acc end kernels
	748	!
	749	!-- Compute possible PE-sum of divergences for flow_statistics
[1221]	750	!$acc parallel loop collapse(3) present( d ) reduction(+:threadsum)
[1113]	751	DO i = nxl, nxr
	752	DO j = nys, nyn
[1221]	753	DO k = nzb+1, nzt
[1]	754	threadsum = threadsum + ABS( d(k,j,i) )
	755	ENDDO
	756	ENDDO
	757	ENDDO
[1257]	758	!$acc end parallel loop
[1]	759	#endif
[667]	760
[1]	761	localsum = localsum + threadsum
	762	!$OMP END PARALLEL
	763
	764	!
	765	!-- For completeness, set the divergence sum of all statistic regions to those
	766	!-- of the total domain
	767	sums_divnew_l(0:statistic_regions) = localsum
	768
	769	CALL cpu_log( log_point_s(1), 'divergence', 'stop' )
	770
	771	CALL cpu_log( log_point(8), 'pres', 'stop' )
	772
	773
	774	END SUBROUTINE pres

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |