source: palm/trunk/SOURCE/init_pegrid.f90 @ 3897

Last change on this file since 3897 was 3897, checked in by suehring, 2 years ago

2D output of emission fluxes in chemistry; revise check for multigrid coarsening and give a warning instead of an error

  • Property svn:keywords set to Id
File size: 55.5 KB
Line 
1!> @file init_pegrid.f90
2!------------------------------------------------------------------------------!
3! This file is part of the PALM model system.
4!
5! PALM is free software: you can redistribute it and/or modify it under the
6! terms of the GNU General Public License as published by the Free Software
7! Foundation, either version 3 of the License, or (at your option) any later
8! 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 1997-2019 Leibniz Universitaet Hannover
18!------------------------------------------------------------------------------!
19!
20! Current revisions:
21! ------------------
22!
23!
24! Former revisions:
25! -----------------
26! $Id: init_pegrid.f90 3897 2019-04-15 11:51:14Z suehring $
27! Minor revision of multigrid check; give warning instead of an abort.
28!
29! 3890 2019-04-12 15:59:20Z suehring
30! Check if grid coarsening is possible on subdomain, in order to avoid that
31! multigrid approach effectively reduces to a Gauss-Seidel scheme.
32!
33! 3885 2019-04-11 11:29:34Z kanani
34! Changes related to global restructuring of location messages and introduction
35! of additional debug messages
36!
37! 3884 2019-04-10 13:31:55Z Giersch
38! id_recycling is only calculated in case of tubulent inflow
39!
40! 3761 2019-02-25 15:31:42Z raasch
41! unused variable removed
42!
43! 3655 2019-01-07 16:51:22Z knoop
44! variables documented
45!
46! 3552 2018-11-22 10:28:35Z suehring
47! Introduce new MPI-datatype for ghostpoint exchange of 2D 8-bit Integer arrays
48!
49! 3542 2018-11-20 17:04:13Z suehring
50! Bugfix in setting number of ghost layers in neutral case
51!
52! 3341 2018-10-15 10:31:27Z suehring
53! unused variables removed
54!
55! 3183 2018-07-27 14:25:55Z suehring
56! Rename variables and boundary conditions in mesoscale-offline nesting mode
57!
58! 3182 2018-07-27 13:36:03Z suehring
59! bugfix: wrong error number in r3057 revised
60!
61! 3057 2018-06-05 09:03:41Z raasch
62! bugfix: check that nz is even in case that optimized multigrid is used
63!
64! 3049 2018-05-29 13:52:36Z Giersch
65! Error messages revised
66!
67! 3045 2018-05-28 07:55:41Z Giersch
68! Error messages revised
69!
70! 2938 2018-03-27 15:52:42Z suehring
71! - No checks for domain decomposition in case of turbulence generator
72!  (is done in stg module)
73! - Introduce ids to indicate lateral processors for turbulence generator
74!
75! 2936 2018-03-27 14:49:27Z suehring
76! Variable use_synthetic_turbulence_generator has been abbreviated
77!
78! 2718 2018-01-02 08:49:38Z maronga
79! Corrected "Former revisions" section
80!
81! 2696 2017-12-14 17:12:51Z kanani
82! Change in file header (GPL part)
83! 3D-Integer exchange on multigrid level (MS)
84! Forcing implemented (MS)
85!
86! 2600 2017-11-01 14:11:20Z raasch
87! calculation of block-I/O quantitites removed (is now done in parin)
88!
89! 2516 2017-10-04 11:03:04Z suehring
90! Remove tabs
91!
92! 2514 2017-10-04 09:52:37Z suehring
93! Redundant preprocessor directives removed
94!
95! 2372 2017-08-25 12:37:32Z sward
96! Shifted cyclic boundary conditions implemented
97!
98! 2365 2017-08-21 14:59:59Z kanani
99! Vertical nesting implemented (SadiqHuq)
100!
101! 2300 2017-06-29 13:31:14Z raasch
102! host-specific settings removed
103!
104! 2298 2017-06-29 09:28:18Z raasch
105! MPI2 related parts removed
106!
107! 2271 2017-06-09 12:34:55Z sward
108! Error message changed
109!
110! 2259 2017-06-08 09:09:11Z gronemeier
111! Implemented synthetic turbulence generator
112!
113! 2238 2017-05-31 16:49:16Z suehring
114! Remove unnecessary module load of pmc_interface
115!
116! 2231 2017-05-30 16:44:33Z suehring
117!
118! 2200 2017-04-11 11:37:51Z suehring
119! monotonic_adjustment removed
120!
121! 2197 2017-03-24 02:25:00Z raasch
122! bugfix: do not allow odd values for nz at the coarsest grid level in case of
123! optimized multigrid solver
124!
125! 2180 2017-03-17 13:33:05Z hellstea
126! Checks to ensure (2178) that pdims match the grid dimensions in the
127! automatic determination of pdims are canceled as unnecessary
128!
129! 2178 2017-03-17 11:07:39Z hellstea
130! Checks to ensure that pdims match the grid dimensions are added in the
131! automatic determination of pdims
132!
133! 2050 2016-11-08 15:00:55Z gronemeier
134! Implement turbulent outflow condition
135!
136! 2000 2016-08-20 18:09:15Z knoop
137! Forced header and separation lines into 80 columns
138!
139! 1968 2016-07-18 12:01:49Z suehring
140! Extent MPI-datatypes for exchange of 2D-INTEGER arrays on coarser multigrid
141! level 
142!
143! 1964 2016-07-14 15:35:18Z hellstea
144! Bugfix: erroneous setting of nest_bound_l/r/s/n = .TRUE. for vertical nesting mode removed.
145!
146! 1923 2016-05-31 16:37:07Z boeske
147! Initial version of purely vertical nesting introduced.
148!
149! 1922 2016-05-31 16:36:08Z boeske
150! Bugfix: array transposition checks restricted to cases if a fourier
151! transform is used , removed unused variable nnx_z
152!
153! 1833 2016-04-07 14:23:03Z raasch
154! spectra related variables moved to spectra_mod
155!
156! 1815 2016-04-06 13:49:59Z raasch
157! cpp-directives for intel openmp bug removed
158!
159! 1804 2016-04-05 16:30:18Z maronga
160! Removed code for parameter file check (__check)
161!
162! 1779 2016-03-03 08:01:28Z raasch
163! changes regarding nested domain removed: virtual PE grid will be automatically
164! calculated for nested runs too
165!
166! 1764 2016-02-28 12:45:19Z raasch
167! cpp-statements for nesting removed
168!
169! 1762 2016-02-25 12:31:13Z hellstea
170! Introduction of nested domain feature
171!
172! 1682 2015-10-07 23:56:08Z knoop
173! Code annotations made doxygen readable
174!
175! 1677 2015-10-02 13:25:23Z boeske
176! New MPI-data types for exchange of 3D integer arrays.
177!
178! 1575 2015-03-27 09:56:27Z raasch
179! adjustments for psolver-queries, calculation of ngp_xz added
180!
181! 1565 2015-03-09 20:59:31Z suehring
182! Refine if-clause for setting nbgp.
183!
184! 1557 2015-03-05 16:43:04Z suehring
185! Adjustment for monotonic limiter
186!
187! 1468 2014-09-24 14:06:57Z maronga
188! Adapted for use on up to 6-digit processor cores
189!
190! 1435 2014-07-21 10:37:02Z keck
191! bugfix: added missing parameter coupling_mode_remote to ONLY-attribute
192!
193! 1402 2014-05-09 14:25:13Z raasch
194! location messages modified
195!
196! 1384 2014-05-02 14:31:06Z raasch
197! location messages added
198!
199! 1353 2014-04-08 15:21:23Z heinze
200! REAL constants provided with KIND-attribute
201!
202! 1322 2014-03-20 16:38:49Z raasch
203! REAL functions provided with KIND-attribute
204!
205! 1320 2014-03-20 08:40:49Z raasch
206! ONLY-attribute added to USE-statements,
207! kind-parameters added to all INTEGER and REAL declaration statements,
208! kinds are defined in new module kinds,
209! revision history before 2012 removed,
210! comment fields (!:) to be used for variable explanations added to
211! all variable declaration statements
212!
213! 1304 2014-03-12 10:29:42Z raasch
214! bugfix: single core MPI runs missed some settings of transpose indices
215!
216! 1212 2013-08-15 08:46:27Z raasch
217! error message for poisfft_hybrid removed
218!
219! 1159 2013-05-21 11:58:22Z fricke
220! dirichlet/neumann and neumann/dirichlet removed
221!
222! 1139 2013-04-18 07:25:03Z raasch
223! bugfix for calculating the id of the PE carrying the recycling plane
224!
225! 1111 2013-03-08 23:54:10Z raasch
226! initialization of poisfft moved to module poisfft
227!
228! 1092 2013-02-02 11:24:22Z raasch
229! unused variables removed
230!
231! 1056 2012-11-16 15:28:04Z raasch
232! Indices for arrays n.._mg start from zero due to definition of arrays f2 and
233! p2 as automatic arrays in recursive subroutine next_mg_level
234!
235! 1041 2012-11-06 02:36:29Z raasch
236! a 2d virtual processor topology is used by default for all machines
237!
238! 1036 2012-10-22 13:43:42Z raasch
239! code put under GPL (PALM 3.9)
240!
241! 1003 2012-09-14 14:35:53Z raasch
242! subdomains must have identical size (grid matching = "match" removed)
243!
244! 1001 2012-09-13 14:08:46Z raasch
245! all actions concerning upstream-spline-method removed
246!
247! 978 2012-08-09 08:28:32Z fricke
248! dirichlet/neumann and neumann/dirichlet added
249! nxlu and nysv are also calculated for inflow boundary
250!
251! 809 2012-01-30 13:32:58Z maronga
252! Bugfix: replaced .AND. and .NOT. with && and ! in the preprocessor directives
253!
254! 807 2012-01-25 11:53:51Z maronga
255! New cpp directive "__check" implemented which is used by check_namelist_files
256!
257! Revision 1.1  1997/07/24 11:15:09  raasch
258! Initial revision
259!
260!
261! Description:
262! ------------
263!> Determination of the virtual processor topology (if not prescribed by the
264!> user)and computation of the grid point number and array bounds of the local
265!> domains.
266!> @todo: remove MPI-data types for 2D exchange on coarse multigrid level (not
267!>        used any more)
268!------------------------------------------------------------------------------!
269 SUBROUTINE init_pegrid
270 
271
272    USE control_parameters,                                                    &
273        ONLY:  bc_dirichlet_l, bc_dirichlet_n, bc_dirichlet_r, bc_dirichlet_s, &
274               bc_lr, bc_ns, bc_radiation_l, bc_radiation_n, bc_radiation_r,   &
275               bc_radiation_s, coupling_mode, coupling_topology, gathered_size,&
276               grid_level, grid_level_count, maximum_grid_level,               &
277               message_string, mg_switch_to_pe0_level, momentum_advec,         &
278               psolver, outflow_source_plane, recycling_width, scalar_advec,   &
279               subdomain_size, turbulent_inflow, turbulent_outflow, y_shift
280
281    USE grid_variables,                                                        &
282        ONLY:  dx
283       
284    USE indices,                                                               &
285        ONLY:  mg_loc_ind, nbgp, nnx, nny, nnz, nx, nx_a, nx_o, nxl, nxl_mg,   &
286               nxlu, nxr, nxr_mg, ny, ny_a, ny_o, nyn, nyn_mg, nys, nys_mg,    &
287               nysv, nz, nzb, nzt, nzt_mg, wall_flags_1, wall_flags_2,         &
288               wall_flags_3, wall_flags_4, wall_flags_5, wall_flags_6,         &
289               wall_flags_7, wall_flags_8, wall_flags_9, wall_flags_10
290
291    USE kinds
292     
293    USE pegrid
294     
295    USE spectra_mod,                                                           &
296        ONLY:  calculate_spectra, dt_dosp
297
298    USE synthetic_turbulence_generator_mod,                                    &
299        ONLY:  id_stg_left, id_stg_north, id_stg_right, id_stg_south,          &
300               use_syn_turb_gen
301
302    USE transpose_indices,                                                     &
303        ONLY:  nxl_y, nxl_yd, nxl_z, nxr_y, nxr_yd, nxr_z, nyn_x, nyn_z, nys_x,&
304               nys_z, nzb_x, nzb_y, nzb_yd, nzt_x, nzt_yd, nzt_y
305
306    USE vertical_nesting_mod,                                                  &
307        ONLY:  vnested, vnest_init_pegrid_domain, vnest_init_pegrid_rank
308
309    IMPLICIT NONE
310
311    INTEGER(iwp) ::  i                        !< running index over number of processors or number of multigrid level
312    INTEGER(iwp) ::  id_inflow_l              !< ID indicating processors located at the left inflow boundary
313    INTEGER(iwp) ::  id_outflow_l             !< local value of id_outflow
314    INTEGER(iwp) ::  id_outflow_source_l      !< local value of id_outflow_source
315    INTEGER(iwp) ::  id_recycling_l           !< ID indicating processors located at the recycling plane
316    INTEGER(iwp) ::  id_stg_left_l            !< left lateral boundary local core id in case of turbulence generator 
317    INTEGER(iwp) ::  id_stg_north_l           !< north lateral boundary local core id in case of turbulence generator 
318    INTEGER(iwp) ::  id_stg_right_l           !< right lateral boundary local core id in case of turbulence generator 
319    INTEGER(iwp) ::  id_stg_south_l           !< south lateral boundary local core id in case of turbulence generator 
320    INTEGER(iwp) ::  ind(5)                   !< array containing the subdomain bounds
321    INTEGER(iwp) ::  j                        !< running index, used for various loops
322    INTEGER(iwp) ::  k                        !< number of vertical grid points in different multigrid level
323    INTEGER(iwp) ::  maximum_grid_level_l     !< maximum number of grid level without switching to PE 0
324    INTEGER(iwp) ::  mg_levels_x              !< maximum number of grid level allowed along x-direction
325    INTEGER(iwp) ::  mg_levels_y              !< maximum number of grid level allowed along y-direction
326    INTEGER(iwp) ::  mg_levels_z              !< maximum number of grid level allowed along z-direction
327    INTEGER(iwp) ::  mg_switch_to_pe0_level_l !< maximum number of grid level with switching to PE 0
328    INTEGER(iwp) ::  nnx_y                    !< quotient of number of grid points along x-direction and number of PEs used along y-direction
329    INTEGER(iwp) ::  nny_x                    !< quotient of number of grid points along y-direction and number of PEs used along x-direction
330    INTEGER(iwp) ::  nny_z                    !< quotient of number of grid points along y-direction and number of PEs used along x-direction
331    INTEGER(iwp) ::  nnz_x                    !< quotient of number of grid points along z-direction and number of PEs used along x-direction
332    INTEGER(iwp) ::  nnz_y                    !< quotient of number of grid points along z-direction and number of PEs used along x-direction
333    INTEGER(iwp) ::  numproc_sqr              !< square root of the number of processors
334    INTEGER(iwp) ::  nxl_l                    !< lower index bound along x-direction on subdomain and different multigrid level
335    INTEGER(iwp) ::  nxr_l                    !< upper index bound along x-direction on subdomain and different multigrid level
336    INTEGER(iwp) ::  nyn_l                    !< lower index bound along y-direction on subdomain and different multigrid level
337    INTEGER(iwp) ::  nys_l                    !< upper index bound along y-direction on subdomain and different multigrid level
338    INTEGER(iwp) ::  nzb_l                    !< lower index bound along z-direction on subdomain and different multigrid level
339    INTEGER(iwp) ::  nzt_l                    !< upper index bound along z-direction on subdomain and different multigrid level
340!$  INTEGER(iwp) ::  omp_get_num_threads      !< number of OpenMP threads
341
342    INTEGER(iwp), DIMENSION(:), ALLOCATABLE ::  ind_all !< dummy array containing index bounds on subdomain, used for gathering
343    INTEGER(iwp), DIMENSION(:), ALLOCATABLE ::  nxlf    !< lower index bound allong x-direction for every PE
344    INTEGER(iwp), DIMENSION(:), ALLOCATABLE ::  nxrf    !< upper index bound allong x-direction for every PE
345    INTEGER(iwp), DIMENSION(:), ALLOCATABLE ::  nynf    !< lower index bound allong y-direction for every PE
346    INTEGER(iwp), DIMENSION(:), ALLOCATABLE ::  nysf    !< lower index bound allong y-direction for every PE
347
348    INTEGER(iwp), DIMENSION(2) ::  pdims_remote         !< number of PEs used for coupled model (only in atmospher-ocean coupling)
349    INTEGER(iwp)               ::  lcoord(2)            !< PE coordinates of left neighbor along x and y
350    INTEGER(iwp)               ::  rcoord(2)            !< PE coordinates of right neighbor along x and y
351
352!
353!-- Get the number of OpenMP threads
354    !$OMP PARALLEL
355!$  threads_per_task = omp_get_num_threads()
356    !$OMP END PARALLEL
357
358
359#if defined( __parallel )
360
361    CALL location_message( 'creating virtual PE grids + MPI derived data types', 'start' )
362
363!
364!-- Determine the processor topology or check it, if prescribed by the user
365    IF ( npex == -1  .AND.  npey == -1 )  THEN
366
367!
368!--    Automatic determination of the topology
369       numproc_sqr = SQRT( REAL( numprocs, KIND=wp ) )
370       pdims(1)    = MAX( numproc_sqr , 1 )
371       DO  WHILE ( MOD( numprocs , pdims(1) ) /= 0 )
372          pdims(1) = pdims(1) - 1
373       ENDDO
374       pdims(2) = numprocs / pdims(1)
375
376    ELSEIF ( npex /= -1  .AND.  npey /= -1 )  THEN
377
378!
379!--    Prescribed by user. Number of processors on the prescribed topology
380!--    must be equal to the number of PEs available to the job
381       IF ( ( npex * npey ) /= numprocs )  THEN
382          WRITE( message_string, * ) 'number of PEs of the prescribed ',       &
383              'topology (', npex*npey,') does not match & the number of ',     &
384              'PEs available to the job (', numprocs, ')'
385          CALL message( 'init_pegrid', 'PA0221', 1, 2, 0, 6, 0 )
386       ENDIF
387       pdims(1) = npex
388       pdims(2) = npey
389
390    ELSE
391!
392!--    If the processor topology is prescribed by the user, the number of
393!--    PEs must be given in both directions
394       message_string = 'if the processor topology is prescribed by th' //     &
395                'e user & both values of "npex" and "npey" must be given' //   &
396                ' in the &NAMELIST-parameter file'
397       CALL message( 'init_pegrid', 'PA0222', 1, 2, 0, 6, 0 )
398
399    ENDIF
400
401!
402!-- If necessary, set horizontal boundary conditions to non-cyclic
403    IF ( bc_lr /= 'cyclic' )  cyclic(1) = .FALSE.
404    IF ( bc_ns /= 'cyclic' )  cyclic(2) = .FALSE.
405
406
407!
408!-- Create the virtual processor grid
409    CALL MPI_CART_CREATE( comm_palm, ndim, pdims, cyclic, reorder, &
410                          comm2d, ierr )
411    CALL MPI_COMM_RANK( comm2d, myid, ierr )
412    WRITE (myid_char,'(''_'',I6.6)')  myid
413
414    CALL MPI_CART_COORDS( comm2d, myid, ndim, pcoord, ierr )
415    CALL MPI_CART_SHIFT( comm2d, 0, 1, pleft, pright, ierr )
416    CALL MPI_CART_SHIFT( comm2d, 1, 1, psouth, pnorth, ierr )
417!
418!-- In case of cyclic boundary conditions, a y-shift at the boundaries in
419!-- x-direction can be introduced via parameter y_shift. The shift is done
420!-- by modifying the processor grid in such a way that processors located
421!-- at the x-boundary communicate across it to processors with y-coordinate
422!-- shifted by y_shift relative to their own. This feature can not be used
423!-- in combination with an fft pressure solver. It has been implemented to
424!-- counter the effect of streak structures in case of cyclic boundary
425!-- conditions. For a description of these see Munters
426!-- (2016; dx.doi.org/10.1063/1.4941912)
427!--
428!-- Get coordinates of left and right neighbor on PE grid
429    IF ( y_shift /= 0 ) THEN
430
431       IF ( bc_lr /= 'cyclic'  .OR.  bc_ns /= 'cyclic' )  THEN
432          message_string = 'y_shift /= 0 is only allowed for cyclic ' //       &
433                           'boundary conditions in both directions '
434          CALL message( 'check_parameters', 'PA0467', 1, 2, 0, 6, 0 )
435       ENDIF
436       IF ( TRIM( psolver ) /= 'multigrid' .AND.                               &
437            TRIM( psolver ) /= 'multigrid_noopt')                              &
438       THEN
439          message_string = 'y_shift /= 0 requires a multigrid pressure solver '
440          CALL message( 'check_parameters', 'PA0468', 1, 2, 0, 6, 0 )
441       ENDIF
442
443       CALL MPI_CART_COORDS( comm2d, pright, ndim, rcoord, ierr )
444       CALL MPI_CART_COORDS( comm2d, pleft, ndim, lcoord, ierr )
445
446!
447!--    If the x(y)-coordinate of the right (left) neighbor is smaller (greater)
448!--    than that of the calling process, then the calling process is located on
449!--    the right (left) boundary of the processor grid. In that case,
450!--    the y-coordinate of that neighbor is increased (decreased) by y_shift.
451!--    The rank of the process with that coordinate is then inquired and the
452!--    neighbor rank for MPI_SENDRECV, pright (pleft) is set to it.
453!--    In this way, the calling process receives a new right (left) neighbor
454!--    for all future MPI_SENDRECV calls. That neighbor has a y-coordinate
455!--    of y+(-)y_shift, where y is the original right (left) neighbor's
456!--    y-coordinate. The modulo-operation ensures that if the neighbor's
457!--    y-coordinate exceeds the grid-boundary, it will be relocated to
458!--    the opposite part of the grid cyclicly.
459       IF ( rcoord(1) < pcoord(1) ) THEN
460          rcoord(2) = MODULO( rcoord(2) + y_shift, pdims(2) )
461          CALL MPI_CART_RANK( comm2d, rcoord, pright, ierr )
462       ENDIF
463
464       IF ( lcoord(1) > pcoord(1) ) THEN
465          lcoord(2) = MODULO( lcoord(2) - y_shift, pdims(2) )
466          CALL MPI_CART_RANK( comm2d, lcoord, pleft, ierr )
467       ENDIF
468    ENDIF
469!
470!-- Vertical nesting: store four lists that identify partner ranks to exchange
471!-- data
472    IF ( vnested )  CALL vnest_init_pegrid_rank
473
474!
475!-- Determine sub-topologies for transpositions
476!-- Transposition from z to x:
477    remain_dims(1) = .TRUE.
478    remain_dims(2) = .FALSE.
479    CALL MPI_CART_SUB( comm2d, remain_dims, comm1dx, ierr )
480    CALL MPI_COMM_RANK( comm1dx, myidx, ierr )
481!
482!-- Transposition from x to y
483    remain_dims(1) = .FALSE.
484    remain_dims(2) = .TRUE.
485    CALL MPI_CART_SUB( comm2d, remain_dims, comm1dy, ierr )
486    CALL MPI_COMM_RANK( comm1dy, myidy, ierr )
487
488
489!
490!-- Calculate array bounds along x-direction for every PE.
491    ALLOCATE( nxlf(0:pdims(1)-1), nxrf(0:pdims(1)-1), nynf(0:pdims(2)-1),      &
492              nysf(0:pdims(2)-1) )
493
494    IF ( MOD( nx+1 , pdims(1) ) /= 0 )  THEN
495       WRITE( message_string, * ) 'x-direction: gridpoint number (',nx+1,') ', &
496                               'is not an& integral divisor of the number ',    &
497                               'of processors (', pdims(1),')'
498       CALL message( 'init_pegrid', 'PA0225', 1, 2, 0, 6, 0 )
499    ELSE
500       nnx  = ( nx + 1 ) / pdims(1)
501    ENDIF   
502
503!
504!-- Left and right array bounds, number of gridpoints
505    DO  i = 0, pdims(1)-1
506       nxlf(i)   = i * nnx
507       nxrf(i)   = ( i + 1 ) * nnx - 1
508    ENDDO
509
510!
511!-- Calculate array bounds in y-direction for every PE.
512    IF ( MOD( ny+1 , pdims(2) ) /= 0 )  THEN
513       WRITE( message_string, * ) 'y-direction: gridpoint number (',ny+1,') ', &
514                           'is not an& integral divisor of the number of',      &
515                           'processors (', pdims(2),')'
516       CALL message( 'init_pegrid', 'PA0227', 1, 2, 0, 6, 0 )
517    ELSE
518       nny  = ( ny + 1 ) / pdims(2)
519    ENDIF   
520
521!
522!-- South and north array bounds
523    DO  j = 0, pdims(2)-1
524       nysf(j)   = j * nny
525       nynf(j)   = ( j + 1 ) * nny - 1
526    ENDDO
527
528!
529!-- Local array bounds of the respective PEs
530    nxl = nxlf(pcoord(1))
531    nxr = nxrf(pcoord(1))
532    nys = nysf(pcoord(2))
533    nyn = nynf(pcoord(2))
534    nzb = 0
535    nzt = nz
536    nnz = nz
537
538!
539!-- Set switches to define if the PE is situated at the border of the virtual
540!-- processor grid
541    IF ( nxl == 0 )   left_border_pe  = .TRUE.
542    IF ( nxr == nx )  right_border_pe = .TRUE.
543    IF ( nys == 0 )   south_border_pe = .TRUE.
544    IF ( nyn == ny )  north_border_pe = .TRUE.
545
546!
547!-- Calculate array bounds and gridpoint numbers for the transposed arrays
548!-- (needed in the pressure solver)
549!-- For the transposed arrays, cyclic boundaries as well as top and bottom
550!-- boundaries are omitted, because they are obstructive to the transposition
551
552!
553!-- 1. transposition  z --> x
554!-- This transposition is not neccessary in case of a 1d-decomposition along x
555    IF ( psolver == 'poisfft'  .OR.  calculate_spectra )  THEN
556
557       IF ( pdims(2) /= 1 )  THEN
558          IF ( MOD( nz , pdims(1) ) /= 0 )  THEN
559             WRITE( message_string, * ) 'transposition z --> x:',              &
560                       '& nz=',nz,' is not an integral divisior of pdims(1)=', &
561                                                                   pdims(1)
562             CALL message( 'init_pegrid', 'PA0230', 1, 2, 0, 6, 0 )
563          ENDIF
564       ENDIF
565
566       nys_x = nys
567       nyn_x = nyn
568       nny_x = nny
569       nnz_x = nz / pdims(1)
570       nzb_x = 1 + myidx * nnz_x
571       nzt_x = ( myidx + 1 ) * nnz_x
572       sendrecvcount_zx = nnx * nny * nnz_x
573
574    ENDIF
575
576
577    IF ( psolver == 'poisfft' )  THEN 
578!
579!--    2. transposition  x --> y
580       IF ( MOD( nx+1 , pdims(2) ) /= 0 )  THEN
581          WRITE( message_string, * ) 'transposition x --> y:',                 &
582                            '& nx+1=',nx+1,' is not an integral divisor of ',  &
583                            'pdims(2)=',pdims(2)
584          CALL message( 'init_pegrid', 'PA0231', 1, 2, 0, 6, 0 )
585       ENDIF
586
587       nnz_y = nnz_x
588       nzb_y = nzb_x
589       nzt_y = nzt_x
590       nnx_y = (nx+1) / pdims(2)
591       nxl_y = myidy * nnx_y
592       nxr_y = ( myidy + 1 ) * nnx_y - 1
593       sendrecvcount_xy = nnx_y * nny_x * nnz_y
594!
595!--    3. transposition  y --> z 
596!--    (ELSE:  x --> y  in case of 1D-decomposition along x)
597       nxl_z = nxl_y
598       nxr_z = nxr_y
599       nny_z = (ny+1) / pdims(1)
600       nys_z = myidx * nny_z
601       nyn_z = ( myidx + 1 ) * nny_z - 1
602       sendrecvcount_yz = nnx_y * nny_z * nnz_y
603
604       IF ( pdims(2) /= 1 )  THEN
605!
606!--       y --> z
607!--       This transposition is not neccessary in case of a 1d-decomposition
608!--       along x, except that the uptream-spline method is switched on
609          IF ( MOD( ny+1 , pdims(1) ) /= 0 )  THEN
610             WRITE( message_string, * ) 'transposition y --> z:',              &
611                               '& ny+1=',ny+1,' is not an integral divisor of',&
612                               ' pdims(1)=',pdims(1)
613             CALL message( 'init_pegrid', 'PA0232', 1, 2, 0, 6, 0 )
614          ENDIF
615
616       ELSE
617!
618!--       x --> y
619!--       This condition must be fulfilled for a 1D-decomposition along x
620          IF ( MOD( ny+1 , pdims(1) ) /= 0 )  THEN
621             WRITE( message_string, * ) 'transposition x --> y:',              &
622                               '& ny+1=',ny+1,' is not an integral divisor of',&
623                               ' pdims(1)=',pdims(1)
624             CALL message( 'init_pegrid', 'PA0233', 1, 2, 0, 6, 0 )
625          ENDIF
626
627       ENDIF
628
629    ENDIF
630
631!
632!-- Indices for direct transpositions z --> y (used for calculating spectra)
633    IF ( calculate_spectra )  THEN
634       IF ( MOD( nz, pdims(2) ) /= 0 )  THEN
635          WRITE( message_string, * ) 'direct transposition z --> y (needed ',  &
636                    'for spectra): nz=',nz,' is not an integral divisor of ',  &
637                    'pdims(2)=',pdims(2)
638          CALL message( 'init_pegrid', 'PA0234', 1, 2, 0, 6, 0 )
639       ELSE
640          nxl_yd = nxl
641          nxr_yd = nxr
642          nzb_yd = 1 + myidy * ( nz / pdims(2) )
643          nzt_yd = ( myidy + 1 ) * ( nz / pdims(2) )
644          sendrecvcount_zyd = nnx * nny * ( nz / pdims(2) )
645       ENDIF
646    ENDIF
647
648    IF ( psolver == 'poisfft'  .OR.  calculate_spectra )  THEN
649!
650!--    Indices for direct transpositions y --> x
651!--    (they are only possible in case of a 1d-decomposition along x)
652       IF ( pdims(2) == 1 )  THEN
653          nny_x = nny / pdims(1)
654          nys_x = myid * nny_x
655          nyn_x = ( myid + 1 ) * nny_x - 1
656          nzb_x = 1
657          nzt_x = nz
658          sendrecvcount_xy = nnx * nny_x * nz
659       ENDIF
660
661    ENDIF
662
663    IF ( psolver == 'poisfft' )  THEN
664!
665!--    Indices for direct transpositions x --> y
666!--    (they are only possible in case of a 1d-decomposition along y)
667       IF ( pdims(1) == 1 )  THEN
668          nnx_y = nnx / pdims(2)
669          nxl_y = myid * nnx_y
670          nxr_y = ( myid + 1 ) * nnx_y - 1
671          nzb_y = 1
672          nzt_y = nz
673          sendrecvcount_xy = nnx_y * nny * nz
674       ENDIF
675
676    ENDIF
677
678!
679!-- Arrays for storing the array bounds are needed any more
680    DEALLOCATE( nxlf , nxrf , nynf , nysf )
681
682
683!
684!-- Collect index bounds from other PEs (to be written to restart file later)
685    ALLOCATE( hor_index_bounds(4,0:numprocs-1) )
686
687    IF ( myid == 0 )  THEN
688
689       hor_index_bounds(1,0) = nxl
690       hor_index_bounds(2,0) = nxr
691       hor_index_bounds(3,0) = nys
692       hor_index_bounds(4,0) = nyn
693
694!
695!--    Receive data from all other PEs
696       DO  i = 1, numprocs-1
697          CALL MPI_RECV( ibuf, 4, MPI_INTEGER, i, MPI_ANY_TAG, comm2d, status, &
698                         ierr )
699          hor_index_bounds(:,i) = ibuf(1:4)
700       ENDDO
701
702    ELSE
703!
704!--    Send index bounds to PE0
705       ibuf(1) = nxl
706       ibuf(2) = nxr
707       ibuf(3) = nys
708       ibuf(4) = nyn
709       CALL MPI_SEND( ibuf, 4, MPI_INTEGER, 0, myid, comm2d, ierr )
710
711    ENDIF
712
713
714#if defined( __print )
715!
716!-- Control output
717    IF ( myid == 0 )  THEN
718       PRINT*, '*** processor topology ***'
719       PRINT*, ' '
720       PRINT*, 'myid   pcoord    left right  south north  idx idy   nxl: nxr',&
721               &'   nys: nyn'
722       PRINT*, '------------------------------------------------------------',&
723               &'-----------'
724       WRITE (*,1000)  0, pcoord(1), pcoord(2), pleft, pright, psouth, pnorth, &
725                       myidx, myidy, nxl, nxr, nys, nyn
7261000   FORMAT (I4,2X,'(',I3,',',I3,')',3X,I4,2X,I4,3X,I4,2X,I4,2X,I3,1X,I3, &
727               2(2X,I4,':',I4))
728
729!
730!--    Receive data from the other PEs
731       DO  i = 1,numprocs-1
732          CALL MPI_RECV( ibuf, 12, MPI_INTEGER, i, MPI_ANY_TAG, comm2d, status, &
733                         ierr )
734          WRITE (*,1000)  i, ( ibuf(j) , j = 1,12 )
735       ENDDO
736    ELSE
737
738!
739!--    Send data to PE0
740       ibuf(1) = pcoord(1); ibuf(2) = pcoord(2); ibuf(3) = pleft
741       ibuf(4) = pright; ibuf(5) = psouth; ibuf(6) = pnorth; ibuf(7) = myidx
742       ibuf(8) = myidy; ibuf(9) = nxl; ibuf(10) = nxr; ibuf(11) = nys
743       ibuf(12) = nyn
744       CALL MPI_SEND( ibuf, 12, MPI_INTEGER, 0, myid, comm2d, ierr )       
745    ENDIF
746#endif
747
748!
749!-- Determine the number of ghost point layers
750    IF ( scalar_advec   == 'ws-scheme'  .OR.                                   &
751         momentum_advec == 'ws-scheme' )  THEN
752       nbgp = 3
753    ELSE
754       nbgp = 1
755    ENDIF
756
757!
758!-- Create a new MPI derived datatype for the exchange of surface (xy) data,
759!-- which is needed for coupled atmosphere-ocean runs.
760!-- First, calculate number of grid points of an xy-plane.
761    ngp_xy  = ( nxr - nxl + 1 + 2 * nbgp ) * ( nyn - nys + 1 + 2 * nbgp )
762    CALL MPI_TYPE_VECTOR( ngp_xy, 1, nzt-nzb+2, MPI_REAL, type_xy, ierr )
763    CALL MPI_TYPE_COMMIT( type_xy, ierr )
764
765    IF ( TRIM( coupling_mode ) /= 'uncoupled' .AND. .NOT. vnested )  THEN
766   
767!
768!--    Pass the number of grid points of the atmosphere model to
769!--    the ocean model and vice versa
770       IF ( coupling_mode == 'atmosphere_to_ocean' )  THEN
771
772          nx_a = nx
773          ny_a = ny
774
775          IF ( myid == 0 )  THEN
776
777             CALL MPI_SEND( nx_a, 1, MPI_INTEGER, numprocs, 1, comm_inter,  &
778                            ierr )
779             CALL MPI_SEND( ny_a, 1, MPI_INTEGER, numprocs, 2, comm_inter,  &
780                            ierr )
781             CALL MPI_SEND( pdims, 2, MPI_INTEGER, numprocs, 3, comm_inter, &
782                            ierr )
783             CALL MPI_RECV( nx_o, 1, MPI_INTEGER, numprocs, 4, comm_inter,  &
784                            status, ierr )
785             CALL MPI_RECV( ny_o, 1, MPI_INTEGER, numprocs, 5, comm_inter,  &
786                            status, ierr )
787             CALL MPI_RECV( pdims_remote, 2, MPI_INTEGER, numprocs, 6,      &
788                            comm_inter, status, ierr )
789          ENDIF
790
791          CALL MPI_BCAST( nx_o, 1, MPI_INTEGER, 0, comm2d, ierr )
792          CALL MPI_BCAST( ny_o, 1, MPI_INTEGER, 0, comm2d, ierr ) 
793          CALL MPI_BCAST( pdims_remote, 2, MPI_INTEGER, 0, comm2d, ierr )
794       
795       ELSEIF ( coupling_mode == 'ocean_to_atmosphere' )  THEN
796
797          nx_o = nx
798          ny_o = ny
799
800          IF ( myid == 0 ) THEN
801
802             CALL MPI_RECV( nx_a, 1, MPI_INTEGER, 0, 1, comm_inter, status, &
803                            ierr )
804             CALL MPI_RECV( ny_a, 1, MPI_INTEGER, 0, 2, comm_inter, status, &
805                            ierr )
806             CALL MPI_RECV( pdims_remote, 2, MPI_INTEGER, 0, 3, comm_inter, &
807                            status, ierr )
808             CALL MPI_SEND( nx_o, 1, MPI_INTEGER, 0, 4, comm_inter, ierr )
809             CALL MPI_SEND( ny_o, 1, MPI_INTEGER, 0, 5, comm_inter, ierr )
810             CALL MPI_SEND( pdims, 2, MPI_INTEGER, 0, 6, comm_inter, ierr )
811          ENDIF
812
813          CALL MPI_BCAST( nx_a, 1, MPI_INTEGER, 0, comm2d, ierr)
814          CALL MPI_BCAST( ny_a, 1, MPI_INTEGER, 0, comm2d, ierr) 
815          CALL MPI_BCAST( pdims_remote, 2, MPI_INTEGER, 0, comm2d, ierr) 
816
817       ENDIF
818 
819       ngp_a = ( nx_a+1 + 2 * nbgp ) * ( ny_a+1 + 2 * nbgp )
820       ngp_o = ( nx_o+1 + 2 * nbgp ) * ( ny_o+1 + 2 * nbgp )
821
822!
823!--    Determine if the horizontal grid and the number of PEs in ocean and
824!--    atmosphere is same or not
825       IF ( nx_o == nx_a  .AND.  ny_o == ny_a  .AND.  &
826            pdims(1) == pdims_remote(1) .AND. pdims(2) == pdims_remote(2) ) &
827       THEN
828          coupling_topology = 0
829       ELSE
830          coupling_topology = 1
831       ENDIF
832
833!
834!--    Determine the target PEs for the exchange between ocean and
835!--    atmosphere (comm2d)
836       IF ( coupling_topology == 0 )  THEN
837!
838!--       In case of identical topologies, every atmosphere PE has exactly one
839!--       ocean PE counterpart and vice versa
840          IF ( TRIM( coupling_mode ) == 'atmosphere_to_ocean' ) THEN
841             target_id = myid + numprocs
842          ELSE
843             target_id = myid
844          ENDIF
845
846       ELSE
847!
848!--       In case of nonequivalent topology in ocean and atmosphere only for
849!--       PE0 in ocean and PE0 in atmosphere a target_id is needed, since
850!--       data echxchange between ocean and atmosphere will be done only
851!--       between these PEs.   
852          IF ( myid == 0 )  THEN
853
854             IF ( TRIM( coupling_mode ) == 'atmosphere_to_ocean' )  THEN
855                target_id = numprocs
856             ELSE
857                target_id = 0
858             ENDIF
859
860          ENDIF
861
862       ENDIF
863
864    ENDIF
865
866!
867!-- Store partner grid point co-ordinates as lists.
868!-- Create custom MPI vector datatypes for contiguous data transfer
869    IF ( vnested )  CALL vnest_init_pegrid_domain
870
871#else
872
873!
874!-- Array bounds when running on a single PE (respectively a non-parallel
875!-- machine)
876    nxl = 0
877    nxr = nx
878    nnx = nxr - nxl + 1
879    nys = 0
880    nyn = ny
881    nny = nyn - nys + 1
882    nzb = 0
883    nzt = nz
884    nnz = nz
885
886    ALLOCATE( hor_index_bounds(4,0:0) )
887    hor_index_bounds(1,0) = nxl
888    hor_index_bounds(2,0) = nxr
889    hor_index_bounds(3,0) = nys
890    hor_index_bounds(4,0) = nyn
891
892!
893!-- Array bounds for the pressure solver (in the parallel code, these bounds
894!-- are the ones for the transposed arrays)
895    nys_x = nys
896    nyn_x = nyn
897    nzb_x = nzb + 1
898    nzt_x = nzt
899
900    nxl_y = nxl
901    nxr_y = nxr
902    nzb_y = nzb + 1
903    nzt_y = nzt
904
905    nxl_z = nxl
906    nxr_z = nxr
907    nys_z = nys
908    nyn_z = nyn
909
910#endif
911
912!
913!-- Calculate number of grid levels necessary for the multigrid poisson solver
914!-- as well as the gridpoint indices on each level
915    IF ( psolver(1:9) == 'multigrid' )  THEN
916
917!
918!--    First calculate number of possible grid levels for the subdomains
919       mg_levels_x = 1
920       mg_levels_y = 1
921       mg_levels_z = 1
922
923       i = nnx
924       DO WHILE ( MOD( i, 2 ) == 0  .AND.  i /= 2 )
925          i = i / 2
926          mg_levels_x = mg_levels_x + 1
927       ENDDO
928
929       j = nny
930       DO WHILE ( MOD( j, 2 ) == 0  .AND.  j /= 2 )
931          j = j / 2
932          mg_levels_y = mg_levels_y + 1
933       ENDDO
934
935       k = nz    ! do not use nnz because it might be > nz due to transposition
936                 ! requirements
937       DO WHILE ( MOD( k, 2 ) == 0  .AND.  k /= 2 )
938          k = k / 2
939          mg_levels_z = mg_levels_z + 1
940       ENDDO
941!
942!--    The optimized MG-solver does not allow odd values for nz at the coarsest
943!--    grid level
944       IF ( TRIM( psolver ) /= 'multigrid_noopt' )  THEN
945          IF ( MOD( k, 2 ) /= 0 )  mg_levels_z = mg_levels_z - 1
946!
947!--       An odd value of nz does not work. The finest level must have an even
948!--       value.
949          IF (  mg_levels_z == 0 )  THEN
950             message_string = 'optimized multigrid method requires nz to be even'
951             CALL message( 'init_pegrid', 'PA0495', 1, 2, 0, 6, 0 )
952          ENDIF
953       ENDIF
954
955       maximum_grid_level = MIN( mg_levels_x, mg_levels_y, mg_levels_z )
956!       
957!--    Check if subdomain sizes prevents any coarsening.
958!--    This case, the maximum number of grid levels is 1, i.e. effectively
959!--    a Gauss-Seidel scheme is applied rather than a multigrid approach.
960!--    Give a warning in this case.
961       IF ( maximum_grid_level == 1  .AND.  mg_switch_to_pe0_level == -1 )  THEN
962          message_string = 'No grid coarsening possible, multigrid ' //        &
963                           'approach effectively reduces to a Gauss-Seidel ' //&
964                           'scheme.'
965         
966          CALL message( 'poismg', 'PA0648', 0, 1, 0, 6, 0 )
967       ENDIF
968
969!
970!--    Find out, if the total domain allows more levels. These additional
971!--    levels are identically processed on all PEs.
972       IF ( numprocs > 1  .AND.  mg_switch_to_pe0_level /= -1 )  THEN
973
974          IF ( mg_levels_z > MIN( mg_levels_x, mg_levels_y ) )  THEN
975
976             mg_switch_to_pe0_level_l = maximum_grid_level
977
978             mg_levels_x = 1
979             mg_levels_y = 1
980
981             i = nx+1
982             DO WHILE ( MOD( i, 2 ) == 0  .AND.  i /= 2 )
983                i = i / 2
984                mg_levels_x = mg_levels_x + 1
985             ENDDO
986
987             j = ny+1
988             DO WHILE ( MOD( j, 2 ) == 0  .AND.  j /= 2 )
989                j = j / 2
990                mg_levels_y = mg_levels_y + 1
991             ENDDO
992
993             maximum_grid_level_l = MIN( mg_levels_x, mg_levels_y, mg_levels_z )
994
995             IF ( maximum_grid_level_l > mg_switch_to_pe0_level_l )  THEN
996                mg_switch_to_pe0_level_l = maximum_grid_level_l - &
997                                           mg_switch_to_pe0_level_l + 1
998             ELSE
999                mg_switch_to_pe0_level_l = 0
1000             ENDIF
1001
1002          ELSE
1003
1004             mg_switch_to_pe0_level_l = 0
1005             maximum_grid_level_l = maximum_grid_level
1006
1007          ENDIF
1008
1009!
1010!--       Use switch level calculated above only if it is not pre-defined
1011!--       by user
1012          IF ( mg_switch_to_pe0_level == 0 )  THEN
1013             IF ( mg_switch_to_pe0_level_l /= 0 )  THEN
1014                mg_switch_to_pe0_level = mg_switch_to_pe0_level_l
1015                maximum_grid_level     = maximum_grid_level_l
1016             ENDIF
1017
1018          ELSE
1019!
1020!--          Check pre-defined value and reset to default, if neccessary
1021             IF ( mg_switch_to_pe0_level < mg_switch_to_pe0_level_l  .OR.      &
1022                  mg_switch_to_pe0_level >= maximum_grid_level_l )  THEN
1023                message_string = 'mg_switch_to_pe0_level ' //                  &
1024                                 'out of range and reset to 0'
1025                CALL message( 'init_pegrid', 'PA0235', 0, 1, 0, 6, 0 )
1026                mg_switch_to_pe0_level = 0
1027             ELSE
1028!
1029!--             Use the largest number of possible levels anyway and recalculate
1030!--             the switch level to this largest number of possible values
1031                maximum_grid_level = maximum_grid_level_l
1032
1033             ENDIF
1034
1035          ENDIF
1036
1037       ENDIF
1038
1039       ALLOCATE( grid_level_count(maximum_grid_level),                       &
1040                 nxl_mg(0:maximum_grid_level), nxr_mg(0:maximum_grid_level), &
1041                 nyn_mg(0:maximum_grid_level), nys_mg(0:maximum_grid_level), &
1042                 nzt_mg(0:maximum_grid_level) )
1043
1044       grid_level_count = 0
1045!
1046!--    Index zero required as dummy due to definition of arrays f2 and p2 in
1047!--    recursive subroutine next_mg_level
1048       nxl_mg(0) = 0; nxr_mg(0) = 0; nyn_mg(0) = 0; nys_mg(0) = 0; nzt_mg(0) = 0
1049
1050       nxl_l = nxl; nxr_l = nxr; nys_l = nys; nyn_l = nyn; nzt_l = nzt
1051
1052       DO  i = maximum_grid_level, 1 , -1
1053
1054          IF ( i == mg_switch_to_pe0_level )  THEN
1055#if defined( __parallel )
1056!
1057!--          Save the grid size of the subdomain at the switch level, because
1058!--          it is needed in poismg.
1059             ind(1) = nxl_l; ind(2) = nxr_l
1060             ind(3) = nys_l; ind(4) = nyn_l
1061             ind(5) = nzt_l
1062             ALLOCATE( ind_all(5*numprocs), mg_loc_ind(5,0:numprocs-1) )
1063             CALL MPI_ALLGATHER( ind, 5, MPI_INTEGER, ind_all, 5, &
1064                                 MPI_INTEGER, comm2d, ierr )
1065             DO  j = 0, numprocs-1
1066                DO  k = 1, 5
1067                   mg_loc_ind(k,j) = ind_all(k+j*5)
1068                ENDDO
1069             ENDDO
1070             DEALLOCATE( ind_all )
1071!
1072!--          Calculate the grid size of the total domain
1073             nxr_l = ( nxr_l-nxl_l+1 ) * pdims(1) - 1
1074             nxl_l = 0
1075             nyn_l = ( nyn_l-nys_l+1 ) * pdims(2) - 1
1076             nys_l = 0
1077!
1078!--          The size of this gathered array must not be larger than the
1079!--          array tend, which is used in the multigrid scheme as a temporary
1080!--          array. Therefore the subdomain size of an PE is calculated and
1081!--          the size of the gathered grid. These values are used in 
1082!--          routines pres and poismg
1083             subdomain_size = ( nxr - nxl + 2 * nbgp + 1 ) * &
1084                              ( nyn - nys + 2 * nbgp + 1 ) * ( nzt - nzb + 2 )
1085             gathered_size  = ( nxr_l - nxl_l + 3 ) * ( nyn_l - nys_l + 3 ) *  &
1086                              ( nzt_l - nzb + 2 )
1087
1088#else
1089             message_string = 'multigrid gather/scatter impossible ' //        &
1090                          'in non parallel mode'
1091             CALL message( 'init_pegrid', 'PA0237', 1, 2, 0, 6, 0 )
1092#endif
1093          ENDIF
1094
1095          nxl_mg(i) = nxl_l
1096          nxr_mg(i) = nxr_l
1097          nys_mg(i) = nys_l
1098          nyn_mg(i) = nyn_l
1099          nzt_mg(i) = nzt_l
1100
1101          nxl_l = nxl_l / 2 
1102          nxr_l = nxr_l / 2
1103          nys_l = nys_l / 2 
1104          nyn_l = nyn_l / 2 
1105          nzt_l = nzt_l / 2 
1106         
1107       ENDDO
1108
1109!
1110!--    Temporary problem: Currently calculation of maxerror in routine poismg crashes
1111!--    if grid data are collected on PE0 already on the finest grid level.
1112!--    To be solved later.
1113       IF ( maximum_grid_level == mg_switch_to_pe0_level )  THEN
1114          message_string = 'grid coarsening on subdomain level cannot be performed'
1115          CALL message( 'poismg', 'PA0236', 1, 2, 0, 6, 0 )
1116       ENDIF
1117
1118    ELSE
1119
1120       maximum_grid_level = 0
1121
1122    ENDIF
1123
1124!
1125!-- Default level 0 tells exchange_horiz that all ghost planes have to be
1126!-- exchanged. grid_level is adjusted in poismg, where only one ghost plane
1127!-- is required.
1128    grid_level = 0
1129
1130#if defined( __parallel )
1131!
1132!-- Gridpoint number for the exchange of ghost points (y-line for 2D-arrays)
1133    ngp_y  = nyn - nys + 1 + 2 * nbgp
1134
1135!
1136!-- Define new MPI derived datatypes for the exchange of ghost points in
1137!-- x- and y-direction for 2D-arrays (line)
1138    CALL MPI_TYPE_VECTOR( nxr-nxl+1+2*nbgp, nbgp, ngp_y, MPI_REAL, type_x,     &
1139                          ierr )
1140    CALL MPI_TYPE_COMMIT( type_x, ierr )
1141
1142    CALL MPI_TYPE_VECTOR( nbgp, ngp_y, ngp_y, MPI_REAL, type_y, ierr )
1143    CALL MPI_TYPE_COMMIT( type_y, ierr )
1144!
1145!-- Define new MPI derived datatypes for the exchange of ghost points in
1146!-- x- and y-direction for 2D-INTEGER arrays (line) - on normal grid.
1147!-- Define types for 32-bit and 8-bit Integer. The 8-bit Integer are only
1148!-- required on normal grid, while 32-bit Integer may be also required on
1149!-- coarser grid level in case of multigrid solver.
1150!
1151!-- 8-bit Integer
1152    CALL MPI_TYPE_VECTOR( nxr-nxl+1+2*nbgp, nbgp, ngp_y, MPI_BYTE,             &
1153                          type_x_byte, ierr )
1154    CALL MPI_TYPE_COMMIT( type_x_byte, ierr )
1155
1156    CALL MPI_TYPE_VECTOR( nbgp, ngp_y, ngp_y, MPI_BYTE,                        &
1157                          type_y_byte, ierr )
1158    CALL MPI_TYPE_COMMIT( type_y_byte, ierr )
1159!
1160!-- 32-bit Integer
1161    ALLOCATE( type_x_int(0:maximum_grid_level),                                &
1162              type_y_int(0:maximum_grid_level) )
1163             
1164    CALL MPI_TYPE_VECTOR( nxr-nxl+1+2*nbgp, nbgp, ngp_y, MPI_INTEGER,          &
1165                          type_x_int(0), ierr )
1166    CALL MPI_TYPE_COMMIT( type_x_int(0), ierr )
1167
1168    CALL MPI_TYPE_VECTOR( nbgp, ngp_y, ngp_y, MPI_INTEGER, type_y_int(0), ierr )
1169    CALL MPI_TYPE_COMMIT( type_y_int(0), ierr )
1170!
1171!-- Calculate gridpoint numbers for the exchange of ghost points along x
1172!-- (yz-plane for 3D-arrays) and define MPI derived data type(s) for the
1173!-- exchange of ghost points in y-direction (xz-plane).
1174!-- Do these calculations for the model grid and (if necessary) also
1175!-- for the coarser grid levels used in the multigrid method
1176    ALLOCATE ( ngp_xz(0:maximum_grid_level),                                   &
1177               ngp_xz_int(0:maximum_grid_level),                               &
1178               ngp_yz(0:maximum_grid_level),                                   &
1179               ngp_yz_int(0:maximum_grid_level),                               &
1180               type_xz(0:maximum_grid_level),                                  &
1181               type_xz_int(0:maximum_grid_level),                              &
1182               type_yz(0:maximum_grid_level),                                  &
1183               type_yz_int(0:maximum_grid_level) )
1184
1185    nxl_l = nxl; nxr_l = nxr; nys_l = nys; nyn_l = nyn; nzb_l = nzb; nzt_l = nzt
1186
1187!
1188!-- Discern between the model grid, which needs nbgp ghost points and
1189!-- grid levels for the multigrid scheme. In the latter case only one
1190!-- ghost point is necessary.
1191!-- First definition of MPI-datatypes for exchange of ghost layers on normal
1192!-- grid. The following loop is needed for data exchange in poismg.f90.
1193!
1194!-- Determine number of grid points of yz-layer for exchange
1195    ngp_yz(0) = (nzt - nzb + 2) * (nyn - nys + 1 + 2 * nbgp)
1196
1197!
1198!-- Define an MPI-datatype for the exchange of left/right boundaries.
1199!-- Although data are contiguous in physical memory (which does not
1200!-- necessarily require an MPI-derived datatype), the data exchange between
1201!-- left and right PE's using the MPI-derived type is 10% faster than without.
1202    CALL MPI_TYPE_VECTOR( nxr-nxl+1+2*nbgp, nbgp*(nzt-nzb+2), ngp_yz(0), &
1203                          MPI_REAL, type_xz(0), ierr )
1204    CALL MPI_TYPE_COMMIT( type_xz(0), ierr )
1205
1206    CALL MPI_TYPE_VECTOR( nbgp, ngp_yz(0), ngp_yz(0), MPI_REAL, type_yz(0), &
1207                          ierr ) 
1208    CALL MPI_TYPE_COMMIT( type_yz(0), ierr )
1209
1210!
1211!-- Define data types for exchange of 3D Integer arrays.
1212    ngp_yz_int(0) = (nzt - nzb + 2) * (nyn - nys + 1 + 2 * nbgp)
1213
1214    CALL MPI_TYPE_VECTOR( nxr-nxl+1+2*nbgp, nbgp*(nzt-nzb+2), ngp_yz_int(0),   &
1215                          MPI_INTEGER, type_xz_int(0), ierr )
1216    CALL MPI_TYPE_COMMIT( type_xz_int(0), ierr )
1217
1218    CALL MPI_TYPE_VECTOR( nbgp, ngp_yz_int(0), ngp_yz_int(0), MPI_INTEGER,     &
1219                          type_yz_int(0), ierr )
1220    CALL MPI_TYPE_COMMIT( type_yz_int(0), ierr )
1221
1222!
1223!-- Definition of MPI-datatypes for multigrid method (coarser level grids)
1224    IF ( psolver(1:9) == 'multigrid' )  THEN
1225!   
1226!--    Definition of MPI-datatyoe as above, but only 1 ghost level is used
1227       DO  i = maximum_grid_level, 1 , -1
1228!
1229!--       For 3D-exchange on different multigrid level, one ghost point for
1230!--       REAL arrays, two ghost points for INTEGER arrays
1231          ngp_xz(i) = (nzt_l - nzb_l + 2) * (nxr_l - nxl_l + 3)
1232          ngp_yz(i) = (nzt_l - nzb_l + 2) * (nyn_l - nys_l + 3)
1233
1234          ngp_xz_int(i) = (nzt_l - nzb_l + 2) * (nxr_l - nxl_l + 3)
1235          ngp_yz_int(i) = (nzt_l - nzb_l + 2) * (nyn_l - nys_l + 3)
1236!
1237!--       MPI data type for REAL arrays, for xz-layers
1238          CALL MPI_TYPE_VECTOR( nxr_l-nxl_l+3, nzt_l-nzb_l+2, ngp_yz(i),       &
1239                                MPI_REAL, type_xz(i), ierr )
1240          CALL MPI_TYPE_COMMIT( type_xz(i), ierr )
1241
1242!
1243!--       MPI data type for INTEGER arrays, for xz-layers
1244          CALL MPI_TYPE_VECTOR( nxr_l-nxl_l+3, nzt_l-nzb_l+2, ngp_yz_int(i),   &
1245                                MPI_INTEGER, type_xz_int(i), ierr )
1246          CALL MPI_TYPE_COMMIT( type_xz_int(i), ierr )
1247
1248!
1249!--       MPI data type for REAL arrays, for yz-layers
1250          CALL MPI_TYPE_VECTOR( 1, ngp_yz(i), ngp_yz(i), MPI_REAL, type_yz(i), &
1251                                ierr )
1252          CALL MPI_TYPE_COMMIT( type_yz(i), ierr )
1253!
1254!--       MPI data type for INTEGER arrays, for yz-layers
1255          CALL MPI_TYPE_VECTOR( 1, ngp_yz_int(i), ngp_yz_int(i), MPI_INTEGER,  &
1256                                type_yz_int(i), ierr )
1257          CALL MPI_TYPE_COMMIT( type_yz_int(i), ierr )
1258
1259
1260!--       For 2D-exchange of INTEGER arrays on coarser grid level, where 2 ghost
1261!--       points need to be exchanged. Only required for 32-bit Integer arrays.
1262          CALL MPI_TYPE_VECTOR( nxr_l-nxl_l+5, 2, nyn_l-nys_l+5, MPI_INTEGER,  &
1263                                type_x_int(i), ierr )
1264          CALL MPI_TYPE_COMMIT( type_x_int(i), ierr )
1265
1266
1267          CALL MPI_TYPE_VECTOR( 2, nyn_l-nys_l+5, nyn_l-nys_l+5, MPI_INTEGER,  &
1268                                type_y_int(i), ierr )
1269          CALL MPI_TYPE_COMMIT( type_y_int(i), ierr )
1270
1271          nxl_l = nxl_l / 2
1272          nxr_l = nxr_l / 2
1273          nys_l = nys_l / 2
1274          nyn_l = nyn_l / 2
1275          nzt_l = nzt_l / 2
1276
1277       ENDDO
1278
1279    ENDIF
1280
1281#endif
1282
1283#if defined( __parallel )
1284!
1285!-- Setting of flags for inflow/outflow/nesting conditions.
1286    IF ( pleft == MPI_PROC_NULL )  THEN
1287       IF ( bc_lr == 'dirichlet/radiation'  .OR.  bc_lr == 'nested'  .OR.      &
1288            bc_lr == 'nesting_offline' )  THEN
1289          bc_dirichlet_l  = .TRUE.
1290       ELSEIF ( bc_lr == 'radiation/dirichlet' )  THEN
1291          bc_radiation_l = .TRUE.
1292       ENDIF
1293    ENDIF
1294 
1295    IF ( pright == MPI_PROC_NULL )  THEN
1296       IF ( bc_lr == 'dirichlet/radiation' )  THEN
1297          bc_radiation_r = .TRUE.
1298       ELSEIF ( bc_lr == 'radiation/dirichlet'  .OR.  bc_lr == 'nested'  .OR.  &
1299                bc_lr == 'nesting_offline' )  THEN
1300          bc_dirichlet_r  = .TRUE.
1301       ENDIF
1302    ENDIF
1303
1304    IF ( psouth == MPI_PROC_NULL )  THEN
1305       IF ( bc_ns == 'dirichlet/radiation' )  THEN
1306          bc_radiation_s = .TRUE.
1307       ELSEIF ( bc_ns == 'radiation/dirichlet'  .OR.  bc_ns == 'nested'  .OR.  &
1308                bc_ns == 'nesting_offline' )  THEN
1309          bc_dirichlet_s  = .TRUE.
1310       ENDIF
1311    ENDIF
1312
1313    IF ( pnorth == MPI_PROC_NULL )  THEN
1314       IF ( bc_ns == 'dirichlet/radiation'  .OR.  bc_ns == 'nested'  .OR.      &
1315            bc_ns == 'nesting_offline' )  THEN
1316          bc_dirichlet_n  = .TRUE.
1317       ELSEIF ( bc_ns == 'radiation/dirichlet' )  THEN
1318          bc_radiation_n = .TRUE.
1319       ENDIF
1320    ENDIF
1321!
1322!-- In case of synthetic turbulence geneartor determine ids.
1323!-- Please note, if no forcing or nesting is applied, the generator is applied
1324!-- only at the left lateral boundary.
1325    IF ( use_syn_turb_gen )  THEN
1326       IF ( bc_dirichlet_l )  THEN
1327          id_stg_left_l = myidx
1328       ELSE
1329          id_stg_left_l = 0
1330       ENDIF
1331       IF ( bc_dirichlet_r )  THEN
1332          id_stg_right_l = myidx
1333       ELSE
1334          id_stg_right_l = 0
1335       ENDIF
1336       IF ( bc_dirichlet_s )  THEN
1337          id_stg_south_l = myidy
1338       ELSE
1339          id_stg_south_l = 0
1340       ENDIF
1341       IF ( bc_dirichlet_n )  THEN
1342          id_stg_north_l = myidy
1343       ELSE
1344          id_stg_north_l = 0
1345       ENDIF
1346
1347       IF ( collective_wait )  CALL MPI_BARRIER( comm2d, ierr )
1348       CALL MPI_ALLREDUCE( id_stg_left_l, id_stg_left,   1, MPI_INTEGER,       &
1349                           MPI_SUM, comm1dx, ierr )
1350
1351       IF ( collective_wait )  CALL MPI_BARRIER( comm2d, ierr )
1352       CALL MPI_ALLREDUCE( id_stg_right_l, id_stg_right, 1, MPI_INTEGER,       &
1353                           MPI_SUM, comm1dx, ierr )
1354
1355       IF ( collective_wait )  CALL MPI_BARRIER( comm2d, ierr )
1356       CALL MPI_ALLREDUCE( id_stg_south_l, id_stg_south, 1, MPI_INTEGER,       &
1357                           MPI_SUM, comm1dy, ierr )
1358
1359       IF ( collective_wait )  CALL MPI_BARRIER( comm2d, ierr )
1360       CALL MPI_ALLREDUCE( id_stg_north_l, id_stg_north, 1, MPI_INTEGER,       &
1361                           MPI_SUM, comm1dy, ierr )
1362
1363    ENDIF
1364 
1365!
1366!-- Broadcast the id of the inflow PE
1367    IF ( bc_dirichlet_l )  THEN
1368       id_inflow_l = myidx
1369    ELSE
1370       id_inflow_l = 0
1371    ENDIF
1372    IF ( collective_wait )  CALL MPI_BARRIER( comm2d, ierr )
1373    CALL MPI_ALLREDUCE( id_inflow_l, id_inflow, 1, MPI_INTEGER, MPI_SUM, &
1374                        comm1dx, ierr )
1375
1376!
1377!-- Broadcast the id of the recycling plane
1378!-- WARNING: needs to be adjusted in case of inflows other than from left side!
1379    IF ( turbulent_inflow ) THEN
1380   
1381       IF ( NINT( recycling_width / dx ) >= nxl  .AND. &
1382            NINT( recycling_width / dx ) <= nxr )  THEN
1383          id_recycling_l = myidx
1384       ELSE
1385          id_recycling_l = 0
1386       ENDIF
1387       IF ( collective_wait )  CALL MPI_BARRIER( comm2d, ierr )
1388       CALL MPI_ALLREDUCE( id_recycling_l, id_recycling, 1, MPI_INTEGER, MPI_SUM, &
1389                           comm1dx, ierr )
1390                           
1391    ENDIF
1392
1393!
1394!-- Broadcast the id of the outflow PE and outflow-source plane
1395    IF ( turbulent_outflow )  THEN
1396
1397       IF ( bc_radiation_r )  THEN
1398          id_outflow_l = myidx
1399       ELSE
1400          id_outflow_l = 0
1401       ENDIF
1402       IF ( collective_wait )  CALL MPI_BARRIER( comm2d, ierr )
1403       CALL MPI_ALLREDUCE( id_outflow_l, id_outflow, 1, MPI_INTEGER, MPI_SUM, &
1404                           comm1dx, ierr )
1405
1406       IF ( NINT( outflow_source_plane / dx ) >= nxl  .AND. &
1407            NINT( outflow_source_plane / dx ) <= nxr )  THEN
1408          id_outflow_source_l = myidx
1409       ELSE
1410          id_outflow_source_l = 0
1411       ENDIF
1412       IF ( collective_wait )  CALL MPI_BARRIER( comm2d, ierr )
1413       CALL MPI_ALLREDUCE( id_outflow_source_l, id_outflow_source, 1, &
1414                           MPI_INTEGER, MPI_SUM, comm1dx, ierr )
1415
1416    ENDIF
1417
1418    CALL location_message( 'creating virtual PE grids + MPI derived data types', 'finished' )
1419
1420#else
1421    IF ( bc_lr == 'dirichlet/radiation' )  THEN
1422       bc_dirichlet_l = .TRUE.
1423       bc_radiation_r = .TRUE.
1424    ELSEIF ( bc_lr == 'radiation/dirichlet' )  THEN
1425       bc_radiation_l = .TRUE.
1426       bc_dirichlet_r = .TRUE.
1427    ENDIF
1428
1429    IF ( bc_ns == 'dirichlet/radiation' )  THEN
1430       bc_dirichlet_n = .TRUE.
1431       bc_radiation_s = .TRUE.
1432    ELSEIF ( bc_ns == 'radiation/dirichlet' )  THEN
1433       bc_radiation_n = .TRUE.
1434       bc_dirichlet_s = .TRUE.
1435    ENDIF
1436#endif
1437
1438!
1439!-- At the inflow or outflow, u or v, respectively, have to be calculated for
1440!-- one more grid point.
1441    IF ( bc_dirichlet_l  .OR.  bc_radiation_l )  THEN
1442       nxlu = nxl + 1
1443    ELSE
1444       nxlu = nxl
1445    ENDIF
1446    IF ( bc_dirichlet_s  .OR.  bc_radiation_s )  THEN
1447       nysv = nys + 1
1448    ELSE
1449       nysv = nys
1450    ENDIF
1451
1452!
1453!-- Allocate wall flag arrays used in the multigrid solver
1454    IF ( psolver(1:9) == 'multigrid' )  THEN
1455
1456       DO  i = maximum_grid_level, 1, -1
1457
1458           SELECT CASE ( i )
1459
1460              CASE ( 1 )
1461                 ALLOCATE( wall_flags_1(nzb:nzt_mg(i)+1,         &
1462                                        nys_mg(i)-1:nyn_mg(i)+1, &
1463                                        nxl_mg(i)-1:nxr_mg(i)+1) )
1464
1465              CASE ( 2 )
1466                 ALLOCATE( wall_flags_2(nzb:nzt_mg(i)+1,         &
1467                                        nys_mg(i)-1:nyn_mg(i)+1, &
1468                                        nxl_mg(i)-1:nxr_mg(i)+1) )
1469
1470              CASE ( 3 )
1471                 ALLOCATE( wall_flags_3(nzb:nzt_mg(i)+1,         &
1472                                        nys_mg(i)-1:nyn_mg(i)+1, &
1473                                        nxl_mg(i)-1:nxr_mg(i)+1) )
1474
1475              CASE ( 4 )
1476                 ALLOCATE( wall_flags_4(nzb:nzt_mg(i)+1,         &
1477                                        nys_mg(i)-1:nyn_mg(i)+1, &
1478                                        nxl_mg(i)-1:nxr_mg(i)+1) )
1479
1480              CASE ( 5 )
1481                 ALLOCATE( wall_flags_5(nzb:nzt_mg(i)+1,         &
1482                                        nys_mg(i)-1:nyn_mg(i)+1, &
1483                                        nxl_mg(i)-1:nxr_mg(i)+1) )
1484
1485              CASE ( 6 )
1486                 ALLOCATE( wall_flags_6(nzb:nzt_mg(i)+1,         &
1487                                        nys_mg(i)-1:nyn_mg(i)+1, &
1488                                        nxl_mg(i)-1:nxr_mg(i)+1) )
1489
1490              CASE ( 7 )
1491                 ALLOCATE( wall_flags_7(nzb:nzt_mg(i)+1,         &
1492                                        nys_mg(i)-1:nyn_mg(i)+1, &
1493                                        nxl_mg(i)-1:nxr_mg(i)+1) )
1494
1495              CASE ( 8 )
1496                 ALLOCATE( wall_flags_8(nzb:nzt_mg(i)+1,         &
1497                                        nys_mg(i)-1:nyn_mg(i)+1, &
1498                                        nxl_mg(i)-1:nxr_mg(i)+1) )
1499
1500              CASE ( 9 )
1501                 ALLOCATE( wall_flags_9(nzb:nzt_mg(i)+1,         &
1502                                        nys_mg(i)-1:nyn_mg(i)+1, &
1503                                        nxl_mg(i)-1:nxr_mg(i)+1) )
1504
1505              CASE ( 10 )
1506                 ALLOCATE( wall_flags_10(nzb:nzt_mg(i)+1,        &
1507                                        nys_mg(i)-1:nyn_mg(i)+1, &
1508                                        nxl_mg(i)-1:nxr_mg(i)+1) )
1509
1510              CASE DEFAULT
1511                 message_string = 'more than 10 multigrid levels'
1512                 CALL message( 'init_pegrid', 'PA0238', 1, 2, 0, 6, 0 )
1513
1514          END SELECT
1515
1516       ENDDO
1517
1518    ENDIF
1519
1520 END SUBROUTINE init_pegrid
Note: See TracBrowser for help on using the repository browser.