[1575] | 1 | MODULE poismg_mod |
---|
| 2 | |
---|
| 3 | !--------------------------------------------------------------------------------! |
---|
| 4 | ! This file is part of PALM. |
---|
| 5 | ! |
---|
| 6 | ! PALM is free software: you can redistribute it and/or modify it under the terms |
---|
| 7 | ! of the GNU General Public License as published by the Free Software Foundation, |
---|
| 8 | ! either version 3 of the License, or (at your option) any later version. |
---|
| 9 | ! |
---|
| 10 | ! PALM is distributed in the hope that it will be useful, but WITHOUT ANY |
---|
| 11 | ! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR |
---|
| 12 | ! A PARTICULAR PURPOSE. See the GNU General Public License for more details. |
---|
| 13 | ! |
---|
| 14 | ! You should have received a copy of the GNU General Public License along with |
---|
| 15 | ! PALM. If not, see <http://www.gnu.org/licenses/>. |
---|
| 16 | ! |
---|
| 17 | ! Copyright 1997-2014 Leibniz Universitaet Hannover |
---|
| 18 | !--------------------------------------------------------------------------------! |
---|
| 19 | ! |
---|
| 20 | ! Attention: Loop unrolling and cache optimization in SOR-Red/Black method |
---|
| 21 | ! still does not give the expected speedup! Further work required. |
---|
| 22 | ! |
---|
| 23 | ! Current revisions: |
---|
| 24 | ! ----------------- |
---|
| 25 | ! Initial revision. |
---|
| 26 | ! Routine re-written and optimised based on poismg. |
---|
| 27 | ! |
---|
| 28 | ! Following optimisations have been made: |
---|
| 29 | ! - vectorisation (for Intel-CPUs) of the red-black algorithm by resorting |
---|
| 30 | ! array elements with even and odd indices |
---|
| 31 | ! - explicit boundary conditions for building walls removed (solver is |
---|
| 32 | ! running through the buildings |
---|
| 33 | ! - reduced data transfer in case of ghost point exchange, because only |
---|
| 34 | ! "red" or "black" data points need to be exchanged. This is not applied |
---|
| 35 | ! for coarser grid levels, since for then the transfer time is latency bound |
---|
| 36 | ! |
---|
| 37 | ! Former revisions: |
---|
| 38 | ! ----------------- |
---|
| 39 | ! $Id: poismg_fast.f90 1575 2015-03-27 09:56:27Z raasch $ |
---|
| 40 | ! |
---|
| 41 | ! |
---|
| 42 | ! |
---|
| 43 | ! Description: |
---|
| 44 | ! ------------ |
---|
| 45 | ! Solves the Poisson equation for the perturbation pressure with a multigrid |
---|
| 46 | ! V- or W-Cycle scheme. |
---|
| 47 | ! |
---|
| 48 | ! This multigrid method was originally developed for PALM by Joerg Uhlenbrock, |
---|
| 49 | ! September 2000 - July 2001. It has been optimised for speed by Klaus |
---|
| 50 | ! Ketelsen in November 2014. |
---|
| 51 | !------------------------------------------------------------------------------! |
---|
| 52 | |
---|
| 53 | USE cpulog, & |
---|
| 54 | ONLY: cpu_log, log_point_s |
---|
| 55 | USE kinds |
---|
| 56 | USE MPI |
---|
| 57 | USE pegrid |
---|
| 58 | |
---|
| 59 | PRIVATE |
---|
| 60 | |
---|
| 61 | INTEGER, SAVE :: ind_even_odd !: border index between even and odd k index |
---|
| 62 | INTEGER, DIMENSION(:), SAVE, ALLOCATABLE :: even_odd_level !: stores ind_even_odd for all MG levels |
---|
| 63 | |
---|
| 64 | REAL(wp), DIMENSION(:,:), SAVE, ALLOCATABLE :: f1_mg_b, f2_mg_b, f3_mg_b !: blocked version of f1_mg ... |
---|
| 65 | |
---|
| 66 | INTERFACE poismg_fast |
---|
| 67 | MODULE PROCEDURE poismg_fast |
---|
| 68 | END INTERFACE poismg_fast |
---|
| 69 | |
---|
| 70 | INTERFACE sort_k_to_even_odd_blocks |
---|
| 71 | MODULE PROCEDURE sort_k_to_even_odd_blocks |
---|
| 72 | MODULE PROCEDURE sort_k_to_even_odd_blocks_int |
---|
| 73 | MODULE PROCEDURE sort_k_to_even_odd_blocks_1d |
---|
| 74 | END INTERFACE sort_k_to_even_odd_blocks |
---|
| 75 | |
---|
| 76 | PUBLIC poismg_fast |
---|
| 77 | |
---|
| 78 | CONTAINS |
---|
| 79 | |
---|
| 80 | SUBROUTINE poismg_fast( r ) |
---|
| 81 | |
---|
| 82 | USE arrays_3d, & |
---|
| 83 | ONLY: d, p_loc |
---|
| 84 | |
---|
| 85 | USE control_parameters, & |
---|
| 86 | ONLY: gathered_size, grid_level, grid_level_count, & |
---|
| 87 | maximum_grid_level, message_string, mgcycles, mg_cycles, & |
---|
| 88 | mg_switch_to_pe0_level, residual_limit, subdomain_size |
---|
| 89 | |
---|
| 90 | USE cpulog, & |
---|
| 91 | ONLY: cpu_log, log_point_s |
---|
| 92 | |
---|
| 93 | USE indices, & |
---|
| 94 | ONLY: nxl, nxlg, nxl_mg, nxr, nxrg, nxr_mg, nys, nysg, nys_mg, nyn,& |
---|
| 95 | nyng, nyn_mg, nzb, nzt, nzt_mg |
---|
| 96 | |
---|
| 97 | IMPLICIT NONE |
---|
| 98 | |
---|
| 99 | REAL(wp) :: maxerror !: |
---|
| 100 | REAL(wp) :: maximum_mgcycles !: |
---|
| 101 | REAL(wp) :: residual_norm !: |
---|
| 102 | |
---|
| 103 | REAL(wp), DIMENSION(nzb:nzt+1,nys-1:nyn+1,nxl-1:nxr+1) :: r !: |
---|
| 104 | |
---|
| 105 | REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: p3 !: |
---|
| 106 | |
---|
| 107 | |
---|
| 108 | CALL cpu_log( log_point_s(29), 'poismg_fast', 'start' ) |
---|
| 109 | ! |
---|
| 110 | !-- Initialize arrays and variables used in this subroutine |
---|
| 111 | |
---|
| 112 | !-- If the number of grid points of the gathered grid, which is collected |
---|
| 113 | !-- on PE0, is larger than the number of grid points of an PE, than array |
---|
| 114 | !-- p3 will be enlarged. |
---|
| 115 | IF ( gathered_size > subdomain_size ) THEN |
---|
| 116 | ALLOCATE( p3(nzb:nzt_mg(mg_switch_to_pe0_level)+1,nys_mg( & |
---|
| 117 | mg_switch_to_pe0_level)-1:nyn_mg(mg_switch_to_pe0_level)+1,& |
---|
| 118 | nxl_mg(mg_switch_to_pe0_level)-1:nxr_mg( & |
---|
| 119 | mg_switch_to_pe0_level)+1) ) |
---|
| 120 | ELSE |
---|
| 121 | ALLOCATE ( p3(nzb:nzt+1,nysg:nyng,nxlg:nxrg) ) |
---|
| 122 | ENDIF |
---|
| 123 | |
---|
| 124 | p3 = 0.0_wp |
---|
| 125 | |
---|
| 126 | |
---|
| 127 | ! |
---|
| 128 | !-- Ghost boundaries have to be added to divergence array. |
---|
| 129 | !-- Exchange routine needs to know the grid level! |
---|
| 130 | grid_level = maximum_grid_level |
---|
| 131 | CALL exchange_horiz( d, 1) |
---|
| 132 | d(nzb,:,:) = d(nzb+1,:,:) |
---|
| 133 | |
---|
| 134 | ! |
---|
| 135 | !-- Initiation of the multigrid scheme. Does n cycles until the |
---|
| 136 | !-- residual is smaller than the given limit. The accuracy of the solution |
---|
| 137 | !-- of the poisson equation will increase with the number of cycles. |
---|
| 138 | !-- If the number of cycles is preset by the user, this number will be |
---|
| 139 | !-- carried out regardless of the accuracy. |
---|
| 140 | grid_level_count = 0 |
---|
| 141 | mgcycles = 0 |
---|
| 142 | IF ( mg_cycles == -1 ) THEN |
---|
| 143 | maximum_mgcycles = 0 |
---|
| 144 | residual_norm = 1.0_wp |
---|
| 145 | ELSE |
---|
| 146 | maximum_mgcycles = mg_cycles |
---|
| 147 | residual_norm = 0.0_wp |
---|
| 148 | ENDIF |
---|
| 149 | |
---|
| 150 | ! |
---|
| 151 | !-- Initial settings for sorting k-dimension from sequential order (alternate |
---|
| 152 | !-- even/odd) into blocks of even and odd or vice versa |
---|
| 153 | CALL init_even_odd_blocks |
---|
| 154 | |
---|
| 155 | ! |
---|
| 156 | !-- Sort input arrays in even/odd blocks along k-dimension |
---|
| 157 | CALL sort_k_to_even_odd_blocks( d, grid_level ) |
---|
| 158 | CALL sort_k_to_even_odd_blocks( p_loc, grid_level ) |
---|
| 159 | |
---|
| 160 | ! |
---|
| 161 | !-- The complete multigrid cycles are running in block mode, i.e. over |
---|
| 162 | !-- seperate data blocks of even and odd indices |
---|
| 163 | DO WHILE ( residual_norm > residual_limit .OR. & |
---|
| 164 | mgcycles < maximum_mgcycles ) |
---|
| 165 | |
---|
| 166 | CALL next_mg_level_fast( d, p_loc, p3, r) |
---|
| 167 | |
---|
| 168 | ! |
---|
| 169 | !-- Calculate the residual if the user has not preset the number of |
---|
| 170 | !-- cycles to be performed |
---|
| 171 | IF ( maximum_mgcycles == 0 ) THEN |
---|
| 172 | CALL resid_fast( d, p_loc, r ) |
---|
| 173 | maxerror = SUM( r(nzb+1:nzt,nys:nyn,nxl:nxr)**2 ) |
---|
| 174 | |
---|
| 175 | #if defined( __parallel ) |
---|
| 176 | IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr ) |
---|
| 177 | CALL MPI_ALLREDUCE( maxerror, residual_norm, 1, MPI_REAL, & |
---|
| 178 | MPI_SUM, comm2d, ierr) |
---|
| 179 | #else |
---|
| 180 | residual_norm = maxerror |
---|
| 181 | #endif |
---|
| 182 | residual_norm = SQRT( residual_norm ) |
---|
| 183 | ENDIF |
---|
| 184 | |
---|
| 185 | mgcycles = mgcycles + 1 |
---|
| 186 | |
---|
| 187 | ! |
---|
| 188 | !-- If the user has not limited the number of cycles, stop the run in case |
---|
| 189 | !-- of insufficient convergence |
---|
| 190 | IF ( mgcycles > 1000 .AND. mg_cycles == -1 ) THEN |
---|
| 191 | message_string = 'no sufficient convergence within 1000 cycles' |
---|
| 192 | CALL message( 'poismg_fast', 'PA0283', 1, 2, 0, 6, 0 ) |
---|
| 193 | ENDIF |
---|
| 194 | |
---|
| 195 | ENDDO |
---|
| 196 | |
---|
| 197 | DEALLOCATE( p3 ) |
---|
| 198 | ! |
---|
| 199 | !-- Result has to be sorted back from even/odd blocks to sequential order |
---|
| 200 | CALL sort_k_to_sequential( p_loc ) |
---|
| 201 | ! |
---|
| 202 | !-- Unset the grid level. Variable is used to determine the MPI datatypes for |
---|
| 203 | !-- ghost point exchange |
---|
| 204 | grid_level = 0 |
---|
| 205 | |
---|
| 206 | CALL cpu_log( log_point_s(29), 'poismg_fast', 'stop' ) |
---|
| 207 | |
---|
| 208 | END SUBROUTINE poismg_fast |
---|
| 209 | |
---|
| 210 | |
---|
| 211 | |
---|
| 212 | SUBROUTINE resid_fast( f_mg, p_mg, r ) |
---|
| 213 | |
---|
| 214 | !------------------------------------------------------------------------------! |
---|
| 215 | ! Description: |
---|
| 216 | ! ------------ |
---|
| 217 | ! Computes the residual of the perturbation pressure. |
---|
| 218 | !------------------------------------------------------------------------------! |
---|
| 219 | |
---|
| 220 | USE arrays_3d, & |
---|
| 221 | ONLY: f1_mg, f2_mg, f3_mg |
---|
| 222 | |
---|
| 223 | USE control_parameters, & |
---|
| 224 | ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, inflow_l,& |
---|
| 225 | inflow_n, inflow_r, inflow_s, masking_method, outflow_l, & |
---|
| 226 | outflow_n, outflow_r, outflow_s, topography |
---|
| 227 | |
---|
| 228 | USE grid_variables, & |
---|
| 229 | ONLY: ddx2_mg, ddy2_mg |
---|
| 230 | |
---|
| 231 | USE indices, & |
---|
| 232 | ONLY: flags, wall_flags_1, wall_flags_2, wall_flags_3, & |
---|
| 233 | wall_flags_4, wall_flags_5, wall_flags_6, wall_flags_7, & |
---|
| 234 | wall_flags_8, wall_flags_9, wall_flags_10, nxl_mg, nxr_mg, & |
---|
| 235 | nys_mg, nyn_mg, nzb, nzt_mg |
---|
| 236 | |
---|
| 237 | IMPLICIT NONE |
---|
| 238 | |
---|
| 239 | INTEGER(iwp) :: i |
---|
| 240 | INTEGER(iwp) :: j |
---|
| 241 | INTEGER(iwp) :: k |
---|
| 242 | INTEGER(iwp) :: l |
---|
| 243 | INTEGER(iwp) :: km1 !: |
---|
| 244 | INTEGER(iwp) :: kp1 !: |
---|
| 245 | |
---|
| 246 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 247 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 248 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f_mg !: |
---|
| 249 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 250 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 251 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p_mg !: |
---|
| 252 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 253 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 254 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: r !: |
---|
| 255 | |
---|
| 256 | ! |
---|
| 257 | !-- Calculate the residual |
---|
| 258 | l = grid_level |
---|
| 259 | |
---|
| 260 | CALL cpu_log( log_point_s(53), 'resid', 'start' ) |
---|
| 261 | IF ( topography == 'flat' .OR. masking_method ) THEN |
---|
| 262 | !$OMP PARALLEL PRIVATE (i,j,k,km1,kp1) |
---|
| 263 | !$OMP DO |
---|
| 264 | DO i = nxl_mg(l), nxr_mg(l) |
---|
| 265 | DO j = nys_mg(l), nyn_mg(l) |
---|
| 266 | !DIR$ IVDEP |
---|
| 267 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 268 | km1 = k-ind_even_odd-1 |
---|
| 269 | kp1 = k-ind_even_odd |
---|
| 270 | r(k,j,i) = f_mg(k,j,i) & |
---|
| 271 | - ddx2_mg(l) * & |
---|
| 272 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 273 | - ddy2_mg(l) * & |
---|
| 274 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 275 | - f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 276 | - f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 277 | + f1_mg_b(k,l) * p_mg(k,j,i) |
---|
| 278 | ENDDO |
---|
| 279 | !DIR$ IVDEP |
---|
| 280 | DO k = nzb+1, ind_even_odd |
---|
| 281 | km1 = k+ind_even_odd |
---|
| 282 | kp1 = k+ind_even_odd+1 |
---|
| 283 | r(k,j,i) = f_mg(k,j,i) & |
---|
| 284 | - ddx2_mg(l) * & |
---|
| 285 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 286 | - ddy2_mg(l) * & |
---|
| 287 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 288 | - f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 289 | - f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 290 | + f1_mg_b(k,l) * p_mg(k,j,i) |
---|
| 291 | ENDDO |
---|
| 292 | ENDDO |
---|
| 293 | ENDDO |
---|
| 294 | !$OMP END PARALLEL |
---|
| 295 | ELSE |
---|
| 296 | ! |
---|
| 297 | !-- Choose flag array of this level |
---|
| 298 | SELECT CASE ( l ) |
---|
| 299 | CASE ( 1 ) |
---|
| 300 | flags => wall_flags_1 |
---|
| 301 | CASE ( 2 ) |
---|
| 302 | flags => wall_flags_2 |
---|
| 303 | CASE ( 3 ) |
---|
| 304 | flags => wall_flags_3 |
---|
| 305 | CASE ( 4 ) |
---|
| 306 | flags => wall_flags_4 |
---|
| 307 | CASE ( 5 ) |
---|
| 308 | flags => wall_flags_5 |
---|
| 309 | CASE ( 6 ) |
---|
| 310 | flags => wall_flags_6 |
---|
| 311 | CASE ( 7 ) |
---|
| 312 | flags => wall_flags_7 |
---|
| 313 | CASE ( 8 ) |
---|
| 314 | flags => wall_flags_8 |
---|
| 315 | CASE ( 9 ) |
---|
| 316 | flags => wall_flags_9 |
---|
| 317 | CASE ( 10 ) |
---|
| 318 | flags => wall_flags_10 |
---|
| 319 | END SELECT |
---|
| 320 | |
---|
| 321 | !$OMP PARALLEL PRIVATE (i,j,k,km1,kp1) |
---|
| 322 | !$OMP DO |
---|
| 323 | DO i = nxl_mg(l), nxr_mg(l) |
---|
| 324 | DO j = nys_mg(l), nyn_mg(l) |
---|
| 325 | |
---|
| 326 | !DIR$ IVDEP |
---|
| 327 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 328 | km1 = k-ind_even_odd-1 |
---|
| 329 | kp1 = k-ind_even_odd |
---|
| 330 | |
---|
| 331 | r(k,j,i) = f_mg(k,j,i) & |
---|
| 332 | - ddx2_mg(l) * & |
---|
| 333 | ( MERGE(p_mg(k,j,i),p_mg(k,j,i+1),BTEST( flags(k,j,i), 5 )) & |
---|
| 334 | + MERGE(p_mg(k,j,i),p_mg(k,j,i-1),BTEST( flags(k,j,i), 4 )) ) & |
---|
| 335 | - ddy2_mg(l) & |
---|
| 336 | * ( MERGE(p_mg(k,j,i),p_mg(k,j+1,i),BTEST( flags(k,j,i), 3 )) & |
---|
| 337 | + MERGE(p_mg(k,j,i),p_mg(k,j-1,i),BTEST( flags(k,j,i), 2 )) ) & |
---|
| 338 | - f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 339 | - f3_mg(k,l) * & |
---|
| 340 | MERGE(p_mg(k,j,i),p_mg(km1,j,i),BTEST( flags(k,j,i), 0 )) & |
---|
| 341 | + f1_mg(k,l) * p_mg(k,j,i) |
---|
| 342 | ENDDO |
---|
| 343 | |
---|
| 344 | !DIR$ IVDEP |
---|
| 345 | DO k = nzb+1, ind_even_odd |
---|
| 346 | km1 = k+ind_even_odd |
---|
| 347 | kp1 = k+ind_even_odd+1 |
---|
| 348 | |
---|
| 349 | r(k,j,i) = f_mg(k,j,i) & |
---|
| 350 | - ddx2_mg(l) * & |
---|
| 351 | ( MERGE(p_mg(k,j,i),p_mg(k,j,i+1),BTEST( flags(k,j,i), 5 )) & |
---|
| 352 | + MERGE(p_mg(k,j,i),p_mg(k,j,i-1),BTEST( flags(k,j,i), 4 )) ) & |
---|
| 353 | - ddy2_mg(l) & |
---|
| 354 | * ( MERGE(p_mg(k,j,i),p_mg(k,j+1,i),BTEST( flags(k,j,i), 3 )) & |
---|
| 355 | + MERGE(p_mg(k,j,i),p_mg(k,j-1,i),BTEST( flags(k,j,i), 2 )) ) & |
---|
| 356 | - f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 357 | - f3_mg(k,l) * & |
---|
| 358 | MERGE(p_mg(k,j,i),p_mg(km1,j,i),BTEST( flags(k,j,i), 0 )) & |
---|
| 359 | + f1_mg(k,l) * p_mg(k,j,i) |
---|
| 360 | ENDDO |
---|
| 361 | ! |
---|
| 362 | !-- The residual within topography should be zero |
---|
| 363 | WHERE( BTEST(flags(nzb+1:nzt_mg(l),j,i), 6) ) |
---|
| 364 | r(nzb+1:nzt_mg(l),j,i) = 0.0 |
---|
| 365 | END WHERE |
---|
| 366 | |
---|
| 367 | ENDDO |
---|
| 368 | ENDDO |
---|
| 369 | !$OMP END PARALLEL |
---|
| 370 | |
---|
| 371 | ENDIF |
---|
| 372 | |
---|
| 373 | ! |
---|
| 374 | !-- Horizontal boundary conditions |
---|
| 375 | CALL exchange_horiz( r, 1) |
---|
| 376 | |
---|
| 377 | IF ( .NOT. bc_lr_cyc ) THEN |
---|
| 378 | IF ( inflow_l .OR. outflow_l ) r(:,:,nxl_mg(l)-1) = r(:,:,nxl_mg(l)) |
---|
| 379 | IF ( inflow_r .OR. outflow_r ) r(:,:,nxr_mg(l)+1) = r(:,:,nxr_mg(l)) |
---|
| 380 | ENDIF |
---|
| 381 | |
---|
| 382 | IF ( .NOT. bc_ns_cyc ) THEN |
---|
| 383 | IF ( inflow_n .OR. outflow_n ) r(:,nyn_mg(l)+1,:) = r(:,nyn_mg(l),:) |
---|
| 384 | IF ( inflow_s .OR. outflow_s ) r(:,nys_mg(l)-1,:) = r(:,nys_mg(l),:) |
---|
| 385 | ENDIF |
---|
| 386 | |
---|
| 387 | ! |
---|
| 388 | !-- Boundary conditions at bottom and top of the domain.outflow_l, outflow_n, |
---|
| 389 | !-- These points are not handled by the above loop. Points may be within |
---|
| 390 | !-- buildings, but that doesn't matter. |
---|
| 391 | IF ( ibc_p_b == 1 ) THEN |
---|
| 392 | r(nzb,:,: ) = r(nzb+1,:,:) |
---|
| 393 | ELSE |
---|
| 394 | r(nzb,:,: ) = 0.0_wp |
---|
| 395 | ENDIF |
---|
| 396 | |
---|
| 397 | IF ( ibc_p_t == 1 ) THEN |
---|
| 398 | r(nzt_mg(l)+1,:,: ) = r(nzt_mg(l),:,:) |
---|
| 399 | ELSE |
---|
| 400 | r(nzt_mg(l)+1,:,: ) = 0.0_wp |
---|
| 401 | ENDIF |
---|
| 402 | |
---|
| 403 | CALL cpu_log( log_point_s(53), 'resid', 'stop' ) |
---|
| 404 | |
---|
| 405 | END SUBROUTINE resid_fast |
---|
| 406 | |
---|
| 407 | |
---|
| 408 | |
---|
| 409 | SUBROUTINE restrict_fast( f_mg, r ) |
---|
| 410 | |
---|
| 411 | !------------------------------------------------------------------------------! |
---|
| 412 | ! Description: |
---|
| 413 | ! ------------ |
---|
| 414 | ! Interpolates the residual on the next coarser grid with "full weighting" |
---|
| 415 | ! scheme |
---|
| 416 | !------------------------------------------------------------------------------! |
---|
| 417 | |
---|
| 418 | USE control_parameters, & |
---|
| 419 | ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, inflow_l,& |
---|
| 420 | inflow_n, inflow_r, inflow_s, masking_method, outflow_l, & |
---|
| 421 | outflow_n, outflow_r, outflow_s, topography |
---|
| 422 | |
---|
| 423 | USE indices, & |
---|
| 424 | ONLY: flags, wall_flags_1, wall_flags_2, wall_flags_3, & |
---|
| 425 | wall_flags_4, wall_flags_5, wall_flags_6, wall_flags_7, & |
---|
| 426 | wall_flags_8, wall_flags_9, wall_flags_10, nxl_mg, nxr_mg, & |
---|
| 427 | nys_mg, nyn_mg, nzb, nzt_mg |
---|
| 428 | |
---|
| 429 | IMPLICIT NONE |
---|
| 430 | |
---|
| 431 | INTEGER(iwp) :: i !: |
---|
| 432 | INTEGER(iwp) :: ic !: |
---|
| 433 | INTEGER(iwp) :: j !: |
---|
| 434 | INTEGER(iwp) :: jc !: |
---|
| 435 | INTEGER(iwp) :: k !: |
---|
| 436 | INTEGER(iwp) :: kc !: |
---|
| 437 | INTEGER(iwp) :: l !: |
---|
| 438 | INTEGER(iwp) :: km1 !: |
---|
| 439 | INTEGER(iwp) :: kp1 !: |
---|
| 440 | |
---|
| 441 | REAL(wp) :: rkjim !: |
---|
| 442 | REAL(wp) :: rkjip !: |
---|
| 443 | REAL(wp) :: rkjmi !: |
---|
| 444 | REAL(wp) :: rkjmim !: |
---|
| 445 | REAL(wp) :: rkjmip !: |
---|
| 446 | REAL(wp) :: rkjpi !: |
---|
| 447 | REAL(wp) :: rkjpim !: |
---|
| 448 | REAL(wp) :: rkjpip !: |
---|
| 449 | REAL(wp) :: rkmji !: |
---|
| 450 | REAL(wp) :: rkmjim !: |
---|
| 451 | REAL(wp) :: rkmjip !: |
---|
| 452 | REAL(wp) :: rkmjmi !: |
---|
| 453 | REAL(wp) :: rkmjmim !: |
---|
| 454 | REAL(wp) :: rkmjmip !: |
---|
| 455 | REAL(wp) :: rkmjpi !: |
---|
| 456 | REAL(wp) :: rkmjpim !: |
---|
| 457 | REAL(wp) :: rkmjpip !: |
---|
| 458 | |
---|
| 459 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 460 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 461 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f_mg !: |
---|
| 462 | |
---|
| 463 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level+1)+1, & |
---|
| 464 | nys_mg(grid_level+1)-1:nyn_mg(grid_level+1)+1, & |
---|
| 465 | nxl_mg(grid_level+1)-1:nxr_mg(grid_level+1)+1) :: r !: |
---|
| 466 | |
---|
| 467 | ! |
---|
| 468 | !-- Interpolate the residual |
---|
| 469 | l = grid_level |
---|
| 470 | |
---|
| 471 | CALL cpu_log( log_point_s(54), 'restrict', 'start' ) |
---|
| 472 | |
---|
| 473 | IF ( topography == 'flat' .OR. masking_method ) THEN |
---|
| 474 | ! |
---|
| 475 | !-- No wall treatment |
---|
| 476 | !$OMP PARALLEL PRIVATE (i,j,k,ic,jc,kc,km1,kp1) |
---|
| 477 | DO ic = nxl_mg(l), nxr_mg(l) |
---|
| 478 | i = 2*ic |
---|
| 479 | !$OMP DO SCHEDULE( STATIC ) |
---|
| 480 | DO jc = nys_mg(l), nyn_mg(l) |
---|
| 481 | ! |
---|
| 482 | !-- Calculation for the first point along k |
---|
| 483 | j = 2*jc |
---|
| 484 | k = ind_even_odd+1 |
---|
| 485 | kp1 = k-ind_even_odd |
---|
| 486 | kc = k-ind_even_odd |
---|
| 487 | f_mg(kc,jc,ic) = 1.0_wp / 64.0_wp * ( & |
---|
| 488 | 8.0_wp * r(k,j,i) & |
---|
| 489 | + 4.0_wp * ( r(k,j,i-1) + r(k,j,i+1) + & |
---|
| 490 | r(k,j+1,i) + r(k,j-1,i) ) & |
---|
| 491 | + 2.0_wp * ( r(k,j-1,i-1) + r(k,j+1,i-1) + & |
---|
| 492 | r(k,j-1,i+1) + r(k,j+1,i+1) ) & |
---|
| 493 | + 16.0_wp * r(k,j,i) & |
---|
| 494 | + 4.0_wp * r(kp1,j,i) & |
---|
| 495 | + 2.0_wp * ( r(kp1,j,i-1) + r(kp1,j,i+1) + & |
---|
| 496 | r(kp1,j+1,i) + r(kp1,j-1,i) ) & |
---|
| 497 | + ( r(kp1,j-1,i-1) + r(kp1,j+1,i-1) + & |
---|
| 498 | r(kp1,j-1,i+1) + r(kp1,j+1,i+1) ) & |
---|
| 499 | ) |
---|
| 500 | ! |
---|
| 501 | !-- Calculation for the other points along k |
---|
| 502 | !DIR$ IVDEP |
---|
| 503 | DO k = ind_even_odd+2, nzt_mg(l+1) ! Fine grid at this point |
---|
| 504 | km1 = k-ind_even_odd-1 |
---|
| 505 | kp1 = k-ind_even_odd |
---|
| 506 | kc = k-ind_even_odd ! Coarse grid index |
---|
| 507 | |
---|
| 508 | f_mg(kc,jc,ic) = 1.0_wp / 64.0_wp * ( & |
---|
| 509 | 8.0_wp * r(k,j,i) & |
---|
| 510 | + 4.0_wp * ( r(k,j,i-1) + r(k,j,i+1) + & |
---|
| 511 | r(k,j+1,i) + r(k,j-1,i) ) & |
---|
| 512 | + 2.0_wp * ( r(k,j-1,i-1) + r(k,j+1,i-1) + & |
---|
| 513 | r(k,j-1,i+1) + r(k,j+1,i+1) ) & |
---|
| 514 | + 4.0_wp * r(km1,j,i) & |
---|
| 515 | + 2.0_wp * ( r(km1,j,i-1) + r(km1,j,i+1) + & |
---|
| 516 | r(km1,j+1,i) + r(km1,j-1,i) ) & |
---|
| 517 | + ( r(km1,j-1,i-1) + r(km1,j+1,i-1) + & |
---|
| 518 | r(km1,j-1,i+1) + r(km1,j+1,i+1) ) & |
---|
| 519 | + 4.0_wp * r(kp1,j,i) & |
---|
| 520 | + 2.0_wp * ( r(kp1,j,i-1) + r(kp1,j,i+1) + & |
---|
| 521 | r(kp1,j+1,i) + r(kp1,j-1,i) ) & |
---|
| 522 | + ( r(kp1,j-1,i-1) + r(kp1,j+1,i-1) + & |
---|
| 523 | r(kp1,j-1,i+1) + r(kp1,j+1,i+1) ) & |
---|
| 524 | ) |
---|
| 525 | ENDDO |
---|
| 526 | ENDDO |
---|
| 527 | !$OMP ENDDO nowait |
---|
| 528 | ENDDO |
---|
| 529 | !$OMP END PARALLEL |
---|
| 530 | |
---|
| 531 | ELSE |
---|
| 532 | ! |
---|
| 533 | !-- Choose flag array of the upper level |
---|
| 534 | SELECT CASE ( l+1 ) |
---|
| 535 | CASE ( 1 ) |
---|
| 536 | flags => wall_flags_1 |
---|
| 537 | CASE ( 2 ) |
---|
| 538 | flags => wall_flags_2 |
---|
| 539 | CASE ( 3 ) |
---|
| 540 | flags => wall_flags_3 |
---|
| 541 | CASE ( 4 ) |
---|
| 542 | flags => wall_flags_4 |
---|
| 543 | CASE ( 5 ) |
---|
| 544 | flags => wall_flags_5 |
---|
| 545 | CASE ( 6 ) |
---|
| 546 | flags => wall_flags_6 |
---|
| 547 | CASE ( 7 ) |
---|
| 548 | flags => wall_flags_7 |
---|
| 549 | CASE ( 8 ) |
---|
| 550 | flags => wall_flags_8 |
---|
| 551 | CASE ( 9 ) |
---|
| 552 | flags => wall_flags_9 |
---|
| 553 | CASE ( 10 ) |
---|
| 554 | flags => wall_flags_10 |
---|
| 555 | END SELECT |
---|
| 556 | |
---|
| 557 | !$OMP PARALLEL PRIVATE (i,j,k,ic,jc,kc,km1,kp1,rkjim,rkjip,rkjpi, & |
---|
| 558 | !$OMP rkjmim,rkjpim,rkjmip,rkjpip,rkmji,rkmjim, & |
---|
| 559 | !$OMP rkmjip,rkmjpi,rkmjmi,rkmjmim,rkmjpim, & |
---|
| 560 | !$OMP rkmjmip,rkmjpip) |
---|
| 561 | !$OMP DO |
---|
| 562 | DO ic = nxl_mg(l), nxr_mg(l) |
---|
| 563 | i = 2*ic |
---|
| 564 | DO jc = nys_mg(l), nyn_mg(l) |
---|
| 565 | j = 2*jc |
---|
| 566 | !DIR$ IVDEP |
---|
| 567 | DO k = ind_even_odd+1, nzt_mg(l+1) !fine grid at this point |
---|
| 568 | |
---|
| 569 | km1 = k-ind_even_odd-1 |
---|
| 570 | kp1 = k-ind_even_odd |
---|
| 571 | kc = k-ind_even_odd !Coarse grid index |
---|
| 572 | ! |
---|
| 573 | !-- Use implicit Neumann BCs if the respective gridpoint is |
---|
| 574 | !-- inside the building |
---|
| 575 | rkjim = MERGE(r(k,j,i),r(k,j,i-1),BTEST( flags(k,j,i-1), 6)) |
---|
| 576 | rkjip = MERGE(r(k,j,i),r(k,j,i+1),BTEST( flags(k,j,i+1), 6)) |
---|
| 577 | rkjpi = MERGE(r(k,j,i),r(k,j+1,i),BTEST( flags(k,j+1,i), 6)) |
---|
| 578 | rkjmi = MERGE(r(k,j,i),r(k,j-1,i),BTEST( flags(k,j-1,i), 6)) |
---|
| 579 | |
---|
| 580 | rkjmim = MERGE(r(k,j,i),r(k,j-1,i-1),BTEST( flags(k,j-1,i-1), 6)) |
---|
| 581 | rkjpim = MERGE(r(k,j,i),r(k,j+1,i-1),BTEST( flags(k,j+1,i-1), 6)) |
---|
| 582 | rkjmip = MERGE(r(k,j,i),r(k,j-1,i+1),BTEST( flags(k,j-1,i+1), 6)) |
---|
| 583 | rkjpip = MERGE(r(k,j,i),r(k,j+1,i+1),BTEST( flags(k,j+1,i+1), 6)) |
---|
| 584 | |
---|
| 585 | rkmji = MERGE(r(k,j,i),r(km1,j,i) ,BTEST( flags(km1,j,i) , 6)) |
---|
| 586 | rkmjim = MERGE(r(k,j,i),r(km1,j,i-1),BTEST( flags(km1,j,i-1), 6)) |
---|
| 587 | rkmjip = MERGE(r(k,j,i),r(km1,j,i+1),BTEST( flags(km1,j,i+1), 6)) |
---|
| 588 | rkmjpi = MERGE(r(k,j,i),r(km1,j+1,i),BTEST( flags(km1,j+1,i), 6)) |
---|
| 589 | rkmjmi = MERGE(r(k,j,i),r(km1,j-1,i),BTEST( flags(km1,j-1,i), 6)) |
---|
| 590 | |
---|
| 591 | rkmjmim = MERGE(r(k,j,i),r(km1,j-1,i-1),BTEST( flags(km1,j-1,i-1), 6)) |
---|
| 592 | rkmjpim = MERGE(r(k,j,i),r(km1,j+1,i-1),BTEST( flags(km1,j+1,i-1), 6)) |
---|
| 593 | rkmjmip = MERGE(r(k,j,i),r(km1,j-1,i+1),BTEST( flags(km1,j-1,i+1), 6)) |
---|
| 594 | rkmjpip = MERGE(r(k,j,i),r(km1,j+1,i+1),BTEST( flags(km1,j+1,i+1), 6)) |
---|
| 595 | |
---|
| 596 | f_mg(kc,jc,ic) = 1.0_wp / 64.0_wp * ( & |
---|
| 597 | 8.0_wp * r(k,j,i) & |
---|
| 598 | + 4.0_wp * ( rkjim + rkjip + rkjpi + rkjmi ) & |
---|
| 599 | + 2.0_wp * ( rkjmim + rkjpim + rkjmip + rkjpip ) & |
---|
| 600 | + 4.0_wp * rkmji & |
---|
| 601 | + 2.0_wp * ( rkmjim + rkmjip + rkmjpi + rkmjmi ) & |
---|
| 602 | + ( rkmjmim + rkmjpim + rkmjmip + rkmjpip ) & |
---|
| 603 | + 4.0_wp * r(kp1,j,i) & |
---|
| 604 | + 2.0_wp * ( r(kp1,j,i-1) + r(kp1,j,i+1) + & |
---|
| 605 | r(kp1,j+1,i) + r(kp1,j-1,i) ) & |
---|
| 606 | + ( r(kp1,j-1,i-1) + r(kp1,j+1,i-1) + & |
---|
| 607 | r(kp1,j-1,i+1) + r(kp1,j+1,i+1) ) & |
---|
| 608 | ) |
---|
| 609 | ENDDO |
---|
| 610 | ENDDO |
---|
| 611 | ENDDO |
---|
| 612 | !$OMP END PARALLEL |
---|
| 613 | |
---|
| 614 | ENDIF |
---|
| 615 | |
---|
| 616 | ! |
---|
| 617 | !-- Horizontal boundary conditions |
---|
| 618 | CALL exchange_horiz( f_mg, 1) |
---|
| 619 | |
---|
| 620 | IF ( .NOT. bc_lr_cyc ) THEN |
---|
| 621 | IF (inflow_l .OR. outflow_l) f_mg(:,:,nxl_mg(l)-1) = f_mg(:,:,nxl_mg(l)) |
---|
| 622 | IF (inflow_r .OR. outflow_r) f_mg(:,:,nxr_mg(l)+1) = f_mg(:,:,nxr_mg(l)) |
---|
| 623 | ENDIF |
---|
| 624 | |
---|
| 625 | IF ( .NOT. bc_ns_cyc ) THEN |
---|
| 626 | IF (inflow_n .OR. outflow_n) f_mg(:,nyn_mg(l)+1,:) = f_mg(:,nyn_mg(l),:) |
---|
| 627 | IF (inflow_s .OR. outflow_s) f_mg(:,nys_mg(l)-1,:) = f_mg(:,nys_mg(l),:) |
---|
| 628 | ENDIF |
---|
| 629 | |
---|
| 630 | ! |
---|
| 631 | !-- Boundary conditions at bottom and top of the domain. |
---|
| 632 | !-- These points are not handled by the above loop. Points may be within |
---|
| 633 | !-- buildings, but that doesn't matter. |
---|
| 634 | IF ( ibc_p_b == 1 ) THEN |
---|
| 635 | f_mg(nzb,:,: ) = f_mg(nzb+1,:,:) |
---|
| 636 | ELSE |
---|
| 637 | f_mg(nzb,:,: ) = 0.0_wp |
---|
| 638 | ENDIF |
---|
| 639 | |
---|
| 640 | IF ( ibc_p_t == 1 ) THEN |
---|
| 641 | f_mg(nzt_mg(l)+1,:,: ) = f_mg(nzt_mg(l),:,:) |
---|
| 642 | ELSE |
---|
| 643 | f_mg(nzt_mg(l)+1,:,: ) = 0.0_wp |
---|
| 644 | ENDIF |
---|
| 645 | |
---|
| 646 | CALL cpu_log( log_point_s(54), 'restrict', 'stop' ) |
---|
| 647 | ! |
---|
| 648 | !-- Why do we need a sorting here? |
---|
| 649 | CALL sort_k_to_even_odd_blocks( f_mg , l) |
---|
| 650 | |
---|
| 651 | END SUBROUTINE restrict_fast |
---|
| 652 | |
---|
| 653 | |
---|
| 654 | |
---|
| 655 | SUBROUTINE prolong_fast( p, temp ) |
---|
| 656 | |
---|
| 657 | !------------------------------------------------------------------------------! |
---|
| 658 | ! Description: |
---|
| 659 | ! ------------ |
---|
| 660 | ! Interpolates the correction of the perturbation pressure |
---|
| 661 | ! to the next finer grid. |
---|
| 662 | !------------------------------------------------------------------------------! |
---|
| 663 | |
---|
| 664 | USE control_parameters, & |
---|
| 665 | ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, inflow_l,& |
---|
| 666 | inflow_n, inflow_r, inflow_s, outflow_l, outflow_n, & |
---|
| 667 | outflow_r, outflow_s |
---|
| 668 | |
---|
| 669 | USE indices, & |
---|
| 670 | ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
---|
| 671 | |
---|
| 672 | IMPLICIT NONE |
---|
| 673 | |
---|
| 674 | INTEGER(iwp) :: i !: |
---|
| 675 | INTEGER(iwp) :: j !: |
---|
| 676 | INTEGER(iwp) :: k !: |
---|
| 677 | INTEGER(iwp) :: l !: |
---|
| 678 | INTEGER(iwp) :: kp1 !: |
---|
| 679 | INTEGER(iwp) :: ke !: Index for prolog even |
---|
| 680 | INTEGER(iwp) :: ko !: Index for prolog odd |
---|
| 681 | |
---|
| 682 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, & |
---|
| 683 | nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, & |
---|
| 684 | nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1 ) :: p !: |
---|
| 685 | |
---|
| 686 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 687 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 688 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: temp !: |
---|
| 689 | |
---|
| 690 | |
---|
| 691 | CALL cpu_log( log_point_s(55), 'prolong', 'start' ) |
---|
| 692 | |
---|
| 693 | ! |
---|
| 694 | !-- First, store elements of the coarser grid on the next finer grid |
---|
| 695 | l = grid_level |
---|
| 696 | ind_even_odd = even_odd_level(grid_level-1) |
---|
| 697 | |
---|
| 698 | !$OMP PARALLEL PRIVATE (i,j,k,kp1,ke,ko) |
---|
| 699 | !$OMP DO |
---|
| 700 | DO i = nxl_mg(l-1), nxr_mg(l-1) |
---|
| 701 | DO j = nys_mg(l-1), nyn_mg(l-1) |
---|
| 702 | |
---|
| 703 | !DIR$ IVDEP |
---|
| 704 | DO k = ind_even_odd+1, nzt_mg(l-1) |
---|
| 705 | kp1 = k - ind_even_odd |
---|
| 706 | ke = 2 * ( k-ind_even_odd - 1 ) + 1 |
---|
| 707 | ko = 2 * k - 1 |
---|
| 708 | ! |
---|
| 709 | !-- Points of the coarse grid are directly stored on the next finer |
---|
| 710 | !-- grid |
---|
| 711 | temp(ko,2*j,2*i) = p(k,j,i) |
---|
| 712 | ! |
---|
| 713 | !-- Points between two coarse-grid points |
---|
| 714 | temp(ko,2*j,2*i+1) = 0.5_wp * ( p(k,j,i) + p(k,j,i+1) ) |
---|
| 715 | temp(ko,2*j+1,2*i) = 0.5_wp * ( p(k,j,i) + p(k,j+1,i) ) |
---|
| 716 | temp(ke,2*j,2*i) = 0.5_wp * ( p(k,j,i) + p(kp1,j,i) ) |
---|
| 717 | ! |
---|
| 718 | !-- Points in the center of the planes stretched by four points |
---|
| 719 | !-- of the coarse grid cube |
---|
| 720 | temp(ko,2*j+1,2*i+1) = 0.25_wp * ( p(k,j,i) + p(k,j,i+1) + & |
---|
| 721 | p(k,j+1,i) + p(k,j+1,i+1) ) |
---|
| 722 | temp(ke,2*j,2*i+1) = 0.25_wp * ( p(k,j,i) + p(k,j,i+1) + & |
---|
| 723 | p(kp1,j,i) + p(kp1,j,i+1) ) |
---|
| 724 | temp(ke,2*j+1,2*i) = 0.25_wp * ( p(k,j,i) + p(k,j+1,i) + & |
---|
| 725 | p(kp1,j,i) + p(kp1,j+1,i) ) |
---|
| 726 | ! |
---|
| 727 | !-- Points in the middle of coarse grid cube |
---|
| 728 | temp(ke,2*j+1,2*i+1) = 0.125_wp * & |
---|
| 729 | ( p(k,j,i) + p(k,j,i+1) + & |
---|
| 730 | p(k,j+1,i) + p(k,j+1,i+1) + & |
---|
| 731 | p(kp1,j,i) + p(kp1,j,i+1) + & |
---|
| 732 | p(kp1,j+1,i) + p(kp1,j+1,i+1) ) |
---|
| 733 | ENDDO |
---|
| 734 | |
---|
| 735 | !DIR$ IVDEP |
---|
| 736 | DO k = nzb+1, ind_even_odd |
---|
| 737 | kp1 = k + ind_even_odd + 1 |
---|
| 738 | ke = 2 * k |
---|
| 739 | ko = 2 * ( k + ind_even_odd ) |
---|
| 740 | ! |
---|
| 741 | !-- Points of the coarse grid are directly stored on the next finer |
---|
| 742 | !-- grid |
---|
| 743 | temp(ko,2*j,2*i) = p(k,j,i) |
---|
| 744 | ! |
---|
| 745 | !-- Points between two coarse-grid points |
---|
| 746 | temp(ko,2*j,2*i+1) = 0.5_wp * ( p(k,j,i) + p(k,j,i+1) ) |
---|
| 747 | temp(ko,2*j+1,2*i) = 0.5_wp * ( p(k,j,i) + p(k,j+1,i) ) |
---|
| 748 | temp(ke,2*j,2*i) = 0.5_wp * ( p(k,j,i) + p(kp1,j,i) ) |
---|
| 749 | ! |
---|
| 750 | !-- Points in the center of the planes stretched by four points |
---|
| 751 | !-- of the coarse grid cube |
---|
| 752 | temp(ko,2*j+1,2*i+1) = 0.25_wp * ( p(k,j,i) + p(k,j,i+1) + & |
---|
| 753 | p(k,j+1,i) + p(k,j+1,i+1) ) |
---|
| 754 | temp(ke,2*j,2*i+1) = 0.25_wp * ( p(k,j,i) + p(k,j,i+1) + & |
---|
| 755 | p(kp1,j,i) + p(kp1,j,i+1) ) |
---|
| 756 | temp(ke,2*j+1,2*i) = 0.25_wp * ( p(k,j,i) + p(k,j+1,i) + & |
---|
| 757 | p(kp1,j,i) + p(kp1,j+1,i) ) |
---|
| 758 | ! |
---|
| 759 | !-- Points in the middle of coarse grid cube |
---|
| 760 | temp(ke,2*j+1,2*i+1) = 0.125_wp * & |
---|
| 761 | ( p(k,j,i) + p(k,j,i+1) + & |
---|
| 762 | p(k,j+1,i) + p(k,j+1,i+1) + & |
---|
| 763 | p(kp1,j,i) + p(kp1,j,i+1) + & |
---|
| 764 | p(kp1,j+1,i) + p(kp1,j+1,i+1) ) |
---|
| 765 | ENDDO |
---|
| 766 | |
---|
| 767 | ENDDO |
---|
| 768 | ENDDO |
---|
| 769 | !$OMP END PARALLEL |
---|
| 770 | |
---|
| 771 | ind_even_odd = even_odd_level(grid_level) |
---|
| 772 | ! |
---|
| 773 | !-- Horizontal boundary conditions |
---|
| 774 | CALL exchange_horiz( temp, 1) |
---|
| 775 | |
---|
| 776 | IF ( .NOT. bc_lr_cyc ) THEN |
---|
| 777 | IF (inflow_l .OR. outflow_l) temp(:,:,nxl_mg(l)-1) = temp(:,:,nxl_mg(l)) |
---|
| 778 | IF (inflow_r .OR. outflow_r) temp(:,:,nxr_mg(l)+1) = temp(:,:,nxr_mg(l)) |
---|
| 779 | ENDIF |
---|
| 780 | |
---|
| 781 | IF ( .NOT. bc_ns_cyc ) THEN |
---|
| 782 | IF (inflow_n .OR. outflow_n) temp(:,nyn_mg(l)+1,:) = temp(:,nyn_mg(l),:) |
---|
| 783 | IF (inflow_s .OR. outflow_s) temp(:,nys_mg(l)-1,:) = temp(:,nys_mg(l),:) |
---|
| 784 | ENDIF |
---|
| 785 | |
---|
| 786 | ! |
---|
| 787 | !-- Bottom and top boundary conditions |
---|
| 788 | IF ( ibc_p_b == 1 ) THEN |
---|
| 789 | temp(nzb,:,: ) = temp(ind_even_odd+1,:,:) |
---|
| 790 | ELSE |
---|
| 791 | temp(nzb,:,: ) = 0.0_wp |
---|
| 792 | ENDIF |
---|
| 793 | |
---|
| 794 | IF ( ibc_p_t == 1 ) THEN |
---|
| 795 | temp(nzt_mg(l)+1,:,: ) = temp(ind_even_odd,:,:) |
---|
| 796 | ELSE |
---|
| 797 | temp(nzt_mg(l)+1,:,: ) = 0.0_wp |
---|
| 798 | ENDIF |
---|
| 799 | |
---|
| 800 | CALL cpu_log( log_point_s(55), 'prolong', 'stop' ) |
---|
| 801 | |
---|
| 802 | END SUBROUTINE prolong_fast |
---|
| 803 | |
---|
| 804 | |
---|
| 805 | |
---|
| 806 | SUBROUTINE redblack_fast( f_mg, p_mg ) |
---|
| 807 | |
---|
| 808 | !------------------------------------------------------------------------------! |
---|
| 809 | ! Description: |
---|
| 810 | ! ------------ |
---|
| 811 | ! Relaxation method for the multigrid scheme. A Gauss-Seidel iteration with |
---|
| 812 | ! 3D-Red-Black decomposition (GS-RB) is used. |
---|
| 813 | !------------------------------------------------------------------------------! |
---|
| 814 | |
---|
| 815 | USE arrays_3d, & |
---|
| 816 | ONLY: f1_mg, f2_mg, f3_mg |
---|
| 817 | |
---|
| 818 | USE control_parameters, & |
---|
| 819 | ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, inflow_l,& |
---|
| 820 | inflow_n, inflow_r, inflow_s, masking_method, ngsrb, & |
---|
| 821 | outflow_l, outflow_n, outflow_r, outflow_s, topography |
---|
| 822 | |
---|
| 823 | USE grid_variables, & |
---|
| 824 | ONLY: ddx2_mg, ddy2_mg |
---|
| 825 | |
---|
| 826 | USE indices, & |
---|
| 827 | ONLY: flags, wall_flags_1, wall_flags_2, wall_flags_3, & |
---|
| 828 | wall_flags_4, wall_flags_5, wall_flags_6, wall_flags_7, & |
---|
| 829 | wall_flags_8, wall_flags_9, wall_flags_10, nxl_mg, nxr_mg, & |
---|
| 830 | nys_mg, nyn_mg, nzb, nzt_mg |
---|
| 831 | |
---|
| 832 | IMPLICIT NONE |
---|
| 833 | |
---|
| 834 | INTEGER(iwp) :: color !: |
---|
| 835 | INTEGER(iwp) :: i !: |
---|
| 836 | INTEGER(iwp) :: ic !: |
---|
| 837 | INTEGER(iwp) :: j !: |
---|
| 838 | INTEGER(iwp) :: jc !: |
---|
| 839 | INTEGER(iwp) :: jj !: |
---|
| 840 | INTEGER(iwp) :: k !: |
---|
| 841 | INTEGER(iwp) :: l !: |
---|
| 842 | INTEGER(iwp) :: n !: |
---|
| 843 | INTEGER(iwp) :: km1 !: |
---|
| 844 | INTEGER(iwp) :: kp1 !: |
---|
| 845 | |
---|
| 846 | LOGICAL :: unroll !: |
---|
| 847 | |
---|
| 848 | REAL(wp) :: wall_left !: |
---|
| 849 | REAL(wp) :: wall_north !: |
---|
| 850 | REAL(wp) :: wall_right !: |
---|
| 851 | REAL(wp) :: wall_south !: |
---|
| 852 | REAL(wp) :: wall_total !: |
---|
| 853 | REAL(wp) :: wall_top !: |
---|
| 854 | |
---|
| 855 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 856 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 857 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f_mg !: |
---|
| 858 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 859 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 860 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p_mg !: |
---|
| 861 | |
---|
| 862 | l = grid_level |
---|
| 863 | |
---|
| 864 | ! |
---|
| 865 | !-- Choose flag array of this level |
---|
| 866 | IF ( topography /= 'flat' .AND. .NOT. masking_method ) THEN |
---|
| 867 | |
---|
| 868 | SELECT CASE ( l ) |
---|
| 869 | CASE ( 1 ) |
---|
| 870 | flags => wall_flags_1 |
---|
| 871 | CASE ( 2 ) |
---|
| 872 | flags => wall_flags_2 |
---|
| 873 | CASE ( 3 ) |
---|
| 874 | flags => wall_flags_3 |
---|
| 875 | CASE ( 4 ) |
---|
| 876 | flags => wall_flags_4 |
---|
| 877 | CASE ( 5 ) |
---|
| 878 | flags => wall_flags_5 |
---|
| 879 | CASE ( 6 ) |
---|
| 880 | flags => wall_flags_6 |
---|
| 881 | CASE ( 7 ) |
---|
| 882 | flags => wall_flags_7 |
---|
| 883 | CASE ( 8 ) |
---|
| 884 | flags => wall_flags_8 |
---|
| 885 | CASE ( 9 ) |
---|
| 886 | flags => wall_flags_9 |
---|
| 887 | CASE ( 10 ) |
---|
| 888 | flags => wall_flags_10 |
---|
| 889 | END SELECT |
---|
| 890 | |
---|
| 891 | ENDIF |
---|
| 892 | |
---|
| 893 | unroll = ( MOD( nyn_mg(l)-nys_mg(l)+1, 4 ) == 0 .AND. & |
---|
| 894 | MOD( nxr_mg(l)-nxl_mg(l)+1, 2 ) == 0 ) |
---|
| 895 | |
---|
| 896 | DO n = 1, ngsrb |
---|
| 897 | |
---|
| 898 | DO color = 1, 2 |
---|
| 899 | ! |
---|
| 900 | !-- No wall treatment in case of flat topography |
---|
| 901 | IF ( topography == 'flat' .OR. masking_method ) THEN |
---|
| 902 | |
---|
| 903 | IF ( .NOT. unroll ) THEN |
---|
| 904 | |
---|
| 905 | CALL cpu_log( log_point_s(36), 'redblack_no_unroll_f', 'start' ) |
---|
| 906 | ! |
---|
| 907 | !-- Without unrolling of loops, no cache optimization |
---|
| 908 | !$OMP PARALLEL PRIVATE (i,j,k,km1,kp1) |
---|
| 909 | !$OMP DO |
---|
| 910 | DO i = nxl_mg(l), nxr_mg(l), 2 |
---|
| 911 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
---|
| 912 | !DIR$ IVDEP |
---|
| 913 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 914 | km1 = k-ind_even_odd-1 |
---|
| 915 | kp1 = k-ind_even_odd |
---|
| 916 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
---|
| 917 | ddx2_mg(l) * & |
---|
| 918 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 919 | + ddy2_mg(l) * & |
---|
| 920 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 921 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 922 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 923 | - f_mg(k,j,i) ) |
---|
| 924 | ENDDO |
---|
| 925 | ENDDO |
---|
| 926 | ENDDO |
---|
| 927 | |
---|
| 928 | !$OMP DO |
---|
| 929 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
---|
| 930 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
---|
| 931 | !DIR$ IVDEP |
---|
| 932 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 933 | km1 = k-ind_even_odd-1 |
---|
| 934 | kp1 = k-ind_even_odd |
---|
| 935 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
---|
| 936 | ddx2_mg(l) * & |
---|
| 937 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 938 | + ddy2_mg(l) * & |
---|
| 939 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 940 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 941 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 942 | - f_mg(k,j,i) ) |
---|
| 943 | ENDDO |
---|
| 944 | ENDDO |
---|
| 945 | ENDDO |
---|
| 946 | |
---|
| 947 | !$OMP DO |
---|
| 948 | DO i = nxl_mg(l), nxr_mg(l), 2 |
---|
| 949 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
---|
| 950 | !DIR$ IVDEP |
---|
| 951 | DO k = nzb+1, ind_even_odd |
---|
| 952 | km1 = k+ind_even_odd |
---|
| 953 | kp1 = k+ind_even_odd+1 |
---|
| 954 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
---|
| 955 | ddx2_mg(l) * & |
---|
| 956 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 957 | + ddy2_mg(l) * & |
---|
| 958 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 959 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 960 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 961 | - f_mg(k,j,i) ) |
---|
| 962 | ENDDO |
---|
| 963 | ENDDO |
---|
| 964 | ENDDO |
---|
| 965 | |
---|
| 966 | !$OMP DO |
---|
| 967 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
---|
| 968 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
---|
| 969 | !DIR$ IVDEP |
---|
| 970 | DO k = nzb+1, ind_even_odd |
---|
| 971 | km1 = k+ind_even_odd |
---|
| 972 | kp1 = k+ind_even_odd+1 |
---|
| 973 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
---|
| 974 | ddx2_mg(l) * & |
---|
| 975 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 976 | + ddy2_mg(l) * & |
---|
| 977 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 978 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 979 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 980 | - f_mg(k,j,i) ) |
---|
| 981 | ENDDO |
---|
| 982 | ENDDO |
---|
| 983 | ENDDO |
---|
| 984 | !$OMP END PARALLEL |
---|
| 985 | |
---|
| 986 | CALL cpu_log( log_point_s(36), 'redblack_no_unroll_f', 'stop' ) |
---|
| 987 | |
---|
| 988 | ELSE |
---|
| 989 | ! |
---|
| 990 | !-- Loop unrolling along y, only one i loop for better cache use |
---|
| 991 | CALL cpu_log( log_point_s(38), 'redblack_unroll_f', 'start' ) |
---|
| 992 | |
---|
| 993 | !$OMP PARALLEL PRIVATE (i,j,k,ic,jc,km1,kp1,jj) |
---|
| 994 | !$OMP DO |
---|
| 995 | DO ic = nxl_mg(l), nxr_mg(l), 2 |
---|
| 996 | DO jc = nys_mg(l), nyn_mg(l), 4 |
---|
| 997 | i = ic |
---|
| 998 | jj = jc+2-color |
---|
| 999 | !DIR$ IVDEP |
---|
| 1000 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 1001 | km1 = k-ind_even_odd-1 |
---|
| 1002 | kp1 = k-ind_even_odd |
---|
| 1003 | j = jj |
---|
| 1004 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
---|
| 1005 | ddx2_mg(l) * & |
---|
| 1006 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 1007 | + ddy2_mg(l) * & |
---|
| 1008 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 1009 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 1010 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 1011 | - f_mg(k,j,i) ) |
---|
| 1012 | j = jj+2 |
---|
| 1013 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
---|
| 1014 | ddx2_mg(l) * & |
---|
| 1015 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 1016 | + ddy2_mg(l) * & |
---|
| 1017 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 1018 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 1019 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 1020 | - f_mg(k,j,i) ) |
---|
| 1021 | ENDDO |
---|
| 1022 | |
---|
| 1023 | i = ic+1 |
---|
| 1024 | jj = jc+color-1 |
---|
| 1025 | !DIR$ IVDEP |
---|
| 1026 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 1027 | km1 = k-ind_even_odd-1 |
---|
| 1028 | kp1 = k-ind_even_odd |
---|
| 1029 | j = jj |
---|
| 1030 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
---|
| 1031 | ddx2_mg(l) * & |
---|
| 1032 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 1033 | + ddy2_mg(l) * & |
---|
| 1034 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 1035 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 1036 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 1037 | - f_mg(k,j,i) ) |
---|
| 1038 | j = jj+2 |
---|
| 1039 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
---|
| 1040 | ddx2_mg(l) * & |
---|
| 1041 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 1042 | + ddy2_mg(l) * & |
---|
| 1043 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 1044 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 1045 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 1046 | - f_mg(k,j,i) ) |
---|
| 1047 | ENDDO |
---|
| 1048 | |
---|
| 1049 | i = ic |
---|
| 1050 | jj = jc+color-1 |
---|
| 1051 | !DIR$ IVDEP |
---|
| 1052 | DO k = nzb+1, ind_even_odd |
---|
| 1053 | km1 = k+ind_even_odd |
---|
| 1054 | kp1 = k+ind_even_odd+1 |
---|
| 1055 | j = jj |
---|
| 1056 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
---|
| 1057 | ddx2_mg(l) * & |
---|
| 1058 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 1059 | + ddy2_mg(l) * & |
---|
| 1060 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 1061 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 1062 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 1063 | - f_mg(k,j,i) ) |
---|
| 1064 | j = jj+2 |
---|
| 1065 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
---|
| 1066 | ddx2_mg(l) * & |
---|
| 1067 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 1068 | + ddy2_mg(l) * & |
---|
| 1069 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 1070 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 1071 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 1072 | - f_mg(k,j,i) ) |
---|
| 1073 | ENDDO |
---|
| 1074 | |
---|
| 1075 | i = ic+1 |
---|
| 1076 | jj = jc+2-color |
---|
| 1077 | !DIR$ IVDEP |
---|
| 1078 | DO k = nzb+1, ind_even_odd |
---|
| 1079 | km1 = k+ind_even_odd |
---|
| 1080 | kp1 = k+ind_even_odd+1 |
---|
| 1081 | j = jj |
---|
| 1082 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
---|
| 1083 | ddx2_mg(l) * & |
---|
| 1084 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 1085 | + ddy2_mg(l) * & |
---|
| 1086 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 1087 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 1088 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 1089 | - f_mg(k,j,i) ) |
---|
| 1090 | j = jj+2 |
---|
| 1091 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
---|
| 1092 | ddx2_mg(l) * & |
---|
| 1093 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
---|
| 1094 | + ddy2_mg(l) * & |
---|
| 1095 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
---|
| 1096 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
---|
| 1097 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
---|
| 1098 | - f_mg(k,j,i) ) |
---|
| 1099 | ENDDO |
---|
| 1100 | |
---|
| 1101 | ENDDO |
---|
| 1102 | ENDDO |
---|
| 1103 | !$OMP END PARALLEL |
---|
| 1104 | |
---|
| 1105 | CALL cpu_log( log_point_s(38), 'redblack_unroll_f', 'stop' ) |
---|
| 1106 | |
---|
| 1107 | ENDIF |
---|
| 1108 | |
---|
| 1109 | ELSE |
---|
| 1110 | |
---|
| 1111 | IF ( .NOT. unroll ) THEN |
---|
| 1112 | |
---|
| 1113 | CALL cpu_log( log_point_s(36), 'redblack_no_unroll', 'start' ) |
---|
| 1114 | ! |
---|
| 1115 | !-- Without unrolling of loops, no cache optimization / |
---|
| 1116 | !-- vectorization. Therefore, the non-vectorized IBITS function |
---|
| 1117 | !-- is still used. |
---|
| 1118 | !$OMP PARALLEL PRIVATE (i,j,k,km1,kp1) |
---|
| 1119 | !$OMP DO |
---|
| 1120 | DO i = nxl_mg(l), nxr_mg(l), 2 |
---|
| 1121 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
---|
| 1122 | !DIR$ IVDEP |
---|
| 1123 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 1124 | km1 = k-ind_even_odd-1 |
---|
| 1125 | kp1 = k-ind_even_odd |
---|
| 1126 | |
---|
| 1127 | p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( & |
---|
| 1128 | ddx2_mg(l) * & |
---|
| 1129 | ( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * & |
---|
| 1130 | ( p_mg(k,j,i) - p_mg(k,j,i+1) ) + & |
---|
| 1131 | p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * & |
---|
| 1132 | ( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) & |
---|
| 1133 | + ddy2_mg(l) * & |
---|
| 1134 | ( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * & |
---|
| 1135 | ( p_mg(k,j,i) - p_mg(k,j+1,i) ) + & |
---|
| 1136 | p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * & |
---|
| 1137 | ( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) & |
---|
| 1138 | + f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 1139 | + f3_mg(k,l) * & |
---|
| 1140 | ( p_mg(km1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * & |
---|
| 1141 | ( p_mg(k,j,i) - p_mg(km1,j,i) ) ) & |
---|
| 1142 | - f_mg(k,j,i) ) |
---|
| 1143 | ENDDO |
---|
| 1144 | ENDDO |
---|
| 1145 | ENDDO |
---|
| 1146 | |
---|
| 1147 | !$OMP DO |
---|
| 1148 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
---|
| 1149 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
---|
| 1150 | !DIR$ IVDEP |
---|
| 1151 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 1152 | km1 = k-ind_even_odd-1 |
---|
| 1153 | kp1 = k-ind_even_odd |
---|
| 1154 | |
---|
| 1155 | p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( & |
---|
| 1156 | ddx2_mg(l) * & |
---|
| 1157 | ( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * & |
---|
| 1158 | ( p_mg(k,j,i) - p_mg(k,j,i+1) ) + & |
---|
| 1159 | p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * & |
---|
| 1160 | ( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) & |
---|
| 1161 | + ddy2_mg(l) * & |
---|
| 1162 | ( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * & |
---|
| 1163 | ( p_mg(k,j,i) - p_mg(k,j+1,i) ) + & |
---|
| 1164 | p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * & |
---|
| 1165 | ( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) & |
---|
| 1166 | + f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 1167 | + f3_mg(k,l) * & |
---|
| 1168 | ( p_mg(km1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * & |
---|
| 1169 | ( p_mg(k,j,i) - p_mg(km1,j,i) ) ) & |
---|
| 1170 | - f_mg(k,j,i) ) |
---|
| 1171 | ENDDO |
---|
| 1172 | ENDDO |
---|
| 1173 | ENDDO |
---|
| 1174 | |
---|
| 1175 | !$OMP DO |
---|
| 1176 | DO i = nxl_mg(l), nxr_mg(l), 2 |
---|
| 1177 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
---|
| 1178 | !DIR$ IVDEP |
---|
| 1179 | DO k = nzb+1, ind_even_odd |
---|
| 1180 | km1 = k+ind_even_odd |
---|
| 1181 | kp1 = k+ind_even_odd+1 |
---|
| 1182 | p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( & |
---|
| 1183 | ddx2_mg(l) * & |
---|
| 1184 | ( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * & |
---|
| 1185 | ( p_mg(k,j,i) - p_mg(k,j,i+1) ) + & |
---|
| 1186 | p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * & |
---|
| 1187 | ( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) & |
---|
| 1188 | + ddy2_mg(l) * & |
---|
| 1189 | ( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * & |
---|
| 1190 | ( p_mg(k,j,i) - p_mg(k,j+1,i) ) + & |
---|
| 1191 | p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * & |
---|
| 1192 | ( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) & |
---|
| 1193 | + f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 1194 | + f3_mg(k,l) * & |
---|
| 1195 | ( p_mg(km1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * & |
---|
| 1196 | ( p_mg(k,j,i) - p_mg(km1,j,i) ) ) & |
---|
| 1197 | - f_mg(k,j,i) ) |
---|
| 1198 | ENDDO |
---|
| 1199 | ENDDO |
---|
| 1200 | ENDDO |
---|
| 1201 | |
---|
| 1202 | !$OMP DO |
---|
| 1203 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
---|
| 1204 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
---|
| 1205 | !DIR$ IVDEP |
---|
| 1206 | DO k = nzb+1, ind_even_odd |
---|
| 1207 | km1 = k+ind_even_odd |
---|
| 1208 | kp1 = k+ind_even_odd+1 |
---|
| 1209 | |
---|
| 1210 | p_mg(k,j,i) = 1.0_wp / f1_mg(k,l) * ( & |
---|
| 1211 | ddx2_mg(l) * & |
---|
| 1212 | ( p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * & |
---|
| 1213 | ( p_mg(k,j,i) - p_mg(k,j,i+1) ) + & |
---|
| 1214 | p_mg(k,j,i-1) + IBITS( flags(k,j,i), 4, 1 ) * & |
---|
| 1215 | ( p_mg(k,j,i) - p_mg(k,j,i-1) ) ) & |
---|
| 1216 | + ddy2_mg(l) * & |
---|
| 1217 | ( p_mg(k,j+1,i) + IBITS( flags(k,j,i), 3, 1 ) * & |
---|
| 1218 | ( p_mg(k,j,i) - p_mg(k,j+1,i) ) + & |
---|
| 1219 | p_mg(k,j-1,i) + IBITS( flags(k,j,i), 2, 1 ) * & |
---|
| 1220 | ( p_mg(k,j,i) - p_mg(k,j-1,i) ) ) & |
---|
| 1221 | + f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 1222 | + f3_mg(k,l) * & |
---|
| 1223 | ( p_mg(km1,j,i) + IBITS( flags(k,j,i), 0, 1 ) * & |
---|
| 1224 | ( p_mg(k,j,i) - p_mg(km1,j,i) ) ) & |
---|
| 1225 | - f_mg(k,j,i) ) |
---|
| 1226 | ENDDO |
---|
| 1227 | ENDDO |
---|
| 1228 | ENDDO |
---|
| 1229 | !$OMP END PARALLEL |
---|
| 1230 | |
---|
| 1231 | CALL cpu_log( log_point_s(36), 'redblack_no_unroll', 'stop' ) |
---|
| 1232 | |
---|
| 1233 | ELSE |
---|
| 1234 | ! |
---|
| 1235 | !-- Loop unrolling along y, only one i loop for better cache use |
---|
| 1236 | CALL cpu_log( log_point_s(38), 'redblack_unroll', 'start' ) |
---|
| 1237 | |
---|
| 1238 | !$OMP PARALLEL PRIVATE (i,j,k,ic,jc,km1,kp1,jj) |
---|
| 1239 | !$OMP DO |
---|
| 1240 | DO ic = nxl_mg(l), nxr_mg(l), 2 |
---|
| 1241 | DO jc = nys_mg(l), nyn_mg(l), 4 |
---|
| 1242 | i = ic |
---|
| 1243 | jj = jc+2-color |
---|
| 1244 | !DIR$ IVDEP |
---|
| 1245 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 1246 | |
---|
| 1247 | km1 = k-ind_even_odd-1 |
---|
| 1248 | kp1 = k-ind_even_odd |
---|
| 1249 | |
---|
| 1250 | j = jj |
---|
| 1251 | p_mg(k,j,i) = 1.0 / f1_mg(k,l) * ( & |
---|
| 1252 | ddx2_mg(l) * & |
---|
| 1253 | ( MERGE(p_mg(k,j,i),p_mg(k,j,i+1),BTEST( flags(k,j,i), 5)) & |
---|
| 1254 | + MERGE(p_mg(k,j,i),p_mg(k,j,i-1),BTEST( flags(k,j,i), 4)) ) & |
---|
| 1255 | + ddy2_mg(l) * & |
---|
| 1256 | ( MERGE(p_mg(k,j,i),p_mg(k,j+1,i),BTEST( flags(k,j,i), 3)) & |
---|
| 1257 | + MERGE(p_mg(k,j,i),p_mg(k,j-1,i),BTEST( flags(k,j,i), 2)) ) & |
---|
| 1258 | + f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 1259 | + f3_mg(k,l) * & |
---|
| 1260 | MERGE(p_mg(k,j,i),p_mg(km1,j,i),BTEST( flags(k,j,i), 0)) & |
---|
| 1261 | - f_mg(k,j,i) ) |
---|
| 1262 | j = jj+2 |
---|
| 1263 | p_mg(k,j,i) = 1.0 / f1_mg(k,l) * ( & |
---|
| 1264 | ddx2_mg(l) * & |
---|
| 1265 | ( MERGE(p_mg(k,j,i),p_mg(k,j,i+1),BTEST( flags(k,j,i), 5)) & |
---|
| 1266 | + MERGE(p_mg(k,j,i),p_mg(k,j,i-1),BTEST( flags(k,j,i), 4)) ) & |
---|
| 1267 | + ddy2_mg(l) * & |
---|
| 1268 | ( MERGE(p_mg(k,j,i),p_mg(k,j+1,i),BTEST( flags(k,j,i), 3)) & |
---|
| 1269 | + MERGE(p_mg(k,j,i),p_mg(k,j-1,i),BTEST( flags(k,j,i), 2)) ) & |
---|
| 1270 | + f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 1271 | + f3_mg(k,l) * & |
---|
| 1272 | MERGE(p_mg(k,j,i),p_mg(km1,j,i),BTEST( flags(k,j,i), 0)) & |
---|
| 1273 | - f_mg(k,j,i) ) |
---|
| 1274 | |
---|
| 1275 | ENDDO |
---|
| 1276 | |
---|
| 1277 | i = ic+1 |
---|
| 1278 | jj = jc+color-1 |
---|
| 1279 | !DIR$ IVDEP |
---|
| 1280 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 1281 | |
---|
| 1282 | km1 = k-ind_even_odd-1 |
---|
| 1283 | kp1 = k-ind_even_odd |
---|
| 1284 | |
---|
| 1285 | j =jj |
---|
| 1286 | p_mg(k,j,i) = 1.0 / f1_mg(k,l) * ( & |
---|
| 1287 | ddx2_mg(l) * & |
---|
| 1288 | ( MERGE(p_mg(k,j,i),p_mg(k,j,i+1),BTEST( flags(k,j,i), 5)) & |
---|
| 1289 | + MERGE(p_mg(k,j,i),p_mg(k,j,i-1),BTEST( flags(k,j,i), 4)) ) & |
---|
| 1290 | + ddy2_mg(l) * & |
---|
| 1291 | ( MERGE(p_mg(k,j,i),p_mg(k,j+1,i),BTEST( flags(k,j,i), 3)) & |
---|
| 1292 | + MERGE(p_mg(k,j,i),p_mg(k,j-1,i),BTEST( flags(k,j,i), 2)) ) & |
---|
| 1293 | + f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 1294 | + f3_mg(k,l) * & |
---|
| 1295 | MERGE(p_mg(k,j,i),p_mg(km1,j,i),BTEST( flags(k,j,i), 0)) & |
---|
| 1296 | - f_mg(k,j,i) ) |
---|
| 1297 | |
---|
| 1298 | j = jj+2 |
---|
| 1299 | p_mg(k,j,i) = 1.0 / f1_mg(k,l) * ( & |
---|
| 1300 | ddx2_mg(l) * & |
---|
| 1301 | ( MERGE(p_mg(k,j,i),p_mg(k,j,i+1),BTEST( flags(k,j,i), 5)) & |
---|
| 1302 | + MERGE(p_mg(k,j,i),p_mg(k,j,i-1),BTEST( flags(k,j,i), 4)) ) & |
---|
| 1303 | + ddy2_mg(l) * & |
---|
| 1304 | ( MERGE(p_mg(k,j,i),p_mg(k,j+1,i),BTEST( flags(k,j,i), 3)) & |
---|
| 1305 | + MERGE(p_mg(k,j,i),p_mg(k,j-1,i),BTEST( flags(k,j,i), 2)) ) & |
---|
| 1306 | + f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 1307 | + f3_mg(k,l) * & |
---|
| 1308 | MERGE(p_mg(k,j,i),p_mg(km1,j,i),BTEST( flags(k,j,i), 0)) & |
---|
| 1309 | - f_mg(k,j,i) ) |
---|
| 1310 | |
---|
| 1311 | ENDDO |
---|
| 1312 | |
---|
| 1313 | i = ic |
---|
| 1314 | jj = jc+color-1 |
---|
| 1315 | !DIR$ IVDEP |
---|
| 1316 | DO k = nzb+1, ind_even_odd |
---|
| 1317 | |
---|
| 1318 | km1 = k+ind_even_odd |
---|
| 1319 | kp1 = k+ind_even_odd+1 |
---|
| 1320 | |
---|
| 1321 | j =jj |
---|
| 1322 | p_mg(k,j,i) = 1.0 / f1_mg(k,l) * ( & |
---|
| 1323 | ddx2_mg(l) * & |
---|
| 1324 | ( MERGE(p_mg(k,j,i),p_mg(k,j,i+1),BTEST( flags(k,j,i), 5)) & |
---|
| 1325 | + MERGE(p_mg(k,j,i),p_mg(k,j,i-1),BTEST( flags(k,j,i), 4)) ) & |
---|
| 1326 | + ddy2_mg(l) * & |
---|
| 1327 | ( MERGE(p_mg(k,j,i),p_mg(k,j+1,i),BTEST( flags(k,j,i), 3)) & |
---|
| 1328 | + MERGE(p_mg(k,j,i),p_mg(k,j-1,i),BTEST( flags(k,j,i), 2)) ) & |
---|
| 1329 | + f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 1330 | + f3_mg(k,l) * & |
---|
| 1331 | MERGE(p_mg(k,j,i),p_mg(km1,j,i),BTEST( flags(k,j,i), 0)) & |
---|
| 1332 | - f_mg(k,j,i) ) |
---|
| 1333 | |
---|
| 1334 | j = jj+2 |
---|
| 1335 | p_mg(k,j,i) = 1.0 / f1_mg(k,l) * ( & |
---|
| 1336 | ddx2_mg(l) * & |
---|
| 1337 | ( MERGE(p_mg(k,j,i),p_mg(k,j,i+1),BTEST( flags(k,j,i), 5)) & |
---|
| 1338 | + MERGE(p_mg(k,j,i),p_mg(k,j,i-1),BTEST( flags(k,j,i), 4)) ) & |
---|
| 1339 | + ddy2_mg(l) * & |
---|
| 1340 | ( MERGE(p_mg(k,j,i),p_mg(k,j+1,i),BTEST( flags(k,j,i), 3)) & |
---|
| 1341 | + MERGE(p_mg(k,j,i),p_mg(k,j-1,i),BTEST( flags(k,j,i), 2)) ) & |
---|
| 1342 | + f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 1343 | + f3_mg(k,l) * & |
---|
| 1344 | MERGE(p_mg(k,j,i),p_mg(km1,j,i),BTEST( flags(k,j,i), 0)) & |
---|
| 1345 | - f_mg(k,j,i) ) |
---|
| 1346 | |
---|
| 1347 | ENDDO |
---|
| 1348 | |
---|
| 1349 | i = ic+1 |
---|
| 1350 | jj = jc+2-color |
---|
| 1351 | !DIR$ IVDEP |
---|
| 1352 | DO k = nzb+1, ind_even_odd |
---|
| 1353 | |
---|
| 1354 | km1 = k+ind_even_odd |
---|
| 1355 | kp1 = k+ind_even_odd+1 |
---|
| 1356 | |
---|
| 1357 | j =jj |
---|
| 1358 | p_mg(k,j,i) = 1.0 / f1_mg(k,l) * ( & |
---|
| 1359 | ddx2_mg(l) * & |
---|
| 1360 | ( MERGE(p_mg(k,j,i),p_mg(k,j,i+1),BTEST( flags(k,j,i), 5)) & |
---|
| 1361 | + MERGE(p_mg(k,j,i),p_mg(k,j,i-1),BTEST( flags(k,j,i), 4)) ) & |
---|
| 1362 | + ddy2_mg(l) * & |
---|
| 1363 | ( MERGE(p_mg(k,j,i),p_mg(k,j+1,i),BTEST( flags(k,j,i), 3)) & |
---|
| 1364 | + MERGE(p_mg(k,j,i),p_mg(k,j-1,i),BTEST( flags(k,j,i), 2)) ) & |
---|
| 1365 | + f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 1366 | + f3_mg(k,l) * & |
---|
| 1367 | MERGE(p_mg(k,j,i),p_mg(km1,j,i),BTEST( flags(k,j,i), 0)) & |
---|
| 1368 | - f_mg(k,j,i) ) |
---|
| 1369 | |
---|
| 1370 | j = jj+2 |
---|
| 1371 | p_mg(k,j,i) = 1.0 / f1_mg(k,l) * ( & |
---|
| 1372 | ddx2_mg(l) * & |
---|
| 1373 | ( MERGE(p_mg(k,j,i),p_mg(k,j,i+1),BTEST( flags(k,j,i), 5)) & |
---|
| 1374 | + MERGE(p_mg(k,j,i),p_mg(k,j,i-1),BTEST( flags(k,j,i), 4)) ) & |
---|
| 1375 | + ddy2_mg(l) * & |
---|
| 1376 | ( MERGE(p_mg(k,j,i),p_mg(k,j+1,i),BTEST( flags(k,j,i), 3)) & |
---|
| 1377 | + MERGE(p_mg(k,j,i),p_mg(k,j-1,i),BTEST( flags(k,j,i), 2)) ) & |
---|
| 1378 | + f2_mg(k,l) * p_mg(kp1,j,i) & |
---|
| 1379 | + f3_mg(k,l) * & |
---|
| 1380 | MERGE(p_mg(k,j,i),p_mg(km1,j,i),BTEST( flags(k,j,i), 0)) & |
---|
| 1381 | - f_mg(k,j,i) ) |
---|
| 1382 | |
---|
| 1383 | ENDDO |
---|
| 1384 | |
---|
| 1385 | ENDDO |
---|
| 1386 | ENDDO |
---|
| 1387 | !$OMP END PARALLEL |
---|
| 1388 | |
---|
| 1389 | CALL cpu_log( log_point_s(38), 'redblack_unroll', 'stop' ) |
---|
| 1390 | |
---|
| 1391 | ENDIF |
---|
| 1392 | |
---|
| 1393 | ENDIF |
---|
| 1394 | |
---|
| 1395 | ! |
---|
| 1396 | !-- Horizontal boundary conditions |
---|
| 1397 | CALL special_exchange_horiz( p_mg, color ) |
---|
| 1398 | |
---|
| 1399 | IF ( .NOT. bc_lr_cyc ) THEN |
---|
| 1400 | IF ( inflow_l .OR. outflow_l ) THEN |
---|
| 1401 | p_mg(:,:,nxl_mg(l)-1) = p_mg(:,:,nxl_mg(l)) |
---|
| 1402 | ENDIF |
---|
| 1403 | IF ( inflow_r .OR. outflow_r ) THEN |
---|
| 1404 | p_mg(:,:,nxr_mg(l)+1) = p_mg(:,:,nxr_mg(l)) |
---|
| 1405 | ENDIF |
---|
| 1406 | ENDIF |
---|
| 1407 | |
---|
| 1408 | IF ( .NOT. bc_ns_cyc ) THEN |
---|
| 1409 | IF ( inflow_n .OR. outflow_n ) THEN |
---|
| 1410 | p_mg(:,nyn_mg(l)+1,:) = p_mg(:,nyn_mg(l),:) |
---|
| 1411 | ENDIF |
---|
| 1412 | IF ( inflow_s .OR. outflow_s ) THEN |
---|
| 1413 | p_mg(:,nys_mg(l)-1,:) = p_mg(:,nys_mg(l),:) |
---|
| 1414 | ENDIF |
---|
| 1415 | ENDIF |
---|
| 1416 | |
---|
| 1417 | ! |
---|
| 1418 | !-- Bottom and top boundary conditions |
---|
| 1419 | IF ( ibc_p_b == 1 ) THEN |
---|
| 1420 | p_mg(nzb,:,: ) = p_mg(ind_even_odd+1,:,:) |
---|
| 1421 | ELSE |
---|
| 1422 | p_mg(nzb,:,: ) = 0.0_wp |
---|
| 1423 | ENDIF |
---|
| 1424 | |
---|
| 1425 | IF ( ibc_p_t == 1 ) THEN |
---|
| 1426 | p_mg(nzt_mg(l)+1,:,: ) = p_mg(ind_even_odd,:,:) |
---|
| 1427 | ELSE |
---|
| 1428 | p_mg(nzt_mg(l)+1,:,: ) = 0.0_wp |
---|
| 1429 | ENDIF |
---|
| 1430 | |
---|
| 1431 | ENDDO |
---|
| 1432 | ENDDO |
---|
| 1433 | ! |
---|
| 1434 | !-- Set pressure within topography and at the topography surfaces |
---|
| 1435 | IF ( topography /= 'flat' .AND. .NOT. masking_method ) THEN |
---|
| 1436 | |
---|
| 1437 | !$OMP PARALLEL PRIVATE (i,j,k,kp1,wall_left,wall_north,wall_right, & |
---|
| 1438 | !$OMP wall_south,wall_top,wall_total) |
---|
| 1439 | !$OMP DO |
---|
| 1440 | DO i = nxl_mg(l), nxr_mg(l) |
---|
| 1441 | DO j = nys_mg(l), nyn_mg(l) |
---|
| 1442 | !DIR$ IVDEP |
---|
| 1443 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 1444 | km1 = k-ind_even_odd-1 |
---|
| 1445 | kp1 = k-ind_even_odd |
---|
| 1446 | ! |
---|
| 1447 | !-- First, set pressure inside topography to zero |
---|
| 1448 | p_mg(k,j,i) = p_mg(k,j,i) * & |
---|
| 1449 | ( 1.0_wp - IBITS( flags(k,j,i), 6, 1 ) ) |
---|
| 1450 | ! |
---|
| 1451 | !-- Second, determine if the gridpoint inside topography is |
---|
| 1452 | !-- adjacent to a wall and set its value to a value given by the |
---|
| 1453 | !-- average of those values obtained from Neumann boundary |
---|
| 1454 | !-- condition |
---|
| 1455 | wall_left = IBITS( flags(k,j,i-1), 5, 1 ) |
---|
| 1456 | wall_right = IBITS( flags(k,j,i+1), 4, 1 ) |
---|
| 1457 | wall_south = IBITS( flags(k,j-1,i), 3, 1 ) |
---|
| 1458 | wall_north = IBITS( flags(k,j+1,i), 2, 1 ) |
---|
| 1459 | wall_top = IBITS( flags(kp1,j,i), 0, 1 ) |
---|
| 1460 | wall_total = wall_left + wall_right + wall_south + & |
---|
| 1461 | wall_north + wall_top |
---|
| 1462 | |
---|
| 1463 | IF ( wall_total > 0.0_wp ) THEN |
---|
| 1464 | p_mg(k,j,i) = 1.0_wp / wall_total * & |
---|
| 1465 | ( wall_left * p_mg(k,j,i-1) + & |
---|
| 1466 | wall_right * p_mg(k,j,i+1) + & |
---|
| 1467 | wall_south * p_mg(k,j-1,i) + & |
---|
| 1468 | wall_north * p_mg(k,j+1,i) + & |
---|
| 1469 | wall_top * p_mg(kp1,j,i) ) |
---|
| 1470 | ENDIF |
---|
| 1471 | ENDDO |
---|
| 1472 | |
---|
| 1473 | !DIR$ IVDEP |
---|
| 1474 | DO k = nzb+1, ind_even_odd |
---|
| 1475 | km1 = k+ind_even_odd |
---|
| 1476 | kp1 = k+ind_even_odd+1 |
---|
| 1477 | ! |
---|
| 1478 | !-- First, set pressure inside topography to zero |
---|
| 1479 | p_mg(k,j,i) = p_mg(k,j,i) * & |
---|
| 1480 | ( 1.0_wp - IBITS( flags(k,j,i), 6, 1 ) ) |
---|
| 1481 | ! |
---|
| 1482 | !-- Second, determine if the gridpoint inside topography is |
---|
| 1483 | !-- adjacent to a wall and set its value to a value given by the |
---|
| 1484 | !-- average of those values obtained from Neumann boundary |
---|
| 1485 | !-- condition |
---|
| 1486 | wall_left = IBITS( flags(k,j,i-1), 5, 1 ) |
---|
| 1487 | wall_right = IBITS( flags(k,j,i+1), 4, 1 ) |
---|
| 1488 | wall_south = IBITS( flags(k,j-1,i), 3, 1 ) |
---|
| 1489 | wall_north = IBITS( flags(k,j+1,i), 2, 1 ) |
---|
| 1490 | wall_top = IBITS( flags(kp1,j,i), 0, 1 ) |
---|
| 1491 | wall_total = wall_left + wall_right + wall_south + & |
---|
| 1492 | wall_north + wall_top |
---|
| 1493 | |
---|
| 1494 | IF ( wall_total > 0.0_wp ) THEN |
---|
| 1495 | p_mg(k,j,i) = 1.0_wp / wall_total * & |
---|
| 1496 | ( wall_left * p_mg(k,j,i-1) + & |
---|
| 1497 | wall_right * p_mg(k,j,i+1) + & |
---|
| 1498 | wall_south * p_mg(k,j-1,i) + & |
---|
| 1499 | wall_north * p_mg(k,j+1,i) + & |
---|
| 1500 | wall_top * p_mg(kp1,j,i) ) |
---|
| 1501 | ENDIF |
---|
| 1502 | ENDDO |
---|
| 1503 | ENDDO |
---|
| 1504 | ENDDO |
---|
| 1505 | !$OMP END PARALLEL |
---|
| 1506 | |
---|
| 1507 | ! |
---|
| 1508 | !-- One more time horizontal boundary conditions |
---|
| 1509 | CALL exchange_horiz( p_mg, 1) |
---|
| 1510 | |
---|
| 1511 | ENDIF |
---|
| 1512 | |
---|
| 1513 | END SUBROUTINE redblack_fast |
---|
| 1514 | |
---|
| 1515 | |
---|
| 1516 | |
---|
| 1517 | SUBROUTINE sort_k_to_even_odd_blocks( p_mg , glevel ) |
---|
| 1518 | |
---|
| 1519 | !------------------------------------------------------------------------------! |
---|
| 1520 | ! Description: |
---|
| 1521 | ! ------------ |
---|
| 1522 | ! Sort k-Dimension from sequential into blocks of even and odd. |
---|
| 1523 | ! This is required to vectorize the red-black subroutine. |
---|
| 1524 | ! Version for 3D-REAL arrays |
---|
| 1525 | !------------------------------------------------------------------------------! |
---|
| 1526 | |
---|
| 1527 | USE control_parameters, & |
---|
| 1528 | ONLY: grid_level |
---|
| 1529 | |
---|
| 1530 | USE indices, & |
---|
| 1531 | ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
---|
| 1532 | |
---|
| 1533 | IMPLICIT NONE |
---|
| 1534 | |
---|
| 1535 | INTEGER(iwp), INTENT(IN) :: glevel |
---|
| 1536 | |
---|
| 1537 | REAL(wp), DIMENSION(nzb:nzt_mg(glevel)+1, & |
---|
| 1538 | nys_mg(glevel)-1:nyn_mg(glevel)+1, & |
---|
| 1539 | nxl_mg(glevel)-1:nxr_mg(glevel)+1) :: p_mg !: |
---|
| 1540 | ! |
---|
| 1541 | !-- Local variables |
---|
| 1542 | INTEGER(iwp) :: i !: |
---|
| 1543 | INTEGER(iwp) :: j !: |
---|
| 1544 | INTEGER(iwp) :: k !: |
---|
| 1545 | INTEGER(iwp) :: l !: |
---|
| 1546 | INTEGER(iwp) :: ind !: |
---|
| 1547 | REAL(wp), DIMENSION(nzb:nzt_mg(glevel)+1) :: tmp !: |
---|
| 1548 | |
---|
| 1549 | |
---|
| 1550 | CALL cpu_log( log_point_s(52), 'sort_k_to_even_odd', 'start' ) |
---|
| 1551 | |
---|
| 1552 | l = glevel |
---|
| 1553 | ind_even_odd = even_odd_level(l) |
---|
| 1554 | |
---|
| 1555 | !$OMP PARALLEL PRIVATE (i,j,k,ind,tmp) |
---|
| 1556 | !$OMP DO |
---|
| 1557 | DO i = nxl_mg(l)-1, nxr_mg(l)+1 |
---|
| 1558 | DO j = nys_mg(l)-1, nyn_mg(l)+1 |
---|
| 1559 | |
---|
| 1560 | ! |
---|
| 1561 | !-- Sort the data with even k index |
---|
| 1562 | ind = nzb-1 |
---|
| 1563 | DO k = nzb, nzt_mg(l), 2 |
---|
| 1564 | ind = ind + 1 |
---|
| 1565 | tmp(ind) = p_mg(k,j,i) |
---|
| 1566 | ENDDO |
---|
| 1567 | ! |
---|
| 1568 | !-- Sort the data with odd k index |
---|
| 1569 | DO k = nzb+1, nzt_mg(l)+1, 2 |
---|
| 1570 | ind = ind + 1 |
---|
| 1571 | tmp(ind) = p_mg(k,j,i) |
---|
| 1572 | ENDDO |
---|
| 1573 | |
---|
| 1574 | p_mg(:,j,i) = tmp |
---|
| 1575 | |
---|
| 1576 | ENDDO |
---|
| 1577 | ENDDO |
---|
| 1578 | !$OMP END PARALLEL |
---|
| 1579 | |
---|
| 1580 | CALL cpu_log( log_point_s(52), 'sort_k_to_even_odd', 'stop' ) |
---|
| 1581 | |
---|
| 1582 | END SUBROUTINE sort_k_to_even_odd_blocks |
---|
| 1583 | |
---|
| 1584 | |
---|
| 1585 | |
---|
| 1586 | SUBROUTINE sort_k_to_even_odd_blocks_1d( f_mg, f_mg_b, glevel ) |
---|
| 1587 | |
---|
| 1588 | !------------------------------------------------------------------------------! |
---|
| 1589 | ! Description: |
---|
| 1590 | ! ------------ |
---|
| 1591 | ! Sort k-Dimension from sequential into blocks of even and odd. |
---|
| 1592 | ! This is required to vectorize the red-black subroutine. |
---|
| 1593 | ! Version for 1D-REAL arrays |
---|
| 1594 | !------------------------------------------------------------------------------! |
---|
| 1595 | |
---|
| 1596 | USE indices, & |
---|
| 1597 | ONLY: nzb, nzt_mg |
---|
| 1598 | |
---|
| 1599 | IMPLICIT NONE |
---|
| 1600 | |
---|
| 1601 | INTEGER(iwp), INTENT(IN) :: glevel |
---|
| 1602 | |
---|
| 1603 | REAL(wp), DIMENSION(nzb+1:nzt_mg(glevel)) :: f_mg |
---|
| 1604 | REAL(wp), DIMENSION(nzb:nzt_mg(glevel)+1) :: f_mg_b |
---|
| 1605 | |
---|
| 1606 | ! |
---|
| 1607 | !-- Local variables |
---|
| 1608 | INTEGER(iwp) :: ind !: |
---|
| 1609 | INTEGER(iwp) :: k !: |
---|
| 1610 | |
---|
| 1611 | |
---|
| 1612 | ind = nzb - 1 |
---|
| 1613 | ! |
---|
| 1614 | !-- Sort the data with even k index |
---|
| 1615 | DO k = nzb, nzt_mg(glevel), 2 |
---|
| 1616 | ind = ind + 1 |
---|
| 1617 | IF ( k >= nzb+1 .AND. k <= nzt_mg(glevel) ) THEN |
---|
| 1618 | f_mg_b(ind) = f_mg(k) |
---|
| 1619 | ENDIF |
---|
| 1620 | ENDDO |
---|
| 1621 | ! |
---|
| 1622 | !-- Sort the data with odd k index |
---|
| 1623 | DO k = nzb+1, nzt_mg(glevel)+1, 2 |
---|
| 1624 | ind = ind + 1 |
---|
| 1625 | IF( k >= nzb+1 .AND. k <= nzt_mg(glevel) ) THEN |
---|
| 1626 | f_mg_b(ind) = f_mg(k) |
---|
| 1627 | ENDIF |
---|
| 1628 | ENDDO |
---|
| 1629 | |
---|
| 1630 | END SUBROUTINE sort_k_to_even_odd_blocks_1d |
---|
| 1631 | |
---|
| 1632 | |
---|
| 1633 | |
---|
| 1634 | SUBROUTINE sort_k_to_even_odd_blocks_int( i_mg , glevel ) |
---|
| 1635 | |
---|
| 1636 | !------------------------------------------------------------------------------! |
---|
| 1637 | ! Description: |
---|
| 1638 | ! ------------ |
---|
| 1639 | ! Sort k-Dimension from sequential into blocks of even and odd. |
---|
| 1640 | ! This is required to vectorize the red-black subroutine. |
---|
| 1641 | ! Version for 2D-INTEGER arrays |
---|
| 1642 | !------------------------------------------------------------------------------! |
---|
| 1643 | |
---|
| 1644 | USE control_parameters, & |
---|
| 1645 | ONLY: grid_level |
---|
| 1646 | |
---|
| 1647 | USE indices, & |
---|
| 1648 | ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
---|
| 1649 | |
---|
| 1650 | IMPLICIT NONE |
---|
| 1651 | |
---|
| 1652 | INTEGER(iwp), INTENT(IN) :: glevel |
---|
| 1653 | |
---|
| 1654 | INTEGER(iwp), DIMENSION(nzb:nzt_mg(glevel)+1, & |
---|
| 1655 | nys_mg(glevel)-1:nyn_mg(glevel)+1, & |
---|
| 1656 | nxl_mg(glevel)-1:nxr_mg(glevel)+1) :: i_mg !: |
---|
| 1657 | ! |
---|
| 1658 | !-- Local variables |
---|
| 1659 | INTEGER(iwp) :: i !: |
---|
| 1660 | INTEGER(iwp) :: j !: |
---|
| 1661 | INTEGER(iwp) :: k !: |
---|
| 1662 | INTEGER(iwp) :: l !: |
---|
| 1663 | INTEGER(iwp) :: ind !: |
---|
| 1664 | INTEGER(iwp),DIMENSION(nzb:nzt_mg(glevel)+1) :: tmp |
---|
| 1665 | |
---|
| 1666 | |
---|
| 1667 | CALL cpu_log( log_point_s(52), 'sort_k_to_even_odd', 'start' ) |
---|
| 1668 | |
---|
| 1669 | l = glevel |
---|
| 1670 | ind_even_odd = even_odd_level(l) |
---|
| 1671 | |
---|
| 1672 | DO i = nxl_mg(l)-1, nxr_mg(l)+1 |
---|
| 1673 | DO j = nys_mg(l)-1, nyn_mg(l)+1 |
---|
| 1674 | |
---|
| 1675 | ! |
---|
| 1676 | !-- Sort the data with even k index |
---|
| 1677 | ind = nzb-1 |
---|
| 1678 | DO k = nzb, nzt_mg(l), 2 |
---|
| 1679 | ind = ind + 1 |
---|
| 1680 | tmp(ind) = i_mg(k,j,i) |
---|
| 1681 | ENDDO |
---|
| 1682 | ! |
---|
| 1683 | !++ ATTENTION: Check reason for this error. Remove it or replace WRITE |
---|
| 1684 | !++ by PALM message |
---|
| 1685 | IF ( ind /= ind_even_odd ) THEN |
---|
| 1686 | WRITE (0,*) 'ERROR ==> illegal ind_even_odd ',ind,ind_even_odd,l |
---|
| 1687 | CALL MPI_ABORT(MPI_COMM_WORLD,i,j) |
---|
| 1688 | ENDIF |
---|
| 1689 | ! |
---|
| 1690 | !-- Sort the data with odd k index |
---|
| 1691 | DO k = nzb+1, nzt_mg(l)+1, 2 |
---|
| 1692 | ind = ind + 1 |
---|
| 1693 | tmp(ind) = i_mg(k,j,i) |
---|
| 1694 | ENDDO |
---|
| 1695 | |
---|
| 1696 | i_mg(:,j,i) = tmp |
---|
| 1697 | |
---|
| 1698 | ENDDO |
---|
| 1699 | ENDDO |
---|
| 1700 | |
---|
| 1701 | CALL cpu_log( log_point_s(52), 'sort_k_to_even_odd', 'stop' ) |
---|
| 1702 | |
---|
| 1703 | END SUBROUTINE sort_k_to_even_odd_blocks_int |
---|
| 1704 | |
---|
| 1705 | |
---|
| 1706 | |
---|
| 1707 | SUBROUTINE sort_k_to_sequential( p_mg ) |
---|
| 1708 | |
---|
| 1709 | !------------------------------------------------------------------------------! |
---|
| 1710 | ! Description: |
---|
| 1711 | ! ------------ |
---|
| 1712 | ! Sort k-dimension from blocks of even and odd into sequential |
---|
| 1713 | !------------------------------------------------------------------------------! |
---|
| 1714 | |
---|
| 1715 | USE control_parameters, & |
---|
| 1716 | ONLY: grid_level |
---|
| 1717 | |
---|
| 1718 | USE indices, & |
---|
| 1719 | ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
---|
| 1720 | |
---|
| 1721 | IMPLICIT NONE |
---|
| 1722 | |
---|
| 1723 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 1724 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 1725 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p_mg !: |
---|
| 1726 | ! |
---|
| 1727 | !-- Local variables |
---|
| 1728 | INTEGER(iwp) :: i !: |
---|
| 1729 | INTEGER(iwp) :: j !: |
---|
| 1730 | INTEGER(iwp) :: k !: |
---|
| 1731 | INTEGER(iwp) :: l !: |
---|
| 1732 | INTEGER(iwp) :: ind !: |
---|
| 1733 | |
---|
| 1734 | REAL(wp),DIMENSION(nzb:nzt_mg(grid_level)+1) :: tmp |
---|
| 1735 | |
---|
| 1736 | |
---|
| 1737 | l = grid_level |
---|
| 1738 | |
---|
| 1739 | !$OMP PARALLEL PRIVATE (i,j,k,ind,tmp) |
---|
| 1740 | !$OMP DO |
---|
| 1741 | DO i = nxl_mg(l)-1, nxr_mg(l)+1 |
---|
| 1742 | DO j = nys_mg(l)-1, nyn_mg(l)+1 |
---|
| 1743 | |
---|
| 1744 | ind = nzb - 1 |
---|
| 1745 | tmp = p_mg(:,j,i) |
---|
| 1746 | DO k = nzb, nzt_mg(l), 2 |
---|
| 1747 | ind = ind + 1 |
---|
| 1748 | p_mg(k,j,i) = tmp(ind) |
---|
| 1749 | ENDDO |
---|
| 1750 | |
---|
| 1751 | DO k = nzb+1, nzt_mg(l)+1, 2 |
---|
| 1752 | ind = ind + 1 |
---|
| 1753 | p_mg(k,j,i) = tmp(ind) |
---|
| 1754 | ENDDO |
---|
| 1755 | ENDDO |
---|
| 1756 | ENDDO |
---|
| 1757 | !$OMP END PARALLEL |
---|
| 1758 | |
---|
| 1759 | END SUBROUTINE sort_k_to_sequential |
---|
| 1760 | |
---|
| 1761 | |
---|
| 1762 | |
---|
| 1763 | SUBROUTINE mg_gather_fast( f2, f2_sub ) |
---|
| 1764 | |
---|
| 1765 | USE control_parameters, & |
---|
| 1766 | ONLY: grid_level |
---|
| 1767 | |
---|
| 1768 | USE cpulog, & |
---|
| 1769 | ONLY: cpu_log, log_point_s |
---|
| 1770 | |
---|
| 1771 | USE indices, & |
---|
| 1772 | ONLY: mg_loc_ind, nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
---|
| 1773 | |
---|
| 1774 | IMPLICIT NONE |
---|
| 1775 | |
---|
| 1776 | INTEGER(iwp) :: i !: |
---|
| 1777 | INTEGER(iwp) :: il !: |
---|
| 1778 | INTEGER(iwp) :: ir !: |
---|
| 1779 | INTEGER(iwp) :: j !: |
---|
| 1780 | INTEGER(iwp) :: jn !: |
---|
| 1781 | INTEGER(iwp) :: js !: |
---|
| 1782 | INTEGER(iwp) :: k !: |
---|
| 1783 | INTEGER(iwp) :: nwords !: |
---|
| 1784 | |
---|
| 1785 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 1786 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 1787 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f2 !: |
---|
| 1788 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 1789 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 1790 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f2_l !: |
---|
| 1791 | |
---|
| 1792 | REAL(wp), DIMENSION(nzb:mg_loc_ind(5,myid)+1, & |
---|
| 1793 | mg_loc_ind(3,myid)-1:mg_loc_ind(4,myid)+1, & |
---|
| 1794 | mg_loc_ind(1,myid)-1:mg_loc_ind(2,myid)+1) :: f2_sub !: |
---|
| 1795 | |
---|
| 1796 | |
---|
| 1797 | #if defined( __parallel ) |
---|
| 1798 | CALL cpu_log( log_point_s(34), 'mg_gather', 'start' ) |
---|
| 1799 | |
---|
| 1800 | f2_l = 0.0_wp |
---|
| 1801 | |
---|
| 1802 | ! |
---|
| 1803 | !-- Store the local subdomain array on the total array |
---|
| 1804 | js = mg_loc_ind(3,myid) |
---|
| 1805 | IF ( south_border_pe ) js = js - 1 |
---|
| 1806 | jn = mg_loc_ind(4,myid) |
---|
| 1807 | IF ( north_border_pe ) jn = jn + 1 |
---|
| 1808 | il = mg_loc_ind(1,myid) |
---|
| 1809 | IF ( left_border_pe ) il = il - 1 |
---|
| 1810 | ir = mg_loc_ind(2,myid) |
---|
| 1811 | IF ( right_border_pe ) ir = ir + 1 |
---|
| 1812 | DO i = il, ir |
---|
| 1813 | DO j = js, jn |
---|
| 1814 | DO k = nzb, nzt_mg(grid_level)+1 |
---|
| 1815 | f2_l(k,j,i) = f2_sub(k,j,i) |
---|
| 1816 | ENDDO |
---|
| 1817 | ENDDO |
---|
| 1818 | ENDDO |
---|
| 1819 | |
---|
| 1820 | ! |
---|
| 1821 | !-- Find out the number of array elements of the total array |
---|
| 1822 | nwords = SIZE( f2 ) |
---|
| 1823 | |
---|
| 1824 | ! |
---|
| 1825 | !-- Gather subdomain data from all PEs |
---|
| 1826 | IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr ) |
---|
| 1827 | CALL MPI_ALLREDUCE( f2_l(nzb,nys_mg(grid_level)-1,nxl_mg(grid_level)-1), & |
---|
| 1828 | f2(nzb,nys_mg(grid_level)-1,nxl_mg(grid_level)-1), & |
---|
| 1829 | nwords, MPI_REAL, MPI_SUM, comm2d, ierr ) |
---|
| 1830 | |
---|
| 1831 | CALL cpu_log( log_point_s(34), 'mg_gather', 'stop' ) |
---|
| 1832 | #endif |
---|
| 1833 | |
---|
| 1834 | END SUBROUTINE mg_gather_fast |
---|
| 1835 | |
---|
| 1836 | |
---|
| 1837 | |
---|
| 1838 | SUBROUTINE mg_scatter_fast( p2, p2_sub ) |
---|
| 1839 | ! |
---|
| 1840 | !-- TODO: It might be possible to improve the speed of this routine by using |
---|
| 1841 | !-- non-blocking communication |
---|
| 1842 | |
---|
| 1843 | USE control_parameters, & |
---|
| 1844 | ONLY: grid_level |
---|
| 1845 | |
---|
| 1846 | USE cpulog, & |
---|
| 1847 | ONLY: cpu_log, log_point_s |
---|
| 1848 | |
---|
| 1849 | USE indices, & |
---|
| 1850 | ONLY: mg_loc_ind, nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
---|
| 1851 | |
---|
| 1852 | IMPLICIT NONE |
---|
| 1853 | |
---|
| 1854 | INTEGER(iwp) :: nwords !: |
---|
| 1855 | |
---|
| 1856 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, & |
---|
| 1857 | nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, & |
---|
| 1858 | nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1) :: p2 !: |
---|
| 1859 | |
---|
| 1860 | REAL(wp), DIMENSION(nzb:mg_loc_ind(5,myid)+1, & |
---|
| 1861 | mg_loc_ind(3,myid)-1:mg_loc_ind(4,myid)+1, & |
---|
| 1862 | mg_loc_ind(1,myid)-1:mg_loc_ind(2,myid)+1) :: p2_sub !: |
---|
| 1863 | |
---|
| 1864 | ! |
---|
| 1865 | !-- Find out the number of array elements of the subdomain array |
---|
| 1866 | nwords = SIZE( p2_sub ) |
---|
| 1867 | |
---|
| 1868 | #if defined( __parallel ) |
---|
| 1869 | CALL cpu_log( log_point_s(35), 'mg_scatter', 'start' ) |
---|
| 1870 | |
---|
| 1871 | p2_sub = p2(:,mg_loc_ind(3,myid)-1:mg_loc_ind(4,myid)+1, & |
---|
| 1872 | mg_loc_ind(1,myid)-1:mg_loc_ind(2,myid)+1) |
---|
| 1873 | |
---|
| 1874 | CALL cpu_log( log_point_s(35), 'mg_scatter', 'stop' ) |
---|
| 1875 | #endif |
---|
| 1876 | |
---|
| 1877 | END SUBROUTINE mg_scatter_fast |
---|
| 1878 | |
---|
| 1879 | |
---|
| 1880 | |
---|
| 1881 | RECURSIVE SUBROUTINE next_mg_level_fast( f_mg, p_mg, p3, r ) |
---|
| 1882 | |
---|
| 1883 | !------------------------------------------------------------------------------! |
---|
| 1884 | ! Description: |
---|
| 1885 | ! ------------ |
---|
| 1886 | ! This is where the multigrid technique takes place. V- and W- Cycle are |
---|
| 1887 | ! implemented and steered by the parameter "gamma". Parameter "nue" determines |
---|
| 1888 | ! the convergence of the multigrid iterative solution. There are nue times |
---|
| 1889 | ! RB-GS iterations. It should be set to "1" or "2", considering the time effort |
---|
| 1890 | ! one would like to invest. Last choice shows a very good converging factor, |
---|
| 1891 | ! but leads to an increase in computing time. |
---|
| 1892 | !------------------------------------------------------------------------------! |
---|
| 1893 | |
---|
| 1894 | USE control_parameters, & |
---|
| 1895 | ONLY: bc_lr_dirrad, bc_lr_raddir, bc_ns_dirrad, bc_ns_raddir, & |
---|
| 1896 | gamma_mg, grid_level, grid_level_count, ibc_p_b, ibc_p_t, & |
---|
| 1897 | inflow_l, inflow_n, inflow_r, inflow_s, maximum_grid_level, & |
---|
| 1898 | mg_switch_to_pe0_level, mg_switch_to_pe0, ngsrb, outflow_l, & |
---|
| 1899 | outflow_n, outflow_r, outflow_s |
---|
| 1900 | |
---|
| 1901 | USE indices, & |
---|
| 1902 | ONLY: mg_loc_ind, nxl, nxl_mg, nxr, nxr_mg, nys, nys_mg, nyn, & |
---|
| 1903 | nyn_mg, nzb, nzt, nzt_mg |
---|
| 1904 | |
---|
| 1905 | IMPLICIT NONE |
---|
| 1906 | |
---|
| 1907 | INTEGER(iwp) :: i !: |
---|
| 1908 | INTEGER(iwp) :: j !: |
---|
| 1909 | INTEGER(iwp) :: k !: |
---|
| 1910 | INTEGER(iwp) :: nxl_mg_save !: |
---|
| 1911 | INTEGER(iwp) :: nxr_mg_save !: |
---|
| 1912 | INTEGER(iwp) :: nyn_mg_save !: |
---|
| 1913 | INTEGER(iwp) :: nys_mg_save !: |
---|
| 1914 | INTEGER(iwp) :: nzt_mg_save !: |
---|
| 1915 | |
---|
| 1916 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 1917 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 1918 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f_mg !: |
---|
| 1919 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 1920 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 1921 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p_mg !: |
---|
| 1922 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 1923 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 1924 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p3 !: |
---|
| 1925 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 1926 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 1927 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: r !: |
---|
| 1928 | |
---|
| 1929 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, & |
---|
| 1930 | nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, & |
---|
| 1931 | nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1) :: f2 !: |
---|
| 1932 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, & |
---|
| 1933 | nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, & |
---|
| 1934 | nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1) :: p2 !: |
---|
| 1935 | |
---|
| 1936 | REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: f2_sub !: |
---|
| 1937 | REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: p2_sub !: |
---|
| 1938 | |
---|
| 1939 | ! |
---|
| 1940 | !-- Restriction to the coarsest grid |
---|
| 1941 | 10 IF ( grid_level == 1 ) THEN |
---|
| 1942 | |
---|
| 1943 | ! |
---|
| 1944 | !-- Solution on the coarsest grid. Double the number of Gauss-Seidel |
---|
| 1945 | !-- iterations in order to get a more accurate solution. |
---|
| 1946 | ngsrb = 2 * ngsrb |
---|
| 1947 | |
---|
| 1948 | ind_even_odd = even_odd_level(grid_level) |
---|
| 1949 | |
---|
| 1950 | CALL redblack_fast( f_mg, p_mg ) |
---|
| 1951 | |
---|
| 1952 | ngsrb = ngsrb / 2 |
---|
| 1953 | |
---|
| 1954 | |
---|
| 1955 | ELSEIF ( grid_level /= 1 ) THEN |
---|
| 1956 | |
---|
| 1957 | grid_level_count(grid_level) = grid_level_count(grid_level) + 1 |
---|
| 1958 | |
---|
| 1959 | ! |
---|
| 1960 | !-- Solution on the actual grid level |
---|
| 1961 | ind_even_odd = even_odd_level(grid_level) |
---|
| 1962 | |
---|
| 1963 | CALL redblack_fast( f_mg, p_mg ) |
---|
| 1964 | |
---|
| 1965 | ! |
---|
| 1966 | !-- Determination of the actual residual |
---|
| 1967 | CALL resid_fast( f_mg, p_mg, r ) |
---|
| 1968 | |
---|
| 1969 | !-- Restriction of the residual (finer grid values!) to the next coarser |
---|
| 1970 | !-- grid. Therefore, the grid level has to be decremented now. nxl..nzt have |
---|
| 1971 | !-- to be set to the coarse grid values, because these variables are needed |
---|
| 1972 | !-- for the exchange of ghost points in routine exchange_horiz |
---|
| 1973 | grid_level = grid_level - 1 |
---|
| 1974 | |
---|
| 1975 | nxl = nxl_mg(grid_level) |
---|
| 1976 | nys = nys_mg(grid_level) |
---|
| 1977 | nxr = nxr_mg(grid_level) |
---|
| 1978 | nyn = nyn_mg(grid_level) |
---|
| 1979 | nzt = nzt_mg(grid_level) |
---|
| 1980 | |
---|
| 1981 | IF ( grid_level == mg_switch_to_pe0_level ) THEN |
---|
| 1982 | |
---|
| 1983 | ! |
---|
| 1984 | !-- From this level on, calculations are done on PE0 only. |
---|
| 1985 | !-- First, carry out restriction on the subdomain. |
---|
| 1986 | !-- Therefore, indices of the level have to be changed to subdomain |
---|
| 1987 | !-- values in between (otherwise, the restrict routine would expect |
---|
| 1988 | !-- the gathered array) |
---|
| 1989 | |
---|
| 1990 | nxl_mg_save = nxl_mg(grid_level) |
---|
| 1991 | nxr_mg_save = nxr_mg(grid_level) |
---|
| 1992 | nys_mg_save = nys_mg(grid_level) |
---|
| 1993 | nyn_mg_save = nyn_mg(grid_level) |
---|
| 1994 | nzt_mg_save = nzt_mg(grid_level) |
---|
| 1995 | nxl_mg(grid_level) = mg_loc_ind(1,myid) |
---|
| 1996 | nxr_mg(grid_level) = mg_loc_ind(2,myid) |
---|
| 1997 | nys_mg(grid_level) = mg_loc_ind(3,myid) |
---|
| 1998 | nyn_mg(grid_level) = mg_loc_ind(4,myid) |
---|
| 1999 | nzt_mg(grid_level) = mg_loc_ind(5,myid) |
---|
| 2000 | nxl = mg_loc_ind(1,myid) |
---|
| 2001 | nxr = mg_loc_ind(2,myid) |
---|
| 2002 | nys = mg_loc_ind(3,myid) |
---|
| 2003 | nyn = mg_loc_ind(4,myid) |
---|
| 2004 | nzt = mg_loc_ind(5,myid) |
---|
| 2005 | |
---|
| 2006 | ALLOCATE( f2_sub(nzb:nzt_mg(grid_level)+1, & |
---|
| 2007 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 2008 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) ) |
---|
| 2009 | |
---|
| 2010 | CALL restrict_fast( f2_sub, r ) |
---|
| 2011 | |
---|
| 2012 | ! |
---|
| 2013 | !-- Restore the correct indices of this level |
---|
| 2014 | nxl_mg(grid_level) = nxl_mg_save |
---|
| 2015 | nxr_mg(grid_level) = nxr_mg_save |
---|
| 2016 | nys_mg(grid_level) = nys_mg_save |
---|
| 2017 | nyn_mg(grid_level) = nyn_mg_save |
---|
| 2018 | nzt_mg(grid_level) = nzt_mg_save |
---|
| 2019 | nxl = nxl_mg(grid_level) |
---|
| 2020 | nxr = nxr_mg(grid_level) |
---|
| 2021 | nys = nys_mg(grid_level) |
---|
| 2022 | nyn = nyn_mg(grid_level) |
---|
| 2023 | nzt = nzt_mg(grid_level) |
---|
| 2024 | ! |
---|
| 2025 | !-- Gather all arrays from the subdomains on PE0 |
---|
| 2026 | CALL mg_gather_fast( f2, f2_sub ) |
---|
| 2027 | |
---|
| 2028 | ! |
---|
| 2029 | !-- Set switch for routine exchange_horiz, that no ghostpoint exchange |
---|
| 2030 | !-- has to be carried out from now on |
---|
| 2031 | mg_switch_to_pe0 = .TRUE. |
---|
| 2032 | |
---|
| 2033 | ! |
---|
| 2034 | !-- In case of non-cyclic lateral boundary conditions, both in- and |
---|
| 2035 | !-- outflow conditions have to be used on all PEs after the switch, |
---|
| 2036 | !-- because then they have the total domain. |
---|
| 2037 | IF ( bc_lr_dirrad ) THEN |
---|
| 2038 | inflow_l = .TRUE. |
---|
| 2039 | inflow_r = .FALSE. |
---|
| 2040 | outflow_l = .FALSE. |
---|
| 2041 | outflow_r = .TRUE. |
---|
| 2042 | ELSEIF ( bc_lr_raddir ) THEN |
---|
| 2043 | inflow_l = .FALSE. |
---|
| 2044 | inflow_r = .TRUE. |
---|
| 2045 | outflow_l = .TRUE. |
---|
| 2046 | outflow_r = .FALSE. |
---|
| 2047 | ENDIF |
---|
| 2048 | |
---|
| 2049 | IF ( bc_ns_dirrad ) THEN |
---|
| 2050 | inflow_n = .TRUE. |
---|
| 2051 | inflow_s = .FALSE. |
---|
| 2052 | outflow_n = .FALSE. |
---|
| 2053 | outflow_s = .TRUE. |
---|
| 2054 | ELSEIF ( bc_ns_raddir ) THEN |
---|
| 2055 | inflow_n = .FALSE. |
---|
| 2056 | inflow_s = .TRUE. |
---|
| 2057 | outflow_n = .TRUE. |
---|
| 2058 | outflow_s = .FALSE. |
---|
| 2059 | ENDIF |
---|
| 2060 | |
---|
| 2061 | DEALLOCATE( f2_sub ) |
---|
| 2062 | |
---|
| 2063 | ELSE |
---|
| 2064 | |
---|
| 2065 | CALL restrict_fast( f2, r ) |
---|
| 2066 | |
---|
| 2067 | ind_even_odd = even_odd_level(grid_level) ! must be after restrict |
---|
| 2068 | |
---|
| 2069 | ENDIF |
---|
| 2070 | |
---|
| 2071 | p2 = 0.0_wp |
---|
| 2072 | |
---|
| 2073 | ! |
---|
| 2074 | !-- Repeat the same procedure till the coarsest grid is reached |
---|
| 2075 | CALL next_mg_level_fast( f2, p2, p3, r ) |
---|
| 2076 | |
---|
| 2077 | ENDIF |
---|
| 2078 | |
---|
| 2079 | ! |
---|
| 2080 | !-- Now follows the prolongation |
---|
| 2081 | IF ( grid_level >= 2 ) THEN |
---|
| 2082 | |
---|
| 2083 | ! |
---|
| 2084 | !-- Prolongation of the new residual. The values are transferred |
---|
| 2085 | !-- from the coarse to the next finer grid. |
---|
| 2086 | IF ( grid_level == mg_switch_to_pe0_level+1 ) THEN |
---|
| 2087 | |
---|
| 2088 | #if defined( __parallel ) |
---|
| 2089 | ! |
---|
| 2090 | !-- At this level, the new residual first has to be scattered from |
---|
| 2091 | !-- PE0 to the other PEs |
---|
| 2092 | ALLOCATE( p2_sub(nzb:mg_loc_ind(5,myid)+1, & |
---|
| 2093 | mg_loc_ind(3,myid)-1:mg_loc_ind(4,myid)+1, & |
---|
| 2094 | mg_loc_ind(1,myid)-1:mg_loc_ind(2,myid)+1) ) |
---|
| 2095 | |
---|
| 2096 | CALL mg_scatter_fast( p2, p2_sub ) |
---|
| 2097 | |
---|
| 2098 | ! |
---|
| 2099 | !-- Therefore, indices of the previous level have to be changed to |
---|
| 2100 | !-- subdomain values in between (otherwise, the prolong routine would |
---|
| 2101 | !-- expect the gathered array) |
---|
| 2102 | nxl_mg_save = nxl_mg(grid_level-1) |
---|
| 2103 | nxr_mg_save = nxr_mg(grid_level-1) |
---|
| 2104 | nys_mg_save = nys_mg(grid_level-1) |
---|
| 2105 | nyn_mg_save = nyn_mg(grid_level-1) |
---|
| 2106 | nzt_mg_save = nzt_mg(grid_level-1) |
---|
| 2107 | nxl_mg(grid_level-1) = mg_loc_ind(1,myid) |
---|
| 2108 | nxr_mg(grid_level-1) = mg_loc_ind(2,myid) |
---|
| 2109 | nys_mg(grid_level-1) = mg_loc_ind(3,myid) |
---|
| 2110 | nyn_mg(grid_level-1) = mg_loc_ind(4,myid) |
---|
| 2111 | nzt_mg(grid_level-1) = mg_loc_ind(5,myid) |
---|
| 2112 | |
---|
| 2113 | ! |
---|
| 2114 | !-- Set switch for routine exchange_horiz, that ghostpoint exchange |
---|
| 2115 | !-- has to be carried again out from now on |
---|
| 2116 | mg_switch_to_pe0 = .FALSE. |
---|
| 2117 | |
---|
| 2118 | ! |
---|
| 2119 | !-- For non-cyclic lateral boundary conditions, restore the |
---|
| 2120 | !-- in-/outflow conditions |
---|
| 2121 | inflow_l = .FALSE.; inflow_r = .FALSE. |
---|
| 2122 | inflow_n = .FALSE.; inflow_s = .FALSE. |
---|
| 2123 | outflow_l = .FALSE.; outflow_r = .FALSE. |
---|
| 2124 | outflow_n = .FALSE.; outflow_s = .FALSE. |
---|
| 2125 | |
---|
| 2126 | IF ( pleft == MPI_PROC_NULL ) THEN |
---|
| 2127 | IF ( bc_lr_dirrad ) THEN |
---|
| 2128 | inflow_l = .TRUE. |
---|
| 2129 | ELSEIF ( bc_lr_raddir ) THEN |
---|
| 2130 | outflow_l = .TRUE. |
---|
| 2131 | ENDIF |
---|
| 2132 | ENDIF |
---|
| 2133 | |
---|
| 2134 | IF ( pright == MPI_PROC_NULL ) THEN |
---|
| 2135 | IF ( bc_lr_dirrad ) THEN |
---|
| 2136 | outflow_r = .TRUE. |
---|
| 2137 | ELSEIF ( bc_lr_raddir ) THEN |
---|
| 2138 | inflow_r = .TRUE. |
---|
| 2139 | ENDIF |
---|
| 2140 | ENDIF |
---|
| 2141 | |
---|
| 2142 | IF ( psouth == MPI_PROC_NULL ) THEN |
---|
| 2143 | IF ( bc_ns_dirrad ) THEN |
---|
| 2144 | outflow_s = .TRUE. |
---|
| 2145 | ELSEIF ( bc_ns_raddir ) THEN |
---|
| 2146 | inflow_s = .TRUE. |
---|
| 2147 | ENDIF |
---|
| 2148 | ENDIF |
---|
| 2149 | |
---|
| 2150 | IF ( pnorth == MPI_PROC_NULL ) THEN |
---|
| 2151 | IF ( bc_ns_dirrad ) THEN |
---|
| 2152 | inflow_n = .TRUE. |
---|
| 2153 | ELSEIF ( bc_ns_raddir ) THEN |
---|
| 2154 | outflow_n = .TRUE. |
---|
| 2155 | ENDIF |
---|
| 2156 | ENDIF |
---|
| 2157 | |
---|
| 2158 | CALL prolong_fast( p2_sub, p3 ) |
---|
| 2159 | |
---|
| 2160 | ! |
---|
| 2161 | !-- Restore the correct indices of the previous level |
---|
| 2162 | nxl_mg(grid_level-1) = nxl_mg_save |
---|
| 2163 | nxr_mg(grid_level-1) = nxr_mg_save |
---|
| 2164 | nys_mg(grid_level-1) = nys_mg_save |
---|
| 2165 | nyn_mg(grid_level-1) = nyn_mg_save |
---|
| 2166 | nzt_mg(grid_level-1) = nzt_mg_save |
---|
| 2167 | |
---|
| 2168 | DEALLOCATE( p2_sub ) |
---|
| 2169 | #endif |
---|
| 2170 | |
---|
| 2171 | ELSE |
---|
| 2172 | |
---|
| 2173 | CALL prolong_fast( p2, p3 ) |
---|
| 2174 | |
---|
| 2175 | ENDIF |
---|
| 2176 | |
---|
| 2177 | ! |
---|
| 2178 | !-- Computation of the new pressure correction. Therefore, |
---|
| 2179 | !-- values from prior grids are added up automatically stage by stage. |
---|
| 2180 | DO i = nxl_mg(grid_level)-1, nxr_mg(grid_level)+1 |
---|
| 2181 | DO j = nys_mg(grid_level)-1, nyn_mg(grid_level)+1 |
---|
| 2182 | DO k = nzb, nzt_mg(grid_level)+1 |
---|
| 2183 | p_mg(k,j,i) = p_mg(k,j,i) + p3(k,j,i) |
---|
| 2184 | ENDDO |
---|
| 2185 | ENDDO |
---|
| 2186 | ENDDO |
---|
| 2187 | |
---|
| 2188 | ! |
---|
| 2189 | !-- Relaxation of the new solution |
---|
| 2190 | CALL redblack_fast( f_mg, p_mg ) |
---|
| 2191 | |
---|
| 2192 | ENDIF |
---|
| 2193 | |
---|
| 2194 | |
---|
| 2195 | ! |
---|
| 2196 | !-- The following few lines serve the steering of the multigrid scheme |
---|
| 2197 | IF ( grid_level == maximum_grid_level ) THEN |
---|
| 2198 | |
---|
| 2199 | GOTO 20 |
---|
| 2200 | |
---|
| 2201 | ELSEIF ( grid_level /= maximum_grid_level .AND. grid_level /= 1 .AND. & |
---|
| 2202 | grid_level_count(grid_level) /= gamma_mg ) THEN |
---|
| 2203 | |
---|
| 2204 | GOTO 10 |
---|
| 2205 | |
---|
| 2206 | ENDIF |
---|
| 2207 | |
---|
| 2208 | ! |
---|
| 2209 | !-- Reset counter for the next call of poismg_fast |
---|
| 2210 | grid_level_count(grid_level) = 0 |
---|
| 2211 | |
---|
| 2212 | ! |
---|
| 2213 | !-- Continue with the next finer level. nxl..nzt have to be |
---|
| 2214 | !-- set to the finer grid values, because these variables are needed for the |
---|
| 2215 | !-- exchange of ghost points in routine exchange_horiz |
---|
| 2216 | grid_level = grid_level + 1 |
---|
| 2217 | ind_even_odd = even_odd_level(grid_level) |
---|
| 2218 | |
---|
| 2219 | nxl = nxl_mg(grid_level) |
---|
| 2220 | nxr = nxr_mg(grid_level) |
---|
| 2221 | nys = nys_mg(grid_level) |
---|
| 2222 | nyn = nyn_mg(grid_level) |
---|
| 2223 | nzt = nzt_mg(grid_level) |
---|
| 2224 | |
---|
| 2225 | 20 CONTINUE |
---|
| 2226 | |
---|
| 2227 | END SUBROUTINE next_mg_level_fast |
---|
| 2228 | |
---|
| 2229 | |
---|
| 2230 | |
---|
| 2231 | SUBROUTINE init_even_odd_blocks |
---|
| 2232 | |
---|
| 2233 | !------------------------------------------------------------------------------! |
---|
| 2234 | ! Description: |
---|
| 2235 | ! ------------ |
---|
| 2236 | ! Initial settings for sorting k-dimension from sequential order (alternate |
---|
| 2237 | ! even/odd) into blocks of even and odd or vice versa |
---|
| 2238 | !------------------------------------------------------------------------------! |
---|
| 2239 | |
---|
| 2240 | USE arrays_3d, & |
---|
| 2241 | ONLY: f1_mg, f2_mg, f3_mg |
---|
| 2242 | |
---|
| 2243 | USE control_parameters, & |
---|
| 2244 | ONLY: grid_level, masking_method, maximum_grid_level, topography |
---|
| 2245 | |
---|
| 2246 | USE indices, & |
---|
| 2247 | ONLY: nzb, nzt, nzt_mg |
---|
| 2248 | |
---|
| 2249 | USE indices, & |
---|
| 2250 | ONLY: flags, wall_flags_1, wall_flags_2, wall_flags_3, & |
---|
| 2251 | wall_flags_4, wall_flags_5, wall_flags_6, wall_flags_7, & |
---|
| 2252 | wall_flags_8, wall_flags_9, wall_flags_10, nxl_mg, nxr_mg, & |
---|
| 2253 | nys_mg, nyn_mg, nzb, nzt_mg |
---|
| 2254 | |
---|
| 2255 | IMPLICIT NONE |
---|
| 2256 | ! |
---|
| 2257 | !-- Local variables |
---|
| 2258 | INTEGER(iwp) :: i !: |
---|
| 2259 | INTEGER(iwp) :: l !: |
---|
| 2260 | |
---|
| 2261 | LOGICAL, SAVE :: lfirst = .TRUE. |
---|
| 2262 | |
---|
| 2263 | |
---|
| 2264 | IF ( .NOT. lfirst ) RETURN |
---|
| 2265 | |
---|
| 2266 | ALLOCATE( even_odd_level(maximum_grid_level) ) |
---|
| 2267 | |
---|
| 2268 | ALLOCATE( f1_mg_b(nzb:nzt+1,maximum_grid_level), & |
---|
| 2269 | f2_mg_b(nzb:nzt+1,maximum_grid_level), & |
---|
| 2270 | f3_mg_b(nzb:nzt+1,maximum_grid_level) ) |
---|
| 2271 | |
---|
| 2272 | ! |
---|
| 2273 | !-- Set border index between the even and odd block |
---|
| 2274 | DO i = maximum_grid_level, 1, -1 |
---|
| 2275 | even_odd_level(i) = nzt_mg(i) / 2 |
---|
| 2276 | ENDDO |
---|
| 2277 | |
---|
| 2278 | ! |
---|
| 2279 | !-- Sort grid coefficients used in red/black scheme and for calculating the |
---|
| 2280 | !-- residual to block (even/odd) structure |
---|
| 2281 | DO l = maximum_grid_level, 1 , -1 |
---|
| 2282 | CALL sort_k_to_even_odd_blocks( f1_mg(nzb+1:nzt_mg(grid_level),l), & |
---|
| 2283 | f1_mg_b(nzb:nzt_mg(grid_level)+1,l), & |
---|
| 2284 | l ) |
---|
| 2285 | CALL sort_k_to_even_odd_blocks( f2_mg(nzb+1:nzt_mg(grid_level),l), & |
---|
| 2286 | f2_mg_b(nzb:nzt_mg(grid_level)+1,l), & |
---|
| 2287 | l ) |
---|
| 2288 | CALL sort_k_to_even_odd_blocks( f3_mg(nzb+1:nzt_mg(grid_level),l), & |
---|
| 2289 | f3_mg_b(nzb:nzt_mg(grid_level)+1,l), & |
---|
| 2290 | l ) |
---|
| 2291 | ENDDO |
---|
| 2292 | |
---|
| 2293 | ! |
---|
| 2294 | !-- Sort flags array for all levels to block (even/odd) structure |
---|
| 2295 | IF ( topography /= 'flat' .AND. .NOT. masking_method ) THEN |
---|
| 2296 | |
---|
| 2297 | DO l = maximum_grid_level, 1 , -1 |
---|
| 2298 | ! |
---|
| 2299 | !-- Assign the flag level to be calculated |
---|
| 2300 | SELECT CASE ( l ) |
---|
| 2301 | CASE ( 1 ) |
---|
| 2302 | flags => wall_flags_1 |
---|
| 2303 | CASE ( 2 ) |
---|
| 2304 | flags => wall_flags_2 |
---|
| 2305 | CASE ( 3 ) |
---|
| 2306 | flags => wall_flags_3 |
---|
| 2307 | CASE ( 4 ) |
---|
| 2308 | flags => wall_flags_4 |
---|
| 2309 | CASE ( 5 ) |
---|
| 2310 | flags => wall_flags_5 |
---|
| 2311 | CASE ( 6 ) |
---|
| 2312 | flags => wall_flags_6 |
---|
| 2313 | CASE ( 7 ) |
---|
| 2314 | flags => wall_flags_7 |
---|
| 2315 | CASE ( 8 ) |
---|
| 2316 | flags => wall_flags_8 |
---|
| 2317 | CASE ( 9 ) |
---|
| 2318 | flags => wall_flags_9 |
---|
| 2319 | CASE ( 10 ) |
---|
| 2320 | flags => wall_flags_10 |
---|
| 2321 | END SELECT |
---|
| 2322 | |
---|
| 2323 | CALL sort_k_to_even_odd_blocks( flags, l ) |
---|
| 2324 | |
---|
| 2325 | ENDDO |
---|
| 2326 | |
---|
| 2327 | ENDIF |
---|
| 2328 | |
---|
| 2329 | lfirst = .FALSE. |
---|
| 2330 | |
---|
| 2331 | END SUBROUTINE init_even_odd_blocks |
---|
| 2332 | |
---|
| 2333 | |
---|
| 2334 | |
---|
| 2335 | SUBROUTINE special_exchange_horiz ( p_mg, color ) |
---|
| 2336 | |
---|
| 2337 | !------------------------------------------------------------------------------! |
---|
| 2338 | ! Description: |
---|
| 2339 | ! ------------ |
---|
| 2340 | ! Special exchange_horiz subroutine for use in redblack. Transfers only |
---|
| 2341 | ! "red" or "black" data points. |
---|
| 2342 | !------------------------------------------------------------------------------! |
---|
| 2343 | |
---|
| 2344 | USE control_parameters, & |
---|
| 2345 | ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, & |
---|
| 2346 | inflow_l, inflow_n, inflow_r, inflow_s, maximum_grid_level, & |
---|
| 2347 | outflow_l, outflow_n, outflow_r, outflow_s, & |
---|
| 2348 | synchronous_exchange |
---|
| 2349 | |
---|
| 2350 | USE indices, & |
---|
| 2351 | ONLY: mg_loc_ind, nxl, nxl_mg, nxr, nxr_mg, nys, nys_mg, nyn, & |
---|
| 2352 | nyn_mg, nzb, nzt, nzt_mg |
---|
| 2353 | |
---|
| 2354 | USE pegrid, & |
---|
| 2355 | ONLY: ngp_xz, ngp_yz |
---|
| 2356 | |
---|
| 2357 | IMPLICIT NONE |
---|
| 2358 | |
---|
| 2359 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
---|
| 2360 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
---|
| 2361 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p_mg !: |
---|
| 2362 | |
---|
| 2363 | INTEGER(iwp), intent(IN) :: color !: |
---|
| 2364 | ! |
---|
| 2365 | !-- Local variables |
---|
| 2366 | INTEGER(iwp) :: i,i1,i2 !: |
---|
| 2367 | INTEGER(iwp) :: j,j1,j2 !: |
---|
| 2368 | INTEGER(iwp) :: k !: |
---|
| 2369 | INTEGER(iwp) :: l !: |
---|
| 2370 | INTEGER(iwp) :: jys !: |
---|
| 2371 | INTEGER(iwp) :: jyn !: |
---|
| 2372 | INTEGER(iwp) :: ixl !: |
---|
| 2373 | INTEGER(iwp) :: ixr !: |
---|
| 2374 | logical :: sendrecv_in_background_save !: |
---|
| 2375 | logical :: synchronous_exchange_save !: |
---|
| 2376 | |
---|
| 2377 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, & |
---|
| 2378 | nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, & |
---|
| 2379 | nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1) :: temp !: |
---|
| 2380 | |
---|
| 2381 | |
---|
| 2382 | sendrecv_in_background_save = sendrecv_in_background |
---|
| 2383 | sendrecv_in_background = .FALSE. |
---|
| 2384 | synchronous_exchange_save = synchronous_exchange |
---|
| 2385 | synchronous_exchange = .FALSE. |
---|
| 2386 | |
---|
| 2387 | l = grid_level |
---|
| 2388 | |
---|
| 2389 | ind_even_odd = even_odd_level(grid_level) |
---|
| 2390 | |
---|
| 2391 | jys = nys_mg(grid_level-1) |
---|
| 2392 | jyn = nyn_mg(grid_level-1) |
---|
| 2393 | ixl = nxl_mg(grid_level-1) |
---|
| 2394 | ixr = nxr_mg(grid_level-1) |
---|
| 2395 | |
---|
| 2396 | ! |
---|
| 2397 | !-- Restricted transfer only on finer levels with enough data |
---|
| 2398 | IF ( ngp_xz(grid_level) >= 900 .OR. ngp_yz(grid_level) >= 900 ) THEN |
---|
| 2399 | ! |
---|
| 2400 | !-- Handling the even k Values |
---|
| 2401 | !-- Collecting data for the north - south exchange |
---|
| 2402 | !-- Since only every second value has to be transfered, data are stored |
---|
| 2403 | !-- on the next coarser grid level, because the arrays on that level |
---|
| 2404 | !-- have just the required size |
---|
| 2405 | i1 = nxl_mg(grid_level-1) |
---|
| 2406 | i2 = nxl_mg(grid_level-1) |
---|
| 2407 | |
---|
| 2408 | DO i = nxl_mg(l), nxr_mg(l), 2 |
---|
| 2409 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
---|
| 2410 | |
---|
| 2411 | IF ( j == nys_mg(l) ) THEN |
---|
| 2412 | !DIR$ IVDEP |
---|
| 2413 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2414 | temp(k-ind_even_odd,jys,i1) = p_mg(k,j,i) |
---|
| 2415 | ENDDO |
---|
| 2416 | i1 = i1 + 1 |
---|
| 2417 | |
---|
| 2418 | ENDIF |
---|
| 2419 | |
---|
| 2420 | IF ( j == nyn_mg(l) ) THEN |
---|
| 2421 | !DIR$ IVDEP |
---|
| 2422 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2423 | temp(k-ind_even_odd,jyn,i2) = p_mg(k,j,i) |
---|
| 2424 | ENDDO |
---|
| 2425 | i2 = i2 + 1 |
---|
| 2426 | |
---|
| 2427 | ENDIF |
---|
| 2428 | |
---|
| 2429 | ENDDO |
---|
| 2430 | ENDDO |
---|
| 2431 | |
---|
| 2432 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
---|
| 2433 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
---|
| 2434 | |
---|
| 2435 | IF ( j == nys_mg(l) ) THEN |
---|
| 2436 | !DIR$ IVDEP |
---|
| 2437 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2438 | temp(k-ind_even_odd,jys,i1) = p_mg(k,j,i) |
---|
| 2439 | ENDDO |
---|
| 2440 | i1 = i1 + 1 |
---|
| 2441 | |
---|
| 2442 | ENDIF |
---|
| 2443 | |
---|
| 2444 | IF ( j == nyn_mg(l) ) THEN |
---|
| 2445 | !DIR$ IVDEP |
---|
| 2446 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2447 | temp(k-ind_even_odd,jyn,i2) = p_mg(k,j,i) |
---|
| 2448 | ENDDO |
---|
| 2449 | i2 = i2 + 1 |
---|
| 2450 | |
---|
| 2451 | ENDIF |
---|
| 2452 | |
---|
| 2453 | ENDDO |
---|
| 2454 | ENDDO |
---|
| 2455 | |
---|
| 2456 | grid_level = grid_level-1 |
---|
| 2457 | |
---|
| 2458 | nxl = nxl_mg(grid_level) |
---|
| 2459 | nys = nys_mg(grid_level) |
---|
| 2460 | nxr = nxr_mg(grid_level) |
---|
| 2461 | nyn = nyn_mg(grid_level) |
---|
| 2462 | nzt = nzt_mg(grid_level) |
---|
| 2463 | |
---|
| 2464 | send_receive = 'ns' |
---|
| 2465 | CALL exchange_horiz( temp, 1 ) |
---|
| 2466 | |
---|
| 2467 | grid_level = grid_level+1 |
---|
| 2468 | |
---|
| 2469 | i1 = nxl_mg(grid_level-1) |
---|
| 2470 | i2 = nxl_mg(grid_level-1) |
---|
| 2471 | |
---|
| 2472 | DO i = nxl_mg(l), nxr_mg(l), 2 |
---|
| 2473 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
---|
| 2474 | |
---|
| 2475 | IF ( j == nys_mg(l) ) THEN |
---|
| 2476 | !DIR$ IVDEP |
---|
| 2477 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2478 | p_mg(k,nyn_mg(l)+1,i) = temp(k-ind_even_odd,jyn+1,i1) |
---|
| 2479 | ENDDO |
---|
| 2480 | i1 = i1 + 1 |
---|
| 2481 | |
---|
| 2482 | ENDIF |
---|
| 2483 | |
---|
| 2484 | IF ( j == nyn_mg(l) ) THEN |
---|
| 2485 | !DIR$ IVDEP |
---|
| 2486 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2487 | p_mg(k,nys_mg(l)-1,i) = temp(k-ind_even_odd,jys-1,i2) |
---|
| 2488 | ENDDO |
---|
| 2489 | i2 = i2 + 1 |
---|
| 2490 | |
---|
| 2491 | ENDIF |
---|
| 2492 | |
---|
| 2493 | ENDDO |
---|
| 2494 | ENDDO |
---|
| 2495 | |
---|
| 2496 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
---|
| 2497 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
---|
| 2498 | |
---|
| 2499 | IF ( j == nys_mg(l) ) THEN |
---|
| 2500 | !DIR$ IVDEP |
---|
| 2501 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2502 | p_mg(k,nyn_mg(l)+1,i) = temp(k-ind_even_odd,jyn+1,i1) |
---|
| 2503 | ENDDO |
---|
| 2504 | i1 = i1 + 1 |
---|
| 2505 | |
---|
| 2506 | ENDIF |
---|
| 2507 | |
---|
| 2508 | IF ( j == nyn_mg(l) ) THEN |
---|
| 2509 | !DIR$ IVDEP |
---|
| 2510 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2511 | p_mg(k,nys_mg(l)-1,i) = temp(k-ind_even_odd,jys-1,i2) |
---|
| 2512 | ENDDO |
---|
| 2513 | i2 = i2 + 1 |
---|
| 2514 | |
---|
| 2515 | ENDIF |
---|
| 2516 | |
---|
| 2517 | ENDDO |
---|
| 2518 | ENDDO |
---|
| 2519 | |
---|
| 2520 | ! |
---|
| 2521 | !-- Collecting data for the left - right exchange |
---|
| 2522 | !-- Since only every second value has to be transfered, data are stored |
---|
| 2523 | !-- on the next coarser grid level, because the arrays on that level |
---|
| 2524 | !-- have just the required size |
---|
| 2525 | j1 = nys_mg(grid_level-1) |
---|
| 2526 | j2 = nys_mg(grid_level-1) |
---|
| 2527 | |
---|
| 2528 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
---|
| 2529 | DO i = nxl_mg(l), nxr_mg(l), 2 |
---|
| 2530 | |
---|
| 2531 | IF ( i == nxl_mg(l) ) THEN |
---|
| 2532 | !DIR$ IVDEP |
---|
| 2533 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2534 | temp(k-ind_even_odd,j1,ixl) = p_mg(k,j,i) |
---|
| 2535 | ENDDO |
---|
| 2536 | j1 = j1 + 1 |
---|
| 2537 | |
---|
| 2538 | ENDIF |
---|
| 2539 | |
---|
| 2540 | IF ( i == nxr_mg(l) ) THEN |
---|
| 2541 | !DIR$ IVDEP |
---|
| 2542 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2543 | temp(k-ind_even_odd,j2,ixr) = p_mg(k,j,i) |
---|
| 2544 | ENDDO |
---|
| 2545 | j2 = j2 + 1 |
---|
| 2546 | |
---|
| 2547 | ENDIF |
---|
| 2548 | |
---|
| 2549 | ENDDO |
---|
| 2550 | ENDDO |
---|
| 2551 | |
---|
| 2552 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
---|
| 2553 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
---|
| 2554 | |
---|
| 2555 | IF ( i == nxl_mg(l) ) THEN |
---|
| 2556 | !DIR$ IVDEP |
---|
| 2557 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2558 | temp(k-ind_even_odd,j1,ixl) = p_mg(k,j,i) |
---|
| 2559 | ENDDO |
---|
| 2560 | j1 = j1 + 1 |
---|
| 2561 | |
---|
| 2562 | ENDIF |
---|
| 2563 | |
---|
| 2564 | IF ( i == nxr_mg(l) ) THEN |
---|
| 2565 | !DIR$ IVDEP |
---|
| 2566 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2567 | temp(k-ind_even_odd,j2,ixr) = p_mg(k,j,i) |
---|
| 2568 | ENDDO |
---|
| 2569 | j2 = j2 + 1 |
---|
| 2570 | |
---|
| 2571 | ENDIF |
---|
| 2572 | |
---|
| 2573 | ENDDO |
---|
| 2574 | ENDDO |
---|
| 2575 | |
---|
| 2576 | grid_level = grid_level-1 |
---|
| 2577 | send_receive = 'lr' |
---|
| 2578 | |
---|
| 2579 | CALL exchange_horiz( temp, 1 ) |
---|
| 2580 | |
---|
| 2581 | grid_level = grid_level+1 |
---|
| 2582 | |
---|
| 2583 | j1 = nys_mg(grid_level-1) |
---|
| 2584 | j2 = nys_mg(grid_level-1) |
---|
| 2585 | |
---|
| 2586 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
---|
| 2587 | DO i = nxl_mg(l), nxr_mg(l), 2 |
---|
| 2588 | |
---|
| 2589 | IF ( i == nxl_mg(l) ) THEN |
---|
| 2590 | !DIR$ IVDEP |
---|
| 2591 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2592 | p_mg(k,j,nxr_mg(l)+1) = temp(k-ind_even_odd,j1,ixr+1) |
---|
| 2593 | ENDDO |
---|
| 2594 | j1 = j1 + 1 |
---|
| 2595 | |
---|
| 2596 | ENDIF |
---|
| 2597 | |
---|
| 2598 | IF ( i == nxr_mg(l) ) THEN |
---|
| 2599 | !DIR$ IVDEP |
---|
| 2600 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2601 | p_mg(k,j,nxl_mg(l)-1) = temp(k-ind_even_odd,j2,ixl-1) |
---|
| 2602 | ENDDO |
---|
| 2603 | j2 = j2 + 1 |
---|
| 2604 | |
---|
| 2605 | ENDIF |
---|
| 2606 | |
---|
| 2607 | ENDDO |
---|
| 2608 | ENDDO |
---|
| 2609 | |
---|
| 2610 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
---|
| 2611 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
---|
| 2612 | |
---|
| 2613 | IF ( i == nxl_mg(l) ) THEN |
---|
| 2614 | !DIR$ IVDEP |
---|
| 2615 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2616 | p_mg(k,j,nxr_mg(l)+1) = temp(k-ind_even_odd,j1,ixr+1) |
---|
| 2617 | ENDDO |
---|
| 2618 | j1 = j1 + 1 |
---|
| 2619 | |
---|
| 2620 | ENDIF |
---|
| 2621 | |
---|
| 2622 | IF ( i == nxr_mg(l) ) THEN |
---|
| 2623 | !DIR$ IVDEP |
---|
| 2624 | DO k = ind_even_odd+1, nzt_mg(l) |
---|
| 2625 | p_mg(k,j,nxl_mg(l)-1) = temp(k-ind_even_odd,j2,ixl-1) |
---|
| 2626 | ENDDO |
---|
| 2627 | j2 = j2 + 1 |
---|
| 2628 | |
---|
| 2629 | ENDIF |
---|
| 2630 | |
---|
| 2631 | ENDDO |
---|
| 2632 | ENDDO |
---|
| 2633 | |
---|
| 2634 | ! |
---|
| 2635 | !-- Now handling the even k values |
---|
| 2636 | !-- Collecting data for the north - south exchange |
---|
| 2637 | !-- Since only every second value has to be transfered, data are stored |
---|
| 2638 | !-- on the next coarser grid level, because the arrays on that level |
---|
| 2639 | !-- have just the required size |
---|
| 2640 | i1 = nxl_mg(grid_level-1) |
---|
| 2641 | i2 = nxl_mg(grid_level-1) |
---|
| 2642 | |
---|
| 2643 | DO i = nxl_mg(l), nxr_mg(l), 2 |
---|
| 2644 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
---|
| 2645 | |
---|
| 2646 | IF ( j == nys_mg(l) ) THEN |
---|
| 2647 | !DIR$ IVDEP |
---|
| 2648 | DO k = nzb+1, ind_even_odd |
---|
| 2649 | temp(k,jys,i1) = p_mg(k,j,i) |
---|
| 2650 | ENDDO |
---|
| 2651 | i1 = i1 + 1 |
---|
| 2652 | |
---|
| 2653 | ENDIF |
---|
| 2654 | |
---|
| 2655 | IF ( j == nyn_mg(l) ) THEN |
---|
| 2656 | !DIR$ IVDEP |
---|
| 2657 | DO k = nzb+1, ind_even_odd |
---|
| 2658 | temp(k,jyn,i2) = p_mg(k,j,i) |
---|
| 2659 | ENDDO |
---|
| 2660 | i2 = i2 + 1 |
---|
| 2661 | |
---|
| 2662 | ENDIF |
---|
| 2663 | |
---|
| 2664 | ENDDO |
---|
| 2665 | ENDDO |
---|
| 2666 | |
---|
| 2667 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
---|
| 2668 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
---|
| 2669 | |
---|
| 2670 | IF ( j == nys_mg(l) ) THEN |
---|
| 2671 | !DIR$ IVDEP |
---|
| 2672 | DO k = nzb+1, ind_even_odd |
---|
| 2673 | temp(k,jys,i1) = p_mg(k,j,i) |
---|
| 2674 | ENDDO |
---|
| 2675 | i1 = i1 + 1 |
---|
| 2676 | |
---|
| 2677 | ENDIF |
---|
| 2678 | |
---|
| 2679 | IF ( j == nyn_mg(l) ) THEN |
---|
| 2680 | !DIR$ IVDEP |
---|
| 2681 | DO k = nzb+1, ind_even_odd |
---|
| 2682 | temp(k,jyn,i2) = p_mg(k,j,i) |
---|
| 2683 | ENDDO |
---|
| 2684 | i2 = i2 + 1 |
---|
| 2685 | |
---|
| 2686 | ENDIF |
---|
| 2687 | |
---|
| 2688 | ENDDO |
---|
| 2689 | ENDDO |
---|
| 2690 | |
---|
| 2691 | grid_level = grid_level-1 |
---|
| 2692 | |
---|
| 2693 | send_receive = 'ns' |
---|
| 2694 | CALL exchange_horiz( temp, 1 ) |
---|
| 2695 | |
---|
| 2696 | grid_level = grid_level+1 |
---|
| 2697 | |
---|
| 2698 | i1 = nxl_mg(grid_level-1) |
---|
| 2699 | i2 = nxl_mg(grid_level-1) |
---|
| 2700 | |
---|
| 2701 | DO i = nxl_mg(l), nxr_mg(l), 2 |
---|
| 2702 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
---|
| 2703 | |
---|
| 2704 | IF ( j == nys_mg(l) ) THEN |
---|
| 2705 | !DIR$ IVDEP |
---|
| 2706 | DO k = nzb+1, ind_even_odd |
---|
| 2707 | p_mg(k,nyn_mg(l)+1,i) = temp(k,jyn+1,i1) |
---|
| 2708 | ENDDO |
---|
| 2709 | i1 = i1 + 1 |
---|
| 2710 | |
---|
| 2711 | ENDIF |
---|
| 2712 | |
---|
| 2713 | IF ( j == nyn_mg(l) ) THEN |
---|
| 2714 | !DIR$ IVDEP |
---|
| 2715 | DO k = nzb+1, ind_even_odd |
---|
| 2716 | p_mg(k,nys_mg(l)-1,i) = temp(k,jys-1,i2) |
---|
| 2717 | ENDDO |
---|
| 2718 | i2 = i2 + 1 |
---|
| 2719 | |
---|
| 2720 | ENDIF |
---|
| 2721 | |
---|
| 2722 | ENDDO |
---|
| 2723 | ENDDO |
---|
| 2724 | |
---|
| 2725 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
---|
| 2726 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
---|
| 2727 | |
---|
| 2728 | IF ( j == nys_mg(l) ) THEN |
---|
| 2729 | !DIR$ IVDEP |
---|
| 2730 | DO k = nzb+1, ind_even_odd |
---|
| 2731 | p_mg(k,nyn_mg(l)+1,i) = temp(k,jyn+1,i1) |
---|
| 2732 | ENDDO |
---|
| 2733 | i1 = i1 + 1 |
---|
| 2734 | |
---|
| 2735 | ENDIF |
---|
| 2736 | |
---|
| 2737 | IF ( j == nyn_mg(l) ) THEN |
---|
| 2738 | !DIR$ IVDEP |
---|
| 2739 | DO k = nzb+1, ind_even_odd |
---|
| 2740 | p_mg(k,nys_mg(l)-1,i) = temp(k,jys-1,i2) |
---|
| 2741 | ENDDO |
---|
| 2742 | i2 = i2 + 1 |
---|
| 2743 | |
---|
| 2744 | ENDIF |
---|
| 2745 | |
---|
| 2746 | ENDDO |
---|
| 2747 | ENDDO |
---|
| 2748 | |
---|
| 2749 | j1 = nys_mg(grid_level-1) |
---|
| 2750 | j2 = nys_mg(grid_level-1) |
---|
| 2751 | |
---|
| 2752 | DO i = nxl_mg(l), nxr_mg(l), 2 |
---|
| 2753 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
---|
| 2754 | |
---|
| 2755 | IF ( i == nxl_mg(l) ) THEN |
---|
| 2756 | !DIR$ IVDEP |
---|
| 2757 | DO k = nzb+1, ind_even_odd |
---|
| 2758 | temp(k,j1,ixl) = p_mg(k,j,i) |
---|
| 2759 | ENDDO |
---|
| 2760 | j1 = j1 + 1 |
---|
| 2761 | |
---|
| 2762 | ENDIF |
---|
| 2763 | |
---|
| 2764 | IF ( i == nxr_mg(l) ) THEN |
---|
| 2765 | !DIR$ IVDEP |
---|
| 2766 | DO k = nzb+1, ind_even_odd |
---|
| 2767 | temp(k,j2,ixr) = p_mg(k,j,i) |
---|
| 2768 | ENDDO |
---|
| 2769 | j2 = j2 + 1 |
---|
| 2770 | |
---|
| 2771 | ENDIF |
---|
| 2772 | |
---|
| 2773 | ENDDO |
---|
| 2774 | ENDDO |
---|
| 2775 | |
---|
| 2776 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
---|
| 2777 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
---|
| 2778 | |
---|
| 2779 | IF ( i == nxl_mg(l) ) THEN |
---|
| 2780 | !DIR$ IVDEP |
---|
| 2781 | DO k = nzb+1, ind_even_odd |
---|
| 2782 | temp(k,j1,ixl) = p_mg(k,j,i) |
---|
| 2783 | ENDDO |
---|
| 2784 | j1 = j1 + 1 |
---|
| 2785 | |
---|
| 2786 | ENDIF |
---|
| 2787 | |
---|
| 2788 | IF ( i == nxr_mg(l) ) THEN |
---|
| 2789 | !DIR$ IVDEP |
---|
| 2790 | DO k = nzb+1, ind_even_odd |
---|
| 2791 | temp(k,j2,ixr) = p_mg(k,j,i) |
---|
| 2792 | ENDDO |
---|
| 2793 | j2 = j2 + 1 |
---|
| 2794 | |
---|
| 2795 | ENDIF |
---|
| 2796 | |
---|
| 2797 | ENDDO |
---|
| 2798 | ENDDO |
---|
| 2799 | |
---|
| 2800 | grid_level = grid_level-1 |
---|
| 2801 | |
---|
| 2802 | send_receive = 'lr' |
---|
| 2803 | CALL exchange_horiz( temp, 1 ) |
---|
| 2804 | |
---|
| 2805 | grid_level = grid_level+1 |
---|
| 2806 | |
---|
| 2807 | nxl = nxl_mg(grid_level) |
---|
| 2808 | nys = nys_mg(grid_level) |
---|
| 2809 | nxr = nxr_mg(grid_level) |
---|
| 2810 | nyn = nyn_mg(grid_level) |
---|
| 2811 | nzt = nzt_mg(grid_level) |
---|
| 2812 | |
---|
| 2813 | j1 = nys_mg(grid_level-1) |
---|
| 2814 | j2 = nys_mg(grid_level-1) |
---|
| 2815 | |
---|
| 2816 | DO i = nxl_mg(l), nxr_mg(l), 2 |
---|
| 2817 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
---|
| 2818 | |
---|
| 2819 | IF ( i == nxl_mg(l) ) THEN |
---|
| 2820 | !DIR$ IVDEP |
---|
| 2821 | DO k = nzb+1, ind_even_odd |
---|
| 2822 | p_mg(k,j,nxr_mg(l)+1) = temp(k,j1,ixr+1) |
---|
| 2823 | ENDDO |
---|
| 2824 | j1 = j1 + 1 |
---|
| 2825 | |
---|
| 2826 | ENDIF |
---|
| 2827 | |
---|
| 2828 | IF ( i == nxr_mg(l) ) THEN |
---|
| 2829 | !DIR$ IVDEP |
---|
| 2830 | DO k = nzb+1, ind_even_odd |
---|
| 2831 | p_mg(k,j,nxl_mg(l)-1) = temp(k,j2,ixl-1) |
---|
| 2832 | ENDDO |
---|
| 2833 | j2 = j2 + 1 |
---|
| 2834 | |
---|
| 2835 | ENDIF |
---|
| 2836 | |
---|
| 2837 | ENDDO |
---|
| 2838 | ENDDO |
---|
| 2839 | |
---|
| 2840 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
---|
| 2841 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
---|
| 2842 | |
---|
| 2843 | IF ( i == nxl_mg(l) ) THEN |
---|
| 2844 | !DIR$ IVDEP |
---|
| 2845 | DO k = nzb+1, ind_even_odd |
---|
| 2846 | p_mg(k,j,nxr_mg(l)+1) = temp(k,j1,ixr+1) |
---|
| 2847 | ENDDO |
---|
| 2848 | j1 = j1 + 1 |
---|
| 2849 | |
---|
| 2850 | ENDIF |
---|
| 2851 | |
---|
| 2852 | IF ( i == nxr_mg(l) ) THEN |
---|
| 2853 | !DIR$ IVDEP |
---|
| 2854 | DO k = nzb+1, ind_even_odd |
---|
| 2855 | p_mg(k,j,nxl_mg(l)-1) = temp(k,j2,ixl-1) |
---|
| 2856 | ENDDO |
---|
| 2857 | j2 = j2 + 1 |
---|
| 2858 | |
---|
| 2859 | ENDIF |
---|
| 2860 | |
---|
| 2861 | ENDDO |
---|
| 2862 | ENDDO |
---|
| 2863 | |
---|
| 2864 | ELSE |
---|
| 2865 | ! |
---|
| 2866 | !-- Standard horizontal ghost boundary exchange for small coarse grid |
---|
| 2867 | !-- levels, where the transfer time is latency bound |
---|
| 2868 | CALL exchange_horiz( p_mg, 1 ) |
---|
| 2869 | |
---|
| 2870 | ENDIF |
---|
| 2871 | |
---|
| 2872 | ! |
---|
| 2873 | !-- Reset values to default PALM setup |
---|
| 2874 | sendrecv_in_background = sendrecv_in_background_save |
---|
| 2875 | synchronous_exchange = synchronous_exchange_save |
---|
| 2876 | send_receive = 'al' |
---|
| 2877 | |
---|
| 2878 | END SUBROUTINE special_exchange_horiz |
---|
| 2879 | |
---|
| 2880 | END MODULE poismg_mod |
---|