Home

Context Navigation

source: palm/trunk/SOURCE/pres.f90 @ 1321

Last change on this file since 1321 was 1321, checked in by raasch, 10 years ago
last commit documented
Property svn:keywords set to `Id`
File size: 23.8 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 1997-2014 Leibniz Universitaet Hannover
18	!--------------------------------------------------------------------------------!
19	!
20	! Current revisions:
21	! ------------------
22	!
23	!
24	! Former revisions:
25	! -----------------
26	! $Id: pres.f90 1321 2014-03-20 09:40:40Z raasch $
27	!
28	! 1320 2014-03-20 08:40:49Z raasch
29	! ONLY-attribute added to USE-statements,
30	! kind-parameters added to all INTEGER and REAL declaration statements,
31	! kinds are defined in new module kinds,
32	! old module precision_kind is removed,
33	! revision history before 2012 removed,
34	! comment fields (!:) to be used for variable explanations added to
35	! all variable declaration statements
36	!
37	! 1318 2014-03-17 13:35:16Z raasch
38	! module interfaces removed
39	!
40	! 1306 2014-03-13 14:30:59Z raasch
41	! second argument removed from routine poisfft
42	!
43	! 1257 2013-11-08 15:18:40Z raasch
44	! openacc loop and loop vector clauses removed, independent clauses added,
45	! end parallel replaced by end parallel loop
46	!
47	! 1221 2013-09-10 08:59:13Z raasch
48	! openACC porting of reduction operations, loops for calculating d are
49	! using the rflags_s_inner multiply flag instead of the nzb_s_inner loop index
50	!
51	! 1212 2013-08-15 08:46:27Z raasch
52	! call of poisfft_hybrid removed
53	!
54	! 1117 2013-03-27 11:15:36Z suehring
55	! Bugfix in OpenMP parallelization.
56	!
57	! 1113 2013-03-10 02:48:14Z raasch
58	! GPU-porting of several loops, some loops rearranged
59	!
60	! 1111 2013-03-08 23:54:10Z
61	! openACC statements added,
62	! ibc_p_b = 2 removed
63	!
64	! 1092 2013-02-02 11:24:22Z raasch
65	! unused variables removed
66	!
67	! 1036 2012-10-22 13:43:42Z raasch
68	! code put under GPL (PALM 3.9)
69	!
70	! 1003 2012-09-14 14:35:53Z raasch
71	! adjustment of array tend for cases with unequal subdomain sizes removed
72	!
73	! Revision 1.1 1997/07/24 11:24:44 raasch
74	! Initial revision
75	!
76	!
77	! Description:
78	! ------------
79	! Compute the divergence of the provisional velocity field. Solve the Poisson
80	! equation for the perturbation pressure. Compute the final velocities using
81	! this perturbation pressure. Compute the remaining divergence.
82	!------------------------------------------------------------------------------!
83
84	USE arrays_3d, &
85	ONLY: d, ddzu, ddzu_pres, ddzw, dzw, p, p_loc, tend, u, v, w
86
87	USE control_parameters, &
88	ONLY: bc_lr_cyc, bc_ns_cyc, conserve_volume_flow, dt_3d, &
89	gathered_size, ibc_p_b, ibc_p_t, intermediate_timestep_count, &
90	mg_switch_to_pe0_level, on_device, outflow_l, outflow_n, &
91	outflow_r, outflow_s, psolver, simulated_time, subdomain_size, &
92	topography, volume_flow, volume_flow_area, volume_flow_initial
93
94	USE cpulog, &
95	ONLY: cpu_log, log_point, log_point_s
96
97	USE grid_variables, &
98	ONLY: ddx, ddy
99
100	USE indices, &
101	ONLY: nbgp, ngp_2dh_outer, nx, nxl, nxlg, nxl_mg, nxr, nxrg, nxr_mg, &
102	ny, nys, nysg, nys_mg, nyn, nyng, nyn_mg, nzb, nzb_s_inner, &
103	nzb_u_inner, nzb_v_inner, nzb_w_inner, nzb_2d, nzt, nzt_mg, &
104	rflags_s_inner
105
106	USE kinds
107
108	USE pegrid
109
110	USE poisfft_mod, &
111	ONLY: poisfft
112
113	USE statistics, &
114	ONLY: statistic_regions, sums_divnew_l, sums_divold_l, weight_pres, &
115	weight_substep
116
117	IMPLICIT NONE
118
119	INTEGER(iwp) :: i !:
120	INTEGER(iwp) :: j !:
121	INTEGER(iwp) :: k !:
122
123	REAL(wp) :: ddt_3d !:
124	REAL(wp) :: localsum !:
125	REAL(wp) :: threadsum !:
126	REAL(wp) :: d_weight_pres !:
127
128	REAL(wp), DIMENSION(1:2) :: volume_flow_l !:
129	REAL(wp), DIMENSION(1:2) :: volume_flow_offset !:
130	REAL(wp), DIMENSION(1:nzt) :: w_l !:
131	REAL(wp), DIMENSION(1:nzt) :: w_l_l !:
132
133
134	CALL cpu_log( log_point(8), 'pres', 'start' )
135
136
137	ddt_3d = 1.0 / dt_3d
138	d_weight_pres = 1.0 / weight_pres(intermediate_timestep_count)
139
140	!
141	!-- Multigrid method expects array d to have one ghost layer.
142	!--
143	IF ( psolver == 'multigrid' ) THEN
144
145	DEALLOCATE( d )
146	ALLOCATE( d(nzb:nzt+1,nys-1:nyn+1,nxl-1:nxr+1) )
147
148	!
149	!-- Since p is later used to hold the weighted average of the substeps, it
150	!-- cannot be used in the iterative solver. Therefore, its initial value is
151	!-- stored on p_loc, which is then iteratively advanced in every substep.
152	IF ( intermediate_timestep_count == 1 ) THEN
153	DO i = nxl-1, nxr+1
154	DO j = nys-1, nyn+1
155	DO k = nzb, nzt+1
156	p_loc(k,j,i) = p(k,j,i)
157	ENDDO
158	ENDDO
159	ENDDO
160	ENDIF
161
162	ELSEIF ( psolver == 'sor' .AND. intermediate_timestep_count == 1 ) THEN
163
164	!
165	!-- Since p is later used to hold the weighted average of the substeps, it
166	!-- cannot be used in the iterative solver. Therefore, its initial value is
167	!-- stored on p_loc, which is then iteratively advanced in every substep.
168	p_loc = p
169
170	ENDIF
171
172	!
173	!-- Conserve the volume flow at the outflow in case of non-cyclic lateral
174	!-- boundary conditions
175	!-- WARNING: so far, this conservation does not work at the left/south
176	!-- boundary if the topography at the inflow differs from that at the
177	!-- outflow! For this case, volume_flow_area needs adjustment!
178	!
179	!-- Left/right
180	IF ( conserve_volume_flow .AND. ( outflow_l .OR. outflow_r ) ) THEN
181
182	volume_flow(1) = 0.0
183	volume_flow_l(1) = 0.0
184
185	IF ( outflow_l ) THEN
186	i = 0
187	ELSEIF ( outflow_r ) THEN
188	i = nx+1
189	ENDIF
190
191	DO j = nys, nyn
192	!
193	!-- Sum up the volume flow through the south/north boundary
194	DO k = nzb_2d(j,i)+1, nzt
195	volume_flow_l(1) = volume_flow_l(1) + u(k,j,i) * dzw(k)
196	ENDDO
197	ENDDO
198
199	#if defined( __parallel )
200	IF ( collective_wait ) CALL MPI_BARRIER( comm1dy, ierr )
201	CALL MPI_ALLREDUCE( volume_flow_l(1), volume_flow(1), 1, MPI_REAL, &
202	MPI_SUM, comm1dy, ierr )
203	#else
204	volume_flow = volume_flow_l
205	#endif
206	volume_flow_offset(1) = ( volume_flow_initial(1) - volume_flow(1) ) &
207	/ volume_flow_area(1)
208
209	DO j = nysg, nyng
210	DO k = nzb_2d(j,i)+1, nzt
211	u(k,j,i) = u(k,j,i) + volume_flow_offset(1)
212	ENDDO
213	ENDDO
214
215	ENDIF
216
217	!
218	!-- South/north
219	IF ( conserve_volume_flow .AND. ( outflow_n .OR. outflow_s ) ) THEN
220
221	volume_flow(2) = 0.0
222	volume_flow_l(2) = 0.0
223
224	IF ( outflow_s ) THEN
225	j = 0
226	ELSEIF ( outflow_n ) THEN
227	j = ny+1
228	ENDIF
229
230	DO i = nxl, nxr
231	!
232	!-- Sum up the volume flow through the south/north boundary
233	DO k = nzb_2d(j,i)+1, nzt
234	volume_flow_l(2) = volume_flow_l(2) + v(k,j,i) * dzw(k)
235	ENDDO
236	ENDDO
237
238	#if defined( __parallel )
239	IF ( collective_wait ) CALL MPI_BARRIER( comm1dx, ierr )
240	CALL MPI_ALLREDUCE( volume_flow_l(2), volume_flow(2), 1, MPI_REAL, &
241	MPI_SUM, comm1dx, ierr )
242	#else
243	volume_flow = volume_flow_l
244	#endif
245	volume_flow_offset(2) = ( volume_flow_initial(2) - volume_flow(2) ) &
246	/ volume_flow_area(2)
247
248	DO i = nxlg, nxrg
249	DO k = nzb_v_inner(j,i)+1, nzt
250	v(k,j,i) = v(k,j,i) + volume_flow_offset(2)
251	ENDDO
252	ENDDO
253
254	ENDIF
255
256	!
257	!-- Remove mean vertical velocity
258	IF ( ibc_p_b == 1 .AND. ibc_p_t == 1 ) THEN
259	IF ( simulated_time > 0.0 ) THEN ! otherwise nzb_w_inner not yet known
260	w_l = 0.0; w_l_l = 0.0
261	DO i = nxl, nxr
262	DO j = nys, nyn
263	DO k = nzb_w_inner(j,i)+1, nzt
264	w_l_l(k) = w_l_l(k) + w(k,j,i)
265	ENDDO
266	ENDDO
267	ENDDO
268	#if defined( __parallel )
269	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
270	CALL MPI_ALLREDUCE( w_l_l(1), w_l(1), nzt, MPI_REAL, MPI_SUM, &
271	comm2d, ierr )
272	#else
273	w_l = w_l_l
274	#endif
275	DO k = 1, nzt
276	w_l(k) = w_l(k) / ngp_2dh_outer(k,0)
277	ENDDO
278	DO i = nxlg, nxrg
279	DO j = nysg, nyng
280	DO k = nzb_w_inner(j,i)+1, nzt
281	w(k,j,i) = w(k,j,i) - w_l(k)
282	ENDDO
283	ENDDO
284	ENDDO
285	ENDIF
286	ENDIF
287
288	!
289	!-- Compute the divergence of the provisional velocity field.
290	CALL cpu_log( log_point_s(1), 'divergence', 'start' )
291
292	IF ( psolver == 'multigrid' ) THEN
293	!$OMP PARALLEL DO SCHEDULE( STATIC )
294	DO i = nxl-1, nxr+1
295	DO j = nys-1, nyn+1
296	DO k = nzb, nzt+1
297	d(k,j,i) = 0.0
298	ENDDO
299	ENDDO
300	ENDDO
301	ELSE
302	!$OMP PARALLEL DO SCHEDULE( STATIC )
303	!$acc kernels present( d )
304	DO i = nxl, nxr
305	DO j = nys, nyn
306	DO k = nzb+1, nzt
307	d(k,j,i) = 0.0
308	ENDDO
309	ENDDO
310	ENDDO
311	!$acc end kernels
312	ENDIF
313
314	localsum = 0.0
315	threadsum = 0.0
316
317	#if defined( __ibm )
318	!$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum)
319	!$OMP DO SCHEDULE( STATIC )
320	DO i = nxl, nxr
321	DO j = nys, nyn
322	DO k = nzb_s_inner(j,i)+1, nzt
323	d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
324	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
325	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k) ) * ddt_3d &
326	* d_weight_pres
327	ENDDO
328	!
329	!-- Compute possible PE-sum of divergences for flow_statistics
330	DO k = nzb_s_inner(j,i)+1, nzt
331	threadsum = threadsum + ABS( d(k,j,i) )
332	ENDDO
333
334	ENDDO
335	ENDDO
336
337	localsum = localsum + threadsum * dt_3d * &
338	weight_pres(intermediate_timestep_count)
339
340	!$OMP END PARALLEL
341	#else
342
343	!$OMP PARALLEL PRIVATE (i,j,k)
344	!$OMP DO SCHEDULE( STATIC )
345	!$acc kernels present( d, ddzw, rflags_s_inner, u, v, w )
346	!$acc loop collapse( 3 )
347	DO i = nxl, nxr
348	DO j = nys, nyn
349	DO k = 1, nzt
350	d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
351	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
352	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k) ) * ddt_3d &
353	* d_weight_pres * rflags_s_inner(k,j,i)
354	ENDDO
355	ENDDO
356	ENDDO
357	!$acc end kernels
358	!$OMP END PARALLEL
359
360	!
361	!-- Compute possible PE-sum of divergences for flow_statistics
362	!$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum)
363	!$OMP DO SCHEDULE( STATIC )
364	!$acc parallel loop collapse(3) present( d ) reduction(+:threadsum)
365	DO i = nxl, nxr
366	DO j = nys, nyn
367	DO k = nzb+1, nzt
368	threadsum = threadsum + ABS( d(k,j,i) )
369	ENDDO
370	ENDDO
371	ENDDO
372	!$acc end parallel loop
373	localsum = localsum + threadsum * dt_3d * &
374	weight_pres(intermediate_timestep_count)
375	!$OMP END PARALLEL
376	#endif
377
378	!
379	!-- For completeness, set the divergence sum of all statistic regions to those
380	!-- of the total domain
381	sums_divold_l(0:statistic_regions) = localsum
382
383	CALL cpu_log( log_point_s(1), 'divergence', 'stop' )
384
385	!
386	!-- Compute the pressure perturbation solving the Poisson equation
387	IF ( psolver(1:7) == 'poisfft' ) THEN
388
389	!
390	!-- Solve Poisson equation via FFT and solution of tridiagonal matrices
391	IF ( psolver == 'poisfft' ) THEN
392
393	CALL poisfft( d )
394
395	ENDIF
396
397	!
398	!-- Store computed perturbation pressure and set boundary condition in
399	!-- z-direction
400	!$OMP PARALLEL DO
401	!$acc kernels present( d, tend )
402	DO i = nxl, nxr
403	DO j = nys, nyn
404	DO k = nzb+1, nzt
405	tend(k,j,i) = d(k,j,i)
406	ENDDO
407	ENDDO
408	ENDDO
409	!$acc end kernels
410
411	!
412	!-- Bottom boundary:
413	!-- This condition is only required for internal output. The pressure
414	!-- gradient (dp(nzb+1)-dp(nzb))/dz is not used anywhere else.
415	IF ( ibc_p_b == 1 ) THEN
416	!
417	!-- Neumann (dp/dz = 0)
418	!$OMP PARALLEL DO
419	!$acc kernels present( nzb_s_inner, tend )
420	DO i = nxlg, nxrg
421	DO j = nysg, nyng
422	tend(nzb_s_inner(j,i),j,i) = tend(nzb_s_inner(j,i)+1,j,i)
423	ENDDO
424	ENDDO
425	!$acc end kernels
426
427	ELSE
428	!
429	!-- Dirichlet
430	!$OMP PARALLEL DO
431	!$acc kernels present( tend )
432	DO i = nxlg, nxrg
433	DO j = nysg, nyng
434	tend(nzb_s_inner(j,i),j,i) = 0.0
435	ENDDO
436	ENDDO
437	!$acc end kernels
438
439	ENDIF
440
441	!
442	!-- Top boundary
443	IF ( ibc_p_t == 1 ) THEN
444	!
445	!-- Neumann
446	!$OMP PARALLEL DO
447	!$acc kernels present( tend )
448	DO i = nxlg, nxrg
449	DO j = nysg, nyng
450	tend(nzt+1,j,i) = tend(nzt,j,i)
451	ENDDO
452	ENDDO
453	!$acc end kernels
454
455	ELSE
456	!
457	!-- Dirichlet
458	!$OMP PARALLEL DO
459	!$acc kernels present( tend )
460	DO i = nxlg, nxrg
461	DO j = nysg, nyng
462	tend(nzt+1,j,i) = 0.0
463	ENDDO
464	ENDDO
465	!$acc end kernels
466
467	ENDIF
468
469	!
470	!-- Exchange boundaries for p
471	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
472	on_device = .TRUE. ! to be removed after complete porting
473	ELSE ! of ghost point exchange
474	!$acc update host( tend )
475	ENDIF
476	CALL exchange_horiz( tend, nbgp )
477	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
478	on_device = .FALSE. ! to be removed after complete porting
479	ELSE ! of ghost point exchange
480	!$acc update device( tend )
481	ENDIF
482
483	ELSEIF ( psolver == 'sor' ) THEN
484
485	!
486	!-- Solve Poisson equation for perturbation pressure using SOR-Red/Black
487	!-- scheme
488	CALL sor( d, ddzu_pres, ddzw, p_loc )
489	tend = p_loc
490
491	ELSEIF ( psolver == 'multigrid' ) THEN
492
493	!
494	!-- Solve Poisson equation for perturbation pressure using Multigrid scheme,
495	!-- array tend is used to store the residuals, logical exchange_mg is used
496	!-- to discern data exchange in multigrid ( 1 ghostpoint ) and normal grid
497	!-- ( nbgp ghost points ).
498
499	!-- If the number of grid points of the gathered grid, which is collected
500	!-- on PE0, is larger than the number of grid points of an PE, than array
501	!-- tend will be enlarged.
502	IF ( gathered_size > subdomain_size ) THEN
503	DEALLOCATE( tend )
504	ALLOCATE( tend(nzb:nzt_mg(mg_switch_to_pe0_level)+1,nys_mg( &
505	mg_switch_to_pe0_level)-1:nyn_mg(mg_switch_to_pe0_level)+1,&
506	nxl_mg(mg_switch_to_pe0_level)-1:nxr_mg( &
507	mg_switch_to_pe0_level)+1) )
508	ENDIF
509
510	CALL poismg( tend )
511
512	IF ( gathered_size > subdomain_size ) THEN
513	DEALLOCATE( tend )
514	ALLOCATE( tend(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
515	ENDIF
516
517	!
518	!-- Restore perturbation pressure on tend because this array is used
519	!-- further below to correct the velocity fields
520	DO i = nxl-1, nxr+1
521	DO j = nys-1, nyn+1
522	DO k = nzb, nzt+1
523	tend(k,j,i) = p_loc(k,j,i)
524	ENDDO
525	ENDDO
526	ENDDO
527
528	ENDIF
529
530	!
531	!-- Store perturbation pressure on array p, used for pressure data output.
532	!-- Ghost layers are added in the output routines (except sor-method: see below)
533	IF ( intermediate_timestep_count == 1 ) THEN
534	!$OMP PARALLEL PRIVATE (i,j,k)
535	!$OMP DO
536	!$acc kernels present( p, tend, weight_substep )
537	!$acc loop independent
538	DO i = nxl-1, nxr+1
539	!$acc loop independent
540	DO j = nys-1, nyn+1
541	!$acc loop independent
542	DO k = nzb, nzt+1
543	p(k,j,i) = tend(k,j,i) * &
544	weight_substep(intermediate_timestep_count)
545	ENDDO
546	ENDDO
547	ENDDO
548	!$acc end kernels
549	!$OMP END PARALLEL
550
551	ELSE
552	!$OMP PARALLEL PRIVATE (i,j,k)
553	!$OMP DO
554	!$acc kernels present( p, tend, weight_substep )
555	!$acc loop independent
556	DO i = nxl-1, nxr+1
557	!$acc loop independent
558	DO j = nys-1, nyn+1
559	!$acc loop independent
560	DO k = nzb, nzt+1
561	p(k,j,i) = p(k,j,i) + tend(k,j,i) * &
562	weight_substep(intermediate_timestep_count)
563	ENDDO
564	ENDDO
565	ENDDO
566	!$acc end kernels
567	!$OMP END PARALLEL
568
569	ENDIF
570
571	!
572	!-- SOR-method needs ghost layers for the next timestep
573	IF ( psolver == 'sor' ) CALL exchange_horiz( p, nbgp )
574
575	!
576	!-- Correction of the provisional velocities with the current perturbation
577	!-- pressure just computed
578	IF ( conserve_volume_flow .AND. ( bc_lr_cyc .OR. bc_ns_cyc ) ) THEN
579	volume_flow_l(1) = 0.0
580	volume_flow_l(2) = 0.0
581	ENDIF
582
583	!$OMP PARALLEL PRIVATE (i,j,k)
584	!$OMP DO
585	!$acc kernels present( ddzu, nzb_u_inner, nzb_v_inner, nzb_w_inner, tend, u, v, w, weight_pres )
586	!$acc loop independent
587	DO i = nxl, nxr
588	!$acc loop independent
589	DO j = nys, nyn
590	!$acc loop independent
591	DO k = 1, nzt
592	IF ( k > nzb_w_inner(j,i) ) THEN
593	w(k,j,i) = w(k,j,i) - dt_3d * &
594	( tend(k+1,j,i) - tend(k,j,i) ) * ddzu(k+1) * &
595	weight_pres(intermediate_timestep_count)
596	ENDIF
597	ENDDO
598	!$acc loop independent
599	DO k = 1, nzt
600	IF ( k > nzb_u_inner(j,i) ) THEN
601	u(k,j,i) = u(k,j,i) - dt_3d * &
602	( tend(k,j,i) - tend(k,j,i-1) ) * ddx * &
603	weight_pres(intermediate_timestep_count)
604	ENDIF
605	ENDDO
606	!$acc loop independent
607	DO k = 1, nzt
608	IF ( k > nzb_v_inner(j,i) ) THEN
609	v(k,j,i) = v(k,j,i) - dt_3d * &
610	( tend(k,j,i) - tend(k,j-1,i) ) * ddy * &
611	weight_pres(intermediate_timestep_count)
612	ENDIF
613	ENDDO
614
615	ENDDO
616	ENDDO
617	!$acc end kernels
618	!$OMP END PARALLEL
619
620	!
621	!-- Sum up the volume flow through the right and north boundary
622	IF ( conserve_volume_flow .AND. bc_lr_cyc .AND. bc_ns_cyc .AND. &
623	nxr == nx ) THEN
624
625	!$OMP PARALLEL PRIVATE (j,k)
626	!$OMP DO
627	DO j = nys, nyn
628	!$OMP CRITICAL
629	DO k = nzb_2d(j,nx) + 1, nzt
630	volume_flow_l(1) = volume_flow_l(1) + u(k,j,nx) * dzw(k)
631	ENDDO
632	!$OMP END CRITICAL
633	ENDDO
634	!$OMP END PARALLEL
635
636	ENDIF
637
638	IF ( conserve_volume_flow .AND. bc_ns_cyc .AND. bc_lr_cyc .AND. &
639	nyn == ny ) THEN
640
641	!$OMP PARALLEL PRIVATE (i,k)
642	!$OMP DO
643	DO i = nxl, nxr
644	!$OMP CRITICAL
645	DO k = nzb_2d(ny,i) + 1, nzt
646	volume_flow_l(2) = volume_flow_l(2) + v(k,ny,i) * dzw(k)
647	ENDDO
648	!$OMP END CRITICAL
649	ENDDO
650	!$OMP END PARALLEL
651
652	ENDIF
653
654	!
655	!-- Conserve the volume flow
656	IF ( conserve_volume_flow .AND. ( bc_lr_cyc .AND. bc_ns_cyc ) ) THEN
657
658	#if defined( __parallel )
659	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
660	CALL MPI_ALLREDUCE( volume_flow_l(1), volume_flow(1), 2, MPI_REAL, &
661	MPI_SUM, comm2d, ierr )
662	#else
663	volume_flow = volume_flow_l
664	#endif
665
666	volume_flow_offset = ( volume_flow_initial - volume_flow ) / &
667	volume_flow_area
668
669	!$OMP PARALLEL PRIVATE (i,j,k)
670	!$OMP DO
671	DO i = nxl, nxr
672	DO j = nys, nyn
673	DO k = nzb_u_inner(j,i) + 1, nzt
674	u(k,j,i) = u(k,j,i) + volume_flow_offset(1)
675	ENDDO
676	DO k = nzb_v_inner(j,i) + 1, nzt
677	v(k,j,i) = v(k,j,i) + volume_flow_offset(2)
678	ENDDO
679	ENDDO
680	ENDDO
681
682	!$OMP END PARALLEL
683
684	ENDIF
685
686	!
687	!-- Exchange of boundaries for the velocities
688	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
689	on_device = .TRUE. ! to be removed after complete porting
690	ELSE ! of ghost point exchange
691	!$acc update host( u, v, w )
692	ENDIF
693	CALL exchange_horiz( u, nbgp )
694	CALL exchange_horiz( v, nbgp )
695	CALL exchange_horiz( w, nbgp )
696	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
697	on_device = .FALSE. ! to be removed after complete porting
698	ELSE ! of ghost point exchange
699	!$acc update device( u, v, w )
700	ENDIF
701
702	!
703	!-- Compute the divergence of the corrected velocity field,
704	!-- a possible PE-sum is computed in flow_statistics
705	CALL cpu_log( log_point_s(1), 'divergence', 'start' )
706	sums_divnew_l = 0.0
707
708	!
709	!-- d must be reset to zero because it can contain nonzero values below the
710	!-- topography
711	IF ( topography /= 'flat' ) d = 0.0
712
713	localsum = 0.0
714	threadsum = 0.0
715
716	!$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum)
717	!$OMP DO SCHEDULE( STATIC )
718	#if defined( __ibm )
719	DO i = nxl, nxr
720	DO j = nys, nyn
721	DO k = nzb_s_inner(j,i)+1, nzt
722	d(k,j,i) = ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
723	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
724	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k)
725	ENDDO
726	DO k = nzb+1, nzt
727	threadsum = threadsum + ABS( d(k,j,i) )
728	ENDDO
729	ENDDO
730	ENDDO
731	#else
732	!$acc kernels present( d, ddzw, rflags_s_inner, u, v, w )
733	!$acc loop collapse( 3 )
734	DO i = nxl, nxr
735	DO j = nys, nyn
736	DO k = 1, nzt
737	d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
738	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
739	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k) &
740	) * rflags_s_inner(k,j,i)
741	ENDDO
742	ENDDO
743	ENDDO
744	!$acc end kernels
745	!
746	!-- Compute possible PE-sum of divergences for flow_statistics
747	!$acc parallel loop collapse(3) present( d ) reduction(+:threadsum)
748	DO i = nxl, nxr
749	DO j = nys, nyn
750	DO k = nzb+1, nzt
751	threadsum = threadsum + ABS( d(k,j,i) )
752	ENDDO
753	ENDDO
754	ENDDO
755	!$acc end parallel loop
756	#endif
757
758	localsum = localsum + threadsum
759	!$OMP END PARALLEL
760
761	!
762	!-- For completeness, set the divergence sum of all statistic regions to those
763	!-- of the total domain
764	sums_divnew_l(0:statistic_regions) = localsum
765
766	CALL cpu_log( log_point_s(1), 'divergence', 'stop' )
767
768	CALL cpu_log( log_point(8), 'pres', 'stop' )
769
770
771	END SUBROUTINE pres

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |