Home

Context Navigation

source: palm/trunk/SOURCE/data_output_particle_mod.f90 @ 4798

Last change on this file since 4798 was 4792, checked in by raasch, 4 years ago
files re-formatted to follow the PALM coding standard
Property svn:keywords set to `Id`
File size: 88.8 KB

Line
1	!> @file data_output_particle_mod.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 terms of the GNU General
6	! Public License as published by the Free Software Foundation, either version 3 of the License, or
7	! (at your option) any later version.
8	!
9	! PALM is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the
10	! implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
11	! Public License for more details.
12	!
13	! You should have received a copy of the GNU General Public License along with PALM. If not, see
14	! <http://www.gnu.org/licenses/>.
15	!
16	! Copyright 1997-2020 Leibniz Universitaet Hannover
17	!--------------------------------------------------------------------------------------------------!
18	!
19	! Current revisions:
20	! -----------------
21	!
22	!
23	! Former revisions:
24	! -----------------
25	! $Id: data_output_particle_mod.f90 4792 2020-11-23 13:02:38Z schwenkel $
26	! file re-formatted to follow the PALM coding standard
27	!
28	! 4779 2020-11-09 17:45:22Z suehring
29	! Avoid overlong lines and unused variables
30	!
31	! 4778 2020-11-09 13:40:05Z raasch
32	! Initial implementation (K. Ketelsen)
33	!
34	!
35	!--------------------------------------------------------------------------------------------------!
36	! Description:
37	! ------------
38	!> Output of particle time series
39	!--------------------------------------------------------------------------------------------------!
40	MODULE data_output_particle_mod
41
42	#if defined( __parallel )
43	USE MPI
44	#endif
45
46	#if defined( __netcdf4 )
47	USE NETCDF
48	#endif
49
50	USE, INTRINSIC :: ISO_C_BINDING
51
52	USE kinds, &
53	ONLY: dp, idp, isp, iwp, sp, wp
54
55	USE indices, &
56	ONLY: nbgp, nnx, nny, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, nz, nzb, nzt
57
58	USE control_parameters, &
59	ONLY: dt_dopts, end_time, simulated_time
60
61	USE pegrid, &
62	ONLY: comm1dx, comm1dy, comm2d, myid, myidx, myidy, npex, npey, numprocs, pdims
63
64	USE particle_attributes, &
65	ONLY: data_output_pts, grid_particles, grid_particle_def, number_of_output_particles, &
66	particles, particle_type, prt_count, pts_id_file, pts_increment, pts_percentage, &
67	oversize, unlimited_dimension
68
69	USE shared_memory_io_mod, &
70	ONLY: sm_class
71
72	USE data_output_netcdf4_module, &
73	ONLY: netcdf4_finalize, netcdf4_get_error_message, netcdf4_init_dimension, &
74	netcdf4_init_module, netcdf4_init_variable, netcdf4_open_file, &
75	netcdf4_stop_file_header_definition, netcdf4_write_attribute, netcdf4_write_variable
76
77	USE cpulog, &
78	ONLY: cpu_log, log_point_s
79
80	IMPLICIT NONE
81
82	PRIVATE
83	SAVE
84
85
86	LOGICAL, PUBLIC :: dop_active = .FALSE.
87
88	!
89	!-- Variables for restart
90	INTEGER(iwp), PUBLIC :: dop_prt_axis_dimension
91	INTEGER(iwp), PUBLIC :: dop_last_active_particle
92
93	!
94	!-- kk Private module variables can be declared inside the submodule
95	INTEGER, PARAMETER :: max_nr_variables = 128
96	INTEGER, PARAMETER :: nr_fixed_variables = 16
97
98	CHARACTER(LEN=32) :: file_name !< Name of NetCDF file
99
100	INTEGER(iwp) :: end_local_numbering
101	INTEGER(iwp) :: io_end_index
102	INTEGER(iwp) :: io_start_index !< Start index of the output area on IO thread
103	INTEGER(iwp) :: nr_particles_pe !< Number of particles assigned for output on this thread
104	INTEGER(iwp) :: nr_particles_rest !< Numbere of rest Particle (ireg. distribution)
105	INTEGER(iwp) :: nr_particles_out !< Total number od particles assigned for output
106	INTEGER(iwp) :: nr_time_values !< Number of values on time axis
107	INTEGER(iwp) :: pe_end_index
108	INTEGER(iwp) :: pe_start_index !< Start index of the output area on this thread
109	INTEGER(iwp) :: start_local_numbering !< Start increment 1 numbering on this thread
110
111	INTEGER(iwp), ALLOCATABLE, DIMENSION(:,:) :: io_indices !< Indices on IO processes
112	INTEGER(iwp), ALLOCATABLE, DIMENSION(:,:) :: mo_indices !< Indices for model communicator
113	INTEGER(iwp), ALLOCATABLE, DIMENSION(:,:) :: remote_nr_particles
114	INTEGER(iwp), ALLOCATABLE, DIMENSION(:,:) :: rma_particles !< Start address and number of remote particles
115	INTEGER(iwp), ALLOCATABLE, DIMENSION(:,:) :: sh_indices !< Indices in shared memory group
116
117	LOGICAL :: irregular_distribubtion !< irregular distribution of output particlesexit
118
119	REAL(sp), ALLOCATABLE, DIMENSION(:) :: time_axis_values !< time axis Values
120
121	TYPE var_def
122	INTEGER(iwp) :: var_id
123	CHARACTER(LEN=32) :: name
124	CHARACTER(LEN=32) :: units
125	LOGICAL :: is_integer = .FALSE.
126	END TYPE var_def
127
128	TYPE(sm_class) :: part_io !< manage communicator for particle IO
129
130	TYPE(var_def), DIMENSION(nr_fixed_variables) :: fix_variables
131	TYPE(var_def), DIMENSION(max_nr_variables) :: variables
132
133	!
134	!-- NetCDF
135	INTEGER(iwp) :: file_id = -1 !< id of Netcdf file
136	INTEGER(iwp) :: nr_fix_variables !< Number of fixed variables scheduled for output
137	INTEGER(iwp) :: nr_variables !< Number of variables scheduled for output
138
139	TYPE dimension_id
140	INTEGER(iwp) :: prt
141	INTEGER(iwp) :: time
142	END TYPE dimension_id
143
144	TYPE variable_id
145	INTEGER(iwp) :: prt
146	INTEGER(iwp) :: time
147	END TYPE variable_id
148
149	TYPE(dimension_id) :: did
150	TYPE(variable_id) :: var_id
151
152	!
153	!-- Shared memory buffer
154	INTEGER(iwp) :: win_prt_i = -1 !< integer MPI shared Memory window
155	INTEGER(iwp) :: win_prt_r = -1 !< real MPI shared Memory window
156	INTEGER(iwp), POINTER, CONTIGUOUS, DIMENSION(:) :: out_buf_i !< integer output buffer
157	REAL(sp), POINTER, CONTIGUOUS, DIMENSION(:) :: out_buf_r !< real output buffer
158
159	!
160	!-- Particle list in file
161	INTEGER(idp), ALLOCATABLE, DIMENSION(:) :: part_id_list_file
162
163	LOGICAL :: part_list_in_file
164
165	!
166	!-- RMA window
167	#if defined( __parallel )
168	INTEGER(iwp) :: win_rma_buf_i
169	INTEGER(iwp) :: win_rma_buf_r
170	#endif
171
172	INTEGER(iwp) :: initial_number_of_active_particles
173	INTEGER(iwp), ALLOCATABLE, DIMENSION(:) :: remote_indices !< particle nubmer of the remodte partices,
174	!< used as indices in the output array
175	INTEGER(iwp), POINTER, DIMENSION(:) :: transfer_buffer_i !< rma window to provide data, which can be
176	!< fetch via MPI_Get
177	REAL(sp), POINTER, DIMENSION(:) :: transfer_buffer_r !< Same for REAL
178
179	!
180	!-- Public subroutine Interface
181	INTERFACE dop_init
182	MODULE PROCEDURE dop_init
183	END INTERFACE dop_init
184
185	INTERFACE dop_output_tseries
186	MODULE PROCEDURE dop_output_tseries
187	END INTERFACE dop_output_tseries
188
189	INTERFACE dop_finalize
190	MODULE PROCEDURE dop_finalize
191	END INTERFACE dop_finalize
192
193	#if defined( __parallel )
194	INTERFACE dop_alloc_rma_mem
195	MODULE PROCEDURE dop_alloc_rma_mem_i1
196	MODULE PROCEDURE dop_alloc_rma_mem_r1
197	END INTERFACE dop_alloc_rma_mem
198	#endif
199
200	PUBLIC dop_init, dop_output_tseries, dop_finalize
201	#if defined( __parallel )
202	PUBLIC dop_alloc_rma_mem ! Must be PUBLIC on NEC, although it is only used in submodule
203	#endif
204
205
206	CONTAINS
207
208
209	!--------------------------------------------------------------------------------------------------!
210	! Description:
211	! ------------
212	!
213	!--------------------------------------------------------------------------------------------------!
214	SUBROUTINE dop_init( read_restart )
215	IMPLICIT NONE
216
217	INTEGER(iwp) :: i !<
218	#if defined( __parallel )
219	INTEGER(iwp) :: ierr !< MPI error code
220	#endif
221	INTEGER(iwp) :: nr_local_last_pe !< Number of output particles on myid == numprocs-2
222	INTEGER(idp) :: nr_particles_8 !< Total number of particles in 64 bit
223	INTEGER(iwp) :: nr_particles_local !< total number of particles scheduled for output on this thread
224
225	INTEGER(idp), DIMENSION(0:numprocs-1) :: nr_particles_all_8
226	INTEGER(iwp), DIMENSION(0:numprocs-1) :: nr_particles_all_s
227	INTEGER(iwp), DIMENSION(0:numprocs-1) :: nr_particles_all_r
228
229	INTEGER(iwp), ALLOCATABLE, DIMENSION(:,:) :: io_indices_s
230	INTEGER(iwp), ALLOCATABLE, DIMENSION(:,:) :: mo_indices_s
231	INTEGER(iwp), ALLOCATABLE, DIMENSION(:,:) :: sh_indices_s
232
233	LOGICAL, INTENT(IN) :: read_restart
234
235	REAL(dp) :: xnr_part !< Must be 64 Bit REAL
236
237
238	part_list_in_file = (pts_id_file /= ' ')
239
240	IF ( .NOT. part_list_in_file ) THEN
241	IF ( .NOT. read_restart) THEN
242
243	CALL set_indef_particle_nr
244
245	CALL count_output_particles( nr_particles_local )
246
247	nr_particles_all_s = 0
248	nr_particles_all_s(myid) = nr_particles_local
249
250	#if defined( __parallel )
251	CALL MPI_ALLREDUCE( nr_particles_all_s, nr_particles_all_r, SIZE( nr_particles_all_s ), &
252	MPI_INTEGER, MPI_SUM, comm2d, ierr )
253	#else
254	nr_particles_all_r = nr_particles_all_s
255	#endif
256
257	initial_number_of_active_particles = SUM( nr_particles_all_r )
258
259	start_local_numbering = 1
260	end_local_numbering = nr_particles_all_r(0)
261
262	IF ( myid > 0 ) THEN
263	DO i = 1, numprocs-1
264	start_local_numbering = start_local_numbering+nr_particles_all_r(i-1)
265	end_local_numbering = start_local_numbering+nr_particles_all_r(i)-1
266	IF ( myid == i ) EXIT
267	ENDDO
268	ENDIF
269
270	nr_particles_all_8 = nr_particles_all_r ! Use 64 INTEGER, maybe more than 2 GB particles
271	nr_particles_8 = SUM( nr_particles_all_8 )
272	IF ( nr_particles_8 > 2000000000_idp ) THEN ! This module works only with 32 Bit INTEGER
273	WRITE( 9, * ) 'Number of particles too large ', nr_particles_8 ; FLUSH( 9 )
274	#if defined( __parallel )
275	CALL MPI_ABORT( MPI_COMM_WORLD, 1, ierr )
276	#endif
277	ENDIF
278	nr_particles_out = nr_particles_8 ! Total number of particles scheduled for output
279
280	!
281	!-- Reserve space for additional particle
282	IF ( number_of_output_particles > 0 ) THEN
283	nr_particles_out = MAX( number_of_output_particles, nr_particles_out )
284	ELSEIF ( oversize > 100.0_wp ) THEN
285	xnr_part = nr_particles_out * oversize / 100.0_wp
286	nr_particles_out = xnr_part
287	ENDIF
288	ELSE
289	nr_particles_out = dop_prt_axis_dimension ! Get Number from restart file
290	initial_number_of_active_particles = dop_last_active_particle
291	ENDIF
292
293	ELSE
294	CALL set_indef_particle_nr
295	CALL dop_read_output_particle_list( nr_particles_out )
296	ENDIF
297	dop_prt_axis_dimension = nr_particles_out
298
299	!
300	!-- The number of particles must be at least the number of MPI processes
301	nr_particles_out = MAX( nr_particles_out, numprocs )
302
303
304	nr_particles_pe = ( nr_particles_out + numprocs - 1 ) / numprocs ! Number of paricles scheduled for ouput on this thread
305
306	pe_start_index = myid * nr_particles_pe + 1 !Output numberimng on this thread
307	pe_end_index = MIN( ( myid + 1 ) * nr_particles_pe, nr_particles_out )
308
309	irregular_distribubtion = .FALSE.
310	#if defined( __parallel )
311	!
312	!-- In case of few particles, it can happen that not only the last thread gets fewer output
313	!-- particles. In this case, the local number of particles on thread numprocs-1 will be < 1.
314	!-- If this happens, irregular distribution of output particles will be used.
315	IF ( myid == numprocs-1 ) THEN
316	nr_local_last_pe = pe_end_index-pe_start_index + 1
317	ELSE
318	nr_local_last_pe = 0
319	ENDIF
320	CALL MPI_BCAST( nr_local_last_pe, 1, MPI_INTEGER, numprocs-1, comm2d, ierr )
321	#else
322	nr_local_last_pe = nr_particles_pe
323	#endif
324	IF ( nr_local_last_pe < 1 ) THEN
325	irregular_distribubtion = .TRUE.
326	CALL dop_setup_ireg_distribution
327	ENDIF
328
329
330	IF ( .NOT. read_restart .AND. .NOT. part_list_in_file ) THEN
331	CALL set_particle_number
332	ENDIF
333
334	IF ( part_list_in_file ) THEN
335	CALL dop_find_particle_in_outlist
336	ENDIF
337
338	CALL part_io%sm_init_data_output_particles( )
339
340	ALLOCATE( sh_indices_s(2,0:part_io%sh_npes-1) )
341	ALLOCATE( sh_indices(2,0:part_io%sh_npes-1) )
342
343
344	#if defined( __parallel )
345	sh_indices_s = 0
346	sh_indices_s(1,part_io%sh_rank) = pe_start_index
347	sh_indices_s(2,part_io%sh_rank) = pe_end_index
348
349	CALL MPI_ALLREDUCE( sh_indices_s, sh_indices, 2*part_io%sh_npes, MPI_INTEGER, MPI_SUM, &
350	part_io%comm_shared, ierr )
351
352	io_start_index = sh_indices(1,0) ! output numbering on actual IO thread
353	io_end_index = sh_indices(2,part_io%sh_npes-1)
354	#else
355	io_start_index = pe_start_index ! output numbering
356	io_end_index = pe_end_index
357	#endif
358
359
360	#if defined( __parallel )
361	CALL MPI_BCAST( part_io%io_npes, 1, MPI_INTEGER, 0, part_io%comm_shared, ierr )
362	#endif
363	ALLOCATE( io_indices(2,0:part_io%io_npes-1) )
364	IF ( part_io%iam_io_pe ) THEN
365	ALLOCATE( io_indices_s(2,0:part_io%io_npes-1) )
366
367	io_indices_s = 0
368	io_indices_s(1,part_io%io_rank) = io_start_index
369	io_indices_s(2,part_io%io_rank) = io_end_index
370
371	#if defined( __parallel )
372	CALL MPI_ALLREDUCE( io_indices_s, io_indices, 2 * part_io%io_npes, MPI_INTEGER, MPI_SUM, &
373	part_io%comm_io, ierr )
374	#else
375	io_indices = io_indices_s
376	#endif
377
378	ENDIF
379
380	#if defined( __parallel )
381	CALL MPI_BCAST( io_indices, SIZE( io_indices ), MPI_INTEGER, 0, part_io%comm_shared, ierr )
382	#endif
383
384	ALLOCATE( remote_nr_particles(2,0:numprocs-1) )
385	ALLOCATE( rma_particles(2,0:numprocs-1) )
386
387	ALLOCATE( mo_indices(2,0:numprocs-1) )
388	ALLOCATE( mo_indices_s(2,0:numprocs-1) )
389
390	mo_indices_s = 0
391	mo_indices_s(1,myid) = pe_start_index
392	mo_indices_s(2,myid) = pe_end_index
393
394	#if defined( __parallel )
395	CALL MPI_ALLREDUCE( mo_indices_s, mo_indices, 2 * numprocs, MPI_INTEGER, MPI_SUM, comm2d, ierr )
396	#else
397	mo_indices = mo_indices_s
398	#endif
399
400	!-- Allocate output buffer
401	#if defined( __parallel )
402	CALL part_io%sm_allocate_shared( out_buf_r, io_start_index, io_end_index, win_prt_r )
403	CALL part_io%sm_allocate_shared( out_buf_i, io_start_index, io_end_index, win_prt_i )
404	#else
405	ALLOCATE( out_buf_r(io_start_index:io_end_index) )
406	ALLOCATE( out_buf_i(io_start_index:io_end_index) )
407	#endif
408
409	!
410	!-- NetCDF
411	CALL dop_netcdf_setup( )
412
413	#if defined( __parallel )
414	CALL MPI_BCAST( nr_fix_variables, 1, MPI_INTEGER, 0, part_io%comm_shared, ierr )
415	CALL MPI_BCAST( nr_variables, 1, MPI_INTEGER, 0, part_io%comm_shared, ierr )
416	#endif
417
418	CALL dop_count_remote_particles
419
420	CALL dop_write_fixed_variables
421
422	CALL deallocate_and_free
423
424	CALL dop_output_tseries
425
426	RETURN
427
428	CONTAINS
429
430
431	!--------------------------------------------------------------------------------------------------!
432	! Description:
433	! ------------
434	!
435	!--------------------------------------------------------------------------------------------------!
436	SUBROUTINE dop_setup_ireg_distribution
437
438	IMPLICIT NONE
439
440	nr_particles_pe = ( nr_particles_out ) / numprocs ! Number of paricles scheduled for ouput on this thread
441
442	nr_particles_rest = nr_particles_out - numprocs * nr_particles_pe
443
444	nr_particles_pe = nr_particles_pe + 1
445	IF ( myid < nr_particles_rest ) THEN
446	pe_start_index = myid * nr_particles_pe + 1 ! Output numberimng on this thread
447	pe_end_index = MIN( ( myid + 1 ) * nr_particles_pe, nr_particles_out)
448	ELSE
449	pe_start_index = nr_particles_rest * ( nr_particles_pe ) &
450	+ ( myid - nr_particles_rest ) * ( nr_particles_pe - 1 ) + 1 !Output numberimng on this thread
451	pe_end_index = MIN( pe_start_index + nr_particles_pe - 2, nr_particles_out )
452	ENDIF
453
454
455	END SUBROUTINE dop_setup_ireg_distribution
456
457	END SUBROUTINE dop_init
458
459
460	!--------------------------------------------------------------------------------------------------!
461	! Description:
462	! ------------
463	!> Particle output on selected time steps
464	!--------------------------------------------------------------------------------------------------!
465	SUBROUTINE dop_output_tseries
466
467	IMPLICIT NONE
468
469	INTEGER(iwp) :: i !<
470	INTEGER(iwp) :: return_value !< Return value data_output_netcdf4 .. routines
471	#if defined( __parallel )
472	INTEGER(iwp) :: ierr !< MPI error code
473	#endif
474	INTEGER(iwp), SAVE :: icount=0 !< count output steps
475
476	INTEGER(iwp), DIMENSION(2) :: bounds_origin, bounds_start, value_counts
477
478	REAl(wp), POINTER, CONTIGUOUS, DIMENSION(:) :: my_time
479
480	icount = icount + 1
481
482	CALL dop_delete_particle_number
483
484	IF ( part_list_in_file ) THEN
485
486	CALL dop_find_particle_in_outlist
487	ELSE
488	CALL dop_newly_generated_particles
489	ENDIF
490
491	CALL dop_count_remote_particles
492
493	bounds_origin = 1
494
495	bounds_start(1) = io_start_index
496	bounds_start(2) = icount
497
498	value_counts(1) = io_end_index-io_start_index + 1
499	value_counts(2) = 1
500
501	DO i = 1, nr_variables
502	#if defined( __netcdf4 )
503	IF ( variables(i)%is_integer ) THEN
504	out_buf_i = NF90_FILL_INT
505	ELSE
506	out_buf_r = NF90_FILL_REAL
507	ENDIF
508	#endif
509
510	CALL cpu_log( log_point_s(99), 'dop_fill_out_buf', 'start' )
511	CALL dop_fill_out_buf (variables(i))
512	CALL cpu_log( log_point_s(99), 'dop_fill_out_buf', 'stop' )
513
514	CALL cpu_log( log_point_s(88), 'dop_get_remote_particle', 'start' )
515	#if defined( __parallel )
516	CALL dop_get_remote_particle( variables(i)%is_integer )
517	#endif
518	CALL cpu_log( log_point_s(88), 'dop_get_remote_particle', 'stop' )
519
520	CALL cpu_log( log_point_s(89), 'particle NetCDF output', 'start' )
521	CALL part_io%sm_node_barrier( )
522	IF ( part_io%iam_io_pe ) THEN
523	IF ( variables(i)%is_integer) THEN
524	CALL netcdf4_write_variable( 'parallel', file_id, variables(i)%var_id, bounds_start, &
525	value_counts, bounds_origin, .FALSE., &
526	values_int32_1d=out_buf_i, return_value=return_value )
527	ELSE
528	CALL netcdf4_write_variable( 'parallel', file_id, variables(i)%var_id, bounds_start, &
529	value_counts, bounds_origin, .FALSE., &
530	values_real32_1d=out_buf_r, return_value=return_value)
531	ENDIF
532	ENDIF
533	CALL part_io%sm_node_barrier( )
534	CALL cpu_log( log_point_s(89), 'particle NetCDF output', 'stop' )
535
536	#if defined( __parallel )
537	CALL MPI_BARRIER( comm2d, ierr ) !kk This Barrier is necessary, not sure why
538	#endif
539	ENDDO
540
541	CALL deallocate_and_free
542
543	!
544	!-- Write Time value
545	IF ( myid == 0 ) THEN
546	ALLOCATE( my_time(1) )
547	bounds_start(1) = icount
548	value_counts(1) = 1
549	my_time(1) = simulated_time
550	IF ( unlimited_dimension ) THEN ! For workaround described in dop finalize
551	time_axis_values(icount) = my_time(1)
552	ELSE
553	CALL netcdf4_write_variable( 'parallel', file_id, var_id%time, bounds_start(1:1), &
554	value_counts(1:1), bounds_origin(1:1), .TRUE., &
555	values_realwp_1d=my_time, return_value=return_value )
556	ENDIF
557	DEALLOCATE( my_time )
558	ENDIF
559
560	RETURN
561	END SUBROUTINE dop_output_tseries
562
563	!--------------------------------------------------------------------------------------------------!
564	! Description:
565	! ------------
566	!
567	!--------------------------------------------------------------------------------------------------!
568	SUBROUTINE dop_finalize
569
570	IMPLICIT NONE
571
572	#if defined( __netcdf4 )
573	INTEGER(iwp) :: ierr !< MPI error code
574	#endif
575	INTEGER(iwp) :: return_value !< Return value data_output_netcdf4 .. routines
576	#if defined( __netcdf4 )
577	INTEGER(iwp) :: var_len
578	#endif
579
580	IF ( win_prt_i /= -1 ) THEN
581	CALL part_io%sm_free_shared( win_prt_i )
582	ENDIF
583	IF ( win_prt_r /= -1 ) THEN
584	CALL part_io%sm_free_shared( win_prt_r )
585	ENDIF
586
587	IF ( file_id /= -1 .AND. part_io%iam_io_pe ) THEN
588	CALL netcdf4_finalize( 'parallel', file_id, return_value )
589	file_id = -1
590
591	#if defined( __netcdf4 )
592	!
593	!-- For yet unknown reasons it is not possible to write in parallel mode the time values to
594	!-- NetCDF variable time.
595	!-- This workaround closes the parallel file and writes the time values in sequential mode on
596	!-- PE0.
597	!-- kk This is a real quick and dirty workaround and the problem should be solved in the final
598	!-- version !!!
599	IF ( myid == 0 .AND. unlimited_dimension ) THEN
600	ierr = NF90_OPEN( TRIM( file_name ), NF90_WRITE, file_id )
601	ierr = NF90_INQUIRE_DIMENSION( file_id, did%time, LEN = var_len)
602
603	ierr = NF90_PUT_VAR( file_id, var_id%time, time_axis_values(1:var_len) )
604
605	ierr = NF90_CLOSE( file_id )
606	ENDIF
607	#endif
608	ENDIF
609
610	RETURN
611	END SUBROUTINE dop_finalize
612
613	!
614	!--Private subroutines
615
616	!--------------------------------------------------------------------------------------------------!
617	! Description:
618	! ------------
619	!
620	!--------------------------------------------------------------------------------------------------!
621	!
622	!-- kk Not sure if necessary, but set ALL particle number to -1
623	SUBROUTINE set_indef_particle_nr
624
625	IMPLICIT NONE
626
627	INTEGER(iwp) :: i !<
628	INTEGER(iwp) :: j !<
629	INTEGER(iwp) :: k !<
630	INTEGER(iwp) :: n !<
631
632	DO i = nxl, nxr
633	DO j = nys, nyn
634	DO k = nzb+1, nzt
635	DO n = 1, SIZE( grid_particles(k,j,i)%particles )
636	grid_particles(k,j,i)%particles(n)%particle_nr = -1
637	ENDDO
638	ENDDO
639	ENDDO
640	ENDDO
641
642	RETURN
643	END SUBROUTINE set_indef_particle_nr
644
645	!--------------------------------------------------------------------------------------------------!
646	! Description:
647	! ------------
648	!> Count particles scheduled for output.
649	!> Here pts_increment and pts_percentage are used to select output particles.
650	!--------------------------------------------------------------------------------------------------!
651	SUBROUTINE count_output_particles (pcount)
652
653	IMPLICIT NONE
654
655	INTEGER(iwp) :: i !<
656	INTEGER(iwp) :: j !<
657	INTEGER(iwp) :: k !<
658	INTEGER(iwp) :: n !<
659	INTEGER(iwp) :: n_all !< count all particles for MOD function
660	INTEGER(iwp), INTENT(OUT) :: pcount !<
661
662	REAL(dp) :: finc
663	REAL(dp) :: fcount
664
665
666	pcount = 0
667	IF ( pts_increment == 1 ) THEN
668	pcount = SUM( prt_count )
669	ELSE
670	n_all = 0
671	DO i = nxl, nxr
672	DO j = nys, nyn
673	DO k = nzb+1, nzt
674	DO n = 1, prt_count(k,j,i)
675	IF ( MOD( n_all, pts_increment ) == 0 ) THEN
676	pcount = pcount + 1
677	ENDIF
678	n_all = n_all + 1
679	ENDDO
680	ENDDO
681	ENDDO
682	ENDDO
683	ENDIF
684
685	IF ( pts_percentage < 100.0_wp ) THEN
686
687	finc = pts_percentage / 100
688
689	pcount = 0
690	fcount = 0.0_wp
691
692	DO i = nxl, nxr
693	DO j = nys, nyn
694	DO k = nzb+1, nzt
695	DO n = 1, prt_count(k,j,i)
696	fcount = fcount + finc
697	IF ( pcount < INT( fcount ) ) THEN
698	pcount = pcount + 1
699	ENDIF
700	ENDDO
701	ENDDO
702	ENDDO
703	ENDDO
704
705	ENDIF
706
707	RETURN
708	END SUBROUTINE count_output_particles
709
710	!--------------------------------------------------------------------------------------------------!
711	! Description:
712	! ------------
713	!
714	!--------------------------------------------------------------------------------------------------!
715	SUBROUTINE dop_read_output_particle_list( nr_particles_out )
716
717	IMPLICIT NONE
718
719	INTEGER(idp) :: dummy !<
720	INTEGER(iwp) :: i !<
721	INTEGER(iwp) :: istat !<
722	INTEGER(iwp) :: iu !<
723	INTEGER(iwp), INTENT(OUT) :: nr_particles_out
724
725
726	iu = 345
727	istat = 0
728	nr_particles_out = 0
729
730	OPEN( UNIT=iu, FILE=TRIM( pts_id_file ) ) !kk should be changed to check_open
731
732	!
733	!-- First stridem cout output particle
734	DO WHILE( istat == 0 )
735	READ( iu, *, iostat=istat ) dummy
736	nr_particles_out = nr_particles_out + 1
737	ENDDO
738
739	nr_particles_out = nr_particles_out - 1 ! subtract 1 for end of file read
740
741	ALLOCATE( part_id_list_file(nr_particles_out) )
742
743	REWIND( iu )
744	!
745	!-- Second stride, read particle ids for scheduled output particle.
746	DO i = 1, nr_particles_out
747	READ( iu, * ) part_id_list_file(i)
748	ENDDO
749
750	CLOSE( iu )
751
752	END SUBROUTINE dop_read_output_particle_list
753
754	!--------------------------------------------------------------------------------------------------!
755	! Description:
756	! ------------
757	!> Set output particle number for selected active particles
758	!--------------------------------------------------------------------------------------------------!
759	SUBROUTINE set_particle_number
760
761	IMPLICIT NONE
762
763	INTEGER(iwp) :: i !<
764	INTEGER(iwp) :: j !<
765	INTEGER(iwp) :: k !<
766	INTEGER(iwp) :: n !<
767	INTEGER(iwp) :: n_all !< count all particles for MOD function
768	INTEGER(iwp) :: particle_nr !< output particle number
769	INTEGER(iwp) :: pcount !< local particle count in case of pts_percentage
770
771	REAL(dp) :: fcount !< partical progress in %/100
772	REAL(dp) :: finc !< increment of particle
773
774
775	pcount = 0
776	fcount = 0.0
777
778	particle_nr = start_local_numbering
779	n_all = 0
780	DO i = nxl, nxr
781	DO j = nys, nyn
782	DO k = nzb+1, nzt
783	IF ( pts_increment > 1 ) THEN
784	DO n = 1, prt_count(k,j,i)
785	IF ( MOD( n_all, pts_increment ) == 0 ) THEN
786	grid_particles(k,j,i)%particles(n)%particle_nr = particle_nr
787	particle_nr = particle_nr + 1
788	ELSE
789	grid_particles(k,j,i)%particles(n)%particle_nr = -2
790	ENDIF
791	n_all = n_all + 1
792	ENDDO
793	ELSEIF ( pts_percentage < 100.0_wp ) THEN
794	finc = pts_percentage / 100
795
796	DO n = 1, prt_count(k,j,i)
797	!
798	!-- Eeach particle moves fraction on particle axis (i.e part percent == 80; move 0.8)
799	!-- if increases next whole number, the particle is taken for output.
800	fcount = fcount + finc
801	IF ( pcount < INT( fcount ) ) THEN
802	pcount = pcount + 1
803	grid_particles(k,j,i)%particles(n)%particle_nr = particle_nr
804	particle_nr = particle_nr + 1
805	ELSE
806	grid_particles(k,j,i)%particles(n)%particle_nr = -2
807	ENDIF
808	ENDDO
809	ELSE
810	DO n = 1, prt_count(k,j,i)
811	grid_particles(k,j,i)%particles(n)%particle_nr = particle_nr
812	particle_nr = particle_nr + 1
813	ENDDO
814	ENDIF
815	ENDDO
816	ENDDO
817	ENDDO
818
819	RETURN
820	END SUBROUTINE set_particle_number
821
822	!--------------------------------------------------------------------------------------------------!
823	! Description:
824	! ------------
825	!
826	!--------------------------------------------------------------------------------------------------!
827	SUBROUTINE dop_find_particle_in_outlist
828
829	IMPLICIT NONE
830
831	INTEGER(iwp) :: i !<
832	#if defined( __parallel )
833	INTEGER(iwp) :: ierr !< MPI error code
834	#endif
835	INTEGER(iwp) :: j !<
836	INTEGER(iwp) :: k !<
837	INTEGER(iwp) :: l !<
838	INTEGER(iwp) :: n !<
839	INTEGER(iwp) :: nr_part !<
840	! INTEGER, save :: icount=0
841
842
843	nr_part = 0
844
845	!
846	!-- If there is a long particle output list, for performance reason it may become necessary to
847	!-- optimize the following loop.
848	!
849	!-- Decode the particle id in i,j,k,n.
850	!-- Split serach, i.e search first in i, then in j ...
851	DO i = nxl, nxr
852	DO j = nys, nyn
853	DO k = nzb+1, nzt
854	DO n = 1, prt_count(k,j,i)
855	DO l = 1, SIZE( part_id_list_file )
856	IF ( grid_particles(k,j,i)%particles(n)%id == part_id_list_file(l) ) THEN
857	grid_particles(k,j,i)%particles(n)%particle_nr = l
858	nr_part = nr_part + 1
859	ENDIF
860	ENDDO
861	ENDDO
862	ENDDO
863	ENDDO
864	ENDDO
865
866	#if defined( __parallel )
867	CALL MPI_ALLREDUCE( nr_part, initial_number_of_active_particles, 1, MPI_INTEGER, MPI_SUM, &
868	comm2d, ierr )
869	#else
870	initial_number_of_active_particles = nr_part
871	#endif
872
873	END SUBROUTINE dop_find_particle_in_outlist
874
875	!
876	!-- Netcdf Setup
877	!
878	!--------------------------------------------------------------------------------------------------!
879	! Description:
880	! ------------
881	!> Open NetCDF File DATA_1D_PTS_NETCDF
882	!> Define Dimensions and variables
883	!> Write constant variables
884	!--------------------------------------------------------------------------------------------------!
885	SUBROUTINE dop_netcdf_setup
886
887	IMPLICIT NONE
888
889	INTEGER, PARAMETER :: global_id_in_file = -1
890
891	INTEGER(iwp) :: i
892	INTEGER(iwp) :: fix_ind
893	INTEGER(iwp) :: return_value
894	INTEGER(iwp) :: var_ind
895
896	INTEGER, DIMENSION(2) :: dimension_ids
897
898	LOGICAL :: const_flag
899
900
901	nr_time_values = end_time / dt_dopts + 1 !kk has to be adapted to formular of 3d output
902
903	IF ( part_io%iam_io_pe ) THEN
904	CALL netcdf4_init_module( "", part_io%comm_io, 0, 9, .TRUE., -1 )
905
906	file_name = 'DATA_1D_PTS_NETCDF'
907	#if defined( __parallel )
908	CALL netcdf4_open_file( 'parallel', TRIM( file_name ), file_id, return_value )
909	#else
910	CALL netcdf4_open_file( 'serial', TRIM( file_name ), file_id, return_value )
911	#endif
912
913	!
914	!-- Global attributes
915	CALL netcdf4_write_attribute( 'parallel', file_id, global_id_in_file, 'comment', &
916	'Particle ouput created by PALM module data_output_particle', &
917	return_value=return_value )
918
919	CALL netcdf4_write_attribute( 'parallel', file_id, global_id_in_file, &
920	'initial_nr_particles', &
921	value_int32=initial_number_of_active_particles, &
922	return_value=return_value )
923	!
924	!-- Define dimensions
925	CALL netcdf4_init_dimension( 'parallel', file_id, did%prt, var_id%prt, 'prt','int32' , &
926	nr_particles_out, .TRUE., return_value )
927
928	IF ( unlimited_dimension ) THEN
929	CALL netcdf4_init_dimension( 'parallel', file_id, did%time, var_id%time, 'time','real32',&
930	-1, .TRUE., return_value )
931	ALLOCATE( time_axis_values(nr_time_values) )
932	ELSE
933	CALL netcdf4_init_dimension( 'parallel', file_id, did%time, var_id%time, 'time','real32',&
934	nr_time_values, .TRUE., return_value )
935	ENDIF
936	ENDIF
937	!
938	!-- Variables without time axis
939	!-- These variables will always be written only once at the beginning of the file
940	dimension_ids(1) = did%prt
941
942	fix_ind = 1
943	fix_variables(fix_ind)%name = 'origin_x'
944	fix_variables(fix_ind)%units = 'meter'
945
946	IF ( part_io%iam_io_pe ) THEN
947	CALL netcdf4_init_variable( 'parallel', file_id, fix_variables(fix_ind)%var_id, &
948	fix_variables(fix_ind)%name, 'real32', dimension_ids(1:1), &
949	.FALSE., return_value )
950
951	CALL netcdf4_write_attribute( 'parallel', file_id, fix_variables(fix_ind)%var_id, 'units', &
952	value_char=TRIM( fix_variables(fix_ind)%units ), &
953	return_value=return_value )
954	ENDIF
955
956	fix_ind = fix_ind + 1
957	fix_variables(fix_ind)%name = 'origin_y'
958	fix_variables(fix_ind)%units = 'meter'
959
960	IF ( part_io%iam_io_pe ) THEN
961	CALL netcdf4_init_variable( 'parallel', file_id, fix_variables(fix_ind)%var_id, &
962	fix_variables(fix_ind)%name, 'real32', dimension_ids(1:1), &
963	.FALSE., return_value )
964
965	CALL netcdf4_write_attribute( 'parallel', file_id, fix_variables(fix_ind)%var_id, 'units', &
966	value_char=TRIM( fix_variables(fix_ind)%units ), &
967	return_value=return_value )
968
969	ENDIF
970
971	fix_ind = fix_ind + 1
972	fix_variables(fix_ind)%name = 'origin_z'
973	fix_variables(fix_ind)%units = 'meter'
974
975	IF ( part_io%iam_io_pe ) THEN
976	CALL netcdf4_init_variable( 'parallel', file_id, fix_variables(fix_ind)%var_id, &
977	fix_variables(fix_ind)%name, 'real32', dimension_ids(1:1), &
978	.FALSE., return_value )
979
980	CALL netcdf4_write_attribute( 'parallel', file_id, fix_variables(fix_ind)%var_id, 'units', &
981	value_char=TRIM( fix_variables(fix_ind)%units ), &
982	return_value=return_value )
983	ENDIF
984
985	!
986	!-- These variables are written if name end with _const'
987	DO i = 1, SIZE( data_output_pts )
988	const_flag = ( INDEX( TRIM( data_output_pts(i) ), '_const' ) > 0 )
989	IF ( LEN( TRIM( data_output_pts(i) ) ) > 0 .AND. const_flag) THEN
990	fix_ind = fix_ind + 1
991	fix_variables(fix_ind)%name = TRIM( data_output_pts(i) )
992
993	SELECT CASE ( TRIM( fix_variables(fix_ind)%name ) )
994	CASE ( 'radius_const' )
995	fix_variables(fix_ind)%units = 'meter'
996	fix_variables(fix_ind)%is_integer = .FALSE.
997	CASE ( 'aux1_const' )
998	fix_variables(fix_ind)%units = 'depend_on_setup'
999	fix_variables(fix_ind)%is_integer = .FALSE.
1000	CASE ( 'aux2_const' )
1001	fix_variables(fix_ind)%units = 'depend_on_setup'
1002	fix_variables(fix_ind)%is_integer = .FALSE.
1003	CASE ( 'rvar1_const' )
1004	fix_variables(fix_ind)%units = 'depend_on_setup'
1005	fix_variables(fix_ind)%is_integer = .FALSE.
1006	CASE ( 'rvar2_const' )
1007	fix_variables(fix_ind)%units = 'depend_on_setup'
1008	fix_variables(fix_ind)%is_integer = .FALSE.
1009	CASE ( 'rvar3_const' )
1010	fix_variables(fix_ind)%units = 'depend_on_setup'
1011	fix_variables(fix_ind)%is_integer = .FALSE.
1012	END SELECT
1013
1014	IF ( part_io%iam_io_pe ) THEN
1015	IF ( fix_variables(fix_ind)%is_integer ) THEN
1016	CALL netcdf4_init_variable( 'parallel', file_id, fix_variables(fix_ind)%var_id, &
1017	fix_variables(fix_ind)%name, 'int32', &
1018	dimension_ids(1:1), .FALSE., return_value )
1019	ELSE
1020	CALL netcdf4_init_variable( 'parallel', file_id, fix_variables(fix_ind)%var_id, &
1021	fix_variables(fix_ind)%name, 'real32', &
1022	dimension_ids(1:1), .FALSE., return_value )
1023	ENDIF
1024
1025	CALL netcdf4_write_attribute( 'parallel', file_id, fix_variables(fix_ind)%var_id, &
1026	'units', &
1027	value_char=TRIM( fix_variables(fix_ind)%units ), &
1028	return_value=return_value )
1029	ENDIF
1030
1031	ENDIF
1032	ENDDO
1033
1034	nr_fix_variables = fix_ind
1035	!
1036	!-- Variables time axis
1037	!-- These variables will always be written in the time loop
1038	dimension_ids(1) = did%prt
1039	dimension_ids(2) = did%time
1040
1041	var_ind = 0
1042
1043	DO i = 1, SIZE( data_output_pts )
1044	const_flag = ( INDEX( TRIM( data_output_pts(i) ), '_const' ) > 0 )
1045	IF ( LEN( TRIM( data_output_pts(i) ) ) > 0 .AND. .NOT. const_flag ) THEN
1046	var_ind = var_ind + 1
1047	variables(var_ind)%name = TRIM( data_output_pts(i) )
1048
1049	SELECT CASE ( TRIM( variables(var_ind)%name) )
1050	CASE ( 'id' )
1051	variables(var_ind)%name = TRIM( data_output_pts(i) ) // '_low'
1052	variables(var_ind)%units = 'Number'
1053	variables(var_ind)%is_integer = .TRUE.
1054	IF ( part_io%iam_io_pe ) THEN
1055	CALL netcdf4_init_variable( 'parallel', file_id, variables(var_ind)%var_id, &
1056	variables(var_ind)%name, 'int32', &
1057	dimension_ids(1:2), .FALSE., return_value )
1058	CALL netcdf4_write_attribute( 'parallel', file_id, variables(var_ind)%var_id, &
1059	'units', &
1060	value_char=TRIM( variables(var_ind)%units ), &
1061	return_value=return_value )
1062	ENDIF
1063
1064	var_ind = var_ind + 1
1065	variables(var_ind)%name = TRIM( data_output_pts(i) ) // '_high'
1066	variables(var_ind)%units = 'Number'
1067	variables(var_ind)%is_integer = .TRUE.
1068	CASE ( 'particle_nr' )
1069	variables(var_ind)%units = 'Number'
1070	variables(var_ind)%is_integer = .TRUE.
1071	CASE ( 'class' )
1072	variables(var_ind)%units = 'Number'
1073	variables(var_ind)%is_integer = .TRUE.
1074	CASE ( 'group' )
1075	variables(var_ind)%units = 'Number'
1076	variables(var_ind)%is_integer = .TRUE.
1077	CASE ( 'x' )
1078	variables(var_ind)%units = 'meter'
1079	variables(var_ind)%is_integer = .FALSE.
1080	CASE ( 'y' )
1081	variables(var_ind)%units = 'meter'
1082	variables(var_ind)%is_integer = .FALSE.
1083	CASE ( 'z' )
1084	variables(var_ind)%units = 'meter'
1085	variables(var_ind)%is_integer = .FALSE.
1086	CASE ( 'speed_x' )
1087	variables(var_ind)%units = 'm/s'
1088	variables(var_ind)%is_integer = .FALSE.
1089	CASE ( 'speed_y' )
1090	variables(var_ind)%units = 'm/s'
1091	variables(var_ind)%is_integer = .FALSE.
1092	CASE ( 'speed_z' )
1093	variables(var_ind)%units = 'm/s'
1094	variables(var_ind)%is_integer = .FALSE.
1095	CASE ( 'radius' )
1096	variables(var_ind)%units = 'meter'
1097	variables(var_ind)%is_integer = .FALSE.
1098	CASE ( 'age' )
1099	variables(var_ind)%units = 'sec'
1100	variables(var_ind)%is_integer = .FALSE.
1101	CASE ( 'age_m' )
1102	variables(var_ind)%units = 'sec'
1103	variables(var_ind)%is_integer = .FALSE.
1104	CASE ( 'dt_sum' )
1105	variables(var_ind)%units = 'sec'
1106	variables(var_ind)%is_integer = .FALSE.
1107	CASE ( 'e_m' )
1108	variables(var_ind)%units = 'Ws'
1109	variables(var_ind)%is_integer = .FALSE.
1110	CASE( 'weight_factor' )
1111	variables(var_ind)%units = 'factor'
1112	variables(var_ind)%is_integer = .FALSE.
1113	CASE ( 'aux1' )
1114	variables(var_ind)%units = 'depend_on_setup'
1115	variables(var_ind)%is_integer = .FALSE.
1116	CASE ( 'aux2' )
1117	variables(var_ind)%units = 'depend_on_setup'
1118	variables(var_ind)%is_integer = .FALSE.
1119	CASE ( 'rvar1' )
1120	variables(var_ind)%units = 'depend_on_setup'
1121	variables(var_ind)%is_integer = .FALSE.
1122	CASE ( 'rvar2' )
1123	variables(var_ind)%units = 'depend_on_setup'
1124	variables(var_ind)%is_integer = .FALSE.
1125	CASE ( 'rvar3' )
1126	variables(var_ind)%units = 'depend_on_setup'
1127	variables(var_ind)%is_integer = .FALSE.
1128	END SELECT
1129
1130	IF ( part_io%iam_io_pe ) THEN
1131	IF ( variables(var_ind)%is_integer ) THEN
1132	CALL netcdf4_init_variable( 'parallel', file_id, variables(var_ind)%var_id, &
1133	variables(var_ind)%name, 'int32', dimension_ids(1:2), &
1134	.FALSE., return_value )
1135	ELSE
1136	CALL netcdf4_init_variable( 'parallel', file_id, variables(var_ind)%var_id, &
1137	variables(var_ind)%name, 'real32', dimension_ids(1:2), &
1138	.FALSE., return_value )
1139	ENDIF
1140
1141	CALL netcdf4_write_attribute( 'parallel', file_id, variables(var_ind)%var_id, 'units',&
1142	value_char=TRIM( variables(var_ind)%units ), &
1143	return_value=return_value )
1144	ENDIF
1145
1146	ENDIF
1147	ENDDO
1148
1149	nr_variables = var_ind
1150
1151	IF ( part_io%iam_io_pe ) THEN
1152	CALL netcdf4_stop_file_header_definition( 'parallel', file_id, return_value )
1153	ENDIF
1154
1155	CALL dop_write_axis
1156
1157	RETURN
1158
1159	CONTAINS
1160
1161	!--------------------------------------------------------------------------------------------------!
1162	! Description:
1163	! ------------
1164	!
1165	!--------------------------------------------------------------------------------------------------!
1166	SUBROUTINE dop_write_axis
1167
1168	IMPLICIT NONE
1169
1170	INTEGER(iwp) :: i
1171	INTEGER(iwp) :: return_value !< Return value data_output_netcdf4 .. routines
1172
1173	INTEGER,DIMENSION(1) :: bounds_origin, bounds_start, value_counts
1174
1175	INTEGER, POINTER, CONTIGUOUS, DIMENSION(:) :: prt_val
1176
1177
1178	bounds_origin = 1
1179	bounds_start(1) = 1
1180
1181	IF ( myid == 0 ) THEN
1182
1183	ALLOCATE( prt_val(nr_particles_out) )
1184	DO i = 1, nr_particles_out
1185	prt_val(i) = i
1186	ENDDO
1187	value_counts(1) = nr_particles_out
1188
1189	CALL netcdf4_write_variable( 'parallel', file_id, var_id%prt, bounds_start, &
1190	value_counts, bounds_origin, .TRUE., &
1191	values_int32_1d=prt_val, return_value=return_value)
1192
1193	DEALLOCATE( prt_val )
1194	ENDIF
1195
1196	RETURN
1197	END SUBROUTINE dop_write_axis
1198
1199	END SUBROUTINE dop_netcdf_setup
1200
1201	!--------------------------------------------------------------------------------------------------!
1202	! Description:
1203	! ------------
1204	!> Write constant variables
1205	!--------------------------------------------------------------------------------------------------!
1206	SUBROUTINE dop_write_fixed_variables
1207
1208	IMPLICIT NONE
1209
1210	INTEGER(iwp) :: i !
1211	INTEGER(iwp) :: return_value !
1212
1213	INTEGER(iwp),DIMENSION(1) :: bounds_origin, bounds_start, value_counts
1214
1215
1216	bounds_origin(1) = 1
1217	bounds_start(1) = io_start_index
1218	value_counts(1) = io_end_index - io_start_index + 1
1219
1220
1221	DO i = 1, nr_fix_variables
1222
1223	#if defined( __netcdf4 )
1224	IF ( fix_variables(i)%is_integer ) THEN
1225	out_buf_i = NF90_FILL_INT
1226	ELSE
1227	out_buf_r = NF90_FILL_REAL
1228	ENDIF
1229	#endif
1230
1231	CALL dop_fill_out_buf( fix_variables(i) )
1232
1233	CALL dop_get_remote_particle( fix_variables(i)%is_integer )
1234
1235	CALL part_io%sm_node_barrier( )
1236	IF ( part_io%iam_io_pe ) THEN
1237	IF ( fix_variables(i)%is_integer ) THEN
1238	CALL netcdf4_write_variable( 'parallel', file_id, fix_variables(i)%var_id, &
1239	bounds_start, value_counts, bounds_origin, .FALSE., &
1240	values_int32_1d=out_buf_i, return_value=return_value)
1241	ELSE
1242	CALL netcdf4_write_variable( 'parallel', file_id, fix_variables(i)%var_id, &
1243	bounds_start, value_counts, bounds_origin, .FALSE., &
1244	values_real32_1d=out_buf_r, return_value=return_value)
1245	ENDIF
1246	ENDIF
1247	CALL part_io%sm_node_barrier( )
1248	ENDDO
1249
1250	RETURN
1251	END SUBROUTINE dop_write_fixed_variables
1252
1253	!--------------------------------------------------------------------------------------------------!
1254	! Description:
1255	! ------------
1256	!
1257	!--------------------------------------------------------------------------------------------------!
1258	SUBROUTINE dop_delete_particle_number
1259
1260	IMPLICIT NONE
1261
1262	INTEGER(iwp) :: i !<
1263	INTEGER(iwp) :: j !<
1264	INTEGER(iwp) :: k !<
1265	INTEGER(iwp) :: n !<
1266
1267	!
1268	!-- Delete inactive particles, i.e. all particles with particle_mask == .FALSE..
1269	!-- Get an output particle nr of -1.
1270	!-- kk Not sure if it is required here or already done in lagrangian_particle_model_mod before
1271	!-- this call.
1272	remote_nr_particles = 0
1273	DO i = nxl, nxr
1274	DO j = nys, nyn
1275	DO k = nzb+1, nzt
1276	DO n = 1, prt_count(k,j,i)
1277	IF ( .NOT. grid_particles(k,j,i)%particles(n)%particle_mask ) THEN
1278	grid_particles(k,j,i)%particles(n)%particle_nr = -1
1279	ENDIF
1280	ENDDO
1281	ENDDO
1282	ENDDO
1283	ENDDO
1284
1285	RETURN
1286	END SUBROUTINE dop_delete_particle_number
1287
1288	!--------------------------------------------------------------------------------------------------!
1289	! Description:
1290	! ------------
1291	!
1292	!--------------------------------------------------------------------------------------------------!
1293	SUBROUTINE dop_newly_generated_particles
1294
1295	IMPLICIT NONE
1296
1297	INTEGER(iwp) :: i !<
1298	#if defined( __parallel )
1299	INTEGER(iwp) :: ierr !< MPI error code
1300	#endif
1301	INTEGER(iwp) :: j !<
1302	INTEGER(iwp) :: k !<
1303	INTEGER(iwp) :: n !<
1304	INTEGER(iwp) :: n_all !< count all particles for MOD function
1305	INTEGER(iwp) :: nr_new_particle !<
1306	INTEGER(iwp) :: particle_nr !<
1307	INTEGER(iwp) :: start_new_numbering
1308
1309	INTEGER(iwp),DIMENSION(0:numprocs-1) :: nr_particles_new_r
1310	INTEGER(iwp),DIMENSION(0:numprocs-1) :: nr_particles_new_s
1311
1312	REAL(dp) :: fcount
1313	REAL(dp) :: finc
1314
1315
1316	!
1317	!-- Count Number of Newly Generated particles.
1318	!-- Condition for a newparticle: particle_mask = .TRUE. and particle nr of -1.
1319	!
1320	!-- For performance reasons, this subroutine may be combined later with dop_delete_particle_number.
1321	nr_new_particle = 0
1322
1323	IF ( pts_increment > 1 ) THEN
1324	n_all = 0
1325	DO i = nxl, nxr
1326	DO j = nys, nyn
1327	DO k = nzb+1, nzt
1328	DO n = 1, prt_count(k,j,i)
1329	IF ( grid_particles(k,j,i)%particles(n)%particle_mask ) THEN
1330	IF ( grid_particles(k,j,i)%particles(n)%particle_nr == -1 ) THEN
1331	IF ( MOD( n_all, pts_increment ) == 0 ) THEN
1332	nr_new_particle = nr_new_particle + 1
1333	ENDIF
1334	n_all = n_all + 1
1335	ENDIF
1336	ENDIF
1337	ENDDO
1338	ENDDO
1339	ENDDO
1340	ENDDO
1341	ELSEIF ( pts_percentage < 100.0_wp ) THEN
1342	finc = pts_percentage / 100
1343	fcount = 0.0
1344
1345	DO i = nxl, nxr
1346	DO j = nys, nyn
1347	DO k = nzb+1, nzt
1348	DO n = 1, prt_count(k,j,i)
1349	IF ( grid_particles(k,j,i)%particles(n)%particle_mask ) THEN
1350	IF ( grid_particles(k,j,i)%particles(n)%particle_nr == -1 ) THEN
1351	fcount = fcount + finc
1352	IF ( nr_new_particle < INT( fcount ) ) THEN
1353	nr_new_particle = nr_new_particle + 1
1354	ENDIF
1355	ENDIF
1356	ENDIF
1357	ENDDO
1358	ENDDO
1359	ENDDO
1360	ENDDO
1361
1362	ELSE
1363	DO i = nxl, nxr
1364	DO j = nys, nyn
1365	DO k = nzb+1, nzt
1366	DO n = 1, prt_count(k,j,i)
1367	IF ( grid_particles(k,j,i)%particles(n)%particle_mask ) THEN
1368	IF ( grid_particles(k,j,i)%particles(n)%particle_nr == -1 ) THEN
1369	nr_new_particle = nr_new_particle + 1
1370	ENDIF
1371	ENDIF
1372	ENDDO
1373	ENDDO
1374	ENDDO
1375	ENDDO
1376	ENDIF
1377	!
1378	!-- Determine start number of new particles on eac thread.
1379	nr_particles_new_s = 0
1380	nr_particles_new_s(myid) = nr_new_particle
1381
1382	#if defined( __parallel )
1383	CALL MPI_ALLREDUCE( nr_particles_new_s, nr_particles_new_r, SIZE( nr_particles_new_s ), &
1384	MPI_INTEGER, MPI_SUM, comm2d, ierr )
1385	#else
1386	nr_particles_new_r = nr_particles_new_s
1387	#endif
1388	!
1389	!-- Abortion if selected particles from new particle set would exceed particle axis of output
1390	!-- changed by JS.
1391	IF ( ( SUM( nr_particles_new_r ) + initial_number_of_active_particles ) > nr_particles_out ) &
1392	THEN
1393	RETURN
1394	ENDIF
1395
1396	start_new_numbering = initial_number_of_active_particles + 1
1397
1398	IF ( myid > 0) THEN
1399	DO i = 1, numprocs-1
1400	start_new_numbering = start_new_numbering + nr_particles_new_r(i-1)
1401	IF ( myid == i ) EXIT
1402	ENDDO
1403	ENDIF
1404
1405	initial_number_of_active_particles = initial_number_of_active_particles + &
1406	SUM( nr_particles_new_r )
1407
1408	dop_last_active_particle = initial_number_of_active_particles
1409	!
1410	!-- Set number of new particles
1411	particle_nr = start_new_numbering
1412	nr_new_particle = 0
1413
1414	IF ( pts_increment > 1 ) THEN
1415	n_all = 0
1416	DO i = nxl, nxr
1417	DO j = nys, nyn
1418	DO k = nzb+1, nzt
1419	DO n = 1, prt_count(k,j,i)
1420	IF ( grid_particles(k,j,i)%particles(n)%particle_mask) THEN
1421	IF ( grid_particles(k,j,i)%particles(n)%particle_nr == -1 ) THEN
1422	IF ( MOD( n_all, pts_increment ) == 0 ) THEN
1423	grid_particles(k,j,i)%particles(n)%particle_nr = particle_nr
1424	particle_nr = particle_nr + 1
1425	ELSE
1426	grid_particles(k,j,i)%particles(n)%particle_nr = -2
1427	ENDIF
1428	n_all = n_all + 1
1429	ENDIF
1430	ENDIF
1431	ENDDO
1432	ENDDO
1433	ENDDO
1434	ENDDO
1435	ELSEIF ( pts_percentage < 100.0_wp ) THEN
1436	finc = pts_percentage / 100
1437	fcount = 0.0_wp
1438
1439	DO i = nxl, nxr
1440	DO j = nys, nyn
1441	DO k = nzb+1, nzt
1442	DO n = 1, prt_count(k,j,i)
1443	IF ( grid_particles(k,j,i)%particles(n)%particle_mask ) THEN
1444	IF ( grid_particles(k,j,i)%particles(n)%particle_nr == -1 ) THEN
1445	fcount = fcount + finc
1446	IF ( nr_new_particle < INT( fcount ) ) THEN
1447	grid_particles(k,j,i)%particles(n)%particle_nr = particle_nr
1448	particle_nr = particle_nr + 1
1449	nr_new_particle = nr_new_particle + 1
1450	ELSE
1451	grid_particles(k,j,i)%particles(n)%particle_nr = -2
1452	ENDIF
1453	ENDIF
1454	ENDIF
1455	ENDDO
1456	ENDDO
1457	ENDDO
1458	ENDDO
1459	ELSE
1460	DO i = nxl, nxr
1461	DO j = nys, nyn
1462	DO k = nzb+1, nzt
1463	DO n = 1, prt_count(k,j,i)
1464	IF ( grid_particles(k,j,i)%particles(n)%particle_mask ) THEN
1465	IF ( grid_particles(k,j,i)%particles(n)%particle_nr == -1 ) THEN
1466	grid_particles(k,j,i)%particles(n)%particle_nr = particle_nr
1467	particle_nr = particle_nr + 1
1468	ENDIF
1469	ENDIF
1470	ENDDO
1471	ENDDO
1472	ENDDO
1473	ENDDO
1474
1475
1476	ENDIF
1477
1478
1479
1480	RETURN
1481	END SUBROUTINE dop_newly_generated_particles
1482
1483	!--------------------------------------------------------------------------------------------------!
1484	! Description:
1485	! ------------
1486	!
1487	!--------------------------------------------------------------------------------------------------!
1488	SUBROUTINE dop_count_remote_particles
1489
1490	IMPLICIT NONE
1491
1492	#if defined( __parallel )
1493	INTEGER(iwp) :: i !<
1494	INTEGER(iwp) :: ierr !< MPI error code
1495	INTEGER(iwp) :: j !<
1496	INTEGER(iwp) :: k !<
1497	INTEGER(iwp) :: n !<
1498	INTEGER(iwp) :: iop !<
1499	INTEGER(iwp) :: particle_nr
1500	INTEGER(iwp) :: pe_nr
1501	INTEGER(iwp) :: win_size
1502
1503	INTEGER(iwp), DIMENSION(0:numprocs-1) :: part_ind
1504
1505	!
1506	!-- Count remote particles
1507	remote_nr_particles = 0
1508	DO i = nxl, nxr
1509	DO j = nys, nyn
1510	DO k = nzb+1, nzt
1511	DO n = 1, prt_count(k,j,i)
1512	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1513	IF ( particle_nr > 0 ) THEN
1514	DO iop = 0, numprocs-1 !kk this loop has to be optimized
1515	!
1516	!-- Although the counting is local PE based, the following if is io processor
1517	!-- based because particles in MPI shared memory do not have to be transfered.
1518	IF ( particle_nr < io_start_index .OR. particle_nr > io_end_index ) THEN
1519	IF ( particle_nr >= mo_indices(1,iop) .AND. &
1520	particle_nr <= mo_indices(2,iop) ) THEN
1521	remote_nr_particles(2,iop) = remote_nr_particles(2,iop) + 1
1522	ENDIF
1523	ENDIF
1524	ENDDO
1525	ENDIF
1526	ENDDO
1527	ENDDO
1528	ENDDO
1529	ENDDO
1530
1531	remote_nr_particles(1,0) = 0
1532	DO i = 1, numprocs-1
1533	remote_nr_particles(1,i) = remote_nr_particles(1,i-1) + remote_nr_particles(2,i-1)
1534	ENDDO
1535
1536	win_size = SUM( remote_nr_particles(2,:) )
1537	CALL dop_alloc_rma_mem( transfer_buffer_i, win_size, win_rma_buf_i )
1538	CALL dop_alloc_rma_mem( transfer_buffer_r, win_size, win_rma_buf_r )
1539
1540	CALL MPI_ALLTOALL( remote_nr_particles, 2, MPI_INTEGER, rma_particles, 2, MPI_INTEGER, comm2d, &
1541	ierr)
1542
1543	!
1544	!-- The particles indices are the same for all output variables during one time step therefore, the
1545	!-- indices are transfered only once here.
1546	part_ind = remote_nr_particles(1,:)
1547	transfer_buffer_i = -9999
1548
1549	DO i = nxl, nxr
1550	DO j = nys, nyn
1551	DO k = nzb+1, nzt
1552
1553	DO n = 1, prt_count(k,j,i)
1554	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1555	IF ( particle_nr < io_start_index .OR. particle_nr > io_end_index ) THEN
1556	IF ( particle_nr > 0 ) THEN
1557	pe_nr = find_pe_from_particle_nr( particle_nr )
1558	transfer_buffer_i(part_ind(pe_nr)) = particle_nr
1559	part_ind(pe_nr) = part_ind(pe_nr) + 1
1560	ENDIF
1561	ENDIF
1562	ENDDO
1563	ENDDO
1564	ENDDO
1565	ENDDO
1566
1567	CALL MPI_BARRIER( MPI_COMM_WORLD, ierr )
1568
1569	CALL dop_get_remote_indices
1570
1571	#endif
1572	RETURN
1573	END SUBROUTINE dop_count_remote_particles
1574
1575	!--------------------------------------------------------------------------------------------------!
1576	! Description:
1577	! ------------
1578	!> Fill ouput buffer.
1579	!> Local variables values are copied into output buffer. Local here means local to shared memory
1580	!> group.
1581	!> Remote variable values are copied into transfer buffer this is done by all threads.
1582	!--------------------------------------------------------------------------------------------------!
1583	SUBROUTINE dop_fill_out_buf (var)
1584
1585	IMPLICIT NONE
1586
1587	CHARACTER(LEN=32) :: local_name
1588
1589	INTEGER(iwp) :: i !<
1590	INTEGER(iwp) :: j !<
1591	INTEGER(iwp) :: k !<
1592	INTEGER(iwp) :: local_len !<
1593	INTEGER(iwp) :: n !<
1594	INTEGER(iwp) :: particle_nr !<
1595	INTEGER(iwp) :: pe_nr !<
1596	INTEGER(idp) :: pval !<
1597
1598	INTEGER(iwp), DIMENSION(0:numprocs-1) :: part_ind
1599
1600	TYPE(var_def), INTENT(IN) :: var
1601
1602
1603	part_ind = remote_nr_particles(1,:)
1604	transfer_buffer_i = -9998
1605	!
1606	!-- Filling output buffer is the same for variable name and variable name_const, therefore set
1607	!-- local_name without _const
1608	local_len = INDEX( TRIM( var%name ), '_const' )
1609	IF ( local_len == 0 ) THEN
1610	local_name = var%name
1611	ELSE
1612	local_name = var%name(1:local_len-1)
1613	ENDIF
1614	!
1615	!-- In this subroutine the particles are seperated:
1616	!
1617	!-- All particles which are located in the share memory area of the respective IO thread are copied
1618	!-- into the output buffer. The other output particle are copied into the transfer buffer.
1619	SELECT CASE ( TRIM( local_name ) )
1620	CASE ( 'origin_x' )
1621	DO i = nxl, nxr
1622	DO j = nys, nyn
1623	DO k = nzb+1, nzt
1624	DO n =1, prt_count(k,j,i)
1625	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1626	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1627	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%origin_x
1628	ELSEIF ( particle_nr > 0 ) THEN
1629	pe_nr = find_pe_from_particle_nr(particle_nr)
1630	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
1631	%particles(n)%origin_x
1632	part_ind(pe_nr) = part_ind(pe_nr) + 1
1633	ENDIF
1634	ENDDO
1635	ENDDO
1636	ENDDO
1637	ENDDO
1638	CASE ( 'origin_y' )
1639	DO i = nxl, nxr
1640	DO j = nys, nyn
1641	DO k = nzb+1, nzt
1642	DO n = 1, prt_count(k,j,i)
1643	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1644	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1645	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%origin_y
1646	ELSEIF ( particle_nr > 0 ) THEN
1647	pe_nr = find_pe_from_particle_nr (particle_nr)
1648	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
1649	%particles(n)%origin_y
1650	part_ind(pe_nr) = part_ind(pe_nr) + 1
1651	ENDIF
1652	ENDDO
1653	ENDDO
1654	ENDDO
1655	ENDDO
1656	CASE ( 'origin_z' )
1657	DO i = nxl, nxr
1658	DO j = nys, nyn
1659	DO k = nzb+1, nzt
1660	DO n = 1, prt_count(k,j,i)
1661	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1662	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1663	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%origin_z
1664	ELSEIF ( particle_nr > 0 ) THEN
1665	pe_nr = find_pe_from_particle_nr (particle_nr)
1666	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
1667	%particles(n)%origin_z
1668	part_ind(pe_nr) = part_ind(pe_nr) + 1
1669	ENDIF
1670	ENDDO
1671	ENDDO
1672	ENDDO
1673	ENDDO
1674	CASE ( 'id_low' )
1675	DO i = nxl, nxr
1676	DO j = nys, nyn
1677	DO k = nzb+1, nzt
1678	DO n = 1, prt_count(k,j,i)
1679	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1680	pval = IBITS( grid_particles(k,j,i)%particles(n)%id,0,32 )
1681	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1682	out_buf_i(particle_nr) = INT(pval,4)
1683	ELSEIF ( particle_nr > 0 ) THEN
1684	pe_nr = find_pe_from_particle_nr(particle_nr)
1685	transfer_buffer_i(part_ind(pe_nr)) = INT( pval, 4 )
1686	part_ind(pe_nr) = part_ind(pe_nr) + 1
1687	ENDIF
1688	ENDDO
1689	ENDDO
1690	ENDDO
1691	ENDDO
1692	CASE ( 'id_high' )
1693	DO i = nxl, nxr
1694	DO j = nys, nyn
1695	DO k = nzb+1, nzt
1696	DO n = 1, prt_count(k,j,i)
1697	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1698	pval = IBITS( grid_particles(k,j,i)%particles(n)%id,32,32 )
1699	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1700	out_buf_i(particle_nr) = INT(pval,4)
1701	ELSEIF ( particle_nr > 0 ) THEN
1702	pe_nr = find_pe_from_particle_nr(particle_nr)
1703	transfer_buffer_i(part_ind(pe_nr)) = INT( pval, 4 )
1704	part_ind(pe_nr) = part_ind(pe_nr) + 1
1705	ENDIF
1706	ENDDO
1707	ENDDO
1708	ENDDO
1709	ENDDO
1710	CASE ( 'particle_nr' )
1711	DO i = nxl, nxr
1712	DO j = nys, nyn
1713	DO k = nzb+1, nzt
1714	DO n = 1, prt_count(k,j,i)
1715	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1716	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1717	out_buf_i(particle_nr) = grid_particles(k,j,i)%particles(n)%particle_nr
1718	ELSEIF ( particle_nr > 0 ) THEN
1719	pe_nr = find_pe_from_particle_nr(particle_nr)
1720	transfer_buffer_i(part_ind(pe_nr)) = grid_particles(k,j,i) &
1721	%particles(n)%particle_nr
1722	part_ind(pe_nr) = part_ind(pe_nr) + 1
1723	ENDIF
1724	ENDDO
1725	ENDDO
1726	ENDDO
1727	ENDDO
1728	CASE ( 'class' )
1729	DO i = nxl, nxr
1730	DO j = nys, nyn
1731	DO k = nzb+1, nzt
1732	DO n = 1, prt_count(k,j,i)
1733	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1734	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1735	out_buf_i(particle_nr) = grid_particles(k,j,i)%particles(n)%class
1736	ELSEIF ( particle_nr > 0 ) THEN
1737	pe_nr = find_pe_from_particle_nr(particle_nr)
1738	transfer_buffer_i(part_ind(pe_nr)) = grid_particles(k,j,i) &
1739	%particles(n)%class
1740	part_ind(pe_nr) = part_ind(pe_nr) + 1
1741	ENDIF
1742	ENDDO
1743	ENDDO
1744	ENDDO
1745	ENDDO
1746	CASE ( 'group' )
1747	DO i = nxl, nxr
1748	DO j = nys, nyn
1749	DO k = nzb+1, nzt
1750	DO n = 1, prt_count(k,j,i)
1751	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1752	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1753	out_buf_i(particle_nr) = grid_particles(k,j,i)%particles(n)%group
1754	ELSEIF ( particle_nr > 0 ) THEN
1755	pe_nr = find_pe_from_particle_nr(particle_nr)
1756	transfer_buffer_i(part_ind(pe_nr)) = grid_particles(k,j,i) &
1757	%particles(n)%group
1758	part_ind(pe_nr) = part_ind(pe_nr) + 1
1759	ENDIF
1760	ENDDO
1761	ENDDO
1762	ENDDO
1763	ENDDO
1764	CASE ( 'x' )
1765	DO i = nxl, nxr
1766	DO j = nys, nyn
1767	DO k = nzb+1, nzt
1768	DO n = 1, prt_count(k,j,i)
1769	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1770	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1771	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%x
1772	ELSEIF ( particle_nr > 0 ) THEN
1773	pe_nr = find_pe_from_particle_nr(particle_nr)
1774	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i)%particles(n)%x
1775	part_ind(pe_nr) = part_ind(pe_nr) + 1
1776	ENDIF
1777	ENDDO
1778	ENDDO
1779	ENDDO
1780	ENDDO
1781	CASE ( 'y' )
1782	DO i = nxl, nxr
1783	DO j = nys, nyn
1784	DO k = nzb+1, nzt
1785	DO n = 1, prt_count(k,j,i)
1786	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1787	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1788	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%y
1789	ELSEIF ( particle_nr > 0 ) THEN
1790	pe_nr = find_pe_from_particle_nr(particle_nr)
1791	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i)%particles(n)%y
1792	part_ind(pe_nr) = part_ind(pe_nr) + 1
1793	ENDIF
1794	ENDDO
1795	ENDDO
1796	ENDDO
1797	ENDDO
1798	CASE ( 'z' )
1799	DO i = nxl, nxr
1800	DO j = nys, nyn
1801	DO k = nzb+1, nzt
1802	DO n = 1, prt_count(k,j,i)
1803	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1804	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1805	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%z
1806	ELSEIF ( particle_nr > 0 ) THEN
1807	pe_nr = find_pe_from_particle_nr(particle_nr)
1808	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i)%particles(n)%z
1809	part_ind(pe_nr) = part_ind(pe_nr) + 1
1810	ENDIF
1811	ENDDO
1812	ENDDO
1813	ENDDO
1814	ENDDO
1815	CASE ( 'speed_x' )
1816	DO i = nxl, nxr
1817	DO j = nys, nyn
1818	DO k = nzb+1, nzt
1819	DO n = 1, prt_count(k,j,i)
1820	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1821	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1822	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%speed_x
1823	ELSEIF ( particle_nr > 0 ) THEN
1824	pe_nr = find_pe_from_particle_nr(particle_nr)
1825	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
1826	%particles(n)%speed_x
1827	part_ind(pe_nr) = part_ind(pe_nr) + 1
1828	ENDIF
1829	ENDDO
1830	ENDDO
1831	ENDDO
1832	ENDDO
1833	CASE ( 'speed_y' )
1834	DO i = nxl, nxr
1835	DO j = nys, nyn
1836	DO k = nzb+1, nzt
1837	DO n = 1, prt_count(k,j,i)
1838	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1839	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1840	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%speed_y
1841	ELSEIF ( particle_nr > 0 ) THEN
1842	pe_nr = find_pe_from_particle_nr(particle_nr)
1843	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
1844	%particles(n)%speed_y
1845	part_ind(pe_nr) = part_ind(pe_nr) + 1
1846	ENDIF
1847	ENDDO
1848	ENDDO
1849	ENDDO
1850	ENDDO
1851	CASE ( 'speed_z' )
1852	DO i = nxl, nxr
1853	DO j = nys, nyn
1854	DO k = nzb+1, nzt
1855	DO n = 1, prt_count(k,j,i)
1856	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1857	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1858	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%speed_z
1859	ELSEIF ( particle_nr > 0 ) THEN
1860	pe_nr = find_pe_from_particle_nr (particle_nr)
1861	transfer_buffer_r (part_ind(pe_nr)) = grid_particles(k,j,i)%particles(n)%speed_z
1862	part_ind(pe_nr) = part_ind(pe_nr) + 1
1863	ENDIF
1864	ENDDO
1865	ENDDO
1866	ENDDO
1867	ENDDO
1868	CASE ( 'radius' )
1869	DO i=nxl,nxr
1870	DO j=nys,nyn
1871	DO k=nzb+1,nzt
1872	DO n=1,prt_count(k,j,i)
1873	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1874	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1875	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%radius
1876	ELSEIF ( particle_nr > 0 ) THEN
1877	pe_nr = find_pe_from_particle_nr(particle_nr)
1878	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
1879	%particles(n)%radius
1880	part_ind(pe_nr) = part_ind(pe_nr) + 1
1881	ENDIF
1882	ENDDO
1883	ENDDO
1884	ENDDO
1885	ENDDO
1886	CASE ( 'age' )
1887	DO i = nxl, nxr
1888	DO j = nys, nyn
1889	DO k = nzb+1, nzt
1890	DO n = 1, prt_count(k,j,i)
1891	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1892	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1893	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%age
1894	ELSEIF ( particle_nr > 0 ) THEN
1895	pe_nr = find_pe_from_particle_nr(particle_nr)
1896	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i)%particles(n)%age
1897	part_ind(pe_nr) = part_ind(pe_nr) + 1
1898	ENDIF
1899	ENDDO
1900	ENDDO
1901	ENDDO
1902	ENDDO
1903	CASE ( 'age_m' )
1904	DO i = nxl, nxr
1905	DO j = nys, nyn
1906	DO k = nzb+1, nzt
1907	DO n = 1, prt_count(k,j,i)
1908	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1909	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1910	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%age_m
1911	ELSEIF ( particle_nr > 0 ) THEN
1912	pe_nr = find_pe_from_particle_nr(particle_nr)
1913	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
1914	%particles(n)%age_m
1915	part_ind(pe_nr) = part_ind(pe_nr) + 1
1916	ENDIF
1917	ENDDO
1918	ENDDO
1919	ENDDO
1920	ENDDO
1921	CASE ( 'dt_sum' )
1922	DO i = nxl, nxr
1923	DO j = nys, nyn
1924	DO k = nzb+1, nzt
1925	DO n = 1, prt_count(k,j,i)
1926	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1927	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1928	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%dt_sum
1929	ELSEIF ( particle_nr > 0 ) THEN
1930	pe_nr = find_pe_from_particle_nr(particle_nr)
1931	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
1932	%particles(n)%dt_sum
1933	part_ind(pe_nr) = part_ind(pe_nr) + 1
1934	ENDIF
1935	ENDDO
1936	ENDDO
1937	ENDDO
1938	ENDDO
1939	CASE( 'e_m' )
1940	DO i = nxl, nxr
1941	DO j = nys, nyn
1942	DO k = nzb+1, nzt
1943	DO n = 1, prt_count(k,j,i)
1944	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1945	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1946	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%e_m
1947	ELSEIF ( particle_nr > 0 ) THEN
1948	pe_nr = find_pe_from_particle_nr(particle_nr)
1949	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i)%particles(n)%e_m
1950	part_ind(pe_nr) = part_ind(pe_nr) + 1
1951	ENDIF
1952	ENDDO
1953	ENDDO
1954	ENDDO
1955	ENDDO
1956	CASE ( 'weight_factor' )
1957	DO i = nxl, nxr
1958	DO j = nys, nyn
1959	DO k = nzb+1, nzt
1960	DO n = 1, prt_count(k,j,i)
1961	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1962	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1963	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%weight_factor
1964	ELSEIF ( particle_nr > 0 ) THEN
1965	pe_nr = find_pe_from_particle_nr(particle_nr)
1966	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
1967	%particles(n)%weight_factor
1968	part_ind(pe_nr) = part_ind(pe_nr) + 1
1969	ENDIF
1970	ENDDO
1971	ENDDO
1972	ENDDO
1973	ENDDO
1974	CASE ( 'aux1' )
1975	DO i = nxl, nxr
1976	DO j = nys, nyn
1977	DO k = nzb+1, nzt
1978	DO n = 1, prt_count(k,j,i)
1979	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1980	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1981	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%aux1
1982	ELSEIF ( particle_nr > 0 ) THEN
1983	pe_nr = find_pe_from_particle_nr(particle_nr)
1984	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
1985	%particles(n)%aux1
1986	part_ind(pe_nr) = part_ind(pe_nr) + 1
1987	ENDIF
1988	ENDDO
1989	ENDDO
1990	ENDDO
1991	ENDDO
1992	CASE ( 'aux2' )
1993	DO i = nxl, nxr
1994	DO j = nys, nyn
1995	DO k = nzb+1, nzt
1996	DO n = 1, prt_count(k,j,i)
1997	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
1998	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
1999	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%aux2
2000	ELSEIF ( particle_nr > 0 ) THEN
2001	pe_nr = find_pe_from_particle_nr(particle_nr)
2002	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
2003	%particles(n)%aux2
2004	part_ind(pe_nr) = part_ind(pe_nr) + 1
2005	ENDIF
2006	ENDDO
2007	ENDDO
2008	ENDDO
2009	ENDDO
2010	CASE ( 'rvar1' )
2011	DO i = nxl, nxr
2012	DO j = nys, nyn
2013	DO k = nzb+1, nzt
2014	DO n = 1, prt_count(k,j,i)
2015	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
2016	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
2017	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%rvar1
2018	ELSEIF ( particle_nr > 0 ) THEN
2019	pe_nr = find_pe_from_particle_nr(particle_nr)
2020	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
2021	%particles(n)%rvar1
2022	part_ind(pe_nr) = part_ind(pe_nr) + 1
2023	ENDIF
2024	ENDDO
2025	ENDDO
2026	ENDDO
2027	ENDDO
2028	CASE ( 'rvar2' )
2029	DO i = nxl, nxr
2030	DO j = nys, nyn
2031	DO k = nzb+1, nzt
2032	DO n = 1, prt_count(k,j,i)
2033	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
2034	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
2035	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%rvar2
2036	ELSEIF ( particle_nr > 0 ) THEN
2037	pe_nr = find_pe_from_particle_nr (particle_nr)
2038	transfer_buffer_r (part_ind(pe_nr)) = grid_particles(k,j,i) &
2039	%particles(n)%rvar2
2040	part_ind(pe_nr) = part_ind(pe_nr) + 1
2041	ENDIF
2042	ENDDO
2043	ENDDO
2044	ENDDO
2045	ENDDO
2046	CASE ( 'rvar3' )
2047	DO i = nxl, nxr
2048	DO j = nys, nyn
2049	DO k = nzb+1, nzt
2050	DO n = 1, prt_count(k,j,i)
2051	particle_nr = grid_particles(k,j,i)%particles(n)%particle_nr
2052	IF ( particle_nr >= io_start_index .AND. particle_nr <= io_end_index ) THEN
2053	out_buf_r(particle_nr) = grid_particles(k,j,i)%particles(n)%rvar3
2054	ELSEIF ( particle_nr > 0 ) THEN
2055	pe_nr = find_pe_from_particle_nr(particle_nr)
2056	transfer_buffer_r(part_ind(pe_nr)) = grid_particles(k,j,i) &
2057	%particles(n)%rvar3
2058	part_ind(pe_nr) = part_ind(pe_nr) + 1
2059	ENDIF
2060	ENDDO
2061	ENDDO
2062	ENDDO
2063	ENDDO
2064	END SELECT
2065
2066	RETURN
2067	END SUBROUTINE dop_fill_out_buf
2068
2069	#if defined( __parallel )
2070
2071	!--------------------------------------------------------------------------------------------------!
2072	! Description:
2073	! ------------
2074	!> Get indices (displacement) of remot particles
2075	!--------------------------------------------------------------------------------------------------!
2076	SUBROUTINE dop_get_remote_indices
2077
2078	IMPLICIT NONE
2079
2080	INTEGER(iwp) :: bufsize !< size of remote indices array
2081	INTEGER(iwp) :: i !<
2082	INTEGER(iwp) :: ierr !< MPI error code
2083	INTEGER(iwp) :: ind_local !< index in remore indices array
2084	INTEGER(KIND=MPI_ADDRESS_KIND) :: disp !< displacement in RMA window
2085
2086
2087	bufsize = SUM( rma_particles(2,:) )
2088	ALLOCATE( remote_indices(0:bufsize) )
2089	remote_indices = -1
2090
2091	ind_local = 0
2092	CALL MPI_WIN_FENCE( 0, win_rma_buf_i, ierr )
2093	DO i = 0, numprocs-1
2094	IF ( rma_particles(2,i) > 0 ) THEN
2095	disp = rma_particles(1,i)
2096	IF ( rma_particles(2,i) > 0 ) THEN
2097	CALL MPI_GET( remote_indices(ind_local), rma_particles(2,i), MPI_INTEGER, i, disp, &
2098	rma_particles(2,i), MPI_INTEGER, win_rma_buf_i, ierr )
2099	ind_local = ind_local + rma_particles(2,i)
2100	ENDIF
2101	ENDIF
2102	ENDDO
2103	CALL MPI_WIN_FENCE( 0, win_rma_buf_i, ierr )
2104
2105	RETURN
2106	END SUBROUTINE dop_get_remote_indices
2107	#endif
2108
2109	!--------------------------------------------------------------------------------------------------!
2110	! Description:
2111	! ------------
2112	!--------------------------------------------------------------------------------------------------!
2113	SUBROUTINE dop_get_remote_particle (is_integer)
2114
2115	IMPLICIT NONE
2116
2117	LOGICAL, INTENT(IN) :: is_integer
2118
2119
2120	#if defined( __parallel )
2121	INTEGER(iwp) :: bufsize !< size of remote data array
2122	INTEGER(iwp) :: i !<
2123	INTEGER(iwp) :: ierr !< MPI error code
2124	INTEGER(iwp) :: ind_local !< index in remore indices array
2125	INTEGER(iwp) :: j !<
2126	INTEGER(iwp) :: particle_nr !< particle number
2127	INTEGER(KIND=MPI_ADDRESS_KIND) :: disp !< displacement in RMA window
2128
2129	INTEGER(iwp), ALLOCATABLE, DIMENSION(:) :: rma_buf_i !< buffer to receive remote data (INTEGER)
2130
2131	REAL(sp), ALLOCATABLE, DIMENSION(:) :: rma_buf_r !< buffer to receive remote data (REAL)
2132
2133
2134	bufsize = SUM( rma_particles(2,:) )
2135	ind_local = 0
2136	ALLOCATE( rma_buf_r(0:bufsize-1) )
2137	ALLOCATE( rma_buf_i(0:bufsize-1) )
2138
2139	IF ( is_integer ) THEN
2140	CALL MPI_WIN_FENCE( 0, win_rma_buf_i, ierr )
2141	ELSE
2142	CALL MPI_WIN_FENCE( 0, win_rma_buf_r, ierr )
2143	ENDIF
2144	DO i = 0, numprocs-1
2145	IF ( rma_particles(2,i) > 0 ) THEN
2146	IF ( is_integer ) THEN
2147	disp = rma_particles(1,i)
2148	CALL MPI_GET( rma_buf_i(ind_local), rma_particles(2,i), MPI_INTEGER, i, disp, &
2149	rma_particles(2,i), MPI_INTEGER, win_rma_buf_i, ierr )
2150	ind_local = ind_local + rma_particles(2,i)
2151	ELSE
2152	disp = rma_particles(1,i)
2153	CALL MPI_GET( rma_buf_r(ind_local), rma_particles(2,i), MPI_real, i, disp, &
2154	rma_particles(2,i), MPI_real, win_rma_buf_r, ierr )
2155	ind_local = ind_local + rma_particles(2,i)
2156	ENDIF
2157	ENDIF
2158	ENDDO
2159	IF ( is_integer ) THEN
2160	CALL MPI_WIN_FENCE( 0, win_rma_buf_i, ierr )
2161	ELSE
2162	CALL MPI_WIN_FENCE( 0, win_rma_buf_r, ierr )
2163	ENDIF
2164
2165	ind_local = 0
2166
2167	DO i = 0, numprocs-1
2168	IF ( rma_particles(2,i) > 0 ) THEN
2169	IF ( is_integer ) THEN
2170	!
2171	!-- Copy data from remote PEs into output array
2172	DO j = 0, rma_particles(2,i)-1
2173	particle_nr = remote_indices(ind_local)
2174	out_buf_i(particle_nr) = rma_buf_i(ind_local)
2175	ind_local = ind_local + 1
2176	ENDDO
2177	ELSE
2178	!
2179	!-- Copy data from remote PEs into output array
2180	DO j = 0, rma_particles(2,i)-1
2181	particle_nr = remote_indices(ind_local)
2182	out_buf_r(particle_nr) = rma_buf_r(ind_local)
2183	ind_local = ind_local + 1
2184	ENDDO
2185	ENDIF
2186	ENDIF
2187	ENDDO
2188
2189	IF ( ALLOCATED( rma_buf_r) ) DEALLOCATE( rma_buf_r )
2190	IF ( ALLOCATED( rma_buf_i) ) DEALLOCATE( rma_buf_i )
2191	#else
2192	IF ( is_integer ) THEN
2193	ENDIF
2194	#endif
2195
2196	RETURN
2197	END SUBROUTINE dop_get_remote_particle
2198
2199	#if defined( __parallel )
2200	!--------------------------------------------------------------------------------------------------!
2201	! Description:
2202	! ------------
2203	!> Allocate memory and cread window for one-sided communication (INTEGER 1-D array)
2204	!--------------------------------------------------------------------------------------------------!
2205	SUBROUTINE dop_alloc_rma_mem_i1( array, idim1, win )
2206
2207	IMPLICIT NONE
2208
2209	INTEGER(iwp), INTENT(IN) :: idim1 !<
2210	INTEGER :: ierr !< MPI error code
2211	INTEGER(iwp), INTENT(OUT) :: win !<
2212	INTEGER(KIND=MPI_ADDRESS_KIND) :: winsize !< size of RMA window
2213
2214	INTEGER(isp), DIMENSION(:), POINTER, INTENT(INOUT) :: array !<
2215
2216
2217	winsize = MAX( idim1, 2 )
2218
2219	ALLOCATE( array(0:winsize-1) )
2220
2221	winsize = winsize * isp
2222
2223	CALL MPI_WIN_CREATE( array, winsize, isp, MPI_INFO_NULL, comm2d, win, ierr )
2224
2225	array = -1
2226
2227	CALL MPI_WIN_FENCE( 0, win, ierr )
2228
2229	END SUBROUTINE dop_alloc_rma_mem_i1
2230	#endif
2231
2232	#if defined( __parallel )
2233	!--------------------------------------------------------------------------------------------------!
2234	! Description:
2235	! ------------
2236	!> Allocate memory and cread window for one-sided communication (REAL 1-D array)
2237	!--------------------------------------------------------------------------------------------------!
2238	SUBROUTINE dop_alloc_rma_mem_r1( array, idim1, win )
2239
2240	IMPLICIT NONE
2241
2242	INTEGER(iwp), INTENT(IN) :: idim1 !<
2243	INTEGER :: ierr !< MPI error code
2244	INTEGER(iwp), INTENT(OUT) :: win !<
2245	INTEGER(KIND=MPI_ADDRESS_KIND) :: winsize !< size of RMA window
2246
2247	REAL(sp), DIMENSION(:), POINTER, INTENT(INOUT) :: array !<
2248
2249
2250	winsize = MAX( idim1, 2 )
2251
2252	ALLOCATE( array(0:winsize-1) )
2253
2254	winsize = winsize * sp
2255
2256	CALL MPI_WIN_CREATE( array, winsize, sp, MPI_INFO_NULL, comm2d, win, ierr )
2257
2258	array = -1.0_wp
2259
2260	CALL MPI_WIN_FENCE( 0, win, ierr )
2261
2262	END SUBROUTINE dop_alloc_rma_mem_r1
2263	#endif
2264
2265	!--------------------------------------------------------------------------------------------------!
2266	! Description:
2267	! ------------
2268	!--------------------------------------------------------------------------------------------------!
2269	SUBROUTINE deallocate_and_free
2270
2271	IMPLICIT NONE
2272
2273	#if defined( __parallel )
2274	INTEGER :: ierr !< MPI error code
2275	#endif
2276
2277	#if defined( __parallel )
2278	CALL MPI_WIN_FREE( win_rma_buf_i, ierr )
2279	CALL MPI_WIN_FREE( win_rma_buf_r, ierr )
2280	#endif
2281	IF ( ALLOCATED( remote_indices ) ) DEALLOCATE( remote_indices )
2282
2283	DEALLOCATE( transfer_buffer_i )
2284	DEALLOCATE( transfer_buffer_r )
2285
2286	RETURN
2287
2288	END SUBROUTINE deallocate_and_free
2289
2290
2291	FUNCTION find_pe_from_particle_nr( particle_nr ) RESULT( pe_nr )
2292	IMPLICIT NONE
2293
2294	INTEGER(iwp) :: base !<
2295	INTEGER(iwp), INTENT(IN) :: particle_nr
2296	INTEGER(iwp) :: pe_nr !<
2297	INTEGER(iwp) :: pnr !<
2298
2299	IF ( irregular_distribubtion ) THEN
2300	IF ( particle_nr <= nr_particles_rest * nr_particles_pe ) THEN
2301	pe_nr = ( particle_nr - 1 ) / nr_particles_pe
2302	ELSE
2303	base = nr_particles_rest * nr_particles_pe
2304	pnr = particle_nr - base
2305	pe_nr = ( pnr - 1 ) / ( nr_particles_pe - 1 )
2306	pe_nr = pe_nr + nr_particles_rest
2307	ENDIF
2308	ELSE
2309	pe_nr = ( particle_nr - 1 ) / nr_particles_pe
2310	ENDIF
2311
2312
2313	!-- kk This error test is to detect programming errors. For performance reasons it can be removed
2314	!-- kk in the final, stable version.
2315
2316	END FUNCTION find_pe_from_particle_nr
2317
2318	END MODULE data_output_particle_mod

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |