source: palm/trunk/SOURCE/palm.f90 @ 1482

 PROGRAM palm

!--------------------------------------------------------------------------------!
! This file is part of PALM.
!
! PALM is free software: you can redistribute it and/or modify it under the terms
! of the GNU General Public License as published by the Free Software Foundation,
! either version 3 of the License, or (at your option) any later version.
!
! PALM is distributed in the hope that it will be useful, but WITHOUT ANY
! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
! A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
!
! You should have received a copy of the GNU General Public License along with
! PALM. If not, see <http://www.gnu.org/licenses/>.
!
! Copyright 1997-2014 Leibniz Universitaet Hannover
!--------------------------------------------------------------------------------!
!
! Current revisions:
! -----------------
! adjustments for using CUDA-aware OpenMPI
! 
! Former revisions:
! -----------------
! $Id: palm.f90 1482 2014-10-18 12:34:45Z raasch $
!
! 1468 2014-09-24 14:06:57Z maronga
! Adapted for use on up to 6-digit processor cores
! 
! 1402 2014-05-09 14:25:13Z raasch
! location messages added
! 
! 1374 2014-04-25 12:55:07Z raasch
! bugfix: various modules added
!
! 1320 2014-03-20 08:40:49Z raasch
! ONLY-attribute added to USE-statements,
! kind-parameters added to all INTEGER and REAL declaration statements,
! kinds are defined in new module kinds,
! old module precision_kind is removed,
! revision history before 2012 removed,
! comment fields (!:) to be used for variable explanations added to
! all variable declaration statements
!
! 1318 2014-03-17 13:35:16Z raasch
! module interfaces removed
!
! 1241 2013-10-30 11:36:58Z heinze
! initialization of nuding and large scale forcing from external file
!
! 1221 2013-09-10 08:59:13Z raasch
! +wall_flags_00, rflags_invers, rflags_s_inner in copyin statement
!
! 1212 2013-08-15 08:46:27Z raasch
! +tri in copyin statement
!
! 1179 2013-06-14 05:57:58Z raasch
! ref_state added to copyin-list
!
! 1113 2013-03-10 02:48:14Z raasch
! openACC statements modified
!
! 1111 2013-03-08 23:54:10Z raasch
! openACC statements updated
!
! 1092 2013-02-02 11:24:22Z raasch
! unused variables removed
!
! 1036 2012-10-22 13:43:42Z raasch
! code put under GPL (PALM 3.9)
!
! 1015 2012-09-27 09:23:24Z raasch
! Version number changed from 3.8 to 3.8a.
! OpenACC statements added + code changes required for GPU optimization
!
! 849 2012-03-15 10:35:09Z raasch
! write_particles renamed lpm_write_restart_file
!
! Revision 1.1  1997/07/24 11:23:35  raasch
! Initial revision
!
!
! Description:
! ------------
! Large-Eddy Simulation (LES) model for the convective boundary layer, 
! optimized for use on parallel machines (implementation realized using the
! Message Passing Interface (MPI)). The model can also be run on vector machines
! (less well optimized) and workstations. Versions for the different types of 
! machines are controlled via cpp-directives.
! Model runs are only feasible using the ksh-script mrun.
!------------------------------------------------------------------------------!

    USE arrays_3d

    USE control_parameters,                                                    &
        ONLY:  coupling_char, coupling_mode, do2d_at_begin, do3d_at_begin,     &
               io_blocks, io_group, large_scale_forcing, nudging,              &
               simulated_time, simulated_time_chr, version, wall_heatflux, write_binary

    USE cpulog,                                                                &
        ONLY:  cpu_log, log_point, cpu_statistics

    USE grid_variables,                                                        &
        ONLY:  fxm, fxp, fym, fyp, fwxm, fwxp, fwym, fwyp, wall_e_x, wall_e_y, &
               wall_u, wall_v, wall_w_x, wall_w_y

    USE indices,                                                               &
        ONLY:  ngp_2dh, ngp_2dh_s_inner, nzb_diff_s_inner, nzb_diff_s_outer, nzb_diff_u, nzb_diff_v,     &
               nzb_s_inner, nzb_s_outer, nzb_u_inner, nzb_u_outer, nzb_v_inner,&
               nzb_v_outer, nzb_w_inner, nzb_w_outer, rflags_invers,           &
               rflags_s_inner, wall_flags_0, wall_flags_00

    USE kinds

    USE ls_forcing_mod,                                                        &
        ONLY:  init_ls_forcing

    USE nudge_mod,                                                             &
        ONLY:  init_nudge

    USE particle_attributes,                                                   &
        ONLY:  particle_advection

    USE pegrid

    USE statistics,                                                            &
        ONLY:  hom, rmask, weight_pres, weight_substep

#if defined( __openacc )
    USE OPENACC
#endif

    IMPLICIT NONE

!
!-- Local variables
    CHARACTER(LEN=9)  ::  time_to_string  !:
    CHARACTER(LEN=10) ::  env_string      !: to store string of environment var
    INTEGER(iwp)      ::  env_stat        !: to hold status of GET_ENV
    INTEGER(iwp)      ::  i               !:
    INTEGER(iwp)      ::  myid_openmpi    !: OpenMPI local rank for CUDA aware MPI
#if defined( __openacc )
    REAL(wp), DIMENSION(100) ::  acc_dum     !:
#endif

    version = 'PALM 3.10'

#if defined( __parallel )
!
!-- MPI initialisation. comm2d is preliminary set, because
!-- it will be defined in init_pegrid but is used before in cpu_log.
    CALL MPI_INIT( ierr )
    CALL MPI_COMM_SIZE( MPI_COMM_WORLD, numprocs, ierr )
    CALL MPI_COMM_RANK( MPI_COMM_WORLD, myid, ierr )
    comm_palm = MPI_COMM_WORLD
    comm2d    = MPI_COMM_WORLD

!
!-- Initialize PE topology in case of coupled runs
    CALL init_coupling
#endif

#if defined( __openacc )
!
!-- Get the local MPI rank in case of CUDA aware OpenMPI. Important, if there
!-- is more than one accelerator board on the node
    CALL GET_ENVIRONMENT_VARIABLE('OMPI_COMM_WORLD_LOCAL_RANK',                &
         VALUE=env_string, STATUS=env_stat )
    READ( env_string, '(I1)' )  myid_openmpi
    PRINT*, '### local_rank = ', myid_openmpi, '  status=',env_stat
!
!-- Get the number of accelerator boards per node and assign the MPI processes
!-- to these boards
    PRINT*, '*** ACC_DEVICE_NVIDIA = ', ACC_DEVICE_NVIDIA
    num_acc_per_node  = ACC_GET_NUM_DEVICES( ACC_DEVICE_NVIDIA )
    IF ( numprocs == 1  .AND.  num_acc_per_node > 0 )  num_acc_per_node = 1
    PRINT*, '*** myid = ', myid_openmpi, ' num_acc_per_node = ', num_acc_per_node
    acc_rank = MOD( myid_openmpi, num_acc_per_node )
    CALL ACC_SET_DEVICE_NUM ( acc_rank, ACC_DEVICE_NVIDIA )
!
!-- Test output (to be removed later)
    WRITE (*,'(A,I6,A,I3,A,I3,A,I3)') '*** Connect MPI-Task ', myid_openmpi,   &
                                      ' to CPU ', acc_rank, ' Devices: ',      &
                                      num_acc_per_node, ' connected to:',      &
                                      ACC_GET_DEVICE_NUM( ACC_DEVICE_NVIDIA )
#endif

!
!-- Ensure that OpenACC first attaches the GPU devices by copying a dummy data
!-- region
    !$acc data copyin( acc_dum )

!
!-- Initialize measuring of the CPU-time remaining to the run
    CALL local_tremain_ini

!
!-- Start of total CPU time measuring.
    CALL cpu_log( log_point(1), 'total', 'start' )
    CALL cpu_log( log_point(2), 'initialisation', 'start' )

!
!-- Open a file for debug output
    WRITE (myid_char,'(''_'',I6.6)')  myid
    OPEN( 9, FILE='DEBUG'//TRIM( coupling_char )//myid_char, FORM='FORMATTED' )

!
!-- Initialize dvrp logging. Also, one PE maybe split from the global
!-- communicator for doing the dvrp output. In that case, the number of
!-- PEs available for PALM is reduced by one and communicator comm_palm
!-- is changed respectively.
#if defined( __parallel )
    CALL MPI_COMM_RANK( comm_palm, myid, ierr )
!
!-- TEST OUTPUT (TO BE REMOVED)
    WRITE(9,*) '*** coupling_mode = "', TRIM( coupling_mode ), '"'
    CALL LOCAL_FLUSH( 9 )
    IF ( TRIM( coupling_mode ) /= 'uncoupled' )  THEN
       PRINT*, '*** PE', myid, ' Global target PE:', target_id, &
               TRIM( coupling_mode )
    ENDIF
#endif

    CALL init_dvrp_logging

!
!-- Read control parameters from NAMELIST files and read environment-variables
    CALL parin

!
!-- Determine processor topology and local array indices
    CALL init_pegrid

!
!-- Generate grid parameters
    CALL init_grid

!
!-- Initialize nudging if required
    IF ( nudging )  THEN
       CALL init_nudge
    ENDIF

!
!-- Initialize reading of large scale forcing from external file - if required
    IF ( large_scale_forcing )  THEN
       CALL init_ls_forcing
    ENDIF

!
!-- Check control parameters and deduce further quantities
    CALL check_parameters

!
!-- Initialize all necessary variables
    CALL init_3d_model

!
!-- Output of program header
    IF ( myid == 0 )  CALL header

    CALL cpu_log( log_point(2), 'initialisation', 'stop' )

!
!-- Set start time in format hh:mm:ss
    simulated_time_chr = time_to_string( simulated_time )

!
!-- If required, output of initial arrays
    IF ( do2d_at_begin )  THEN
       CALL data_output_2d( 'xy', 0 )
       CALL data_output_2d( 'xz', 0 )
       CALL data_output_2d( 'yz', 0 )
    ENDIF
    IF ( do3d_at_begin )  THEN
       CALL data_output_3d( 0 )
    ENDIF

!
!-- Declare and initialize variables in the accelerator memory with their
!-- host values
    !$acc  data copyin( d, diss, e, e_p, kh, km, p, pt, pt_p, q, ql, tend, te_m, tpt_m, tu_m, tv_m, tw_m, u, u_p, v, vpt, v_p, w, w_p )          &
    !$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 )   &
    !$acc       copyin( hom, qs, qsws, qswst, rif, rif_wall, shf, ts, tswst, us, usws, uswst, vsws, vswst, z0, z0h )      &
    !$acc       copyin( fxm, fxp, fym, fyp, fwxm, fwxp, fwym, fwyp, nzb_diff_s_inner, nzb_diff_s_outer, nzb_diff_u )       &
    !$acc       copyin( nzb_diff_v, nzb_s_inner, nzb_s_outer, nzb_u_inner )    &
    !$acc       copyin( nzb_u_outer, nzb_v_inner, nzb_v_outer, nzb_w_inner )   &
    !$acc       copyin( nzb_w_outer, rflags_invers, rflags_s_inner, rmask, wall_heatflux, wall_e_x, wall_e_y, wall_u, wall_v, wall_w_x, wall_w_y, wall_flags_0, wall_flags_00 )  &
    !$acc       copyin( ngp_2dh, ngp_2dh_s_inner )  &
    !$acc       copyin( weight_pres, weight_substep )
!
!-- Integration of the model equations using timestep-scheme
    CALL time_integration

!
!-- If required, write binary data for restart runs
    IF ( write_binary(1:4) == 'true' )  THEN

       CALL cpu_log( log_point(22), 'write_3d_binary', 'start' )

       CALL location_message( 'writing restart data', .FALSE. )

       CALL check_open( 14 )

       DO  i = 0, io_blocks-1
          IF ( i == io_group )  THEN
!
!--          Write flow field data
             CALL write_3d_binary
          ENDIF
#if defined( __parallel )
          CALL MPI_BARRIER( comm2d, ierr )
#endif
       ENDDO

       CALL location_message( 'finished', .TRUE. )

       CALL cpu_log( log_point(22), 'write_3d_binary', 'stop' )

!
!--    If required, write particle data
       IF ( particle_advection )  CALL lpm_write_restart_file
    ENDIF

!
!-- If required, repeat output of header including the required CPU-time
    IF ( myid == 0 )  CALL header
!
!-- If required, final user-defined actions, and
!-- last actions on the open files and close files. Unit 14 was opened
!-- in write_3d_binary but it is closed here, to allow writing on this
!-- unit in routine user_last_actions.
    CALL cpu_log( log_point(4), 'last actions', 'start' )
    DO  i = 0, io_blocks-1
       IF ( i == io_group )  THEN
          CALL user_last_actions
          IF ( write_binary(1:4) == 'true' )  CALL close_file( 14 )
       ENDIF
#if defined( __parallel )
       CALL MPI_BARRIER( comm2d, ierr )
#endif
    ENDDO
    CALL close_file( 0 )
    CALL close_dvrp
    CALL cpu_log( log_point(4), 'last actions', 'stop' )

#if defined( __mpi2 )
!
!-- Test exchange via intercommunicator in case of a MPI-2 coupling
    IF ( coupling_mode == 'atmosphere_to_ocean' )  THEN
       i = 12345 + myid
       CALL MPI_SEND( i, 1, MPI_INTEGER, myid, 11, comm_inter, ierr )
    ELSEIF ( coupling_mode == 'ocean_to_atmosphere' )  THEN
       CALL MPI_RECV( i, 1, MPI_INTEGER, myid, 11, comm_inter, status, ierr )
       PRINT*, '### myid: ', myid, '   received from atmosphere:  i = ', i
    ENDIF
#endif

!
!-- Close the OpenACC dummy data region
    !$acc end data
    !$acc end data

!
!-- Take final CPU-time for CPU-time analysis
    CALL cpu_log( log_point(1), 'total', 'stop' )
    CALL cpu_statistics

#if defined( __parallel )
    CALL MPI_FINALIZE( ierr )
#endif

 END PROGRAM palm