Home

Context Navigation

source: palm/trunk/SOURCE/pres.f90 @ 1175

Last change on this file since 1175 was 1118, checked in by suehring, 12 years ago
Last commit documented.
Property svn:keywords set to `Id`
File size: 24.1 KB

Line
1	SUBROUTINE pres
2
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	!
17	! Copyright 1997-2012 Leibniz University Hannover
18	!--------------------------------------------------------------------------------!
19	!
20	! Current revisions:
21	! -----------------
22	!
23	! Former revisions:
24	! -----------------
25	! $Id: pres.f90 1118 2013-03-27 11:17:44Z gryschka $
26	!
27	! 1117 2013-03-27 11:15:36Z suehring
28	! Bugfix in OpenMP parallelization.
29	!
30	! 1113 2013-03-10 02:48:14Z raasch
31	! GPU-porting of several loops, some loops rearranged
32	!
33	! 1111 2013-03-08 23:54:10Z
34	! openACC statements added,
35	! ibc_p_b = 2 removed
36	!
37	! 1092 2013-02-02 11:24:22Z raasch
38	! unused variables removed
39	!
40	! 1036 2012-10-22 13:43:42Z raasch
41	! code put under GPL (PALM 3.9)
42	!
43	! 1003 2012-09-14 14:35:53Z raasch
44	! adjustment of array tend for cases with unequal subdomain sizes removed
45	!
46	! 778 2011-11-07 14:18:25Z fricke
47	! New allocation of tend when multigrid is used and the collected field on PE0
48	! has more grid points than the subdomain of an PE.
49	!
50	! 719 2011-04-06 13:05:23Z gryschka
51	! Bugfix in volume flow control for double cyclic boundary conditions
52	!
53	! 709 2011-03-30 09:31:40Z raasch
54	! formatting adjustments
55	!
56	! 707 2011-03-29 11:39:40Z raasch
57	! Calculation of weighted average of p is now handled in the same way
58	! regardless of the number of ghost layers (advection scheme),
59	! multigrid and sor method are using p_loc for iterative advancements of
60	! pressure,
61	! localsum calculation modified for proper OpenMP reduction,
62	! bc_lr/ns replaced by bc_lr/ns_cyc
63	!
64	! 693 2011-03-08 09:..:..Z raasch
65	! bugfix: weighting coefficient added to ibm branch
66	!
67	! 680 2011-02-04 23:16:06Z gryschka
68	! bugfix: collective_wait
69	!
70	! 675 2011-01-19 10:56:55Z suehring
71	! Removed bugfix while copying tend.
72	!
73	! 673 2011-01-18 16:19:48Z suehring
74	! Weighting coefficients added for right computation of the pressure during
75	! Runge-Kutta substeps.
76	!
77	! 667 2010-12-23 12:06:00Z suehring/gryschka
78	! New allocation of tend when ws-scheme and multigrid is used. This is due to
79	! reasons of perforance of the data_exchange. The same is done with p after
80	! poismg is called.
81	! nxl-1, nxr+1, nys-1, nyn+1 replaced by nxlg, nxrg, nysg, nyng when no
82	! multigrid is used. Calls of exchange_horiz are modified.
83	! bugfix: After pressure correction no volume flow correction in case of
84	! non-cyclic boundary conditions
85	! (has to be done only before pressure correction)
86	! Call of SOR routine is referenced with ddzu_pres.
87	!
88	! 622 2010-12-10 08:08:13Z raasch
89	! optional barriers included in order to speed up collective operations
90	!
91	! 151 2008-03-07 13:42:18Z raasch
92	! Bugfix in volume flow control for non-cyclic boundary conditions
93	!
94	! 106 2007-08-16 14:30:26Z raasch
95	! Volume flow conservation added for the remaining three outflow boundaries
96	!
97	! 85 2007-05-11 09:35:14Z raasch
98	! Division through dt_3d replaced by multiplication of the inverse.
99	! For performance optimisation, this is done in the loop calculating the
100	! divergence instead of using a seperate loop.
101	!
102	! 75 2007-03-22 09:54:05Z raasch
103	! Volume flow control for non-cyclic boundary conditions added (currently only
104	! for the north boundary!!), 2nd+3rd argument removed from exchange horiz,
105	! mean vertical velocity is removed in case of Neumann boundary conditions
106	! both at the bottom and the top
107	!
108	! RCS Log replace by Id keyword, revision history cleaned up
109	!
110	! Revision 1.25 2006/04/26 13:26:12 raasch
111	! OpenMP optimization (+localsum, threadsum)
112	!
113	! Revision 1.1 1997/07/24 11:24:44 raasch
114	! Initial revision
115	!
116	!
117	! Description:
118	! ------------
119	! Compute the divergence of the provisional velocity field. Solve the Poisson
120	! equation for the perturbation pressure. Compute the final velocities using
121	! this perturbation pressure. Compute the remaining divergence.
122	!------------------------------------------------------------------------------!
123
124	USE arrays_3d
125	USE constants
126	USE control_parameters
127	USE cpulog
128	USE grid_variables
129	USE indices
130	USE interfaces
131	USE pegrid
132	USE poisfft_mod
133	USE poisfft_hybrid_mod
134	USE statistics
135
136	IMPLICIT NONE
137
138	INTEGER :: i, j, k
139
140	REAL :: ddt_3d, localsum, threadsum, d_weight_pres
141
142	REAL, DIMENSION(1:2) :: volume_flow_l, volume_flow_offset
143	REAL, DIMENSION(1:nzt) :: w_l, w_l_l
144
145
146	CALL cpu_log( log_point(8), 'pres', 'start' )
147
148
149	ddt_3d = 1.0 / dt_3d
150	d_weight_pres = 1.0 / weight_pres(intermediate_timestep_count)
151
152	!
153	!-- Multigrid method expects array d to have one ghost layer.
154	!--
155	IF ( psolver == 'multigrid' ) THEN
156
157	DEALLOCATE( d )
158	ALLOCATE( d(nzb:nzt+1,nys-1:nyn+1,nxl-1:nxr+1) )
159
160	!
161	!-- Since p is later used to hold the weighted average of the substeps, it
162	!-- cannot be used in the iterative solver. Therefore, its initial value is
163	!-- stored on p_loc, which is then iteratively advanced in every substep.
164	IF ( intermediate_timestep_count == 1 ) THEN
165	DO i = nxl-1, nxr+1
166	DO j = nys-1, nyn+1
167	DO k = nzb, nzt+1
168	p_loc(k,j,i) = p(k,j,i)
169	ENDDO
170	ENDDO
171	ENDDO
172	ENDIF
173
174	ELSEIF ( psolver == 'sor' .AND. intermediate_timestep_count == 1 ) THEN
175
176	!
177	!-- Since p is later used to hold the weighted average of the substeps, it
178	!-- cannot be used in the iterative solver. Therefore, its initial value is
179	!-- stored on p_loc, which is then iteratively advanced in every substep.
180	p_loc = p
181
182	ENDIF
183
184	!
185	!-- Conserve the volume flow at the outflow in case of non-cyclic lateral
186	!-- boundary conditions
187	!-- WARNING: so far, this conservation does not work at the left/south
188	!-- boundary if the topography at the inflow differs from that at the
189	!-- outflow! For this case, volume_flow_area needs adjustment!
190	!
191	!-- Left/right
192	IF ( conserve_volume_flow .AND. ( outflow_l .OR. outflow_r ) ) THEN
193
194	volume_flow(1) = 0.0
195	volume_flow_l(1) = 0.0
196
197	IF ( outflow_l ) THEN
198	i = 0
199	ELSEIF ( outflow_r ) THEN
200	i = nx+1
201	ENDIF
202
203	DO j = nys, nyn
204	!
205	!-- Sum up the volume flow through the south/north boundary
206	DO k = nzb_2d(j,i)+1, nzt
207	volume_flow_l(1) = volume_flow_l(1) + u(k,j,i) * dzw(k)
208	ENDDO
209	ENDDO
210
211	#if defined( __parallel )
212	IF ( collective_wait ) CALL MPI_BARRIER( comm1dy, ierr )
213	CALL MPI_ALLREDUCE( volume_flow_l(1), volume_flow(1), 1, MPI_REAL, &
214	MPI_SUM, comm1dy, ierr )
215	#else
216	volume_flow = volume_flow_l
217	#endif
218	volume_flow_offset(1) = ( volume_flow_initial(1) - volume_flow(1) ) &
219	/ volume_flow_area(1)
220
221	DO j = nysg, nyng
222	DO k = nzb_2d(j,i)+1, nzt
223	u(k,j,i) = u(k,j,i) + volume_flow_offset(1)
224	ENDDO
225	ENDDO
226
227	ENDIF
228
229	!
230	!-- South/north
231	IF ( conserve_volume_flow .AND. ( outflow_n .OR. outflow_s ) ) THEN
232
233	volume_flow(2) = 0.0
234	volume_flow_l(2) = 0.0
235
236	IF ( outflow_s ) THEN
237	j = 0
238	ELSEIF ( outflow_n ) THEN
239	j = ny+1
240	ENDIF
241
242	DO i = nxl, nxr
243	!
244	!-- Sum up the volume flow through the south/north boundary
245	DO k = nzb_2d(j,i)+1, nzt
246	volume_flow_l(2) = volume_flow_l(2) + v(k,j,i) * dzw(k)
247	ENDDO
248	ENDDO
249
250	#if defined( __parallel )
251	IF ( collective_wait ) CALL MPI_BARRIER( comm1dx, ierr )
252	CALL MPI_ALLREDUCE( volume_flow_l(2), volume_flow(2), 1, MPI_REAL, &
253	MPI_SUM, comm1dx, ierr )
254	#else
255	volume_flow = volume_flow_l
256	#endif
257	volume_flow_offset(2) = ( volume_flow_initial(2) - volume_flow(2) ) &
258	/ volume_flow_area(2)
259
260	DO i = nxlg, nxrg
261	DO k = nzb_v_inner(j,i)+1, nzt
262	v(k,j,i) = v(k,j,i) + volume_flow_offset(2)
263	ENDDO
264	ENDDO
265
266	ENDIF
267
268	!
269	!-- Remove mean vertical velocity
270	IF ( ibc_p_b == 1 .AND. ibc_p_t == 1 ) THEN
271	IF ( simulated_time > 0.0 ) THEN ! otherwise nzb_w_inner not yet known
272	w_l = 0.0; w_l_l = 0.0
273	DO i = nxl, nxr
274	DO j = nys, nyn
275	DO k = nzb_w_inner(j,i)+1, nzt
276	w_l_l(k) = w_l_l(k) + w(k,j,i)
277	ENDDO
278	ENDDO
279	ENDDO
280	#if defined( __parallel )
281	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
282	CALL MPI_ALLREDUCE( w_l_l(1), w_l(1), nzt, MPI_REAL, MPI_SUM, &
283	comm2d, ierr )
284	#else
285	w_l = w_l_l
286	#endif
287	DO k = 1, nzt
288	w_l(k) = w_l(k) / ngp_2dh_outer(k,0)
289	ENDDO
290	DO i = nxlg, nxrg
291	DO j = nysg, nyng
292	DO k = nzb_w_inner(j,i)+1, nzt
293	w(k,j,i) = w(k,j,i) - w_l(k)
294	ENDDO
295	ENDDO
296	ENDDO
297	ENDIF
298	ENDIF
299
300	!
301	!-- Compute the divergence of the provisional velocity field.
302	CALL cpu_log( log_point_s(1), 'divergence', 'start' )
303
304	IF ( psolver == 'multigrid' ) THEN
305	!$OMP PARALLEL DO SCHEDULE( STATIC )
306	DO i = nxl-1, nxr+1
307	DO j = nys-1, nyn+1
308	DO k = nzb, nzt+1
309	d(k,j,i) = 0.0
310	ENDDO
311	ENDDO
312	ENDDO
313	ELSE
314	!$OMP PARALLEL DO SCHEDULE( STATIC )
315	!$acc kernels present( d )
316	!$acc loop
317	DO i = nxl, nxr
318	DO j = nys, nyn
319	!$acc loop vector(32)
320	DO k = nzb+1, nzt
321	d(k,j,i) = 0.0
322	ENDDO
323	ENDDO
324	ENDDO
325	!$acc end kernels
326	ENDIF
327
328	localsum = 0.0
329	threadsum = 0.0
330
331	#if defined( __ibm )
332	!$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum)
333	!$OMP DO SCHEDULE( STATIC )
334	DO i = nxl, nxr
335	DO j = nys, nyn
336	DO k = nzb_s_inner(j,i)+1, nzt
337	d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
338	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
339	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k) ) * ddt_3d &
340	* d_weight_pres
341	ENDDO
342	!
343	!-- Compute possible PE-sum of divergences for flow_statistics
344	DO k = nzb_s_inner(j,i)+1, nzt
345	threadsum = threadsum + ABS( d(k,j,i) )
346	ENDDO
347
348	ENDDO
349	ENDDO
350
351	localsum = localsum + threadsum * dt_3d * &
352	weight_pres(intermediate_timestep_count)
353
354	!$OMP END PARALLEL
355	#else
356
357	!$OMP PARALLEL PRIVATE (i,j,k)
358	!$OMP DO SCHEDULE( STATIC )
359	!$acc kernels present( d, ddzw, nzb_s_inner, u, v, w )
360	!$acc loop
361	DO i = nxl, nxr
362	DO j = nys, nyn
363	!$acc loop vector(32)
364	DO k = 1, nzt
365	IF ( k > nzb_s_inner(j,i) ) THEN
366	d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
367	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
368	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k) ) * ddt_3d &
369	* d_weight_pres
370	ENDIF
371	ENDDO
372	ENDDO
373	ENDDO
374	!$acc end kernels
375	!$OMP END PARALLEL
376
377	!
378	!-- Compute possible PE-sum of divergences for flow_statistics
379	!$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum)
380	!$OMP DO SCHEDULE( STATIC )
381	DO i = nxl, nxr
382	DO j = nys, nyn
383	DO k = nzb+1, nzt
384	threadsum = threadsum + ABS( d(k,j,i) )
385	ENDDO
386	ENDDO
387	ENDDO
388	localsum = localsum + threadsum * dt_3d * &
389	weight_pres(intermediate_timestep_count)
390	!$OMP END PARALLEL
391	#endif
392
393	!
394	!-- For completeness, set the divergence sum of all statistic regions to those
395	!-- of the total domain
396	sums_divold_l(0:statistic_regions) = localsum
397
398	CALL cpu_log( log_point_s(1), 'divergence', 'stop' )
399
400	!
401	!-- Compute the pressure perturbation solving the Poisson equation
402	IF ( psolver(1:7) == 'poisfft' ) THEN
403
404	!
405	!-- Solve Poisson equation via FFT and solution of tridiagonal matrices
406	IF ( psolver == 'poisfft' ) THEN
407	!
408	!-- Solver for 2d-decomposition
409	CALL poisfft( d, tend )
410
411	ELSEIF ( psolver == 'poisfft_hybrid' ) THEN
412	!
413	!-- Solver for 1d-decomposition (using MPI and OpenMP).
414	!-- The old hybrid-solver is still included here, as long as there
415	!-- are some optimization problems in poisfft
416	CALL poisfft_hybrid( d )
417
418	ENDIF
419
420	!
421	!-- Store computed perturbation pressure and set boundary condition in
422	!-- z-direction
423	!$OMP PARALLEL DO
424	!$acc kernels present( d, tend )
425	!$acc loop
426	DO i = nxl, nxr
427	DO j = nys, nyn
428	!$acc loop vector( 32 )
429	DO k = nzb+1, nzt
430	tend(k,j,i) = d(k,j,i)
431	ENDDO
432	ENDDO
433	ENDDO
434	!$acc end kernels
435
436	!
437	!-- Bottom boundary:
438	!-- This condition is only required for internal output. The pressure
439	!-- gradient (dp(nzb+1)-dp(nzb))/dz is not used anywhere else.
440	IF ( ibc_p_b == 1 ) THEN
441	!
442	!-- Neumann (dp/dz = 0)
443	!$OMP PARALLEL DO
444	!$acc kernels present( nzb_s_inner, tend )
445	DO i = nxlg, nxrg
446	DO j = nysg, nyng
447	tend(nzb_s_inner(j,i),j,i) = tend(nzb_s_inner(j,i)+1,j,i)
448	ENDDO
449	ENDDO
450	!$acc end kernels
451
452	ELSE
453	!
454	!-- Dirichlet
455	!$OMP PARALLEL DO
456	!$acc kernels present( tend )
457	DO i = nxlg, nxrg
458	DO j = nysg, nyng
459	tend(nzb_s_inner(j,i),j,i) = 0.0
460	ENDDO
461	ENDDO
462	!$acc end kernels
463
464	ENDIF
465
466	!
467	!-- Top boundary
468	IF ( ibc_p_t == 1 ) THEN
469	!
470	!-- Neumann
471	!$OMP PARALLEL DO
472	!$acc kernels present( tend )
473	DO i = nxlg, nxrg
474	DO j = nysg, nyng
475	tend(nzt+1,j,i) = tend(nzt,j,i)
476	ENDDO
477	ENDDO
478	!$acc end kernels
479
480	ELSE
481	!
482	!-- Dirichlet
483	!$OMP PARALLEL DO
484	!$acc kernels present( tend )
485	DO i = nxlg, nxrg
486	DO j = nysg, nyng
487	tend(nzt+1,j,i) = 0.0
488	ENDDO
489	ENDDO
490	!$acc end kernels
491
492	ENDIF
493
494	!
495	!-- Exchange boundaries for p
496	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
497	on_device = .TRUE. ! to be removed after complete porting
498	ELSE ! of ghost point exchange
499	!$acc update host( tend )
500	ENDIF
501	CALL exchange_horiz( tend, nbgp )
502	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
503	on_device = .FALSE. ! to be removed after complete porting
504	ELSE ! of ghost point exchange
505	!$acc update device( tend )
506	ENDIF
507
508	ELSEIF ( psolver == 'sor' ) THEN
509
510	!
511	!-- Solve Poisson equation for perturbation pressure using SOR-Red/Black
512	!-- scheme
513	CALL sor( d, ddzu_pres, ddzw, p_loc )
514	tend = p_loc
515
516	ELSEIF ( psolver == 'multigrid' ) THEN
517
518	!
519	!-- Solve Poisson equation for perturbation pressure using Multigrid scheme,
520	!-- array tend is used to store the residuals, logical exchange_mg is used
521	!-- to discern data exchange in multigrid ( 1 ghostpoint ) and normal grid
522	!-- ( nbgp ghost points ).
523
524	!-- If the number of grid points of the gathered grid, which is collected
525	!-- on PE0, is larger than the number of grid points of an PE, than array
526	!-- tend will be enlarged.
527	IF ( gathered_size > subdomain_size ) THEN
528	DEALLOCATE( tend )
529	ALLOCATE( tend(nzb:nzt_mg(mg_switch_to_pe0_level)+1,nys_mg( &
530	mg_switch_to_pe0_level)-1:nyn_mg(mg_switch_to_pe0_level)+1,&
531	nxl_mg(mg_switch_to_pe0_level)-1:nxr_mg( &
532	mg_switch_to_pe0_level)+1) )
533	ENDIF
534
535	CALL poismg( tend )
536
537	IF ( gathered_size > subdomain_size ) THEN
538	DEALLOCATE( tend )
539	ALLOCATE( tend(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
540	ENDIF
541
542	!
543	!-- Restore perturbation pressure on tend because this array is used
544	!-- further below to correct the velocity fields
545	DO i = nxl-1, nxr+1
546	DO j = nys-1, nyn+1
547	DO k = nzb, nzt+1
548	tend(k,j,i) = p_loc(k,j,i)
549	ENDDO
550	ENDDO
551	ENDDO
552
553	ENDIF
554
555	!
556	!-- Store perturbation pressure on array p, used for pressure data output.
557	!-- Ghost layers are added in the output routines (except sor-method: see below)
558	IF ( intermediate_timestep_count == 1 ) THEN
559	!$OMP PARALLEL PRIVATE (i,j,k)
560	!$OMP DO
561	!$acc kernels present( p, tend, weight_substep )
562	!$acc loop
563	DO i = nxl-1, nxr+1
564	DO j = nys-1, nyn+1
565	!$acc loop vector( 32 )
566	DO k = nzb, nzt+1
567	p(k,j,i) = tend(k,j,i) * &
568	weight_substep(intermediate_timestep_count)
569	ENDDO
570	ENDDO
571	ENDDO
572	!$acc end kernels
573	!$OMP END PARALLEL
574
575	ELSE
576	!$OMP PARALLEL PRIVATE (i,j,k)
577	!$OMP DO
578	!$acc kernels present( p, tend, weight_substep )
579	!$acc loop
580	DO i = nxl-1, nxr+1
581	DO j = nys-1, nyn+1
582	!$acc loop vector( 32 )
583	DO k = nzb, nzt+1
584	p(k,j,i) = p(k,j,i) + tend(k,j,i) * &
585	weight_substep(intermediate_timestep_count)
586	ENDDO
587	ENDDO
588	ENDDO
589	!$acc end kernels
590	!$OMP END PARALLEL
591
592	ENDIF
593
594	!
595	!-- SOR-method needs ghost layers for the next timestep
596	IF ( psolver == 'sor' ) CALL exchange_horiz( p, nbgp )
597
598	!
599	!-- Correction of the provisional velocities with the current perturbation
600	!-- pressure just computed
601	!$acc update host( u, v, w )
602	IF ( conserve_volume_flow .AND. ( bc_lr_cyc .OR. bc_ns_cyc ) ) THEN
603	volume_flow_l(1) = 0.0
604	volume_flow_l(2) = 0.0
605	ENDIF
606
607	!$OMP PARALLEL PRIVATE (i,j,k)
608	!$OMP DO
609	!$acc kernels present( ddzu, nzb_u_inner, nzb_v_inner, nzb_w_inner, tend, u, v, w, weight_pres )
610	!$acc loop
611	DO i = nxl, nxr
612	DO j = nys, nyn
613	!$acc loop vector( 32 )
614	DO k = 1, nzt
615	IF ( k > nzb_w_inner(j,i) ) THEN
616	w(k,j,i) = w(k,j,i) - dt_3d * &
617	( tend(k+1,j,i) - tend(k,j,i) ) * ddzu(k+1) * &
618	weight_pres(intermediate_timestep_count)
619	ENDIF
620	ENDDO
621	!$acc loop vector( 32 )
622	DO k = 1, nzt
623	IF ( k > nzb_u_inner(j,i) ) THEN
624	u(k,j,i) = u(k,j,i) - dt_3d * &
625	( tend(k,j,i) - tend(k,j,i-1) ) * ddx * &
626	weight_pres(intermediate_timestep_count)
627	ENDIF
628	ENDDO
629	!$acc loop vector( 32 )
630	DO k = 1, nzt
631	IF ( k > nzb_v_inner(j,i) ) THEN
632	v(k,j,i) = v(k,j,i) - dt_3d * &
633	( tend(k,j,i) - tend(k,j-1,i) ) * ddy * &
634	weight_pres(intermediate_timestep_count)
635	ENDIF
636	ENDDO
637
638	ENDDO
639	ENDDO
640	!$acc end kernels
641	!$OMP END PARALLEL
642
643	!
644	!-- Sum up the volume flow through the right and north boundary
645	IF ( conserve_volume_flow .AND. bc_lr_cyc .AND. bc_ns_cyc .AND. &
646	nxr == nx ) THEN
647
648	!$OMP PARALLEL PRIVATE (j,k)
649	!$OMP DO
650	DO j = nys, nyn
651	!$OMP CRITICAL
652	DO k = nzb_2d(j,nx) + 1, nzt
653	volume_flow_l(1) = volume_flow_l(1) + u(k,j,nx) * dzw(k)
654	ENDDO
655	!$OMP END CRITICAL
656	ENDDO
657	!$OMP END PARALLEL
658
659	ENDIF
660
661	IF ( conserve_volume_flow .AND. bc_ns_cyc .AND. bc_lr_cyc .AND. &
662	nyn == ny ) THEN
663
664	!$OMP PARALLEL PRIVATE (i,k)
665	!$OMP DO
666	DO i = nxl, nxr
667	!$OMP CRITICAL
668	DO k = nzb_2d(ny,i) + 1, nzt
669	volume_flow_l(2) = volume_flow_l(2) + v(k,ny,i) * dzw(k)
670	ENDDO
671	!$OMP END CRITICAL
672	ENDDO
673	!$OMP END PARALLEL
674
675	ENDIF
676
677	!
678	!-- Conserve the volume flow
679	IF ( conserve_volume_flow .AND. ( bc_lr_cyc .AND. bc_ns_cyc ) ) THEN
680
681	#if defined( __parallel )
682	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
683	CALL MPI_ALLREDUCE( volume_flow_l(1), volume_flow(1), 2, MPI_REAL, &
684	MPI_SUM, comm2d, ierr )
685	#else
686	volume_flow = volume_flow_l
687	#endif
688
689	volume_flow_offset = ( volume_flow_initial - volume_flow ) / &
690	volume_flow_area
691
692	!$OMP PARALLEL PRIVATE (i,j,k)
693	!$OMP DO
694	DO i = nxl, nxr
695	DO j = nys, nyn
696	DO k = nzb_u_inner(j,i) + 1, nzt
697	u(k,j,i) = u(k,j,i) + volume_flow_offset(1)
698	ENDDO
699	DO k = nzb_v_inner(j,i) + 1, nzt
700	v(k,j,i) = v(k,j,i) + volume_flow_offset(2)
701	ENDDO
702	ENDDO
703	ENDDO
704
705	!$OMP END PARALLEL
706
707	ENDIF
708
709	!
710	!-- Exchange of boundaries for the velocities
711	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
712	on_device = .TRUE. ! to be removed after complete porting
713	ELSE ! of ghost point exchange
714	!$acc update host( u, v, w )
715	ENDIF
716	CALL exchange_horiz( u, nbgp )
717	CALL exchange_horiz( v, nbgp )
718	CALL exchange_horiz( w, nbgp )
719	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
720	on_device = .FALSE. ! to be removed after complete porting
721	ELSE ! of ghost point exchange
722	!$acc update device( u, v, w )
723	ENDIF
724
725	!
726	!-- Compute the divergence of the corrected velocity field,
727	!-- a possible PE-sum is computed in flow_statistics
728	CALL cpu_log( log_point_s(1), 'divergence', 'start' )
729	sums_divnew_l = 0.0
730
731	!
732	!-- d must be reset to zero because it can contain nonzero values below the
733	!-- topography
734	IF ( topography /= 'flat' ) d = 0.0
735
736	localsum = 0.0
737	threadsum = 0.0
738
739	!$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum)
740	!$OMP DO SCHEDULE( STATIC )
741	#if defined( __ibm )
742	DO i = nxl, nxr
743	DO j = nys, nyn
744	DO k = nzb_s_inner(j,i)+1, nzt
745	d(k,j,i) = ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
746	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
747	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k)
748	ENDDO
749	DO k = nzb+1, nzt
750	threadsum = threadsum + ABS( d(k,j,i) )
751	ENDDO
752	ENDDO
753	ENDDO
754	#else
755	!$acc kernels present( d, ddzw, nzb_s_inner, u, v, w )
756	!$acc loop
757	DO i = nxl, nxr
758	DO j = nys, nyn
759	!$acc loop vector( 32 )
760	DO k = 1, nzt
761	IF ( k > nzb_s_inner(j,i) ) THEN
762	d(k,j,i) = ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
763	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
764	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k)
765	ENDIF
766	ENDDO
767	ENDDO
768	ENDDO
769	!$acc end kernels
770	!
771	!-- Compute possible PE-sum of divergences for flow_statistics
772	DO i = nxl, nxr
773	DO j = nys, nyn
774	DO k = nzb_s_inner(j,i)+1, nzt
775	threadsum = threadsum + ABS( d(k,j,i) )
776	ENDDO
777	ENDDO
778	ENDDO
779	#endif
780
781	localsum = localsum + threadsum
782	!$OMP END PARALLEL
783
784	!
785	!-- For completeness, set the divergence sum of all statistic regions to those
786	!-- of the total domain
787	sums_divnew_l(0:statistic_regions) = localsum
788
789	CALL cpu_log( log_point_s(1), 'divergence', 'stop' )
790
791	CALL cpu_log( log_point(8), 'pres', 'stop' )
792
793
794	END SUBROUTINE pres

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |