source: palm/trunk/SOURCE/palm.f90 @ 1763

Last change on this file since 1763 was 1763, checked in by hellstea, 5 years ago

last commit documented

  • Property svn:keywords set to Id
File size: 14.5 KB
Line 
1!> @file palm.f90
2!--------------------------------------------------------------------------------!
3! This file is part of PALM.
4!
5! PALM is free software: you can redistribute it and/or modify it under the terms
6! of the GNU General Public License as published by the Free Software Foundation,
7! either version 3 of the License, or (at your option) any later version.
8!
9! PALM is distributed in the hope that it will be useful, but WITHOUT ANY
10! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
11! A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
12!
13! You should have received a copy of the GNU General Public License along with
14! PALM. If not, see <http://www.gnu.org/licenses/>.
15!
16! Copyright 1997-2014 Leibniz Universitaet Hannover
17!--------------------------------------------------------------------------------!
18!
19! Current revisions:
20! -----------------
21!
22!
23! Former revisions:
24! -----------------
25! $Id: palm.f90 1763 2016-02-25 13:00:19Z hellstea $
26!
27! 1762 2016-02-25 12:31:13Z hellstea
28! Introduction of nested domain feature
29!
30! 1747 2016-02-08 12:25:53Z raasch
31! OpenACC-adjustment for new surface layer parameterization
32!
33! 1682 2015-10-07 23:56:08Z knoop
34! Code annotations made doxygen readable
35!
36! 1668 2015-09-23 13:45:36Z raasch
37! warning replaced by abort in case of failed user interface check
38!
39! 1666 2015-09-23 07:31:10Z raasch
40! check for user's interface version added
41!
42! 1482 2014-10-18 12:34:45Z raasch
43! adjustments for using CUDA-aware OpenMPI
44!
45! 1468 2014-09-24 14:06:57Z maronga
46! Adapted for use on up to 6-digit processor cores
47!
48! 1402 2014-05-09 14:25:13Z raasch
49! location messages added
50!
51! 1374 2014-04-25 12:55:07Z raasch
52! bugfix: various modules added
53!
54! 1320 2014-03-20 08:40:49Z raasch
55! ONLY-attribute added to USE-statements,
56! kind-parameters added to all INTEGER and REAL declaration statements,
57! kinds are defined in new module kinds,
58! old module precision_kind is removed,
59! revision history before 2012 removed,
60! comment fields (!:) to be used for variable explanations added to
61! all variable declaration statements
62!
63! 1318 2014-03-17 13:35:16Z raasch
64! module interfaces removed
65!
66! 1241 2013-10-30 11:36:58Z heinze
67! initialization of nuding and large scale forcing from external file
68!
69! 1221 2013-09-10 08:59:13Z raasch
70! +wall_flags_00, rflags_invers, rflags_s_inner in copyin statement
71!
72! 1212 2013-08-15 08:46:27Z raasch
73! +tri in copyin statement
74!
75! 1179 2013-06-14 05:57:58Z raasch
76! ref_state added to copyin-list
77!
78! 1113 2013-03-10 02:48:14Z raasch
79! openACC statements modified
80!
81! 1111 2013-03-08 23:54:10Z raasch
82! openACC statements updated
83!
84! 1092 2013-02-02 11:24:22Z raasch
85! unused variables removed
86!
87! 1036 2012-10-22 13:43:42Z raasch
88! code put under GPL (PALM 3.9)
89!
90! 1015 2012-09-27 09:23:24Z raasch
91! Version number changed from 3.8 to 3.8a.
92! OpenACC statements added + code changes required for GPU optimization
93!
94! 849 2012-03-15 10:35:09Z raasch
95! write_particles renamed lpm_write_restart_file
96!
97! Revision 1.1  1997/07/24 11:23:35  raasch
98! Initial revision
99!
100!
101! Description:
102! ------------
103!> Large-Eddy Simulation (LES) model for the convective boundary layer,
104!> optimized for use on parallel machines (implementation realized using the
105!> Message Passing Interface (MPI)). The model can also be run on vector machines
106!> (less well optimized) and workstations. Versions for the different types of
107!> machines are controlled via cpp-directives.
108!> Model runs are only feasible using the ksh-script mrun.
109!------------------------------------------------------------------------------!
110 PROGRAM palm
111 
112
113    USE arrays_3d
114
115    USE control_parameters,                                                    &
116        ONLY:  coupling_char, coupling_mode, do2d_at_begin, do3d_at_begin,     &
117               io_blocks, io_group, large_scale_forcing, message_string,       &
118               nest_domain, nudging, simulated_time, simulated_time_chr,       &
119               user_interface_current_revision,                                &
120               user_interface_required_revision, version, wall_heatflux,       &
121               write_binary
122
123    USE cpulog,                                                                &
124        ONLY:  cpu_log, log_point, cpu_statistics
125
126    USE grid_variables,                                                        &
127        ONLY:  fxm, fxp, fym, fyp, fwxm, fwxp, fwym, fwyp, wall_e_x, wall_e_y, &
128               wall_u, wall_v, wall_w_x, wall_w_y
129
130    USE indices,                                                               &
131        ONLY:  ngp_2dh, ngp_2dh_s_inner, nzb_diff_s_inner, nzb_diff_s_outer,   &
132               nzb_diff_u, nzb_diff_v, nzb_s_inner, nzb_s_outer, nzb_u_inner,  &
133               nzb_u_outer, nzb_v_inner, nzb_v_outer, nzb_w_inner,             &
134               nzb_w_outer, rflags_invers, rflags_s_inner, wall_flags_0,       &
135               wall_flags_00
136
137    USE kinds
138
139    USE ls_forcing_mod,                                                        &
140        ONLY:  init_ls_forcing
141
142    USE nudge_mod,                                                             &
143        ONLY:  init_nudge
144
145    USE particle_attributes,                                                   &
146        ONLY:  particle_advection
147
148    USE pegrid
149
150#if defined( PMC_ACTIVE )
151    USE pmc_interface,                                                         &
152        ONLY:  cpl_id, pmci_init, pmci_modelconfiguration
153#endif
154
155    USE statistics,                                                            &
156        ONLY:  hom, rmask, weight_pres, weight_substep
157
158    USE surface_layer_fluxes_mod,                                              &
159        ONLY:  pt1, qv1, uv_total
160
161#if defined( __openacc )
162    USE OPENACC
163#endif
164
165    IMPLICIT NONE
166
167!
168!-- Local variables
169    CHARACTER(LEN=9)  ::  time_to_string  !<
170    CHARACTER(LEN=10) ::  env_string      !< to store string of environment var
171    INTEGER(iwp)      ::  env_stat        !< to hold status of GET_ENV
172    INTEGER(iwp)      ::  i               !<
173    INTEGER(iwp)      ::  myid_openmpi    !< OpenMPI local rank for CUDA aware MPI
174#if defined( __openacc )
175    REAL(wp), DIMENSION(100) ::  acc_dum     !<
176#endif
177
178    version = 'PALM 4.0'
179    user_interface_required_revision = 'r1663'
180
181#if defined( __parallel )
182!
183!-- MPI initialisation. comm2d is preliminary set, because
184!-- it will be defined in init_pegrid but is used before in cpu_log.
185    CALL MPI_INIT( ierr )
186
187#if defined( PMC_ACTIVE )
188!
189!-- Initialize the coupling for nested-domain runs
190    CALL pmci_init( comm_palm )
191    comm2d = comm_palm
192
193    IF ( cpl_id >= 2 )  THEN
194       nest_domain = .TRUE.
195       WRITE( coupling_char, '(A1,I1.1)') '_', cpl_id
196    ENDIF
197
198    CALL MPI_COMM_SIZE( comm_palm, numprocs, ierr )
199    CALL MPI_COMM_RANK( comm_palm, myid, ierr )
200#else
201    CALL MPI_COMM_SIZE( MPI_COMM_WORLD, numprocs, ierr )
202    CALL MPI_COMM_RANK( MPI_COMM_WORLD, myid, ierr )
203    comm_palm = MPI_COMM_WORLD
204    comm2d    = MPI_COMM_WORLD
205!
206!-- Initialize PE topology in case of coupled atmosphere-ocean runs (comm_palm
207!-- will be splitted in init_coupling)
208    CALL init_coupling
209#endif
210#endif
211
212#if defined( __openacc )
213!
214!-- Get the local MPI rank in case of CUDA aware OpenMPI. Important, if there
215!-- is more than one accelerator board on the node
216    CALL GET_ENVIRONMENT_VARIABLE('OMPI_COMM_WORLD_LOCAL_RANK',                &
217         VALUE=env_string, STATUS=env_stat )
218    READ( env_string, '(I1)' )  myid_openmpi
219    PRINT*, '### local_rank = ', myid_openmpi, '  status=',env_stat
220!
221!-- Get the number of accelerator boards per node and assign the MPI processes
222!-- to these boards
223    PRINT*, '*** ACC_DEVICE_NVIDIA = ', ACC_DEVICE_NVIDIA
224    num_acc_per_node  = ACC_GET_NUM_DEVICES( ACC_DEVICE_NVIDIA )
225    IF ( numprocs == 1  .AND.  num_acc_per_node > 0 )  num_acc_per_node = 1
226    PRINT*, '*** myid = ', myid_openmpi, ' num_acc_per_node = ', num_acc_per_node
227    acc_rank = MOD( myid_openmpi, num_acc_per_node )
228    CALL ACC_SET_DEVICE_NUM ( acc_rank, ACC_DEVICE_NVIDIA )
229!
230!-- Test output (to be removed later)
231    WRITE (*,'(A,I6,A,I3,A,I3,A,I3)') '*** Connect MPI-Task ', myid_openmpi,   &
232                                      ' to CPU ', acc_rank, ' Devices: ',      &
233                                      num_acc_per_node, ' connected to:',      &
234                                      ACC_GET_DEVICE_NUM( ACC_DEVICE_NVIDIA )
235#endif
236
237!
238!-- Ensure that OpenACC first attaches the GPU devices by copying a dummy data
239!-- region
240    !$acc data copyin( acc_dum )
241
242!
243!-- Initialize measuring of the CPU-time remaining to the run
244    CALL local_tremain_ini
245
246!
247!-- Start of total CPU time measuring.
248    CALL cpu_log( log_point(1), 'total', 'start' )
249    CALL cpu_log( log_point(2), 'initialisation', 'start' )
250
251!
252!-- Open a file for debug output
253    WRITE (myid_char,'(''_'',I6.6)')  myid
254    OPEN( 9, FILE='DEBUG'//TRIM( coupling_char )//myid_char, FORM='FORMATTED' )
255
256!
257!-- Initialize dvrp logging. Also, one PE maybe split from the global
258!-- communicator for doing the dvrp output. In that case, the number of
259!-- PEs available for PALM is reduced by one and communicator comm_palm
260!-- is changed respectively.
261#if defined( __parallel )
262    CALL MPI_COMM_RANK( comm_palm, myid, ierr )
263!
264!-- TEST OUTPUT (TO BE REMOVED)
265    WRITE(9,*) '*** coupling_mode = "', TRIM( coupling_mode ), '"'
266    CALL LOCAL_FLUSH( 9 )
267    IF ( TRIM( coupling_mode ) /= 'uncoupled' )  THEN
268       PRINT*, '*** PE', myid, ' Global target PE:', target_id, &
269               TRIM( coupling_mode )
270    ENDIF
271#endif
272
273    CALL init_dvrp_logging
274
275!
276!-- Read control parameters from NAMELIST files and read environment-variables
277    CALL parin
278
279!
280!-- Check for the user's interface version
281    IF ( user_interface_current_revision /= user_interface_required_revision )  &
282    THEN
283       message_string = 'current user-interface revision "' //                  &
284                        TRIM( user_interface_current_revision ) // '" does ' // &
285                        'not match the required revision ' //                   &
286                        TRIM( user_interface_required_revision )
287        CALL message( 'palm', 'PA0169', 1, 2, 0, 6, 0 )
288    ENDIF
289
290!
291!-- Determine processor topology and local array indices
292    CALL init_pegrid
293
294!
295!-- Generate grid parameters
296    CALL init_grid
297
298!
299!-- Initialize nudging if required
300    IF ( nudging )  THEN
301       CALL init_nudge
302    ENDIF
303
304!
305!-- Initialize reading of large scale forcing from external file - if required
306    IF ( large_scale_forcing )  THEN
307       CALL init_ls_forcing
308    ENDIF
309
310!
311!-- Check control parameters and deduce further quantities
312    CALL check_parameters
313
314!
315!-- Initialize all necessary variables
316    CALL init_3d_model
317
318#if defined( PMC_ACTIVE )
319!
320!-- Coupling protocol setup for nested-domain runs
321    CALL pmci_modelconfiguration
322#endif
323
324!
325!-- Output of program header
326    IF ( myid == 0 )  CALL header
327
328    CALL cpu_log( log_point(2), 'initialisation', 'stop' )
329
330!
331!-- Set start time in format hh:mm:ss
332    simulated_time_chr = time_to_string( simulated_time )
333
334!
335!-- If required, output of initial arrays
336    IF ( do2d_at_begin )  THEN
337       CALL data_output_2d( 'xy', 0 )
338       CALL data_output_2d( 'xz', 0 )
339       CALL data_output_2d( 'yz', 0 )
340    ENDIF
341    IF ( do3d_at_begin )  THEN
342       CALL data_output_3d( 0 )
343    ENDIF
344
345!
346!-- Declare and initialize variables in the accelerator memory with their
347!-- host values
348    !$acc  data copyin( d, diss, e, e_p, kh, km, p, pt, pt_p, q, ql, tend, te_m, tpt_m, tu_m, tv_m, tw_m, u, u_p, v, vpt, v_p, w, w_p )          &
349    !$acc       copyin( tri, tric, dzu, ddzu, ddzw, dd2zu, l_grid, l_wall, ptdf_x, ptdf_y, pt_init, rdf, rdf_sc, ref_state, ug, u_init, vg, v_init, zu, zw )   &
350    !$acc       copyin( hom, ol, pt1, qs, qsws, qswst, qv1, rif_wall, shf, ts, tswst, us, usws, uswst, uv_total, vsws, vswst, z0, z0h )      &
351    !$acc       copyin( fxm, fxp, fym, fyp, fwxm, fwxp, fwym, fwyp, nzb_diff_s_inner, nzb_diff_s_outer, nzb_diff_u )       &
352    !$acc       copyin( nzb_diff_v, nzb_s_inner, nzb_s_outer, nzb_u_inner )    &
353    !$acc       copyin( nzb_u_outer, nzb_v_inner, nzb_v_outer, nzb_w_inner )   &
354    !$acc       copyin( nzb_w_outer, rflags_invers, rflags_s_inner, rmask, wall_heatflux, wall_e_x, wall_e_y, wall_u, wall_v, wall_w_x, wall_w_y, wall_flags_0, wall_flags_00 )  &
355    !$acc       copyin( ngp_2dh, ngp_2dh_s_inner )  &
356    !$acc       copyin( weight_pres, weight_substep )
357!
358!-- Integration of the model equations using timestep-scheme
359    CALL time_integration
360
361!
362!-- If required, write binary data for restart runs
363    IF ( write_binary(1:4) == 'true' )  THEN
364
365       CALL cpu_log( log_point(22), 'write_3d_binary', 'start' )
366
367       CALL location_message( 'writing restart data', .FALSE. )
368
369       CALL check_open( 14 )
370
371       DO  i = 0, io_blocks-1
372          IF ( i == io_group )  THEN
373!
374!--          Write flow field data
375             CALL write_3d_binary
376          ENDIF
377#if defined( __parallel )
378          CALL MPI_BARRIER( comm2d, ierr )
379#endif
380       ENDDO
381
382       CALL location_message( 'finished', .TRUE. )
383
384       CALL cpu_log( log_point(22), 'write_3d_binary', 'stop' )
385
386!
387!--    If required, write particle data
388       IF ( particle_advection )  CALL lpm_write_restart_file
389    ENDIF
390
391!
392!-- If required, repeat output of header including the required CPU-time
393    IF ( myid == 0 )  CALL header
394!
395!-- If required, final user-defined actions, and
396!-- last actions on the open files and close files. Unit 14 was opened
397!-- in write_3d_binary but it is closed here, to allow writing on this
398!-- unit in routine user_last_actions.
399    CALL cpu_log( log_point(4), 'last actions', 'start' )
400    DO  i = 0, io_blocks-1
401       IF ( i == io_group )  THEN
402          CALL user_last_actions
403          IF ( write_binary(1:4) == 'true' )  CALL close_file( 14 )
404       ENDIF
405#if defined( __parallel )
406       CALL MPI_BARRIER( comm2d, ierr )
407#endif
408    ENDDO
409    CALL close_file( 0 )
410    CALL close_dvrp
411    CALL cpu_log( log_point(4), 'last actions', 'stop' )
412
413#if defined( __mpi2 )
414!
415!-- Test exchange via intercommunicator in case of a MPI-2 coupling
416    IF ( coupling_mode == 'atmosphere_to_ocean' )  THEN
417       i = 12345 + myid
418       CALL MPI_SEND( i, 1, MPI_INTEGER, myid, 11, comm_inter, ierr )
419    ELSEIF ( coupling_mode == 'ocean_to_atmosphere' )  THEN
420       CALL MPI_RECV( i, 1, MPI_INTEGER, myid, 11, comm_inter, status, ierr )
421       PRINT*, '### myid: ', myid, '   received from atmosphere:  i = ', i
422    ENDIF
423#endif
424
425!
426!-- Close the OpenACC dummy data region
427    !$acc end data
428    !$acc end data
429
430!
431!-- Take final CPU-time for CPU-time analysis
432    CALL cpu_log( log_point(1), 'total', 'stop' )
433    CALL cpu_statistics
434
435#if defined( __parallel )
436    CALL MPI_FINALIZE( ierr )
437#endif
438
439 END PROGRAM palm
Note: See TracBrowser for help on using the repository browser.