| 1 |
guez |
30 |
MODULE histwrite_m |
| 2 |
|
|
|
| 3 |
guez |
31 |
! From histcom.f90, version 2.1 2004/04/21 09:27:10 |
| 4 |
guez |
30 |
|
| 5 |
guez |
92 |
USE errioipsl, ONLY: histerr |
| 6 |
|
|
USE histcom_var, ONLY: datasz_in, datasz_max, date0, deltat, & |
| 7 |
|
|
freq_opp, freq_wrt, fuchnbout, last_opp, last_opp_chk, last_wrt, & |
| 8 |
|
|
last_wrt_chk, missing_val, nbopp, nb_files, scal, scsize, sopps, & |
| 9 |
|
|
topp |
| 10 |
|
|
use histvar_seq_m, only: histvar_seq |
| 11 |
|
|
use histwrite_real_m, only: histwrite_real |
| 12 |
|
|
use isittime_m, only: isittime |
| 13 |
|
|
USE mathop_m, ONLY: mathop |
| 14 |
|
|
|
| 15 |
guez |
30 |
implicit none |
| 16 |
|
|
|
| 17 |
|
|
INTERFACE histwrite |
| 18 |
guez |
31 |
! The "histwrite" procedures give the data to the input-output system. |
| 19 |
|
|
! They trigger the operations to be performed and the writing to |
| 20 |
|
|
! the file if needed. |
| 21 |
guez |
30 |
|
| 22 |
guez |
31 |
! We test the work to be done at this time here so that at a |
| 23 |
|
|
! later stage we can call different operations and write subroutines |
| 24 |
|
|
! for the REAL and INTEGER interfaces. |
| 25 |
guez |
30 |
|
| 26 |
guez |
62 |
! INTEGER, INTENT(IN):: fileid |
| 27 |
guez |
30 |
! The ID of the file on which this variable is to be written. |
| 28 |
|
|
! The variable should have been defined in this file before. |
| 29 |
|
|
|
| 30 |
guez |
62 |
! CHARACTER(LEN=*), INTENT(IN):: varname |
| 31 |
guez |
30 |
! short name of the variable |
| 32 |
|
|
|
| 33 |
guez |
62 |
! INTEGER, INTENT(IN):: itau |
| 34 |
guez |
30 |
! current timestep |
| 35 |
|
|
|
| 36 |
|
|
! REAL, INTENT(IN):: pdata(:) or (:, :) or (:, :, :) |
| 37 |
|
|
! values of the variable |
| 38 |
|
|
|
| 39 |
|
|
! The difference between the procedures is the rank of "pdata". |
| 40 |
|
|
|
| 41 |
|
|
MODULE PROCEDURE histwrite_r1d, histwrite_r2d, histwrite_r3d |
| 42 |
guez |
67 |
END INTERFACE histwrite |
| 43 |
guez |
30 |
|
| 44 |
guez |
92 |
PRIVATE |
| 45 |
|
|
public histwrite |
| 46 |
guez |
45 |
|
| 47 |
guez |
30 |
CONTAINS |
| 48 |
|
|
|
| 49 |
guez |
62 |
SUBROUTINE histwrite_r1d(fileid, varname, itau, pdata) |
| 50 |
guez |
30 |
|
| 51 |
guez |
62 |
INTEGER, INTENT(IN):: fileid, itau |
| 52 |
|
|
CHARACTER(LEN=*), INTENT(IN):: varname |
| 53 |
guez |
56 |
REAL, INTENT(IN):: pdata(:) |
| 54 |
guez |
30 |
|
| 55 |
|
|
! Variables local to the procedure: |
| 56 |
|
|
integer nbindex, nindex(size(pdata)) |
| 57 |
guez |
56 |
LOGICAL:: do_oper, do_write, largebuf |
| 58 |
|
|
INTEGER:: varid, io, nbpt_in, nbpt_out |
| 59 |
|
|
REAL, ALLOCATABLE, SAVE:: buff_tmp(:) |
| 60 |
|
|
INTEGER, SAVE:: buff_tmp_sz |
| 61 |
|
|
CHARACTER(LEN=7):: tmp_opp |
| 62 |
guez |
30 |
|
| 63 |
guez |
31 |
!-------------------------------------------------------------------- |
| 64 |
guez |
30 |
|
| 65 |
|
|
nbindex = size(nindex) |
| 66 |
|
|
nindex = 0 |
| 67 |
|
|
|
| 68 |
|
|
! 1.0 Try to catch errors like specifying the wrong file ID. |
| 69 |
|
|
|
| 70 |
guez |
67 |
IF ((fileid < 1) .OR. (fileid > nb_files)) THEN |
| 71 |
|
|
CALL histerr(3, "histwrite", & |
| 72 |
|
|
'Illegal file ID in the histwrite of variable', varname, ' ') |
| 73 |
guez |
30 |
ENDIF |
| 74 |
|
|
|
| 75 |
|
|
! 1.1 Find the id of the variable to be written and the real time |
| 76 |
|
|
|
| 77 |
guez |
67 |
CALL histvar_seq(fileid, varname, varid) |
| 78 |
guez |
30 |
|
| 79 |
|
|
! 2.0 do nothing for never operation |
| 80 |
|
|
|
| 81 |
guez |
62 |
tmp_opp = topp(fileid, varid) |
| 82 |
guez |
30 |
|
| 83 |
|
|
IF (TRIM(tmp_opp) == "never") THEN |
| 84 |
guez |
62 |
last_opp_chk(fileid, varid) = -99 |
| 85 |
|
|
last_wrt_chk(fileid, varid) = -99 |
| 86 |
guez |
30 |
ENDIF |
| 87 |
|
|
|
| 88 |
|
|
! 3.0 We check if we need to do an operation |
| 89 |
|
|
|
| 90 |
guez |
62 |
IF (last_opp_chk(fileid, varid) == itau) THEN |
| 91 |
guez |
67 |
CALL histerr(3, "histwrite", & |
| 92 |
|
|
'This variable as already been analysed at the present', & |
| 93 |
|
|
'time step', ' ') |
| 94 |
guez |
30 |
ENDIF |
| 95 |
|
|
|
| 96 |
guez |
62 |
CALL isittime(itau, date0(fileid), deltat(fileid), & |
| 97 |
|
|
freq_opp(fileid, varid), last_opp(fileid, varid), & |
| 98 |
|
|
last_opp_chk(fileid, varid), do_oper) |
| 99 |
guez |
30 |
|
| 100 |
|
|
! 4.0 We check if we need to write the data |
| 101 |
|
|
|
| 102 |
guez |
62 |
IF (last_wrt_chk(fileid, varid) == itau) THEN |
| 103 |
guez |
67 |
CALL histerr(3, "histwrite", & |
| 104 |
|
|
'This variable as already been written for the present', & |
| 105 |
|
|
'time step', ' ') |
| 106 |
guez |
30 |
ENDIF |
| 107 |
|
|
|
| 108 |
guez |
67 |
CALL isittime(itau, date0(fileid), deltat(fileid), & |
| 109 |
|
|
freq_wrt(fileid, varid), last_wrt(fileid, varid), & |
| 110 |
|
|
last_wrt_chk(fileid, varid), do_write) |
| 111 |
guez |
30 |
|
| 112 |
|
|
! 5.0 histwrite called |
| 113 |
|
|
|
| 114 |
guez |
67 |
IF (do_oper .OR. do_write) THEN |
| 115 |
|
|
! 5.1 Get the sizes of the data we will handle |
| 116 |
guez |
30 |
|
| 117 |
guez |
62 |
IF (datasz_in(fileid, varid, 1) <= 0) THEN |
| 118 |
guez |
67 |
! There is the risk here that the user has over-sized the array. |
| 119 |
|
|
! But how can we catch this ? |
| 120 |
|
|
! In the worst case we will do impossible operations |
| 121 |
|
|
! on part of the data ! |
| 122 |
guez |
62 |
datasz_in(fileid, varid, 1) = SIZE(pdata) |
| 123 |
|
|
datasz_in(fileid, varid, 2) = -1 |
| 124 |
|
|
datasz_in(fileid, varid, 3) = -1 |
| 125 |
guez |
30 |
ENDIF |
| 126 |
|
|
|
| 127 |
guez |
67 |
! 5.2 The maximum size of the data will give the size of the buffer |
| 128 |
guez |
30 |
|
| 129 |
guez |
62 |
IF (datasz_max(fileid, varid) <= 0) THEN |
| 130 |
guez |
30 |
largebuf = .FALSE. |
| 131 |
guez |
62 |
DO io=1, nbopp(fileid, varid) |
| 132 |
|
|
IF (INDEX(fuchnbout, sopps(fileid, varid, io)) > 0) THEN |
| 133 |
guez |
30 |
largebuf = .TRUE. |
| 134 |
|
|
ENDIF |
| 135 |
|
|
ENDDO |
| 136 |
|
|
IF (largebuf) THEN |
| 137 |
guez |
67 |
datasz_max(fileid, varid) = scsize(fileid, varid, 1) & |
| 138 |
|
|
* scsize(fileid, varid, 2) *scsize(fileid, varid, 3) |
| 139 |
guez |
30 |
ELSE |
| 140 |
guez |
67 |
datasz_max(fileid, varid) = datasz_in(fileid, varid, 1) |
| 141 |
guez |
30 |
ENDIF |
| 142 |
|
|
ENDIF |
| 143 |
|
|
|
| 144 |
|
|
IF (.NOT.ALLOCATED(buff_tmp)) THEN |
| 145 |
guez |
67 |
ALLOCATE(buff_tmp(datasz_max(fileid, varid))) |
| 146 |
guez |
62 |
buff_tmp_sz = datasz_max(fileid, varid) |
| 147 |
|
|
ELSE IF (datasz_max(fileid, varid) > buff_tmp_sz) THEN |
| 148 |
guez |
67 |
DEALLOCATE(buff_tmp) |
| 149 |
|
|
ALLOCATE(buff_tmp(datasz_max(fileid, varid))) |
| 150 |
guez |
62 |
buff_tmp_sz = datasz_max(fileid, varid) |
| 151 |
guez |
30 |
ENDIF |
| 152 |
|
|
|
| 153 |
guez |
67 |
! We have to do the first operation anyway. Thus we do it here |
| 154 |
|
|
! and change the ranke of the data at the same time. This |
| 155 |
|
|
! should speed up things. |
| 156 |
guez |
30 |
|
| 157 |
guez |
62 |
nbpt_in = datasz_in(fileid, varid, 1) |
| 158 |
|
|
nbpt_out = datasz_max(fileid, varid) |
| 159 |
guez |
67 |
CALL mathop(sopps(fileid, varid, 1), nbpt_in, pdata, missing_val, & |
| 160 |
|
|
nbindex, nindex, scal(fileid, varid, 1), nbpt_out, buff_tmp) |
| 161 |
|
|
CALL histwrite_real(fileid, varid, itau, nbpt_out, buff_tmp, nbindex, & |
| 162 |
|
|
nindex, do_oper, do_write) |
| 163 |
guez |
30 |
ENDIF |
| 164 |
|
|
|
| 165 |
|
|
! 6.0 Manage time steps |
| 166 |
|
|
|
| 167 |
guez |
67 |
IF ((TRIM(tmp_opp) /= "once") .AND. (TRIM(tmp_opp) /= "never")) THEN |
| 168 |
guez |
62 |
last_opp_chk(fileid, varid) = itau |
| 169 |
|
|
last_wrt_chk(fileid, varid) = itau |
| 170 |
guez |
30 |
ELSE |
| 171 |
guez |
62 |
last_opp_chk(fileid, varid) = -99 |
| 172 |
|
|
last_wrt_chk(fileid, varid) = -99 |
| 173 |
guez |
30 |
ENDIF |
| 174 |
guez |
56 |
|
| 175 |
guez |
30 |
END SUBROUTINE histwrite_r1d |
| 176 |
|
|
|
| 177 |
guez |
56 |
!************************************************************************ |
| 178 |
guez |
30 |
|
| 179 |
guez |
62 |
SUBROUTINE histwrite_r2d (fileid, varname, itau, pdata) |
| 180 |
guez |
30 |
|
| 181 |
guez |
62 |
INTEGER, INTENT(IN):: fileid, itau |
| 182 |
guez |
56 |
REAL, INTENT(IN):: pdata(:, :) |
| 183 |
guez |
62 |
CHARACTER(LEN=*), INTENT(IN):: varname |
| 184 |
guez |
30 |
|
| 185 |
|
|
integer nbindex, nindex(size(pdata)) |
| 186 |
guez |
56 |
LOGICAL:: do_oper, do_write, largebuf |
| 187 |
|
|
INTEGER:: varid, io, nbpt_in(1:2), nbpt_out |
| 188 |
|
|
REAL, ALLOCATABLE, SAVE:: buff_tmp(:) |
| 189 |
|
|
INTEGER, SAVE:: buff_tmp_sz |
| 190 |
|
|
CHARACTER(LEN=7):: tmp_opp |
| 191 |
guez |
30 |
|
| 192 |
guez |
31 |
!-------------------------------------------------------------------- |
| 193 |
guez |
30 |
|
| 194 |
|
|
nbindex = size(nindex) |
| 195 |
|
|
nindex = 0 |
| 196 |
|
|
|
| 197 |
|
|
! 1.0 Try to catch errors like specifying the wrong file ID. |
| 198 |
|
|
|
| 199 |
guez |
62 |
IF ( (fileid < 1).OR.(fileid > nb_files) ) THEN |
| 200 |
guez |
30 |
CALL histerr (3, "histwrite", & |
| 201 |
guez |
67 |
'Illegal file ID in the histwrite of variable', varname, ' ') |
| 202 |
guez |
30 |
ENDIF |
| 203 |
|
|
|
| 204 |
|
|
! 1.1 Find the id of the variable to be written and the real time |
| 205 |
|
|
|
| 206 |
guez |
62 |
CALL histvar_seq (fileid, varname, varid) |
| 207 |
guez |
30 |
|
| 208 |
|
|
! 2.0 do nothing for never operation |
| 209 |
|
|
|
| 210 |
guez |
62 |
tmp_opp = topp(fileid, varid) |
| 211 |
guez |
30 |
|
| 212 |
|
|
IF (TRIM(tmp_opp) == "never") THEN |
| 213 |
guez |
62 |
last_opp_chk(fileid, varid) = -99 |
| 214 |
|
|
last_wrt_chk(fileid, varid) = -99 |
| 215 |
guez |
30 |
ENDIF |
| 216 |
|
|
|
| 217 |
|
|
! 3.0 We check if we need to do an operation |
| 218 |
|
|
|
| 219 |
guez |
68 |
IF (last_opp_chk(fileid, varid) == itau) CALL histerr (3, "histwrite", & |
| 220 |
|
|
'This variable as already been analysed at the present', & |
| 221 |
|
|
'time step', ' ') |
| 222 |
guez |
30 |
|
| 223 |
guez |
68 |
CALL isittime(itau, date0(fileid), deltat(fileid), & |
| 224 |
|
|
freq_opp(fileid, varid), last_opp(fileid, varid), & |
| 225 |
|
|
last_opp_chk(fileid, varid), do_oper) |
| 226 |
guez |
30 |
|
| 227 |
|
|
! 4.0 We check if we need to write the data |
| 228 |
|
|
|
| 229 |
guez |
62 |
IF (last_wrt_chk(fileid, varid) == itau) THEN |
| 230 |
guez |
30 |
CALL histerr (3, "histwrite", & |
| 231 |
guez |
67 |
'This variable as already been written for the present', & |
| 232 |
|
|
'time step', ' ') |
| 233 |
guez |
30 |
ENDIF |
| 234 |
|
|
|
| 235 |
|
|
CALL isittime & |
| 236 |
guez |
67 |
(itau, date0(fileid), deltat(fileid), freq_wrt(fileid, varid), & |
| 237 |
|
|
last_wrt(fileid, varid), last_wrt_chk(fileid, varid), do_write) |
| 238 |
guez |
30 |
|
| 239 |
|
|
! 5.0 histwrite called |
| 240 |
|
|
|
| 241 |
|
|
IF (do_oper.OR.do_write) THEN |
| 242 |
|
|
|
| 243 |
guez |
31 |
!- 5.1 Get the sizes of the data we will handle |
| 244 |
guez |
30 |
|
| 245 |
guez |
62 |
IF (datasz_in(fileid, varid, 1) <= 0) THEN |
| 246 |
guez |
31 |
!--- There is the risk here that the user has over-sized the array. |
| 247 |
|
|
!--- But how can we catch this ? |
| 248 |
|
|
!--- In the worst case we will do impossible operations |
| 249 |
|
|
!--- on part of the data ! |
| 250 |
guez |
62 |
datasz_in(fileid, varid, 1) = SIZE(pdata, DIM=1) |
| 251 |
|
|
datasz_in(fileid, varid, 2) = SIZE(pdata, DIM=2) |
| 252 |
|
|
datasz_in(fileid, varid, 3) = -1 |
| 253 |
guez |
30 |
ENDIF |
| 254 |
|
|
|
| 255 |
guez |
31 |
!- 5.2 The maximum size of the data will give the size of the buffer |
| 256 |
guez |
30 |
|
| 257 |
guez |
62 |
IF (datasz_max(fileid, varid) <= 0) THEN |
| 258 |
guez |
30 |
largebuf = .FALSE. |
| 259 |
guez |
62 |
DO io=1, nbopp(fileid, varid) |
| 260 |
|
|
IF (INDEX(fuchnbout, sopps(fileid, varid, io)) > 0) THEN |
| 261 |
guez |
30 |
largebuf = .TRUE. |
| 262 |
|
|
ENDIF |
| 263 |
|
|
ENDDO |
| 264 |
|
|
IF (largebuf) THEN |
| 265 |
guez |
62 |
datasz_max(fileid, varid) = & |
| 266 |
guez |
67 |
scsize(fileid, varid, 1) & |
| 267 |
|
|
*scsize(fileid, varid, 2) & |
| 268 |
|
|
*scsize(fileid, varid, 3) |
| 269 |
guez |
30 |
ELSE |
| 270 |
guez |
62 |
datasz_max(fileid, varid) = & |
| 271 |
guez |
67 |
datasz_in(fileid, varid, 1) & |
| 272 |
|
|
*datasz_in(fileid, varid, 2) |
| 273 |
guez |
30 |
ENDIF |
| 274 |
|
|
ENDIF |
| 275 |
|
|
|
| 276 |
|
|
IF (.NOT.ALLOCATED(buff_tmp)) THEN |
| 277 |
guez |
62 |
ALLOCATE (buff_tmp(datasz_max(fileid, varid))) |
| 278 |
|
|
buff_tmp_sz = datasz_max(fileid, varid) |
| 279 |
|
|
ELSE IF (datasz_max(fileid, varid) > buff_tmp_sz) THEN |
| 280 |
guez |
30 |
DEALLOCATE (buff_tmp) |
| 281 |
guez |
62 |
ALLOCATE (buff_tmp(datasz_max(fileid, varid))) |
| 282 |
|
|
buff_tmp_sz = datasz_max(fileid, varid) |
| 283 |
guez |
30 |
ENDIF |
| 284 |
|
|
|
| 285 |
guez |
31 |
!- We have to do the first operation anyway. |
| 286 |
|
|
!- Thus we do it here and change the ranke |
| 287 |
|
|
!- of the data at the same time. This should speed up things. |
| 288 |
guez |
30 |
|
| 289 |
guez |
62 |
nbpt_in(1:2) = datasz_in(fileid, varid, 1:2) |
| 290 |
|
|
nbpt_out = datasz_max(fileid, varid) |
| 291 |
|
|
CALL mathop (sopps(fileid, varid, 1), nbpt_in, pdata, & |
| 292 |
guez |
67 |
missing_val, nbindex, nindex, & |
| 293 |
|
|
scal(fileid, varid, 1), nbpt_out, buff_tmp) |
| 294 |
guez |
62 |
CALL histwrite_real (fileid, varid, itau, nbpt_out, & |
| 295 |
guez |
67 |
buff_tmp, nbindex, nindex, do_oper, do_write) |
| 296 |
guez |
30 |
ENDIF |
| 297 |
|
|
|
| 298 |
|
|
! 6.0 Manage time steps |
| 299 |
|
|
|
| 300 |
|
|
IF ((TRIM(tmp_opp) /= "once").AND.(TRIM(tmp_opp) /= "never")) THEN |
| 301 |
guez |
62 |
last_opp_chk(fileid, varid) = itau |
| 302 |
|
|
last_wrt_chk(fileid, varid) = itau |
| 303 |
guez |
30 |
ELSE |
| 304 |
guez |
62 |
last_opp_chk(fileid, varid) = -99 |
| 305 |
|
|
last_wrt_chk(fileid, varid) = -99 |
| 306 |
guez |
30 |
ENDIF |
| 307 |
guez |
56 |
|
| 308 |
guez |
30 |
END SUBROUTINE histwrite_r2d |
| 309 |
|
|
|
| 310 |
guez |
56 |
!************************************************************************ |
| 311 |
guez |
30 |
|
| 312 |
guez |
62 |
SUBROUTINE histwrite_r3d (fileid, varname, itau, pdata) |
| 313 |
guez |
30 |
|
| 314 |
guez |
62 |
INTEGER, INTENT(IN):: fileid, itau |
| 315 |
guez |
56 |
REAL, DIMENSION(:, :, :), INTENT(IN):: pdata |
| 316 |
guez |
62 |
CHARACTER(LEN=*), INTENT(IN):: varname |
| 317 |
guez |
30 |
|
| 318 |
|
|
integer nbindex, nindex(size(pdata)) |
| 319 |
guez |
56 |
LOGICAL:: do_oper, do_write, largebuf |
| 320 |
|
|
INTEGER:: varid, io, nbpt_in(1:3), nbpt_out |
| 321 |
|
|
REAL, ALLOCATABLE, SAVE:: buff_tmp(:) |
| 322 |
|
|
INTEGER, SAVE:: buff_tmp_sz |
| 323 |
|
|
CHARACTER(LEN=7):: tmp_opp |
| 324 |
guez |
30 |
|
| 325 |
guez |
31 |
!-------------------------------------------------------------------- |
| 326 |
guez |
30 |
|
| 327 |
|
|
nbindex = size(nindex) |
| 328 |
|
|
nindex = 0 |
| 329 |
|
|
|
| 330 |
|
|
! 1.0 Try to catch errors like specifying the wrong file ID. |
| 331 |
|
|
! Thanks Marine for showing us what errors users can make ! |
| 332 |
|
|
|
| 333 |
guez |
62 |
IF ( (fileid < 1).OR.(fileid > nb_files) ) THEN |
| 334 |
guez |
30 |
CALL histerr (3, "histwrite", & |
| 335 |
guez |
67 |
'Illegal file ID in the histwrite of variable', varname, ' ') |
| 336 |
guez |
30 |
ENDIF |
| 337 |
|
|
|
| 338 |
|
|
! 1.1 Find the id of the variable to be written and the real time |
| 339 |
|
|
|
| 340 |
guez |
62 |
CALL histvar_seq (fileid, varname, varid) |
| 341 |
guez |
30 |
|
| 342 |
|
|
! 2.0 do nothing for never operation |
| 343 |
|
|
|
| 344 |
guez |
62 |
tmp_opp = topp(fileid, varid) |
| 345 |
guez |
30 |
|
| 346 |
|
|
IF (TRIM(tmp_opp) == "never") THEN |
| 347 |
guez |
62 |
last_opp_chk(fileid, varid) = -99 |
| 348 |
|
|
last_wrt_chk(fileid, varid) = -99 |
| 349 |
guez |
30 |
ENDIF |
| 350 |
|
|
|
| 351 |
|
|
! 3.0 We check if we need to do an operation |
| 352 |
|
|
|
| 353 |
guez |
62 |
IF (last_opp_chk(fileid, varid) == itau) THEN |
| 354 |
guez |
30 |
CALL histerr (3, "histwrite", & |
| 355 |
guez |
67 |
'This variable as already been analysed at the present', & |
| 356 |
|
|
'time step', ' ') |
| 357 |
guez |
30 |
ENDIF |
| 358 |
|
|
|
| 359 |
|
|
CALL isittime & |
| 360 |
guez |
67 |
(itau, date0(fileid), deltat(fileid), freq_opp(fileid, varid), & |
| 361 |
|
|
last_opp(fileid, varid), last_opp_chk(fileid, varid), do_oper) |
| 362 |
guez |
30 |
|
| 363 |
|
|
! 4.0 We check if we need to write the data |
| 364 |
|
|
|
| 365 |
guez |
62 |
IF (last_wrt_chk(fileid, varid) == itau) THEN |
| 366 |
guez |
30 |
CALL histerr (3, "histwrite", & |
| 367 |
guez |
67 |
'This variable as already been written for the present', & |
| 368 |
|
|
'time step', ' ') |
| 369 |
guez |
30 |
ENDIF |
| 370 |
|
|
|
| 371 |
|
|
CALL isittime & |
| 372 |
guez |
67 |
(itau, date0(fileid), deltat(fileid), freq_wrt(fileid, varid), & |
| 373 |
|
|
last_wrt(fileid, varid), last_wrt_chk(fileid, varid), do_write) |
| 374 |
guez |
30 |
|
| 375 |
|
|
! 5.0 histwrite called |
| 376 |
|
|
|
| 377 |
|
|
IF (do_oper.OR.do_write) THEN |
| 378 |
|
|
|
| 379 |
guez |
31 |
!- 5.1 Get the sizes of the data we will handle |
| 380 |
guez |
30 |
|
| 381 |
guez |
62 |
IF (datasz_in(fileid, varid, 1) <= 0) THEN |
| 382 |
guez |
31 |
!--- There is the risk here that the user has over-sized the array. |
| 383 |
|
|
!--- But how can we catch this ? |
| 384 |
|
|
!--- In the worst case we will do impossible operations |
| 385 |
|
|
!--- on part of the data ! |
| 386 |
guez |
62 |
datasz_in(fileid, varid, 1) = SIZE(pdata, DIM=1) |
| 387 |
|
|
datasz_in(fileid, varid, 2) = SIZE(pdata, DIM=2) |
| 388 |
|
|
datasz_in(fileid, varid, 3) = SIZE(pdata, DIM=3) |
| 389 |
guez |
30 |
ENDIF |
| 390 |
|
|
|
| 391 |
guez |
31 |
!- 5.2 The maximum size of the data will give the size of the buffer |
| 392 |
guez |
30 |
|
| 393 |
guez |
62 |
IF (datasz_max(fileid, varid) <= 0) THEN |
| 394 |
guez |
30 |
largebuf = .FALSE. |
| 395 |
guez |
62 |
DO io =1, nbopp(fileid, varid) |
| 396 |
|
|
IF (INDEX(fuchnbout, sopps(fileid, varid, io)) > 0) THEN |
| 397 |
guez |
30 |
largebuf = .TRUE. |
| 398 |
|
|
ENDIF |
| 399 |
|
|
ENDDO |
| 400 |
|
|
IF (largebuf) THEN |
| 401 |
guez |
62 |
datasz_max(fileid, varid) = & |
| 402 |
guez |
67 |
scsize(fileid, varid, 1) & |
| 403 |
|
|
*scsize(fileid, varid, 2) & |
| 404 |
|
|
*scsize(fileid, varid, 3) |
| 405 |
guez |
30 |
ELSE |
| 406 |
guez |
62 |
datasz_max(fileid, varid) = & |
| 407 |
guez |
67 |
datasz_in(fileid, varid, 1) & |
| 408 |
|
|
*datasz_in(fileid, varid, 2) & |
| 409 |
|
|
*datasz_in(fileid, varid, 3) |
| 410 |
guez |
30 |
ENDIF |
| 411 |
|
|
ENDIF |
| 412 |
|
|
|
| 413 |
|
|
IF (.NOT.ALLOCATED(buff_tmp)) THEN |
| 414 |
guez |
62 |
ALLOCATE (buff_tmp(datasz_max(fileid, varid))) |
| 415 |
|
|
buff_tmp_sz = datasz_max(fileid, varid) |
| 416 |
|
|
ELSE IF (datasz_max(fileid, varid) > buff_tmp_sz) THEN |
| 417 |
guez |
30 |
DEALLOCATE (buff_tmp) |
| 418 |
guez |
62 |
ALLOCATE (buff_tmp(datasz_max(fileid, varid))) |
| 419 |
|
|
buff_tmp_sz = datasz_max(fileid, varid) |
| 420 |
guez |
30 |
ENDIF |
| 421 |
|
|
|
| 422 |
guez |
31 |
!- We have to do the first operation anyway. |
| 423 |
|
|
!- Thus we do it here and change the ranke |
| 424 |
|
|
!- of the data at the same time. This should speed up things. |
| 425 |
guez |
30 |
|
| 426 |
guez |
62 |
nbpt_in(1:3) = datasz_in(fileid, varid, 1:3) |
| 427 |
|
|
nbpt_out = datasz_max(fileid, varid) |
| 428 |
|
|
CALL mathop (sopps(fileid, varid, 1), nbpt_in, pdata, & |
| 429 |
guez |
67 |
missing_val, nbindex, nindex, & |
| 430 |
|
|
scal(fileid, varid, 1), nbpt_out, buff_tmp) |
| 431 |
guez |
62 |
CALL histwrite_real (fileid, varid, itau, nbpt_out, & |
| 432 |
guez |
67 |
buff_tmp, nbindex, nindex, do_oper, do_write) |
| 433 |
guez |
30 |
ENDIF |
| 434 |
|
|
|
| 435 |
|
|
! 6.0 Manage time steps |
| 436 |
|
|
|
| 437 |
|
|
IF ((TRIM(tmp_opp) /= "once").AND.(TRIM(tmp_opp) /= "never")) THEN |
| 438 |
guez |
62 |
last_opp_chk(fileid, varid) = itau |
| 439 |
|
|
last_wrt_chk(fileid, varid) = itau |
| 440 |
guez |
30 |
ELSE |
| 441 |
guez |
62 |
last_opp_chk(fileid, varid) = -99 |
| 442 |
|
|
last_wrt_chk(fileid, varid) = -99 |
| 443 |
guez |
30 |
ENDIF |
| 444 |
guez |
56 |
|
| 445 |
guez |
30 |
END SUBROUTINE histwrite_r3d |
| 446 |
|
|
|
| 447 |
|
|
END MODULE histwrite_m |
Parent Directory
| 
Revision Log