Home

Context Navigation

source: palm/trunk/SOURCE/init_pegrid.f90 @ 1509

Last change on this file since 1509 was 1469, checked in by maronga, 10 years ago
last commit documented
Property svn:keywords set to `Id`
File size: 41.2 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 1469 2014-09-24 14:09:56Z heinze $
27	!
28	! 1468 2014-09-24 14:06:57Z maronga
29	! Adapted for use on up to 6-digit processor cores
30	!
31	! 1435 2014-07-21 10:37:02Z keck
32	! bugfix: added missing parameter coupling_mode_remote to ONLY-attribute
33	!
34	! 1402 2014-05-09 14:25:13Z raasch
35	! location messages modified
36	!
37	! 1384 2014-05-02 14:31:06Z raasch
38	! location messages added
39	!
40	! 1353 2014-04-08 15:21:23Z heinze
41	! REAL constants provided with KIND-attribute
42	!
43	! 1322 2014-03-20 16:38:49Z raasch
44	! REAL functions provided with KIND-attribute
45	!
46	! 1320 2014-03-20 08:40:49Z raasch
47	! ONLY-attribute added to USE-statements,
48	! kind-parameters added to all INTEGER and REAL declaration statements,
49	! kinds are defined in new module kinds,
50	! revision history before 2012 removed,
51	! comment fields (!:) to be used for variable explanations added to
52	! all variable declaration statements
53	!
54	! 1304 2014-03-12 10:29:42Z raasch
55	! bugfix: single core MPI runs missed some settings of transpose indices
56	!
57	! 1212 2013-08-15 08:46:27Z raasch
58	! error message for poisfft_hybrid removed
59	!
60	! 1159 2013-05-21 11:58:22Z fricke
61	! dirichlet/neumann and neumann/dirichlet removed
62	!
63	! 1139 2013-04-18 07:25:03Z raasch
64	! bugfix for calculating the id of the PE carrying the recycling plane
65	!
66	! 1111 2013-03-08 23:54:10Z raasch
67	! initialization of poisfft moved to module poisfft
68	!
69	! 1092 2013-02-02 11:24:22Z raasch
70	! unused variables removed
71	!
72	! 1056 2012-11-16 15:28:04Z raasch
73	! Indices for arrays n.._mg start from zero due to definition of arrays f2 and
74	! p2 as automatic arrays in recursive subroutine next_mg_level
75	!
76	! 1041 2012-11-06 02:36:29Z raasch
77	! a 2d virtual processor topology is used by default for all machines
78	!
79	! 1036 2012-10-22 13:43:42Z raasch
80	! code put under GPL (PALM 3.9)
81	!
82	! 1003 2012-09-14 14:35:53Z raasch
83	! subdomains must have identical size (grid matching = "match" removed)
84	!
85	! 1001 2012-09-13 14:08:46Z raasch
86	! all actions concerning upstream-spline-method removed
87	!
88	! 978 2012-08-09 08:28:32Z fricke
89	! dirichlet/neumann and neumann/dirichlet added
90	! nxlu and nysv are also calculated for inflow boundary
91	!
92	! 809 2012-01-30 13:32:58Z maronga
93	! Bugfix: replaced .AND. and .NOT. with && and ! in the preprocessor directives
94	!
95	! 807 2012-01-25 11:53:51Z maronga
96	! New cpp directive "__check" implemented which is used by check_namelist_files
97	!
98	! Revision 1.1 1997/07/24 11:15:09 raasch
99	! Initial revision
100	!
101	!
102	! Description:
103	! ------------
104	! Determination of the virtual processor topology (if not prescribed by the
105	! user)and computation of the grid point number and array bounds of the local
106	! domains.
107	!------------------------------------------------------------------------------!
108
109	USE control_parameters, &
110	ONLY: bc_lr, bc_ns, coupling_mode, coupling_mode_remote, &
111	coupling_topology, dt_dosp, gathered_size, grid_level, &
112	grid_level_count, host, inflow_l, inflow_n, inflow_r, inflow_s, &
113	io_blocks, io_group, maximum_grid_level, &
114	maximum_parallel_io_streams, message_string, &
115	mg_switch_to_pe0_level, momentum_advec, psolver, outflow_l, &
116	outflow_n, outflow_r, outflow_s, recycling_width, scalar_advec, &
117	subdomain_size
118
119	USE grid_variables, &
120	ONLY: dx
121
122	USE indices, &
123	ONLY: mg_loc_ind, nbgp, nnx, nny, nnz, nx, nx_a, nx_o, nxl, nxl_mg, &
124	nxlu, nxr, nxr_mg, ny, ny_a, ny_o, nyn, nyn_mg, nys, nys_mg, &
125	nysv, nz, nzb, nzt, nzt_mg, wall_flags_1, wall_flags_2, &
126	wall_flags_3, wall_flags_4, wall_flags_5, wall_flags_6, &
127	wall_flags_7, wall_flags_8, wall_flags_9, wall_flags_10
128
129	USE kinds
130
131	USE pegrid
132
133	USE transpose_indices, &
134	ONLY: nxl_y, nxl_yd, nxl_z, nxr_y, nxr_yd, nxr_z, nyn_x, nyn_z, nys_x,&
135	nys_z, nzb_x, nzb_y, nzb_yd, nzt_x, nzt_yd, nzt_y
136
137	IMPLICIT NONE
138
139	INTEGER(iwp) :: i !:
140	INTEGER(iwp) :: id_inflow_l !:
141	INTEGER(iwp) :: id_recycling_l !:
142	INTEGER(iwp) :: ind(5) !:
143	INTEGER(iwp) :: j !:
144	INTEGER(iwp) :: k !:
145	INTEGER(iwp) :: maximum_grid_level_l !:
146	INTEGER(iwp) :: mg_levels_x !:
147	INTEGER(iwp) :: mg_levels_y !:
148	INTEGER(iwp) :: mg_levels_z !:
149	INTEGER(iwp) :: mg_switch_to_pe0_level_l !:
150	INTEGER(iwp) :: nnx_y !:
151	INTEGER(iwp) :: nnx_z !:
152	INTEGER(iwp) :: nny_x !:
153	INTEGER(iwp) :: nny_z !:
154	INTEGER(iwp) :: nnz_x !:
155	INTEGER(iwp) :: nnz_y !:
156	INTEGER(iwp) :: numproc_sqr !:
157	INTEGER(iwp) :: nxl_l !:
158	INTEGER(iwp) :: nxr_l !:
159	INTEGER(iwp) :: nyn_l !:
160	INTEGER(iwp) :: nys_l !:
161	INTEGER(iwp) :: nzb_l !:
162	INTEGER(iwp) :: nzt_l !:
163	INTEGER(iwp) :: omp_get_num_threads !:
164
165	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: ind_all !:
166	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: nxlf !:
167	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: nxrf !:
168	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: nynf !:
169	INTEGER(iwp), DIMENSION(:), ALLOCATABLE :: nysf !:
170
171	INTEGER(iwp), DIMENSION(2) :: pdims_remote !:
172
173	#if defined( __mpi2 )
174	LOGICAL :: found !:
175	#endif
176
177	!
178	!-- Get the number of OpenMP threads
179	!$OMP PARALLEL
180	#if defined( __intel_openmp_bug )
181	threads_per_task = omp_get_num_threads()
182	#else
183	!$ threads_per_task = omp_get_num_threads()
184	#endif
185	!$OMP END PARALLEL
186
187
188	#if defined( __parallel )
189
190	CALL location_message( 'creating virtual PE grids + MPI derived data types', &
191	.FALSE. )
192	!
193	!-- Determine the processor topology or check it, if prescribed by the user
194	IF ( npex == -1 .AND. npey == -1 ) THEN
195
196	!
197	!-- Automatic determination of the topology
198	numproc_sqr = SQRT( REAL( numprocs, KIND=wp ) )
199	pdims(1) = MAX( numproc_sqr , 1 )
200	DO WHILE ( MOD( numprocs , pdims(1) ) /= 0 )
201	pdims(1) = pdims(1) - 1
202	ENDDO
203	pdims(2) = numprocs / pdims(1)
204
205	ELSEIF ( npex /= -1 .AND. npey /= -1 ) THEN
206
207	!
208	!-- Prescribed by user. Number of processors on the prescribed topology
209	!-- must be equal to the number of PEs available to the job
210	IF ( ( npex * npey ) /= numprocs ) THEN
211	WRITE( message_string, * ) 'number of PEs of the prescribed ', &
212	'topology (', npex*npey,') does not match & the number of ', &
213	'PEs available to the job (', numprocs, ')'
214	CALL message( 'init_pegrid', 'PA0221', 1, 2, 0, 6, 0 )
215	ENDIF
216	pdims(1) = npex
217	pdims(2) = npey
218
219	ELSE
220	!
221	!-- If the processor topology is prescribed by the user, the number of
222	!-- PEs must be given in both directions
223	message_string = 'if the processor topology is prescribed by the, ' // &
224	' user& both values of "npex" and "npey" must be given ' // &
225	'in the &NAMELIST-parameter file'
226	CALL message( 'init_pegrid', 'PA0222', 1, 2, 0, 6, 0 )
227
228	ENDIF
229
230	!
231	!-- For communication speedup, set barriers in front of collective
232	!-- communications by default on SGI-type systems
233	IF ( host(3:5) == 'sgi' ) collective_wait = .TRUE.
234
235	!
236	!-- If necessary, set horizontal boundary conditions to non-cyclic
237	IF ( bc_lr /= 'cyclic' ) cyclic(1) = .FALSE.
238	IF ( bc_ns /= 'cyclic' ) cyclic(2) = .FALSE.
239
240
241	#if ! defined( __check)
242	!
243	!-- Create the virtual processor grid
244	CALL MPI_CART_CREATE( comm_palm, ndim, pdims, cyclic, reorder, &
245	comm2d, ierr )
246	CALL MPI_COMM_RANK( comm2d, myid, ierr )
247	WRITE (myid_char,'(''_'',I6.6)') myid
248
249	CALL MPI_CART_COORDS( comm2d, myid, ndim, pcoord, ierr )
250	CALL MPI_CART_SHIFT( comm2d, 0, 1, pleft, pright, ierr )
251	CALL MPI_CART_SHIFT( comm2d, 1, 1, psouth, pnorth, ierr )
252
253	!
254	!-- Determine sub-topologies for transpositions
255	!-- Transposition from z to x:
256	remain_dims(1) = .TRUE.
257	remain_dims(2) = .FALSE.
258	CALL MPI_CART_SUB( comm2d, remain_dims, comm1dx, ierr )
259	CALL MPI_COMM_RANK( comm1dx, myidx, ierr )
260	!
261	!-- Transposition from x to y
262	remain_dims(1) = .FALSE.
263	remain_dims(2) = .TRUE.
264	CALL MPI_CART_SUB( comm2d, remain_dims, comm1dy, ierr )
265	CALL MPI_COMM_RANK( comm1dy, myidy, ierr )
266
267	#endif
268
269	!
270	!-- Calculate array bounds along x-direction for every PE.
271	ALLOCATE( nxlf(0:pdims(1)-1), nxrf(0:pdims(1)-1), nynf(0:pdims(2)-1), &
272	nysf(0:pdims(2)-1) )
273
274	IF ( MOD( nx+1 , pdims(1) ) /= 0 ) THEN
275	WRITE( message_string, * ) 'x-direction: gridpoint number (',nx+1,') ',&
276	'is not an& integral divisor of the number ', &
277	'processors (', pdims(1),')'
278	CALL message( 'init_pegrid', 'PA0225', 1, 2, 0, 6, 0 )
279	ELSE
280	nnx = ( nx + 1 ) / pdims(1)
281	IF ( nnx*pdims(1) - ( nx + 1) > nnx ) THEN
282	WRITE( message_string, * ) 'x-direction: nx does not match the', &
283	'requirements given by the number of PEs &used', &
284	'& please use nx = ', nx - ( pdims(1) - ( nnx*pdims(1) &
285	- ( nx + 1 ) ) ), ' instead of nx =', nx
286	CALL message( 'init_pegrid', 'PA0226', 1, 2, 0, 6, 0 )
287	ENDIF
288	ENDIF
289
290	!
291	!-- Left and right array bounds, number of gridpoints
292	DO i = 0, pdims(1)-1
293	nxlf(i) = i * nnx
294	nxrf(i) = ( i + 1 ) * nnx - 1
295	ENDDO
296
297	!
298	!-- Calculate array bounds in y-direction for every PE.
299	IF ( MOD( ny+1 , pdims(2) ) /= 0 ) THEN
300	WRITE( message_string, * ) 'y-direction: gridpoint number (',ny+1,') ', &
301	'is not an& integral divisor of the number of', &
302	'processors (', pdims(2),')'
303	CALL message( 'init_pegrid', 'PA0227', 1, 2, 0, 6, 0 )
304	ELSE
305	nny = ( ny + 1 ) / pdims(2)
306	IF ( nny*pdims(2) - ( ny + 1) > nny ) THEN
307	WRITE( message_string, * ) 'y-direction: ny does not match the', &
308	'requirements given by the number of PEs &used ', &
309	'& please use ny = ', ny - ( pdims(2) - ( nnx*pdims(2) &
310	- ( ny + 1 ) ) ), ' instead of ny =', ny
311	CALL message( 'init_pegrid', 'PA0228', 1, 2, 0, 6, 0 )
312	ENDIF
313	ENDIF
314
315	!
316	!-- South and north array bounds
317	DO j = 0, pdims(2)-1
318	nysf(j) = j * nny
319	nynf(j) = ( j + 1 ) * nny - 1
320	ENDDO
321
322	!
323	!-- Local array bounds of the respective PEs
324	nxl = nxlf(pcoord(1))
325	nxr = nxrf(pcoord(1))
326	nys = nysf(pcoord(2))
327	nyn = nynf(pcoord(2))
328	nzb = 0
329	nzt = nz
330	nnz = nz
331
332	!
333	!-- Set switches to define if the PE is situated at the border of the virtual
334	!-- processor grid
335	IF ( nxl == 0 ) left_border_pe = .TRUE.
336	IF ( nxr == nx ) right_border_pe = .TRUE.
337	IF ( nys == 0 ) south_border_pe = .TRUE.
338	IF ( nyn == ny ) north_border_pe = .TRUE.
339
340	!
341	!-- Calculate array bounds and gridpoint numbers for the transposed arrays
342	!-- (needed in the pressure solver)
343	!-- For the transposed arrays, cyclic boundaries as well as top and bottom
344	!-- boundaries are omitted, because they are obstructive to the transposition
345
346	!
347	!-- 1. transposition z --> x
348	!-- This transposition is not neccessary in case of a 1d-decomposition along x
349	nys_x = nys
350	nyn_x = nyn
351	nny_x = nny
352	nnz_x = nz / pdims(1)
353	nzb_x = 1 + myidx * nnz_x
354	nzt_x = ( myidx + 1 ) * nnz_x
355	sendrecvcount_zx = nnx * nny * nnz_x
356
357	IF ( pdims(2) /= 1 ) THEN
358	IF ( MOD( nz , pdims(1) ) /= 0 ) THEN
359	WRITE( message_string, * ) 'transposition z --> x:', &
360	'&nz=',nz,' is not an integral divisior of pdims(1)=', &
361	pdims(1)
362	CALL message( 'init_pegrid', 'PA0230', 1, 2, 0, 6, 0 )
363	ENDIF
364	ENDIF
365
366	!
367	!-- 2. transposition x --> y
368	nnz_y = nnz_x
369	nzb_y = nzb_x
370	nzt_y = nzt_x
371	IF ( MOD( nx+1 , pdims(2) ) /= 0 ) THEN
372	WRITE( message_string, * ) 'transposition x --> y:', &
373	'&nx+1=',nx+1,' is not an integral divisor of ',&
374	'pdims(2)=',pdims(2)
375	CALL message( 'init_pegrid', 'PA0231', 1, 2, 0, 6, 0 )
376	ENDIF
377	nnx_y = (nx+1) / pdims(2)
378	nxl_y = myidy * nnx_y
379	nxr_y = ( myidy + 1 ) * nnx_y - 1
380	sendrecvcount_xy = nnx_y * nny_x * nnz_y
381
382	!
383	!-- 3. transposition y --> z (ELSE: x --> y in case of 1D-decomposition
384	!-- along x)
385	nnx_z = nnx_y
386	nxl_z = nxl_y
387	nxr_z = nxr_y
388	nny_z = (ny+1) / pdims(1)
389	nys_z = myidx * nny_z
390	nyn_z = ( myidx + 1 ) * nny_z - 1
391	sendrecvcount_yz = nnx_y * nny_z * nnz_y
392
393	IF ( pdims(2) /= 1 ) THEN
394	!
395	!-- y --> z
396	!-- This transposition is not neccessary in case of a 1d-decomposition
397	!-- along x, except that the uptream-spline method is switched on
398	IF ( MOD( ny+1 , pdims(1) ) /= 0 ) THEN
399	WRITE( message_string, * ) 'transposition y --> z:', &
400	'& ny+1=',ny+1,' is not an integral divisor of ',&
401	'pdims(1)=',pdims(1)
402	CALL message( 'init_pegrid', 'PA0232', 1, 2, 0, 6, 0 )
403	ENDIF
404
405	ELSE
406	!
407	!-- x --> y. This condition must be fulfilled for a 1D-decomposition along x
408	IF ( MOD( ny+1 , pdims(1) ) /= 0 ) THEN
409	WRITE( message_string, * ) 'transposition x --> y:', &
410	'& ny+1=',ny+1,' is not an integral divisor of ',&
411	'pdims(1)=',pdims(1)
412	CALL message( 'init_pegrid', 'PA0233', 1, 2, 0, 6, 0 )
413	ENDIF
414
415	ENDIF
416
417	!
418	!-- Indices for direct transpositions z --> y (used for calculating spectra)
419	IF ( dt_dosp /= 9999999.9_wp ) THEN
420	IF ( MOD( nz, pdims(2) ) /= 0 ) THEN
421	WRITE( message_string, * ) 'direct transposition z --> y (needed ', &
422	'for spectra):& nz=',nz,' is not an integral divisor of ',&
423	'pdims(2)=',pdims(2)
424	CALL message( 'init_pegrid', 'PA0234', 1, 2, 0, 6, 0 )
425	ELSE
426	nxl_yd = nxl
427	nxr_yd = nxr
428	nzb_yd = 1 + myidy * ( nz / pdims(2) )
429	nzt_yd = ( myidy + 1 ) * ( nz / pdims(2) )
430	sendrecvcount_zyd = nnx * nny * ( nz / pdims(2) )
431	ENDIF
432	ENDIF
433
434	!
435	!-- Indices for direct transpositions y --> x (they are only possible in case
436	!-- of a 1d-decomposition along x)
437	IF ( pdims(2) == 1 ) THEN
438	nny_x = nny / pdims(1)
439	nys_x = myid * nny_x
440	nyn_x = ( myid + 1 ) * nny_x - 1
441	nzb_x = 1
442	nzt_x = nz
443	sendrecvcount_xy = nnx * nny_x * nz
444	ENDIF
445
446	!
447	!-- Indices for direct transpositions x --> y (they are only possible in case
448	!-- of a 1d-decomposition along y)
449	IF ( pdims(1) == 1 ) THEN
450	nnx_y = nnx / pdims(2)
451	nxl_y = myid * nnx_y
452	nxr_y = ( myid + 1 ) * nnx_y - 1
453	nzb_y = 1
454	nzt_y = nz
455	sendrecvcount_xy = nnx_y * nny * nz
456	ENDIF
457
458	!
459	!-- Arrays for storing the array bounds are needed any more
460	DEALLOCATE( nxlf , nxrf , nynf , nysf )
461
462
463	#if ! defined( __check)
464	!
465	!-- Collect index bounds from other PEs (to be written to restart file later)
466	ALLOCATE( hor_index_bounds(4,0:numprocs-1) )
467
468	IF ( myid == 0 ) THEN
469
470	hor_index_bounds(1,0) = nxl
471	hor_index_bounds(2,0) = nxr
472	hor_index_bounds(3,0) = nys
473	hor_index_bounds(4,0) = nyn
474
475	!
476	!-- Receive data from all other PEs
477	DO i = 1, numprocs-1
478	CALL MPI_RECV( ibuf, 4, MPI_INTEGER, i, MPI_ANY_TAG, comm2d, status, &
479	ierr )
480	hor_index_bounds(:,i) = ibuf(1:4)
481	ENDDO
482
483	ELSE
484	!
485	!-- Send index bounds to PE0
486	ibuf(1) = nxl
487	ibuf(2) = nxr
488	ibuf(3) = nys
489	ibuf(4) = nyn
490	CALL MPI_SEND( ibuf, 4, MPI_INTEGER, 0, myid, comm2d, ierr )
491
492	ENDIF
493
494	#endif
495
496	#if defined( __print )
497	!
498	!-- Control output
499	IF ( myid == 0 ) THEN
500	PRINT, ' processor topology *'
501	PRINT*, ' '
502	PRINT*, 'myid pcoord left right south north idx idy nxl: nxr',&
503	&' nys: nyn'
504	PRINT*, '------------------------------------------------------------',&
505	&'-----------'
506	WRITE (*,1000) 0, pcoord(1), pcoord(2), pleft, pright, psouth, pnorth, &
507	myidx, myidy, nxl, nxr, nys, nyn
508	1000 FORMAT (I4,2X,'(',I3,',',I3,')',3X,I4,2X,I4,3X,I4,2X,I4,2X,I3,1X,I3, &
509	2(2X,I4,':',I4))
510
511	!
512	!-- Receive data from the other PEs
513	DO i = 1,numprocs-1
514	CALL MPI_RECV( ibuf, 12, MPI_INTEGER, i, MPI_ANY_TAG, comm2d, status, &
515	ierr )
516	WRITE (*,1000) i, ( ibuf(j) , j = 1,12 )
517	ENDDO
518	ELSE
519
520	!
521	!-- Send data to PE0
522	ibuf(1) = pcoord(1); ibuf(2) = pcoord(2); ibuf(3) = pleft
523	ibuf(4) = pright; ibuf(5) = psouth; ibuf(6) = pnorth; ibuf(7) = myidx
524	ibuf(8) = myidy; ibuf(9) = nxl; ibuf(10) = nxr; ibuf(11) = nys
525	ibuf(12) = nyn
526	CALL MPI_SEND( ibuf, 12, MPI_INTEGER, 0, myid, comm2d, ierr )
527	ENDIF
528	#endif
529
530	#if defined( __parallel ) && ! defined( __check)
531	#if defined( __mpi2 )
532	!
533	!-- In case of coupled runs, get the port name on PE0 of the atmosphere model
534	!-- and pass it to PE0 of the ocean model
535	IF ( myid == 0 ) THEN
536
537	IF ( coupling_mode == 'atmosphere_to_ocean' ) THEN
538
539	CALL MPI_OPEN_PORT( MPI_INFO_NULL, port_name, ierr )
540
541	CALL MPI_PUBLISH_NAME( 'palm_coupler', MPI_INFO_NULL, port_name, &
542	ierr )
543
544	!
545	!-- Write a flag file for the ocean model and the other atmosphere
546	!-- processes.
547	!-- There seems to be a bug in MPICH2 which causes hanging processes
548	!-- in case that execution of LOOKUP_NAME is continued too early
549	!-- (i.e. before the port has been created)
550	OPEN( 90, FILE='COUPLING_PORT_OPENED', FORM='FORMATTED' )
551	WRITE ( 90, '(''TRUE'')' )
552	CLOSE ( 90 )
553
554	ELSEIF ( coupling_mode == 'ocean_to_atmosphere' ) THEN
555
556	!
557	!-- Continue only if the atmosphere model has created the port.
558	!-- There seems to be a bug in MPICH2 which causes hanging processes
559	!-- in case that execution of LOOKUP_NAME is continued too early
560	!-- (i.e. before the port has been created)
561	INQUIRE( FILE='COUPLING_PORT_OPENED', EXIST=found )
562	DO WHILE ( .NOT. found )
563	INQUIRE( FILE='COUPLING_PORT_OPENED', EXIST=found )
564	ENDDO
565
566	CALL MPI_LOOKUP_NAME( 'palm_coupler', MPI_INFO_NULL, port_name, ierr )
567
568	ENDIF
569
570	ENDIF
571
572	!
573	!-- In case of coupled runs, establish the connection between the atmosphere
574	!-- and the ocean model and define the intercommunicator (comm_inter)
575	CALL MPI_BARRIER( comm2d, ierr )
576	IF ( coupling_mode == 'atmosphere_to_ocean' ) THEN
577
578	CALL MPI_COMM_ACCEPT( port_name, MPI_INFO_NULL, 0, MPI_COMM_WORLD, &
579	comm_inter, ierr )
580	coupling_mode_remote = 'ocean_to_atmosphere'
581
582	ELSEIF ( coupling_mode == 'ocean_to_atmosphere' ) THEN
583
584	CALL MPI_COMM_CONNECT( port_name, MPI_INFO_NULL, 0, MPI_COMM_WORLD, &
585	comm_inter, ierr )
586	coupling_mode_remote = 'atmosphere_to_ocean'
587
588	ENDIF
589	#endif
590
591	!
592	!-- Determine the number of ghost point layers
593	IF ( scalar_advec == 'ws-scheme' .OR. momentum_advec == 'ws-scheme' ) THEN
594	nbgp = 3
595	ELSE
596	nbgp = 1
597	ENDIF
598
599	!
600	!-- Create a new MPI derived datatype for the exchange of surface (xy) data,
601	!-- which is needed for coupled atmosphere-ocean runs.
602	!-- First, calculate number of grid points of an xy-plane.
603	ngp_xy = ( nxr - nxl + 1 + 2 * nbgp ) * ( nyn - nys + 1 + 2 * nbgp )
604	CALL MPI_TYPE_VECTOR( ngp_xy, 1, nzt-nzb+2, MPI_REAL, type_xy, ierr )
605	CALL MPI_TYPE_COMMIT( type_xy, ierr )
606
607	IF ( TRIM( coupling_mode ) /= 'uncoupled' ) THEN
608
609	!
610	!-- Pass the number of grid points of the atmosphere model to
611	!-- the ocean model and vice versa
612	IF ( coupling_mode == 'atmosphere_to_ocean' ) THEN
613
614	nx_a = nx
615	ny_a = ny
616
617	IF ( myid == 0 ) THEN
618
619	CALL MPI_SEND( nx_a, 1, MPI_INTEGER, numprocs, 1, comm_inter, &
620	ierr )
621	CALL MPI_SEND( ny_a, 1, MPI_INTEGER, numprocs, 2, comm_inter, &
622	ierr )
623	CALL MPI_SEND( pdims, 2, MPI_INTEGER, numprocs, 3, comm_inter, &
624	ierr )
625	CALL MPI_RECV( nx_o, 1, MPI_INTEGER, numprocs, 4, comm_inter, &
626	status, ierr )
627	CALL MPI_RECV( ny_o, 1, MPI_INTEGER, numprocs, 5, comm_inter, &
628	status, ierr )
629	CALL MPI_RECV( pdims_remote, 2, MPI_INTEGER, numprocs, 6, &
630	comm_inter, status, ierr )
631	ENDIF
632
633	CALL MPI_BCAST( nx_o, 1, MPI_INTEGER, 0, comm2d, ierr )
634	CALL MPI_BCAST( ny_o, 1, MPI_INTEGER, 0, comm2d, ierr )
635	CALL MPI_BCAST( pdims_remote, 2, MPI_INTEGER, 0, comm2d, ierr )
636
637	ELSEIF ( coupling_mode == 'ocean_to_atmosphere' ) THEN
638
639	nx_o = nx
640	ny_o = ny
641
642	IF ( myid == 0 ) THEN
643
644	CALL MPI_RECV( nx_a, 1, MPI_INTEGER, 0, 1, comm_inter, status, &
645	ierr )
646	CALL MPI_RECV( ny_a, 1, MPI_INTEGER, 0, 2, comm_inter, status, &
647	ierr )
648	CALL MPI_RECV( pdims_remote, 2, MPI_INTEGER, 0, 3, comm_inter, &
649	status, ierr )
650	CALL MPI_SEND( nx_o, 1, MPI_INTEGER, 0, 4, comm_inter, ierr )
651	CALL MPI_SEND( ny_o, 1, MPI_INTEGER, 0, 5, comm_inter, ierr )
652	CALL MPI_SEND( pdims, 2, MPI_INTEGER, 0, 6, comm_inter, ierr )
653	ENDIF
654
655	CALL MPI_BCAST( nx_a, 1, MPI_INTEGER, 0, comm2d, ierr)
656	CALL MPI_BCAST( ny_a, 1, MPI_INTEGER, 0, comm2d, ierr)
657	CALL MPI_BCAST( pdims_remote, 2, MPI_INTEGER, 0, comm2d, ierr)
658
659	ENDIF
660
661	ngp_a = ( nx_a+1 + 2 * nbgp ) * ( ny_a+1 + 2 * nbgp )
662	ngp_o = ( nx_o+1 + 2 * nbgp ) * ( ny_o+1 + 2 * nbgp )
663
664	!
665	!-- Determine if the horizontal grid and the number of PEs in ocean and
666	!-- atmosphere is same or not
667	IF ( nx_o == nx_a .AND. ny_o == ny_a .AND. &
668	pdims(1) == pdims_remote(1) .AND. pdims(2) == pdims_remote(2) ) &
669	THEN
670	coupling_topology = 0
671	ELSE
672	coupling_topology = 1
673	ENDIF
674
675	!
676	!-- Determine the target PEs for the exchange between ocean and
677	!-- atmosphere (comm2d)
678	IF ( coupling_topology == 0 ) THEN
679	!
680	!-- In case of identical topologies, every atmosphere PE has exactly one
681	!-- ocean PE counterpart and vice versa
682	IF ( TRIM( coupling_mode ) == 'atmosphere_to_ocean' ) THEN
683	target_id = myid + numprocs
684	ELSE
685	target_id = myid
686	ENDIF
687
688	ELSE
689	!
690	!-- In case of nonequivalent topology in ocean and atmosphere only for
691	!-- PE0 in ocean and PE0 in atmosphere a target_id is needed, since
692	!-- data echxchange between ocean and atmosphere will be done only
693	!-- between these PEs.
694	IF ( myid == 0 ) THEN
695
696	IF ( TRIM( coupling_mode ) == 'atmosphere_to_ocean' ) THEN
697	target_id = numprocs
698	ELSE
699	target_id = 0
700	ENDIF
701
702	ENDIF
703
704	ENDIF
705
706	ENDIF
707
708
709	#endif
710
711	#else
712
713	!
714	!-- Array bounds when running on a single PE (respectively a non-parallel
715	!-- machine)
716	nxl = 0
717	nxr = nx
718	nnx = nxr - nxl + 1
719	nys = 0
720	nyn = ny
721	nny = nyn - nys + 1
722	nzb = 0
723	nzt = nz
724	nnz = nz
725
726	ALLOCATE( hor_index_bounds(4,0:0) )
727	hor_index_bounds(1,0) = nxl
728	hor_index_bounds(2,0) = nxr
729	hor_index_bounds(3,0) = nys
730	hor_index_bounds(4,0) = nyn
731
732	!
733	!-- Array bounds for the pressure solver (in the parallel code, these bounds
734	!-- are the ones for the transposed arrays)
735	nys_x = nys
736	nyn_x = nyn
737	nzb_x = nzb + 1
738	nzt_x = nzt
739
740	nxl_y = nxl
741	nxr_y = nxr
742	nzb_y = nzb + 1
743	nzt_y = nzt
744
745	nxl_z = nxl
746	nxr_z = nxr
747	nys_z = nys
748	nyn_z = nyn
749
750	#endif
751
752	!
753	!-- Calculate number of grid levels necessary for the multigrid poisson solver
754	!-- as well as the gridpoint indices on each level
755	IF ( psolver == 'multigrid' ) THEN
756
757	!
758	!-- First calculate number of possible grid levels for the subdomains
759	mg_levels_x = 1
760	mg_levels_y = 1
761	mg_levels_z = 1
762
763	i = nnx
764	DO WHILE ( MOD( i, 2 ) == 0 .AND. i /= 2 )
765	i = i / 2
766	mg_levels_x = mg_levels_x + 1
767	ENDDO
768
769	j = nny
770	DO WHILE ( MOD( j, 2 ) == 0 .AND. j /= 2 )
771	j = j / 2
772	mg_levels_y = mg_levels_y + 1
773	ENDDO
774
775	k = nz ! do not use nnz because it might be > nz due to transposition
776	! requirements
777	DO WHILE ( MOD( k, 2 ) == 0 .AND. k /= 2 )
778	k = k / 2
779	mg_levels_z = mg_levels_z + 1
780	ENDDO
781
782	maximum_grid_level = MIN( mg_levels_x, mg_levels_y, mg_levels_z )
783
784	!
785	!-- Find out, if the total domain allows more levels. These additional
786	!-- levels are identically processed on all PEs.
787	IF ( numprocs > 1 .AND. mg_switch_to_pe0_level /= -1 ) THEN
788
789	IF ( mg_levels_z > MIN( mg_levels_x, mg_levels_y ) ) THEN
790
791	mg_switch_to_pe0_level_l = maximum_grid_level
792
793	mg_levels_x = 1
794	mg_levels_y = 1
795
796	i = nx+1
797	DO WHILE ( MOD( i, 2 ) == 0 .AND. i /= 2 )
798	i = i / 2
799	mg_levels_x = mg_levels_x + 1
800	ENDDO
801
802	j = ny+1
803	DO WHILE ( MOD( j, 2 ) == 0 .AND. j /= 2 )
804	j = j / 2
805	mg_levels_y = mg_levels_y + 1
806	ENDDO
807
808	maximum_grid_level_l = MIN( mg_levels_x, mg_levels_y, mg_levels_z )
809
810	IF ( maximum_grid_level_l > mg_switch_to_pe0_level_l ) THEN
811	mg_switch_to_pe0_level_l = maximum_grid_level_l - &
812	mg_switch_to_pe0_level_l + 1
813	ELSE
814	mg_switch_to_pe0_level_l = 0
815	ENDIF
816
817	ELSE
818	mg_switch_to_pe0_level_l = 0
819	maximum_grid_level_l = maximum_grid_level
820
821	ENDIF
822
823	!
824	!-- Use switch level calculated above only if it is not pre-defined
825	!-- by user
826	IF ( mg_switch_to_pe0_level == 0 ) THEN
827	IF ( mg_switch_to_pe0_level_l /= 0 ) THEN
828	mg_switch_to_pe0_level = mg_switch_to_pe0_level_l
829	maximum_grid_level = maximum_grid_level_l
830	ENDIF
831
832	ELSE
833	!
834	!-- Check pre-defined value and reset to default, if neccessary
835	IF ( mg_switch_to_pe0_level < mg_switch_to_pe0_level_l .OR. &
836	mg_switch_to_pe0_level >= maximum_grid_level_l ) THEN
837	message_string = 'mg_switch_to_pe0_level ' // &
838	'out of range and reset to default (=0)'
839	CALL message( 'init_pegrid', 'PA0235', 0, 1, 0, 6, 0 )
840	mg_switch_to_pe0_level = 0
841	ELSE
842	!
843	!-- Use the largest number of possible levels anyway and recalculate
844	!-- the switch level to this largest number of possible values
845	maximum_grid_level = maximum_grid_level_l
846
847	ENDIF
848
849	ENDIF
850
851	ENDIF
852
853	ALLOCATE( grid_level_count(maximum_grid_level), &
854	nxl_mg(0:maximum_grid_level), nxr_mg(0:maximum_grid_level), &
855	nyn_mg(0:maximum_grid_level), nys_mg(0:maximum_grid_level), &
856	nzt_mg(0:maximum_grid_level) )
857
858	grid_level_count = 0
859	!
860	!-- Index zero required as dummy due to definition of arrays f2 and p2 in
861	!-- recursive subroutine next_mg_level
862	nxl_mg(0) = 0; nxr_mg(0) = 0; nyn_mg(0) = 0; nys_mg(0) = 0; nzt_mg(0) = 0
863
864	nxl_l = nxl; nxr_l = nxr; nys_l = nys; nyn_l = nyn; nzt_l = nzt
865
866	DO i = maximum_grid_level, 1 , -1
867
868	IF ( i == mg_switch_to_pe0_level ) THEN
869	#if defined( __parallel ) && ! defined( __check )
870	!
871	!-- Save the grid size of the subdomain at the switch level, because
872	!-- it is needed in poismg.
873	ind(1) = nxl_l; ind(2) = nxr_l
874	ind(3) = nys_l; ind(4) = nyn_l
875	ind(5) = nzt_l
876	ALLOCATE( ind_all(5*numprocs), mg_loc_ind(5,0:numprocs-1) )
877	CALL MPI_ALLGATHER( ind, 5, MPI_INTEGER, ind_all, 5, &
878	MPI_INTEGER, comm2d, ierr )
879	DO j = 0, numprocs-1
880	DO k = 1, 5
881	mg_loc_ind(k,j) = ind_all(k+j*5)
882	ENDDO
883	ENDDO
884	DEALLOCATE( ind_all )
885	!
886	!-- Calculate the grid size of the total domain
887	nxr_l = ( nxr_l-nxl_l+1 ) * pdims(1) - 1
888	nxl_l = 0
889	nyn_l = ( nyn_l-nys_l+1 ) * pdims(2) - 1
890	nys_l = 0
891	!
892	!-- The size of this gathered array must not be larger than the
893	!-- array tend, which is used in the multigrid scheme as a temporary
894	!-- array. Therefore the subdomain size of an PE is calculated and
895	!-- the size of the gathered grid. These values are used in
896	!-- routines pres and poismg
897	subdomain_size = ( nxr - nxl + 2 * nbgp + 1 ) * &
898	( nyn - nys + 2 * nbgp + 1 ) * ( nzt - nzb + 2 )
899	gathered_size = ( nxr_l - nxl_l + 3 ) * ( nyn_l - nys_l + 3 ) * &
900	( nzt_l - nzb + 2 )
901
902	#elif ! defined ( __parallel )
903	message_string = 'multigrid gather/scatter impossible ' // &
904	'in non parallel mode'
905	CALL message( 'init_pegrid', 'PA0237', 1, 2, 0, 6, 0 )
906	#endif
907	ENDIF
908
909	nxl_mg(i) = nxl_l
910	nxr_mg(i) = nxr_l
911	nys_mg(i) = nys_l
912	nyn_mg(i) = nyn_l
913	nzt_mg(i) = nzt_l
914
915	nxl_l = nxl_l / 2
916	nxr_l = nxr_l / 2
917	nys_l = nys_l / 2
918	nyn_l = nyn_l / 2
919	nzt_l = nzt_l / 2
920
921	ENDDO
922
923	!
924	!-- Temporary problem: Currently calculation of maxerror iin routine poismg crashes
925	!-- if grid data are collected on PE0 already on the finest grid level.
926	!-- To be solved later.
927	IF ( maximum_grid_level == mg_switch_to_pe0_level ) THEN
928	message_string = 'grid coarsening on subdomain level cannot be performed'
929	CALL message( 'poismg', 'PA0236', 1, 2, 0, 6, 0 )
930	ENDIF
931
932	ELSE
933
934	maximum_grid_level = 0
935
936	ENDIF
937
938	!
939	!-- Default level 0 tells exchange_horiz that all ghost planes have to be
940	!-- exchanged. grid_level is adjusted in poismg, where only one ghost plane
941	!-- is required.
942	grid_level = 0
943
944	#if defined( __parallel ) && ! defined ( __check )
945	!
946	!-- Gridpoint number for the exchange of ghost points (y-line for 2D-arrays)
947	ngp_y = nyn - nys + 1 + 2 * nbgp
948
949	!
950	!-- Define new MPI derived datatypes for the exchange of ghost points in
951	!-- x- and y-direction for 2D-arrays (line)
952	CALL MPI_TYPE_VECTOR( nxr-nxl+1+2*nbgp, nbgp, ngp_y, MPI_REAL, type_x, &
953	ierr )
954	CALL MPI_TYPE_COMMIT( type_x, ierr )
955	CALL MPI_TYPE_VECTOR( nxr-nxl+1+2*nbgp, nbgp, ngp_y, MPI_INTEGER, &
956	type_x_int, ierr )
957	CALL MPI_TYPE_COMMIT( type_x_int, ierr )
958
959	CALL MPI_TYPE_VECTOR( nbgp, ngp_y, ngp_y, MPI_REAL, type_y, ierr )
960	CALL MPI_TYPE_COMMIT( type_y, ierr )
961	CALL MPI_TYPE_VECTOR( nbgp, ngp_y, ngp_y, MPI_INTEGER, type_y_int, ierr )
962	CALL MPI_TYPE_COMMIT( type_y_int, ierr )
963
964
965	!
966	!-- Calculate gridpoint numbers for the exchange of ghost points along x
967	!-- (yz-plane for 3D-arrays) and define MPI derived data type(s) for the
968	!-- exchange of ghost points in y-direction (xz-plane).
969	!-- Do these calculations for the model grid and (if necessary) also
970	!-- for the coarser grid levels used in the multigrid method
971	ALLOCATE ( ngp_yz(0:maximum_grid_level), type_xz(0:maximum_grid_level),&
972	type_yz(0:maximum_grid_level) )
973
974	nxl_l = nxl; nxr_l = nxr; nys_l = nys; nyn_l = nyn; nzb_l = nzb; nzt_l = nzt
975
976	!
977	!-- Discern between the model grid, which needs nbgp ghost points and
978	!-- grid levels for the multigrid scheme. In the latter case only one
979	!-- ghost point is necessary.
980	!-- First definition of MPI-datatypes for exchange of ghost layers on normal
981	!-- grid. The following loop is needed for data exchange in poismg.f90.
982	!
983	!-- Determine number of grid points of yz-layer for exchange
984	ngp_yz(0) = (nzt - nzb + 2) * (nyn - nys + 1 + 2 * nbgp)
985
986	!
987	!-- Define an MPI-datatype for the exchange of left/right boundaries.
988	!-- Although data are contiguous in physical memory (which does not
989	!-- necessarily require an MPI-derived datatype), the data exchange between
990	!-- left and right PE's using the MPI-derived type is 10% faster than without.
991	CALL MPI_TYPE_VECTOR( nxr-nxl+1+2nbgp, nbgp(nzt-nzb+2), ngp_yz(0), &
992	MPI_REAL, type_xz(0), ierr )
993	CALL MPI_TYPE_COMMIT( type_xz(0), ierr )
994
995	CALL MPI_TYPE_VECTOR( nbgp, ngp_yz(0), ngp_yz(0), MPI_REAL, type_yz(0), &
996	ierr )
997	CALL MPI_TYPE_COMMIT( type_yz(0), ierr )
998
999	!
1000	!-- Definition of MPI-datatypes for multigrid method (coarser level grids)
1001	IF ( psolver == 'multigrid' ) THEN
1002	!
1003	!-- Definition of MPI-datatyoe as above, but only 1 ghost level is used
1004	DO i = maximum_grid_level, 1 , -1
1005
1006	ngp_yz(i) = (nzt_l - nzb_l + 2) * (nyn_l - nys_l + 3)
1007
1008	CALL MPI_TYPE_VECTOR( nxr_l-nxl_l+3, nzt_l-nzb_l+2, ngp_yz(i), &
1009	MPI_REAL, type_xz(i), ierr )
1010	CALL MPI_TYPE_COMMIT( type_xz(i), ierr )
1011
1012	CALL MPI_TYPE_VECTOR( 1, ngp_yz(i), ngp_yz(i), MPI_REAL, type_yz(i), &
1013	ierr )
1014	CALL MPI_TYPE_COMMIT( type_yz(i), ierr )
1015
1016	nxl_l = nxl_l / 2
1017	nxr_l = nxr_l / 2
1018	nys_l = nys_l / 2
1019	nyn_l = nyn_l / 2
1020	nzt_l = nzt_l / 2
1021
1022	ENDDO
1023
1024	ENDIF
1025	#endif
1026
1027	#if defined( __parallel ) && ! defined ( __check )
1028	!
1029	!-- Setting of flags for inflow/outflow conditions in case of non-cyclic
1030	!-- horizontal boundary conditions.
1031	IF ( pleft == MPI_PROC_NULL ) THEN
1032	IF ( bc_lr == 'dirichlet/radiation' ) THEN
1033	inflow_l = .TRUE.
1034	ELSEIF ( bc_lr == 'radiation/dirichlet' ) THEN
1035	outflow_l = .TRUE.
1036	ENDIF
1037	ENDIF
1038
1039	IF ( pright == MPI_PROC_NULL ) THEN
1040	IF ( bc_lr == 'dirichlet/radiation' ) THEN
1041	outflow_r = .TRUE.
1042	ELSEIF ( bc_lr == 'radiation/dirichlet' ) THEN
1043	inflow_r = .TRUE.
1044	ENDIF
1045	ENDIF
1046
1047	IF ( psouth == MPI_PROC_NULL ) THEN
1048	IF ( bc_ns == 'dirichlet/radiation' ) THEN
1049	outflow_s = .TRUE.
1050	ELSEIF ( bc_ns == 'radiation/dirichlet' ) THEN
1051	inflow_s = .TRUE.
1052	ENDIF
1053	ENDIF
1054
1055	IF ( pnorth == MPI_PROC_NULL ) THEN
1056	IF ( bc_ns == 'dirichlet/radiation' ) THEN
1057	inflow_n = .TRUE.
1058	ELSEIF ( bc_ns == 'radiation/dirichlet' ) THEN
1059	outflow_n = .TRUE.
1060	ENDIF
1061	ENDIF
1062
1063	!
1064	!-- Broadcast the id of the inflow PE
1065	IF ( inflow_l ) THEN
1066	id_inflow_l = myidx
1067	ELSE
1068	id_inflow_l = 0
1069	ENDIF
1070	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1071	CALL MPI_ALLREDUCE( id_inflow_l, id_inflow, 1, MPI_INTEGER, MPI_SUM, &
1072	comm1dx, ierr )
1073
1074	!
1075	!-- Broadcast the id of the recycling plane
1076	!-- WARNING: needs to be adjusted in case of inflows other than from left side!
1077	IF ( NINT( recycling_width / dx ) >= nxl .AND. &
1078	NINT( recycling_width / dx ) <= nxr ) THEN
1079	id_recycling_l = myidx
1080	ELSE
1081	id_recycling_l = 0
1082	ENDIF
1083	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1084	CALL MPI_ALLREDUCE( id_recycling_l, id_recycling, 1, MPI_INTEGER, MPI_SUM, &
1085	comm1dx, ierr )
1086
1087	CALL location_message( 'finished', .TRUE. )
1088
1089	#elif ! defined ( __parallel )
1090	IF ( bc_lr == 'dirichlet/radiation' ) THEN
1091	inflow_l = .TRUE.
1092	outflow_r = .TRUE.
1093	ELSEIF ( bc_lr == 'radiation/dirichlet' ) THEN
1094	outflow_l = .TRUE.
1095	inflow_r = .TRUE.
1096	ENDIF
1097
1098	IF ( bc_ns == 'dirichlet/radiation' ) THEN
1099	inflow_n = .TRUE.
1100	outflow_s = .TRUE.
1101	ELSEIF ( bc_ns == 'radiation/dirichlet' ) THEN
1102	outflow_n = .TRUE.
1103	inflow_s = .TRUE.
1104	ENDIF
1105	#endif
1106
1107	!
1108	!-- At the inflow or outflow, u or v, respectively, have to be calculated for
1109	!-- one more grid point.
1110	IF ( inflow_l .OR. outflow_l ) THEN
1111	nxlu = nxl + 1
1112	ELSE
1113	nxlu = nxl
1114	ENDIF
1115	IF ( inflow_s .OR. outflow_s ) THEN
1116	nysv = nys + 1
1117	ELSE
1118	nysv = nys
1119	ENDIF
1120
1121	!
1122	!-- Allocate wall flag arrays used in the multigrid solver
1123	IF ( psolver == 'multigrid' ) THEN
1124
1125	DO i = maximum_grid_level, 1, -1
1126
1127	SELECT CASE ( i )
1128
1129	CASE ( 1 )
1130	ALLOCATE( wall_flags_1(nzb:nzt_mg(i)+1, &
1131	nys_mg(i)-1:nyn_mg(i)+1, &
1132	nxl_mg(i)-1:nxr_mg(i)+1) )
1133
1134	CASE ( 2 )
1135	ALLOCATE( wall_flags_2(nzb:nzt_mg(i)+1, &
1136	nys_mg(i)-1:nyn_mg(i)+1, &
1137	nxl_mg(i)-1:nxr_mg(i)+1) )
1138
1139	CASE ( 3 )
1140	ALLOCATE( wall_flags_3(nzb:nzt_mg(i)+1, &
1141	nys_mg(i)-1:nyn_mg(i)+1, &
1142	nxl_mg(i)-1:nxr_mg(i)+1) )
1143
1144	CASE ( 4 )
1145	ALLOCATE( wall_flags_4(nzb:nzt_mg(i)+1, &
1146	nys_mg(i)-1:nyn_mg(i)+1, &
1147	nxl_mg(i)-1:nxr_mg(i)+1) )
1148
1149	CASE ( 5 )
1150	ALLOCATE( wall_flags_5(nzb:nzt_mg(i)+1, &
1151	nys_mg(i)-1:nyn_mg(i)+1, &
1152	nxl_mg(i)-1:nxr_mg(i)+1) )
1153
1154	CASE ( 6 )
1155	ALLOCATE( wall_flags_6(nzb:nzt_mg(i)+1, &
1156	nys_mg(i)-1:nyn_mg(i)+1, &
1157	nxl_mg(i)-1:nxr_mg(i)+1) )
1158
1159	CASE ( 7 )
1160	ALLOCATE( wall_flags_7(nzb:nzt_mg(i)+1, &
1161	nys_mg(i)-1:nyn_mg(i)+1, &
1162	nxl_mg(i)-1:nxr_mg(i)+1) )
1163
1164	CASE ( 8 )
1165	ALLOCATE( wall_flags_8(nzb:nzt_mg(i)+1, &
1166	nys_mg(i)-1:nyn_mg(i)+1, &
1167	nxl_mg(i)-1:nxr_mg(i)+1) )
1168
1169	CASE ( 9 )
1170	ALLOCATE( wall_flags_9(nzb:nzt_mg(i)+1, &
1171	nys_mg(i)-1:nyn_mg(i)+1, &
1172	nxl_mg(i)-1:nxr_mg(i)+1) )
1173
1174	CASE ( 10 )
1175	ALLOCATE( wall_flags_10(nzb:nzt_mg(i)+1, &
1176	nys_mg(i)-1:nyn_mg(i)+1, &
1177	nxl_mg(i)-1:nxr_mg(i)+1) )
1178
1179	CASE DEFAULT
1180	message_string = 'more than 10 multigrid levels'
1181	CALL message( 'init_pegrid', 'PA0238', 1, 2, 0, 6, 0 )
1182
1183	END SELECT
1184
1185	ENDDO
1186
1187	ENDIF
1188
1189	!
1190	!-- Calculate the number of groups into which parallel I/O is split.
1191	!-- The default for files which are opened by all PEs (or where each
1192	!-- PE opens his own independent file) is, that all PEs are doing input/output
1193	!-- in parallel at the same time. This might cause performance or even more
1194	!-- severe problems depending on the configuration of the underlying file
1195	!-- system.
1196	!-- First, set the default:
1197	IF ( maximum_parallel_io_streams == -1 .OR. &
1198	maximum_parallel_io_streams > numprocs ) THEN
1199	maximum_parallel_io_streams = numprocs
1200	ENDIF
1201
1202	!
1203	!-- Now calculate the number of io_blocks and the io_group to which the
1204	!-- respective PE belongs. I/O of the groups is done in serial, but in parallel
1205	!-- for all PEs belonging to the same group. A preliminary setting with myid
1206	!-- based on MPI_COMM_WORLD has been done in parin.
1207	io_blocks = numprocs / maximum_parallel_io_streams
1208	io_group = MOD( myid+1, io_blocks )
1209
1210
1211	END SUBROUTINE init_pegrid

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |