Home

Context Navigation

source: palm/trunk/SOURCE/init_pegrid.f90 @ 1848

Last change on this file since 1848 was 1834, checked in by raasch, 9 years ago
last commit documented
Property svn:keywords set to `Id`
File size: 43.1 KB

Line
1	!> @file init_pegrid.f90
2	!--------------------------------------------------------------------------------!
3	! This file is part of PALM.
4	!
5	! PALM is free software: you can redistribute it and/or modify it under the terms
6	! of the GNU General Public License as published by the Free Software Foundation,
7	! either version 3 of the License, or (at your option) any later version.
8	!
9	! PALM is distributed in the hope that it will be useful, but WITHOUT ANY
10	! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
11	! A PARTICULAR PURPOSE. See the GNU General Public License for more details.
12	!
13	! You should have received a copy of the GNU General Public License along with
14	! PALM. If not, see <http://www.gnu.org/licenses/>.
15	!
16	! Copyright 1997-2016 Leibniz Universitaet Hannover
17	!--------------------------------------------------------------------------------!
18	!
19	! Current revisions:
20	! ------------------
21	!
22	!
23	! Former revisions:
24	! -----------------
25	! $Id: init_pegrid.f90 1834 2016-04-07 14:34:20Z witha $
26	!
27	! 1833 2016-04-07 14:23:03Z raasch
28	! spectra related variables moved to spectra_mod
29	!
30	! 1815 2016-04-06 13:49:59Z raasch
31	! cpp-directives for intel openmp bug removed
32	!
33	! 1804 2016-04-05 16:30:18Z maronga
34	! Removed code for parameter file check (__check)
35	!
36	! 1779 2016-03-03 08:01:28Z raasch
37	! changes regarding nested domain removed: virtual PE grid will be automatically
38	! calculated for nested runs too
39	!
40	! 1764 2016-02-28 12:45:19Z raasch
41	! cpp-statements for nesting removed
42	!
43	! 1762 2016-02-25 12:31:13Z hellstea
44	! Introduction of nested domain feature
45	!
46	! 1682 2015-10-07 23:56:08Z knoop
47	! Code annotations made doxygen readable
48	!
49	! 1677 2015-10-02 13:25:23Z boeske
50	! New MPI-data types for exchange of 3D integer arrays.
51	!
52	! 1575 2015-03-27 09:56:27Z raasch
53	! adjustments for psolver-queries, calculation of ngp_xz added
54	!
55	! 1565 2015-03-09 20:59:31Z suehring
56	! Refine if-clause for setting nbgp.
57	!
58	! 1557 2015-03-05 16:43:04Z suehring
59	! Adjustment for monotonic limiter
60	!
61	! 1468 2014-09-24 14:06:57Z maronga
62	! Adapted for use on up to 6-digit processor cores
63	!
64	! 1435 2014-07-21 10:37:02Z keck
65	! bugfix: added missing parameter coupling_mode_remote to ONLY-attribute
66	!
67	! 1402 2014-05-09 14:25:13Z raasch
68	! location messages modified
69	!
70	! 1384 2014-05-02 14:31:06Z raasch
71	! location messages added
72	!
73	! 1353 2014-04-08 15:21:23Z heinze
74	! REAL constants provided with KIND-attribute
75	!
76	! 1322 2014-03-20 16:38:49Z raasch
77	! REAL functions provided with KIND-attribute
78	!
79	! 1320 2014-03-20 08:40:49Z raasch
80	! ONLY-attribute added to USE-statements,
81	! kind-parameters added to all INTEGER and REAL declaration statements,
82	! kinds are defined in new module kinds,
83	! revision history before 2012 removed,
84	! comment fields (!:) to be used for variable explanations added to
85	! all variable declaration statements
86	!
87	! 1304 2014-03-12 10:29:42Z raasch
88	! bugfix: single core MPI runs missed some settings of transpose indices
89	!
90	! 1212 2013-08-15 08:46:27Z raasch
91	! error message for poisfft_hybrid removed
92	!
93	! 1159 2013-05-21 11:58:22Z fricke
94	! dirichlet/neumann and neumann/dirichlet removed
95	!
96	! 1139 2013-04-18 07:25:03Z raasch
97	! bugfix for calculating the id of the PE carrying the recycling plane
98	!
99	! 1111 2013-03-08 23:54:10Z raasch
100	! initialization of poisfft moved to module poisfft
101	!
102	! 1092 2013-02-02 11:24:22Z raasch
103	! unused variables removed
104	!
105	! 1056 2012-11-16 15:28:04Z raasch
106	! Indices for arrays n.._mg start from zero due to definition of arrays f2 and
107	! p2 as automatic arrays in recursive subroutine next_mg_level
108	!
109	! 1041 2012-11-06 02:36:29Z raasch
110	! a 2d virtual processor topology is used by default for all machines
111	!
112	! 1036 2012-10-22 13:43:42Z raasch
113	! code put under GPL (PALM 3.9)
114	!
115	! 1003 2012-09-14 14:35:53Z raasch
116	! subdomains must have identical size (grid matching = "match" removed)
117	!
118	! 1001 2012-09-13 14:08:46Z raasch
119	! all actions concerning upstream-spline-method removed
120	!
121	! 978 2012-08-09 08:28:32Z fricke
122	! dirichlet/neumann and neumann/dirichlet added
123	! nxlu and nysv are also calculated for inflow boundary
124	!
125	! 809 2012-01-30 13:32:58Z maronga
126	! Bugfix: replaced .AND. and .NOT. with && and ! in the preprocessor directives
127	!
128	! 807 2012-01-25 11:53:51Z maronga
129	! New cpp directive "__check" implemented which is used by check_namelist_files
130	!
131	! Revision 1.1 1997/07/24 11:15:09 raasch
132	! Initial revision
133	!
134	!
135	! Description:
136	! ------------
137	!> Determination of the virtual processor topology (if not prescribed by the
138	!> user)and computation of the grid point number and array bounds of the local
139	!> domains.
140	!------------------------------------------------------------------------------!
141	SUBROUTINE init_pegrid
142
143
144	USE control_parameters, &
145	ONLY: bc_lr, bc_ns, coupling_mode, coupling_mode_remote, &
146	coupling_topology, gathered_size, grid_level, &
147	grid_level_count, host, inflow_l, inflow_n, inflow_r, inflow_s, &
148	io_blocks, io_group, maximum_grid_level, &
149	maximum_parallel_io_streams, message_string, &
150	mg_switch_to_pe0_level, momentum_advec, nest_bound_l, &
151	nest_bound_n, nest_bound_r, nest_bound_s, neutral, psolver, &
152	outflow_l, outflow_n, outflow_r, outflow_s, recycling_width, &
153	scalar_advec, subdomain_size
154
155	USE grid_variables, &
156	ONLY: dx
157
158	USE indices, &
159	ONLY: mg_loc_ind, nbgp, nnx, nny, nnz, nx, nx_a, nx_o, nxl, nxl_mg, &
160	nxlu, nxr, nxr_mg, ny, ny_a, ny_o, nyn, nyn_mg, nys, nys_mg, &
161	nysv, nz, nzb, nzt, nzt_mg, wall_flags_1, wall_flags_2, &
162	wall_flags_3, wall_flags_4, wall_flags_5, wall_flags_6, &
163	wall_flags_7, wall_flags_8, wall_flags_9, wall_flags_10
164
165	USE kinds
166
167	USE pegrid
168
169	USE spectra_mod, &
170	ONLY: dt_dosp
171
172	USE transpose_indices, &
173	ONLY: nxl_y, nxl_yd, nxl_z, nxr_y, nxr_yd, nxr_z, nyn_x, nyn_z, nys_x,&
174	nys_z, nzb_x, nzb_y, nzb_yd, nzt_x, nzt_yd, nzt_y
175
176	IMPLICIT NONE
177
178	INTEGER(iwp) :: i !<
179	INTEGER(iwp) :: id_inflow_l !<
180	INTEGER(iwp) :: id_recycling_l !<
181	INTEGER(iwp) :: ind(5) !<
182	INTEGER(iwp) :: j !<
183	INTEGER(iwp) :: k !<
184	INTEGER(iwp) :: maximum_grid_level_l !<
185	INTEGER(iwp) :: mg_levels_x !<
186	INTEGER(iwp) :: mg_levels_y !<
187	INTEGER(iwp) :: mg_levels_z !<
188	INTEGER(iwp) :: mg_switch_to_pe0_level_l !<
189	INTEGER(iwp) :: nnx_y !<
190	INTEGER(iwp) :: nnx_z !<
191	INTEGER(iwp) :: nny_x !<
192	INTEGER(iwp) :: nny_z !<
193	INTEGER(iwp) :: nnz_x !<
194	INTEGER(iwp) :: nnz_y !<
195	INTEGER(iwp) :: numproc_sqr !<
196	INTEGER(iwp) :: nxl_l !<
197	INTEGER(iwp) :: nxr_l !<
198	INTEGER(iwp) :: nyn_l !<
199	INTEGER(iwp) :: nys_l !<
200	INTEGER(iwp) :: nzb_l !<
201	INTEGER(iwp) :: nzt_l !<
202	INTEGER(iwp) :: omp_get_num_threads !<
203
204	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: ind_all !<
205	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: nxlf !<
206	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: nxrf !<
207	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: nynf !<
208	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: nysf !<
209
210	INTEGER(iwp), DIMENSION(2) :: pdims_remote !<
211
212	#if defined( __mpi2 )
213	LOGICAL :: found !<
214	#endif
215
216	!
217	!-- Get the number of OpenMP threads
218	!$OMP PARALLEL
219	!$ threads_per_task = omp_get_num_threads()
220	!$OMP END PARALLEL
221
222
223	#if defined( __parallel )
224
225	CALL location_message( 'creating virtual PE grids + MPI derived data types', &
226	.FALSE. )
227
228	!
229	!-- Determine the processor topology or check it, if prescribed by the user
230	IF ( npex == -1 .AND. npey == -1 ) THEN
231
232	!
233	!-- Automatic determination of the topology
234	numproc_sqr = SQRT( REAL( numprocs, KIND=wp ) )
235	pdims(1) = MAX( numproc_sqr , 1 )
236	DO WHILE ( MOD( numprocs , pdims(1) ) /= 0 )
237	pdims(1) = pdims(1) - 1
238	ENDDO
239	pdims(2) = numprocs / pdims(1)
240
241	ELSEIF ( npex /= -1 .AND. npey /= -1 ) THEN
242
243	!
244	!-- Prescribed by user. Number of processors on the prescribed topology
245	!-- must be equal to the number of PEs available to the job
246	IF ( ( npex * npey ) /= numprocs ) THEN
247	WRITE( message_string, * ) 'number of PEs of the prescribed ', &
248	'topology (', npex*npey,') does not match & the number of ', &
249	'PEs available to the job (', numprocs, ')'
250	CALL message( 'init_pegrid', 'PA0221', 1, 2, 0, 6, 0 )
251	ENDIF
252	pdims(1) = npex
253	pdims(2) = npey
254
255	ELSE
256	!
257	!-- If the processor topology is prescribed by the user, the number of
258	!-- PEs must be given in both directions
259	message_string = 'if the processor topology is prescribed by th' // &
260	'e user& both values of "npex" and "npey" must be given' // &
261	' in the &NAMELIST-parameter file'
262	CALL message( 'init_pegrid', 'PA0222', 1, 2, 0, 6, 0 )
263
264	ENDIF
265
266	!
267	!-- For communication speedup, set barriers in front of collective
268	!-- communications by default on SGI-type systems
269	IF ( host(3:5) == 'sgi' ) collective_wait = .TRUE.
270
271	!
272	!-- If necessary, set horizontal boundary conditions to non-cyclic
273	IF ( bc_lr /= 'cyclic' ) cyclic(1) = .FALSE.
274	IF ( bc_ns /= 'cyclic' ) cyclic(2) = .FALSE.
275
276
277	!
278	!-- Create the virtual processor grid
279	CALL MPI_CART_CREATE( comm_palm, ndim, pdims, cyclic, reorder, &
280	comm2d, ierr )
281	CALL MPI_COMM_RANK( comm2d, myid, ierr )
282	WRITE (myid_char,'(''_'',I6.6)') myid
283
284	CALL MPI_CART_COORDS( comm2d, myid, ndim, pcoord, ierr )
285	CALL MPI_CART_SHIFT( comm2d, 0, 1, pleft, pright, ierr )
286	CALL MPI_CART_SHIFT( comm2d, 1, 1, psouth, pnorth, ierr )
287
288	!
289	!-- Determine sub-topologies for transpositions
290	!-- Transposition from z to x:
291	remain_dims(1) = .TRUE.
292	remain_dims(2) = .FALSE.
293	CALL MPI_CART_SUB( comm2d, remain_dims, comm1dx, ierr )
294	CALL MPI_COMM_RANK( comm1dx, myidx, ierr )
295	!
296	!-- Transposition from x to y
297	remain_dims(1) = .FALSE.
298	remain_dims(2) = .TRUE.
299	CALL MPI_CART_SUB( comm2d, remain_dims, comm1dy, ierr )
300	CALL MPI_COMM_RANK( comm1dy, myidy, ierr )
301
302
303	!
304	!-- Calculate array bounds along x-direction for every PE.
305	ALLOCATE( nxlf(0:pdims(1)-1), nxrf(0:pdims(1)-1), nynf(0:pdims(2)-1), &
306	nysf(0:pdims(2)-1) )
307
308	IF ( MOD( nx+1 , pdims(1) ) /= 0 ) THEN
309	WRITE( message_string, * ) 'x-direction: gridpoint number (',nx+1,') ',&
310	'is not an& integral divisor of the number ', &
311	'processors (', pdims(1),')'
312	CALL message( 'init_pegrid', 'PA0225', 1, 2, 0, 6, 0 )
313	ELSE
314	nnx = ( nx + 1 ) / pdims(1)
315	IF ( nnx*pdims(1) - ( nx + 1) > nnx ) THEN
316	WRITE( message_string, * ) 'x-direction: nx does not match the', &
317	'requirements given by the number of PEs &used', &
318	'& please use nx = ', nx - ( pdims(1) - ( nnx*pdims(1) &
319	- ( nx + 1 ) ) ), ' instead of nx =', nx
320	CALL message( 'init_pegrid', 'PA0226', 1, 2, 0, 6, 0 )
321	ENDIF
322	ENDIF
323
324	!
325	!-- Left and right array bounds, number of gridpoints
326	DO i = 0, pdims(1)-1
327	nxlf(i) = i * nnx
328	nxrf(i) = ( i + 1 ) * nnx - 1
329	ENDDO
330
331	!
332	!-- Calculate array bounds in y-direction for every PE.
333	IF ( MOD( ny+1 , pdims(2) ) /= 0 ) THEN
334	WRITE( message_string, * ) 'y-direction: gridpoint number (',ny+1,') ', &
335	'is not an& integral divisor of the number of', &
336	'processors (', pdims(2),')'
337	CALL message( 'init_pegrid', 'PA0227', 1, 2, 0, 6, 0 )
338	ELSE
339	nny = ( ny + 1 ) / pdims(2)
340	IF ( nny*pdims(2) - ( ny + 1) > nny ) THEN
341	WRITE( message_string, * ) 'y-direction: ny does not match the', &
342	'requirements given by the number of PEs &used ', &
343	'& please use ny = ', ny - ( pdims(2) - ( nnx*pdims(2) &
344	- ( ny + 1 ) ) ), ' instead of ny =', ny
345	CALL message( 'init_pegrid', 'PA0228', 1, 2, 0, 6, 0 )
346	ENDIF
347	ENDIF
348
349	!
350	!-- South and north array bounds
351	DO j = 0, pdims(2)-1
352	nysf(j) = j * nny
353	nynf(j) = ( j + 1 ) * nny - 1
354	ENDDO
355
356	!
357	!-- Local array bounds of the respective PEs
358	nxl = nxlf(pcoord(1))
359	nxr = nxrf(pcoord(1))
360	nys = nysf(pcoord(2))
361	nyn = nynf(pcoord(2))
362	nzb = 0
363	nzt = nz
364	nnz = nz
365
366	!
367	!-- Set switches to define if the PE is situated at the border of the virtual
368	!-- processor grid
369	IF ( nxl == 0 ) left_border_pe = .TRUE.
370	IF ( nxr == nx ) right_border_pe = .TRUE.
371	IF ( nys == 0 ) south_border_pe = .TRUE.
372	IF ( nyn == ny ) north_border_pe = .TRUE.
373
374	!
375	!-- Calculate array bounds and gridpoint numbers for the transposed arrays
376	!-- (needed in the pressure solver)
377	!-- For the transposed arrays, cyclic boundaries as well as top and bottom
378	!-- boundaries are omitted, because they are obstructive to the transposition
379
380	!
381	!-- 1. transposition z --> x
382	!-- This transposition is not neccessary in case of a 1d-decomposition along x
383	nys_x = nys
384	nyn_x = nyn
385	nny_x = nny
386	nnz_x = nz / pdims(1)
387	nzb_x = 1 + myidx * nnz_x
388	nzt_x = ( myidx + 1 ) * nnz_x
389	sendrecvcount_zx = nnx * nny * nnz_x
390
391	IF ( pdims(2) /= 1 ) THEN
392	IF ( MOD( nz , pdims(1) ) /= 0 ) THEN
393	WRITE( message_string, * ) 'transposition z --> x:', &
394	'&nz=',nz,' is not an integral divisior of pdims(1)=', &
395	pdims(1)
396	CALL message( 'init_pegrid', 'PA0230', 1, 2, 0, 6, 0 )
397	ENDIF
398	ENDIF
399
400	!
401	!-- 2. transposition x --> y
402	nnz_y = nnz_x
403	nzb_y = nzb_x
404	nzt_y = nzt_x
405	IF ( MOD( nx+1 , pdims(2) ) /= 0 ) THEN
406	WRITE( message_string, * ) 'transposition x --> y:', &
407	'&nx+1=',nx+1,' is not an integral divisor of ',&
408	'pdims(2)=',pdims(2)
409	CALL message( 'init_pegrid', 'PA0231', 1, 2, 0, 6, 0 )
410	ENDIF
411	nnx_y = (nx+1) / pdims(2)
412	nxl_y = myidy * nnx_y
413	nxr_y = ( myidy + 1 ) * nnx_y - 1
414	sendrecvcount_xy = nnx_y * nny_x * nnz_y
415
416	!
417	!-- 3. transposition y --> z (ELSE: x --> y in case of 1D-decomposition
418	!-- along x)
419	nnx_z = nnx_y
420	nxl_z = nxl_y
421	nxr_z = nxr_y
422	nny_z = (ny+1) / pdims(1)
423	nys_z = myidx * nny_z
424	nyn_z = ( myidx + 1 ) * nny_z - 1
425	sendrecvcount_yz = nnx_y * nny_z * nnz_y
426
427	IF ( pdims(2) /= 1 ) THEN
428	!
429	!-- y --> z
430	!-- This transposition is not neccessary in case of a 1d-decomposition
431	!-- along x, except that the uptream-spline method is switched on
432	IF ( MOD( ny+1 , pdims(1) ) /= 0 ) THEN
433	WRITE( message_string, * ) 'transposition y --> z:', &
434	'& ny+1=',ny+1,' is not an integral divisor of ',&
435	'pdims(1)=',pdims(1)
436	CALL message( 'init_pegrid', 'PA0232', 1, 2, 0, 6, 0 )
437	ENDIF
438
439	ELSE
440	!
441	!-- x --> y. This condition must be fulfilled for a 1D-decomposition along x
442	IF ( MOD( ny+1 , pdims(1) ) /= 0 ) THEN
443	WRITE( message_string, * ) 'transposition x --> y:', &
444	'& ny+1=',ny+1,' is not an integral divisor of ',&
445	'pdims(1)=',pdims(1)
446	CALL message( 'init_pegrid', 'PA0233', 1, 2, 0, 6, 0 )
447	ENDIF
448
449	ENDIF
450
451	!
452	!-- Indices for direct transpositions z --> y (used for calculating spectra)
453	IF ( dt_dosp /= 9999999.9_wp ) THEN
454	IF ( MOD( nz, pdims(2) ) /= 0 ) THEN
455	WRITE( message_string, * ) 'direct transposition z --> y (needed ', &
456	'for spectra):& nz=',nz,' is not an integral divisor of ',&
457	'pdims(2)=',pdims(2)
458	CALL message( 'init_pegrid', 'PA0234', 1, 2, 0, 6, 0 )
459	ELSE
460	nxl_yd = nxl
461	nxr_yd = nxr
462	nzb_yd = 1 + myidy * ( nz / pdims(2) )
463	nzt_yd = ( myidy + 1 ) * ( nz / pdims(2) )
464	sendrecvcount_zyd = nnx * nny * ( nz / pdims(2) )
465	ENDIF
466	ENDIF
467
468	!
469	!-- Indices for direct transpositions y --> x (they are only possible in case
470	!-- of a 1d-decomposition along x)
471	IF ( pdims(2) == 1 ) THEN
472	nny_x = nny / pdims(1)
473	nys_x = myid * nny_x
474	nyn_x = ( myid + 1 ) * nny_x - 1
475	nzb_x = 1
476	nzt_x = nz
477	sendrecvcount_xy = nnx * nny_x * nz
478	ENDIF
479
480	!
481	!-- Indices for direct transpositions x --> y (they are only possible in case
482	!-- of a 1d-decomposition along y)
483	IF ( pdims(1) == 1 ) THEN
484	nnx_y = nnx / pdims(2)
485	nxl_y = myid * nnx_y
486	nxr_y = ( myid + 1 ) * nnx_y - 1
487	nzb_y = 1
488	nzt_y = nz
489	sendrecvcount_xy = nnx_y * nny * nz
490	ENDIF
491
492	!
493	!-- Arrays for storing the array bounds are needed any more
494	DEALLOCATE( nxlf , nxrf , nynf , nysf )
495
496
497	!
498	!-- Collect index bounds from other PEs (to be written to restart file later)
499	ALLOCATE( hor_index_bounds(4,0:numprocs-1) )
500
501	IF ( myid == 0 ) THEN
502
503	hor_index_bounds(1,0) = nxl
504	hor_index_bounds(2,0) = nxr
505	hor_index_bounds(3,0) = nys
506	hor_index_bounds(4,0) = nyn
507
508	!
509	!-- Receive data from all other PEs
510	DO i = 1, numprocs-1
511	CALL MPI_RECV( ibuf, 4, MPI_INTEGER, i, MPI_ANY_TAG, comm2d, status, &
512	ierr )
513	hor_index_bounds(:,i) = ibuf(1:4)
514	ENDDO
515
516	ELSE
517	!
518	!-- Send index bounds to PE0
519	ibuf(1) = nxl
520	ibuf(2) = nxr
521	ibuf(3) = nys
522	ibuf(4) = nyn
523	CALL MPI_SEND( ibuf, 4, MPI_INTEGER, 0, myid, comm2d, ierr )
524
525	ENDIF
526
527
528	#if defined( __print )
529	!
530	!-- Control output
531	IF ( myid == 0 ) THEN
532	PRINT, ' processor topology *'
533	PRINT*, ' '
534	PRINT*, 'myid pcoord left right south north idx idy nxl: nxr',&
535	&' nys: nyn'
536	PRINT*, '------------------------------------------------------------',&
537	&'-----------'
538	WRITE (*,1000) 0, pcoord(1), pcoord(2), pleft, pright, psouth, pnorth, &
539	myidx, myidy, nxl, nxr, nys, nyn
540	1000 FORMAT (I4,2X,'(',I3,',',I3,')',3X,I4,2X,I4,3X,I4,2X,I4,2X,I3,1X,I3, &
541	2(2X,I4,':',I4))
542
543	!
544	!-- Receive data from the other PEs
545	DO i = 1,numprocs-1
546	CALL MPI_RECV( ibuf, 12, MPI_INTEGER, i, MPI_ANY_TAG, comm2d, status, &
547	ierr )
548	WRITE (*,1000) i, ( ibuf(j) , j = 1,12 )
549	ENDDO
550	ELSE
551
552	!
553	!-- Send data to PE0
554	ibuf(1) = pcoord(1); ibuf(2) = pcoord(2); ibuf(3) = pleft
555	ibuf(4) = pright; ibuf(5) = psouth; ibuf(6) = pnorth; ibuf(7) = myidx
556	ibuf(8) = myidy; ibuf(9) = nxl; ibuf(10) = nxr; ibuf(11) = nys
557	ibuf(12) = nyn
558	CALL MPI_SEND( ibuf, 12, MPI_INTEGER, 0, myid, comm2d, ierr )
559	ENDIF
560	#endif
561
562	#if defined( __parallel )
563	#if defined( __mpi2 )
564	!
565	!-- In case of coupled runs, get the port name on PE0 of the atmosphere model
566	!-- and pass it to PE0 of the ocean model
567	IF ( myid == 0 ) THEN
568
569	IF ( coupling_mode == 'atmosphere_to_ocean' ) THEN
570
571	CALL MPI_OPEN_PORT( MPI_INFO_NULL, port_name, ierr )
572
573	CALL MPI_PUBLISH_NAME( 'palm_coupler', MPI_INFO_NULL, port_name, &
574	ierr )
575
576	!
577	!-- Write a flag file for the ocean model and the other atmosphere
578	!-- processes.
579	!-- There seems to be a bug in MPICH2 which causes hanging processes
580	!-- in case that execution of LOOKUP_NAME is continued too early
581	!-- (i.e. before the port has been created)
582	OPEN( 90, FILE='COUPLING_PORT_OPENED', FORM='FORMATTED' )
583	WRITE ( 90, '(''TRUE'')' )
584	CLOSE ( 90 )
585
586	ELSEIF ( coupling_mode == 'ocean_to_atmosphere' ) THEN
587
588	!
589	!-- Continue only if the atmosphere model has created the port.
590	!-- There seems to be a bug in MPICH2 which causes hanging processes
591	!-- in case that execution of LOOKUP_NAME is continued too early
592	!-- (i.e. before the port has been created)
593	INQUIRE( FILE='COUPLING_PORT_OPENED', EXIST=found )
594	DO WHILE ( .NOT. found )
595	INQUIRE( FILE='COUPLING_PORT_OPENED', EXIST=found )
596	ENDDO
597
598	CALL MPI_LOOKUP_NAME( 'palm_coupler', MPI_INFO_NULL, port_name, ierr )
599
600	ENDIF
601
602	ENDIF
603
604	!
605	!-- In case of coupled runs, establish the connection between the atmosphere
606	!-- and the ocean model and define the intercommunicator (comm_inter)
607	CALL MPI_BARRIER( comm2d, ierr )
608	IF ( coupling_mode == 'atmosphere_to_ocean' ) THEN
609
610	CALL MPI_COMM_ACCEPT( port_name, MPI_INFO_NULL, 0, MPI_COMM_WORLD, &
611	comm_inter, ierr )
612	coupling_mode_remote = 'ocean_to_atmosphere'
613
614	ELSEIF ( coupling_mode == 'ocean_to_atmosphere' ) THEN
615
616	CALL MPI_COMM_CONNECT( port_name, MPI_INFO_NULL, 0, MPI_COMM_WORLD, &
617	comm_inter, ierr )
618	coupling_mode_remote = 'atmosphere_to_ocean'
619
620	ENDIF
621	#endif
622
623	!
624	!-- Determine the number of ghost point layers
625	IF ( ( scalar_advec == 'ws-scheme' .AND. .NOT. neutral ) .OR. &
626	scalar_advec == 'ws-scheme-mono' .OR. &
627	momentum_advec == 'ws-scheme' ) THEN
628	nbgp = 3
629	ELSE
630	nbgp = 1
631	ENDIF
632
633	!
634	!-- Create a new MPI derived datatype for the exchange of surface (xy) data,
635	!-- which is needed for coupled atmosphere-ocean runs.
636	!-- First, calculate number of grid points of an xy-plane.
637	ngp_xy = ( nxr - nxl + 1 + 2 * nbgp ) * ( nyn - nys + 1 + 2 * nbgp )
638	CALL MPI_TYPE_VECTOR( ngp_xy, 1, nzt-nzb+2, MPI_REAL, type_xy, ierr )
639	CALL MPI_TYPE_COMMIT( type_xy, ierr )
640
641	IF ( TRIM( coupling_mode ) /= 'uncoupled' ) THEN
642
643	!
644	!-- Pass the number of grid points of the atmosphere model to
645	!-- the ocean model and vice versa
646	IF ( coupling_mode == 'atmosphere_to_ocean' ) THEN
647
648	nx_a = nx
649	ny_a = ny
650
651	IF ( myid == 0 ) THEN
652
653	CALL MPI_SEND( nx_a, 1, MPI_INTEGER, numprocs, 1, comm_inter, &
654	ierr )
655	CALL MPI_SEND( ny_a, 1, MPI_INTEGER, numprocs, 2, comm_inter, &
656	ierr )
657	CALL MPI_SEND( pdims, 2, MPI_INTEGER, numprocs, 3, comm_inter, &
658	ierr )
659	CALL MPI_RECV( nx_o, 1, MPI_INTEGER, numprocs, 4, comm_inter, &
660	status, ierr )
661	CALL MPI_RECV( ny_o, 1, MPI_INTEGER, numprocs, 5, comm_inter, &
662	status, ierr )
663	CALL MPI_RECV( pdims_remote, 2, MPI_INTEGER, numprocs, 6, &
664	comm_inter, status, ierr )
665	ENDIF
666
667	CALL MPI_BCAST( nx_o, 1, MPI_INTEGER, 0, comm2d, ierr )
668	CALL MPI_BCAST( ny_o, 1, MPI_INTEGER, 0, comm2d, ierr )
669	CALL MPI_BCAST( pdims_remote, 2, MPI_INTEGER, 0, comm2d, ierr )
670
671	ELSEIF ( coupling_mode == 'ocean_to_atmosphere' ) THEN
672
673	nx_o = nx
674	ny_o = ny
675
676	IF ( myid == 0 ) THEN
677
678	CALL MPI_RECV( nx_a, 1, MPI_INTEGER, 0, 1, comm_inter, status, &
679	ierr )
680	CALL MPI_RECV( ny_a, 1, MPI_INTEGER, 0, 2, comm_inter, status, &
681	ierr )
682	CALL MPI_RECV( pdims_remote, 2, MPI_INTEGER, 0, 3, comm_inter, &
683	status, ierr )
684	CALL MPI_SEND( nx_o, 1, MPI_INTEGER, 0, 4, comm_inter, ierr )
685	CALL MPI_SEND( ny_o, 1, MPI_INTEGER, 0, 5, comm_inter, ierr )
686	CALL MPI_SEND( pdims, 2, MPI_INTEGER, 0, 6, comm_inter, ierr )
687	ENDIF
688
689	CALL MPI_BCAST( nx_a, 1, MPI_INTEGER, 0, comm2d, ierr)
690	CALL MPI_BCAST( ny_a, 1, MPI_INTEGER, 0, comm2d, ierr)
691	CALL MPI_BCAST( pdims_remote, 2, MPI_INTEGER, 0, comm2d, ierr)
692
693	ENDIF
694
695	ngp_a = ( nx_a+1 + 2 * nbgp ) * ( ny_a+1 + 2 * nbgp )
696	ngp_o = ( nx_o+1 + 2 * nbgp ) * ( ny_o+1 + 2 * nbgp )
697
698	!
699	!-- Determine if the horizontal grid and the number of PEs in ocean and
700	!-- atmosphere is same or not
701	IF ( nx_o == nx_a .AND. ny_o == ny_a .AND. &
702	pdims(1) == pdims_remote(1) .AND. pdims(2) == pdims_remote(2) ) &
703	THEN
704	coupling_topology = 0
705	ELSE
706	coupling_topology = 1
707	ENDIF
708
709	!
710	!-- Determine the target PEs for the exchange between ocean and
711	!-- atmosphere (comm2d)
712	IF ( coupling_topology == 0 ) THEN
713	!
714	!-- In case of identical topologies, every atmosphere PE has exactly one
715	!-- ocean PE counterpart and vice versa
716	IF ( TRIM( coupling_mode ) == 'atmosphere_to_ocean' ) THEN
717	target_id = myid + numprocs
718	ELSE
719	target_id = myid
720	ENDIF
721
722	ELSE
723	!
724	!-- In case of nonequivalent topology in ocean and atmosphere only for
725	!-- PE0 in ocean and PE0 in atmosphere a target_id is needed, since
726	!-- data echxchange between ocean and atmosphere will be done only
727	!-- between these PEs.
728	IF ( myid == 0 ) THEN
729
730	IF ( TRIM( coupling_mode ) == 'atmosphere_to_ocean' ) THEN
731	target_id = numprocs
732	ELSE
733	target_id = 0
734	ENDIF
735
736	ENDIF
737
738	ENDIF
739
740	ENDIF
741
742
743	#endif
744
745	#else
746
747	!
748	!-- Array bounds when running on a single PE (respectively a non-parallel
749	!-- machine)
750	nxl = 0
751	nxr = nx
752	nnx = nxr - nxl + 1
753	nys = 0
754	nyn = ny
755	nny = nyn - nys + 1
756	nzb = 0
757	nzt = nz
758	nnz = nz
759
760	ALLOCATE( hor_index_bounds(4,0:0) )
761	hor_index_bounds(1,0) = nxl
762	hor_index_bounds(2,0) = nxr
763	hor_index_bounds(3,0) = nys
764	hor_index_bounds(4,0) = nyn
765
766	!
767	!-- Array bounds for the pressure solver (in the parallel code, these bounds
768	!-- are the ones for the transposed arrays)
769	nys_x = nys
770	nyn_x = nyn
771	nzb_x = nzb + 1
772	nzt_x = nzt
773
774	nxl_y = nxl
775	nxr_y = nxr
776	nzb_y = nzb + 1
777	nzt_y = nzt
778
779	nxl_z = nxl
780	nxr_z = nxr
781	nys_z = nys
782	nyn_z = nyn
783
784	#endif
785
786	!
787	!-- Calculate number of grid levels necessary for the multigrid poisson solver
788	!-- as well as the gridpoint indices on each level
789	IF ( psolver(1:9) == 'multigrid' ) THEN
790
791	!
792	!-- First calculate number of possible grid levels for the subdomains
793	mg_levels_x = 1
794	mg_levels_y = 1
795	mg_levels_z = 1
796
797	i = nnx
798	DO WHILE ( MOD( i, 2 ) == 0 .AND. i /= 2 )
799	i = i / 2
800	mg_levels_x = mg_levels_x + 1
801	ENDDO
802
803	j = nny
804	DO WHILE ( MOD( j, 2 ) == 0 .AND. j /= 2 )
805	j = j / 2
806	mg_levels_y = mg_levels_y + 1
807	ENDDO
808
809	k = nz ! do not use nnz because it might be > nz due to transposition
810	! requirements
811	DO WHILE ( MOD( k, 2 ) == 0 .AND. k /= 2 )
812	k = k / 2
813	mg_levels_z = mg_levels_z + 1
814	ENDDO
815
816	maximum_grid_level = MIN( mg_levels_x, mg_levels_y, mg_levels_z )
817
818	!
819	!-- Find out, if the total domain allows more levels. These additional
820	!-- levels are identically processed on all PEs.
821	IF ( numprocs > 1 .AND. mg_switch_to_pe0_level /= -1 ) THEN
822
823	IF ( mg_levels_z > MIN( mg_levels_x, mg_levels_y ) ) THEN
824
825	mg_switch_to_pe0_level_l = maximum_grid_level
826
827	mg_levels_x = 1
828	mg_levels_y = 1
829
830	i = nx+1
831	DO WHILE ( MOD( i, 2 ) == 0 .AND. i /= 2 )
832	i = i / 2
833	mg_levels_x = mg_levels_x + 1
834	ENDDO
835
836	j = ny+1
837	DO WHILE ( MOD( j, 2 ) == 0 .AND. j /= 2 )
838	j = j / 2
839	mg_levels_y = mg_levels_y + 1
840	ENDDO
841
842	maximum_grid_level_l = MIN( mg_levels_x, mg_levels_y, mg_levels_z )
843
844	IF ( maximum_grid_level_l > mg_switch_to_pe0_level_l ) THEN
845	mg_switch_to_pe0_level_l = maximum_grid_level_l - &
846	mg_switch_to_pe0_level_l + 1
847	ELSE
848	mg_switch_to_pe0_level_l = 0
849	ENDIF
850
851	ELSE
852	mg_switch_to_pe0_level_l = 0
853	maximum_grid_level_l = maximum_grid_level
854
855	ENDIF
856
857	!
858	!-- Use switch level calculated above only if it is not pre-defined
859	!-- by user
860	IF ( mg_switch_to_pe0_level == 0 ) THEN
861	IF ( mg_switch_to_pe0_level_l /= 0 ) THEN
862	mg_switch_to_pe0_level = mg_switch_to_pe0_level_l
863	maximum_grid_level = maximum_grid_level_l
864	ENDIF
865
866	ELSE
867	!
868	!-- Check pre-defined value and reset to default, if neccessary
869	IF ( mg_switch_to_pe0_level < mg_switch_to_pe0_level_l .OR. &
870	mg_switch_to_pe0_level >= maximum_grid_level_l ) THEN
871	message_string = 'mg_switch_to_pe0_level ' // &
872	'out of range and reset to default (=0)'
873	CALL message( 'init_pegrid', 'PA0235', 0, 1, 0, 6, 0 )
874	mg_switch_to_pe0_level = 0
875	ELSE
876	!
877	!-- Use the largest number of possible levels anyway and recalculate
878	!-- the switch level to this largest number of possible values
879	maximum_grid_level = maximum_grid_level_l
880
881	ENDIF
882
883	ENDIF
884
885	ENDIF
886
887	ALLOCATE( grid_level_count(maximum_grid_level), &
888	nxl_mg(0:maximum_grid_level), nxr_mg(0:maximum_grid_level), &
889	nyn_mg(0:maximum_grid_level), nys_mg(0:maximum_grid_level), &
890	nzt_mg(0:maximum_grid_level) )
891
892	grid_level_count = 0
893	!
894	!-- Index zero required as dummy due to definition of arrays f2 and p2 in
895	!-- recursive subroutine next_mg_level
896	nxl_mg(0) = 0; nxr_mg(0) = 0; nyn_mg(0) = 0; nys_mg(0) = 0; nzt_mg(0) = 0
897
898	nxl_l = nxl; nxr_l = nxr; nys_l = nys; nyn_l = nyn; nzt_l = nzt
899
900	DO i = maximum_grid_level, 1 , -1
901
902	IF ( i == mg_switch_to_pe0_level ) THEN
903	#if defined( __parallel )
904	!
905	!-- Save the grid size of the subdomain at the switch level, because
906	!-- it is needed in poismg.
907	ind(1) = nxl_l; ind(2) = nxr_l
908	ind(3) = nys_l; ind(4) = nyn_l
909	ind(5) = nzt_l
910	ALLOCATE( ind_all(5*numprocs), mg_loc_ind(5,0:numprocs-1) )
911	CALL MPI_ALLGATHER( ind, 5, MPI_INTEGER, ind_all, 5, &
912	MPI_INTEGER, comm2d, ierr )
913	DO j = 0, numprocs-1
914	DO k = 1, 5
915	mg_loc_ind(k,j) = ind_all(k+j*5)
916	ENDDO
917	ENDDO
918	DEALLOCATE( ind_all )
919	!
920	!-- Calculate the grid size of the total domain
921	nxr_l = ( nxr_l-nxl_l+1 ) * pdims(1) - 1
922	nxl_l = 0
923	nyn_l = ( nyn_l-nys_l+1 ) * pdims(2) - 1
924	nys_l = 0
925	!
926	!-- The size of this gathered array must not be larger than the
927	!-- array tend, which is used in the multigrid scheme as a temporary
928	!-- array. Therefore the subdomain size of an PE is calculated and
929	!-- the size of the gathered grid. These values are used in
930	!-- routines pres and poismg
931	subdomain_size = ( nxr - nxl + 2 * nbgp + 1 ) * &
932	( nyn - nys + 2 * nbgp + 1 ) * ( nzt - nzb + 2 )
933	gathered_size = ( nxr_l - nxl_l + 3 ) * ( nyn_l - nys_l + 3 ) * &
934	( nzt_l - nzb + 2 )
935
936	#else
937	message_string = 'multigrid gather/scatter impossible ' // &
938	'in non parallel mode'
939	CALL message( 'init_pegrid', 'PA0237', 1, 2, 0, 6, 0 )
940	#endif
941	ENDIF
942
943	nxl_mg(i) = nxl_l
944	nxr_mg(i) = nxr_l
945	nys_mg(i) = nys_l
946	nyn_mg(i) = nyn_l
947	nzt_mg(i) = nzt_l
948
949	nxl_l = nxl_l / 2
950	nxr_l = nxr_l / 2
951	nys_l = nys_l / 2
952	nyn_l = nyn_l / 2
953	nzt_l = nzt_l / 2
954
955	ENDDO
956
957	!
958	!-- Temporary problem: Currently calculation of maxerror iin routine poismg crashes
959	!-- if grid data are collected on PE0 already on the finest grid level.
960	!-- To be solved later.
961	IF ( maximum_grid_level == mg_switch_to_pe0_level ) THEN
962	message_string = 'grid coarsening on subdomain level cannot be performed'
963	CALL message( 'poismg', 'PA0236', 1, 2, 0, 6, 0 )
964	ENDIF
965
966	ELSE
967
968	maximum_grid_level = 0
969
970	ENDIF
971
972	!
973	!-- Default level 0 tells exchange_horiz that all ghost planes have to be
974	!-- exchanged. grid_level is adjusted in poismg, where only one ghost plane
975	!-- is required.
976	grid_level = 0
977
978	#if defined( __parallel )
979	!
980	!-- Gridpoint number for the exchange of ghost points (y-line for 2D-arrays)
981	ngp_y = nyn - nys + 1 + 2 * nbgp
982
983	!
984	!-- Define new MPI derived datatypes for the exchange of ghost points in
985	!-- x- and y-direction for 2D-arrays (line)
986	CALL MPI_TYPE_VECTOR( nxr-nxl+1+2*nbgp, nbgp, ngp_y, MPI_REAL, type_x, &
987	ierr )
988	CALL MPI_TYPE_COMMIT( type_x, ierr )
989	CALL MPI_TYPE_VECTOR( nxr-nxl+1+2*nbgp, nbgp, ngp_y, MPI_INTEGER, &
990	type_x_int, ierr )
991	CALL MPI_TYPE_COMMIT( type_x_int, ierr )
992
993	CALL MPI_TYPE_VECTOR( nbgp, ngp_y, ngp_y, MPI_REAL, type_y, ierr )
994	CALL MPI_TYPE_COMMIT( type_y, ierr )
995	CALL MPI_TYPE_VECTOR( nbgp, ngp_y, ngp_y, MPI_INTEGER, type_y_int, ierr )
996	CALL MPI_TYPE_COMMIT( type_y_int, ierr )
997
998
999	!
1000	!-- Calculate gridpoint numbers for the exchange of ghost points along x
1001	!-- (yz-plane for 3D-arrays) and define MPI derived data type(s) for the
1002	!-- exchange of ghost points in y-direction (xz-plane).
1003	!-- Do these calculations for the model grid and (if necessary) also
1004	!-- for the coarser grid levels used in the multigrid method
1005	ALLOCATE ( ngp_xz(0:maximum_grid_level), ngp_yz(0:maximum_grid_level), &
1006	type_xz(0:maximum_grid_level), type_yz(0:maximum_grid_level) )
1007
1008	nxl_l = nxl; nxr_l = nxr; nys_l = nys; nyn_l = nyn; nzb_l = nzb; nzt_l = nzt
1009
1010	!
1011	!-- Discern between the model grid, which needs nbgp ghost points and
1012	!-- grid levels for the multigrid scheme. In the latter case only one
1013	!-- ghost point is necessary.
1014	!-- First definition of MPI-datatypes for exchange of ghost layers on normal
1015	!-- grid. The following loop is needed for data exchange in poismg.f90.
1016	!
1017	!-- Determine number of grid points of yz-layer for exchange
1018	ngp_yz(0) = (nzt - nzb + 2) * (nyn - nys + 1 + 2 * nbgp)
1019
1020	!
1021	!-- Define an MPI-datatype for the exchange of left/right boundaries.
1022	!-- Although data are contiguous in physical memory (which does not
1023	!-- necessarily require an MPI-derived datatype), the data exchange between
1024	!-- left and right PE's using the MPI-derived type is 10% faster than without.
1025	CALL MPI_TYPE_VECTOR( nxr-nxl+1+2nbgp, nbgp(nzt-nzb+2), ngp_yz(0), &
1026	MPI_REAL, type_xz(0), ierr )
1027	CALL MPI_TYPE_COMMIT( type_xz(0), ierr )
1028
1029	CALL MPI_TYPE_VECTOR( nbgp, ngp_yz(0), ngp_yz(0), MPI_REAL, type_yz(0), &
1030	ierr )
1031	CALL MPI_TYPE_COMMIT( type_yz(0), ierr )
1032
1033	!
1034	!-- Definition of MPI-datatypes for multigrid method (coarser level grids)
1035	IF ( psolver(1:9) == 'multigrid' ) THEN
1036	!
1037	!-- Definition of MPI-datatyoe as above, but only 1 ghost level is used
1038	DO i = maximum_grid_level, 1 , -1
1039
1040	ngp_xz(i) = (nzt_l - nzb_l + 2) * (nxr_l - nxl_l + 3)
1041	ngp_yz(i) = (nzt_l - nzb_l + 2) * (nyn_l - nys_l + 3)
1042
1043	CALL MPI_TYPE_VECTOR( nxr_l-nxl_l+3, nzt_l-nzb_l+2, ngp_yz(i), &
1044	MPI_REAL, type_xz(i), ierr )
1045	CALL MPI_TYPE_COMMIT( type_xz(i), ierr )
1046
1047	CALL MPI_TYPE_VECTOR( 1, ngp_yz(i), ngp_yz(i), MPI_REAL, type_yz(i), &
1048	ierr )
1049	CALL MPI_TYPE_COMMIT( type_yz(i), ierr )
1050
1051	nxl_l = nxl_l / 2
1052	nxr_l = nxr_l / 2
1053	nys_l = nys_l / 2
1054	nyn_l = nyn_l / 2
1055	nzt_l = nzt_l / 2
1056
1057	ENDDO
1058
1059	ENDIF
1060	!
1061	!-- Define data types for exchange of 3D Integer arrays.
1062	ngp_yz_int = (nzt - nzb + 2) * (nyn - nys + 1 + 2 * nbgp)
1063
1064	CALL MPI_TYPE_VECTOR( nxr-nxl+1+2nbgp, nbgp(nzt-nzb+2), ngp_yz_int, &
1065	MPI_INTEGER, type_xz_int, ierr )
1066	CALL MPI_TYPE_COMMIT( type_xz_int, ierr )
1067
1068	CALL MPI_TYPE_VECTOR( nbgp, ngp_yz_int, ngp_yz_int, MPI_INTEGER, type_yz_int, &
1069	ierr )
1070	CALL MPI_TYPE_COMMIT( type_yz_int, ierr )
1071
1072	#endif
1073
1074	#if defined( __parallel )
1075	!
1076	!-- Setting of flags for inflow/outflow/nesting conditions in case of non-cyclic
1077	!-- horizontal boundary conditions.
1078	IF ( pleft == MPI_PROC_NULL ) THEN
1079	IF ( bc_lr == 'dirichlet/radiation' ) THEN
1080	inflow_l = .TRUE.
1081	ELSEIF ( bc_lr == 'radiation/dirichlet' ) THEN
1082	outflow_l = .TRUE.
1083	ELSEIF ( bc_lr == 'nested' ) THEN
1084	nest_bound_l = .TRUE.
1085	ENDIF
1086	ENDIF
1087
1088	IF ( pright == MPI_PROC_NULL ) THEN
1089	IF ( bc_lr == 'dirichlet/radiation' ) THEN
1090	outflow_r = .TRUE.
1091	ELSEIF ( bc_lr == 'radiation/dirichlet' ) THEN
1092	inflow_r = .TRUE.
1093	ELSEIF ( bc_lr == 'nested' ) THEN
1094	nest_bound_r = .TRUE.
1095	ENDIF
1096	ENDIF
1097
1098	IF ( psouth == MPI_PROC_NULL ) THEN
1099	IF ( bc_ns == 'dirichlet/radiation' ) THEN
1100	outflow_s = .TRUE.
1101	ELSEIF ( bc_ns == 'radiation/dirichlet' ) THEN
1102	inflow_s = .TRUE.
1103	ELSEIF ( bc_ns == 'nested' ) THEN
1104	nest_bound_s = .TRUE.
1105	ENDIF
1106	ENDIF
1107
1108	IF ( pnorth == MPI_PROC_NULL ) THEN
1109	IF ( bc_ns == 'dirichlet/radiation' ) THEN
1110	inflow_n = .TRUE.
1111	ELSEIF ( bc_ns == 'radiation/dirichlet' ) THEN
1112	outflow_n = .TRUE.
1113	ELSEIF ( bc_ns == 'nested' ) THEN
1114	nest_bound_n = .TRUE.
1115	ENDIF
1116	ENDIF
1117
1118	!
1119	!-- Broadcast the id of the inflow PE
1120	IF ( inflow_l ) THEN
1121	id_inflow_l = myidx
1122	ELSE
1123	id_inflow_l = 0
1124	ENDIF
1125	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1126	CALL MPI_ALLREDUCE( id_inflow_l, id_inflow, 1, MPI_INTEGER, MPI_SUM, &
1127	comm1dx, ierr )
1128
1129	!
1130	!-- Broadcast the id of the recycling plane
1131	!-- WARNING: needs to be adjusted in case of inflows other than from left side!
1132	IF ( NINT( recycling_width / dx ) >= nxl .AND. &
1133	NINT( recycling_width / dx ) <= nxr ) THEN
1134	id_recycling_l = myidx
1135	ELSE
1136	id_recycling_l = 0
1137	ENDIF
1138	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1139	CALL MPI_ALLREDUCE( id_recycling_l, id_recycling, 1, MPI_INTEGER, MPI_SUM, &
1140	comm1dx, ierr )
1141
1142	CALL location_message( 'finished', .TRUE. )
1143
1144	#else
1145	IF ( bc_lr == 'dirichlet/radiation' ) THEN
1146	inflow_l = .TRUE.
1147	outflow_r = .TRUE.
1148	ELSEIF ( bc_lr == 'radiation/dirichlet' ) THEN
1149	outflow_l = .TRUE.
1150	inflow_r = .TRUE.
1151	ENDIF
1152
1153	IF ( bc_ns == 'dirichlet/radiation' ) THEN
1154	inflow_n = .TRUE.
1155	outflow_s = .TRUE.
1156	ELSEIF ( bc_ns == 'radiation/dirichlet' ) THEN
1157	outflow_n = .TRUE.
1158	inflow_s = .TRUE.
1159	ENDIF
1160	#endif
1161
1162	!
1163	!-- At the inflow or outflow, u or v, respectively, have to be calculated for
1164	!-- one more grid point.
1165	IF ( inflow_l .OR. outflow_l .OR. nest_bound_l ) THEN
1166	nxlu = nxl + 1
1167	ELSE
1168	nxlu = nxl
1169	ENDIF
1170	IF ( inflow_s .OR. outflow_s .OR. nest_bound_s ) THEN
1171	nysv = nys + 1
1172	ELSE
1173	nysv = nys
1174	ENDIF
1175
1176	!
1177	!-- Allocate wall flag arrays used in the multigrid solver
1178	IF ( psolver(1:9) == 'multigrid' ) THEN
1179
1180	DO i = maximum_grid_level, 1, -1
1181
1182	SELECT CASE ( i )
1183
1184	CASE ( 1 )
1185	ALLOCATE( wall_flags_1(nzb:nzt_mg(i)+1, &
1186	nys_mg(i)-1:nyn_mg(i)+1, &
1187	nxl_mg(i)-1:nxr_mg(i)+1) )
1188
1189	CASE ( 2 )
1190	ALLOCATE( wall_flags_2(nzb:nzt_mg(i)+1, &
1191	nys_mg(i)-1:nyn_mg(i)+1, &
1192	nxl_mg(i)-1:nxr_mg(i)+1) )
1193
1194	CASE ( 3 )
1195	ALLOCATE( wall_flags_3(nzb:nzt_mg(i)+1, &
1196	nys_mg(i)-1:nyn_mg(i)+1, &
1197	nxl_mg(i)-1:nxr_mg(i)+1) )
1198
1199	CASE ( 4 )
1200	ALLOCATE( wall_flags_4(nzb:nzt_mg(i)+1, &
1201	nys_mg(i)-1:nyn_mg(i)+1, &
1202	nxl_mg(i)-1:nxr_mg(i)+1) )
1203
1204	CASE ( 5 )
1205	ALLOCATE( wall_flags_5(nzb:nzt_mg(i)+1, &
1206	nys_mg(i)-1:nyn_mg(i)+1, &
1207	nxl_mg(i)-1:nxr_mg(i)+1) )
1208
1209	CASE ( 6 )
1210	ALLOCATE( wall_flags_6(nzb:nzt_mg(i)+1, &
1211	nys_mg(i)-1:nyn_mg(i)+1, &
1212	nxl_mg(i)-1:nxr_mg(i)+1) )
1213
1214	CASE ( 7 )
1215	ALLOCATE( wall_flags_7(nzb:nzt_mg(i)+1, &
1216	nys_mg(i)-1:nyn_mg(i)+1, &
1217	nxl_mg(i)-1:nxr_mg(i)+1) )
1218
1219	CASE ( 8 )
1220	ALLOCATE( wall_flags_8(nzb:nzt_mg(i)+1, &
1221	nys_mg(i)-1:nyn_mg(i)+1, &
1222	nxl_mg(i)-1:nxr_mg(i)+1) )
1223
1224	CASE ( 9 )
1225	ALLOCATE( wall_flags_9(nzb:nzt_mg(i)+1, &
1226	nys_mg(i)-1:nyn_mg(i)+1, &
1227	nxl_mg(i)-1:nxr_mg(i)+1) )
1228
1229	CASE ( 10 )
1230	ALLOCATE( wall_flags_10(nzb:nzt_mg(i)+1, &
1231	nys_mg(i)-1:nyn_mg(i)+1, &
1232	nxl_mg(i)-1:nxr_mg(i)+1) )
1233
1234	CASE DEFAULT
1235	message_string = 'more than 10 multigrid levels'
1236	CALL message( 'init_pegrid', 'PA0238', 1, 2, 0, 6, 0 )
1237
1238	END SELECT
1239
1240	ENDDO
1241
1242	ENDIF
1243
1244	!
1245	!-- Calculate the number of groups into which parallel I/O is split.
1246	!-- The default for files which are opened by all PEs (or where each
1247	!-- PE opens his own independent file) is, that all PEs are doing input/output
1248	!-- in parallel at the same time. This might cause performance or even more
1249	!-- severe problems depending on the configuration of the underlying file
1250	!-- system.
1251	!-- First, set the default:
1252	IF ( maximum_parallel_io_streams == -1 .OR. &
1253	maximum_parallel_io_streams > numprocs ) THEN
1254	maximum_parallel_io_streams = numprocs
1255	ENDIF
1256
1257	!
1258	!-- Now calculate the number of io_blocks and the io_group to which the
1259	!-- respective PE belongs. I/O of the groups is done in serial, but in parallel
1260	!-- for all PEs belonging to the same group. A preliminary setting with myid
1261	!-- based on MPI_COMM_WORLD has been done in parin.
1262	io_blocks = numprocs / maximum_parallel_io_streams
1263	io_group = MOD( myid+1, io_blocks )
1264
1265
1266	END SUBROUTINE init_pegrid

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |