[1682] | 1 | !> @file fft_xy.f90 |
---|
[1036] | 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 | ! |
---|
[1310] | 16 | ! Copyright 1997-2014 Leibniz Universitaet Hannover |
---|
[1322] | 17 | !------------------------------------------------------------------------------! |
---|
[1036] | 18 | ! |
---|
[254] | 19 | ! Current revisions: |
---|
[1] | 20 | ! ----------------- |
---|
[1682] | 21 | ! Code annotations made doxygen readable |
---|
[1483] | 22 | ! |
---|
[1321] | 23 | ! Former revisions: |
---|
| 24 | ! ----------------- |
---|
| 25 | ! $Id: fft_xy.f90 1682 2015-10-07 23:56:08Z knoop $ |
---|
| 26 | ! |
---|
[1601] | 27 | ! 1600 2015-06-11 15:50:12Z raasch |
---|
| 28 | ! bugfix: openMP threadprivate statement moved after variable declaration |
---|
| 29 | ! |
---|
[1483] | 30 | ! 1482 2014-10-18 12:34:45Z raasch |
---|
| 31 | ! cudafft workaround for data declaration of ar_tmp because of PGI 14.1 bug |
---|
| 32 | ! |
---|
[1403] | 33 | ! 1402 2014-05-09 14:25:13Z raasch |
---|
| 34 | ! fortran bugfix for r1392 |
---|
| 35 | ! |
---|
[1399] | 36 | ! 1398 2014-05-07 11:15:00Z heinze |
---|
| 37 | ! bugfix: typo removed for KIND in CMPLX function |
---|
| 38 | ! |
---|
[1393] | 39 | ! 1392 2014-05-06 09:10:05Z raasch |
---|
| 40 | ! bugfix: KIND attribute added to CMPLX functions |
---|
| 41 | ! |
---|
[1375] | 42 | ! 1374 2014-04-25 12:55:07Z raasch |
---|
| 43 | ! bugfixes: missing variables added to ONLY list, dpk renamed dp |
---|
| 44 | ! |
---|
[1373] | 45 | ! 1372 2014-04-24 06:29:32Z raasch |
---|
| 46 | ! openMP-bugfix for fftw: some arrays defined as threadprivate |
---|
| 47 | ! |
---|
[1354] | 48 | ! 1353 2014-04-08 15:21:23Z heinze |
---|
| 49 | ! REAL constants provided with KIND-attribute |
---|
| 50 | ! |
---|
[1343] | 51 | ! 1342 2014-03-26 17:04:47Z kanani |
---|
| 52 | ! REAL constants defined as wp-kind |
---|
| 53 | ! |
---|
[1323] | 54 | ! 1322 2014-03-20 16:38:49Z raasch |
---|
| 55 | ! REAL functions provided with KIND-attribute |
---|
| 56 | ! |
---|
[1321] | 57 | ! 1320 2014-03-20 08:40:49Z raasch |
---|
[1320] | 58 | ! ONLY-attribute added to USE-statements, |
---|
| 59 | ! kind-parameters added to all INTEGER and REAL declaration statements, |
---|
| 60 | ! kinds are defined in new module kinds, |
---|
| 61 | ! old module precision_kind is removed, |
---|
| 62 | ! revision history before 2012 removed, |
---|
| 63 | ! comment fields (!:) to be used for variable explanations added to |
---|
| 64 | ! all variable declaration statements |
---|
[1] | 65 | ! |
---|
[1305] | 66 | ! 1304 2014-03-12 10:29:42Z raasch |
---|
| 67 | ! openmp bugfix: work1 used in Temperton algorithm must be private |
---|
| 68 | ! |
---|
[1258] | 69 | ! 1257 2013-11-08 15:18:40Z raasch |
---|
| 70 | ! openacc loop and loop vector clauses removed, declare create moved after |
---|
| 71 | ! the FORTRAN declaration statement |
---|
| 72 | ! |
---|
[1220] | 73 | ! 1219 2013-08-30 09:33:18Z heinze |
---|
| 74 | ! bugfix: use own branch for fftw |
---|
| 75 | ! |
---|
[1217] | 76 | ! 1216 2013-08-26 09:31:42Z raasch |
---|
| 77 | ! fft_x and fft_y modified for parallel / ovverlapping execution of fft and |
---|
| 78 | ! transpositions, |
---|
| 79 | ! fftw implemented for 1d-decomposition (fft_x_1d, fft_y_1d) |
---|
| 80 | ! |
---|
[1211] | 81 | ! 1210 2013-08-14 10:58:20Z raasch |
---|
| 82 | ! fftw added |
---|
| 83 | ! |
---|
[1167] | 84 | ! 1166 2013-05-24 13:55:44Z raasch |
---|
| 85 | ! C_DOUBLE/COMPLEX reset to dpk |
---|
| 86 | ! |
---|
[1154] | 87 | ! 1153 2013-05-10 14:33:08Z raasch |
---|
| 88 | ! code adjustment of data types for CUDA fft required by PGI 12.3 / CUDA 5.0 |
---|
| 89 | ! |
---|
[1112] | 90 | ! 1111 2013-03-08 23:54:10Z raasch |
---|
| 91 | ! further openACC statements added, CUDA branch completely runs on GPU |
---|
| 92 | ! bugfix: CUDA fft plans adjusted for domain decomposition (before they always |
---|
| 93 | ! used total domain) |
---|
| 94 | ! |
---|
[1107] | 95 | ! 1106 2013-03-04 05:31:38Z raasch |
---|
| 96 | ! CUDA fft added |
---|
| 97 | ! array_kind renamed precision_kind, 3D- instead of 1D-loops in fft_x and fft_y |
---|
| 98 | ! old fft_x, fft_y become fft_x_1d, fft_y_1d and are used for 1D-decomposition |
---|
| 99 | ! |
---|
[1093] | 100 | ! 1092 2013-02-02 11:24:22Z raasch |
---|
| 101 | ! variable sizw declared for NEC case only |
---|
| 102 | ! |
---|
[1037] | 103 | ! 1036 2012-10-22 13:43:42Z raasch |
---|
| 104 | ! code put under GPL (PALM 3.9) |
---|
| 105 | ! |
---|
[1] | 106 | ! Revision 1.1 2002/06/11 13:00:49 raasch |
---|
| 107 | ! Initial revision |
---|
| 108 | ! |
---|
| 109 | ! |
---|
| 110 | ! Description: |
---|
| 111 | ! ------------ |
---|
[1682] | 112 | !> Fast Fourier transformation along x and y for 1d domain decomposition along x. |
---|
| 113 | !> Original version: Klaus Ketelsen (May 2002) |
---|
[1] | 114 | !------------------------------------------------------------------------------! |
---|
[1682] | 115 | MODULE fft_xy |
---|
| 116 | |
---|
[1] | 117 | |
---|
[1320] | 118 | USE control_parameters, & |
---|
| 119 | ONLY: fft_method, message_string |
---|
| 120 | |
---|
| 121 | USE indices, & |
---|
| 122 | ONLY: nx, ny, nz |
---|
| 123 | |
---|
[1153] | 124 | #if defined( __cuda_fft ) |
---|
| 125 | USE ISO_C_BINDING |
---|
[1210] | 126 | #elif defined( __fftw ) |
---|
| 127 | USE, INTRINSIC :: ISO_C_BINDING |
---|
[1153] | 128 | #endif |
---|
[1320] | 129 | |
---|
| 130 | USE kinds |
---|
| 131 | |
---|
| 132 | USE singleton, & |
---|
| 133 | ONLY: fftn |
---|
| 134 | |
---|
[1] | 135 | USE temperton_fft |
---|
[1320] | 136 | |
---|
| 137 | USE transpose_indices, & |
---|
[1374] | 138 | ONLY: nxl_y, nxr_y, nyn_x, nys_x, nzb_x, nzb_y, nzt_x, nzt_y |
---|
[1] | 139 | |
---|
| 140 | IMPLICIT NONE |
---|
| 141 | |
---|
| 142 | PRIVATE |
---|
[1106] | 143 | PUBLIC fft_x, fft_x_1d, fft_y, fft_y_1d, fft_init, fft_x_m, fft_y_m |
---|
[1] | 144 | |
---|
[1682] | 145 | INTEGER(iwp), DIMENSION(:), ALLOCATABLE, SAVE :: ifax_x !< |
---|
| 146 | INTEGER(iwp), DIMENSION(:), ALLOCATABLE, SAVE :: ifax_y !< |
---|
[1] | 147 | |
---|
[1682] | 148 | LOGICAL, SAVE :: init_fft = .FALSE. !< |
---|
[1] | 149 | |
---|
[1682] | 150 | REAL(wp), SAVE :: dnx !< |
---|
| 151 | REAL(wp), SAVE :: dny !< |
---|
| 152 | REAL(wp), SAVE :: sqr_dnx !< |
---|
| 153 | REAL(wp), SAVE :: sqr_dny !< |
---|
[1320] | 154 | |
---|
[1682] | 155 | REAL(wp), DIMENSION(:), ALLOCATABLE, SAVE :: trigs_x !< |
---|
| 156 | REAL(wp), DIMENSION(:), ALLOCATABLE, SAVE :: trigs_y !< |
---|
[1] | 157 | |
---|
| 158 | #if defined( __ibm ) |
---|
[1682] | 159 | INTEGER(iwp), PARAMETER :: nau1 = 20000 !< |
---|
| 160 | INTEGER(iwp), PARAMETER :: nau2 = 22000 !< |
---|
[1] | 161 | ! |
---|
| 162 | !-- The following working arrays contain tables and have to be "save" and |
---|
| 163 | !-- shared in OpenMP sense |
---|
[1682] | 164 | REAL(wp), DIMENSION(nau1), SAVE :: aux1 !< |
---|
| 165 | REAL(wp), DIMENSION(nau1), SAVE :: auy1 !< |
---|
| 166 | REAL(wp), DIMENSION(nau1), SAVE :: aux3 !< |
---|
| 167 | REAL(wp), DIMENSION(nau1), SAVE :: auy3 !< |
---|
[1320] | 168 | |
---|
[1] | 169 | #elif defined( __nec ) |
---|
[1682] | 170 | INTEGER(iwp), SAVE :: nz1 !< |
---|
[1320] | 171 | |
---|
[1682] | 172 | REAL(wp), DIMENSION(:), ALLOCATABLE, SAVE :: trig_xb !< |
---|
| 173 | REAL(wp), DIMENSION(:), ALLOCATABLE, SAVE :: trig_xf !< |
---|
| 174 | REAL(wp), DIMENSION(:), ALLOCATABLE, SAVE :: trig_yb !< |
---|
| 175 | REAL(wp), DIMENSION(:), ALLOCATABLE, SAVE :: trig_yf !< |
---|
[1320] | 176 | |
---|
[1106] | 177 | #elif defined( __cuda_fft ) |
---|
[1682] | 178 | INTEGER(C_INT), SAVE :: plan_xf !< |
---|
| 179 | INTEGER(C_INT), SAVE :: plan_xi !< |
---|
| 180 | INTEGER(C_INT), SAVE :: plan_yf !< |
---|
| 181 | INTEGER(C_INT), SAVE :: plan_yi !< |
---|
[1320] | 182 | |
---|
[1682] | 183 | INTEGER(iwp), SAVE :: total_points_x_transpo !< |
---|
| 184 | INTEGER(iwp), SAVE :: total_points_y_transpo !< |
---|
[1219] | 185 | #endif |
---|
| 186 | |
---|
| 187 | #if defined( __fftw ) |
---|
[1210] | 188 | INCLUDE 'fftw3.f03' |
---|
[1682] | 189 | INTEGER(KIND=C_INT) :: nx_c !< |
---|
| 190 | INTEGER(KIND=C_INT) :: ny_c !< |
---|
[1320] | 191 | |
---|
[1682] | 192 | COMPLEX(KIND=C_DOUBLE_COMPLEX), DIMENSION(:), ALLOCATABLE, SAVE :: x_out !< |
---|
[1320] | 193 | COMPLEX(KIND=C_DOUBLE_COMPLEX), DIMENSION(:), ALLOCATABLE, SAVE :: & |
---|
[1682] | 194 | y_out !< |
---|
[1320] | 195 | |
---|
| 196 | REAL(KIND=C_DOUBLE), DIMENSION(:), ALLOCATABLE, SAVE :: & |
---|
[1682] | 197 | x_in !< |
---|
[1320] | 198 | REAL(KIND=C_DOUBLE), DIMENSION(:), ALLOCATABLE, SAVE :: & |
---|
[1682] | 199 | y_in !< |
---|
[1600] | 200 | !$OMP THREADPRIVATE( x_out, y_out, x_in, y_in ) |
---|
[1320] | 201 | |
---|
| 202 | |
---|
[1210] | 203 | TYPE(C_PTR), SAVE :: plan_xf, plan_xi, plan_yf, plan_yi |
---|
[1] | 204 | #endif |
---|
| 205 | |
---|
| 206 | ! |
---|
| 207 | !-- Public interfaces |
---|
| 208 | INTERFACE fft_init |
---|
| 209 | MODULE PROCEDURE fft_init |
---|
| 210 | END INTERFACE fft_init |
---|
| 211 | |
---|
| 212 | INTERFACE fft_x |
---|
| 213 | MODULE PROCEDURE fft_x |
---|
| 214 | END INTERFACE fft_x |
---|
| 215 | |
---|
[1106] | 216 | INTERFACE fft_x_1d |
---|
| 217 | MODULE PROCEDURE fft_x_1d |
---|
| 218 | END INTERFACE fft_x_1d |
---|
| 219 | |
---|
[1] | 220 | INTERFACE fft_y |
---|
| 221 | MODULE PROCEDURE fft_y |
---|
| 222 | END INTERFACE fft_y |
---|
| 223 | |
---|
[1106] | 224 | INTERFACE fft_y_1d |
---|
| 225 | MODULE PROCEDURE fft_y_1d |
---|
| 226 | END INTERFACE fft_y_1d |
---|
| 227 | |
---|
[1] | 228 | INTERFACE fft_x_m |
---|
| 229 | MODULE PROCEDURE fft_x_m |
---|
| 230 | END INTERFACE fft_x_m |
---|
| 231 | |
---|
| 232 | INTERFACE fft_y_m |
---|
| 233 | MODULE PROCEDURE fft_y_m |
---|
| 234 | END INTERFACE fft_y_m |
---|
| 235 | |
---|
| 236 | CONTAINS |
---|
| 237 | |
---|
| 238 | |
---|
[1682] | 239 | !------------------------------------------------------------------------------! |
---|
| 240 | ! Description: |
---|
| 241 | ! ------------ |
---|
| 242 | !> @todo Missing subroutine description. |
---|
| 243 | !------------------------------------------------------------------------------! |
---|
[1] | 244 | SUBROUTINE fft_init |
---|
| 245 | |
---|
[1106] | 246 | USE cuda_fft_interfaces |
---|
| 247 | |
---|
[1] | 248 | IMPLICIT NONE |
---|
| 249 | |
---|
| 250 | ! |
---|
| 251 | !-- The following temporary working arrays have to be on stack or private |
---|
| 252 | !-- in OpenMP sense |
---|
| 253 | #if defined( __ibm ) |
---|
[1682] | 254 | REAL(wp), DIMENSION(0:nx+2) :: workx !< |
---|
| 255 | REAL(wp), DIMENSION(0:ny+2) :: worky !< |
---|
| 256 | REAL(wp), DIMENSION(nau2) :: aux2 !< |
---|
| 257 | REAL(wp), DIMENSION(nau2) :: auy2 !< |
---|
| 258 | REAL(wp), DIMENSION(nau2) :: aux4 !< |
---|
| 259 | REAL(wp), DIMENSION(nau2) :: auy4 !< |
---|
[1] | 260 | #elif defined( __nec ) |
---|
[1682] | 261 | REAL(wp), DIMENSION(0:nx+3,nz+1) :: work_x !< |
---|
| 262 | REAL(wp), DIMENSION(0:ny+3,nz+1) :: work_y !< |
---|
| 263 | REAL(wp), DIMENSION(6*(nx+3),nz+1) :: workx !< |
---|
| 264 | REAL(wp), DIMENSION(6*(ny+3),nz+1) :: worky !< |
---|
[1] | 265 | #endif |
---|
| 266 | |
---|
| 267 | ! |
---|
| 268 | !-- Return, if already called |
---|
| 269 | IF ( init_fft ) THEN |
---|
| 270 | RETURN |
---|
| 271 | ELSE |
---|
| 272 | init_fft = .TRUE. |
---|
| 273 | ENDIF |
---|
| 274 | |
---|
| 275 | IF ( fft_method == 'system-specific' ) THEN |
---|
| 276 | |
---|
[1342] | 277 | dnx = 1.0_wp / ( nx + 1.0_wp ) |
---|
| 278 | dny = 1.0_wp / ( ny + 1.0_wp ) |
---|
[1106] | 279 | sqr_dnx = SQRT( dnx ) |
---|
| 280 | sqr_dny = SQRT( dny ) |
---|
[1] | 281 | #if defined( __ibm ) && ! defined( __ibmy_special ) |
---|
| 282 | ! |
---|
| 283 | !-- Initialize tables for fft along x |
---|
[1106] | 284 | CALL DRCFT( 1, workx, 1, workx, 1, nx+1, 1, 1, sqr_dnx, aux1, nau1, & |
---|
[1] | 285 | aux2, nau2 ) |
---|
[1106] | 286 | CALL DCRFT( 1, workx, 1, workx, 1, nx+1, 1, -1, sqr_dnx, aux3, nau1, & |
---|
[1] | 287 | aux4, nau2 ) |
---|
| 288 | ! |
---|
| 289 | !-- Initialize tables for fft along y |
---|
[1106] | 290 | CALL DRCFT( 1, worky, 1, worky, 1, ny+1, 1, 1, sqr_dny, auy1, nau1, & |
---|
[1] | 291 | auy2, nau2 ) |
---|
[1106] | 292 | CALL DCRFT( 1, worky, 1, worky, 1, ny+1, 1, -1, sqr_dny, auy3, nau1, & |
---|
[1] | 293 | auy4, nau2 ) |
---|
| 294 | #elif defined( __nec ) |
---|
[254] | 295 | message_string = 'fft method "' // TRIM( fft_method) // & |
---|
| 296 | '" currently does not work on NEC' |
---|
| 297 | CALL message( 'fft_init', 'PA0187', 1, 2, 0, 6, 0 ) |
---|
[1] | 298 | |
---|
[1320] | 299 | ALLOCATE( trig_xb(2*(nx+1)), trig_xf(2*(nx+1)), & |
---|
[1] | 300 | trig_yb(2*(ny+1)), trig_yf(2*(ny+1)) ) |
---|
| 301 | |
---|
[1342] | 302 | work_x = 0.0_wp |
---|
| 303 | work_y = 0.0_wp |
---|
[1] | 304 | nz1 = nz + MOD( nz+1, 2 ) ! odd nz slows down fft significantly |
---|
| 305 | ! when using the NEC ffts |
---|
| 306 | |
---|
| 307 | ! |
---|
| 308 | !-- Initialize tables for fft along x (non-vector and vector case (M)) |
---|
[1106] | 309 | CALL DZFFT( 0, nx+1, sqr_dnx, work_x, work_x, trig_xf, workx, 0 ) |
---|
| 310 | CALL ZDFFT( 0, nx+1, sqr_dnx, work_x, work_x, trig_xb, workx, 0 ) |
---|
[1320] | 311 | CALL DZFFTM( 0, nx+1, nz1, sqr_dnx, work_x, nx+4, work_x, nx+4, & |
---|
[1] | 312 | trig_xf, workx, 0 ) |
---|
[1320] | 313 | CALL ZDFFTM( 0, nx+1, nz1, sqr_dnx, work_x, nx+4, work_x, nx+4, & |
---|
[1] | 314 | trig_xb, workx, 0 ) |
---|
| 315 | ! |
---|
| 316 | !-- Initialize tables for fft along y (non-vector and vector case (M)) |
---|
[1106] | 317 | CALL DZFFT( 0, ny+1, sqr_dny, work_y, work_y, trig_yf, worky, 0 ) |
---|
| 318 | CALL ZDFFT( 0, ny+1, sqr_dny, work_y, work_y, trig_yb, worky, 0 ) |
---|
[1320] | 319 | CALL DZFFTM( 0, ny+1, nz1, sqr_dny, work_y, ny+4, work_y, ny+4, & |
---|
[1] | 320 | trig_yf, worky, 0 ) |
---|
[1320] | 321 | CALL ZDFFTM( 0, ny+1, nz1, sqr_dny, work_y, ny+4, work_y, ny+4, & |
---|
[1] | 322 | trig_yb, worky, 0 ) |
---|
[1106] | 323 | #elif defined( __cuda_fft ) |
---|
| 324 | total_points_x_transpo = (nx+1) * (nyn_x-nys_x+1) * (nzt_x-nzb_x+1) |
---|
| 325 | total_points_y_transpo = (ny+1) * (nxr_y-nxl_y+1) * (nzt_y-nzb_y+1) |
---|
[1111] | 326 | CALL CUFFTPLAN1D( plan_xf, nx+1, CUFFT_D2Z, (nyn_x-nys_x+1) * (nzt_x-nzb_x+1) ) |
---|
| 327 | CALL CUFFTPLAN1D( plan_xi, nx+1, CUFFT_Z2D, (nyn_x-nys_x+1) * (nzt_x-nzb_x+1) ) |
---|
| 328 | CALL CUFFTPLAN1D( plan_yf, ny+1, CUFFT_D2Z, (nxr_y-nxl_y+1) * (nzt_y-nzb_y+1) ) |
---|
| 329 | CALL CUFFTPLAN1D( plan_yi, ny+1, CUFFT_Z2D, (nxr_y-nxl_y+1) * (nzt_y-nzb_y+1) ) |
---|
[1] | 330 | #else |
---|
[254] | 331 | message_string = 'no system-specific fft-call available' |
---|
| 332 | CALL message( 'fft_init', 'PA0188', 1, 2, 0, 6, 0 ) |
---|
[1] | 333 | #endif |
---|
| 334 | ELSEIF ( fft_method == 'temperton-algorithm' ) THEN |
---|
| 335 | ! |
---|
| 336 | !-- Temperton-algorithm |
---|
| 337 | !-- Initialize tables for fft along x and y |
---|
| 338 | ALLOCATE( ifax_x(nx+1), ifax_y(ny+1), trigs_x(nx+1), trigs_y(ny+1) ) |
---|
| 339 | |
---|
| 340 | CALL set99( trigs_x, ifax_x, nx+1 ) |
---|
| 341 | CALL set99( trigs_y, ifax_y, ny+1 ) |
---|
| 342 | |
---|
[1210] | 343 | ELSEIF ( fft_method == 'fftw' ) THEN |
---|
| 344 | ! |
---|
| 345 | !-- FFTW |
---|
| 346 | #if defined( __fftw ) |
---|
| 347 | nx_c = nx+1 |
---|
| 348 | ny_c = ny+1 |
---|
[1372] | 349 | !$OMP PARALLEL |
---|
[1320] | 350 | ALLOCATE( x_in(0:nx+2), y_in(0:ny+2), x_out(0:(nx+1)/2), & |
---|
[1210] | 351 | y_out(0:(ny+1)/2) ) |
---|
[1372] | 352 | !$OMP END PARALLEL |
---|
[1210] | 353 | plan_xf = FFTW_PLAN_DFT_R2C_1D( nx_c, x_in, x_out, FFTW_ESTIMATE ) |
---|
| 354 | plan_xi = FFTW_PLAN_DFT_C2R_1D( nx_c, x_out, x_in, FFTW_ESTIMATE ) |
---|
| 355 | plan_yf = FFTW_PLAN_DFT_R2C_1D( ny_c, y_in, y_out, FFTW_ESTIMATE ) |
---|
| 356 | plan_yi = FFTW_PLAN_DFT_C2R_1D( ny_c, y_out, y_in, FFTW_ESTIMATE ) |
---|
| 357 | #else |
---|
| 358 | message_string = 'preprocessor switch for fftw is missing' |
---|
| 359 | CALL message( 'fft_init', 'PA0080', 1, 2, 0, 6, 0 ) |
---|
| 360 | #endif |
---|
| 361 | |
---|
[1] | 362 | ELSEIF ( fft_method == 'singleton-algorithm' ) THEN |
---|
| 363 | |
---|
| 364 | CONTINUE |
---|
| 365 | |
---|
| 366 | ELSE |
---|
| 367 | |
---|
[254] | 368 | message_string = 'fft method "' // TRIM( fft_method) // & |
---|
| 369 | '" not available' |
---|
| 370 | CALL message( 'fft_init', 'PA0189', 1, 2, 0, 6, 0 ) |
---|
[1] | 371 | ENDIF |
---|
| 372 | |
---|
| 373 | END SUBROUTINE fft_init |
---|
| 374 | |
---|
| 375 | |
---|
[1682] | 376 | !------------------------------------------------------------------------------! |
---|
| 377 | ! Description: |
---|
| 378 | ! ------------ |
---|
| 379 | !> Fourier-transformation along x-direction. |
---|
| 380 | !> Version for 2D-decomposition. |
---|
| 381 | !> It uses internal algorithms (Singleton or Temperton) or |
---|
| 382 | !> system-specific routines, if they are available |
---|
| 383 | !------------------------------------------------------------------------------! |
---|
| 384 | |
---|
[1216] | 385 | SUBROUTINE fft_x( ar, direction, ar_2d ) |
---|
[1] | 386 | |
---|
| 387 | |
---|
[1106] | 388 | USE cuda_fft_interfaces |
---|
[1153] | 389 | #if defined( __cuda_fft ) |
---|
| 390 | USE ISO_C_BINDING |
---|
| 391 | #endif |
---|
[1106] | 392 | |
---|
[1] | 393 | IMPLICIT NONE |
---|
| 394 | |
---|
[1682] | 395 | CHARACTER (LEN=*) :: direction !< |
---|
[1320] | 396 | |
---|
[1682] | 397 | COMPLEX(wp), DIMENSION(:), ALLOCATABLE :: cwork !< |
---|
[1106] | 398 | |
---|
[1682] | 399 | INTEGER(iwp) :: i !< |
---|
| 400 | INTEGER(iwp) :: ishape(1) !< |
---|
| 401 | INTEGER(iwp) :: j !< |
---|
| 402 | INTEGER(iwp) :: k !< |
---|
[1106] | 403 | |
---|
[1682] | 404 | LOGICAL :: forward_fft !< |
---|
[1320] | 405 | |
---|
[1682] | 406 | REAL(wp), DIMENSION(0:nx+2) :: work !< |
---|
| 407 | REAL(wp), DIMENSION(nx+2) :: work1 !< |
---|
[1320] | 408 | |
---|
[1106] | 409 | #if defined( __ibm ) |
---|
[1682] | 410 | REAL(wp), DIMENSION(nau2) :: aux2 !< |
---|
| 411 | REAL(wp), DIMENSION(nau2) :: aux4 !< |
---|
[1106] | 412 | #elif defined( __nec ) |
---|
[1682] | 413 | REAL(wp), DIMENSION(6*(nx+1)) :: work2 !< |
---|
[1106] | 414 | #elif defined( __cuda_fft ) |
---|
[1374] | 415 | COMPLEX(dp), DIMENSION(0:(nx+1)/2,nys_x:nyn_x,nzb_x:nzt_x) :: & |
---|
[1682] | 416 | ar_tmp !< |
---|
[1482] | 417 | ! following does not work for PGI 14.1 -> to be removed later |
---|
| 418 | ! !$acc declare create( ar_tmp ) |
---|
[1106] | 419 | #endif |
---|
| 420 | |
---|
[1320] | 421 | REAL(wp), DIMENSION(0:nx,nys_x:nyn_x), OPTIONAL :: & |
---|
[1682] | 422 | ar_2d !< |
---|
[1320] | 423 | REAL(wp), DIMENSION(0:nx,nys_x:nyn_x,nzb_x:nzt_x) :: & |
---|
[1682] | 424 | ar !< |
---|
[1320] | 425 | |
---|
[1106] | 426 | IF ( direction == 'forward' ) THEN |
---|
| 427 | forward_fft = .TRUE. |
---|
| 428 | ELSE |
---|
| 429 | forward_fft = .FALSE. |
---|
| 430 | ENDIF |
---|
| 431 | |
---|
| 432 | IF ( fft_method == 'singleton-algorithm' ) THEN |
---|
| 433 | |
---|
| 434 | ! |
---|
| 435 | !-- Performing the fft with singleton's software works on every system, |
---|
| 436 | !-- since it is part of the model |
---|
| 437 | ALLOCATE( cwork(0:nx) ) |
---|
| 438 | |
---|
| 439 | IF ( forward_fft ) then |
---|
| 440 | |
---|
| 441 | !$OMP PARALLEL PRIVATE ( cwork, i, ishape, j, k ) |
---|
| 442 | !$OMP DO |
---|
| 443 | DO k = nzb_x, nzt_x |
---|
| 444 | DO j = nys_x, nyn_x |
---|
| 445 | |
---|
| 446 | DO i = 0, nx |
---|
[1392] | 447 | cwork(i) = CMPLX( ar(i,j,k), KIND=wp ) |
---|
[1106] | 448 | ENDDO |
---|
| 449 | |
---|
| 450 | ishape = SHAPE( cwork ) |
---|
| 451 | CALL FFTN( cwork, ishape ) |
---|
| 452 | |
---|
| 453 | DO i = 0, (nx+1)/2 |
---|
[1322] | 454 | ar(i,j,k) = REAL( cwork(i), KIND=wp ) |
---|
[1106] | 455 | ENDDO |
---|
| 456 | DO i = 1, (nx+1)/2 - 1 |
---|
| 457 | ar(nx+1-i,j,k) = -AIMAG( cwork(i) ) |
---|
| 458 | ENDDO |
---|
| 459 | |
---|
| 460 | ENDDO |
---|
| 461 | ENDDO |
---|
| 462 | !$OMP END PARALLEL |
---|
| 463 | |
---|
| 464 | ELSE |
---|
| 465 | |
---|
| 466 | !$OMP PARALLEL PRIVATE ( cwork, i, ishape, j, k ) |
---|
| 467 | !$OMP DO |
---|
| 468 | DO k = nzb_x, nzt_x |
---|
| 469 | DO j = nys_x, nyn_x |
---|
| 470 | |
---|
[1392] | 471 | cwork(0) = CMPLX( ar(0,j,k), 0.0_wp, KIND=wp ) |
---|
[1106] | 472 | DO i = 1, (nx+1)/2 - 1 |
---|
[1392] | 473 | cwork(i) = CMPLX( ar(i,j,k), -ar(nx+1-i,j,k), & |
---|
| 474 | KIND=wp ) |
---|
| 475 | cwork(nx+1-i) = CMPLX( ar(i,j,k), ar(nx+1-i,j,k), & |
---|
| 476 | KIND=wp ) |
---|
[1106] | 477 | ENDDO |
---|
[1392] | 478 | cwork((nx+1)/2) = CMPLX( ar((nx+1)/2,j,k), 0.0_wp, KIND=wp ) |
---|
[1106] | 479 | |
---|
| 480 | ishape = SHAPE( cwork ) |
---|
| 481 | CALL FFTN( cwork, ishape, inv = .TRUE. ) |
---|
| 482 | |
---|
| 483 | DO i = 0, nx |
---|
[1322] | 484 | ar(i,j,k) = REAL( cwork(i), KIND=wp ) |
---|
[1106] | 485 | ENDDO |
---|
| 486 | |
---|
| 487 | ENDDO |
---|
| 488 | ENDDO |
---|
| 489 | !$OMP END PARALLEL |
---|
| 490 | |
---|
| 491 | ENDIF |
---|
| 492 | |
---|
| 493 | DEALLOCATE( cwork ) |
---|
| 494 | |
---|
| 495 | ELSEIF ( fft_method == 'temperton-algorithm' ) THEN |
---|
| 496 | |
---|
| 497 | ! |
---|
| 498 | !-- Performing the fft with Temperton's software works on every system, |
---|
| 499 | !-- since it is part of the model |
---|
| 500 | IF ( forward_fft ) THEN |
---|
| 501 | |
---|
[1304] | 502 | !$OMP PARALLEL PRIVATE ( work, work1, i, j, k ) |
---|
[1106] | 503 | !$OMP DO |
---|
| 504 | DO k = nzb_x, nzt_x |
---|
| 505 | DO j = nys_x, nyn_x |
---|
| 506 | |
---|
| 507 | work(0:nx) = ar(0:nx,j,k) |
---|
| 508 | CALL fft991cy( work, work1, trigs_x, ifax_x, 1, nx+1, nx+1, 1, -1 ) |
---|
| 509 | |
---|
| 510 | DO i = 0, (nx+1)/2 |
---|
| 511 | ar(i,j,k) = work(2*i) |
---|
| 512 | ENDDO |
---|
| 513 | DO i = 1, (nx+1)/2 - 1 |
---|
| 514 | ar(nx+1-i,j,k) = work(2*i+1) |
---|
| 515 | ENDDO |
---|
| 516 | |
---|
| 517 | ENDDO |
---|
| 518 | ENDDO |
---|
| 519 | !$OMP END PARALLEL |
---|
| 520 | |
---|
| 521 | ELSE |
---|
| 522 | |
---|
[1304] | 523 | !$OMP PARALLEL PRIVATE ( work, work1, i, j, k ) |
---|
[1106] | 524 | !$OMP DO |
---|
| 525 | DO k = nzb_x, nzt_x |
---|
| 526 | DO j = nys_x, nyn_x |
---|
| 527 | |
---|
| 528 | DO i = 0, (nx+1)/2 |
---|
| 529 | work(2*i) = ar(i,j,k) |
---|
| 530 | ENDDO |
---|
| 531 | DO i = 1, (nx+1)/2 - 1 |
---|
| 532 | work(2*i+1) = ar(nx+1-i,j,k) |
---|
| 533 | ENDDO |
---|
[1342] | 534 | work(1) = 0.0_wp |
---|
| 535 | work(nx+2) = 0.0_wp |
---|
[1106] | 536 | |
---|
| 537 | CALL fft991cy( work, work1, trigs_x, ifax_x, 1, nx+1, nx+1, 1, 1 ) |
---|
| 538 | ar(0:nx,j,k) = work(0:nx) |
---|
| 539 | |
---|
| 540 | ENDDO |
---|
| 541 | ENDDO |
---|
| 542 | !$OMP END PARALLEL |
---|
| 543 | |
---|
| 544 | ENDIF |
---|
| 545 | |
---|
[1210] | 546 | ELSEIF ( fft_method == 'fftw' ) THEN |
---|
| 547 | |
---|
| 548 | #if defined( __fftw ) |
---|
| 549 | IF ( forward_fft ) THEN |
---|
| 550 | |
---|
| 551 | !$OMP PARALLEL PRIVATE ( work, i, j, k ) |
---|
| 552 | !$OMP DO |
---|
| 553 | DO k = nzb_x, nzt_x |
---|
| 554 | DO j = nys_x, nyn_x |
---|
| 555 | |
---|
| 556 | x_in(0:nx) = ar(0:nx,j,k) |
---|
| 557 | CALL FFTW_EXECUTE_DFT_R2C( plan_xf, x_in, x_out ) |
---|
| 558 | |
---|
[1216] | 559 | IF ( PRESENT( ar_2d ) ) THEN |
---|
[1210] | 560 | |
---|
[1216] | 561 | DO i = 0, (nx+1)/2 |
---|
[1322] | 562 | ar_2d(i,j) = REAL( x_out(i), KIND=wp ) / ( nx+1 ) |
---|
[1216] | 563 | ENDDO |
---|
| 564 | DO i = 1, (nx+1)/2 - 1 |
---|
| 565 | ar_2d(nx+1-i,j) = AIMAG( x_out(i) ) / ( nx+1 ) |
---|
| 566 | ENDDO |
---|
| 567 | |
---|
| 568 | ELSE |
---|
| 569 | |
---|
| 570 | DO i = 0, (nx+1)/2 |
---|
[1322] | 571 | ar(i,j,k) = REAL( x_out(i), KIND=wp ) / ( nx+1 ) |
---|
[1216] | 572 | ENDDO |
---|
| 573 | DO i = 1, (nx+1)/2 - 1 |
---|
| 574 | ar(nx+1-i,j,k) = AIMAG( x_out(i) ) / ( nx+1 ) |
---|
| 575 | ENDDO |
---|
| 576 | |
---|
| 577 | ENDIF |
---|
| 578 | |
---|
[1210] | 579 | ENDDO |
---|
| 580 | ENDDO |
---|
| 581 | !$OMP END PARALLEL |
---|
| 582 | |
---|
[1216] | 583 | ELSE |
---|
[1210] | 584 | !$OMP PARALLEL PRIVATE ( work, i, j, k ) |
---|
| 585 | !$OMP DO |
---|
| 586 | DO k = nzb_x, nzt_x |
---|
| 587 | DO j = nys_x, nyn_x |
---|
| 588 | |
---|
[1216] | 589 | IF ( PRESENT( ar_2d ) ) THEN |
---|
[1210] | 590 | |
---|
[1392] | 591 | x_out(0) = CMPLX( ar_2d(0,j), 0.0_wp, KIND=wp ) |
---|
[1216] | 592 | DO i = 1, (nx+1)/2 - 1 |
---|
[1392] | 593 | x_out(i) = CMPLX( ar_2d(i,j), ar_2d(nx+1-i,j), & |
---|
| 594 | KIND=wp ) |
---|
[1216] | 595 | ENDDO |
---|
[1392] | 596 | x_out((nx+1)/2) = CMPLX( ar_2d((nx+1)/2,j), 0.0_wp, & |
---|
| 597 | KIND=wp ) |
---|
[1216] | 598 | |
---|
| 599 | ELSE |
---|
| 600 | |
---|
[1392] | 601 | x_out(0) = CMPLX( ar(0,j,k), 0.0_wp, KIND=wp ) |
---|
[1216] | 602 | DO i = 1, (nx+1)/2 - 1 |
---|
[1392] | 603 | x_out(i) = CMPLX( ar(i,j,k), ar(nx+1-i,j,k), KIND=wp ) |
---|
[1216] | 604 | ENDDO |
---|
[1392] | 605 | x_out((nx+1)/2) = CMPLX( ar((nx+1)/2,j,k), 0.0_wp, & |
---|
| 606 | KIND=wp ) |
---|
[1216] | 607 | |
---|
| 608 | ENDIF |
---|
| 609 | |
---|
[1210] | 610 | CALL FFTW_EXECUTE_DFT_C2R( plan_xi, x_out, x_in) |
---|
| 611 | ar(0:nx,j,k) = x_in(0:nx) |
---|
| 612 | |
---|
| 613 | ENDDO |
---|
| 614 | ENDDO |
---|
| 615 | !$OMP END PARALLEL |
---|
| 616 | |
---|
[1216] | 617 | ENDIF |
---|
[1210] | 618 | #endif |
---|
| 619 | |
---|
[1106] | 620 | ELSEIF ( fft_method == 'system-specific' ) THEN |
---|
| 621 | |
---|
| 622 | #if defined( __ibm ) && ! defined( __ibmy_special ) |
---|
| 623 | IF ( forward_fft ) THEN |
---|
| 624 | |
---|
| 625 | !$OMP PARALLEL PRIVATE ( work, i, j, k ) |
---|
| 626 | !$OMP DO |
---|
| 627 | DO k = nzb_x, nzt_x |
---|
| 628 | DO j = nys_x, nyn_x |
---|
| 629 | |
---|
[1320] | 630 | CALL DRCFT( 0, ar, 1, work, 1, nx+1, 1, 1, sqr_dnx, aux1, & |
---|
| 631 | nau1, aux2, nau2 ) |
---|
[1106] | 632 | |
---|
| 633 | DO i = 0, (nx+1)/2 |
---|
| 634 | ar(i,j,k) = work(2*i) |
---|
| 635 | ENDDO |
---|
| 636 | DO i = 1, (nx+1)/2 - 1 |
---|
| 637 | ar(nx+1-i,j,k) = work(2*i+1) |
---|
| 638 | ENDDO |
---|
| 639 | |
---|
| 640 | ENDDO |
---|
| 641 | ENDDO |
---|
| 642 | !$OMP END PARALLEL |
---|
| 643 | |
---|
| 644 | ELSE |
---|
| 645 | |
---|
| 646 | !$OMP PARALLEL PRIVATE ( work, i, j, k ) |
---|
| 647 | !$OMP DO |
---|
| 648 | DO k = nzb_x, nzt_x |
---|
| 649 | DO j = nys_x, nyn_x |
---|
| 650 | |
---|
| 651 | DO i = 0, (nx+1)/2 |
---|
| 652 | work(2*i) = ar(i,j,k) |
---|
| 653 | ENDDO |
---|
| 654 | DO i = 1, (nx+1)/2 - 1 |
---|
| 655 | work(2*i+1) = ar(nx+1-i,j,k) |
---|
| 656 | ENDDO |
---|
[1342] | 657 | work(1) = 0.0_wp |
---|
| 658 | work(nx+2) = 0.0_wp |
---|
[1106] | 659 | |
---|
[1320] | 660 | CALL DCRFT( 0, work, 1, work, 1, nx+1, 1, -1, sqr_dnx, & |
---|
| 661 | aux3, nau1, aux4, nau2 ) |
---|
[1106] | 662 | |
---|
| 663 | DO i = 0, nx |
---|
| 664 | ar(i,j,k) = work(i) |
---|
| 665 | ENDDO |
---|
| 666 | |
---|
| 667 | ENDDO |
---|
| 668 | ENDDO |
---|
| 669 | !$OMP END PARALLEL |
---|
| 670 | |
---|
| 671 | ENDIF |
---|
| 672 | |
---|
| 673 | #elif defined( __nec ) |
---|
| 674 | |
---|
| 675 | IF ( forward_fft ) THEN |
---|
| 676 | |
---|
| 677 | !$OMP PARALLEL PRIVATE ( work, i, j, k ) |
---|
| 678 | !$OMP DO |
---|
| 679 | DO k = nzb_x, nzt_x |
---|
| 680 | DO j = nys_x, nyn_x |
---|
| 681 | |
---|
| 682 | work(0:nx) = ar(0:nx,j,k) |
---|
| 683 | |
---|
| 684 | CALL DZFFT( 1, nx+1, sqr_dnx, work, work, trig_xf, work2, 0 ) |
---|
| 685 | |
---|
| 686 | DO i = 0, (nx+1)/2 |
---|
| 687 | ar(i,j,k) = work(2*i) |
---|
| 688 | ENDDO |
---|
| 689 | DO i = 1, (nx+1)/2 - 1 |
---|
| 690 | ar(nx+1-i,j,k) = work(2*i+1) |
---|
| 691 | ENDDO |
---|
| 692 | |
---|
| 693 | ENDDO |
---|
| 694 | ENDDO |
---|
| 695 | !$END OMP PARALLEL |
---|
| 696 | |
---|
| 697 | ELSE |
---|
| 698 | |
---|
| 699 | !$OMP PARALLEL PRIVATE ( work, i, j, k ) |
---|
| 700 | !$OMP DO |
---|
| 701 | DO k = nzb_x, nzt_x |
---|
| 702 | DO j = nys_x, nyn_x |
---|
| 703 | |
---|
| 704 | DO i = 0, (nx+1)/2 |
---|
| 705 | work(2*i) = ar(i,j,k) |
---|
| 706 | ENDDO |
---|
| 707 | DO i = 1, (nx+1)/2 - 1 |
---|
| 708 | work(2*i+1) = ar(nx+1-i,j,k) |
---|
| 709 | ENDDO |
---|
[1342] | 710 | work(1) = 0.0_wp |
---|
| 711 | work(nx+2) = 0.0_wp |
---|
[1106] | 712 | |
---|
| 713 | CALL ZDFFT( -1, nx+1, sqr_dnx, work, work, trig_xb, work2, 0 ) |
---|
| 714 | |
---|
| 715 | ar(0:nx,j,k) = work(0:nx) |
---|
| 716 | |
---|
| 717 | ENDDO |
---|
| 718 | ENDDO |
---|
| 719 | !$OMP END PARALLEL |
---|
| 720 | |
---|
| 721 | ENDIF |
---|
| 722 | |
---|
| 723 | #elif defined( __cuda_fft ) |
---|
| 724 | |
---|
[1482] | 725 | !$acc data create( ar_tmp ) |
---|
[1106] | 726 | IF ( forward_fft ) THEN |
---|
| 727 | |
---|
[1111] | 728 | !$acc data present( ar ) |
---|
| 729 | CALL CUFFTEXECD2Z( plan_xf, ar, ar_tmp ) |
---|
[1106] | 730 | |
---|
[1111] | 731 | !$acc kernels |
---|
[1106] | 732 | DO k = nzb_x, nzt_x |
---|
| 733 | DO j = nys_x, nyn_x |
---|
| 734 | |
---|
| 735 | DO i = 0, (nx+1)/2 |
---|
[1322] | 736 | ar(i,j,k) = REAL( ar_tmp(i,j,k), KIND=wp ) * dnx |
---|
[1106] | 737 | ENDDO |
---|
| 738 | |
---|
| 739 | DO i = 1, (nx+1)/2 - 1 |
---|
[1111] | 740 | ar(nx+1-i,j,k) = AIMAG( ar_tmp(i,j,k) ) * dnx |
---|
[1106] | 741 | ENDDO |
---|
| 742 | |
---|
| 743 | ENDDO |
---|
| 744 | ENDDO |
---|
[1111] | 745 | !$acc end kernels |
---|
| 746 | !$acc end data |
---|
[1106] | 747 | |
---|
| 748 | ELSE |
---|
| 749 | |
---|
[1111] | 750 | !$acc data present( ar ) |
---|
| 751 | !$acc kernels |
---|
[1106] | 752 | DO k = nzb_x, nzt_x |
---|
| 753 | DO j = nys_x, nyn_x |
---|
| 754 | |
---|
[1392] | 755 | ar_tmp(0,j,k) = CMPLX( ar(0,j,k), 0.0_wp, KIND=wp ) |
---|
[1106] | 756 | |
---|
| 757 | DO i = 1, (nx+1)/2 - 1 |
---|
[1392] | 758 | ar_tmp(i,j,k) = CMPLX( ar(i,j,k), ar(nx+1-i,j,k), & |
---|
| 759 | KIND=wp ) |
---|
[1106] | 760 | ENDDO |
---|
[1392] | 761 | ar_tmp((nx+1)/2,j,k) = CMPLX( ar((nx+1)/2,j,k), 0.0_wp, & |
---|
| 762 | KIND=wp ) |
---|
[1106] | 763 | |
---|
| 764 | ENDDO |
---|
| 765 | ENDDO |
---|
[1111] | 766 | !$acc end kernels |
---|
[1106] | 767 | |
---|
[1111] | 768 | CALL CUFFTEXECZ2D( plan_xi, ar_tmp, ar ) |
---|
| 769 | !$acc end data |
---|
[1106] | 770 | |
---|
| 771 | ENDIF |
---|
[1482] | 772 | !$acc end data |
---|
[1106] | 773 | |
---|
| 774 | #else |
---|
| 775 | message_string = 'no system-specific fft-call available' |
---|
| 776 | CALL message( 'fft_x', 'PA0188', 1, 2, 0, 6, 0 ) |
---|
| 777 | #endif |
---|
| 778 | |
---|
| 779 | ELSE |
---|
| 780 | |
---|
| 781 | message_string = 'fft method "' // TRIM( fft_method) // & |
---|
| 782 | '" not available' |
---|
| 783 | CALL message( 'fft_x', 'PA0189', 1, 2, 0, 6, 0 ) |
---|
| 784 | |
---|
| 785 | ENDIF |
---|
| 786 | |
---|
| 787 | END SUBROUTINE fft_x |
---|
| 788 | |
---|
[1682] | 789 | !------------------------------------------------------------------------------! |
---|
| 790 | ! Description: |
---|
| 791 | ! ------------ |
---|
| 792 | !> Fourier-transformation along x-direction. |
---|
| 793 | !> Version for 1D-decomposition. |
---|
| 794 | !> It uses internal algorithms (Singleton or Temperton) or |
---|
| 795 | !> system-specific routines, if they are available |
---|
| 796 | !------------------------------------------------------------------------------! |
---|
| 797 | |
---|
[1106] | 798 | SUBROUTINE fft_x_1d( ar, direction ) |
---|
| 799 | |
---|
| 800 | |
---|
| 801 | IMPLICIT NONE |
---|
| 802 | |
---|
[1682] | 803 | CHARACTER (LEN=*) :: direction !< |
---|
[1320] | 804 | |
---|
[1682] | 805 | INTEGER(iwp) :: i !< |
---|
| 806 | INTEGER(iwp) :: ishape(1) !< |
---|
[1] | 807 | |
---|
[1682] | 808 | LOGICAL :: forward_fft !< |
---|
[1106] | 809 | |
---|
[1682] | 810 | REAL(wp), DIMENSION(0:nx) :: ar !< |
---|
| 811 | REAL(wp), DIMENSION(0:nx+2) :: work !< |
---|
| 812 | REAL(wp), DIMENSION(nx+2) :: work1 !< |
---|
[1320] | 813 | |
---|
[1682] | 814 | COMPLEX(wp), DIMENSION(:), ALLOCATABLE :: cwork !< |
---|
[1320] | 815 | |
---|
[1] | 816 | #if defined( __ibm ) |
---|
[1682] | 817 | REAL(wp), DIMENSION(nau2) :: aux2 !< |
---|
| 818 | REAL(wp), DIMENSION(nau2) :: aux4 !< |
---|
[1] | 819 | #elif defined( __nec ) |
---|
[1682] | 820 | REAL(wp), DIMENSION(6*(nx+1)) :: work2 !< |
---|
[1] | 821 | #endif |
---|
| 822 | |
---|
[1106] | 823 | IF ( direction == 'forward' ) THEN |
---|
| 824 | forward_fft = .TRUE. |
---|
| 825 | ELSE |
---|
| 826 | forward_fft = .FALSE. |
---|
| 827 | ENDIF |
---|
| 828 | |
---|
[1] | 829 | IF ( fft_method == 'singleton-algorithm' ) THEN |
---|
| 830 | |
---|
| 831 | ! |
---|
| 832 | !-- Performing the fft with singleton's software works on every system, |
---|
| 833 | !-- since it is part of the model |
---|
| 834 | ALLOCATE( cwork(0:nx) ) |
---|
| 835 | |
---|
[1106] | 836 | IF ( forward_fft ) then |
---|
[1] | 837 | |
---|
| 838 | DO i = 0, nx |
---|
[1392] | 839 | cwork(i) = CMPLX( ar(i), KIND=wp ) |
---|
[1] | 840 | ENDDO |
---|
| 841 | ishape = SHAPE( cwork ) |
---|
| 842 | CALL FFTN( cwork, ishape ) |
---|
| 843 | DO i = 0, (nx+1)/2 |
---|
[1322] | 844 | ar(i) = REAL( cwork(i), KIND=wp ) |
---|
[1] | 845 | ENDDO |
---|
| 846 | DO i = 1, (nx+1)/2 - 1 |
---|
| 847 | ar(nx+1-i) = -AIMAG( cwork(i) ) |
---|
| 848 | ENDDO |
---|
| 849 | |
---|
| 850 | ELSE |
---|
| 851 | |
---|
[1392] | 852 | cwork(0) = CMPLX( ar(0), 0.0_wp, KIND=wp ) |
---|
[1] | 853 | DO i = 1, (nx+1)/2 - 1 |
---|
[1392] | 854 | cwork(i) = CMPLX( ar(i), -ar(nx+1-i), KIND=wp ) |
---|
| 855 | cwork(nx+1-i) = CMPLX( ar(i), ar(nx+1-i), KIND=wp ) |
---|
[1] | 856 | ENDDO |
---|
[1392] | 857 | cwork((nx+1)/2) = CMPLX( ar((nx+1)/2), 0.0_wp, KIND=wp ) |
---|
[1] | 858 | |
---|
| 859 | ishape = SHAPE( cwork ) |
---|
| 860 | CALL FFTN( cwork, ishape, inv = .TRUE. ) |
---|
| 861 | |
---|
| 862 | DO i = 0, nx |
---|
[1322] | 863 | ar(i) = REAL( cwork(i), KIND=wp ) |
---|
[1] | 864 | ENDDO |
---|
| 865 | |
---|
| 866 | ENDIF |
---|
| 867 | |
---|
| 868 | DEALLOCATE( cwork ) |
---|
| 869 | |
---|
| 870 | ELSEIF ( fft_method == 'temperton-algorithm' ) THEN |
---|
| 871 | |
---|
| 872 | ! |
---|
| 873 | !-- Performing the fft with Temperton's software works on every system, |
---|
| 874 | !-- since it is part of the model |
---|
[1106] | 875 | IF ( forward_fft ) THEN |
---|
[1] | 876 | |
---|
| 877 | work(0:nx) = ar |
---|
| 878 | CALL fft991cy( work, work1, trigs_x, ifax_x, 1, nx+1, nx+1, 1, -1 ) |
---|
| 879 | |
---|
| 880 | DO i = 0, (nx+1)/2 |
---|
| 881 | ar(i) = work(2*i) |
---|
| 882 | ENDDO |
---|
| 883 | DO i = 1, (nx+1)/2 - 1 |
---|
| 884 | ar(nx+1-i) = work(2*i+1) |
---|
| 885 | ENDDO |
---|
| 886 | |
---|
| 887 | ELSE |
---|
| 888 | |
---|
| 889 | DO i = 0, (nx+1)/2 |
---|
| 890 | work(2*i) = ar(i) |
---|
| 891 | ENDDO |
---|
| 892 | DO i = 1, (nx+1)/2 - 1 |
---|
| 893 | work(2*i+1) = ar(nx+1-i) |
---|
| 894 | ENDDO |
---|
[1342] | 895 | work(1) = 0.0_wp |
---|
| 896 | work(nx+2) = 0.0_wp |
---|
[1] | 897 | |
---|
| 898 | CALL fft991cy( work, work1, trigs_x, ifax_x, 1, nx+1, nx+1, 1, 1 ) |
---|
| 899 | ar = work(0:nx) |
---|
| 900 | |
---|
| 901 | ENDIF |
---|
| 902 | |
---|
[1216] | 903 | ELSEIF ( fft_method == 'fftw' ) THEN |
---|
| 904 | |
---|
| 905 | #if defined( __fftw ) |
---|
| 906 | IF ( forward_fft ) THEN |
---|
| 907 | |
---|
| 908 | x_in(0:nx) = ar(0:nx) |
---|
| 909 | CALL FFTW_EXECUTE_DFT_R2C( plan_xf, x_in, x_out ) |
---|
| 910 | |
---|
| 911 | DO i = 0, (nx+1)/2 |
---|
[1322] | 912 | ar(i) = REAL( x_out(i), KIND=wp ) / ( nx+1 ) |
---|
[1216] | 913 | ENDDO |
---|
| 914 | DO i = 1, (nx+1)/2 - 1 |
---|
| 915 | ar(nx+1-i) = AIMAG( x_out(i) ) / ( nx+1 ) |
---|
| 916 | ENDDO |
---|
| 917 | |
---|
| 918 | ELSE |
---|
| 919 | |
---|
[1392] | 920 | x_out(0) = CMPLX( ar(0), 0.0_wp, KIND=wp ) |
---|
[1216] | 921 | DO i = 1, (nx+1)/2 - 1 |
---|
[1392] | 922 | x_out(i) = CMPLX( ar(i), ar(nx+1-i), KIND=wp ) |
---|
[1216] | 923 | ENDDO |
---|
[1392] | 924 | x_out((nx+1)/2) = CMPLX( ar((nx+1)/2), 0.0_wp, KIND=wp ) |
---|
[1216] | 925 | |
---|
| 926 | CALL FFTW_EXECUTE_DFT_C2R( plan_xi, x_out, x_in) |
---|
| 927 | ar(0:nx) = x_in(0:nx) |
---|
| 928 | |
---|
| 929 | ENDIF |
---|
| 930 | #endif |
---|
| 931 | |
---|
[1] | 932 | ELSEIF ( fft_method == 'system-specific' ) THEN |
---|
| 933 | |
---|
| 934 | #if defined( __ibm ) && ! defined( __ibmy_special ) |
---|
[1106] | 935 | IF ( forward_fft ) THEN |
---|
[1] | 936 | |
---|
[1320] | 937 | CALL DRCFT( 0, ar, 1, work, 1, nx+1, 1, 1, sqr_dnx, aux1, nau1, & |
---|
[1] | 938 | aux2, nau2 ) |
---|
| 939 | |
---|
| 940 | DO i = 0, (nx+1)/2 |
---|
| 941 | ar(i) = work(2*i) |
---|
| 942 | ENDDO |
---|
| 943 | DO i = 1, (nx+1)/2 - 1 |
---|
| 944 | ar(nx+1-i) = work(2*i+1) |
---|
| 945 | ENDDO |
---|
| 946 | |
---|
| 947 | ELSE |
---|
| 948 | |
---|
| 949 | DO i = 0, (nx+1)/2 |
---|
| 950 | work(2*i) = ar(i) |
---|
| 951 | ENDDO |
---|
| 952 | DO i = 1, (nx+1)/2 - 1 |
---|
| 953 | work(2*i+1) = ar(nx+1-i) |
---|
| 954 | ENDDO |
---|
[1342] | 955 | work(1) = 0.0_wp |
---|
| 956 | work(nx+2) = 0.0_wp |
---|
[1] | 957 | |
---|
[1106] | 958 | CALL DCRFT( 0, work, 1, work, 1, nx+1, 1, -1, sqr_dnx, aux3, nau1, & |
---|
[1] | 959 | aux4, nau2 ) |
---|
| 960 | |
---|
| 961 | DO i = 0, nx |
---|
| 962 | ar(i) = work(i) |
---|
| 963 | ENDDO |
---|
| 964 | |
---|
| 965 | ENDIF |
---|
| 966 | #elif defined( __nec ) |
---|
[1106] | 967 | IF ( forward_fft ) THEN |
---|
[1] | 968 | |
---|
| 969 | work(0:nx) = ar(0:nx) |
---|
| 970 | |
---|
[1106] | 971 | CALL DZFFT( 1, nx+1, sqr_dnx, work, work, trig_xf, work2, 0 ) |
---|
| 972 | |
---|
[1] | 973 | DO i = 0, (nx+1)/2 |
---|
| 974 | ar(i) = work(2*i) |
---|
| 975 | ENDDO |
---|
| 976 | DO i = 1, (nx+1)/2 - 1 |
---|
| 977 | ar(nx+1-i) = work(2*i+1) |
---|
| 978 | ENDDO |
---|
| 979 | |
---|
| 980 | ELSE |
---|
| 981 | |
---|
| 982 | DO i = 0, (nx+1)/2 |
---|
| 983 | work(2*i) = ar(i) |
---|
| 984 | ENDDO |
---|
| 985 | DO i = 1, (nx+1)/2 - 1 |
---|
| 986 | work(2*i+1) = ar(nx+1-i) |
---|
| 987 | ENDDO |
---|
[1342] | 988 | work(1) = 0.0_wp |
---|
| 989 | work(nx+2) = 0.0_wp |
---|
[1] | 990 | |
---|
[1106] | 991 | CALL ZDFFT( -1, nx+1, sqr_dnx, work, work, trig_xb, work2, 0 ) |
---|
[1] | 992 | |
---|
| 993 | ar(0:nx) = work(0:nx) |
---|
| 994 | |
---|
| 995 | ENDIF |
---|
| 996 | #else |
---|
[254] | 997 | message_string = 'no system-specific fft-call available' |
---|
[1106] | 998 | CALL message( 'fft_x_1d', 'PA0188', 1, 2, 0, 6, 0 ) |
---|
[1] | 999 | #endif |
---|
| 1000 | ELSE |
---|
[274] | 1001 | message_string = 'fft method "' // TRIM( fft_method) // & |
---|
| 1002 | '" not available' |
---|
[1106] | 1003 | CALL message( 'fft_x_1d', 'PA0189', 1, 2, 0, 6, 0 ) |
---|
[1] | 1004 | |
---|
| 1005 | ENDIF |
---|
| 1006 | |
---|
[1106] | 1007 | END SUBROUTINE fft_x_1d |
---|
[1] | 1008 | |
---|
[1682] | 1009 | !------------------------------------------------------------------------------! |
---|
| 1010 | ! Description: |
---|
| 1011 | ! ------------ |
---|
| 1012 | !> Fourier-transformation along y-direction. |
---|
| 1013 | !> Version for 2D-decomposition. |
---|
| 1014 | !> It uses internal algorithms (Singleton or Temperton) or |
---|
| 1015 | !> system-specific routines, if they are available. |
---|
| 1016 | !> |
---|
| 1017 | !> direction: 'forward' or 'backward' |
---|
| 1018 | !> ar, ar_tr: 3D data arrays |
---|
| 1019 | !> forward: ar: before ar_tr: after transformation |
---|
| 1020 | !> backward: ar_tr: before ar: after transfosition |
---|
| 1021 | !> |
---|
| 1022 | !> In case of non-overlapping transposition/transformation: |
---|
| 1023 | !> nxl_y_bound = nxl_y_l = nxl_y |
---|
| 1024 | !> nxr_y_bound = nxr_y_l = nxr_y |
---|
| 1025 | !> |
---|
| 1026 | !> In case of overlapping transposition/transformation |
---|
| 1027 | !> - nxl_y_bound and nxr_y_bound have the original values of |
---|
| 1028 | !> nxl_y, nxr_y. ar_tr is dimensioned using these values. |
---|
| 1029 | !> - nxl_y_l = nxr_y_r. ar is dimensioned with these values, so that |
---|
| 1030 | !> transformation is carried out for a 2D-plane only. |
---|
| 1031 | !------------------------------------------------------------------------------! |
---|
| 1032 | |
---|
[1216] | 1033 | SUBROUTINE fft_y( ar, direction, ar_tr, nxl_y_bound, nxr_y_bound, nxl_y_l, & |
---|
| 1034 | nxr_y_l ) |
---|
[1] | 1035 | |
---|
| 1036 | |
---|
[1106] | 1037 | USE cuda_fft_interfaces |
---|
[1153] | 1038 | #if defined( __cuda_fft ) |
---|
| 1039 | USE ISO_C_BINDING |
---|
| 1040 | #endif |
---|
[1106] | 1041 | |
---|
[1] | 1042 | IMPLICIT NONE |
---|
| 1043 | |
---|
[1682] | 1044 | CHARACTER (LEN=*) :: direction !< |
---|
[1320] | 1045 | |
---|
[1682] | 1046 | INTEGER(iwp) :: i !< |
---|
| 1047 | INTEGER(iwp) :: j !< |
---|
| 1048 | INTEGER(iwp) :: jshape(1) !< |
---|
| 1049 | INTEGER(iwp) :: k !< |
---|
| 1050 | INTEGER(iwp) :: nxl_y_bound !< |
---|
| 1051 | INTEGER(iwp) :: nxl_y_l !< |
---|
| 1052 | INTEGER(iwp) :: nxr_y_bound !< |
---|
| 1053 | INTEGER(iwp) :: nxr_y_l !< |
---|
[1106] | 1054 | |
---|
[1682] | 1055 | LOGICAL :: forward_fft !< |
---|
[1106] | 1056 | |
---|
[1682] | 1057 | REAL(wp), DIMENSION(0:ny+2) :: work !< |
---|
| 1058 | REAL(wp), DIMENSION(ny+2) :: work1 !< |
---|
[1320] | 1059 | |
---|
[1682] | 1060 | COMPLEX(wp), DIMENSION(:), ALLOCATABLE :: cwork !< |
---|
[1320] | 1061 | |
---|
[1106] | 1062 | #if defined( __ibm ) |
---|
[1682] | 1063 | REAL(wp), DIMENSION(nau2) :: auy2 !< |
---|
| 1064 | REAL(wp), DIMENSION(nau2) :: auy4 !< |
---|
[1106] | 1065 | #elif defined( __nec ) |
---|
[1682] | 1066 | REAL(wp), DIMENSION(6*(ny+1)) :: work2 !< |
---|
[1106] | 1067 | #elif defined( __cuda_fft ) |
---|
[1374] | 1068 | COMPLEX(dp), DIMENSION(0:(ny+1)/2,nxl_y:nxr_y,nzb_y:nzt_y) :: & |
---|
[1682] | 1069 | ar_tmp !< |
---|
[1482] | 1070 | ! following does not work for PGI 14.1 -> to be removed later |
---|
[1111] | 1071 | !$acc declare create( ar_tmp ) |
---|
[1106] | 1072 | #endif |
---|
| 1073 | |
---|
[1320] | 1074 | REAL(wp), DIMENSION(0:ny,nxl_y_l:nxr_y_l,nzb_y:nzt_y) :: & |
---|
[1682] | 1075 | ar !< |
---|
[1320] | 1076 | REAL(wp), DIMENSION(0:ny,nxl_y_bound:nxr_y_bound,nzb_y:nzt_y) :: & |
---|
[1682] | 1077 | ar_tr !< |
---|
[1320] | 1078 | |
---|
[1106] | 1079 | IF ( direction == 'forward' ) THEN |
---|
| 1080 | forward_fft = .TRUE. |
---|
| 1081 | ELSE |
---|
| 1082 | forward_fft = .FALSE. |
---|
| 1083 | ENDIF |
---|
| 1084 | |
---|
| 1085 | IF ( fft_method == 'singleton-algorithm' ) THEN |
---|
| 1086 | |
---|
| 1087 | ! |
---|
| 1088 | !-- Performing the fft with singleton's software works on every system, |
---|
| 1089 | !-- since it is part of the model |
---|
| 1090 | ALLOCATE( cwork(0:ny) ) |
---|
| 1091 | |
---|
| 1092 | IF ( forward_fft ) then |
---|
| 1093 | |
---|
| 1094 | !$OMP PARALLEL PRIVATE ( cwork, i, jshape, j, k ) |
---|
| 1095 | !$OMP DO |
---|
| 1096 | DO k = nzb_y, nzt_y |
---|
[1216] | 1097 | DO i = nxl_y_l, nxr_y_l |
---|
[1106] | 1098 | |
---|
| 1099 | DO j = 0, ny |
---|
[1392] | 1100 | cwork(j) = CMPLX( ar(j,i,k), KIND=wp ) |
---|
[1106] | 1101 | ENDDO |
---|
| 1102 | |
---|
| 1103 | jshape = SHAPE( cwork ) |
---|
| 1104 | CALL FFTN( cwork, jshape ) |
---|
| 1105 | |
---|
| 1106 | DO j = 0, (ny+1)/2 |
---|
[1322] | 1107 | ar_tr(j,i,k) = REAL( cwork(j), KIND=wp ) |
---|
[1106] | 1108 | ENDDO |
---|
| 1109 | DO j = 1, (ny+1)/2 - 1 |
---|
[1216] | 1110 | ar_tr(ny+1-j,i,k) = -AIMAG( cwork(j) ) |
---|
[1106] | 1111 | ENDDO |
---|
| 1112 | |
---|
| 1113 | ENDDO |
---|
| 1114 | ENDDO |
---|
| 1115 | !$OMP END PARALLEL |
---|
| 1116 | |
---|
| 1117 | ELSE |
---|
| 1118 | |
---|
| 1119 | !$OMP PARALLEL PRIVATE ( cwork, i, jshape, j, k ) |
---|
| 1120 | !$OMP DO |
---|
| 1121 | DO k = nzb_y, nzt_y |
---|
[1216] | 1122 | DO i = nxl_y_l, nxr_y_l |
---|
[1106] | 1123 | |
---|
[1392] | 1124 | cwork(0) = CMPLX( ar_tr(0,i,k), 0.0_wp, KIND=wp ) |
---|
[1106] | 1125 | DO j = 1, (ny+1)/2 - 1 |
---|
[1392] | 1126 | cwork(j) = CMPLX( ar_tr(j,i,k), -ar_tr(ny+1-j,i,k), & |
---|
| 1127 | KIND=wp ) |
---|
| 1128 | cwork(ny+1-j) = CMPLX( ar_tr(j,i,k), ar_tr(ny+1-j,i,k), & |
---|
| 1129 | KIND=wp ) |
---|
[1106] | 1130 | ENDDO |
---|
[1392] | 1131 | cwork((ny+1)/2) = CMPLX( ar_tr((ny+1)/2,i,k), 0.0_wp, & |
---|
| 1132 | KIND=wp ) |
---|
[1106] | 1133 | |
---|
| 1134 | jshape = SHAPE( cwork ) |
---|
| 1135 | CALL FFTN( cwork, jshape, inv = .TRUE. ) |
---|
| 1136 | |
---|
| 1137 | DO j = 0, ny |
---|
[1322] | 1138 | ar(j,i,k) = REAL( cwork(j), KIND=wp ) |
---|
[1106] | 1139 | ENDDO |
---|
| 1140 | |
---|
| 1141 | ENDDO |
---|
| 1142 | ENDDO |
---|
| 1143 | !$OMP END PARALLEL |
---|
| 1144 | |
---|
| 1145 | ENDIF |
---|
| 1146 | |
---|
| 1147 | DEALLOCATE( cwork ) |
---|
| 1148 | |
---|
| 1149 | ELSEIF ( fft_method == 'temperton-algorithm' ) THEN |
---|
| 1150 | |
---|
| 1151 | ! |
---|
| 1152 | !-- Performing the fft with Temperton's software works on every system, |
---|
| 1153 | !-- since it is part of the model |
---|
| 1154 | IF ( forward_fft ) THEN |
---|
| 1155 | |
---|
[1304] | 1156 | !$OMP PARALLEL PRIVATE ( work, work1, i, j, k ) |
---|
[1106] | 1157 | !$OMP DO |
---|
| 1158 | DO k = nzb_y, nzt_y |
---|
[1216] | 1159 | DO i = nxl_y_l, nxr_y_l |
---|
[1106] | 1160 | |
---|
| 1161 | work(0:ny) = ar(0:ny,i,k) |
---|
| 1162 | CALL fft991cy( work, work1, trigs_y, ifax_y, 1, ny+1, ny+1, 1, -1 ) |
---|
| 1163 | |
---|
| 1164 | DO j = 0, (ny+1)/2 |
---|
[1216] | 1165 | ar_tr(j,i,k) = work(2*j) |
---|
[1106] | 1166 | ENDDO |
---|
| 1167 | DO j = 1, (ny+1)/2 - 1 |
---|
[1216] | 1168 | ar_tr(ny+1-j,i,k) = work(2*j+1) |
---|
[1106] | 1169 | ENDDO |
---|
| 1170 | |
---|
| 1171 | ENDDO |
---|
| 1172 | ENDDO |
---|
| 1173 | !$OMP END PARALLEL |
---|
| 1174 | |
---|
| 1175 | ELSE |
---|
| 1176 | |
---|
[1304] | 1177 | !$OMP PARALLEL PRIVATE ( work, work1, i, j, k ) |
---|
[1106] | 1178 | !$OMP DO |
---|
| 1179 | DO k = nzb_y, nzt_y |
---|
[1216] | 1180 | DO i = nxl_y_l, nxr_y_l |
---|
[1106] | 1181 | |
---|
| 1182 | DO j = 0, (ny+1)/2 |
---|
[1216] | 1183 | work(2*j) = ar_tr(j,i,k) |
---|
[1106] | 1184 | ENDDO |
---|
| 1185 | DO j = 1, (ny+1)/2 - 1 |
---|
[1216] | 1186 | work(2*j+1) = ar_tr(ny+1-j,i,k) |
---|
[1106] | 1187 | ENDDO |
---|
[1342] | 1188 | work(1) = 0.0_wp |
---|
| 1189 | work(ny+2) = 0.0_wp |
---|
[1106] | 1190 | |
---|
| 1191 | CALL fft991cy( work, work1, trigs_y, ifax_y, 1, ny+1, ny+1, 1, 1 ) |
---|
| 1192 | ar(0:ny,i,k) = work(0:ny) |
---|
| 1193 | |
---|
| 1194 | ENDDO |
---|
| 1195 | ENDDO |
---|
| 1196 | !$OMP END PARALLEL |
---|
| 1197 | |
---|
| 1198 | ENDIF |
---|
| 1199 | |
---|
[1210] | 1200 | ELSEIF ( fft_method == 'fftw' ) THEN |
---|
| 1201 | |
---|
| 1202 | #if defined( __fftw ) |
---|
| 1203 | IF ( forward_fft ) THEN |
---|
| 1204 | |
---|
| 1205 | !$OMP PARALLEL PRIVATE ( work, i, j, k ) |
---|
| 1206 | !$OMP DO |
---|
| 1207 | DO k = nzb_y, nzt_y |
---|
[1216] | 1208 | DO i = nxl_y_l, nxr_y_l |
---|
[1210] | 1209 | |
---|
| 1210 | y_in(0:ny) = ar(0:ny,i,k) |
---|
| 1211 | CALL FFTW_EXECUTE_DFT_R2C( plan_yf, y_in, y_out ) |
---|
| 1212 | |
---|
| 1213 | DO j = 0, (ny+1)/2 |
---|
[1322] | 1214 | ar_tr(j,i,k) = REAL( y_out(j), KIND=wp ) / (ny+1) |
---|
[1210] | 1215 | ENDDO |
---|
| 1216 | DO j = 1, (ny+1)/2 - 1 |
---|
[1216] | 1217 | ar_tr(ny+1-j,i,k) = AIMAG( y_out(j) ) / (ny+1) |
---|
[1210] | 1218 | ENDDO |
---|
| 1219 | |
---|
| 1220 | ENDDO |
---|
| 1221 | ENDDO |
---|
| 1222 | !$OMP END PARALLEL |
---|
| 1223 | |
---|
| 1224 | ELSE |
---|
| 1225 | |
---|
| 1226 | !$OMP PARALLEL PRIVATE ( work, i, j, k ) |
---|
| 1227 | !$OMP DO |
---|
| 1228 | DO k = nzb_y, nzt_y |
---|
[1216] | 1229 | DO i = nxl_y_l, nxr_y_l |
---|
[1210] | 1230 | |
---|
[1392] | 1231 | y_out(0) = CMPLX( ar_tr(0,i,k), 0.0_wp, KIND=wp ) |
---|
[1210] | 1232 | DO j = 1, (ny+1)/2 - 1 |
---|
[1398] | 1233 | y_out(j) = CMPLX( ar_tr(j,i,k), ar_tr(ny+1-j,i,k), & |
---|
| 1234 | KIND=wp ) |
---|
[1210] | 1235 | ENDDO |
---|
[1392] | 1236 | y_out((ny+1)/2) = CMPLX( ar_tr((ny+1)/2,i,k), 0.0_wp, & |
---|
| 1237 | KIND=wp ) |
---|
[1210] | 1238 | |
---|
| 1239 | CALL FFTW_EXECUTE_DFT_C2R( plan_yi, y_out, y_in ) |
---|
| 1240 | ar(0:ny,i,k) = y_in(0:ny) |
---|
| 1241 | |
---|
| 1242 | ENDDO |
---|
| 1243 | ENDDO |
---|
| 1244 | !$OMP END PARALLEL |
---|
| 1245 | |
---|
| 1246 | ENDIF |
---|
| 1247 | #endif |
---|
| 1248 | |
---|
[1106] | 1249 | ELSEIF ( fft_method == 'system-specific' ) THEN |
---|
| 1250 | |
---|
| 1251 | #if defined( __ibm ) && ! defined( __ibmy_special ) |
---|
| 1252 | IF ( forward_fft) THEN |
---|
| 1253 | |
---|
| 1254 | !$OMP PARALLEL PRIVATE ( work, i, j, k ) |
---|
| 1255 | !$OMP DO |
---|
| 1256 | DO k = nzb_y, nzt_y |
---|
[1216] | 1257 | DO i = nxl_y_l, nxr_y_l |
---|
[1106] | 1258 | |
---|
[1320] | 1259 | CALL DRCFT( 0, ar, 1, work, 1, ny+1, 1, 1, sqr_dny, auy1, & |
---|
| 1260 | nau1, auy2, nau2 ) |
---|
[1106] | 1261 | |
---|
| 1262 | DO j = 0, (ny+1)/2 |
---|
[1216] | 1263 | ar_tr(j,i,k) = work(2*j) |
---|
[1106] | 1264 | ENDDO |
---|
| 1265 | DO j = 1, (ny+1)/2 - 1 |
---|
[1216] | 1266 | ar_tr(ny+1-j,i,k) = work(2*j+1) |
---|
[1106] | 1267 | ENDDO |
---|
| 1268 | |
---|
| 1269 | ENDDO |
---|
| 1270 | ENDDO |
---|
| 1271 | !$OMP END PARALLEL |
---|
| 1272 | |
---|
| 1273 | ELSE |
---|
| 1274 | |
---|
| 1275 | !$OMP PARALLEL PRIVATE ( work, i, j, k ) |
---|
| 1276 | !$OMP DO |
---|
| 1277 | DO k = nzb_y, nzt_y |
---|
[1216] | 1278 | DO i = nxl_y_l, nxr_y_l |
---|
[1106] | 1279 | |
---|
| 1280 | DO j = 0, (ny+1)/2 |
---|
[1216] | 1281 | work(2*j) = ar_tr(j,i,k) |
---|
[1106] | 1282 | ENDDO |
---|
| 1283 | DO j = 1, (ny+1)/2 - 1 |
---|
[1216] | 1284 | work(2*j+1) = ar_tr(ny+1-j,i,k) |
---|
[1106] | 1285 | ENDDO |
---|
[1342] | 1286 | work(1) = 0.0_wp |
---|
| 1287 | work(ny+2) = 0.0_wp |
---|
[1106] | 1288 | |
---|
[1320] | 1289 | CALL DCRFT( 0, work, 1, work, 1, ny+1, 1, -1, sqr_dny, & |
---|
| 1290 | auy3, nau1, auy4, nau2 ) |
---|
[1106] | 1291 | |
---|
| 1292 | DO j = 0, ny |
---|
| 1293 | ar(j,i,k) = work(j) |
---|
| 1294 | ENDDO |
---|
| 1295 | |
---|
| 1296 | ENDDO |
---|
| 1297 | ENDDO |
---|
| 1298 | !$OMP END PARALLEL |
---|
| 1299 | |
---|
| 1300 | ENDIF |
---|
| 1301 | #elif defined( __nec ) |
---|
| 1302 | IF ( forward_fft ) THEN |
---|
| 1303 | |
---|
| 1304 | !$OMP PARALLEL PRIVATE ( work, i, j, k ) |
---|
| 1305 | !$OMP DO |
---|
| 1306 | DO k = nzb_y, nzt_y |
---|
[1216] | 1307 | DO i = nxl_y_l, nxr_y_l |
---|
[1106] | 1308 | |
---|
| 1309 | work(0:ny) = ar(0:ny,i,k) |
---|
| 1310 | |
---|
| 1311 | CALL DZFFT( 1, ny+1, sqr_dny, work, work, trig_yf, work2, 0 ) |
---|
| 1312 | |
---|
| 1313 | DO j = 0, (ny+1)/2 |
---|
[1216] | 1314 | ar_tr(j,i,k) = work(2*j) |
---|
[1106] | 1315 | ENDDO |
---|
| 1316 | DO j = 1, (ny+1)/2 - 1 |
---|
[1216] | 1317 | ar_tr(ny+1-j,i,k) = work(2*j+1) |
---|
[1106] | 1318 | ENDDO |
---|
| 1319 | |
---|
| 1320 | ENDDO |
---|
| 1321 | ENDDO |
---|
| 1322 | !$END OMP PARALLEL |
---|
| 1323 | |
---|
| 1324 | ELSE |
---|
| 1325 | |
---|
| 1326 | !$OMP PARALLEL PRIVATE ( work, i, j, k ) |
---|
| 1327 | !$OMP DO |
---|
| 1328 | DO k = nzb_y, nzt_y |
---|
[1216] | 1329 | DO i = nxl_y_l, nxr_y_l |
---|
[1106] | 1330 | |
---|
| 1331 | DO j = 0, (ny+1)/2 |
---|
[1216] | 1332 | work(2*j) = ar_tr(j,i,k) |
---|
[1106] | 1333 | ENDDO |
---|
| 1334 | DO j = 1, (ny+1)/2 - 1 |
---|
[1216] | 1335 | work(2*j+1) = ar_tr(ny+1-j,i,k) |
---|
[1106] | 1336 | ENDDO |
---|
[1342] | 1337 | work(1) = 0.0_wp |
---|
| 1338 | work(ny+2) = 0.0_wp |
---|
[1106] | 1339 | |
---|
| 1340 | CALL ZDFFT( -1, ny+1, sqr_dny, work, work, trig_yb, work2, 0 ) |
---|
| 1341 | |
---|
| 1342 | ar(0:ny,i,k) = work(0:ny) |
---|
| 1343 | |
---|
| 1344 | ENDDO |
---|
| 1345 | ENDDO |
---|
| 1346 | !$OMP END PARALLEL |
---|
| 1347 | |
---|
| 1348 | ENDIF |
---|
| 1349 | #elif defined( __cuda_fft ) |
---|
| 1350 | |
---|
[1482] | 1351 | !$acc data create( ar_tmp ) |
---|
[1106] | 1352 | IF ( forward_fft ) THEN |
---|
| 1353 | |
---|
[1111] | 1354 | !$acc data present( ar ) |
---|
| 1355 | CALL CUFFTEXECD2Z( plan_yf, ar, ar_tmp ) |
---|
[1106] | 1356 | |
---|
[1111] | 1357 | !$acc kernels |
---|
[1106] | 1358 | DO k = nzb_y, nzt_y |
---|
| 1359 | DO i = nxl_y, nxr_y |
---|
| 1360 | |
---|
| 1361 | DO j = 0, (ny+1)/2 |
---|
[1322] | 1362 | ar(j,i,k) = REAL( ar_tmp(j,i,k), KIND=wp ) * dny |
---|
[1106] | 1363 | ENDDO |
---|
| 1364 | |
---|
| 1365 | DO j = 1, (ny+1)/2 - 1 |
---|
[1111] | 1366 | ar(ny+1-j,i,k) = AIMAG( ar_tmp(j,i,k) ) * dny |
---|
[1106] | 1367 | ENDDO |
---|
| 1368 | |
---|
| 1369 | ENDDO |
---|
| 1370 | ENDDO |
---|
[1111] | 1371 | !$acc end kernels |
---|
| 1372 | !$acc end data |
---|
[1106] | 1373 | |
---|
| 1374 | ELSE |
---|
| 1375 | |
---|
[1111] | 1376 | !$acc data present( ar ) |
---|
| 1377 | !$acc kernels |
---|
[1106] | 1378 | DO k = nzb_y, nzt_y |
---|
| 1379 | DO i = nxl_y, nxr_y |
---|
| 1380 | |
---|
[1392] | 1381 | ar_tmp(0,i,k) = CMPLX( ar(0,i,k), 0.0_wp, KIND=wp ) |
---|
[1106] | 1382 | |
---|
| 1383 | DO j = 1, (ny+1)/2 - 1 |
---|
[1392] | 1384 | ar_tmp(j,i,k) = CMPLX( ar(j,i,k), ar(ny+1-j,i,k), & |
---|
| 1385 | KIND=wp ) |
---|
[1106] | 1386 | ENDDO |
---|
[1392] | 1387 | ar_tmp((ny+1)/2,i,k) = CMPLX( ar((ny+1)/2,i,k), 0.0_wp, & |
---|
| 1388 | KIND=wp ) |
---|
[1106] | 1389 | |
---|
| 1390 | ENDDO |
---|
| 1391 | ENDDO |
---|
[1111] | 1392 | !$acc end kernels |
---|
[1106] | 1393 | |
---|
[1111] | 1394 | CALL CUFFTEXECZ2D( plan_yi, ar_tmp, ar ) |
---|
| 1395 | !$acc end data |
---|
[1106] | 1396 | |
---|
| 1397 | ENDIF |
---|
[1482] | 1398 | !$acc end data |
---|
[1106] | 1399 | |
---|
| 1400 | #else |
---|
| 1401 | message_string = 'no system-specific fft-call available' |
---|
| 1402 | CALL message( 'fft_y', 'PA0188', 1, 2, 0, 6, 0 ) |
---|
| 1403 | #endif |
---|
| 1404 | |
---|
| 1405 | ELSE |
---|
| 1406 | |
---|
| 1407 | message_string = 'fft method "' // TRIM( fft_method) // & |
---|
| 1408 | '" not available' |
---|
| 1409 | CALL message( 'fft_y', 'PA0189', 1, 2, 0, 6, 0 ) |
---|
| 1410 | |
---|
| 1411 | ENDIF |
---|
| 1412 | |
---|
| 1413 | END SUBROUTINE fft_y |
---|
| 1414 | |
---|
[1682] | 1415 | !------------------------------------------------------------------------------! |
---|
| 1416 | ! Description: |
---|
| 1417 | ! ------------ |
---|
| 1418 | !> Fourier-transformation along y-direction. |
---|
| 1419 | !> Version for 1D-decomposition. |
---|
| 1420 | !> It uses internal algorithms (Singleton or Temperton) or |
---|
| 1421 | !> system-specific routines, if they are available. |
---|
| 1422 | !------------------------------------------------------------------------------! |
---|
| 1423 | |
---|
[1106] | 1424 | SUBROUTINE fft_y_1d( ar, direction ) |
---|
| 1425 | |
---|
| 1426 | |
---|
| 1427 | IMPLICIT NONE |
---|
| 1428 | |
---|
| 1429 | CHARACTER (LEN=*) :: direction |
---|
[1320] | 1430 | |
---|
[1682] | 1431 | INTEGER(iwp) :: j !< |
---|
| 1432 | INTEGER(iwp) :: jshape(1) !< |
---|
[1] | 1433 | |
---|
[1682] | 1434 | LOGICAL :: forward_fft !< |
---|
[1106] | 1435 | |
---|
[1682] | 1436 | REAL(wp), DIMENSION(0:ny) :: ar !< |
---|
| 1437 | REAL(wp), DIMENSION(0:ny+2) :: work !< |
---|
| 1438 | REAL(wp), DIMENSION(ny+2) :: work1 !< |
---|
[1320] | 1439 | |
---|
[1682] | 1440 | COMPLEX(wp), DIMENSION(:), ALLOCATABLE :: cwork !< |
---|
[1320] | 1441 | |
---|
[1] | 1442 | #if defined( __ibm ) |
---|
[1682] | 1443 | REAL(wp), DIMENSION(nau2) :: auy2 !< |
---|
| 1444 | REAL(wp), DIMENSION(nau2) :: auy4 !< |
---|
[1] | 1445 | #elif defined( __nec ) |
---|
[1682] | 1446 | REAL(wp), DIMENSION(6*(ny+1)) :: work2 !< |
---|
[1] | 1447 | #endif |
---|
| 1448 | |
---|
[1106] | 1449 | IF ( direction == 'forward' ) THEN |
---|
| 1450 | forward_fft = .TRUE. |
---|
| 1451 | ELSE |
---|
| 1452 | forward_fft = .FALSE. |
---|
| 1453 | ENDIF |
---|
| 1454 | |
---|
[1] | 1455 | IF ( fft_method == 'singleton-algorithm' ) THEN |
---|
| 1456 | |
---|
| 1457 | ! |
---|
| 1458 | !-- Performing the fft with singleton's software works on every system, |
---|
| 1459 | !-- since it is part of the model |
---|
| 1460 | ALLOCATE( cwork(0:ny) ) |
---|
| 1461 | |
---|
[1106] | 1462 | IF ( forward_fft ) THEN |
---|
[1] | 1463 | |
---|
| 1464 | DO j = 0, ny |
---|
[1392] | 1465 | cwork(j) = CMPLX( ar(j), KIND=wp ) |
---|
[1] | 1466 | ENDDO |
---|
| 1467 | |
---|
| 1468 | jshape = SHAPE( cwork ) |
---|
| 1469 | CALL FFTN( cwork, jshape ) |
---|
| 1470 | |
---|
| 1471 | DO j = 0, (ny+1)/2 |
---|
[1322] | 1472 | ar(j) = REAL( cwork(j), KIND=wp ) |
---|
[1] | 1473 | ENDDO |
---|
| 1474 | DO j = 1, (ny+1)/2 - 1 |
---|
| 1475 | ar(ny+1-j) = -AIMAG( cwork(j) ) |
---|
| 1476 | ENDDO |
---|
| 1477 | |
---|
| 1478 | ELSE |
---|
| 1479 | |
---|
[1392] | 1480 | cwork(0) = CMPLX( ar(0), 0.0_wp, KIND=wp ) |
---|
[1] | 1481 | DO j = 1, (ny+1)/2 - 1 |
---|
[1392] | 1482 | cwork(j) = CMPLX( ar(j), -ar(ny+1-j), KIND=wp ) |
---|
| 1483 | cwork(ny+1-j) = CMPLX( ar(j), ar(ny+1-j), KIND=wp ) |
---|
[1] | 1484 | ENDDO |
---|
[1392] | 1485 | cwork((ny+1)/2) = CMPLX( ar((ny+1)/2), 0.0_wp, KIND=wp ) |
---|
[1] | 1486 | |
---|
| 1487 | jshape = SHAPE( cwork ) |
---|
| 1488 | CALL FFTN( cwork, jshape, inv = .TRUE. ) |
---|
| 1489 | |
---|
| 1490 | DO j = 0, ny |
---|
[1322] | 1491 | ar(j) = REAL( cwork(j), KIND=wp ) |
---|
[1] | 1492 | ENDDO |
---|
| 1493 | |
---|
| 1494 | ENDIF |
---|
| 1495 | |
---|
| 1496 | DEALLOCATE( cwork ) |
---|
| 1497 | |
---|
| 1498 | ELSEIF ( fft_method == 'temperton-algorithm' ) THEN |
---|
| 1499 | |
---|
| 1500 | ! |
---|
| 1501 | !-- Performing the fft with Temperton's software works on every system, |
---|
| 1502 | !-- since it is part of the model |
---|
[1106] | 1503 | IF ( forward_fft ) THEN |
---|
[1] | 1504 | |
---|
| 1505 | work(0:ny) = ar |
---|
| 1506 | CALL fft991cy( work, work1, trigs_y, ifax_y, 1, ny+1, ny+1, 1, -1 ) |
---|
| 1507 | |
---|
| 1508 | DO j = 0, (ny+1)/2 |
---|
| 1509 | ar(j) = work(2*j) |
---|
| 1510 | ENDDO |
---|
| 1511 | DO j = 1, (ny+1)/2 - 1 |
---|
| 1512 | ar(ny+1-j) = work(2*j+1) |
---|
| 1513 | ENDDO |
---|
| 1514 | |
---|
| 1515 | ELSE |
---|
| 1516 | |
---|
| 1517 | DO j = 0, (ny+1)/2 |
---|
| 1518 | work(2*j) = ar(j) |
---|
| 1519 | ENDDO |
---|
| 1520 | DO j = 1, (ny+1)/2 - 1 |
---|
| 1521 | work(2*j+1) = ar(ny+1-j) |
---|
| 1522 | ENDDO |
---|
[1342] | 1523 | work(1) = 0.0_wp |
---|
| 1524 | work(ny+2) = 0.0_wp |
---|
[1] | 1525 | |
---|
| 1526 | CALL fft991cy( work, work1, trigs_y, ifax_y, 1, ny+1, ny+1, 1, 1 ) |
---|
| 1527 | ar = work(0:ny) |
---|
| 1528 | |
---|
| 1529 | ENDIF |
---|
| 1530 | |
---|
[1216] | 1531 | ELSEIF ( fft_method == 'fftw' ) THEN |
---|
| 1532 | |
---|
| 1533 | #if defined( __fftw ) |
---|
| 1534 | IF ( forward_fft ) THEN |
---|
| 1535 | |
---|
| 1536 | y_in(0:ny) = ar(0:ny) |
---|
| 1537 | CALL FFTW_EXECUTE_DFT_R2C( plan_yf, y_in, y_out ) |
---|
| 1538 | |
---|
| 1539 | DO j = 0, (ny+1)/2 |
---|
[1322] | 1540 | ar(j) = REAL( y_out(j), KIND=wp ) / (ny+1) |
---|
[1216] | 1541 | ENDDO |
---|
| 1542 | DO j = 1, (ny+1)/2 - 1 |
---|
| 1543 | ar(ny+1-j) = AIMAG( y_out(j) ) / (ny+1) |
---|
| 1544 | ENDDO |
---|
| 1545 | |
---|
| 1546 | ELSE |
---|
| 1547 | |
---|
[1392] | 1548 | y_out(0) = CMPLX( ar(0), 0.0_wp, KIND=wp ) |
---|
[1216] | 1549 | DO j = 1, (ny+1)/2 - 1 |
---|
[1392] | 1550 | y_out(j) = CMPLX( ar(j), ar(ny+1-j), KIND=wp ) |
---|
[1216] | 1551 | ENDDO |
---|
[1392] | 1552 | y_out((ny+1)/2) = CMPLX( ar((ny+1)/2), 0.0_wp, KIND=wp ) |
---|
[1216] | 1553 | |
---|
| 1554 | CALL FFTW_EXECUTE_DFT_C2R( plan_yi, y_out, y_in ) |
---|
| 1555 | ar(0:ny) = y_in(0:ny) |
---|
| 1556 | |
---|
| 1557 | ENDIF |
---|
| 1558 | #endif |
---|
| 1559 | |
---|
[1] | 1560 | ELSEIF ( fft_method == 'system-specific' ) THEN |
---|
| 1561 | |
---|
| 1562 | #if defined( __ibm ) && ! defined( __ibmy_special ) |
---|
[1106] | 1563 | IF ( forward_fft ) THEN |
---|
[1] | 1564 | |
---|
[1320] | 1565 | CALL DRCFT( 0, ar, 1, work, 1, ny+1, 1, 1, sqr_dny, auy1, nau1, & |
---|
[1] | 1566 | auy2, nau2 ) |
---|
| 1567 | |
---|
| 1568 | DO j = 0, (ny+1)/2 |
---|
| 1569 | ar(j) = work(2*j) |
---|
| 1570 | ENDDO |
---|
| 1571 | DO j = 1, (ny+1)/2 - 1 |
---|
| 1572 | ar(ny+1-j) = work(2*j+1) |
---|
| 1573 | ENDDO |
---|
| 1574 | |
---|
| 1575 | ELSE |
---|
| 1576 | |
---|
| 1577 | DO j = 0, (ny+1)/2 |
---|
| 1578 | work(2*j) = ar(j) |
---|
| 1579 | ENDDO |
---|
| 1580 | DO j = 1, (ny+1)/2 - 1 |
---|
| 1581 | work(2*j+1) = ar(ny+1-j) |
---|
| 1582 | ENDDO |
---|
[1342] | 1583 | work(1) = 0.0_wp |
---|
| 1584 | work(ny+2) = 0.0_wp |
---|
[1] | 1585 | |
---|
[1320] | 1586 | CALL DCRFT( 0, work, 1, work, 1, ny+1, 1, -1, sqr_dny, auy3, & |
---|
| 1587 | nau1, auy4, nau2 ) |
---|
[1] | 1588 | |
---|
| 1589 | DO j = 0, ny |
---|
| 1590 | ar(j) = work(j) |
---|
| 1591 | ENDDO |
---|
| 1592 | |
---|
| 1593 | ENDIF |
---|
| 1594 | #elif defined( __nec ) |
---|
[1106] | 1595 | IF ( forward_fft ) THEN |
---|
[1] | 1596 | |
---|
| 1597 | work(0:ny) = ar(0:ny) |
---|
| 1598 | |
---|
[1106] | 1599 | CALL DZFFT( 1, ny+1, sqr_dny, work, work, trig_yf, work2, 0 ) |
---|
[1] | 1600 | |
---|
| 1601 | DO j = 0, (ny+1)/2 |
---|
| 1602 | ar(j) = work(2*j) |
---|
| 1603 | ENDDO |
---|
| 1604 | DO j = 1, (ny+1)/2 - 1 |
---|
| 1605 | ar(ny+1-j) = work(2*j+1) |
---|
| 1606 | ENDDO |
---|
| 1607 | |
---|
| 1608 | ELSE |
---|
| 1609 | |
---|
| 1610 | DO j = 0, (ny+1)/2 |
---|
| 1611 | work(2*j) = ar(j) |
---|
| 1612 | ENDDO |
---|
| 1613 | DO j = 1, (ny+1)/2 - 1 |
---|
| 1614 | work(2*j+1) = ar(ny+1-j) |
---|
| 1615 | ENDDO |
---|
[1342] | 1616 | work(1) = 0.0_wp |
---|
| 1617 | work(ny+2) = 0.0_wp |
---|
[1] | 1618 | |
---|
[1106] | 1619 | CALL ZDFFT( -1, ny+1, sqr_dny, work, work, trig_yb, work2, 0 ) |
---|
[1] | 1620 | |
---|
| 1621 | ar(0:ny) = work(0:ny) |
---|
| 1622 | |
---|
| 1623 | ENDIF |
---|
| 1624 | #else |
---|
[254] | 1625 | message_string = 'no system-specific fft-call available' |
---|
[1106] | 1626 | CALL message( 'fft_y_1d', 'PA0188', 1, 2, 0, 6, 0 ) |
---|
[254] | 1627 | |
---|
[1] | 1628 | #endif |
---|
| 1629 | |
---|
| 1630 | ELSE |
---|
| 1631 | |
---|
[274] | 1632 | message_string = 'fft method "' // TRIM( fft_method) // & |
---|
| 1633 | '" not available' |
---|
[1106] | 1634 | CALL message( 'fft_y_1d', 'PA0189', 1, 2, 0, 6, 0 ) |
---|
[1] | 1635 | |
---|
| 1636 | ENDIF |
---|
| 1637 | |
---|
[1106] | 1638 | END SUBROUTINE fft_y_1d |
---|
[1] | 1639 | |
---|
[1682] | 1640 | !------------------------------------------------------------------------------! |
---|
| 1641 | ! Description: |
---|
| 1642 | ! ------------ |
---|
| 1643 | !> Fourier-transformation along x-direction. |
---|
| 1644 | !> Version for 1d domain decomposition |
---|
| 1645 | !> using multiple 1D FFT from Math Keisan on NEC or Temperton-algorithm |
---|
| 1646 | !> (no singleton-algorithm on NEC because it does not vectorize) |
---|
| 1647 | !------------------------------------------------------------------------------! |
---|
| 1648 | |
---|
[1] | 1649 | SUBROUTINE fft_x_m( ar, direction ) |
---|
| 1650 | |
---|
| 1651 | |
---|
| 1652 | IMPLICIT NONE |
---|
| 1653 | |
---|
[1682] | 1654 | CHARACTER (LEN=*) :: direction !< |
---|
[1320] | 1655 | |
---|
[1682] | 1656 | INTEGER(iwp) :: i !< |
---|
| 1657 | INTEGER(iwp) :: k !< |
---|
| 1658 | INTEGER(iwp) :: siza !< |
---|
[1] | 1659 | |
---|
[1682] | 1660 | REAL(wp), DIMENSION(0:nx,nz) :: ar !< |
---|
| 1661 | REAL(wp), DIMENSION(0:nx+3,nz+1) :: ai !< |
---|
| 1662 | REAL(wp), DIMENSION(6*(nx+4),nz+1) :: work1 !< |
---|
[1320] | 1663 | |
---|
[1] | 1664 | #if defined( __nec ) |
---|
[1682] | 1665 | INTEGER(iwp) :: sizw !< |
---|
[1320] | 1666 | |
---|
[1682] | 1667 | COMPLEX(wp), DIMENSION((nx+4)/2+1,nz+1) :: work !< |
---|
[1] | 1668 | #endif |
---|
| 1669 | |
---|
| 1670 | IF ( fft_method == 'temperton-algorithm' ) THEN |
---|
| 1671 | |
---|
| 1672 | siza = SIZE( ai, 1 ) |
---|
| 1673 | |
---|
| 1674 | IF ( direction == 'forward') THEN |
---|
| 1675 | |
---|
| 1676 | ai(0:nx,1:nz) = ar(0:nx,1:nz) |
---|
[1342] | 1677 | ai(nx+1:,:) = 0.0_wp |
---|
[1] | 1678 | |
---|
| 1679 | CALL fft991cy( ai, work1, trigs_x, ifax_x, 1, siza, nx+1, nz, -1 ) |
---|
| 1680 | |
---|
| 1681 | DO k = 1, nz |
---|
| 1682 | DO i = 0, (nx+1)/2 |
---|
| 1683 | ar(i,k) = ai(2*i,k) |
---|
| 1684 | ENDDO |
---|
| 1685 | DO i = 1, (nx+1)/2 - 1 |
---|
| 1686 | ar(nx+1-i,k) = ai(2*i+1,k) |
---|
| 1687 | ENDDO |
---|
| 1688 | ENDDO |
---|
| 1689 | |
---|
| 1690 | ELSE |
---|
| 1691 | |
---|
| 1692 | DO k = 1, nz |
---|
| 1693 | DO i = 0, (nx+1)/2 |
---|
| 1694 | ai(2*i,k) = ar(i,k) |
---|
| 1695 | ENDDO |
---|
| 1696 | DO i = 1, (nx+1)/2 - 1 |
---|
| 1697 | ai(2*i+1,k) = ar(nx+1-i,k) |
---|
| 1698 | ENDDO |
---|
[1342] | 1699 | ai(1,k) = 0.0_wp |
---|
| 1700 | ai(nx+2,k) = 0.0_wp |
---|
[1] | 1701 | ENDDO |
---|
| 1702 | |
---|
| 1703 | CALL fft991cy( ai, work1, trigs_x, ifax_x, 1, siza, nx+1, nz, 1 ) |
---|
| 1704 | |
---|
| 1705 | ar(0:nx,1:nz) = ai(0:nx,1:nz) |
---|
| 1706 | |
---|
| 1707 | ENDIF |
---|
| 1708 | |
---|
| 1709 | ELSEIF ( fft_method == 'system-specific' ) THEN |
---|
| 1710 | |
---|
| 1711 | #if defined( __nec ) |
---|
| 1712 | siza = SIZE( ai, 1 ) |
---|
| 1713 | sizw = SIZE( work, 1 ) |
---|
| 1714 | |
---|
| 1715 | IF ( direction == 'forward') THEN |
---|
| 1716 | |
---|
| 1717 | ! |
---|
| 1718 | !-- Tables are initialized once more. This call should not be |
---|
| 1719 | !-- necessary, but otherwise program aborts in asymmetric case |
---|
[1320] | 1720 | CALL DZFFTM( 0, nx+1, nz1, sqr_dnx, work, nx+4, work, nx+4, & |
---|
[1] | 1721 | trig_xf, work1, 0 ) |
---|
| 1722 | |
---|
| 1723 | ai(0:nx,1:nz) = ar(0:nx,1:nz) |
---|
| 1724 | IF ( nz1 > nz ) THEN |
---|
[1342] | 1725 | ai(:,nz1) = 0.0_wp |
---|
[1] | 1726 | ENDIF |
---|
| 1727 | |
---|
[1320] | 1728 | CALL DZFFTM( 1, nx+1, nz1, sqr_dnx, ai, siza, work, sizw, & |
---|
[1] | 1729 | trig_xf, work1, 0 ) |
---|
| 1730 | |
---|
| 1731 | DO k = 1, nz |
---|
| 1732 | DO i = 0, (nx+1)/2 |
---|
[1322] | 1733 | ar(i,k) = REAL( work(i+1,k), KIND=wp ) |
---|
[1] | 1734 | ENDDO |
---|
| 1735 | DO i = 1, (nx+1)/2 - 1 |
---|
| 1736 | ar(nx+1-i,k) = AIMAG( work(i+1,k) ) |
---|
| 1737 | ENDDO |
---|
| 1738 | ENDDO |
---|
| 1739 | |
---|
| 1740 | ELSE |
---|
| 1741 | |
---|
| 1742 | ! |
---|
| 1743 | !-- Tables are initialized once more. This call should not be |
---|
| 1744 | !-- necessary, but otherwise program aborts in asymmetric case |
---|
[1320] | 1745 | CALL ZDFFTM( 0, nx+1, nz1, sqr_dnx, work, nx+4, work, nx+4, & |
---|
[1] | 1746 | trig_xb, work1, 0 ) |
---|
| 1747 | |
---|
| 1748 | IF ( nz1 > nz ) THEN |
---|
[1342] | 1749 | work(:,nz1) = 0.0_wp |
---|
[1] | 1750 | ENDIF |
---|
| 1751 | DO k = 1, nz |
---|
[1392] | 1752 | work(1,k) = CMPLX( ar(0,k), 0.0_wp, KIND=wp ) |
---|
[1] | 1753 | DO i = 1, (nx+1)/2 - 1 |
---|
[1392] | 1754 | work(i+1,k) = CMPLX( ar(i,k), ar(nx+1-i,k), KIND=wp ) |
---|
[1] | 1755 | ENDDO |
---|
[1392] | 1756 | work(((nx+1)/2)+1,k) = CMPLX( ar((nx+1)/2,k), 0.0_wp, KIND=wp ) |
---|
[1] | 1757 | ENDDO |
---|
| 1758 | |
---|
[1106] | 1759 | CALL ZDFFTM( -1, nx+1, nz1, sqr_dnx, work, sizw, ai, siza, & |
---|
[1] | 1760 | trig_xb, work1, 0 ) |
---|
| 1761 | |
---|
| 1762 | ar(0:nx,1:nz) = ai(0:nx,1:nz) |
---|
| 1763 | |
---|
| 1764 | ENDIF |
---|
| 1765 | |
---|
| 1766 | #else |
---|
[254] | 1767 | message_string = 'no system-specific fft-call available' |
---|
| 1768 | CALL message( 'fft_x_m', 'PA0188', 1, 2, 0, 6, 0 ) |
---|
[1] | 1769 | #endif |
---|
| 1770 | |
---|
| 1771 | ELSE |
---|
| 1772 | |
---|
[274] | 1773 | message_string = 'fft method "' // TRIM( fft_method) // & |
---|
| 1774 | '" not available' |
---|
[254] | 1775 | CALL message( 'fft_x_m', 'PA0189', 1, 2, 0, 6, 0 ) |
---|
[1] | 1776 | |
---|
| 1777 | ENDIF |
---|
| 1778 | |
---|
| 1779 | END SUBROUTINE fft_x_m |
---|
| 1780 | |
---|
[1682] | 1781 | !------------------------------------------------------------------------------! |
---|
| 1782 | ! Description: |
---|
| 1783 | ! ------------ |
---|
| 1784 | !> Fourier-transformation along y-direction. |
---|
| 1785 | !> Version for 1d domain decomposition |
---|
| 1786 | !> using multiple 1D FFT from Math Keisan on NEC or Temperton-algorithm |
---|
| 1787 | !> (no singleton-algorithm on NEC because it does not vectorize) |
---|
| 1788 | !------------------------------------------------------------------------------! |
---|
| 1789 | |
---|
[1] | 1790 | SUBROUTINE fft_y_m( ar, ny1, direction ) |
---|
| 1791 | |
---|
| 1792 | |
---|
| 1793 | IMPLICIT NONE |
---|
| 1794 | |
---|
[1682] | 1795 | CHARACTER (LEN=*) :: direction !< |
---|
[1320] | 1796 | |
---|
[1682] | 1797 | INTEGER(iwp) :: j !< |
---|
| 1798 | INTEGER(iwp) :: k !< |
---|
| 1799 | INTEGER(iwp) :: ny1 !< |
---|
| 1800 | INTEGER(iwp) :: siza !< |
---|
[1] | 1801 | |
---|
[1682] | 1802 | REAL(wp), DIMENSION(0:ny1,nz) :: ar !< |
---|
| 1803 | REAL(wp), DIMENSION(0:ny+3,nz+1) :: ai !< |
---|
| 1804 | REAL(wp), DIMENSION(6*(ny+4),nz+1) :: work1 !< |
---|
[1320] | 1805 | |
---|
[1] | 1806 | #if defined( __nec ) |
---|
[1682] | 1807 | INTEGER(iwp) :: sizw !< |
---|
[1320] | 1808 | |
---|
[1682] | 1809 | COMPLEX(wp), DIMENSION((ny+4)/2+1,nz+1) :: work !< |
---|
[1] | 1810 | #endif |
---|
| 1811 | |
---|
| 1812 | IF ( fft_method == 'temperton-algorithm' ) THEN |
---|
| 1813 | |
---|
| 1814 | siza = SIZE( ai, 1 ) |
---|
| 1815 | |
---|
| 1816 | IF ( direction == 'forward') THEN |
---|
| 1817 | |
---|
| 1818 | ai(0:ny,1:nz) = ar(0:ny,1:nz) |
---|
[1342] | 1819 | ai(ny+1:,:) = 0.0_wp |
---|
[1] | 1820 | |
---|
| 1821 | CALL fft991cy( ai, work1, trigs_y, ifax_y, 1, siza, ny+1, nz, -1 ) |
---|
| 1822 | |
---|
| 1823 | DO k = 1, nz |
---|
| 1824 | DO j = 0, (ny+1)/2 |
---|
| 1825 | ar(j,k) = ai(2*j,k) |
---|
| 1826 | ENDDO |
---|
| 1827 | DO j = 1, (ny+1)/2 - 1 |
---|
| 1828 | ar(ny+1-j,k) = ai(2*j+1,k) |
---|
| 1829 | ENDDO |
---|
| 1830 | ENDDO |
---|
| 1831 | |
---|
| 1832 | ELSE |
---|
| 1833 | |
---|
| 1834 | DO k = 1, nz |
---|
| 1835 | DO j = 0, (ny+1)/2 |
---|
| 1836 | ai(2*j,k) = ar(j,k) |
---|
| 1837 | ENDDO |
---|
| 1838 | DO j = 1, (ny+1)/2 - 1 |
---|
| 1839 | ai(2*j+1,k) = ar(ny+1-j,k) |
---|
| 1840 | ENDDO |
---|
[1342] | 1841 | ai(1,k) = 0.0_wp |
---|
| 1842 | ai(ny+2,k) = 0.0_wp |
---|
[1] | 1843 | ENDDO |
---|
| 1844 | |
---|
| 1845 | CALL fft991cy( ai, work1, trigs_y, ifax_y, 1, siza, ny+1, nz, 1 ) |
---|
| 1846 | |
---|
| 1847 | ar(0:ny,1:nz) = ai(0:ny,1:nz) |
---|
| 1848 | |
---|
| 1849 | ENDIF |
---|
| 1850 | |
---|
| 1851 | ELSEIF ( fft_method == 'system-specific' ) THEN |
---|
| 1852 | |
---|
| 1853 | #if defined( __nec ) |
---|
| 1854 | siza = SIZE( ai, 1 ) |
---|
| 1855 | sizw = SIZE( work, 1 ) |
---|
| 1856 | |
---|
| 1857 | IF ( direction == 'forward') THEN |
---|
| 1858 | |
---|
| 1859 | ! |
---|
| 1860 | !-- Tables are initialized once more. This call should not be |
---|
| 1861 | !-- necessary, but otherwise program aborts in asymmetric case |
---|
[1106] | 1862 | CALL DZFFTM( 0, ny+1, nz1, sqr_dny, work, ny+4, work, ny+4, & |
---|
[1] | 1863 | trig_yf, work1, 0 ) |
---|
| 1864 | |
---|
| 1865 | ai(0:ny,1:nz) = ar(0:ny,1:nz) |
---|
| 1866 | IF ( nz1 > nz ) THEN |
---|
[1342] | 1867 | ai(:,nz1) = 0.0_wp |
---|
[1] | 1868 | ENDIF |
---|
| 1869 | |
---|
[1106] | 1870 | CALL DZFFTM( 1, ny+1, nz1, sqr_dny, ai, siza, work, sizw, & |
---|
[1] | 1871 | trig_yf, work1, 0 ) |
---|
| 1872 | |
---|
| 1873 | DO k = 1, nz |
---|
| 1874 | DO j = 0, (ny+1)/2 |
---|
[1322] | 1875 | ar(j,k) = REAL( work(j+1,k), KIND=wp ) |
---|
[1] | 1876 | ENDDO |
---|
| 1877 | DO j = 1, (ny+1)/2 - 1 |
---|
| 1878 | ar(ny+1-j,k) = AIMAG( work(j+1,k) ) |
---|
| 1879 | ENDDO |
---|
| 1880 | ENDDO |
---|
| 1881 | |
---|
| 1882 | ELSE |
---|
| 1883 | |
---|
| 1884 | ! |
---|
| 1885 | !-- Tables are initialized once more. This call should not be |
---|
| 1886 | !-- necessary, but otherwise program aborts in asymmetric case |
---|
[1106] | 1887 | CALL ZDFFTM( 0, ny+1, nz1, sqr_dny, work, ny+4, work, ny+4, & |
---|
[1] | 1888 | trig_yb, work1, 0 ) |
---|
| 1889 | |
---|
| 1890 | IF ( nz1 > nz ) THEN |
---|
[1342] | 1891 | work(:,nz1) = 0.0_wp |
---|
[1] | 1892 | ENDIF |
---|
| 1893 | DO k = 1, nz |
---|
[1392] | 1894 | work(1,k) = CMPLX( ar(0,k), 0.0_wp, KIND=wp ) |
---|
[1] | 1895 | DO j = 1, (ny+1)/2 - 1 |
---|
[1392] | 1896 | work(j+1,k) = CMPLX( ar(j,k), ar(ny+1-j,k), KIND=wp ) |
---|
[1] | 1897 | ENDDO |
---|
[1392] | 1898 | work(((ny+1)/2)+1,k) = CMPLX( ar((ny+1)/2,k), 0.0_wp, KIND=wp ) |
---|
[1] | 1899 | ENDDO |
---|
| 1900 | |
---|
[1106] | 1901 | CALL ZDFFTM( -1, ny+1, nz1, sqr_dny, work, sizw, ai, siza, & |
---|
[1] | 1902 | trig_yb, work1, 0 ) |
---|
| 1903 | |
---|
| 1904 | ar(0:ny,1:nz) = ai(0:ny,1:nz) |
---|
| 1905 | |
---|
| 1906 | ENDIF |
---|
| 1907 | |
---|
| 1908 | #else |
---|
[254] | 1909 | message_string = 'no system-specific fft-call available' |
---|
| 1910 | CALL message( 'fft_y_m', 'PA0188', 1, 2, 0, 6, 0 ) |
---|
[1] | 1911 | #endif |
---|
| 1912 | |
---|
| 1913 | ELSE |
---|
[254] | 1914 | |
---|
[274] | 1915 | message_string = 'fft method "' // TRIM( fft_method) // & |
---|
| 1916 | '" not available' |
---|
[254] | 1917 | CALL message( 'fft_x_m', 'PA0189', 1, 2, 0, 6, 0 ) |
---|
[1] | 1918 | |
---|
| 1919 | ENDIF |
---|
| 1920 | |
---|
| 1921 | END SUBROUTINE fft_y_m |
---|
| 1922 | |
---|
[1106] | 1923 | |
---|
[1] | 1924 | END MODULE fft_xy |
---|