Home

Context Navigation

source: palm/trunk/SOURCE/init_pegrid.f90 @ 1413

Last change on this file since 1413 was 1403, checked in by raasch, 11 years ago
last commit documented
Property svn:keywords set to `Id`
File size: 40.9 KB

Line
1	SUBROUTINE init_pegrid
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: init_pegrid.f90 1403 2014-05-09 14:52:24Z boeske $
27	!
28	! 1402 2014-05-09 14:25:13Z raasch
29	! location messages modified
30	!
31	! 1384 2014-05-02 14:31:06Z raasch
32	! location messages added
33	!
34	! 1353 2014-04-08 15:21:23Z heinze
35	! REAL constants provided with KIND-attribute
36	!
37	! 1322 2014-03-20 16:38:49Z raasch
38	! REAL functions provided with KIND-attribute
39	!
40	! 1320 2014-03-20 08:40:49Z raasch
41	! ONLY-attribute added to USE-statements,
42	! kind-parameters added to all INTEGER and REAL declaration statements,
43	! kinds are defined in new module kinds,
44	! revision history before 2012 removed,
45	! comment fields (!:) to be used for variable explanations added to
46	! all variable declaration statements
47	!
48	! 1304 2014-03-12 10:29:42Z raasch
49	! bugfix: single core MPI runs missed some settings of transpose indices
50	!
51	! 1212 2013-08-15 08:46:27Z raasch
52	! error message for poisfft_hybrid removed
53	!
54	! 1159 2013-05-21 11:58:22Z fricke
55	! dirichlet/neumann and neumann/dirichlet removed
56	!
57	! 1139 2013-04-18 07:25:03Z raasch
58	! bugfix for calculating the id of the PE carrying the recycling plane
59	!
60	! 1111 2013-03-08 23:54:10Z raasch
61	! initialization of poisfft moved to module poisfft
62	!
63	! 1092 2013-02-02 11:24:22Z raasch
64	! unused variables removed
65	!
66	! 1056 2012-11-16 15:28:04Z raasch
67	! Indices for arrays n.._mg start from zero due to definition of arrays f2 and
68	! p2 as automatic arrays in recursive subroutine next_mg_level
69	!
70	! 1041 2012-11-06 02:36:29Z raasch
71	! a 2d virtual processor topology is used by default for all machines
72	!
73	! 1036 2012-10-22 13:43:42Z raasch
74	! code put under GPL (PALM 3.9)
75	!
76	! 1003 2012-09-14 14:35:53Z raasch
77	! subdomains must have identical size (grid matching = "match" removed)
78	!
79	! 1001 2012-09-13 14:08:46Z raasch
80	! all actions concerning upstream-spline-method removed
81	!
82	! 978 2012-08-09 08:28:32Z fricke
83	! dirichlet/neumann and neumann/dirichlet added
84	! nxlu and nysv are also calculated for inflow boundary
85	!
86	! 809 2012-01-30 13:32:58Z maronga
87	! Bugfix: replaced .AND. and .NOT. with && and ! in the preprocessor directives
88	!
89	! 807 2012-01-25 11:53:51Z maronga
90	! New cpp directive "__check" implemented which is used by check_namelist_files
91	!
92	! Revision 1.1 1997/07/24 11:15:09 raasch
93	! Initial revision
94	!
95	!
96	! Description:
97	! ------------
98	! Determination of the virtual processor topology (if not prescribed by the
99	! user)and computation of the grid point number and array bounds of the local
100	! domains.
101	!------------------------------------------------------------------------------!
102
103	USE control_parameters, &
104	ONLY: bc_lr, bc_ns, coupling_mode, coupling_topology, dt_dosp, &
105	gathered_size, grid_level, grid_level_count, host, inflow_l, &
106	inflow_n, inflow_r, inflow_s, io_blocks, io_group, &
107	maximum_grid_level, maximum_parallel_io_streams, message_string,&
108	mg_switch_to_pe0_level, momentum_advec, psolver, outflow_l, &
109	outflow_n, outflow_r, outflow_s, recycling_width, scalar_advec, &
110	subdomain_size
111
112	USE grid_variables, &
113	ONLY: dx
114
115	USE indices, &
116	ONLY: mg_loc_ind, nbgp, nnx, nny, nnz, nx, nx_a, nx_o, nxl, nxl_mg, &
117	nxlu, nxr, nxr_mg, ny, ny_a, ny_o, nyn, nyn_mg, nys, nys_mg, &
118	nysv, nz, nzb, nzt, nzt_mg, wall_flags_1, wall_flags_2, &
119	wall_flags_3, wall_flags_4, wall_flags_5, wall_flags_6, &
120	wall_flags_7, wall_flags_8, wall_flags_9, wall_flags_10
121
122	USE kinds
123
124	USE pegrid
125
126	USE transpose_indices, &
127	ONLY: nxl_y, nxl_yd, nxl_z, nxr_y, nxr_yd, nxr_z, nyn_x, nyn_z, nys_x,&
128	nys_z, nzb_x, nzb_y, nzb_yd, nzt_x, nzt_yd, nzt_y
129
130	IMPLICIT NONE
131
132	INTEGER(iwp) :: i !:
133	INTEGER(iwp) :: id_inflow_l !:
134	INTEGER(iwp) :: id_recycling_l !:
135	INTEGER(iwp) :: ind(5) !:
136	INTEGER(iwp) :: j !:
137	INTEGER(iwp) :: k !:
138	INTEGER(iwp) :: maximum_grid_level_l !:
139	INTEGER(iwp) :: mg_levels_x !:
140	INTEGER(iwp) :: mg_levels_y !:
141	INTEGER(iwp) :: mg_levels_z !:
142	INTEGER(iwp) :: mg_switch_to_pe0_level_l !:
143	INTEGER(iwp) :: nnx_y !:
144	INTEGER(iwp) :: nnx_z !:
145	INTEGER(iwp) :: nny_x !:
146	INTEGER(iwp) :: nny_z !:
147	INTEGER(iwp) :: nnz_x !:
148	INTEGER(iwp) :: nnz_y !:
149	INTEGER(iwp) :: numproc_sqr !:
150	INTEGER(iwp) :: nxl_l !:
151	INTEGER(iwp) :: nxr_l !:
152	INTEGER(iwp) :: nyn_l !:
153	INTEGER(iwp) :: nys_l !:
154	INTEGER(iwp) :: nzb_l !:
155	INTEGER(iwp) :: nzt_l !:
156	INTEGER(iwp) :: omp_get_num_threads !:
157
158	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: ind_all !:
159	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: nxlf !:
160	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: nxrf !:
161	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: nynf !:
162	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: nysf !:
163
164	INTEGER(iwp), DIMENSION(2) :: pdims_remote !:
165
166	#if defined( __mpi2 )
167	LOGICAL :: found !:
168	#endif
169
170	!
171	!-- Get the number of OpenMP threads
172	!$OMP PARALLEL
173	#if defined( __intel_openmp_bug )
174	threads_per_task = omp_get_num_threads()
175	#else
176	!$ threads_per_task = omp_get_num_threads()
177	#endif
178	!$OMP END PARALLEL
179
180
181	#if defined( __parallel )
182
183	CALL location_message( 'creating virtual PE grids + MPI derived data types', &
184	.FALSE. )
185	!
186	!-- Determine the processor topology or check it, if prescribed by the user
187	IF ( npex == -1 .AND. npey == -1 ) THEN
188
189	!
190	!-- Automatic determination of the topology
191	numproc_sqr = SQRT( REAL( numprocs, KIND=wp ) )
192	pdims(1) = MAX( numproc_sqr , 1 )
193	DO WHILE ( MOD( numprocs , pdims(1) ) /= 0 )
194	pdims(1) = pdims(1) - 1
195	ENDDO
196	pdims(2) = numprocs / pdims(1)
197
198	ELSEIF ( npex /= -1 .AND. npey /= -1 ) THEN
199
200	!
201	!-- Prescribed by user. Number of processors on the prescribed topology
202	!-- must be equal to the number of PEs available to the job
203	IF ( ( npex * npey ) /= numprocs ) THEN
204	WRITE( message_string, * ) 'number of PEs of the prescribed ', &
205	'topology (', npex*npey,') does not match & the number of ', &
206	'PEs available to the job (', numprocs, ')'
207	CALL message( 'init_pegrid', 'PA0221', 1, 2, 0, 6, 0 )
208	ENDIF
209	pdims(1) = npex
210	pdims(2) = npey
211
212	ELSE
213	!
214	!-- If the processor topology is prescribed by the user, the number of
215	!-- PEs must be given in both directions
216	message_string = 'if the processor topology is prescribed by the, ' // &
217	' user& both values of "npex" and "npey" must be given ' // &
218	'in the &NAMELIST-parameter file'
219	CALL message( 'init_pegrid', 'PA0222', 1, 2, 0, 6, 0 )
220
221	ENDIF
222
223	!
224	!-- For communication speedup, set barriers in front of collective
225	!-- communications by default on SGI-type systems
226	IF ( host(3:5) == 'sgi' ) collective_wait = .TRUE.
227
228	!
229	!-- If necessary, set horizontal boundary conditions to non-cyclic
230	IF ( bc_lr /= 'cyclic' ) cyclic(1) = .FALSE.
231	IF ( bc_ns /= 'cyclic' ) cyclic(2) = .FALSE.
232
233
234	#if ! defined( __check)
235	!
236	!-- Create the virtual processor grid
237	CALL MPI_CART_CREATE( comm_palm, ndim, pdims, cyclic, reorder, &
238	comm2d, ierr )
239	CALL MPI_COMM_RANK( comm2d, myid, ierr )
240	WRITE (myid_char,'(''_'',I4.4)') myid
241
242	CALL MPI_CART_COORDS( comm2d, myid, ndim, pcoord, ierr )
243	CALL MPI_CART_SHIFT( comm2d, 0, 1, pleft, pright, ierr )
244	CALL MPI_CART_SHIFT( comm2d, 1, 1, psouth, pnorth, ierr )
245
246	!
247	!-- Determine sub-topologies for transpositions
248	!-- Transposition from z to x:
249	remain_dims(1) = .TRUE.
250	remain_dims(2) = .FALSE.
251	CALL MPI_CART_SUB( comm2d, remain_dims, comm1dx, ierr )
252	CALL MPI_COMM_RANK( comm1dx, myidx, ierr )
253	!
254	!-- Transposition from x to y
255	remain_dims(1) = .FALSE.
256	remain_dims(2) = .TRUE.
257	CALL MPI_CART_SUB( comm2d, remain_dims, comm1dy, ierr )
258	CALL MPI_COMM_RANK( comm1dy, myidy, ierr )
259
260	#endif
261
262	!
263	!-- Calculate array bounds along x-direction for every PE.
264	ALLOCATE( nxlf(0:pdims(1)-1), nxrf(0:pdims(1)-1), nynf(0:pdims(2)-1), &
265	nysf(0:pdims(2)-1) )
266
267	IF ( MOD( nx+1 , pdims(1) ) /= 0 ) THEN
268	WRITE( message_string, * ) 'x-direction: gridpoint number (',nx+1,') ',&
269	'is not an& integral divisor of the number ', &
270	'processors (', pdims(1),')'
271	CALL message( 'init_pegrid', 'PA0225', 1, 2, 0, 6, 0 )
272	ELSE
273	nnx = ( nx + 1 ) / pdims(1)
274	IF ( nnx*pdims(1) - ( nx + 1) > nnx ) THEN
275	WRITE( message_string, * ) 'x-direction: nx does not match the', &
276	'requirements given by the number of PEs &used', &
277	'& please use nx = ', nx - ( pdims(1) - ( nnx*pdims(1) &
278	- ( nx + 1 ) ) ), ' instead of nx =', nx
279	CALL message( 'init_pegrid', 'PA0226', 1, 2, 0, 6, 0 )
280	ENDIF
281	ENDIF
282
283	!
284	!-- Left and right array bounds, number of gridpoints
285	DO i = 0, pdims(1)-1
286	nxlf(i) = i * nnx
287	nxrf(i) = ( i + 1 ) * nnx - 1
288	ENDDO
289
290	!
291	!-- Calculate array bounds in y-direction for every PE.
292	IF ( MOD( ny+1 , pdims(2) ) /= 0 ) THEN
293	WRITE( message_string, * ) 'y-direction: gridpoint number (',ny+1,') ', &
294	'is not an& integral divisor of the number of', &
295	'processors (', pdims(2),')'
296	CALL message( 'init_pegrid', 'PA0227', 1, 2, 0, 6, 0 )
297	ELSE
298	nny = ( ny + 1 ) / pdims(2)
299	IF ( nny*pdims(2) - ( ny + 1) > nny ) THEN
300	WRITE( message_string, * ) 'y-direction: ny does not match the', &
301	'requirements given by the number of PEs &used ', &
302	'& please use ny = ', ny - ( pdims(2) - ( nnx*pdims(2) &
303	- ( ny + 1 ) ) ), ' instead of ny =', ny
304	CALL message( 'init_pegrid', 'PA0228', 1, 2, 0, 6, 0 )
305	ENDIF
306	ENDIF
307
308	!
309	!-- South and north array bounds
310	DO j = 0, pdims(2)-1
311	nysf(j) = j * nny
312	nynf(j) = ( j + 1 ) * nny - 1
313	ENDDO
314
315	!
316	!-- Local array bounds of the respective PEs
317	nxl = nxlf(pcoord(1))
318	nxr = nxrf(pcoord(1))
319	nys = nysf(pcoord(2))
320	nyn = nynf(pcoord(2))
321	nzb = 0
322	nzt = nz
323	nnz = nz
324
325	!
326	!-- Set switches to define if the PE is situated at the border of the virtual
327	!-- processor grid
328	IF ( nxl == 0 ) left_border_pe = .TRUE.
329	IF ( nxr == nx ) right_border_pe = .TRUE.
330	IF ( nys == 0 ) south_border_pe = .TRUE.
331	IF ( nyn == ny ) north_border_pe = .TRUE.
332
333	!
334	!-- Calculate array bounds and gridpoint numbers for the transposed arrays
335	!-- (needed in the pressure solver)
336	!-- For the transposed arrays, cyclic boundaries as well as top and bottom
337	!-- boundaries are omitted, because they are obstructive to the transposition
338
339	!
340	!-- 1. transposition z --> x
341	!-- This transposition is not neccessary in case of a 1d-decomposition along x
342	nys_x = nys
343	nyn_x = nyn
344	nny_x = nny
345	nnz_x = nz / pdims(1)
346	nzb_x = 1 + myidx * nnz_x
347	nzt_x = ( myidx + 1 ) * nnz_x
348	sendrecvcount_zx = nnx * nny * nnz_x
349
350	IF ( pdims(2) /= 1 ) THEN
351	IF ( MOD( nz , pdims(1) ) /= 0 ) THEN
352	WRITE( message_string, * ) 'transposition z --> x:', &
353	'&nz=',nz,' is not an integral divisior of pdims(1)=', &
354	pdims(1)
355	CALL message( 'init_pegrid', 'PA0230', 1, 2, 0, 6, 0 )
356	ENDIF
357	ENDIF
358
359	!
360	!-- 2. transposition x --> y
361	nnz_y = nnz_x
362	nzb_y = nzb_x
363	nzt_y = nzt_x
364	IF ( MOD( nx+1 , pdims(2) ) /= 0 ) THEN
365	WRITE( message_string, * ) 'transposition x --> y:', &
366	'&nx+1=',nx+1,' is not an integral divisor of ',&
367	'pdims(2)=',pdims(2)
368	CALL message( 'init_pegrid', 'PA0231', 1, 2, 0, 6, 0 )
369	ENDIF
370	nnx_y = (nx+1) / pdims(2)
371	nxl_y = myidy * nnx_y
372	nxr_y = ( myidy + 1 ) * nnx_y - 1
373	sendrecvcount_xy = nnx_y * nny_x * nnz_y
374
375	!
376	!-- 3. transposition y --> z (ELSE: x --> y in case of 1D-decomposition
377	!-- along x)
378	nnx_z = nnx_y
379	nxl_z = nxl_y
380	nxr_z = nxr_y
381	nny_z = (ny+1) / pdims(1)
382	nys_z = myidx * nny_z
383	nyn_z = ( myidx + 1 ) * nny_z - 1
384	sendrecvcount_yz = nnx_y * nny_z * nnz_y
385
386	IF ( pdims(2) /= 1 ) THEN
387	!
388	!-- y --> z
389	!-- This transposition is not neccessary in case of a 1d-decomposition
390	!-- along x, except that the uptream-spline method is switched on
391	IF ( MOD( ny+1 , pdims(1) ) /= 0 ) THEN
392	WRITE( message_string, * ) 'transposition y --> z:', &
393	'& ny+1=',ny+1,' is not an integral divisor of ',&
394	'pdims(1)=',pdims(1)
395	CALL message( 'init_pegrid', 'PA0232', 1, 2, 0, 6, 0 )
396	ENDIF
397
398	ELSE
399	!
400	!-- x --> y. This condition must be fulfilled for a 1D-decomposition along x
401	IF ( MOD( ny+1 , pdims(1) ) /= 0 ) THEN
402	WRITE( message_string, * ) 'transposition x --> y:', &
403	'& ny+1=',ny+1,' is not an integral divisor of ',&
404	'pdims(1)=',pdims(1)
405	CALL message( 'init_pegrid', 'PA0233', 1, 2, 0, 6, 0 )
406	ENDIF
407
408	ENDIF
409
410	!
411	!-- Indices for direct transpositions z --> y (used for calculating spectra)
412	IF ( dt_dosp /= 9999999.9_wp ) THEN
413	IF ( MOD( nz, pdims(2) ) /= 0 ) THEN
414	WRITE( message_string, * ) 'direct transposition z --> y (needed ', &
415	'for spectra):& nz=',nz,' is not an integral divisor of ',&
416	'pdims(2)=',pdims(2)
417	CALL message( 'init_pegrid', 'PA0234', 1, 2, 0, 6, 0 )
418	ELSE
419	nxl_yd = nxl
420	nxr_yd = nxr
421	nzb_yd = 1 + myidy * ( nz / pdims(2) )
422	nzt_yd = ( myidy + 1 ) * ( nz / pdims(2) )
423	sendrecvcount_zyd = nnx * nny * ( nz / pdims(2) )
424	ENDIF
425	ENDIF
426
427	!
428	!-- Indices for direct transpositions y --> x (they are only possible in case
429	!-- of a 1d-decomposition along x)
430	IF ( pdims(2) == 1 ) THEN
431	nny_x = nny / pdims(1)
432	nys_x = myid * nny_x
433	nyn_x = ( myid + 1 ) * nny_x - 1
434	nzb_x = 1
435	nzt_x = nz
436	sendrecvcount_xy = nnx * nny_x * nz
437	ENDIF
438
439	!
440	!-- Indices for direct transpositions x --> y (they are only possible in case
441	!-- of a 1d-decomposition along y)
442	IF ( pdims(1) == 1 ) THEN
443	nnx_y = nnx / pdims(2)
444	nxl_y = myid * nnx_y
445	nxr_y = ( myid + 1 ) * nnx_y - 1
446	nzb_y = 1
447	nzt_y = nz
448	sendrecvcount_xy = nnx_y * nny * nz
449	ENDIF
450
451	!
452	!-- Arrays for storing the array bounds are needed any more
453	DEALLOCATE( nxlf , nxrf , nynf , nysf )
454
455
456	#if ! defined( __check)
457	!
458	!-- Collect index bounds from other PEs (to be written to restart file later)
459	ALLOCATE( hor_index_bounds(4,0:numprocs-1) )
460
461	IF ( myid == 0 ) THEN
462
463	hor_index_bounds(1,0) = nxl
464	hor_index_bounds(2,0) = nxr
465	hor_index_bounds(3,0) = nys
466	hor_index_bounds(4,0) = nyn
467
468	!
469	!-- Receive data from all other PEs
470	DO i = 1, numprocs-1
471	CALL MPI_RECV( ibuf, 4, MPI_INTEGER, i, MPI_ANY_TAG, comm2d, status, &
472	ierr )
473	hor_index_bounds(:,i) = ibuf(1:4)
474	ENDDO
475
476	ELSE
477	!
478	!-- Send index bounds to PE0
479	ibuf(1) = nxl
480	ibuf(2) = nxr
481	ibuf(3) = nys
482	ibuf(4) = nyn
483	CALL MPI_SEND( ibuf, 4, MPI_INTEGER, 0, myid, comm2d, ierr )
484
485	ENDIF
486
487	#endif
488
489	#if defined( __print )
490	!
491	!-- Control output
492	IF ( myid == 0 ) THEN
493	PRINT, ' processor topology *'
494	PRINT*, ' '
495	PRINT*, 'myid pcoord left right south north idx idy nxl: nxr',&
496	&' nys: nyn'
497	PRINT*, '------------------------------------------------------------',&
498	&'-----------'
499	WRITE (*,1000) 0, pcoord(1), pcoord(2), pleft, pright, psouth, pnorth, &
500	myidx, myidy, nxl, nxr, nys, nyn
501	1000 FORMAT (I4,2X,'(',I3,',',I3,')',3X,I4,2X,I4,3X,I4,2X,I4,2X,I3,1X,I3, &
502	2(2X,I4,':',I4))
503
504	!
505	!-- Receive data from the other PEs
506	DO i = 1,numprocs-1
507	CALL MPI_RECV( ibuf, 12, MPI_INTEGER, i, MPI_ANY_TAG, comm2d, status, &
508	ierr )
509	WRITE (*,1000) i, ( ibuf(j) , j = 1,12 )
510	ENDDO
511	ELSE
512
513	!
514	!-- Send data to PE0
515	ibuf(1) = pcoord(1); ibuf(2) = pcoord(2); ibuf(3) = pleft
516	ibuf(4) = pright; ibuf(5) = psouth; ibuf(6) = pnorth; ibuf(7) = myidx
517	ibuf(8) = myidy; ibuf(9) = nxl; ibuf(10) = nxr; ibuf(11) = nys
518	ibuf(12) = nyn
519	CALL MPI_SEND( ibuf, 12, MPI_INTEGER, 0, myid, comm2d, ierr )
520	ENDIF
521	#endif
522
523	#if defined( __parallel ) && ! defined( __check)
524	#if defined( __mpi2 )
525	!
526	!-- In case of coupled runs, get the port name on PE0 of the atmosphere model
527	!-- and pass it to PE0 of the ocean model
528	IF ( myid == 0 ) THEN
529
530	IF ( coupling_mode == 'atmosphere_to_ocean' ) THEN
531
532	CALL MPI_OPEN_PORT( MPI_INFO_NULL, port_name, ierr )
533
534	CALL MPI_PUBLISH_NAME( 'palm_coupler', MPI_INFO_NULL, port_name, &
535	ierr )
536
537	!
538	!-- Write a flag file for the ocean model and the other atmosphere
539	!-- processes.
540	!-- There seems to be a bug in MPICH2 which causes hanging processes
541	!-- in case that execution of LOOKUP_NAME is continued too early
542	!-- (i.e. before the port has been created)
543	OPEN( 90, FILE='COUPLING_PORT_OPENED', FORM='FORMATTED' )
544	WRITE ( 90, '(''TRUE'')' )
545	CLOSE ( 90 )
546
547	ELSEIF ( coupling_mode == 'ocean_to_atmosphere' ) THEN
548
549	!
550	!-- Continue only if the atmosphere model has created the port.
551	!-- There seems to be a bug in MPICH2 which causes hanging processes
552	!-- in case that execution of LOOKUP_NAME is continued too early
553	!-- (i.e. before the port has been created)
554	INQUIRE( FILE='COUPLING_PORT_OPENED', EXIST=found )
555	DO WHILE ( .NOT. found )
556	INQUIRE( FILE='COUPLING_PORT_OPENED', EXIST=found )
557	ENDDO
558
559	CALL MPI_LOOKUP_NAME( 'palm_coupler', MPI_INFO_NULL, port_name, ierr )
560
561	ENDIF
562
563	ENDIF
564
565	!
566	!-- In case of coupled runs, establish the connection between the atmosphere
567	!-- and the ocean model and define the intercommunicator (comm_inter)
568	CALL MPI_BARRIER( comm2d, ierr )
569	IF ( coupling_mode == 'atmosphere_to_ocean' ) THEN
570
571	CALL MPI_COMM_ACCEPT( port_name, MPI_INFO_NULL, 0, MPI_COMM_WORLD, &
572	comm_inter, ierr )
573	coupling_mode_remote = 'ocean_to_atmosphere'
574
575	ELSEIF ( coupling_mode == 'ocean_to_atmosphere' ) THEN
576
577	CALL MPI_COMM_CONNECT( port_name, MPI_INFO_NULL, 0, MPI_COMM_WORLD, &
578	comm_inter, ierr )
579	coupling_mode_remote = 'atmosphere_to_ocean'
580
581	ENDIF
582	#endif
583
584	!
585	!-- Determine the number of ghost point layers
586	IF ( scalar_advec == 'ws-scheme' .OR. momentum_advec == 'ws-scheme' ) THEN
587	nbgp = 3
588	ELSE
589	nbgp = 1
590	ENDIF
591
592	!
593	!-- Create a new MPI derived datatype for the exchange of surface (xy) data,
594	!-- which is needed for coupled atmosphere-ocean runs.
595	!-- First, calculate number of grid points of an xy-plane.
596	ngp_xy = ( nxr - nxl + 1 + 2 * nbgp ) * ( nyn - nys + 1 + 2 * nbgp )
597	CALL MPI_TYPE_VECTOR( ngp_xy, 1, nzt-nzb+2, MPI_REAL, type_xy, ierr )
598	CALL MPI_TYPE_COMMIT( type_xy, ierr )
599
600	IF ( TRIM( coupling_mode ) /= 'uncoupled' ) THEN
601
602	!
603	!-- Pass the number of grid points of the atmosphere model to
604	!-- the ocean model and vice versa
605	IF ( coupling_mode == 'atmosphere_to_ocean' ) THEN
606
607	nx_a = nx
608	ny_a = ny
609
610	IF ( myid == 0 ) THEN
611
612	CALL MPI_SEND( nx_a, 1, MPI_INTEGER, numprocs, 1, comm_inter, &
613	ierr )
614	CALL MPI_SEND( ny_a, 1, MPI_INTEGER, numprocs, 2, comm_inter, &
615	ierr )
616	CALL MPI_SEND( pdims, 2, MPI_INTEGER, numprocs, 3, comm_inter, &
617	ierr )
618	CALL MPI_RECV( nx_o, 1, MPI_INTEGER, numprocs, 4, comm_inter, &
619	status, ierr )
620	CALL MPI_RECV( ny_o, 1, MPI_INTEGER, numprocs, 5, comm_inter, &
621	status, ierr )
622	CALL MPI_RECV( pdims_remote, 2, MPI_INTEGER, numprocs, 6, &
623	comm_inter, status, ierr )
624	ENDIF
625
626	CALL MPI_BCAST( nx_o, 1, MPI_INTEGER, 0, comm2d, ierr )
627	CALL MPI_BCAST( ny_o, 1, MPI_INTEGER, 0, comm2d, ierr )
628	CALL MPI_BCAST( pdims_remote, 2, MPI_INTEGER, 0, comm2d, ierr )
629
630	ELSEIF ( coupling_mode == 'ocean_to_atmosphere' ) THEN
631
632	nx_o = nx
633	ny_o = ny
634
635	IF ( myid == 0 ) THEN
636
637	CALL MPI_RECV( nx_a, 1, MPI_INTEGER, 0, 1, comm_inter, status, &
638	ierr )
639	CALL MPI_RECV( ny_a, 1, MPI_INTEGER, 0, 2, comm_inter, status, &
640	ierr )
641	CALL MPI_RECV( pdims_remote, 2, MPI_INTEGER, 0, 3, comm_inter, &
642	status, ierr )
643	CALL MPI_SEND( nx_o, 1, MPI_INTEGER, 0, 4, comm_inter, ierr )
644	CALL MPI_SEND( ny_o, 1, MPI_INTEGER, 0, 5, comm_inter, ierr )
645	CALL MPI_SEND( pdims, 2, MPI_INTEGER, 0, 6, comm_inter, ierr )
646	ENDIF
647
648	CALL MPI_BCAST( nx_a, 1, MPI_INTEGER, 0, comm2d, ierr)
649	CALL MPI_BCAST( ny_a, 1, MPI_INTEGER, 0, comm2d, ierr)
650	CALL MPI_BCAST( pdims_remote, 2, MPI_INTEGER, 0, comm2d, ierr)
651
652	ENDIF
653
654	ngp_a = ( nx_a+1 + 2 * nbgp ) * ( ny_a+1 + 2 * nbgp )
655	ngp_o = ( nx_o+1 + 2 * nbgp ) * ( ny_o+1 + 2 * nbgp )
656
657	!
658	!-- Determine if the horizontal grid and the number of PEs in ocean and
659	!-- atmosphere is same or not
660	IF ( nx_o == nx_a .AND. ny_o == ny_a .AND. &
661	pdims(1) == pdims_remote(1) .AND. pdims(2) == pdims_remote(2) ) &
662	THEN
663	coupling_topology = 0
664	ELSE
665	coupling_topology = 1
666	ENDIF
667
668	!
669	!-- Determine the target PEs for the exchange between ocean and
670	!-- atmosphere (comm2d)
671	IF ( coupling_topology == 0 ) THEN
672	!
673	!-- In case of identical topologies, every atmosphere PE has exactly one
674	!-- ocean PE counterpart and vice versa
675	IF ( TRIM( coupling_mode ) == 'atmosphere_to_ocean' ) THEN
676	target_id = myid + numprocs
677	ELSE
678	target_id = myid
679	ENDIF
680
681	ELSE
682	!
683	!-- In case of nonequivalent topology in ocean and atmosphere only for
684	!-- PE0 in ocean and PE0 in atmosphere a target_id is needed, since
685	!-- data echxchange between ocean and atmosphere will be done only
686	!-- between these PEs.
687	IF ( myid == 0 ) THEN
688
689	IF ( TRIM( coupling_mode ) == 'atmosphere_to_ocean' ) THEN
690	target_id = numprocs
691	ELSE
692	target_id = 0
693	ENDIF
694
695	ENDIF
696
697	ENDIF
698
699	ENDIF
700
701
702	#endif
703
704	#else
705
706	!
707	!-- Array bounds when running on a single PE (respectively a non-parallel
708	!-- machine)
709	nxl = 0
710	nxr = nx
711	nnx = nxr - nxl + 1
712	nys = 0
713	nyn = ny
714	nny = nyn - nys + 1
715	nzb = 0
716	nzt = nz
717	nnz = nz
718
719	ALLOCATE( hor_index_bounds(4,0:0) )
720	hor_index_bounds(1,0) = nxl
721	hor_index_bounds(2,0) = nxr
722	hor_index_bounds(3,0) = nys
723	hor_index_bounds(4,0) = nyn
724
725	!
726	!-- Array bounds for the pressure solver (in the parallel code, these bounds
727	!-- are the ones for the transposed arrays)
728	nys_x = nys
729	nyn_x = nyn
730	nzb_x = nzb + 1
731	nzt_x = nzt
732
733	nxl_y = nxl
734	nxr_y = nxr
735	nzb_y = nzb + 1
736	nzt_y = nzt
737
738	nxl_z = nxl
739	nxr_z = nxr
740	nys_z = nys
741	nyn_z = nyn
742
743	#endif
744
745	!
746	!-- Calculate number of grid levels necessary for the multigrid poisson solver
747	!-- as well as the gridpoint indices on each level
748	IF ( psolver == 'multigrid' ) THEN
749
750	!
751	!-- First calculate number of possible grid levels for the subdomains
752	mg_levels_x = 1
753	mg_levels_y = 1
754	mg_levels_z = 1
755
756	i = nnx
757	DO WHILE ( MOD( i, 2 ) == 0 .AND. i /= 2 )
758	i = i / 2
759	mg_levels_x = mg_levels_x + 1
760	ENDDO
761
762	j = nny
763	DO WHILE ( MOD( j, 2 ) == 0 .AND. j /= 2 )
764	j = j / 2
765	mg_levels_y = mg_levels_y + 1
766	ENDDO
767
768	k = nz ! do not use nnz because it might be > nz due to transposition
769	! requirements
770	DO WHILE ( MOD( k, 2 ) == 0 .AND. k /= 2 )
771	k = k / 2
772	mg_levels_z = mg_levels_z + 1
773	ENDDO
774
775	maximum_grid_level = MIN( mg_levels_x, mg_levels_y, mg_levels_z )
776
777	!
778	!-- Find out, if the total domain allows more levels. These additional
779	!-- levels are identically processed on all PEs.
780	IF ( numprocs > 1 .AND. mg_switch_to_pe0_level /= -1 ) THEN
781
782	IF ( mg_levels_z > MIN( mg_levels_x, mg_levels_y ) ) THEN
783
784	mg_switch_to_pe0_level_l = maximum_grid_level
785
786	mg_levels_x = 1
787	mg_levels_y = 1
788
789	i = nx+1
790	DO WHILE ( MOD( i, 2 ) == 0 .AND. i /= 2 )
791	i = i / 2
792	mg_levels_x = mg_levels_x + 1
793	ENDDO
794
795	j = ny+1
796	DO WHILE ( MOD( j, 2 ) == 0 .AND. j /= 2 )
797	j = j / 2
798	mg_levels_y = mg_levels_y + 1
799	ENDDO
800
801	maximum_grid_level_l = MIN( mg_levels_x, mg_levels_y, mg_levels_z )
802
803	IF ( maximum_grid_level_l > mg_switch_to_pe0_level_l ) THEN
804	mg_switch_to_pe0_level_l = maximum_grid_level_l - &
805	mg_switch_to_pe0_level_l + 1
806	ELSE
807	mg_switch_to_pe0_level_l = 0
808	ENDIF
809
810	ELSE
811	mg_switch_to_pe0_level_l = 0
812	maximum_grid_level_l = maximum_grid_level
813
814	ENDIF
815
816	!
817	!-- Use switch level calculated above only if it is not pre-defined
818	!-- by user
819	IF ( mg_switch_to_pe0_level == 0 ) THEN
820	IF ( mg_switch_to_pe0_level_l /= 0 ) THEN
821	mg_switch_to_pe0_level = mg_switch_to_pe0_level_l
822	maximum_grid_level = maximum_grid_level_l
823	ENDIF
824
825	ELSE
826	!
827	!-- Check pre-defined value and reset to default, if neccessary
828	IF ( mg_switch_to_pe0_level < mg_switch_to_pe0_level_l .OR. &
829	mg_switch_to_pe0_level >= maximum_grid_level_l ) THEN
830	message_string = 'mg_switch_to_pe0_level ' // &
831	'out of range and reset to default (=0)'
832	CALL message( 'init_pegrid', 'PA0235', 0, 1, 0, 6, 0 )
833	mg_switch_to_pe0_level = 0
834	ELSE
835	!
836	!-- Use the largest number of possible levels anyway and recalculate
837	!-- the switch level to this largest number of possible values
838	maximum_grid_level = maximum_grid_level_l
839
840	ENDIF
841
842	ENDIF
843
844	ENDIF
845
846	ALLOCATE( grid_level_count(maximum_grid_level), &
847	nxl_mg(0:maximum_grid_level), nxr_mg(0:maximum_grid_level), &
848	nyn_mg(0:maximum_grid_level), nys_mg(0:maximum_grid_level), &
849	nzt_mg(0:maximum_grid_level) )
850
851	grid_level_count = 0
852	!
853	!-- Index zero required as dummy due to definition of arrays f2 and p2 in
854	!-- recursive subroutine next_mg_level
855	nxl_mg(0) = 0; nxr_mg(0) = 0; nyn_mg(0) = 0; nys_mg(0) = 0; nzt_mg(0) = 0
856
857	nxl_l = nxl; nxr_l = nxr; nys_l = nys; nyn_l = nyn; nzt_l = nzt
858
859	DO i = maximum_grid_level, 1 , -1
860
861	IF ( i == mg_switch_to_pe0_level ) THEN
862	#if defined( __parallel ) && ! defined( __check )
863	!
864	!-- Save the grid size of the subdomain at the switch level, because
865	!-- it is needed in poismg.
866	ind(1) = nxl_l; ind(2) = nxr_l
867	ind(3) = nys_l; ind(4) = nyn_l
868	ind(5) = nzt_l
869	ALLOCATE( ind_all(5*numprocs), mg_loc_ind(5,0:numprocs-1) )
870	CALL MPI_ALLGATHER( ind, 5, MPI_INTEGER, ind_all, 5, &
871	MPI_INTEGER, comm2d, ierr )
872	DO j = 0, numprocs-1
873	DO k = 1, 5
874	mg_loc_ind(k,j) = ind_all(k+j*5)
875	ENDDO
876	ENDDO
877	DEALLOCATE( ind_all )
878	!
879	!-- Calculate the grid size of the total domain
880	nxr_l = ( nxr_l-nxl_l+1 ) * pdims(1) - 1
881	nxl_l = 0
882	nyn_l = ( nyn_l-nys_l+1 ) * pdims(2) - 1
883	nys_l = 0
884	!
885	!-- The size of this gathered array must not be larger than the
886	!-- array tend, which is used in the multigrid scheme as a temporary
887	!-- array. Therefore the subdomain size of an PE is calculated and
888	!-- the size of the gathered grid. These values are used in
889	!-- routines pres and poismg
890	subdomain_size = ( nxr - nxl + 2 * nbgp + 1 ) * &
891	( nyn - nys + 2 * nbgp + 1 ) * ( nzt - nzb + 2 )
892	gathered_size = ( nxr_l - nxl_l + 3 ) * ( nyn_l - nys_l + 3 ) * &
893	( nzt_l - nzb + 2 )
894
895	#elif ! defined ( __parallel )
896	message_string = 'multigrid gather/scatter impossible ' // &
897	'in non parallel mode'
898	CALL message( 'init_pegrid', 'PA0237', 1, 2, 0, 6, 0 )
899	#endif
900	ENDIF
901
902	nxl_mg(i) = nxl_l
903	nxr_mg(i) = nxr_l
904	nys_mg(i) = nys_l
905	nyn_mg(i) = nyn_l
906	nzt_mg(i) = nzt_l
907
908	nxl_l = nxl_l / 2
909	nxr_l = nxr_l / 2
910	nys_l = nys_l / 2
911	nyn_l = nyn_l / 2
912	nzt_l = nzt_l / 2
913
914	ENDDO
915
916	!
917	!-- Temporary problem: Currently calculation of maxerror iin routine poismg crashes
918	!-- if grid data are collected on PE0 already on the finest grid level.
919	!-- To be solved later.
920	IF ( maximum_grid_level == mg_switch_to_pe0_level ) THEN
921	message_string = 'grid coarsening on subdomain level cannot be performed'
922	CALL message( 'poismg', 'PA0236', 1, 2, 0, 6, 0 )
923	ENDIF
924
925	ELSE
926
927	maximum_grid_level = 0
928
929	ENDIF
930
931	!
932	!-- Default level 0 tells exchange_horiz that all ghost planes have to be
933	!-- exchanged. grid_level is adjusted in poismg, where only one ghost plane
934	!-- is required.
935	grid_level = 0
936
937	#if defined( __parallel ) && ! defined ( __check )
938	!
939	!-- Gridpoint number for the exchange of ghost points (y-line for 2D-arrays)
940	ngp_y = nyn - nys + 1 + 2 * nbgp
941
942	!
943	!-- Define new MPI derived datatypes for the exchange of ghost points in
944	!-- x- and y-direction for 2D-arrays (line)
945	CALL MPI_TYPE_VECTOR( nxr-nxl+1+2*nbgp, nbgp, ngp_y, MPI_REAL, type_x, &
946	ierr )
947	CALL MPI_TYPE_COMMIT( type_x, ierr )
948	CALL MPI_TYPE_VECTOR( nxr-nxl+1+2*nbgp, nbgp, ngp_y, MPI_INTEGER, &
949	type_x_int, ierr )
950	CALL MPI_TYPE_COMMIT( type_x_int, ierr )
951
952	CALL MPI_TYPE_VECTOR( nbgp, ngp_y, ngp_y, MPI_REAL, type_y, ierr )
953	CALL MPI_TYPE_COMMIT( type_y, ierr )
954	CALL MPI_TYPE_VECTOR( nbgp, ngp_y, ngp_y, MPI_INTEGER, type_y_int, ierr )
955	CALL MPI_TYPE_COMMIT( type_y_int, ierr )
956
957
958	!
959	!-- Calculate gridpoint numbers for the exchange of ghost points along x
960	!-- (yz-plane for 3D-arrays) and define MPI derived data type(s) for the
961	!-- exchange of ghost points in y-direction (xz-plane).
962	!-- Do these calculations for the model grid and (if necessary) also
963	!-- for the coarser grid levels used in the multigrid method
964	ALLOCATE ( ngp_yz(0:maximum_grid_level), type_xz(0:maximum_grid_level),&
965	type_yz(0:maximum_grid_level) )
966
967	nxl_l = nxl; nxr_l = nxr; nys_l = nys; nyn_l = nyn; nzb_l = nzb; nzt_l = nzt
968
969	!
970	!-- Discern between the model grid, which needs nbgp ghost points and
971	!-- grid levels for the multigrid scheme. In the latter case only one
972	!-- ghost point is necessary.
973	!-- First definition of MPI-datatypes for exchange of ghost layers on normal
974	!-- grid. The following loop is needed for data exchange in poismg.f90.
975	!
976	!-- Determine number of grid points of yz-layer for exchange
977	ngp_yz(0) = (nzt - nzb + 2) * (nyn - nys + 1 + 2 * nbgp)
978
979	!
980	!-- Define an MPI-datatype for the exchange of left/right boundaries.
981	!-- Although data are contiguous in physical memory (which does not
982	!-- necessarily require an MPI-derived datatype), the data exchange between
983	!-- left and right PE's using the MPI-derived type is 10% faster than without.
984	CALL MPI_TYPE_VECTOR( nxr-nxl+1+2nbgp, nbgp(nzt-nzb+2), ngp_yz(0), &
985	MPI_REAL, type_xz(0), ierr )
986	CALL MPI_TYPE_COMMIT( type_xz(0), ierr )
987
988	CALL MPI_TYPE_VECTOR( nbgp, ngp_yz(0), ngp_yz(0), MPI_REAL, type_yz(0), &
989	ierr )
990	CALL MPI_TYPE_COMMIT( type_yz(0), ierr )
991
992	!
993	!-- Definition of MPI-datatypes for multigrid method (coarser level grids)
994	IF ( psolver == 'multigrid' ) THEN
995	!
996	!-- Definition of MPI-datatyoe as above, but only 1 ghost level is used
997	DO i = maximum_grid_level, 1 , -1
998
999	ngp_yz(i) = (nzt_l - nzb_l + 2) * (nyn_l - nys_l + 3)
1000
1001	CALL MPI_TYPE_VECTOR( nxr_l-nxl_l+3, nzt_l-nzb_l+2, ngp_yz(i), &
1002	MPI_REAL, type_xz(i), ierr )
1003	CALL MPI_TYPE_COMMIT( type_xz(i), ierr )
1004
1005	CALL MPI_TYPE_VECTOR( 1, ngp_yz(i), ngp_yz(i), MPI_REAL, type_yz(i), &
1006	ierr )
1007	CALL MPI_TYPE_COMMIT( type_yz(i), ierr )
1008
1009	nxl_l = nxl_l / 2
1010	nxr_l = nxr_l / 2
1011	nys_l = nys_l / 2
1012	nyn_l = nyn_l / 2
1013	nzt_l = nzt_l / 2
1014
1015	ENDDO
1016
1017	ENDIF
1018	#endif
1019
1020	#if defined( __parallel ) && ! defined ( __check )
1021	!
1022	!-- Setting of flags for inflow/outflow conditions in case of non-cyclic
1023	!-- horizontal boundary conditions.
1024	IF ( pleft == MPI_PROC_NULL ) THEN
1025	IF ( bc_lr == 'dirichlet/radiation' ) THEN
1026	inflow_l = .TRUE.
1027	ELSEIF ( bc_lr == 'radiation/dirichlet' ) THEN
1028	outflow_l = .TRUE.
1029	ENDIF
1030	ENDIF
1031
1032	IF ( pright == MPI_PROC_NULL ) THEN
1033	IF ( bc_lr == 'dirichlet/radiation' ) THEN
1034	outflow_r = .TRUE.
1035	ELSEIF ( bc_lr == 'radiation/dirichlet' ) THEN
1036	inflow_r = .TRUE.
1037	ENDIF
1038	ENDIF
1039
1040	IF ( psouth == MPI_PROC_NULL ) THEN
1041	IF ( bc_ns == 'dirichlet/radiation' ) THEN
1042	outflow_s = .TRUE.
1043	ELSEIF ( bc_ns == 'radiation/dirichlet' ) THEN
1044	inflow_s = .TRUE.
1045	ENDIF
1046	ENDIF
1047
1048	IF ( pnorth == MPI_PROC_NULL ) THEN
1049	IF ( bc_ns == 'dirichlet/radiation' ) THEN
1050	inflow_n = .TRUE.
1051	ELSEIF ( bc_ns == 'radiation/dirichlet' ) THEN
1052	outflow_n = .TRUE.
1053	ENDIF
1054	ENDIF
1055
1056	!
1057	!-- Broadcast the id of the inflow PE
1058	IF ( inflow_l ) THEN
1059	id_inflow_l = myidx
1060	ELSE
1061	id_inflow_l = 0
1062	ENDIF
1063	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1064	CALL MPI_ALLREDUCE( id_inflow_l, id_inflow, 1, MPI_INTEGER, MPI_SUM, &
1065	comm1dx, ierr )
1066
1067	!
1068	!-- Broadcast the id of the recycling plane
1069	!-- WARNING: needs to be adjusted in case of inflows other than from left side!
1070	IF ( NINT( recycling_width / dx ) >= nxl .AND. &
1071	NINT( recycling_width / dx ) <= nxr ) THEN
1072	id_recycling_l = myidx
1073	ELSE
1074	id_recycling_l = 0
1075	ENDIF
1076	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1077	CALL MPI_ALLREDUCE( id_recycling_l, id_recycling, 1, MPI_INTEGER, MPI_SUM, &
1078	comm1dx, ierr )
1079
1080	CALL location_message( 'finished', .TRUE. )
1081
1082	#elif ! defined ( __parallel )
1083	IF ( bc_lr == 'dirichlet/radiation' ) THEN
1084	inflow_l = .TRUE.
1085	outflow_r = .TRUE.
1086	ELSEIF ( bc_lr == 'radiation/dirichlet' ) THEN
1087	outflow_l = .TRUE.
1088	inflow_r = .TRUE.
1089	ENDIF
1090
1091	IF ( bc_ns == 'dirichlet/radiation' ) THEN
1092	inflow_n = .TRUE.
1093	outflow_s = .TRUE.
1094	ELSEIF ( bc_ns == 'radiation/dirichlet' ) THEN
1095	outflow_n = .TRUE.
1096	inflow_s = .TRUE.
1097	ENDIF
1098	#endif
1099
1100	!
1101	!-- At the inflow or outflow, u or v, respectively, have to be calculated for
1102	!-- one more grid point.
1103	IF ( inflow_l .OR. outflow_l ) THEN
1104	nxlu = nxl + 1
1105	ELSE
1106	nxlu = nxl
1107	ENDIF
1108	IF ( inflow_s .OR. outflow_s ) THEN
1109	nysv = nys + 1
1110	ELSE
1111	nysv = nys
1112	ENDIF
1113
1114	!
1115	!-- Allocate wall flag arrays used in the multigrid solver
1116	IF ( psolver == 'multigrid' ) THEN
1117
1118	DO i = maximum_grid_level, 1, -1
1119
1120	SELECT CASE ( i )
1121
1122	CASE ( 1 )
1123	ALLOCATE( wall_flags_1(nzb:nzt_mg(i)+1, &
1124	nys_mg(i)-1:nyn_mg(i)+1, &
1125	nxl_mg(i)-1:nxr_mg(i)+1) )
1126
1127	CASE ( 2 )
1128	ALLOCATE( wall_flags_2(nzb:nzt_mg(i)+1, &
1129	nys_mg(i)-1:nyn_mg(i)+1, &
1130	nxl_mg(i)-1:nxr_mg(i)+1) )
1131
1132	CASE ( 3 )
1133	ALLOCATE( wall_flags_3(nzb:nzt_mg(i)+1, &
1134	nys_mg(i)-1:nyn_mg(i)+1, &
1135	nxl_mg(i)-1:nxr_mg(i)+1) )
1136
1137	CASE ( 4 )
1138	ALLOCATE( wall_flags_4(nzb:nzt_mg(i)+1, &
1139	nys_mg(i)-1:nyn_mg(i)+1, &
1140	nxl_mg(i)-1:nxr_mg(i)+1) )
1141
1142	CASE ( 5 )
1143	ALLOCATE( wall_flags_5(nzb:nzt_mg(i)+1, &
1144	nys_mg(i)-1:nyn_mg(i)+1, &
1145	nxl_mg(i)-1:nxr_mg(i)+1) )
1146
1147	CASE ( 6 )
1148	ALLOCATE( wall_flags_6(nzb:nzt_mg(i)+1, &
1149	nys_mg(i)-1:nyn_mg(i)+1, &
1150	nxl_mg(i)-1:nxr_mg(i)+1) )
1151
1152	CASE ( 7 )
1153	ALLOCATE( wall_flags_7(nzb:nzt_mg(i)+1, &
1154	nys_mg(i)-1:nyn_mg(i)+1, &
1155	nxl_mg(i)-1:nxr_mg(i)+1) )
1156
1157	CASE ( 8 )
1158	ALLOCATE( wall_flags_8(nzb:nzt_mg(i)+1, &
1159	nys_mg(i)-1:nyn_mg(i)+1, &
1160	nxl_mg(i)-1:nxr_mg(i)+1) )
1161
1162	CASE ( 9 )
1163	ALLOCATE( wall_flags_9(nzb:nzt_mg(i)+1, &
1164	nys_mg(i)-1:nyn_mg(i)+1, &
1165	nxl_mg(i)-1:nxr_mg(i)+1) )
1166
1167	CASE ( 10 )
1168	ALLOCATE( wall_flags_10(nzb:nzt_mg(i)+1, &
1169	nys_mg(i)-1:nyn_mg(i)+1, &
1170	nxl_mg(i)-1:nxr_mg(i)+1) )
1171
1172	CASE DEFAULT
1173	message_string = 'more than 10 multigrid levels'
1174	CALL message( 'init_pegrid', 'PA0238', 1, 2, 0, 6, 0 )
1175
1176	END SELECT
1177
1178	ENDDO
1179
1180	ENDIF
1181
1182	!
1183	!-- Calculate the number of groups into which parallel I/O is split.
1184	!-- The default for files which are opened by all PEs (or where each
1185	!-- PE opens his own independent file) is, that all PEs are doing input/output
1186	!-- in parallel at the same time. This might cause performance or even more
1187	!-- severe problems depending on the configuration of the underlying file
1188	!-- system.
1189	!-- First, set the default:
1190	IF ( maximum_parallel_io_streams == -1 .OR. &
1191	maximum_parallel_io_streams > numprocs ) THEN
1192	maximum_parallel_io_streams = numprocs
1193	ENDIF
1194
1195	!
1196	!-- Now calculate the number of io_blocks and the io_group to which the
1197	!-- respective PE belongs. I/O of the groups is done in serial, but in parallel
1198	!-- for all PEs belonging to the same group. A preliminary setting with myid
1199	!-- based on MPI_COMM_WORLD has been done in parin.
1200	io_blocks = numprocs / maximum_parallel_io_streams
1201	io_group = MOD( myid+1, io_blocks )
1202
1203
1204	END SUBROUTINE init_pegrid

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |