source: palm/trunk/SOURCE/palm.f90 @ 2053

Last change on this file since 2053 was 2012, checked in by kanani, 8 years ago

last commit documented

  • Property svn:keywords set to Id
File size: 17.8 KB
RevLine 
[1682]1!> @file palm.f90
[2000]2!------------------------------------------------------------------------------!
[1036]3! This file is part of PALM.
4!
[2000]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.
[1036]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!
[1818]17! Copyright 1997-2016 Leibniz Universitaet Hannover
[2000]18!------------------------------------------------------------------------------!
[1036]19!
[484]20! Current revisions:
[1]21! -----------------
[1834]22!
[2012]23!
[1321]24! Former revisions:
25! -----------------
26! $Id: palm.f90 2012 2016-09-19 17:31:38Z gronemeier $
27!
[2012]28! 2011 2016-09-19 17:29:57Z kanani
29! Flag urban_surface is now defined in module control_parameters.
30!
[2008]31! 2007 2016-08-24 15:47:17Z kanani
32! Temporarily added CALL for writing of restart data for urban surface model
33!
[2001]34! 2000 2016-08-20 18:09:15Z knoop
35! Forced header and separation lines into 80 columns
36!
[1977]37! 1976 2016-07-27 13:28:04Z maronga
38! Added call to radiation_last_actions for binary output of land surface model
39! data
40!
[1973]41! 1972 2016-07-26 07:52:02Z maronga
42! Added call to lsm_last_actions for binary output of land surface model data
43!
[1961]44! 1960 2016-07-12 16:34:24Z suehring
45! Separate humidity and passive scalar
46!
[1933]47! 1834 2016-04-07 14:34:20Z raasch
48! Initial version of purely vertical nesting introduced.
49!
[1834]50! 1833 2016-04-07 14:23:03Z raasch
51! required user interface version changed
52!
[1809]53! 1808 2016-04-05 19:44:00Z raasch
54! routine local_flush replaced by FORTRAN statement
55!
[1784]56! 1783 2016-03-06 18:36:17Z raasch
57! required user interface version changed
58!
[1782]59! 1781 2016-03-03 15:12:23Z raasch
60! pmc initialization moved from time_integration to here
61!
[1780]62! 1779 2016-03-03 08:01:28Z raasch
63! setting of nest_domain and coupling_char moved to the pmci
64!
[1765]65! 1764 2016-02-28 12:45:19Z raasch
66! cpp-statements for nesting removed, communicator settings cleaned up
67!
[1763]68! 1762 2016-02-25 12:31:13Z hellstea
69! Introduction of nested domain feature
70!
[1748]71! 1747 2016-02-08 12:25:53Z raasch
72! OpenACC-adjustment for new surface layer parameterization
73!
[1683]74! 1682 2015-10-07 23:56:08Z knoop
75! Code annotations made doxygen readable
76!
[1669]77! 1668 2015-09-23 13:45:36Z raasch
78! warning replaced by abort in case of failed user interface check
79!
[1667]80! 1666 2015-09-23 07:31:10Z raasch
81! check for user's interface version added
82!
[1483]83! 1482 2014-10-18 12:34:45Z raasch
84! adjustments for using CUDA-aware OpenMPI
85!
[1469]86! 1468 2014-09-24 14:06:57Z maronga
87! Adapted for use on up to 6-digit processor cores
88!
[1403]89! 1402 2014-05-09 14:25:13Z raasch
90! location messages added
91!
[1375]92! 1374 2014-04-25 12:55:07Z raasch
93! bugfix: various modules added
94!
[1321]95! 1320 2014-03-20 08:40:49Z raasch
[1320]96! ONLY-attribute added to USE-statements,
97! kind-parameters added to all INTEGER and REAL declaration statements,
98! kinds are defined in new module kinds,
99! old module precision_kind is removed,
100! revision history before 2012 removed,
101! comment fields (!:) to be used for variable explanations added to
102! all variable declaration statements
[77]103!
[1319]104! 1318 2014-03-17 13:35:16Z raasch
105! module interfaces removed
106!
[1242]107! 1241 2013-10-30 11:36:58Z heinze
108! initialization of nuding and large scale forcing from external file
109!
[1222]110! 1221 2013-09-10 08:59:13Z raasch
111! +wall_flags_00, rflags_invers, rflags_s_inner in copyin statement
112!
[1213]113! 1212 2013-08-15 08:46:27Z raasch
114! +tri in copyin statement
115!
[1182]116! 1179 2013-06-14 05:57:58Z raasch
117! ref_state added to copyin-list
118!
[1114]119! 1113 2013-03-10 02:48:14Z raasch
120! openACC statements modified
121!
[1112]122! 1111 2013-03-08 23:54:10Z raasch
123! openACC statements updated
124!
[1093]125! 1092 2013-02-02 11:24:22Z raasch
126! unused variables removed
127!
[1037]128! 1036 2012-10-22 13:43:42Z raasch
129! code put under GPL (PALM 3.9)
130!
[1017]131! 1015 2012-09-27 09:23:24Z raasch
[863]132! Version number changed from 3.8 to 3.8a.
[1017]133! OpenACC statements added + code changes required for GPU optimization
[863]134!
[850]135! 849 2012-03-15 10:35:09Z raasch
136! write_particles renamed lpm_write_restart_file
137!
[1]138! Revision 1.1  1997/07/24 11:23:35  raasch
139! Initial revision
140!
141!
142! Description:
143! ------------
[1682]144!> Large-Eddy Simulation (LES) model for the convective boundary layer,
145!> optimized for use on parallel machines (implementation realized using the
146!> Message Passing Interface (MPI)). The model can also be run on vector machines
147!> (less well optimized) and workstations. Versions for the different types of
148!> machines are controlled via cpp-directives.
149!> Model runs are only feasible using the ksh-script mrun.
[1976]150!>
151!> @todo create routine last_actions instead of calling lsm_last_actions etc.
[2007]152!> @todo eventually move CALL usm_write_restart_data to suitable location
[1]153!------------------------------------------------------------------------------!
[1682]154 PROGRAM palm
155 
[1]156
[1374]157    USE arrays_3d
[1]158
[1320]159    USE control_parameters,                                                    &
[1781]160        ONLY:  constant_diffusion, coupling_char, coupling_mode,               &
161               do2d_at_begin, do3d_at_begin, humidity, io_blocks, io_group,    &
[1933]162               large_scale_forcing, message_string, nest_domain, neutral,      &
163               nudging, passive_scalar, simulated_time, simulated_time_chr,    &
[2011]164               urban_surface,                                                  &
[1666]165               user_interface_current_revision,                                &
166               user_interface_required_revision, version, wall_heatflux,       &
167               write_binary
[1320]168
169    USE cpulog,                                                                &
170        ONLY:  cpu_log, log_point, cpu_statistics
171
[1374]172    USE grid_variables,                                                        &
173        ONLY:  fxm, fxp, fym, fyp, fwxm, fwxp, fwym, fwyp, wall_e_x, wall_e_y, &
174               wall_u, wall_v, wall_w_x, wall_w_y
175
176    USE indices,                                                               &
[1781]177        ONLY:  nbgp, ngp_2dh, ngp_2dh_s_inner, nzb_diff_s_inner, nzb_diff_s_outer,   &
[1666]178               nzb_diff_u, nzb_diff_v, nzb_s_inner, nzb_s_outer, nzb_u_inner,  &
179               nzb_u_outer, nzb_v_inner, nzb_v_outer, nzb_w_inner,             &
180               nzb_w_outer, rflags_invers, rflags_s_inner, wall_flags_0,       &
181               wall_flags_00
[1374]182
[1320]183    USE kinds
184
[1972]185    USE land_surface_model_mod,                                                &
186        ONLY:  land_surface, lsm_last_actions
187
[1320]188    USE ls_forcing_mod,                                                        &
189        ONLY:  init_ls_forcing
190
191    USE nudge_mod,                                                             &
192        ONLY:  init_nudge
193
194    USE particle_attributes,                                                   &
195        ONLY:  particle_advection
196
[1]197    USE pegrid
198
[1762]199    USE pmc_interface,                                                         &
[1933]200        ONLY:  cpl_id, nested_run, pmci_child_initialize, pmci_init,           &
201               pmci_modelconfiguration, pmci_parent_initialize
[1762]202
[1976]203    USE radiation_model_mod,                                                   &
204        ONLY:  radiation, radiation_last_actions
205
[1374]206    USE statistics,                                                            &
207        ONLY:  hom, rmask, weight_pres, weight_substep
208
[1747]209    USE surface_layer_fluxes_mod,                                              &
210        ONLY:  pt1, qv1, uv_total
[2007]211       
212    USE urban_surface_mod,                                                     &
[2011]213        ONLY:  usm_write_restart_data       
[1747]214
[1015]215#if defined( __openacc )
216    USE OPENACC
217#endif
218
[1]219    IMPLICIT NONE
220
221!
222!-- Local variables
[1682]223    CHARACTER(LEN=9)  ::  time_to_string  !<
224    CHARACTER(LEN=10) ::  env_string      !< to store string of environment var
225    INTEGER(iwp)      ::  env_stat        !< to hold status of GET_ENV
226    INTEGER(iwp)      ::  i               !<
227    INTEGER(iwp)      ::  myid_openmpi    !< OpenMPI local rank for CUDA aware MPI
[1015]228#if defined( __openacc )
[1682]229    REAL(wp), DIMENSION(100) ::  acc_dum     !<
[1015]230#endif
[1]231
[1539]232    version = 'PALM 4.0'
[1833]233    user_interface_required_revision = 'r1819'
[75]234
[1]235#if defined( __parallel )
236!
237!-- MPI initialisation. comm2d is preliminary set, because
238!-- it will be defined in init_pegrid but is used before in cpu_log.
239    CALL MPI_INIT( ierr )
[1762]240
241!
242!-- Initialize the coupling for nested-domain runs
[1764]243!-- comm_palm is the communicator which includes all PEs (MPI processes)
244!-- available for this (nested) model. If it is not a nested run, comm_palm
245!-- is returned as MPI_COMM_WORLD
[1762]246    CALL pmci_init( comm_palm )
247    comm2d = comm_palm
[1764]248!
249!-- Get the (preliminary) number of MPI processes and the local PE-id (in case
250!-- of a further communicator splitting in init_coupling, these numbers will
251!-- be changed in init_pegrid).
252    IF ( nested_run )  THEN
[1762]253
[1764]254       CALL MPI_COMM_SIZE( comm_palm, numprocs, ierr )
255       CALL MPI_COMM_RANK( comm_palm, myid, ierr )
[1762]256
[1764]257    ELSE
258
259       CALL MPI_COMM_SIZE( MPI_COMM_WORLD, numprocs, ierr )
260       CALL MPI_COMM_RANK( MPI_COMM_WORLD, myid, ierr )
[1]261!
[1764]262!--    Initialize PE topology in case of coupled atmosphere-ocean runs (comm_palm
263!--    will be splitted in init_coupling)
264       CALL init_coupling
265    ENDIF
[102]266#endif
267
[1015]268#if defined( __openacc )
[102]269!
[1482]270!-- Get the local MPI rank in case of CUDA aware OpenMPI. Important, if there
271!-- is more than one accelerator board on the node
272    CALL GET_ENVIRONMENT_VARIABLE('OMPI_COMM_WORLD_LOCAL_RANK',                &
273         VALUE=env_string, STATUS=env_stat )
274    READ( env_string, '(I1)' )  myid_openmpi
275    PRINT*, '### local_rank = ', myid_openmpi, '  status=',env_stat
276!
[1015]277!-- Get the number of accelerator boards per node and assign the MPI processes
278!-- to these boards
[1092]279    PRINT*, '*** ACC_DEVICE_NVIDIA = ', ACC_DEVICE_NVIDIA
[1015]280    num_acc_per_node  = ACC_GET_NUM_DEVICES( ACC_DEVICE_NVIDIA )
[1092]281    IF ( numprocs == 1  .AND.  num_acc_per_node > 0 )  num_acc_per_node = 1
[1482]282    PRINT*, '*** myid = ', myid_openmpi, ' num_acc_per_node = ', num_acc_per_node
283    acc_rank = MOD( myid_openmpi, num_acc_per_node )
[1015]284    CALL ACC_SET_DEVICE_NUM ( acc_rank, ACC_DEVICE_NVIDIA )
285!
286!-- Test output (to be removed later)
[1482]287    WRITE (*,'(A,I6,A,I3,A,I3,A,I3)') '*** Connect MPI-Task ', myid_openmpi,   &
288                                      ' to CPU ', acc_rank, ' Devices: ',      &
289                                      num_acc_per_node, ' connected to:',      &
[1015]290                                      ACC_GET_DEVICE_NUM( ACC_DEVICE_NVIDIA )
291#endif
[1092]292
[1015]293!
294!-- Ensure that OpenACC first attaches the GPU devices by copying a dummy data
295!-- region
296    !$acc data copyin( acc_dum )
297
298!
[1]299!-- Initialize measuring of the CPU-time remaining to the run
300    CALL local_tremain_ini
301
302!
303!-- Start of total CPU time measuring.
304    CALL cpu_log( log_point(1), 'total', 'start' )
305    CALL cpu_log( log_point(2), 'initialisation', 'start' )
306
307!
[206]308!-- Open a file for debug output
[1468]309    WRITE (myid_char,'(''_'',I6.6)')  myid
[206]310    OPEN( 9, FILE='DEBUG'//TRIM( coupling_char )//myid_char, FORM='FORMATTED' )
311
312!
[1]313!-- Initialize dvrp logging. Also, one PE maybe split from the global
314!-- communicator for doing the dvrp output. In that case, the number of
315!-- PEs available for PALM is reduced by one and communicator comm_palm
316!-- is changed respectively.
317#if defined( __parallel )
318    CALL MPI_COMM_RANK( comm_palm, myid, ierr )
319!
[102]320!-- TEST OUTPUT (TO BE REMOVED)
321    WRITE(9,*) '*** coupling_mode = "', TRIM( coupling_mode ), '"'
[1808]322    FLUSH( 9 )
[215]323    IF ( TRIM( coupling_mode ) /= 'uncoupled' )  THEN
324       PRINT*, '*** PE', myid, ' Global target PE:', target_id, &
325               TRIM( coupling_mode )
326    ENDIF
[102]327#endif
328
[108]329    CALL init_dvrp_logging
330
[102]331!
[108]332!-- Read control parameters from NAMELIST files and read environment-variables
333    CALL parin
334
335!
[1666]336!-- Check for the user's interface version
337    IF ( user_interface_current_revision /= user_interface_required_revision )  &
338    THEN
339       message_string = 'current user-interface revision "' //                  &
340                        TRIM( user_interface_current_revision ) // '" does ' // &
341                        'not match the required revision ' //                   &
342                        TRIM( user_interface_required_revision )
[1668]343        CALL message( 'palm', 'PA0169', 1, 2, 0, 6, 0 )
[1666]344    ENDIF
345
346!
[108]347!-- Determine processor topology and local array indices
348    CALL init_pegrid
349
350!
[1]351!-- Generate grid parameters
352    CALL init_grid
353
354!
[1241]355!-- Initialize nudging if required
356    IF ( nudging )  THEN
357       CALL init_nudge
358    ENDIF
359
360!
361!-- Initialize reading of large scale forcing from external file - if required
362    IF ( large_scale_forcing )  THEN
363       CALL init_ls_forcing
364    ENDIF
365
366!
[1]367!-- Check control parameters and deduce further quantities
368    CALL check_parameters
369
370!
371!-- Initialize all necessary variables
372    CALL init_3d_model
373
374!
[1762]375!-- Coupling protocol setup for nested-domain runs
[1764]376    IF ( nested_run )  THEN
377       CALL pmci_modelconfiguration
[1781]378!
[1933]379!--    Receive and interpolate initial data on children.
380!--    Child initialization must be made first if the model is both child and
381!--    parent
382       CALL pmci_child_initialize
[1781]383!
[1933]384!--    Send initial condition data from parent to children
385       CALL pmci_parent_initialize
[1781]386!
387!--    Exchange_horiz is needed after the nest initialization
388       IF ( nest_domain )  THEN
389          CALL exchange_horiz( u, nbgp )
390          CALL exchange_horiz( v, nbgp )
391          CALL exchange_horiz( w, nbgp )
[1933]392          IF ( .NOT. neutral )  THEN
393             CALL exchange_horiz( pt, nbgp )
394          ENDIF
[1781]395          IF ( .NOT. constant_diffusion )  CALL exchange_horiz( e, nbgp )
[1960]396          IF ( humidity       )  CALL exchange_horiz( q, nbgp )
397          IF ( passive_scalar )  CALL exchange_horiz( s, nbgp )
[1781]398       ENDIF
399
[1764]400    ENDIF
[1762]401
402!
[1]403!-- Output of program header
404    IF ( myid == 0 )  CALL header
405
406    CALL cpu_log( log_point(2), 'initialisation', 'stop' )
407
408!
409!-- Set start time in format hh:mm:ss
410    simulated_time_chr = time_to_string( simulated_time )
411
412!
413!-- If required, output of initial arrays
414    IF ( do2d_at_begin )  THEN
415       CALL data_output_2d( 'xy', 0 )
416       CALL data_output_2d( 'xz', 0 )
417       CALL data_output_2d( 'yz', 0 )
418    ENDIF
[1976]419
[1]420    IF ( do3d_at_begin )  THEN
421       CALL data_output_3d( 0 )
422    ENDIF
423
424!
[1015]425!-- Declare and initialize variables in the accelerator memory with their
426!-- host values
[1113]427    !$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 )          &
[1212]428    !$acc       copyin( tri, tric, dzu, ddzu, ddzw, dd2zu, l_grid, l_wall, ptdf_x, ptdf_y, pt_init, rdf, rdf_sc, ref_state, ug, u_init, vg, v_init, zu, zw )   &
[1747]429    !$acc       copyin( hom, ol, pt1, qs, qsws, qswst, qv1, rif_wall, shf, ts, tswst, us, usws, uswst, uv_total, vsws, vswst, z0, z0h )      &
[1015]430    !$acc       copyin( fxm, fxp, fym, fyp, fwxm, fwxp, fwym, fwyp, nzb_diff_s_inner, nzb_diff_s_outer, nzb_diff_u )       &
431    !$acc       copyin( nzb_diff_v, nzb_s_inner, nzb_s_outer, nzb_u_inner )    &
432    !$acc       copyin( nzb_u_outer, nzb_v_inner, nzb_v_outer, nzb_w_inner )   &
[1221]433    !$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 )  &
434    !$acc       copyin( ngp_2dh, ngp_2dh_s_inner )  &
[1113]435    !$acc       copyin( weight_pres, weight_substep )
[1015]436!
[495]437!-- Integration of the model equations using timestep-scheme
[1]438    CALL time_integration
439
440!
[495]441!-- If required, write binary data for restart runs
442    IF ( write_binary(1:4) == 'true' )  THEN
[759]443
444       CALL cpu_log( log_point(22), 'write_3d_binary', 'start' )
445
[1402]446       CALL location_message( 'writing restart data', .FALSE. )
447
[759]448       CALL check_open( 14 )
449
450       DO  i = 0, io_blocks-1
451          IF ( i == io_group )  THEN
[1]452!
[759]453!--          Write flow field data
454             CALL write_3d_binary
455          ENDIF
456#if defined( __parallel )
457          CALL MPI_BARRIER( comm2d, ierr )
458#endif
459       ENDDO
460
[1402]461       CALL location_message( 'finished', .TRUE. )
462
[759]463       CALL cpu_log( log_point(22), 'write_3d_binary', 'stop' )
464
[495]465!
466!--    If required, write particle data
[849]467       IF ( particle_advection )  CALL lpm_write_restart_file
[2007]468!
469!--    If required, write urban surface data
470       IF (urban_surface)  CALL usm_write_restart_data
471       
[495]472    ENDIF
[1]473
474!
475!-- If required, repeat output of header including the required CPU-time
476    IF ( myid == 0 )  CALL header
477!
[1972]478!-- If required, final land surface and user-defined actions, and
[1]479!-- last actions on the open files and close files. Unit 14 was opened
480!-- in write_3d_binary but it is closed here, to allow writing on this
481!-- unit in routine user_last_actions.
482    CALL cpu_log( log_point(4), 'last actions', 'start' )
[759]483    DO  i = 0, io_blocks-1
484       IF ( i == io_group )  THEN
[1972]485          IF ( land_surface )  THEN
486             CALL lsm_last_actions
487          ENDIF
[1976]488          IF ( radiation )  THEN
489             CALL radiation_last_actions
490          ENDIF
[759]491          CALL user_last_actions
492          IF ( write_binary(1:4) == 'true' )  CALL close_file( 14 )
493       ENDIF
494#if defined( __parallel )
495       CALL MPI_BARRIER( comm2d, ierr )
496#endif
497    ENDDO
[1]498    CALL close_file( 0 )
499    CALL close_dvrp
500    CALL cpu_log( log_point(4), 'last actions', 'stop' )
501
[102]502#if defined( __mpi2 )
[1]503!
[206]504!-- Test exchange via intercommunicator in case of a MPI-2 coupling
[102]505    IF ( coupling_mode == 'atmosphere_to_ocean' )  THEN
506       i = 12345 + myid
507       CALL MPI_SEND( i, 1, MPI_INTEGER, myid, 11, comm_inter, ierr )
508    ELSEIF ( coupling_mode == 'ocean_to_atmosphere' )  THEN
509       CALL MPI_RECV( i, 1, MPI_INTEGER, myid, 11, comm_inter, status, ierr )
510       PRINT*, '### myid: ', myid, '   received from atmosphere:  i = ', i
511    ENDIF
512#endif
513
514!
[1015]515!-- Close the OpenACC dummy data region
516    !$acc end data
517    !$acc end data
518
519!
[1]520!-- Take final CPU-time for CPU-time analysis
521    CALL cpu_log( log_point(1), 'total', 'stop' )
522    CALL cpu_statistics
523
524#if defined( __parallel )
525    CALL MPI_FINALIZE( ierr )
526#endif
527
528 END PROGRAM palm
Note: See TracBrowser for help on using the repository browser.