| 1 | !> @file poismg_fast_mod.f90 |
|---|
| 2 | !--------------------------------------------------------------------------------! |
|---|
| 3 | ! This file is part of PALM. |
|---|
| 4 | ! |
|---|
| 5 | ! PALM is free software: you can redistribute it and/or modify it under the terms |
|---|
| 6 | ! of the GNU General Public License as published by the Free Software Foundation, |
|---|
| 7 | ! either version 3 of the License, or (at your option) any later version. |
|---|
| 8 | ! |
|---|
| 9 | ! PALM is distributed in the hope that it will be useful, but WITHOUT ANY |
|---|
| 10 | ! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR |
|---|
| 11 | ! A PARTICULAR PURPOSE. See the GNU General Public License for more details. |
|---|
| 12 | ! |
|---|
| 13 | ! You should have received a copy of the GNU General Public License along with |
|---|
| 14 | ! PALM. If not, see <http://www.gnu.org/licenses/>. |
|---|
| 15 | ! |
|---|
| 16 | ! Copyright 1997-2016 Leibniz Universitaet Hannover |
|---|
| 17 | !--------------------------------------------------------------------------------! |
|---|
| 18 | ! |
|---|
| 19 | ! Current revisions: |
|---|
| 20 | ! ----------------- |
|---|
| 21 | ! Bugfix: enable special_exchange_horiz only for finer grid levels. |
|---|
| 22 | ! Some formatting adjustments and variable descriptions. |
|---|
| 23 | ! |
|---|
| 24 | ! Former revisions: |
|---|
| 25 | ! ----------------- |
|---|
| 26 | ! $Id: poismg_fast_mod.f90 1904 2016-05-11 13:06:12Z suehring $ |
|---|
| 27 | ! |
|---|
| 28 | ! 1898 2016-05-03 11:27:17Z suehring |
|---|
| 29 | ! Bugfix: bottom and top boundary condition in resid_fast |
|---|
| 30 | ! Bugfix: restriction at nzb+1 |
|---|
| 31 | ! formatting adjustments, variable descriptions added in some declaration blocks |
|---|
| 32 | ! |
|---|
| 33 | ! 1850 2016-04-08 13:29:27Z maronga |
|---|
| 34 | ! Module renamed |
|---|
| 35 | ! |
|---|
| 36 | ! |
|---|
| 37 | ! 1762 2016-02-25 12:31:13Z hellstea |
|---|
| 38 | ! Introduction of nested domain feature |
|---|
| 39 | ! |
|---|
| 40 | ! 1682 2015-10-07 23:56:08Z knoop |
|---|
| 41 | ! Code annotations made doxygen readable |
|---|
| 42 | ! |
|---|
| 43 | ! 1609 2015-07-03 15:37:58Z maronga |
|---|
| 44 | ! Bugfix: allow compilation without __parallel. |
|---|
| 45 | ! |
|---|
| 46 | ! 1575 2015-03-27 09:56:27Z raasch |
|---|
| 47 | ! Initial revision. |
|---|
| 48 | ! Routine re-written and optimised based on poismg. |
|---|
| 49 | ! |
|---|
| 50 | ! Following optimisations have been made: |
|---|
| 51 | ! - vectorisation (for Intel-CPUs) of the red-black algorithm by resorting |
|---|
| 52 | ! array elements with even and odd indices |
|---|
| 53 | ! - explicit boundary conditions for building walls removed (solver is |
|---|
| 54 | ! running through the buildings |
|---|
| 55 | ! - reduced data transfer in case of ghost point exchange, because only |
|---|
| 56 | ! "red" or "black" data points need to be exchanged. This is not applied |
|---|
| 57 | ! for coarser grid levels, since for then the transfer time is latency bound |
|---|
| 58 | ! |
|---|
| 59 | ! |
|---|
| 60 | ! Description: |
|---|
| 61 | ! ------------ |
|---|
| 62 | !> Solves the Poisson equation for the perturbation pressure with a multigrid |
|---|
| 63 | !> V- or W-Cycle scheme. |
|---|
| 64 | !> |
|---|
| 65 | !> This multigrid method was originally developed for PALM by Joerg Uhlenbrock, |
|---|
| 66 | !> September 2000 - July 2001. It has been optimised for speed by Klaus |
|---|
| 67 | !> Ketelsen in November 2014. |
|---|
| 68 | !> |
|---|
| 69 | !> @attention Loop unrolling and cache optimization in SOR-Red/Black method |
|---|
| 70 | !> still does not give the expected speedup! |
|---|
| 71 | !> |
|---|
| 72 | !> @todo Further work required. |
|---|
| 73 | !------------------------------------------------------------------------------! |
|---|
| 74 | MODULE poismg_mod |
|---|
| 75 | |
|---|
| 76 | |
|---|
| 77 | USE cpulog, & |
|---|
| 78 | ONLY: cpu_log, log_point_s |
|---|
| 79 | |
|---|
| 80 | USE kinds |
|---|
| 81 | |
|---|
| 82 | USE pegrid |
|---|
| 83 | |
|---|
| 84 | PRIVATE |
|---|
| 85 | |
|---|
| 86 | INTEGER, SAVE :: ind_even_odd !< border index between even and odd k index |
|---|
| 87 | INTEGER, DIMENSION(:), SAVE, ALLOCATABLE :: even_odd_level !< stores ind_even_odd for all MG levels |
|---|
| 88 | |
|---|
| 89 | REAL(wp), DIMENSION(:,:), SAVE, ALLOCATABLE :: f1_mg_b, f2_mg_b, f3_mg_b !< blocked version of f1_mg ... |
|---|
| 90 | |
|---|
| 91 | INTERFACE poismg_fast |
|---|
| 92 | MODULE PROCEDURE poismg_fast |
|---|
| 93 | END INTERFACE poismg_fast |
|---|
| 94 | |
|---|
| 95 | INTERFACE sort_k_to_even_odd_blocks |
|---|
| 96 | MODULE PROCEDURE sort_k_to_even_odd_blocks |
|---|
| 97 | MODULE PROCEDURE sort_k_to_even_odd_blocks_int |
|---|
| 98 | MODULE PROCEDURE sort_k_to_even_odd_blocks_1d |
|---|
| 99 | END INTERFACE sort_k_to_even_odd_blocks |
|---|
| 100 | |
|---|
| 101 | PUBLIC poismg_fast |
|---|
| 102 | |
|---|
| 103 | CONTAINS |
|---|
| 104 | |
|---|
| 105 | !------------------------------------------------------------------------------! |
|---|
| 106 | ! Description: |
|---|
| 107 | ! ------------ |
|---|
| 108 | !> Solves the Poisson equation for the perturbation pressure with a multigrid |
|---|
| 109 | !> V- or W-Cycle scheme. |
|---|
| 110 | !------------------------------------------------------------------------------! |
|---|
| 111 | SUBROUTINE poismg_fast( r ) |
|---|
| 112 | |
|---|
| 113 | USE arrays_3d, & |
|---|
| 114 | ONLY: d, p_loc |
|---|
| 115 | |
|---|
| 116 | USE control_parameters, & |
|---|
| 117 | ONLY: gathered_size, grid_level, grid_level_count, & |
|---|
| 118 | maximum_grid_level, message_string, mgcycles, mg_cycles, & |
|---|
| 119 | mg_switch_to_pe0_level, residual_limit, subdomain_size |
|---|
| 120 | |
|---|
| 121 | USE cpulog, & |
|---|
| 122 | ONLY: cpu_log, log_point_s |
|---|
| 123 | |
|---|
| 124 | USE indices, & |
|---|
| 125 | ONLY: nxl, nxlg, nxl_mg, nxr, nxrg, nxr_mg, nys, nysg, nys_mg, nyn,& |
|---|
| 126 | nyng, nyn_mg, nzb, nzt, nzt_mg |
|---|
| 127 | |
|---|
| 128 | IMPLICIT NONE |
|---|
| 129 | |
|---|
| 130 | REAL(wp) :: maxerror !< |
|---|
| 131 | REAL(wp) :: maximum_mgcycles !< |
|---|
| 132 | REAL(wp) :: residual_norm !< |
|---|
| 133 | |
|---|
| 134 | REAL(wp), DIMENSION(nzb:nzt+1,nys-1:nyn+1,nxl-1:nxr+1) :: r !< |
|---|
| 135 | |
|---|
| 136 | REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: p3 !< |
|---|
| 137 | |
|---|
| 138 | |
|---|
| 139 | CALL cpu_log( log_point_s(29), 'poismg_fast', 'start' ) |
|---|
| 140 | ! |
|---|
| 141 | !-- Initialize arrays and variables used in this subroutine |
|---|
| 142 | |
|---|
| 143 | !-- If the number of grid points of the gathered grid, which is collected |
|---|
| 144 | !-- on PE0, is larger than the number of grid points of an PE, than array |
|---|
| 145 | !-- p3 will be enlarged. |
|---|
| 146 | IF ( gathered_size > subdomain_size ) THEN |
|---|
| 147 | ALLOCATE( p3(nzb:nzt_mg(mg_switch_to_pe0_level)+1,nys_mg( & |
|---|
| 148 | mg_switch_to_pe0_level)-1:nyn_mg(mg_switch_to_pe0_level)+1,& |
|---|
| 149 | nxl_mg(mg_switch_to_pe0_level)-1:nxr_mg( & |
|---|
| 150 | mg_switch_to_pe0_level)+1) ) |
|---|
| 151 | ELSE |
|---|
| 152 | ALLOCATE ( p3(nzb:nzt+1,nysg:nyng,nxlg:nxrg) ) |
|---|
| 153 | ENDIF |
|---|
| 154 | |
|---|
| 155 | p3 = 0.0_wp |
|---|
| 156 | |
|---|
| 157 | |
|---|
| 158 | ! |
|---|
| 159 | !-- Ghost boundaries have to be added to divergence array. |
|---|
| 160 | !-- Exchange routine needs to know the grid level! |
|---|
| 161 | grid_level = maximum_grid_level |
|---|
| 162 | CALL exchange_horiz( d, 1) |
|---|
| 163 | d(nzb,:,:) = d(nzb+1,:,:) |
|---|
| 164 | |
|---|
| 165 | ! |
|---|
| 166 | !-- Initiation of the multigrid scheme. Does n cycles until the |
|---|
| 167 | !-- residual is smaller than the given limit. The accuracy of the solution |
|---|
| 168 | !-- of the poisson equation will increase with the number of cycles. |
|---|
| 169 | !-- If the number of cycles is preset by the user, this number will be |
|---|
| 170 | !-- carried out regardless of the accuracy. |
|---|
| 171 | grid_level_count = 0 |
|---|
| 172 | mgcycles = 0 |
|---|
| 173 | IF ( mg_cycles == -1 ) THEN |
|---|
| 174 | maximum_mgcycles = 0 |
|---|
| 175 | residual_norm = 1.0_wp |
|---|
| 176 | ELSE |
|---|
| 177 | maximum_mgcycles = mg_cycles |
|---|
| 178 | residual_norm = 0.0_wp |
|---|
| 179 | ENDIF |
|---|
| 180 | |
|---|
| 181 | ! |
|---|
| 182 | !-- Initial settings for sorting k-dimension from sequential order (alternate |
|---|
| 183 | !-- even/odd) into blocks of even and odd or vice versa |
|---|
| 184 | CALL init_even_odd_blocks |
|---|
| 185 | |
|---|
| 186 | ! |
|---|
| 187 | !-- Sort input arrays in even/odd blocks along k-dimension |
|---|
| 188 | CALL sort_k_to_even_odd_blocks( d, grid_level ) |
|---|
| 189 | CALL sort_k_to_even_odd_blocks( p_loc, grid_level ) |
|---|
| 190 | |
|---|
| 191 | ! |
|---|
| 192 | !-- The complete multigrid cycles are running in block mode, i.e. over |
|---|
| 193 | !-- seperate data blocks of even and odd indices |
|---|
| 194 | DO WHILE ( residual_norm > residual_limit .OR. & |
|---|
| 195 | mgcycles < maximum_mgcycles ) |
|---|
| 196 | |
|---|
| 197 | CALL next_mg_level_fast( d, p_loc, p3, r) |
|---|
| 198 | |
|---|
| 199 | ! |
|---|
| 200 | !-- Calculate the residual if the user has not preset the number of |
|---|
| 201 | !-- cycles to be performed |
|---|
| 202 | IF ( maximum_mgcycles == 0 ) THEN |
|---|
| 203 | CALL resid_fast( d, p_loc, r ) |
|---|
| 204 | maxerror = SUM( r(nzb+1:nzt,nys:nyn,nxl:nxr)**2 ) |
|---|
| 205 | |
|---|
| 206 | #if defined( __parallel ) |
|---|
| 207 | IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr ) |
|---|
| 208 | CALL MPI_ALLREDUCE( maxerror, residual_norm, 1, MPI_REAL, & |
|---|
| 209 | MPI_SUM, comm2d, ierr) |
|---|
| 210 | #else |
|---|
| 211 | residual_norm = maxerror |
|---|
| 212 | #endif |
|---|
| 213 | residual_norm = SQRT( residual_norm ) |
|---|
| 214 | ENDIF |
|---|
| 215 | |
|---|
| 216 | mgcycles = mgcycles + 1 |
|---|
| 217 | |
|---|
| 218 | ! |
|---|
| 219 | !-- If the user has not limited the number of cycles, stop the run in case |
|---|
| 220 | !-- of insufficient convergence |
|---|
| 221 | IF ( mgcycles > 1000 .AND. mg_cycles == -1 ) THEN |
|---|
| 222 | message_string = 'no sufficient convergence within 1000 cycles' |
|---|
| 223 | CALL message( 'poismg_fast', 'PA0283', 1, 2, 0, 6, 0 ) |
|---|
| 224 | ENDIF |
|---|
| 225 | |
|---|
| 226 | ENDDO |
|---|
| 227 | |
|---|
| 228 | DEALLOCATE( p3 ) |
|---|
| 229 | ! |
|---|
| 230 | !-- Result has to be sorted back from even/odd blocks to sequential order |
|---|
| 231 | CALL sort_k_to_sequential( p_loc ) |
|---|
| 232 | ! |
|---|
| 233 | !-- Unset the grid level. Variable is used to determine the MPI datatypes for |
|---|
| 234 | !-- ghost point exchange |
|---|
| 235 | grid_level = 0 |
|---|
| 236 | |
|---|
| 237 | CALL cpu_log( log_point_s(29), 'poismg_fast', 'stop' ) |
|---|
| 238 | |
|---|
| 239 | END SUBROUTINE poismg_fast |
|---|
| 240 | |
|---|
| 241 | |
|---|
| 242 | !------------------------------------------------------------------------------! |
|---|
| 243 | ! Description: |
|---|
| 244 | ! ------------ |
|---|
| 245 | !> Computes the residual of the perturbation pressure. |
|---|
| 246 | !------------------------------------------------------------------------------! |
|---|
| 247 | SUBROUTINE resid_fast( f_mg, p_mg, r ) |
|---|
| 248 | |
|---|
| 249 | |
|---|
| 250 | USE arrays_3d, & |
|---|
| 251 | ONLY: f1_mg, f2_mg, f3_mg |
|---|
| 252 | |
|---|
| 253 | USE control_parameters, & |
|---|
| 254 | ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, inflow_l,& |
|---|
| 255 | inflow_n, inflow_r, inflow_s, nest_bound_l, nest_bound_n, & |
|---|
| 256 | nest_bound_r, nest_bound_s, outflow_l, outflow_n, outflow_r, & |
|---|
| 257 | outflow_s |
|---|
| 258 | |
|---|
| 259 | USE grid_variables, & |
|---|
| 260 | ONLY: ddx2_mg, ddy2_mg |
|---|
| 261 | |
|---|
| 262 | USE indices, & |
|---|
| 263 | ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
|---|
| 264 | |
|---|
| 265 | IMPLICIT NONE |
|---|
| 266 | |
|---|
| 267 | INTEGER(iwp) :: i !< index variable along x |
|---|
| 268 | INTEGER(iwp) :: j !< index variable along y |
|---|
| 269 | INTEGER(iwp) :: k !< index variable along z |
|---|
| 270 | INTEGER(iwp) :: l !< index indicating grid level |
|---|
| 271 | INTEGER(iwp) :: km1 !< index variable along z dimension (k-1) |
|---|
| 272 | INTEGER(iwp) :: kp1 !< index variable along z dimension (k+1) |
|---|
| 273 | |
|---|
| 274 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 275 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 276 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f_mg !< velocity divergence |
|---|
| 277 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 278 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 279 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p_mg !< perturbation pressure |
|---|
| 280 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 281 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 282 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: r !< residuum of perturbation pressure |
|---|
| 283 | |
|---|
| 284 | ! |
|---|
| 285 | !-- Calculate the residual |
|---|
| 286 | l = grid_level |
|---|
| 287 | |
|---|
| 288 | CALL cpu_log( log_point_s(53), 'resid', 'start' ) |
|---|
| 289 | !$OMP PARALLEL PRIVATE (i,j,k,km1,kp1) |
|---|
| 290 | !$OMP DO |
|---|
| 291 | DO i = nxl_mg(l), nxr_mg(l) |
|---|
| 292 | DO j = nys_mg(l), nyn_mg(l) |
|---|
| 293 | !DIR$ IVDEP |
|---|
| 294 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 295 | km1 = k-ind_even_odd-1 |
|---|
| 296 | kp1 = k-ind_even_odd |
|---|
| 297 | r(k,j,i) = f_mg(k,j,i) & |
|---|
| 298 | - ddx2_mg(l) * & |
|---|
| 299 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 300 | - ddy2_mg(l) * & |
|---|
| 301 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 302 | - f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 303 | - f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 304 | + f1_mg_b(k,l) * p_mg(k,j,i) |
|---|
| 305 | ENDDO |
|---|
| 306 | !DIR$ IVDEP |
|---|
| 307 | DO k = nzb+1, ind_even_odd |
|---|
| 308 | km1 = k+ind_even_odd |
|---|
| 309 | kp1 = k+ind_even_odd+1 |
|---|
| 310 | r(k,j,i) = f_mg(k,j,i) & |
|---|
| 311 | - ddx2_mg(l) * & |
|---|
| 312 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 313 | - ddy2_mg(l) * & |
|---|
| 314 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 315 | - f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 316 | - f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 317 | + f1_mg_b(k,l) * p_mg(k,j,i) |
|---|
| 318 | ENDDO |
|---|
| 319 | ENDDO |
|---|
| 320 | ENDDO |
|---|
| 321 | !$OMP END PARALLEL |
|---|
| 322 | ! |
|---|
| 323 | !-- Horizontal boundary conditions |
|---|
| 324 | CALL exchange_horiz( r, 1) |
|---|
| 325 | |
|---|
| 326 | IF ( .NOT. bc_lr_cyc ) THEN |
|---|
| 327 | IF ( inflow_l .OR. outflow_l .OR. nest_bound_l ) THEN |
|---|
| 328 | r(:,:,nxl_mg(l)-1) = r(:,:,nxl_mg(l)) |
|---|
| 329 | ENDIF |
|---|
| 330 | IF ( inflow_r .OR. outflow_r .OR. nest_bound_r ) THEN |
|---|
| 331 | r(:,:,nxr_mg(l)+1) = r(:,:,nxr_mg(l)) |
|---|
| 332 | ENDIF |
|---|
| 333 | ENDIF |
|---|
| 334 | |
|---|
| 335 | IF ( .NOT. bc_ns_cyc ) THEN |
|---|
| 336 | IF ( inflow_n .OR. outflow_n .OR. nest_bound_n ) THEN |
|---|
| 337 | r(:,nyn_mg(l)+1,:) = r(:,nyn_mg(l),:) |
|---|
| 338 | ENDIF |
|---|
| 339 | IF ( inflow_s .OR. outflow_s .OR. nest_bound_s ) THEN |
|---|
| 340 | r(:,nys_mg(l)-1,:) = r(:,nys_mg(l),:) |
|---|
| 341 | ENDIF |
|---|
| 342 | ENDIF |
|---|
| 343 | |
|---|
| 344 | ! |
|---|
| 345 | !-- Boundary conditions at bottom and top of the domain. Points may be within |
|---|
| 346 | !-- buildings, but that doesn't matter. |
|---|
| 347 | IF ( ibc_p_b == 1 ) THEN |
|---|
| 348 | ! |
|---|
| 349 | !-- equivalent to r(nzb,:,: ) = r(nzb+1,:,:) |
|---|
| 350 | r(nzb,:,: ) = r(ind_even_odd+1,:,:) |
|---|
| 351 | ELSE |
|---|
| 352 | r(nzb,:,: ) = 0.0_wp |
|---|
| 353 | ENDIF |
|---|
| 354 | |
|---|
| 355 | IF ( ibc_p_t == 1 ) THEN |
|---|
| 356 | ! |
|---|
| 357 | !-- equivalent to r(nzt_mg(l)+1,:,: ) = r(nzt_mg(l),:,:) |
|---|
| 358 | r(nzt_mg(l)+1,:,: ) = r(ind_even_odd,:,:) |
|---|
| 359 | ELSE |
|---|
| 360 | r(nzt_mg(l)+1,:,: ) = 0.0_wp |
|---|
| 361 | ENDIF |
|---|
| 362 | |
|---|
| 363 | CALL cpu_log( log_point_s(53), 'resid', 'stop' ) |
|---|
| 364 | |
|---|
| 365 | END SUBROUTINE resid_fast |
|---|
| 366 | |
|---|
| 367 | |
|---|
| 368 | !------------------------------------------------------------------------------! |
|---|
| 369 | ! Description: |
|---|
| 370 | ! ------------ |
|---|
| 371 | !> Interpolates the residual on the next coarser grid with "full weighting" |
|---|
| 372 | !> scheme |
|---|
| 373 | !------------------------------------------------------------------------------! |
|---|
| 374 | SUBROUTINE restrict_fast( f_mg, r ) |
|---|
| 375 | |
|---|
| 376 | |
|---|
| 377 | USE control_parameters, & |
|---|
| 378 | ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, inflow_l,& |
|---|
| 379 | inflow_n, inflow_r, inflow_s, nest_bound_l, nest_bound_n, & |
|---|
| 380 | nest_bound_r, nest_bound_s, outflow_l, outflow_n, outflow_r, & |
|---|
| 381 | outflow_s |
|---|
| 382 | |
|---|
| 383 | USE indices, & |
|---|
| 384 | ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
|---|
| 385 | |
|---|
| 386 | IMPLICIT NONE |
|---|
| 387 | |
|---|
| 388 | INTEGER(iwp) :: i !< index variable along x on finer grid |
|---|
| 389 | INTEGER(iwp) :: ic !< index variable along x on coarser grid |
|---|
| 390 | INTEGER(iwp) :: j !< index variable along y on finer grid |
|---|
| 391 | INTEGER(iwp) :: jc !< index variable along y on coarser grid |
|---|
| 392 | INTEGER(iwp) :: k !< index variable along z on finer grid |
|---|
| 393 | INTEGER(iwp) :: kc !< index variable along z on coarser grid |
|---|
| 394 | INTEGER(iwp) :: l !< index indicating finer grid level |
|---|
| 395 | INTEGER(iwp) :: km1 !< index variable along z dimension (k-1 on finer level) |
|---|
| 396 | INTEGER(iwp) :: kp1 !< index variable along z dimension (k+1 on finer level) |
|---|
| 397 | |
|---|
| 398 | |
|---|
| 399 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 400 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 401 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: & |
|---|
| 402 | f_mg !< Residual on coarser grid level |
|---|
| 403 | |
|---|
| 404 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level+1)+1, & |
|---|
| 405 | nys_mg(grid_level+1)-1:nyn_mg(grid_level+1)+1, & |
|---|
| 406 | nxl_mg(grid_level+1)-1:nxr_mg(grid_level+1)+1) :: & |
|---|
| 407 | r !< Residual on finer grid level |
|---|
| 408 | |
|---|
| 409 | ! |
|---|
| 410 | !-- Interpolate the residual |
|---|
| 411 | l = grid_level |
|---|
| 412 | |
|---|
| 413 | CALL cpu_log( log_point_s(54), 'restrict', 'start' ) |
|---|
| 414 | ! |
|---|
| 415 | !-- No wall treatment |
|---|
| 416 | !$OMP PARALLEL PRIVATE (i,j,k,ic,jc,kc,km1,kp1) |
|---|
| 417 | DO ic = nxl_mg(l), nxr_mg(l) |
|---|
| 418 | i = 2*ic |
|---|
| 419 | !$OMP DO SCHEDULE( STATIC ) |
|---|
| 420 | DO jc = nys_mg(l), nyn_mg(l) |
|---|
| 421 | ! |
|---|
| 422 | !-- Calculation for the first point along k |
|---|
| 423 | j = 2*jc |
|---|
| 424 | ! |
|---|
| 425 | !-- Calculation for the other points along k |
|---|
| 426 | !DIR$ IVDEP |
|---|
| 427 | DO k = ind_even_odd+1, nzt_mg(l+1) ! Fine grid at this point |
|---|
| 428 | km1 = k-ind_even_odd-1 |
|---|
| 429 | kp1 = k-ind_even_odd |
|---|
| 430 | kc = k-ind_even_odd ! Coarse grid index |
|---|
| 431 | |
|---|
| 432 | f_mg(kc,jc,ic) = 1.0_wp / 64.0_wp * ( & |
|---|
| 433 | 8.0_wp * r(k,j,i) & |
|---|
| 434 | + 4.0_wp * ( r(k,j,i-1) + r(k,j,i+1) + & |
|---|
| 435 | r(k,j+1,i) + r(k,j-1,i) ) & |
|---|
| 436 | + 2.0_wp * ( r(k,j-1,i-1) + r(k,j+1,i-1) + & |
|---|
| 437 | r(k,j-1,i+1) + r(k,j+1,i+1) ) & |
|---|
| 438 | + 4.0_wp * r(km1,j,i) & |
|---|
| 439 | + 2.0_wp * ( r(km1,j,i-1) + r(km1,j,i+1) + & |
|---|
| 440 | r(km1,j+1,i) + r(km1,j-1,i) ) & |
|---|
| 441 | + ( r(km1,j-1,i-1) + r(km1,j+1,i-1) + & |
|---|
| 442 | r(km1,j-1,i+1) + r(km1,j+1,i+1) ) & |
|---|
| 443 | + 4.0_wp * r(kp1,j,i) & |
|---|
| 444 | + 2.0_wp * ( r(kp1,j,i-1) + r(kp1,j,i+1) + & |
|---|
| 445 | r(kp1,j+1,i) + r(kp1,j-1,i) ) & |
|---|
| 446 | + ( r(kp1,j-1,i-1) + r(kp1,j+1,i-1) + & |
|---|
| 447 | r(kp1,j-1,i+1) + r(kp1,j+1,i+1) ) & |
|---|
| 448 | ) |
|---|
| 449 | ENDDO |
|---|
| 450 | ENDDO |
|---|
| 451 | !$OMP ENDDO nowait |
|---|
| 452 | ENDDO |
|---|
| 453 | !$OMP END PARALLEL |
|---|
| 454 | |
|---|
| 455 | ! |
|---|
| 456 | !-- Ghost point exchange |
|---|
| 457 | CALL exchange_horiz( f_mg, 1) |
|---|
| 458 | ! |
|---|
| 459 | !-- Horizontal boundary conditions |
|---|
| 460 | IF ( .NOT. bc_lr_cyc ) THEN |
|---|
| 461 | IF ( inflow_l .OR. outflow_l .OR. nest_bound_l ) THEN |
|---|
| 462 | f_mg(:,:,nxl_mg(l)-1) = f_mg(:,:,nxl_mg(l)) |
|---|
| 463 | ENDIF |
|---|
| 464 | IF ( inflow_r .OR. outflow_r .OR. nest_bound_r ) THEN |
|---|
| 465 | f_mg(:,:,nxr_mg(l)+1) = f_mg(:,:,nxr_mg(l)) |
|---|
| 466 | ENDIF |
|---|
| 467 | ENDIF |
|---|
| 468 | |
|---|
| 469 | IF ( .NOT. bc_ns_cyc ) THEN |
|---|
| 470 | IF ( inflow_n .OR. outflow_n .OR. nest_bound_n ) THEN |
|---|
| 471 | f_mg(:,nyn_mg(l)+1,:) = f_mg(:,nyn_mg(l),:) |
|---|
| 472 | ENDIF |
|---|
| 473 | IF ( inflow_s .OR. outflow_s .OR. nest_bound_s ) THEN |
|---|
| 474 | f_mg(:,nys_mg(l)-1,:) = f_mg(:,nys_mg(l),:) |
|---|
| 475 | ENDIF |
|---|
| 476 | ENDIF |
|---|
| 477 | |
|---|
| 478 | ! |
|---|
| 479 | !-- Boundary conditions at bottom and top of the domain. |
|---|
| 480 | !-- These points are not handled by the above loop. Points may be within |
|---|
| 481 | !-- buildings, but that doesn't matter. Remark: f_mg is ordered sequentielly |
|---|
| 482 | !-- after interpolation on coarse grid (is ordered in odd-even blocks further |
|---|
| 483 | !-- below). |
|---|
| 484 | IF ( ibc_p_b == 1 ) THEN |
|---|
| 485 | f_mg(nzb,:,: ) = f_mg(nzb+1,:,:) |
|---|
| 486 | ELSE |
|---|
| 487 | f_mg(nzb,:,: ) = 0.0_wp |
|---|
| 488 | ENDIF |
|---|
| 489 | |
|---|
| 490 | IF ( ibc_p_t == 1 ) THEN |
|---|
| 491 | f_mg(nzt_mg(l)+1,:,: ) = f_mg(nzt_mg(l),:,:) |
|---|
| 492 | ELSE |
|---|
| 493 | f_mg(nzt_mg(l)+1,:,: ) = 0.0_wp |
|---|
| 494 | ENDIF |
|---|
| 495 | |
|---|
| 496 | CALL cpu_log( log_point_s(54), 'restrict', 'stop' ) |
|---|
| 497 | ! |
|---|
| 498 | !-- Since residual is in sequential order after interpolation, an additional |
|---|
| 499 | !-- sorting in odd-even blocks along z dimension is required at this point. |
|---|
| 500 | CALL sort_k_to_even_odd_blocks( f_mg , l) |
|---|
| 501 | |
|---|
| 502 | END SUBROUTINE restrict_fast |
|---|
| 503 | |
|---|
| 504 | |
|---|
| 505 | !------------------------------------------------------------------------------! |
|---|
| 506 | ! Description: |
|---|
| 507 | ! ------------ |
|---|
| 508 | !> Interpolates the correction of the perturbation pressure |
|---|
| 509 | !> to the next finer grid. |
|---|
| 510 | !------------------------------------------------------------------------------! |
|---|
| 511 | SUBROUTINE prolong_fast( p, temp ) |
|---|
| 512 | |
|---|
| 513 | |
|---|
| 514 | USE control_parameters, & |
|---|
| 515 | ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, inflow_l,& |
|---|
| 516 | inflow_n, inflow_r, inflow_s, nest_bound_l, nest_bound_n, & |
|---|
| 517 | nest_bound_r, nest_bound_s, outflow_l, outflow_n, & |
|---|
| 518 | outflow_r, outflow_s |
|---|
| 519 | |
|---|
| 520 | USE indices, & |
|---|
| 521 | ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
|---|
| 522 | |
|---|
| 523 | IMPLICIT NONE |
|---|
| 524 | |
|---|
| 525 | INTEGER(iwp) :: i !< index variable along x on coarser grid level |
|---|
| 526 | INTEGER(iwp) :: j !< index variable along y on coarser grid level |
|---|
| 527 | INTEGER(iwp) :: k !< index variable along z on coarser grid level |
|---|
| 528 | INTEGER(iwp) :: l !< index indicating finer grid level |
|---|
| 529 | INTEGER(iwp) :: kp1 !< index variable along z |
|---|
| 530 | INTEGER(iwp) :: ke !< index for prolog even |
|---|
| 531 | INTEGER(iwp) :: ko !< index for prolog odd |
|---|
| 532 | |
|---|
| 533 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, & |
|---|
| 534 | nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, & |
|---|
| 535 | nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1 ) :: & |
|---|
| 536 | p !< perturbation pressure on coarser grid level |
|---|
| 537 | |
|---|
| 538 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 539 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 540 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: & |
|---|
| 541 | temp !< perturbation pressure on finer grid level |
|---|
| 542 | |
|---|
| 543 | |
|---|
| 544 | CALL cpu_log( log_point_s(55), 'prolong', 'start' ) |
|---|
| 545 | |
|---|
| 546 | ! |
|---|
| 547 | !-- First, store elements of the coarser grid on the next finer grid |
|---|
| 548 | l = grid_level |
|---|
| 549 | ind_even_odd = even_odd_level(grid_level-1) |
|---|
| 550 | |
|---|
| 551 | !$OMP PARALLEL PRIVATE (i,j,k,kp1,ke,ko) |
|---|
| 552 | !$OMP DO |
|---|
| 553 | DO i = nxl_mg(l-1), nxr_mg(l-1) |
|---|
| 554 | DO j = nys_mg(l-1), nyn_mg(l-1) |
|---|
| 555 | |
|---|
| 556 | !DIR$ IVDEP |
|---|
| 557 | DO k = ind_even_odd+1, nzt_mg(l-1) |
|---|
| 558 | kp1 = k - ind_even_odd |
|---|
| 559 | ke = 2 * ( k-ind_even_odd - 1 ) + 1 |
|---|
| 560 | ko = 2 * k - 1 |
|---|
| 561 | ! |
|---|
| 562 | !-- Points of the coarse grid are directly stored on the next finer |
|---|
| 563 | !-- grid |
|---|
| 564 | temp(ko,2*j,2*i) = p(k,j,i) |
|---|
| 565 | ! |
|---|
| 566 | !-- Points between two coarse-grid points |
|---|
| 567 | temp(ko,2*j,2*i+1) = 0.5_wp * ( p(k,j,i) + p(k,j,i+1) ) |
|---|
| 568 | temp(ko,2*j+1,2*i) = 0.5_wp * ( p(k,j,i) + p(k,j+1,i) ) |
|---|
| 569 | temp(ke,2*j,2*i) = 0.5_wp * ( p(k,j,i) + p(kp1,j,i) ) |
|---|
| 570 | ! |
|---|
| 571 | !-- Points in the center of the planes stretched by four points |
|---|
| 572 | !-- of the coarse grid cube |
|---|
| 573 | temp(ko,2*j+1,2*i+1) = 0.25_wp * ( p(k,j,i) + p(k,j,i+1) + & |
|---|
| 574 | p(k,j+1,i) + p(k,j+1,i+1) ) |
|---|
| 575 | temp(ke,2*j,2*i+1) = 0.25_wp * ( p(k,j,i) + p(k,j,i+1) + & |
|---|
| 576 | p(kp1,j,i) + p(kp1,j,i+1) ) |
|---|
| 577 | temp(ke,2*j+1,2*i) = 0.25_wp * ( p(k,j,i) + p(k,j+1,i) + & |
|---|
| 578 | p(kp1,j,i) + p(kp1,j+1,i) ) |
|---|
| 579 | ! |
|---|
| 580 | !-- Points in the middle of coarse grid cube |
|---|
| 581 | temp(ke,2*j+1,2*i+1) = 0.125_wp * & |
|---|
| 582 | ( p(k,j,i) + p(k,j,i+1) + & |
|---|
| 583 | p(k,j+1,i) + p(k,j+1,i+1) + & |
|---|
| 584 | p(kp1,j,i) + p(kp1,j,i+1) + & |
|---|
| 585 | p(kp1,j+1,i) + p(kp1,j+1,i+1) ) |
|---|
| 586 | |
|---|
| 587 | ENDDO |
|---|
| 588 | |
|---|
| 589 | !DIR$ IVDEP |
|---|
| 590 | DO k = nzb+1, ind_even_odd |
|---|
| 591 | kp1 = k + ind_even_odd + 1 |
|---|
| 592 | ke = 2 * k |
|---|
| 593 | ko = 2 * ( k + ind_even_odd ) |
|---|
| 594 | ! |
|---|
| 595 | !-- Points of the coarse grid are directly stored on the next finer |
|---|
| 596 | !-- grid |
|---|
| 597 | temp(ko,2*j,2*i) = p(k,j,i) |
|---|
| 598 | ! |
|---|
| 599 | !-- Points between two coarse-grid points |
|---|
| 600 | temp(ko,2*j,2*i+1) = 0.5_wp * ( p(k,j,i) + p(k,j,i+1) ) |
|---|
| 601 | temp(ko,2*j+1,2*i) = 0.5_wp * ( p(k,j,i) + p(k,j+1,i) ) |
|---|
| 602 | temp(ke,2*j,2*i) = 0.5_wp * ( p(k,j,i) + p(kp1,j,i) ) |
|---|
| 603 | ! |
|---|
| 604 | !-- Points in the center of the planes stretched by four points |
|---|
| 605 | !-- of the coarse grid cube |
|---|
| 606 | temp(ko,2*j+1,2*i+1) = 0.25_wp * ( p(k,j,i) + p(k,j,i+1) + & |
|---|
| 607 | p(k,j+1,i) + p(k,j+1,i+1) ) |
|---|
| 608 | temp(ke,2*j,2*i+1) = 0.25_wp * ( p(k,j,i) + p(k,j,i+1) + & |
|---|
| 609 | p(kp1,j,i) + p(kp1,j,i+1) ) |
|---|
| 610 | temp(ke,2*j+1,2*i) = 0.25_wp * ( p(k,j,i) + p(k,j+1,i) + & |
|---|
| 611 | p(kp1,j,i) + p(kp1,j+1,i) ) |
|---|
| 612 | ! |
|---|
| 613 | !-- Points in the middle of coarse grid cube |
|---|
| 614 | temp(ke,2*j+1,2*i+1) = 0.125_wp * & |
|---|
| 615 | ( p(k,j,i) + p(k,j,i+1) + & |
|---|
| 616 | p(k,j+1,i) + p(k,j+1,i+1) + & |
|---|
| 617 | p(kp1,j,i) + p(kp1,j,i+1) + & |
|---|
| 618 | p(kp1,j+1,i) + p(kp1,j+1,i+1) ) |
|---|
| 619 | |
|---|
| 620 | ENDDO |
|---|
| 621 | |
|---|
| 622 | ENDDO |
|---|
| 623 | ENDDO |
|---|
| 624 | !$OMP END PARALLEL |
|---|
| 625 | |
|---|
| 626 | ind_even_odd = even_odd_level(grid_level) |
|---|
| 627 | ! |
|---|
| 628 | !-- Horizontal boundary conditions |
|---|
| 629 | CALL exchange_horiz( temp, 1) |
|---|
| 630 | |
|---|
| 631 | IF ( .NOT. bc_lr_cyc ) THEN |
|---|
| 632 | IF ( inflow_l .OR. outflow_l .OR. nest_bound_l ) THEN |
|---|
| 633 | temp(:,:,nxl_mg(l)-1) = temp(:,:,nxl_mg(l)) |
|---|
| 634 | ENDIF |
|---|
| 635 | IF ( inflow_r .OR. outflow_r .OR. nest_bound_r ) THEN |
|---|
| 636 | temp(:,:,nxr_mg(l)+1) = temp(:,:,nxr_mg(l)) |
|---|
| 637 | ENDIF |
|---|
| 638 | ENDIF |
|---|
| 639 | |
|---|
| 640 | IF ( .NOT. bc_ns_cyc ) THEN |
|---|
| 641 | IF ( inflow_n .OR. outflow_n .OR. nest_bound_n ) THEN |
|---|
| 642 | temp(:,nyn_mg(l)+1,:) = temp(:,nyn_mg(l),:) |
|---|
| 643 | ENDIF |
|---|
| 644 | IF ( inflow_s .OR. outflow_s .OR. nest_bound_s ) THEN |
|---|
| 645 | temp(:,nys_mg(l)-1,:) = temp(:,nys_mg(l),:) |
|---|
| 646 | ENDIF |
|---|
| 647 | ENDIF |
|---|
| 648 | |
|---|
| 649 | ! |
|---|
| 650 | !-- Bottom and top boundary conditions |
|---|
| 651 | IF ( ibc_p_b == 1 ) THEN |
|---|
| 652 | ! |
|---|
| 653 | !-- equivalent to temp(nzb,:,: ) = temp(nzb+1,:,:) |
|---|
| 654 | temp(nzb,:,: ) = temp(ind_even_odd+1,:,:) |
|---|
| 655 | ELSE |
|---|
| 656 | temp(nzb,:,: ) = 0.0_wp |
|---|
| 657 | ENDIF |
|---|
| 658 | |
|---|
| 659 | IF ( ibc_p_t == 1 ) THEN |
|---|
| 660 | ! |
|---|
| 661 | !-- equivalent to temp(nzt_mg(l)+1,:,: ) = temp(nzt_mg(l),:,:) |
|---|
| 662 | temp(nzt_mg(l)+1,:,: ) = temp(ind_even_odd,:,:) |
|---|
| 663 | ELSE |
|---|
| 664 | temp(nzt_mg(l)+1,:,: ) = 0.0_wp |
|---|
| 665 | ENDIF |
|---|
| 666 | |
|---|
| 667 | CALL cpu_log( log_point_s(55), 'prolong', 'stop' ) |
|---|
| 668 | |
|---|
| 669 | END SUBROUTINE prolong_fast |
|---|
| 670 | |
|---|
| 671 | |
|---|
| 672 | !------------------------------------------------------------------------------! |
|---|
| 673 | ! Description: |
|---|
| 674 | ! ------------ |
|---|
| 675 | !> Relaxation method for the multigrid scheme. A Gauss-Seidel iteration with |
|---|
| 676 | !> 3D-Red-Black decomposition (GS-RB) is used. |
|---|
| 677 | !------------------------------------------------------------------------------! |
|---|
| 678 | SUBROUTINE redblack_fast( f_mg, p_mg ) |
|---|
| 679 | |
|---|
| 680 | |
|---|
| 681 | USE arrays_3d, & |
|---|
| 682 | ONLY: f1_mg, f2_mg, f3_mg |
|---|
| 683 | |
|---|
| 684 | USE control_parameters, & |
|---|
| 685 | ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, inflow_l,& |
|---|
| 686 | inflow_n, inflow_r, inflow_s, nest_bound_l, nest_bound_n, & |
|---|
| 687 | nest_bound_r, nest_bound_s, ngsrb, outflow_l, outflow_n, & |
|---|
| 688 | outflow_r, outflow_s |
|---|
| 689 | |
|---|
| 690 | USE grid_variables, & |
|---|
| 691 | ONLY: ddx2_mg, ddy2_mg |
|---|
| 692 | |
|---|
| 693 | USE indices, & |
|---|
| 694 | ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
|---|
| 695 | |
|---|
| 696 | IMPLICIT NONE |
|---|
| 697 | |
|---|
| 698 | INTEGER(iwp) :: color !< grid point color, either red or black |
|---|
| 699 | INTEGER(iwp) :: i !< index variable along x |
|---|
| 700 | INTEGER(iwp) :: ic !< index variable along x |
|---|
| 701 | INTEGER(iwp) :: j !< index variable along y |
|---|
| 702 | INTEGER(iwp) :: jc !< index variable along y |
|---|
| 703 | INTEGER(iwp) :: jj !< index variable along y |
|---|
| 704 | INTEGER(iwp) :: k !< index variable along z |
|---|
| 705 | INTEGER(iwp) :: l !< grid level |
|---|
| 706 | INTEGER(iwp) :: n !< loop variable GauÃ-Seidel iterations |
|---|
| 707 | INTEGER(iwp) :: km1 !< index variable (k-1) |
|---|
| 708 | INTEGER(iwp) :: kp1 !< index variable (k+1) |
|---|
| 709 | |
|---|
| 710 | LOGICAL :: unroll !< flag indicating whether loop unrolling is possible |
|---|
| 711 | |
|---|
| 712 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 713 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 714 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: & |
|---|
| 715 | f_mg !< residual of perturbation pressure |
|---|
| 716 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 717 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 718 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: & |
|---|
| 719 | p_mg !< perturbation pressure |
|---|
| 720 | |
|---|
| 721 | l = grid_level |
|---|
| 722 | |
|---|
| 723 | unroll = ( MOD( nyn_mg(l)-nys_mg(l)+1, 4 ) == 0 .AND. & |
|---|
| 724 | MOD( nxr_mg(l)-nxl_mg(l)+1, 2 ) == 0 ) |
|---|
| 725 | |
|---|
| 726 | DO n = 1, ngsrb |
|---|
| 727 | |
|---|
| 728 | DO color = 1, 2 |
|---|
| 729 | |
|---|
| 730 | IF ( .NOT. unroll ) THEN |
|---|
| 731 | |
|---|
| 732 | CALL cpu_log( log_point_s(36), 'redblack_no_unroll_f', 'start' ) |
|---|
| 733 | ! |
|---|
| 734 | !-- Without unrolling of loops, no cache optimization |
|---|
| 735 | !$OMP PARALLEL PRIVATE (i,j,k,km1,kp1) |
|---|
| 736 | !$OMP DO |
|---|
| 737 | DO i = nxl_mg(l), nxr_mg(l), 2 |
|---|
| 738 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
|---|
| 739 | !DIR$ IVDEP |
|---|
| 740 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 741 | km1 = k-ind_even_odd-1 |
|---|
| 742 | kp1 = k-ind_even_odd |
|---|
| 743 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
|---|
| 744 | ddx2_mg(l) * & |
|---|
| 745 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 746 | + ddy2_mg(l) * & |
|---|
| 747 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 748 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 749 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 750 | - f_mg(k,j,i) ) |
|---|
| 751 | ENDDO |
|---|
| 752 | ENDDO |
|---|
| 753 | ENDDO |
|---|
| 754 | |
|---|
| 755 | !$OMP DO |
|---|
| 756 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
|---|
| 757 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
|---|
| 758 | !DIR$ IVDEP |
|---|
| 759 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 760 | km1 = k-ind_even_odd-1 |
|---|
| 761 | kp1 = k-ind_even_odd |
|---|
| 762 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
|---|
| 763 | ddx2_mg(l) * & |
|---|
| 764 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 765 | + ddy2_mg(l) * & |
|---|
| 766 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 767 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 768 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 769 | - f_mg(k,j,i) ) |
|---|
| 770 | ENDDO |
|---|
| 771 | ENDDO |
|---|
| 772 | ENDDO |
|---|
| 773 | |
|---|
| 774 | !$OMP DO |
|---|
| 775 | DO i = nxl_mg(l), nxr_mg(l), 2 |
|---|
| 776 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
|---|
| 777 | !DIR$ IVDEP |
|---|
| 778 | DO k = nzb+1, ind_even_odd |
|---|
| 779 | km1 = k+ind_even_odd |
|---|
| 780 | kp1 = k+ind_even_odd+1 |
|---|
| 781 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
|---|
| 782 | ddx2_mg(l) * & |
|---|
| 783 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 784 | + ddy2_mg(l) * & |
|---|
| 785 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 786 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 787 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 788 | - f_mg(k,j,i) ) |
|---|
| 789 | ENDDO |
|---|
| 790 | ENDDO |
|---|
| 791 | ENDDO |
|---|
| 792 | |
|---|
| 793 | !$OMP DO |
|---|
| 794 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
|---|
| 795 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
|---|
| 796 | !DIR$ IVDEP |
|---|
| 797 | DO k = nzb+1, ind_even_odd |
|---|
| 798 | km1 = k+ind_even_odd |
|---|
| 799 | kp1 = k+ind_even_odd+1 |
|---|
| 800 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
|---|
| 801 | ddx2_mg(l) * & |
|---|
| 802 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 803 | + ddy2_mg(l) * & |
|---|
| 804 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 805 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 806 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 807 | - f_mg(k,j,i) ) |
|---|
| 808 | ENDDO |
|---|
| 809 | ENDDO |
|---|
| 810 | ENDDO |
|---|
| 811 | !$OMP END PARALLEL |
|---|
| 812 | |
|---|
| 813 | CALL cpu_log( log_point_s(36), 'redblack_no_unroll_f', 'stop' ) |
|---|
| 814 | |
|---|
| 815 | ELSE |
|---|
| 816 | ! |
|---|
| 817 | !-- Loop unrolling along y, only one i loop for better cache use |
|---|
| 818 | CALL cpu_log( log_point_s(38), 'redblack_unroll_f', 'start' ) |
|---|
| 819 | |
|---|
| 820 | !$OMP PARALLEL PRIVATE (i,j,k,ic,jc,km1,kp1,jj) |
|---|
| 821 | !$OMP DO |
|---|
| 822 | DO ic = nxl_mg(l), nxr_mg(l), 2 |
|---|
| 823 | DO jc = nys_mg(l), nyn_mg(l), 4 |
|---|
| 824 | i = ic |
|---|
| 825 | jj = jc+2-color |
|---|
| 826 | !DIR$ IVDEP |
|---|
| 827 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 828 | km1 = k-ind_even_odd-1 |
|---|
| 829 | kp1 = k-ind_even_odd |
|---|
| 830 | j = jj |
|---|
| 831 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
|---|
| 832 | ddx2_mg(l) * & |
|---|
| 833 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 834 | + ddy2_mg(l) * & |
|---|
| 835 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 836 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 837 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 838 | - f_mg(k,j,i) ) |
|---|
| 839 | j = jj+2 |
|---|
| 840 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
|---|
| 841 | ddx2_mg(l) * & |
|---|
| 842 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 843 | + ddy2_mg(l) * & |
|---|
| 844 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 845 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 846 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 847 | - f_mg(k,j,i) ) |
|---|
| 848 | ENDDO |
|---|
| 849 | |
|---|
| 850 | i = ic+1 |
|---|
| 851 | jj = jc+color-1 |
|---|
| 852 | !DIR$ IVDEP |
|---|
| 853 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 854 | km1 = k-ind_even_odd-1 |
|---|
| 855 | kp1 = k-ind_even_odd |
|---|
| 856 | j = jj |
|---|
| 857 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
|---|
| 858 | ddx2_mg(l) * & |
|---|
| 859 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 860 | + ddy2_mg(l) * & |
|---|
| 861 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 862 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 863 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 864 | - f_mg(k,j,i) ) |
|---|
| 865 | j = jj+2 |
|---|
| 866 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
|---|
| 867 | ddx2_mg(l) * & |
|---|
| 868 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 869 | + ddy2_mg(l) * & |
|---|
| 870 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 871 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 872 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 873 | - f_mg(k,j,i) ) |
|---|
| 874 | ENDDO |
|---|
| 875 | |
|---|
| 876 | i = ic |
|---|
| 877 | jj = jc+color-1 |
|---|
| 878 | !DIR$ IVDEP |
|---|
| 879 | DO k = nzb+1, ind_even_odd |
|---|
| 880 | km1 = k+ind_even_odd |
|---|
| 881 | kp1 = k+ind_even_odd+1 |
|---|
| 882 | j = jj |
|---|
| 883 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
|---|
| 884 | ddx2_mg(l) * & |
|---|
| 885 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 886 | + ddy2_mg(l) * & |
|---|
| 887 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 888 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 889 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 890 | - f_mg(k,j,i) ) |
|---|
| 891 | j = jj+2 |
|---|
| 892 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
|---|
| 893 | ddx2_mg(l) * & |
|---|
| 894 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 895 | + ddy2_mg(l) * & |
|---|
| 896 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 897 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 898 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 899 | - f_mg(k,j,i) ) |
|---|
| 900 | ENDDO |
|---|
| 901 | |
|---|
| 902 | i = ic+1 |
|---|
| 903 | jj = jc+2-color |
|---|
| 904 | !DIR$ IVDEP |
|---|
| 905 | DO k = nzb+1, ind_even_odd |
|---|
| 906 | km1 = k+ind_even_odd |
|---|
| 907 | kp1 = k+ind_even_odd+1 |
|---|
| 908 | j = jj |
|---|
| 909 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
|---|
| 910 | ddx2_mg(l) * & |
|---|
| 911 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 912 | + ddy2_mg(l) * & |
|---|
| 913 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 914 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 915 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 916 | - f_mg(k,j,i) ) |
|---|
| 917 | j = jj+2 |
|---|
| 918 | p_mg(k,j,i) = 1.0_wp / f1_mg_b(k,l) * ( & |
|---|
| 919 | ddx2_mg(l) * & |
|---|
| 920 | ( p_mg(k,j,i+1) + p_mg(k,j,i-1) ) & |
|---|
| 921 | + ddy2_mg(l) * & |
|---|
| 922 | ( p_mg(k,j+1,i) + p_mg(k,j-1,i) ) & |
|---|
| 923 | + f2_mg_b(k,l) * p_mg(kp1,j,i) & |
|---|
| 924 | + f3_mg_b(k,l) * p_mg(km1,j,i) & |
|---|
| 925 | - f_mg(k,j,i) ) |
|---|
| 926 | ENDDO |
|---|
| 927 | |
|---|
| 928 | ENDDO |
|---|
| 929 | ENDDO |
|---|
| 930 | !$OMP END PARALLEL |
|---|
| 931 | |
|---|
| 932 | CALL cpu_log( log_point_s(38), 'redblack_unroll_f', 'stop' ) |
|---|
| 933 | |
|---|
| 934 | ENDIF |
|---|
| 935 | |
|---|
| 936 | ! |
|---|
| 937 | !-- Horizontal boundary conditions |
|---|
| 938 | CALL special_exchange_horiz( p_mg, color ) |
|---|
| 939 | |
|---|
| 940 | IF ( .NOT. bc_lr_cyc ) THEN |
|---|
| 941 | IF ( inflow_l .OR. outflow_l .OR. nest_bound_l ) THEN |
|---|
| 942 | p_mg(:,:,nxl_mg(l)-1) = p_mg(:,:,nxl_mg(l)) |
|---|
| 943 | ENDIF |
|---|
| 944 | IF ( inflow_r .OR. outflow_r .OR. nest_bound_r ) THEN |
|---|
| 945 | p_mg(:,:,nxr_mg(l)+1) = p_mg(:,:,nxr_mg(l)) |
|---|
| 946 | ENDIF |
|---|
| 947 | ENDIF |
|---|
| 948 | |
|---|
| 949 | IF ( .NOT. bc_ns_cyc ) THEN |
|---|
| 950 | IF ( inflow_n .OR. outflow_n .OR. nest_bound_n ) THEN |
|---|
| 951 | p_mg(:,nyn_mg(l)+1,:) = p_mg(:,nyn_mg(l),:) |
|---|
| 952 | ENDIF |
|---|
| 953 | IF ( inflow_s .OR. outflow_s .OR. nest_bound_s ) THEN |
|---|
| 954 | p_mg(:,nys_mg(l)-1,:) = p_mg(:,nys_mg(l),:) |
|---|
| 955 | ENDIF |
|---|
| 956 | ENDIF |
|---|
| 957 | |
|---|
| 958 | ! |
|---|
| 959 | !-- Bottom and top boundary conditions |
|---|
| 960 | IF ( ibc_p_b == 1 ) THEN |
|---|
| 961 | ! |
|---|
| 962 | !-- equivalent to p_mg(nzb,:,: ) = p_mg(nzb+1,:,:) |
|---|
| 963 | p_mg(nzb,:,: ) = p_mg(ind_even_odd+1,:,:) |
|---|
| 964 | ELSE |
|---|
| 965 | p_mg(nzb,:,: ) = 0.0_wp |
|---|
| 966 | ENDIF |
|---|
| 967 | |
|---|
| 968 | IF ( ibc_p_t == 1 ) THEN |
|---|
| 969 | ! |
|---|
| 970 | !-- equivalent to p_mg(nzt_mg(l)+1,:,: ) = p_mg(nzt_mg(l),:,:) |
|---|
| 971 | p_mg(nzt_mg(l)+1,:,: ) = p_mg(ind_even_odd,:,:) |
|---|
| 972 | ELSE |
|---|
| 973 | p_mg(nzt_mg(l)+1,:,: ) = 0.0_wp |
|---|
| 974 | ENDIF |
|---|
| 975 | |
|---|
| 976 | ENDDO |
|---|
| 977 | |
|---|
| 978 | ENDDO |
|---|
| 979 | |
|---|
| 980 | END SUBROUTINE redblack_fast |
|---|
| 981 | |
|---|
| 982 | |
|---|
| 983 | !------------------------------------------------------------------------------! |
|---|
| 984 | ! Description: |
|---|
| 985 | ! ------------ |
|---|
| 986 | !> Sort k-Dimension from sequential into blocks of even and odd. |
|---|
| 987 | !> This is required to vectorize the red-black subroutine. |
|---|
| 988 | !> Version for 3D-REAL arrays |
|---|
| 989 | !------------------------------------------------------------------------------! |
|---|
| 990 | SUBROUTINE sort_k_to_even_odd_blocks( p_mg , glevel ) |
|---|
| 991 | |
|---|
| 992 | |
|---|
| 993 | USE control_parameters, & |
|---|
| 994 | ONLY: grid_level |
|---|
| 995 | |
|---|
| 996 | USE indices, & |
|---|
| 997 | ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
|---|
| 998 | |
|---|
| 999 | IMPLICIT NONE |
|---|
| 1000 | |
|---|
| 1001 | INTEGER(iwp), INTENT(IN) :: glevel !< grid level |
|---|
| 1002 | |
|---|
| 1003 | REAL(wp), DIMENSION(nzb:nzt_mg(glevel)+1, & |
|---|
| 1004 | nys_mg(glevel)-1:nyn_mg(glevel)+1, & |
|---|
| 1005 | nxl_mg(glevel)-1:nxr_mg(glevel)+1) :: & |
|---|
| 1006 | p_mg !< array to be sorted |
|---|
| 1007 | ! |
|---|
| 1008 | !-- Local variables |
|---|
| 1009 | INTEGER(iwp) :: i !< index variable along x |
|---|
| 1010 | INTEGER(iwp) :: j !< index variable along y |
|---|
| 1011 | INTEGER(iwp) :: k !< index variable along z |
|---|
| 1012 | INTEGER(iwp) :: l !< grid level |
|---|
| 1013 | INTEGER(iwp) :: ind !< index variable along z |
|---|
| 1014 | REAL(wp), DIMENSION(nzb:nzt_mg(glevel)+1) :: tmp !< odd-even sorted temporary array |
|---|
| 1015 | |
|---|
| 1016 | |
|---|
| 1017 | CALL cpu_log( log_point_s(52), 'sort_k_to_even_odd', 'start' ) |
|---|
| 1018 | |
|---|
| 1019 | l = glevel |
|---|
| 1020 | ind_even_odd = even_odd_level(l) |
|---|
| 1021 | |
|---|
| 1022 | !$OMP PARALLEL PRIVATE (i,j,k,ind,tmp) |
|---|
| 1023 | !$OMP DO |
|---|
| 1024 | DO i = nxl_mg(l)-1, nxr_mg(l)+1 |
|---|
| 1025 | DO j = nys_mg(l)-1, nyn_mg(l)+1 |
|---|
| 1026 | |
|---|
| 1027 | ! |
|---|
| 1028 | !-- Sort the data with even k index |
|---|
| 1029 | ind = nzb-1 |
|---|
| 1030 | DO k = nzb, nzt_mg(l), 2 |
|---|
| 1031 | ind = ind + 1 |
|---|
| 1032 | tmp(ind) = p_mg(k,j,i) |
|---|
| 1033 | ENDDO |
|---|
| 1034 | ! |
|---|
| 1035 | !-- Sort the data with odd k index |
|---|
| 1036 | DO k = nzb+1, nzt_mg(l)+1, 2 |
|---|
| 1037 | ind = ind + 1 |
|---|
| 1038 | tmp(ind) = p_mg(k,j,i) |
|---|
| 1039 | ENDDO |
|---|
| 1040 | |
|---|
| 1041 | p_mg(:,j,i) = tmp |
|---|
| 1042 | |
|---|
| 1043 | ENDDO |
|---|
| 1044 | ENDDO |
|---|
| 1045 | !$OMP END PARALLEL |
|---|
| 1046 | |
|---|
| 1047 | CALL cpu_log( log_point_s(52), 'sort_k_to_even_odd', 'stop' ) |
|---|
| 1048 | |
|---|
| 1049 | END SUBROUTINE sort_k_to_even_odd_blocks |
|---|
| 1050 | |
|---|
| 1051 | |
|---|
| 1052 | !------------------------------------------------------------------------------! |
|---|
| 1053 | ! Description: |
|---|
| 1054 | ! ------------ |
|---|
| 1055 | !> Sort k-Dimension from sequential into blocks of even and odd. |
|---|
| 1056 | !> This is required to vectorize the red-black subroutine. |
|---|
| 1057 | !> Version for 1D-REAL arrays |
|---|
| 1058 | !------------------------------------------------------------------------------! |
|---|
| 1059 | SUBROUTINE sort_k_to_even_odd_blocks_1d( f_mg, f_mg_b, glevel ) |
|---|
| 1060 | |
|---|
| 1061 | |
|---|
| 1062 | USE indices, & |
|---|
| 1063 | ONLY: nzb, nzt_mg |
|---|
| 1064 | |
|---|
| 1065 | IMPLICIT NONE |
|---|
| 1066 | |
|---|
| 1067 | INTEGER(iwp), INTENT(IN) :: glevel !< grid level |
|---|
| 1068 | |
|---|
| 1069 | REAL(wp), DIMENSION(nzb+1:nzt_mg(glevel)) :: f_mg !< 1D input array |
|---|
| 1070 | REAL(wp), DIMENSION(nzb:nzt_mg(glevel)+1) :: f_mg_b !< 1D output array |
|---|
| 1071 | |
|---|
| 1072 | ! |
|---|
| 1073 | !-- Local variables |
|---|
| 1074 | INTEGER(iwp) :: ind !< index variable along z |
|---|
| 1075 | INTEGER(iwp) :: k !< index variable along z |
|---|
| 1076 | |
|---|
| 1077 | |
|---|
| 1078 | ind = nzb - 1 |
|---|
| 1079 | ! |
|---|
| 1080 | !-- Sort the data with even k index |
|---|
| 1081 | DO k = nzb, nzt_mg(glevel), 2 |
|---|
| 1082 | ind = ind + 1 |
|---|
| 1083 | IF ( k >= nzb+1 .AND. k <= nzt_mg(glevel) ) THEN |
|---|
| 1084 | f_mg_b(ind) = f_mg(k) |
|---|
| 1085 | ENDIF |
|---|
| 1086 | ENDDO |
|---|
| 1087 | ! |
|---|
| 1088 | !-- Sort the data with odd k index |
|---|
| 1089 | DO k = nzb+1, nzt_mg(glevel)+1, 2 |
|---|
| 1090 | ind = ind + 1 |
|---|
| 1091 | IF( k >= nzb+1 .AND. k <= nzt_mg(glevel) ) THEN |
|---|
| 1092 | f_mg_b(ind) = f_mg(k) |
|---|
| 1093 | ENDIF |
|---|
| 1094 | ENDDO |
|---|
| 1095 | |
|---|
| 1096 | END SUBROUTINE sort_k_to_even_odd_blocks_1d |
|---|
| 1097 | |
|---|
| 1098 | |
|---|
| 1099 | !------------------------------------------------------------------------------! |
|---|
| 1100 | ! Description: |
|---|
| 1101 | ! ------------ |
|---|
| 1102 | !> Sort k-Dimension from sequential into blocks of even and odd. |
|---|
| 1103 | !> This is required to vectorize the red-black subroutine. |
|---|
| 1104 | !> Version for 2D-INTEGER arrays |
|---|
| 1105 | !------------------------------------------------------------------------------! |
|---|
| 1106 | SUBROUTINE sort_k_to_even_odd_blocks_int( i_mg , glevel ) |
|---|
| 1107 | |
|---|
| 1108 | |
|---|
| 1109 | USE control_parameters, & |
|---|
| 1110 | ONLY: grid_level |
|---|
| 1111 | |
|---|
| 1112 | USE indices, & |
|---|
| 1113 | ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
|---|
| 1114 | |
|---|
| 1115 | IMPLICIT NONE |
|---|
| 1116 | |
|---|
| 1117 | INTEGER(iwp), INTENT(IN) :: glevel !< grid level |
|---|
| 1118 | |
|---|
| 1119 | INTEGER(iwp), DIMENSION(nzb:nzt_mg(glevel)+1, & |
|---|
| 1120 | nys_mg(glevel)-1:nyn_mg(glevel)+1, & |
|---|
| 1121 | nxl_mg(glevel)-1:nxr_mg(glevel)+1) :: & |
|---|
| 1122 | i_mg !< array to be sorted |
|---|
| 1123 | ! |
|---|
| 1124 | !-- Local variables |
|---|
| 1125 | INTEGER(iwp) :: i !< index variabel along x |
|---|
| 1126 | INTEGER(iwp) :: j !< index variable along y |
|---|
| 1127 | INTEGER(iwp) :: k !< index variable along z |
|---|
| 1128 | INTEGER(iwp) :: l !< grid level |
|---|
| 1129 | INTEGER(iwp) :: ind !< index variable along z |
|---|
| 1130 | INTEGER(iwp),DIMENSION(nzb:nzt_mg(glevel)+1) :: tmp !< temporary odd-even sorted array |
|---|
| 1131 | |
|---|
| 1132 | |
|---|
| 1133 | CALL cpu_log( log_point_s(52), 'sort_k_to_even_odd', 'start' ) |
|---|
| 1134 | |
|---|
| 1135 | l = glevel |
|---|
| 1136 | ind_even_odd = even_odd_level(l) |
|---|
| 1137 | |
|---|
| 1138 | DO i = nxl_mg(l)-1, nxr_mg(l)+1 |
|---|
| 1139 | DO j = nys_mg(l)-1, nyn_mg(l)+1 |
|---|
| 1140 | |
|---|
| 1141 | ! |
|---|
| 1142 | !-- Sort the data with even k index |
|---|
| 1143 | ind = nzb-1 |
|---|
| 1144 | DO k = nzb, nzt_mg(l), 2 |
|---|
| 1145 | ind = ind + 1 |
|---|
| 1146 | tmp(ind) = i_mg(k,j,i) |
|---|
| 1147 | ENDDO |
|---|
| 1148 | ! |
|---|
| 1149 | !++ ATTENTION: Check reason for this error. Remove it or replace WRITE |
|---|
| 1150 | !++ by PALM message |
|---|
| 1151 | #if defined ( __parallel ) |
|---|
| 1152 | IF ( ind /= ind_even_odd ) THEN |
|---|
| 1153 | WRITE (0,*) 'ERROR ==> illegal ind_even_odd ',ind,ind_even_odd,l |
|---|
| 1154 | CALL MPI_ABORT(MPI_COMM_WORLD,i,j) |
|---|
| 1155 | ENDIF |
|---|
| 1156 | #endif |
|---|
| 1157 | ! |
|---|
| 1158 | !-- Sort the data with odd k index |
|---|
| 1159 | DO k = nzb+1, nzt_mg(l)+1, 2 |
|---|
| 1160 | ind = ind + 1 |
|---|
| 1161 | tmp(ind) = i_mg(k,j,i) |
|---|
| 1162 | ENDDO |
|---|
| 1163 | |
|---|
| 1164 | i_mg(:,j,i) = tmp |
|---|
| 1165 | |
|---|
| 1166 | ENDDO |
|---|
| 1167 | ENDDO |
|---|
| 1168 | |
|---|
| 1169 | CALL cpu_log( log_point_s(52), 'sort_k_to_even_odd', 'stop' ) |
|---|
| 1170 | |
|---|
| 1171 | END SUBROUTINE sort_k_to_even_odd_blocks_int |
|---|
| 1172 | |
|---|
| 1173 | |
|---|
| 1174 | !------------------------------------------------------------------------------! |
|---|
| 1175 | ! Description: |
|---|
| 1176 | ! ------------ |
|---|
| 1177 | !> Sort k-dimension from blocks of even and odd into sequential |
|---|
| 1178 | !------------------------------------------------------------------------------! |
|---|
| 1179 | SUBROUTINE sort_k_to_sequential( p_mg ) |
|---|
| 1180 | |
|---|
| 1181 | |
|---|
| 1182 | USE control_parameters, & |
|---|
| 1183 | ONLY: grid_level |
|---|
| 1184 | |
|---|
| 1185 | USE indices, & |
|---|
| 1186 | ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
|---|
| 1187 | |
|---|
| 1188 | IMPLICIT NONE |
|---|
| 1189 | |
|---|
| 1190 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 1191 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 1192 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: & |
|---|
| 1193 | p_mg !< array to be sorted |
|---|
| 1194 | ! |
|---|
| 1195 | !-- Local variables |
|---|
| 1196 | INTEGER(iwp) :: i !< index variable along x |
|---|
| 1197 | INTEGER(iwp) :: j !< index variable along y |
|---|
| 1198 | INTEGER(iwp) :: k !< index variable along z |
|---|
| 1199 | INTEGER(iwp) :: l !< grid level |
|---|
| 1200 | INTEGER(iwp) :: ind !< index variable along z |
|---|
| 1201 | |
|---|
| 1202 | REAL(wp),DIMENSION(nzb:nzt_mg(grid_level)+1) :: tmp |
|---|
| 1203 | |
|---|
| 1204 | |
|---|
| 1205 | l = grid_level |
|---|
| 1206 | |
|---|
| 1207 | !$OMP PARALLEL PRIVATE (i,j,k,ind,tmp) |
|---|
| 1208 | !$OMP DO |
|---|
| 1209 | DO i = nxl_mg(l)-1, nxr_mg(l)+1 |
|---|
| 1210 | DO j = nys_mg(l)-1, nyn_mg(l)+1 |
|---|
| 1211 | |
|---|
| 1212 | ind = nzb - 1 |
|---|
| 1213 | tmp = p_mg(:,j,i) |
|---|
| 1214 | DO k = nzb, nzt_mg(l), 2 |
|---|
| 1215 | ind = ind + 1 |
|---|
| 1216 | p_mg(k,j,i) = tmp(ind) |
|---|
| 1217 | ENDDO |
|---|
| 1218 | |
|---|
| 1219 | DO k = nzb+1, nzt_mg(l)+1, 2 |
|---|
| 1220 | ind = ind + 1 |
|---|
| 1221 | p_mg(k,j,i) = tmp(ind) |
|---|
| 1222 | ENDDO |
|---|
| 1223 | ENDDO |
|---|
| 1224 | ENDDO |
|---|
| 1225 | !$OMP END PARALLEL |
|---|
| 1226 | |
|---|
| 1227 | END SUBROUTINE sort_k_to_sequential |
|---|
| 1228 | |
|---|
| 1229 | |
|---|
| 1230 | !------------------------------------------------------------------------------! |
|---|
| 1231 | ! Description: |
|---|
| 1232 | ! ------------ |
|---|
| 1233 | !> Gather subdomain data from all PEs. |
|---|
| 1234 | !------------------------------------------------------------------------------! |
|---|
| 1235 | SUBROUTINE mg_gather_fast( f2, f2_sub ) |
|---|
| 1236 | |
|---|
| 1237 | USE control_parameters, & |
|---|
| 1238 | ONLY: grid_level |
|---|
| 1239 | |
|---|
| 1240 | USE cpulog, & |
|---|
| 1241 | ONLY: cpu_log, log_point_s |
|---|
| 1242 | |
|---|
| 1243 | USE indices, & |
|---|
| 1244 | ONLY: mg_loc_ind, nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
|---|
| 1245 | |
|---|
| 1246 | IMPLICIT NONE |
|---|
| 1247 | |
|---|
| 1248 | INTEGER(iwp) :: i !< |
|---|
| 1249 | INTEGER(iwp) :: il !< |
|---|
| 1250 | INTEGER(iwp) :: ir !< |
|---|
| 1251 | INTEGER(iwp) :: j !< |
|---|
| 1252 | INTEGER(iwp) :: jn !< |
|---|
| 1253 | INTEGER(iwp) :: js !< |
|---|
| 1254 | INTEGER(iwp) :: k !< |
|---|
| 1255 | INTEGER(iwp) :: nwords !< |
|---|
| 1256 | |
|---|
| 1257 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 1258 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 1259 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f2 !< |
|---|
| 1260 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 1261 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 1262 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f2_l !< |
|---|
| 1263 | |
|---|
| 1264 | REAL(wp), DIMENSION(nzb:mg_loc_ind(5,myid)+1, & |
|---|
| 1265 | mg_loc_ind(3,myid)-1:mg_loc_ind(4,myid)+1, & |
|---|
| 1266 | mg_loc_ind(1,myid)-1:mg_loc_ind(2,myid)+1) :: f2_sub !< |
|---|
| 1267 | |
|---|
| 1268 | |
|---|
| 1269 | #if defined( __parallel ) |
|---|
| 1270 | CALL cpu_log( log_point_s(34), 'mg_gather', 'start' ) |
|---|
| 1271 | |
|---|
| 1272 | f2_l = 0.0_wp |
|---|
| 1273 | |
|---|
| 1274 | ! |
|---|
| 1275 | !-- Store the local subdomain array on the total array |
|---|
| 1276 | js = mg_loc_ind(3,myid) |
|---|
| 1277 | IF ( south_border_pe ) js = js - 1 |
|---|
| 1278 | jn = mg_loc_ind(4,myid) |
|---|
| 1279 | IF ( north_border_pe ) jn = jn + 1 |
|---|
| 1280 | il = mg_loc_ind(1,myid) |
|---|
| 1281 | IF ( left_border_pe ) il = il - 1 |
|---|
| 1282 | ir = mg_loc_ind(2,myid) |
|---|
| 1283 | IF ( right_border_pe ) ir = ir + 1 |
|---|
| 1284 | DO i = il, ir |
|---|
| 1285 | DO j = js, jn |
|---|
| 1286 | DO k = nzb, nzt_mg(grid_level)+1 |
|---|
| 1287 | f2_l(k,j,i) = f2_sub(k,j,i) |
|---|
| 1288 | ENDDO |
|---|
| 1289 | ENDDO |
|---|
| 1290 | ENDDO |
|---|
| 1291 | |
|---|
| 1292 | ! |
|---|
| 1293 | !-- Find out the number of array elements of the total array |
|---|
| 1294 | nwords = SIZE( f2 ) |
|---|
| 1295 | |
|---|
| 1296 | ! |
|---|
| 1297 | !-- Gather subdomain data from all PEs |
|---|
| 1298 | IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr ) |
|---|
| 1299 | CALL MPI_ALLREDUCE( f2_l(nzb,nys_mg(grid_level)-1,nxl_mg(grid_level)-1), & |
|---|
| 1300 | f2(nzb,nys_mg(grid_level)-1,nxl_mg(grid_level)-1), & |
|---|
| 1301 | nwords, MPI_REAL, MPI_SUM, comm2d, ierr ) |
|---|
| 1302 | |
|---|
| 1303 | CALL cpu_log( log_point_s(34), 'mg_gather', 'stop' ) |
|---|
| 1304 | #endif |
|---|
| 1305 | |
|---|
| 1306 | END SUBROUTINE mg_gather_fast |
|---|
| 1307 | |
|---|
| 1308 | |
|---|
| 1309 | |
|---|
| 1310 | !------------------------------------------------------------------------------! |
|---|
| 1311 | ! Description: |
|---|
| 1312 | ! ------------ |
|---|
| 1313 | !> @todo It might be possible to improve the speed of this routine by using |
|---|
| 1314 | !> non-blocking communication |
|---|
| 1315 | !------------------------------------------------------------------------------! |
|---|
| 1316 | SUBROUTINE mg_scatter_fast( p2, p2_sub ) |
|---|
| 1317 | |
|---|
| 1318 | USE control_parameters, & |
|---|
| 1319 | ONLY: grid_level |
|---|
| 1320 | |
|---|
| 1321 | USE cpulog, & |
|---|
| 1322 | ONLY: cpu_log, log_point_s |
|---|
| 1323 | |
|---|
| 1324 | USE indices, & |
|---|
| 1325 | ONLY: mg_loc_ind, nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
|---|
| 1326 | |
|---|
| 1327 | IMPLICIT NONE |
|---|
| 1328 | |
|---|
| 1329 | INTEGER(iwp) :: nwords !< |
|---|
| 1330 | |
|---|
| 1331 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, & |
|---|
| 1332 | nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, & |
|---|
| 1333 | nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1) :: p2 !< |
|---|
| 1334 | |
|---|
| 1335 | REAL(wp), DIMENSION(nzb:mg_loc_ind(5,myid)+1, & |
|---|
| 1336 | mg_loc_ind(3,myid)-1:mg_loc_ind(4,myid)+1, & |
|---|
| 1337 | mg_loc_ind(1,myid)-1:mg_loc_ind(2,myid)+1) :: p2_sub !< |
|---|
| 1338 | |
|---|
| 1339 | ! |
|---|
| 1340 | !-- Find out the number of array elements of the subdomain array |
|---|
| 1341 | nwords = SIZE( p2_sub ) |
|---|
| 1342 | |
|---|
| 1343 | #if defined( __parallel ) |
|---|
| 1344 | CALL cpu_log( log_point_s(35), 'mg_scatter', 'start' ) |
|---|
| 1345 | |
|---|
| 1346 | p2_sub = p2(:,mg_loc_ind(3,myid)-1:mg_loc_ind(4,myid)+1, & |
|---|
| 1347 | mg_loc_ind(1,myid)-1:mg_loc_ind(2,myid)+1) |
|---|
| 1348 | |
|---|
| 1349 | CALL cpu_log( log_point_s(35), 'mg_scatter', 'stop' ) |
|---|
| 1350 | #endif |
|---|
| 1351 | |
|---|
| 1352 | END SUBROUTINE mg_scatter_fast |
|---|
| 1353 | |
|---|
| 1354 | |
|---|
| 1355 | !------------------------------------------------------------------------------! |
|---|
| 1356 | ! Description: |
|---|
| 1357 | ! ------------ |
|---|
| 1358 | !> This is where the multigrid technique takes place. V- and W- Cycle are |
|---|
| 1359 | !> implemented and steered by the parameter "gamma". Parameter "nue" determines |
|---|
| 1360 | !> the convergence of the multigrid iterative solution. There are nue times |
|---|
| 1361 | !> RB-GS iterations. It should be set to "1" or "2", considering the time effort |
|---|
| 1362 | !> one would like to invest. Last choice shows a very good converging factor, |
|---|
| 1363 | !> but leads to an increase in computing time. |
|---|
| 1364 | !------------------------------------------------------------------------------! |
|---|
| 1365 | RECURSIVE SUBROUTINE next_mg_level_fast( f_mg, p_mg, p3, r ) |
|---|
| 1366 | |
|---|
| 1367 | USE control_parameters, & |
|---|
| 1368 | ONLY: bc_lr_dirrad, bc_lr_raddir, bc_ns_dirrad, bc_ns_raddir, & |
|---|
| 1369 | gamma_mg, grid_level, grid_level_count, ibc_p_b, ibc_p_t, & |
|---|
| 1370 | inflow_l, inflow_n, inflow_r, inflow_s, maximum_grid_level, & |
|---|
| 1371 | mg_switch_to_pe0_level, mg_switch_to_pe0, nest_domain, & |
|---|
| 1372 | nest_bound_l, nest_bound_n, nest_bound_r, nest_bound_s, & |
|---|
| 1373 | ngsrb, outflow_l, outflow_n, outflow_r, outflow_s |
|---|
| 1374 | |
|---|
| 1375 | USE indices, & |
|---|
| 1376 | ONLY: mg_loc_ind, nxl, nxl_mg, nxr, nxr_mg, nys, nys_mg, nyn, & |
|---|
| 1377 | nyn_mg, nzb, nzt, nzt_mg |
|---|
| 1378 | |
|---|
| 1379 | IMPLICIT NONE |
|---|
| 1380 | |
|---|
| 1381 | INTEGER(iwp) :: i !< index variable along x |
|---|
| 1382 | INTEGER(iwp) :: j !< index variable along y |
|---|
| 1383 | INTEGER(iwp) :: k !< index variable along z |
|---|
| 1384 | INTEGER(iwp) :: nxl_mg_save !< |
|---|
| 1385 | INTEGER(iwp) :: nxr_mg_save !< |
|---|
| 1386 | INTEGER(iwp) :: nyn_mg_save !< |
|---|
| 1387 | INTEGER(iwp) :: nys_mg_save !< |
|---|
| 1388 | INTEGER(iwp) :: nzt_mg_save !< |
|---|
| 1389 | |
|---|
| 1390 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 1391 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 1392 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: f_mg !< |
|---|
| 1393 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 1394 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 1395 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p_mg !< |
|---|
| 1396 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 1397 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 1398 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: p3 !< |
|---|
| 1399 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 1400 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 1401 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: r !< |
|---|
| 1402 | |
|---|
| 1403 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, & |
|---|
| 1404 | nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, & |
|---|
| 1405 | nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1) :: f2 !< |
|---|
| 1406 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level-1)+1, & |
|---|
| 1407 | nys_mg(grid_level-1)-1:nyn_mg(grid_level-1)+1, & |
|---|
| 1408 | nxl_mg(grid_level-1)-1:nxr_mg(grid_level-1)+1) :: p2 !< |
|---|
| 1409 | |
|---|
| 1410 | REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: f2_sub !< |
|---|
| 1411 | REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: p2_sub !< |
|---|
| 1412 | |
|---|
| 1413 | ! |
|---|
| 1414 | !-- Restriction to the coarsest grid |
|---|
| 1415 | 10 IF ( grid_level == 1 ) THEN |
|---|
| 1416 | |
|---|
| 1417 | ! |
|---|
| 1418 | !-- Solution on the coarsest grid. Double the number of Gauss-Seidel |
|---|
| 1419 | !-- iterations in order to get a more accurate solution. |
|---|
| 1420 | ngsrb = 2 * ngsrb |
|---|
| 1421 | |
|---|
| 1422 | ind_even_odd = even_odd_level(grid_level) |
|---|
| 1423 | |
|---|
| 1424 | CALL redblack_fast( f_mg, p_mg ) |
|---|
| 1425 | |
|---|
| 1426 | ngsrb = ngsrb / 2 |
|---|
| 1427 | |
|---|
| 1428 | |
|---|
| 1429 | ELSEIF ( grid_level /= 1 ) THEN |
|---|
| 1430 | |
|---|
| 1431 | grid_level_count(grid_level) = grid_level_count(grid_level) + 1 |
|---|
| 1432 | |
|---|
| 1433 | ! |
|---|
| 1434 | !-- Solution on the actual grid level |
|---|
| 1435 | ind_even_odd = even_odd_level(grid_level) |
|---|
| 1436 | |
|---|
| 1437 | CALL redblack_fast( f_mg, p_mg ) |
|---|
| 1438 | |
|---|
| 1439 | ! |
|---|
| 1440 | !-- Determination of the actual residual |
|---|
| 1441 | CALL resid_fast( f_mg, p_mg, r ) |
|---|
| 1442 | |
|---|
| 1443 | !-- Restriction of the residual (finer grid values!) to the next coarser |
|---|
| 1444 | !-- grid. Therefore, the grid level has to be decremented now. nxl..nzt have |
|---|
| 1445 | !-- to be set to the coarse grid values, because these variables are needed |
|---|
| 1446 | !-- for the exchange of ghost points in routine exchange_horiz |
|---|
| 1447 | grid_level = grid_level - 1 |
|---|
| 1448 | |
|---|
| 1449 | nxl = nxl_mg(grid_level) |
|---|
| 1450 | nys = nys_mg(grid_level) |
|---|
| 1451 | nxr = nxr_mg(grid_level) |
|---|
| 1452 | nyn = nyn_mg(grid_level) |
|---|
| 1453 | nzt = nzt_mg(grid_level) |
|---|
| 1454 | |
|---|
| 1455 | IF ( grid_level == mg_switch_to_pe0_level ) THEN |
|---|
| 1456 | |
|---|
| 1457 | ! |
|---|
| 1458 | !-- From this level on, calculations are done on PE0 only. |
|---|
| 1459 | !-- First, carry out restriction on the subdomain. |
|---|
| 1460 | !-- Therefore, indices of the level have to be changed to subdomain |
|---|
| 1461 | !-- values in between (otherwise, the restrict routine would expect |
|---|
| 1462 | !-- the gathered array) |
|---|
| 1463 | |
|---|
| 1464 | nxl_mg_save = nxl_mg(grid_level) |
|---|
| 1465 | nxr_mg_save = nxr_mg(grid_level) |
|---|
| 1466 | nys_mg_save = nys_mg(grid_level) |
|---|
| 1467 | nyn_mg_save = nyn_mg(grid_level) |
|---|
| 1468 | nzt_mg_save = nzt_mg(grid_level) |
|---|
| 1469 | nxl_mg(grid_level) = mg_loc_ind(1,myid) |
|---|
| 1470 | nxr_mg(grid_level) = mg_loc_ind(2,myid) |
|---|
| 1471 | nys_mg(grid_level) = mg_loc_ind(3,myid) |
|---|
| 1472 | nyn_mg(grid_level) = mg_loc_ind(4,myid) |
|---|
| 1473 | nzt_mg(grid_level) = mg_loc_ind(5,myid) |
|---|
| 1474 | nxl = mg_loc_ind(1,myid) |
|---|
| 1475 | nxr = mg_loc_ind(2,myid) |
|---|
| 1476 | nys = mg_loc_ind(3,myid) |
|---|
| 1477 | nyn = mg_loc_ind(4,myid) |
|---|
| 1478 | nzt = mg_loc_ind(5,myid) |
|---|
| 1479 | |
|---|
| 1480 | ALLOCATE( f2_sub(nzb:nzt_mg(grid_level)+1, & |
|---|
| 1481 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 1482 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) ) |
|---|
| 1483 | |
|---|
| 1484 | CALL restrict_fast( f2_sub, r ) |
|---|
| 1485 | |
|---|
| 1486 | ! |
|---|
| 1487 | !-- Restore the correct indices of this level |
|---|
| 1488 | nxl_mg(grid_level) = nxl_mg_save |
|---|
| 1489 | nxr_mg(grid_level) = nxr_mg_save |
|---|
| 1490 | nys_mg(grid_level) = nys_mg_save |
|---|
| 1491 | nyn_mg(grid_level) = nyn_mg_save |
|---|
| 1492 | nzt_mg(grid_level) = nzt_mg_save |
|---|
| 1493 | nxl = nxl_mg(grid_level) |
|---|
| 1494 | nxr = nxr_mg(grid_level) |
|---|
| 1495 | nys = nys_mg(grid_level) |
|---|
| 1496 | nyn = nyn_mg(grid_level) |
|---|
| 1497 | nzt = nzt_mg(grid_level) |
|---|
| 1498 | ! |
|---|
| 1499 | !-- Gather all arrays from the subdomains on PE0 |
|---|
| 1500 | CALL mg_gather_fast( f2, f2_sub ) |
|---|
| 1501 | |
|---|
| 1502 | ! |
|---|
| 1503 | !-- Set switch for routine exchange_horiz, that no ghostpoint exchange |
|---|
| 1504 | !-- has to be carried out from now on |
|---|
| 1505 | mg_switch_to_pe0 = .TRUE. |
|---|
| 1506 | |
|---|
| 1507 | ! |
|---|
| 1508 | !-- In case of non-cyclic lateral boundary conditions, both in- and |
|---|
| 1509 | !-- outflow conditions have to be used on all PEs after the switch, |
|---|
| 1510 | !-- because then they have the total domain. |
|---|
| 1511 | IF ( bc_lr_dirrad ) THEN |
|---|
| 1512 | inflow_l = .TRUE. |
|---|
| 1513 | inflow_r = .FALSE. |
|---|
| 1514 | outflow_l = .FALSE. |
|---|
| 1515 | outflow_r = .TRUE. |
|---|
| 1516 | ELSEIF ( bc_lr_raddir ) THEN |
|---|
| 1517 | inflow_l = .FALSE. |
|---|
| 1518 | inflow_r = .TRUE. |
|---|
| 1519 | outflow_l = .TRUE. |
|---|
| 1520 | outflow_r = .FALSE. |
|---|
| 1521 | ELSEIF ( nest_domain ) THEN |
|---|
| 1522 | nest_bound_l = .TRUE. |
|---|
| 1523 | nest_bound_r = .TRUE. |
|---|
| 1524 | ENDIF |
|---|
| 1525 | |
|---|
| 1526 | IF ( bc_ns_dirrad ) THEN |
|---|
| 1527 | inflow_n = .TRUE. |
|---|
| 1528 | inflow_s = .FALSE. |
|---|
| 1529 | outflow_n = .FALSE. |
|---|
| 1530 | outflow_s = .TRUE. |
|---|
| 1531 | ELSEIF ( bc_ns_raddir ) THEN |
|---|
| 1532 | inflow_n = .FALSE. |
|---|
| 1533 | inflow_s = .TRUE. |
|---|
| 1534 | outflow_n = .TRUE. |
|---|
| 1535 | outflow_s = .FALSE. |
|---|
| 1536 | ELSEIF ( nest_domain ) THEN |
|---|
| 1537 | nest_bound_s = .TRUE. |
|---|
| 1538 | nest_bound_n = .TRUE. |
|---|
| 1539 | ENDIF |
|---|
| 1540 | |
|---|
| 1541 | DEALLOCATE( f2_sub ) |
|---|
| 1542 | |
|---|
| 1543 | ELSE |
|---|
| 1544 | |
|---|
| 1545 | CALL restrict_fast( f2, r ) |
|---|
| 1546 | |
|---|
| 1547 | ind_even_odd = even_odd_level(grid_level) ! must be after restrict |
|---|
| 1548 | |
|---|
| 1549 | ENDIF |
|---|
| 1550 | |
|---|
| 1551 | p2 = 0.0_wp |
|---|
| 1552 | |
|---|
| 1553 | ! |
|---|
| 1554 | !-- Repeat the same procedure till the coarsest grid is reached |
|---|
| 1555 | CALL next_mg_level_fast( f2, p2, p3, r ) |
|---|
| 1556 | |
|---|
| 1557 | ENDIF |
|---|
| 1558 | |
|---|
| 1559 | ! |
|---|
| 1560 | !-- Now follows the prolongation |
|---|
| 1561 | IF ( grid_level >= 2 ) THEN |
|---|
| 1562 | |
|---|
| 1563 | ! |
|---|
| 1564 | !-- Prolongation of the new residual. The values are transferred |
|---|
| 1565 | !-- from the coarse to the next finer grid. |
|---|
| 1566 | IF ( grid_level == mg_switch_to_pe0_level+1 ) THEN |
|---|
| 1567 | |
|---|
| 1568 | #if defined( __parallel ) |
|---|
| 1569 | ! |
|---|
| 1570 | !-- At this level, the new residual first has to be scattered from |
|---|
| 1571 | !-- PE0 to the other PEs |
|---|
| 1572 | ALLOCATE( p2_sub(nzb:mg_loc_ind(5,myid)+1, & |
|---|
| 1573 | mg_loc_ind(3,myid)-1:mg_loc_ind(4,myid)+1, & |
|---|
| 1574 | mg_loc_ind(1,myid)-1:mg_loc_ind(2,myid)+1) ) |
|---|
| 1575 | |
|---|
| 1576 | CALL mg_scatter_fast( p2, p2_sub ) |
|---|
| 1577 | |
|---|
| 1578 | ! |
|---|
| 1579 | !-- Therefore, indices of the previous level have to be changed to |
|---|
| 1580 | !-- subdomain values in between (otherwise, the prolong routine would |
|---|
| 1581 | !-- expect the gathered array) |
|---|
| 1582 | nxl_mg_save = nxl_mg(grid_level-1) |
|---|
| 1583 | nxr_mg_save = nxr_mg(grid_level-1) |
|---|
| 1584 | nys_mg_save = nys_mg(grid_level-1) |
|---|
| 1585 | nyn_mg_save = nyn_mg(grid_level-1) |
|---|
| 1586 | nzt_mg_save = nzt_mg(grid_level-1) |
|---|
| 1587 | nxl_mg(grid_level-1) = mg_loc_ind(1,myid) |
|---|
| 1588 | nxr_mg(grid_level-1) = mg_loc_ind(2,myid) |
|---|
| 1589 | nys_mg(grid_level-1) = mg_loc_ind(3,myid) |
|---|
| 1590 | nyn_mg(grid_level-1) = mg_loc_ind(4,myid) |
|---|
| 1591 | nzt_mg(grid_level-1) = mg_loc_ind(5,myid) |
|---|
| 1592 | |
|---|
| 1593 | ! |
|---|
| 1594 | !-- Set switch for routine exchange_horiz, that ghostpoint exchange |
|---|
| 1595 | !-- has to be carried again out from now on |
|---|
| 1596 | mg_switch_to_pe0 = .FALSE. |
|---|
| 1597 | |
|---|
| 1598 | ! |
|---|
| 1599 | !-- For non-cyclic lateral boundary conditions, restore the |
|---|
| 1600 | !-- in-/outflow conditions |
|---|
| 1601 | inflow_l = .FALSE.; inflow_r = .FALSE. |
|---|
| 1602 | inflow_n = .FALSE.; inflow_s = .FALSE. |
|---|
| 1603 | outflow_l = .FALSE.; outflow_r = .FALSE. |
|---|
| 1604 | outflow_n = .FALSE.; outflow_s = .FALSE. |
|---|
| 1605 | |
|---|
| 1606 | IF ( pleft == MPI_PROC_NULL ) THEN |
|---|
| 1607 | IF ( bc_lr_dirrad ) THEN |
|---|
| 1608 | inflow_l = .TRUE. |
|---|
| 1609 | ELSEIF ( bc_lr_raddir ) THEN |
|---|
| 1610 | outflow_l = .TRUE. |
|---|
| 1611 | ELSEIF ( nest_domain ) THEN |
|---|
| 1612 | nest_bound_l = .TRUE. |
|---|
| 1613 | ENDIF |
|---|
| 1614 | ENDIF |
|---|
| 1615 | |
|---|
| 1616 | IF ( pright == MPI_PROC_NULL ) THEN |
|---|
| 1617 | IF ( bc_lr_dirrad ) THEN |
|---|
| 1618 | outflow_r = .TRUE. |
|---|
| 1619 | ELSEIF ( bc_lr_raddir ) THEN |
|---|
| 1620 | inflow_r = .TRUE. |
|---|
| 1621 | ELSEIF ( nest_domain ) THEN |
|---|
| 1622 | nest_bound_r = .TRUE. |
|---|
| 1623 | ENDIF |
|---|
| 1624 | ENDIF |
|---|
| 1625 | |
|---|
| 1626 | IF ( psouth == MPI_PROC_NULL ) THEN |
|---|
| 1627 | IF ( bc_ns_dirrad ) THEN |
|---|
| 1628 | outflow_s = .TRUE. |
|---|
| 1629 | ELSEIF ( bc_ns_raddir ) THEN |
|---|
| 1630 | inflow_s = .TRUE. |
|---|
| 1631 | ELSEIF ( nest_domain ) THEN |
|---|
| 1632 | nest_bound_s = .TRUE. |
|---|
| 1633 | ENDIF |
|---|
| 1634 | ENDIF |
|---|
| 1635 | |
|---|
| 1636 | IF ( pnorth == MPI_PROC_NULL ) THEN |
|---|
| 1637 | IF ( bc_ns_dirrad ) THEN |
|---|
| 1638 | inflow_n = .TRUE. |
|---|
| 1639 | ELSEIF ( bc_ns_raddir ) THEN |
|---|
| 1640 | outflow_n = .TRUE. |
|---|
| 1641 | ELSEIF ( nest_domain ) THEN |
|---|
| 1642 | nest_bound_n = .TRUE. |
|---|
| 1643 | ENDIF |
|---|
| 1644 | ENDIF |
|---|
| 1645 | |
|---|
| 1646 | CALL prolong_fast( p2_sub, p3 ) |
|---|
| 1647 | |
|---|
| 1648 | ! |
|---|
| 1649 | !-- Restore the correct indices of the previous level |
|---|
| 1650 | nxl_mg(grid_level-1) = nxl_mg_save |
|---|
| 1651 | nxr_mg(grid_level-1) = nxr_mg_save |
|---|
| 1652 | nys_mg(grid_level-1) = nys_mg_save |
|---|
| 1653 | nyn_mg(grid_level-1) = nyn_mg_save |
|---|
| 1654 | nzt_mg(grid_level-1) = nzt_mg_save |
|---|
| 1655 | |
|---|
| 1656 | DEALLOCATE( p2_sub ) |
|---|
| 1657 | #endif |
|---|
| 1658 | |
|---|
| 1659 | ELSE |
|---|
| 1660 | |
|---|
| 1661 | CALL prolong_fast( p2, p3 ) |
|---|
| 1662 | |
|---|
| 1663 | ENDIF |
|---|
| 1664 | |
|---|
| 1665 | ! |
|---|
| 1666 | !-- Computation of the new pressure correction. Therefore, |
|---|
| 1667 | !-- values from prior grids are added up automatically stage by stage. |
|---|
| 1668 | DO i = nxl_mg(grid_level)-1, nxr_mg(grid_level)+1 |
|---|
| 1669 | DO j = nys_mg(grid_level)-1, nyn_mg(grid_level)+1 |
|---|
| 1670 | DO k = nzb, nzt_mg(grid_level)+1 |
|---|
| 1671 | p_mg(k,j,i) = p_mg(k,j,i) + p3(k,j,i) |
|---|
| 1672 | ENDDO |
|---|
| 1673 | ENDDO |
|---|
| 1674 | ENDDO |
|---|
| 1675 | |
|---|
| 1676 | ! |
|---|
| 1677 | !-- Relaxation of the new solution |
|---|
| 1678 | CALL redblack_fast( f_mg, p_mg ) |
|---|
| 1679 | |
|---|
| 1680 | ENDIF |
|---|
| 1681 | |
|---|
| 1682 | |
|---|
| 1683 | ! |
|---|
| 1684 | !-- The following few lines serve the steering of the multigrid scheme |
|---|
| 1685 | IF ( grid_level == maximum_grid_level ) THEN |
|---|
| 1686 | |
|---|
| 1687 | GOTO 20 |
|---|
| 1688 | |
|---|
| 1689 | ELSEIF ( grid_level /= maximum_grid_level .AND. grid_level /= 1 .AND. & |
|---|
| 1690 | grid_level_count(grid_level) /= gamma_mg ) THEN |
|---|
| 1691 | |
|---|
| 1692 | GOTO 10 |
|---|
| 1693 | |
|---|
| 1694 | ENDIF |
|---|
| 1695 | |
|---|
| 1696 | ! |
|---|
| 1697 | !-- Reset counter for the next call of poismg_fast |
|---|
| 1698 | grid_level_count(grid_level) = 0 |
|---|
| 1699 | |
|---|
| 1700 | ! |
|---|
| 1701 | !-- Continue with the next finer level. nxl..nzt have to be |
|---|
| 1702 | !-- set to the finer grid values, because these variables are needed for the |
|---|
| 1703 | !-- exchange of ghost points in routine exchange_horiz |
|---|
| 1704 | grid_level = grid_level + 1 |
|---|
| 1705 | ind_even_odd = even_odd_level(grid_level) |
|---|
| 1706 | |
|---|
| 1707 | nxl = nxl_mg(grid_level) |
|---|
| 1708 | nxr = nxr_mg(grid_level) |
|---|
| 1709 | nys = nys_mg(grid_level) |
|---|
| 1710 | nyn = nyn_mg(grid_level) |
|---|
| 1711 | nzt = nzt_mg(grid_level) |
|---|
| 1712 | |
|---|
| 1713 | 20 CONTINUE |
|---|
| 1714 | |
|---|
| 1715 | END SUBROUTINE next_mg_level_fast |
|---|
| 1716 | |
|---|
| 1717 | |
|---|
| 1718 | !------------------------------------------------------------------------------! |
|---|
| 1719 | ! Description: |
|---|
| 1720 | ! ------------ |
|---|
| 1721 | !> Initial settings for sorting k-dimension from sequential order (alternate |
|---|
| 1722 | !> even/odd) into blocks of even and odd or vice versa |
|---|
| 1723 | !------------------------------------------------------------------------------! |
|---|
| 1724 | SUBROUTINE init_even_odd_blocks |
|---|
| 1725 | |
|---|
| 1726 | |
|---|
| 1727 | USE arrays_3d, & |
|---|
| 1728 | ONLY: f1_mg, f2_mg, f3_mg |
|---|
| 1729 | |
|---|
| 1730 | USE control_parameters, & |
|---|
| 1731 | ONLY: grid_level, maximum_grid_level |
|---|
| 1732 | |
|---|
| 1733 | USE indices, & |
|---|
| 1734 | ONLY: nzb, nzt, nzt_mg |
|---|
| 1735 | |
|---|
| 1736 | USE indices, & |
|---|
| 1737 | ONLY: nxl_mg, nxr_mg, nys_mg, nyn_mg, nzb, nzt_mg |
|---|
| 1738 | |
|---|
| 1739 | IMPLICIT NONE |
|---|
| 1740 | ! |
|---|
| 1741 | !-- Local variables |
|---|
| 1742 | INTEGER(iwp) :: i !< |
|---|
| 1743 | INTEGER(iwp) :: l !< |
|---|
| 1744 | |
|---|
| 1745 | LOGICAL, SAVE :: lfirst = .TRUE. |
|---|
| 1746 | |
|---|
| 1747 | |
|---|
| 1748 | IF ( .NOT. lfirst ) RETURN |
|---|
| 1749 | |
|---|
| 1750 | ALLOCATE( even_odd_level(maximum_grid_level) ) |
|---|
| 1751 | |
|---|
| 1752 | ALLOCATE( f1_mg_b(nzb:nzt+1,maximum_grid_level), & |
|---|
| 1753 | f2_mg_b(nzb:nzt+1,maximum_grid_level), & |
|---|
| 1754 | f3_mg_b(nzb:nzt+1,maximum_grid_level) ) |
|---|
| 1755 | |
|---|
| 1756 | ! |
|---|
| 1757 | !-- Set border index between the even and odd block |
|---|
| 1758 | DO i = maximum_grid_level, 1, -1 |
|---|
| 1759 | even_odd_level(i) = nzt_mg(i) / 2 |
|---|
| 1760 | ENDDO |
|---|
| 1761 | |
|---|
| 1762 | ! |
|---|
| 1763 | !-- Sort grid coefficients used in red/black scheme and for calculating the |
|---|
| 1764 | !-- residual to block (even/odd) structure |
|---|
| 1765 | DO l = maximum_grid_level, 1 , -1 |
|---|
| 1766 | CALL sort_k_to_even_odd_blocks( f1_mg(nzb+1:nzt_mg(grid_level),l), & |
|---|
| 1767 | f1_mg_b(nzb:nzt_mg(grid_level)+1,l), & |
|---|
| 1768 | l ) |
|---|
| 1769 | CALL sort_k_to_even_odd_blocks( f2_mg(nzb+1:nzt_mg(grid_level),l), & |
|---|
| 1770 | f2_mg_b(nzb:nzt_mg(grid_level)+1,l), & |
|---|
| 1771 | l ) |
|---|
| 1772 | CALL sort_k_to_even_odd_blocks( f3_mg(nzb+1:nzt_mg(grid_level),l), & |
|---|
| 1773 | f3_mg_b(nzb:nzt_mg(grid_level)+1,l), & |
|---|
| 1774 | l ) |
|---|
| 1775 | ENDDO |
|---|
| 1776 | |
|---|
| 1777 | lfirst = .FALSE. |
|---|
| 1778 | |
|---|
| 1779 | END SUBROUTINE init_even_odd_blocks |
|---|
| 1780 | |
|---|
| 1781 | |
|---|
| 1782 | !------------------------------------------------------------------------------! |
|---|
| 1783 | ! Description: |
|---|
| 1784 | ! ------------ |
|---|
| 1785 | !> Special exchange_horiz subroutine for use in redblack. Transfers only |
|---|
| 1786 | !> "red" or "black" data points. |
|---|
| 1787 | !------------------------------------------------------------------------------! |
|---|
| 1788 | SUBROUTINE special_exchange_horiz ( p_mg, color ) |
|---|
| 1789 | |
|---|
| 1790 | |
|---|
| 1791 | USE control_parameters, & |
|---|
| 1792 | ONLY: bc_lr_cyc, bc_ns_cyc, grid_level, ibc_p_b, ibc_p_t, & |
|---|
| 1793 | inflow_l, inflow_n, inflow_r, inflow_s, maximum_grid_level, & |
|---|
| 1794 | mg_switch_to_pe0_level, outflow_l, outflow_n, outflow_r, & |
|---|
| 1795 | outflow_s, synchronous_exchange |
|---|
| 1796 | |
|---|
| 1797 | USE indices, & |
|---|
| 1798 | ONLY: mg_loc_ind, nxl, nxl_mg, nxr, nxr_mg, nys, nys_mg, nyn, & |
|---|
| 1799 | nyn_mg, nzb, nzt, nzt_mg |
|---|
| 1800 | |
|---|
| 1801 | IMPLICIT NONE |
|---|
| 1802 | |
|---|
| 1803 | REAL(wp), DIMENSION(nzb:nzt_mg(grid_level)+1, & |
|---|
| 1804 | nys_mg(grid_level)-1:nyn_mg(grid_level)+1, & |
|---|
| 1805 | nxl_mg(grid_level)-1:nxr_mg(grid_level)+1) :: & |
|---|
| 1806 | p_mg !< treated array |
|---|
| 1807 | |
|---|
| 1808 | INTEGER(iwp), intent(IN) :: color !< flag for grid point type (red or black) |
|---|
| 1809 | ! |
|---|
| 1810 | !-- Local variables |
|---|
| 1811 | INTEGER(iwp) :: i !< index variable along x |
|---|
| 1812 | INTEGER(iwp) :: i1 !< index variable along x on coarse level |
|---|
| 1813 | INTEGER(iwp) :: i2 !< index variable along x on coarse level |
|---|
| 1814 | |
|---|
| 1815 | INTEGER(iwp) :: j !< index variable along y |
|---|
| 1816 | INTEGER(iwp) :: j1 !< index variable along y on coarse level |
|---|
| 1817 | INTEGER(iwp) :: j2 !< index variable along y on coarse level |
|---|
| 1818 | INTEGER(iwp) :: k !< index variable along z |
|---|
| 1819 | INTEGER(iwp) :: l !< short for grid level |
|---|
| 1820 | INTEGER(iwp) :: jys !< index for lower local PE boundary along y |
|---|
| 1821 | INTEGER(iwp) :: jyn !< index for upper local PE boundary along y |
|---|
| 1822 | INTEGER(iwp) :: ixl !< index for lower local PE boundary along x |
|---|
| 1823 | INTEGER(iwp) :: ixr !< index for upper local PE boundary along x |
|---|
| 1824 | |
|---|
| 1825 | LOGICAL :: sendrecv_in_background_save !< dummy to reset sendrecv_in_background to prescribed value |
|---|
| 1826 | LOGICAL :: synchronous_exchange_save !< dummy to reset synchronous_exchange to prescribed value |
|---|
| 1827 | |
|---|
| 1828 | REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: temp !< temporary array on next coarser grid level |
|---|
| 1829 | |
|---|
| 1830 | #if defined ( __parallel ) |
|---|
| 1831 | sendrecv_in_background_save = sendrecv_in_background |
|---|
| 1832 | sendrecv_in_background = .FALSE. |
|---|
| 1833 | synchronous_exchange_save = synchronous_exchange |
|---|
| 1834 | synchronous_exchange = .FALSE. |
|---|
| 1835 | |
|---|
| 1836 | l = grid_level |
|---|
| 1837 | |
|---|
| 1838 | ind_even_odd = even_odd_level(grid_level) |
|---|
| 1839 | |
|---|
| 1840 | ! |
|---|
| 1841 | !-- Restricted transfer only on finer levels with enough data. |
|---|
| 1842 | !-- Restricted transfer is not possible for levels smaller or equal to |
|---|
| 1843 | !-- 'switch to PE0 levels', since array bounds does not fit. Moreover, |
|---|
| 1844 | !-- it is not possible for the coarsest grid level, since the dimensions |
|---|
| 1845 | !-- of temp are not defined. For such cases, normal exchange_horiz is called. |
|---|
| 1846 | IF ( l > 1 .AND. l > mg_switch_to_pe0_level + 1 .AND. & |
|---|
| 1847 | ( ngp_xz(grid_level) >= 900 .OR. ngp_yz(grid_level) >= 900 ) ) THEN |
|---|
| 1848 | |
|---|
| 1849 | jys = nys_mg(grid_level-1) |
|---|
| 1850 | jyn = nyn_mg(grid_level-1) |
|---|
| 1851 | ixl = nxl_mg(grid_level-1) |
|---|
| 1852 | ixr = nxr_mg(grid_level-1) |
|---|
| 1853 | ALLOCATE( temp(nzb:nzt_mg(l-1)+1,jys-1:jyn+1,ixl-1:ixr+1) ) |
|---|
| 1854 | ! |
|---|
| 1855 | !-- Handling the even k Values |
|---|
| 1856 | !-- Collecting data for the north - south exchange |
|---|
| 1857 | !-- Since only every second value has to be transfered, data are stored |
|---|
| 1858 | !-- on the next coarser grid level, because the arrays on that level |
|---|
| 1859 | !-- have just the required size |
|---|
| 1860 | i1 = nxl_mg(grid_level-1) |
|---|
| 1861 | i2 = nxl_mg(grid_level-1) |
|---|
| 1862 | |
|---|
| 1863 | DO i = nxl_mg(l), nxr_mg(l), 2 |
|---|
| 1864 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
|---|
| 1865 | |
|---|
| 1866 | IF ( j == nys_mg(l) ) THEN |
|---|
| 1867 | !DIR$ IVDEP |
|---|
| 1868 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 1869 | temp(k-ind_even_odd,jys,i1) = p_mg(k,j,i) |
|---|
| 1870 | ENDDO |
|---|
| 1871 | i1 = i1 + 1 |
|---|
| 1872 | |
|---|
| 1873 | ENDIF |
|---|
| 1874 | |
|---|
| 1875 | IF ( j == nyn_mg(l) ) THEN |
|---|
| 1876 | !DIR$ IVDEP |
|---|
| 1877 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 1878 | temp(k-ind_even_odd,jyn,i2) = p_mg(k,j,i) |
|---|
| 1879 | ENDDO |
|---|
| 1880 | i2 = i2 + 1 |
|---|
| 1881 | |
|---|
| 1882 | ENDIF |
|---|
| 1883 | |
|---|
| 1884 | ENDDO |
|---|
| 1885 | ENDDO |
|---|
| 1886 | |
|---|
| 1887 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
|---|
| 1888 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
|---|
| 1889 | |
|---|
| 1890 | IF ( j == nys_mg(l) ) THEN |
|---|
| 1891 | !DIR$ IVDEP |
|---|
| 1892 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 1893 | temp(k-ind_even_odd,jys,i1) = p_mg(k,j,i) |
|---|
| 1894 | ENDDO |
|---|
| 1895 | i1 = i1 + 1 |
|---|
| 1896 | |
|---|
| 1897 | ENDIF |
|---|
| 1898 | |
|---|
| 1899 | IF ( j == nyn_mg(l) ) THEN |
|---|
| 1900 | !DIR$ IVDEP |
|---|
| 1901 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 1902 | temp(k-ind_even_odd,jyn,i2) = p_mg(k,j,i) |
|---|
| 1903 | ENDDO |
|---|
| 1904 | i2 = i2 + 1 |
|---|
| 1905 | |
|---|
| 1906 | ENDIF |
|---|
| 1907 | |
|---|
| 1908 | ENDDO |
|---|
| 1909 | ENDDO |
|---|
| 1910 | |
|---|
| 1911 | grid_level = grid_level-1 |
|---|
| 1912 | |
|---|
| 1913 | nxl = nxl_mg(grid_level) |
|---|
| 1914 | nys = nys_mg(grid_level) |
|---|
| 1915 | nxr = nxr_mg(grid_level) |
|---|
| 1916 | nyn = nyn_mg(grid_level) |
|---|
| 1917 | nzt = nzt_mg(grid_level) |
|---|
| 1918 | |
|---|
| 1919 | send_receive = 'ns' |
|---|
| 1920 | CALL exchange_horiz( temp, 1 ) |
|---|
| 1921 | |
|---|
| 1922 | grid_level = grid_level+1 |
|---|
| 1923 | |
|---|
| 1924 | i1 = nxl_mg(grid_level-1) |
|---|
| 1925 | i2 = nxl_mg(grid_level-1) |
|---|
| 1926 | |
|---|
| 1927 | DO i = nxl_mg(l), nxr_mg(l), 2 |
|---|
| 1928 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
|---|
| 1929 | |
|---|
| 1930 | IF ( j == nys_mg(l) ) THEN |
|---|
| 1931 | !DIR$ IVDEP |
|---|
| 1932 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 1933 | p_mg(k,nyn_mg(l)+1,i) = temp(k-ind_even_odd,jyn+1,i1) |
|---|
| 1934 | ENDDO |
|---|
| 1935 | i1 = i1 + 1 |
|---|
| 1936 | |
|---|
| 1937 | ENDIF |
|---|
| 1938 | |
|---|
| 1939 | IF ( j == nyn_mg(l) ) THEN |
|---|
| 1940 | !DIR$ IVDEP |
|---|
| 1941 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 1942 | p_mg(k,nys_mg(l)-1,i) = temp(k-ind_even_odd,jys-1,i2) |
|---|
| 1943 | ENDDO |
|---|
| 1944 | i2 = i2 + 1 |
|---|
| 1945 | |
|---|
| 1946 | ENDIF |
|---|
| 1947 | |
|---|
| 1948 | ENDDO |
|---|
| 1949 | ENDDO |
|---|
| 1950 | |
|---|
| 1951 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
|---|
| 1952 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
|---|
| 1953 | |
|---|
| 1954 | IF ( j == nys_mg(l) ) THEN |
|---|
| 1955 | !DIR$ IVDEP |
|---|
| 1956 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 1957 | p_mg(k,nyn_mg(l)+1,i) = temp(k-ind_even_odd,jyn+1,i1) |
|---|
| 1958 | ENDDO |
|---|
| 1959 | i1 = i1 + 1 |
|---|
| 1960 | |
|---|
| 1961 | ENDIF |
|---|
| 1962 | |
|---|
| 1963 | IF ( j == nyn_mg(l) ) THEN |
|---|
| 1964 | !DIR$ IVDEP |
|---|
| 1965 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 1966 | p_mg(k,nys_mg(l)-1,i) = temp(k-ind_even_odd,jys-1,i2) |
|---|
| 1967 | ENDDO |
|---|
| 1968 | i2 = i2 + 1 |
|---|
| 1969 | |
|---|
| 1970 | ENDIF |
|---|
| 1971 | |
|---|
| 1972 | ENDDO |
|---|
| 1973 | ENDDO |
|---|
| 1974 | |
|---|
| 1975 | ! |
|---|
| 1976 | !-- Collecting data for the left - right exchange |
|---|
| 1977 | !-- Since only every second value has to be transfered, data are stored |
|---|
| 1978 | !-- on the next coarser grid level, because the arrays on that level |
|---|
| 1979 | !-- have just the required size |
|---|
| 1980 | j1 = nys_mg(grid_level-1) |
|---|
| 1981 | j2 = nys_mg(grid_level-1) |
|---|
| 1982 | |
|---|
| 1983 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
|---|
| 1984 | DO i = nxl_mg(l), nxr_mg(l), 2 |
|---|
| 1985 | |
|---|
| 1986 | IF ( i == nxl_mg(l) ) THEN |
|---|
| 1987 | !DIR$ IVDEP |
|---|
| 1988 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 1989 | temp(k-ind_even_odd,j1,ixl) = p_mg(k,j,i) |
|---|
| 1990 | ENDDO |
|---|
| 1991 | j1 = j1 + 1 |
|---|
| 1992 | |
|---|
| 1993 | ENDIF |
|---|
| 1994 | |
|---|
| 1995 | IF ( i == nxr_mg(l) ) THEN |
|---|
| 1996 | !DIR$ IVDEP |
|---|
| 1997 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 1998 | temp(k-ind_even_odd,j2,ixr) = p_mg(k,j,i) |
|---|
| 1999 | ENDDO |
|---|
| 2000 | j2 = j2 + 1 |
|---|
| 2001 | |
|---|
| 2002 | ENDIF |
|---|
| 2003 | |
|---|
| 2004 | ENDDO |
|---|
| 2005 | ENDDO |
|---|
| 2006 | |
|---|
| 2007 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
|---|
| 2008 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
|---|
| 2009 | |
|---|
| 2010 | IF ( i == nxl_mg(l) ) THEN |
|---|
| 2011 | !DIR$ IVDEP |
|---|
| 2012 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 2013 | temp(k-ind_even_odd,j1,ixl) = p_mg(k,j,i) |
|---|
| 2014 | ENDDO |
|---|
| 2015 | j1 = j1 + 1 |
|---|
| 2016 | |
|---|
| 2017 | ENDIF |
|---|
| 2018 | |
|---|
| 2019 | IF ( i == nxr_mg(l) ) THEN |
|---|
| 2020 | !DIR$ IVDEP |
|---|
| 2021 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 2022 | temp(k-ind_even_odd,j2,ixr) = p_mg(k,j,i) |
|---|
| 2023 | ENDDO |
|---|
| 2024 | j2 = j2 + 1 |
|---|
| 2025 | |
|---|
| 2026 | ENDIF |
|---|
| 2027 | |
|---|
| 2028 | ENDDO |
|---|
| 2029 | ENDDO |
|---|
| 2030 | |
|---|
| 2031 | grid_level = grid_level-1 |
|---|
| 2032 | send_receive = 'lr' |
|---|
| 2033 | |
|---|
| 2034 | CALL exchange_horiz( temp, 1 ) |
|---|
| 2035 | |
|---|
| 2036 | grid_level = grid_level+1 |
|---|
| 2037 | |
|---|
| 2038 | j1 = nys_mg(grid_level-1) |
|---|
| 2039 | j2 = nys_mg(grid_level-1) |
|---|
| 2040 | |
|---|
| 2041 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
|---|
| 2042 | DO i = nxl_mg(l), nxr_mg(l), 2 |
|---|
| 2043 | |
|---|
| 2044 | IF ( i == nxl_mg(l) ) THEN |
|---|
| 2045 | !DIR$ IVDEP |
|---|
| 2046 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 2047 | p_mg(k,j,nxr_mg(l)+1) = temp(k-ind_even_odd,j1,ixr+1) |
|---|
| 2048 | ENDDO |
|---|
| 2049 | j1 = j1 + 1 |
|---|
| 2050 | |
|---|
| 2051 | ENDIF |
|---|
| 2052 | |
|---|
| 2053 | IF ( i == nxr_mg(l) ) THEN |
|---|
| 2054 | !DIR$ IVDEP |
|---|
| 2055 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 2056 | p_mg(k,j,nxl_mg(l)-1) = temp(k-ind_even_odd,j2,ixl-1) |
|---|
| 2057 | ENDDO |
|---|
| 2058 | j2 = j2 + 1 |
|---|
| 2059 | |
|---|
| 2060 | ENDIF |
|---|
| 2061 | |
|---|
| 2062 | ENDDO |
|---|
| 2063 | ENDDO |
|---|
| 2064 | |
|---|
| 2065 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
|---|
| 2066 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
|---|
| 2067 | |
|---|
| 2068 | IF ( i == nxl_mg(l) ) THEN |
|---|
| 2069 | !DIR$ IVDEP |
|---|
| 2070 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 2071 | p_mg(k,j,nxr_mg(l)+1) = temp(k-ind_even_odd,j1,ixr+1) |
|---|
| 2072 | ENDDO |
|---|
| 2073 | j1 = j1 + 1 |
|---|
| 2074 | |
|---|
| 2075 | ENDIF |
|---|
| 2076 | |
|---|
| 2077 | IF ( i == nxr_mg(l) ) THEN |
|---|
| 2078 | !DIR$ IVDEP |
|---|
| 2079 | DO k = ind_even_odd+1, nzt_mg(l) |
|---|
| 2080 | p_mg(k,j,nxl_mg(l)-1) = temp(k-ind_even_odd,j2,ixl-1) |
|---|
| 2081 | ENDDO |
|---|
| 2082 | j2 = j2 + 1 |
|---|
| 2083 | |
|---|
| 2084 | ENDIF |
|---|
| 2085 | |
|---|
| 2086 | ENDDO |
|---|
| 2087 | ENDDO |
|---|
| 2088 | |
|---|
| 2089 | ! |
|---|
| 2090 | !-- Now handling the even k values |
|---|
| 2091 | !-- Collecting data for the north - south exchange |
|---|
| 2092 | !-- Since only every second value has to be transfered, data are stored |
|---|
| 2093 | !-- on the next coarser grid level, because the arrays on that level |
|---|
| 2094 | !-- have just the required size |
|---|
| 2095 | i1 = nxl_mg(grid_level-1) |
|---|
| 2096 | i2 = nxl_mg(grid_level-1) |
|---|
| 2097 | |
|---|
| 2098 | DO i = nxl_mg(l), nxr_mg(l), 2 |
|---|
| 2099 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
|---|
| 2100 | |
|---|
| 2101 | IF ( j == nys_mg(l) ) THEN |
|---|
| 2102 | !DIR$ IVDEP |
|---|
| 2103 | DO k = nzb+1, ind_even_odd |
|---|
| 2104 | temp(k,jys,i1) = p_mg(k,j,i) |
|---|
| 2105 | ENDDO |
|---|
| 2106 | i1 = i1 + 1 |
|---|
| 2107 | |
|---|
| 2108 | ENDIF |
|---|
| 2109 | |
|---|
| 2110 | IF ( j == nyn_mg(l) ) THEN |
|---|
| 2111 | !DIR$ IVDEP |
|---|
| 2112 | DO k = nzb+1, ind_even_odd |
|---|
| 2113 | temp(k,jyn,i2) = p_mg(k,j,i) |
|---|
| 2114 | ENDDO |
|---|
| 2115 | i2 = i2 + 1 |
|---|
| 2116 | |
|---|
| 2117 | ENDIF |
|---|
| 2118 | |
|---|
| 2119 | ENDDO |
|---|
| 2120 | ENDDO |
|---|
| 2121 | |
|---|
| 2122 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
|---|
| 2123 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
|---|
| 2124 | |
|---|
| 2125 | IF ( j == nys_mg(l) ) THEN |
|---|
| 2126 | !DIR$ IVDEP |
|---|
| 2127 | DO k = nzb+1, ind_even_odd |
|---|
| 2128 | temp(k,jys,i1) = p_mg(k,j,i) |
|---|
| 2129 | ENDDO |
|---|
| 2130 | i1 = i1 + 1 |
|---|
| 2131 | |
|---|
| 2132 | ENDIF |
|---|
| 2133 | |
|---|
| 2134 | IF ( j == nyn_mg(l) ) THEN |
|---|
| 2135 | !DIR$ IVDEP |
|---|
| 2136 | DO k = nzb+1, ind_even_odd |
|---|
| 2137 | temp(k,jyn,i2) = p_mg(k,j,i) |
|---|
| 2138 | ENDDO |
|---|
| 2139 | i2 = i2 + 1 |
|---|
| 2140 | |
|---|
| 2141 | ENDIF |
|---|
| 2142 | |
|---|
| 2143 | ENDDO |
|---|
| 2144 | ENDDO |
|---|
| 2145 | |
|---|
| 2146 | grid_level = grid_level-1 |
|---|
| 2147 | |
|---|
| 2148 | send_receive = 'ns' |
|---|
| 2149 | CALL exchange_horiz( temp, 1 ) |
|---|
| 2150 | |
|---|
| 2151 | grid_level = grid_level+1 |
|---|
| 2152 | |
|---|
| 2153 | i1 = nxl_mg(grid_level-1) |
|---|
| 2154 | i2 = nxl_mg(grid_level-1) |
|---|
| 2155 | |
|---|
| 2156 | DO i = nxl_mg(l), nxr_mg(l), 2 |
|---|
| 2157 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
|---|
| 2158 | |
|---|
| 2159 | IF ( j == nys_mg(l) ) THEN |
|---|
| 2160 | !DIR$ IVDEP |
|---|
| 2161 | DO k = nzb+1, ind_even_odd |
|---|
| 2162 | p_mg(k,nyn_mg(l)+1,i) = temp(k,jyn+1,i1) |
|---|
| 2163 | ENDDO |
|---|
| 2164 | i1 = i1 + 1 |
|---|
| 2165 | |
|---|
| 2166 | ENDIF |
|---|
| 2167 | |
|---|
| 2168 | IF ( j == nyn_mg(l) ) THEN |
|---|
| 2169 | !DIR$ IVDEP |
|---|
| 2170 | DO k = nzb+1, ind_even_odd |
|---|
| 2171 | p_mg(k,nys_mg(l)-1,i) = temp(k,jys-1,i2) |
|---|
| 2172 | ENDDO |
|---|
| 2173 | i2 = i2 + 1 |
|---|
| 2174 | |
|---|
| 2175 | ENDIF |
|---|
| 2176 | |
|---|
| 2177 | ENDDO |
|---|
| 2178 | ENDDO |
|---|
| 2179 | |
|---|
| 2180 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
|---|
| 2181 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
|---|
| 2182 | |
|---|
| 2183 | IF ( j == nys_mg(l) ) THEN |
|---|
| 2184 | !DIR$ IVDEP |
|---|
| 2185 | DO k = nzb+1, ind_even_odd |
|---|
| 2186 | p_mg(k,nyn_mg(l)+1,i) = temp(k,jyn+1,i1) |
|---|
| 2187 | ENDDO |
|---|
| 2188 | i1 = i1 + 1 |
|---|
| 2189 | |
|---|
| 2190 | ENDIF |
|---|
| 2191 | |
|---|
| 2192 | IF ( j == nyn_mg(l) ) THEN |
|---|
| 2193 | !DIR$ IVDEP |
|---|
| 2194 | DO k = nzb+1, ind_even_odd |
|---|
| 2195 | p_mg(k,nys_mg(l)-1,i) = temp(k,jys-1,i2) |
|---|
| 2196 | ENDDO |
|---|
| 2197 | i2 = i2 + 1 |
|---|
| 2198 | |
|---|
| 2199 | ENDIF |
|---|
| 2200 | |
|---|
| 2201 | ENDDO |
|---|
| 2202 | ENDDO |
|---|
| 2203 | |
|---|
| 2204 | j1 = nys_mg(grid_level-1) |
|---|
| 2205 | j2 = nys_mg(grid_level-1) |
|---|
| 2206 | |
|---|
| 2207 | DO i = nxl_mg(l), nxr_mg(l), 2 |
|---|
| 2208 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
|---|
| 2209 | |
|---|
| 2210 | IF ( i == nxl_mg(l) ) THEN |
|---|
| 2211 | !DIR$ IVDEP |
|---|
| 2212 | DO k = nzb+1, ind_even_odd |
|---|
| 2213 | temp(k,j1,ixl) = p_mg(k,j,i) |
|---|
| 2214 | ENDDO |
|---|
| 2215 | j1 = j1 + 1 |
|---|
| 2216 | |
|---|
| 2217 | ENDIF |
|---|
| 2218 | |
|---|
| 2219 | IF ( i == nxr_mg(l) ) THEN |
|---|
| 2220 | !DIR$ IVDEP |
|---|
| 2221 | DO k = nzb+1, ind_even_odd |
|---|
| 2222 | temp(k,j2,ixr) = p_mg(k,j,i) |
|---|
| 2223 | ENDDO |
|---|
| 2224 | j2 = j2 + 1 |
|---|
| 2225 | |
|---|
| 2226 | ENDIF |
|---|
| 2227 | |
|---|
| 2228 | ENDDO |
|---|
| 2229 | ENDDO |
|---|
| 2230 | |
|---|
| 2231 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
|---|
| 2232 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
|---|
| 2233 | |
|---|
| 2234 | IF ( i == nxl_mg(l) ) THEN |
|---|
| 2235 | !DIR$ IVDEP |
|---|
| 2236 | DO k = nzb+1, ind_even_odd |
|---|
| 2237 | temp(k,j1,ixl) = p_mg(k,j,i) |
|---|
| 2238 | ENDDO |
|---|
| 2239 | j1 = j1 + 1 |
|---|
| 2240 | |
|---|
| 2241 | ENDIF |
|---|
| 2242 | |
|---|
| 2243 | IF ( i == nxr_mg(l) ) THEN |
|---|
| 2244 | !DIR$ IVDEP |
|---|
| 2245 | DO k = nzb+1, ind_even_odd |
|---|
| 2246 | temp(k,j2,ixr) = p_mg(k,j,i) |
|---|
| 2247 | ENDDO |
|---|
| 2248 | j2 = j2 + 1 |
|---|
| 2249 | |
|---|
| 2250 | ENDIF |
|---|
| 2251 | |
|---|
| 2252 | ENDDO |
|---|
| 2253 | ENDDO |
|---|
| 2254 | |
|---|
| 2255 | grid_level = grid_level-1 |
|---|
| 2256 | |
|---|
| 2257 | send_receive = 'lr' |
|---|
| 2258 | CALL exchange_horiz( temp, 1 ) |
|---|
| 2259 | |
|---|
| 2260 | grid_level = grid_level+1 |
|---|
| 2261 | |
|---|
| 2262 | nxl = nxl_mg(grid_level) |
|---|
| 2263 | nys = nys_mg(grid_level) |
|---|
| 2264 | nxr = nxr_mg(grid_level) |
|---|
| 2265 | nyn = nyn_mg(grid_level) |
|---|
| 2266 | nzt = nzt_mg(grid_level) |
|---|
| 2267 | |
|---|
| 2268 | j1 = nys_mg(grid_level-1) |
|---|
| 2269 | j2 = nys_mg(grid_level-1) |
|---|
| 2270 | |
|---|
| 2271 | DO i = nxl_mg(l), nxr_mg(l), 2 |
|---|
| 2272 | DO j = nys_mg(l) + (color-1), nyn_mg(l), 2 |
|---|
| 2273 | |
|---|
| 2274 | IF ( i == nxl_mg(l) ) THEN |
|---|
| 2275 | !DIR$ IVDEP |
|---|
| 2276 | DO k = nzb+1, ind_even_odd |
|---|
| 2277 | p_mg(k,j,nxr_mg(l)+1) = temp(k,j1,ixr+1) |
|---|
| 2278 | ENDDO |
|---|
| 2279 | j1 = j1 + 1 |
|---|
| 2280 | |
|---|
| 2281 | ENDIF |
|---|
| 2282 | |
|---|
| 2283 | IF ( i == nxr_mg(l) ) THEN |
|---|
| 2284 | !DIR$ IVDEP |
|---|
| 2285 | DO k = nzb+1, ind_even_odd |
|---|
| 2286 | p_mg(k,j,nxl_mg(l)-1) = temp(k,j2,ixl-1) |
|---|
| 2287 | ENDDO |
|---|
| 2288 | j2 = j2 + 1 |
|---|
| 2289 | |
|---|
| 2290 | ENDIF |
|---|
| 2291 | |
|---|
| 2292 | ENDDO |
|---|
| 2293 | ENDDO |
|---|
| 2294 | |
|---|
| 2295 | DO i = nxl_mg(l)+1, nxr_mg(l), 2 |
|---|
| 2296 | DO j = nys_mg(l) + 2 - color, nyn_mg(l), 2 |
|---|
| 2297 | |
|---|
| 2298 | IF ( i == nxl_mg(l) ) THEN |
|---|
| 2299 | !DIR$ IVDEP |
|---|
| 2300 | DO k = nzb+1, ind_even_odd |
|---|
| 2301 | p_mg(k,j,nxr_mg(l)+1) = temp(k,j1,ixr+1) |
|---|
| 2302 | ENDDO |
|---|
| 2303 | j1 = j1 + 1 |
|---|
| 2304 | |
|---|
| 2305 | ENDIF |
|---|
| 2306 | |
|---|
| 2307 | IF ( i == nxr_mg(l) ) THEN |
|---|
| 2308 | !DIR$ IVDEP |
|---|
| 2309 | DO k = nzb+1, ind_even_odd |
|---|
| 2310 | p_mg(k,j,nxl_mg(l)-1) = temp(k,j2,ixl-1) |
|---|
| 2311 | ENDDO |
|---|
| 2312 | j2 = j2 + 1 |
|---|
| 2313 | |
|---|
| 2314 | ENDIF |
|---|
| 2315 | |
|---|
| 2316 | ENDDO |
|---|
| 2317 | ENDDO |
|---|
| 2318 | |
|---|
| 2319 | DEALLOCATE( temp ) |
|---|
| 2320 | |
|---|
| 2321 | ELSE |
|---|
| 2322 | |
|---|
| 2323 | ! |
|---|
| 2324 | !-- Standard horizontal ghost boundary exchange for small coarse grid |
|---|
| 2325 | !-- levels, where the transfer time is latency bound |
|---|
| 2326 | CALL exchange_horiz( p_mg, 1 ) |
|---|
| 2327 | |
|---|
| 2328 | ENDIF |
|---|
| 2329 | |
|---|
| 2330 | ! |
|---|
| 2331 | !-- Reset values to default PALM setup |
|---|
| 2332 | sendrecv_in_background = sendrecv_in_background_save |
|---|
| 2333 | synchronous_exchange = synchronous_exchange_save |
|---|
| 2334 | send_receive = 'al' |
|---|
| 2335 | #else |
|---|
| 2336 | |
|---|
| 2337 | ! |
|---|
| 2338 | !-- Standard horizontal ghost boundary exchange for small coarse grid |
|---|
| 2339 | !-- levels, where the transfer time is latency bound |
|---|
| 2340 | CALL exchange_horiz( p_mg, 1 ) |
|---|
| 2341 | #endif |
|---|
| 2342 | |
|---|
| 2343 | END SUBROUTINE special_exchange_horiz |
|---|
| 2344 | |
|---|
| 2345 | END MODULE poismg_mod |
|---|
