[1649] | 1 | #!/bin/ksh |
---|
[1046] | 2 | |
---|
[1059] | 3 | # mrun - script for running PALM jobs |
---|
| 4 | |
---|
[1046] | 5 | #--------------------------------------------------------------------------------# |
---|
| 6 | # This file is part of PALM. |
---|
| 7 | # |
---|
| 8 | # PALM is free software: you can redistribute it and/or modify it under the terms |
---|
| 9 | # of the GNU General Public License as published by the Free Software Foundation, |
---|
| 10 | # either version 3 of the License, or (at your option) any later version. |
---|
| 11 | # |
---|
| 12 | # PALM is distributed in the hope that it will be useful, but WITHOUT ANY |
---|
| 13 | # WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR |
---|
| 14 | # A PARTICULAR PURPOSE. See the GNU General Public License for more details. |
---|
| 15 | # |
---|
| 16 | # You should have received a copy of the GNU General Public License along with |
---|
| 17 | # PALM. If not, see <http://www.gnu.org/licenses/>. |
---|
| 18 | # |
---|
[1753] | 19 | # Copyright 1997-2015 Leibniz Universitaet Hannover |
---|
[1046] | 20 | #--------------------------------------------------------------------------------# |
---|
| 21 | # |
---|
| 22 | # Current revisions: |
---|
[1090] | 23 | # ------------------ |
---|
[1305] | 24 | # |
---|
[1758] | 25 | # |
---|
[1200] | 26 | # Former revisions: |
---|
| 27 | # ----------------- |
---|
| 28 | # $Id: mrun 1758 2016-02-22 15:53:28Z maronga $ |
---|
[1623] | 29 | # |
---|
[1758] | 30 | # 1757 2016-02-22 15:49:32Z maronga |
---|
| 31 | # Bugfix: use random number when PBS system is not working properly at lccray* |
---|
| 32 | # |
---|
[1756] | 33 | # 1755 2016-02-22 13:53:39Z maronga |
---|
| 34 | # Minor bugfix to last commit |
---|
| 35 | # |
---|
[1754] | 36 | # 1753 2016-02-22 13:49:49Z maronga |
---|
| 37 | # Bugfix: use random job number when PBS job nummer retrieval fails on lccray* |
---|
| 38 | # |
---|
[1623] | 39 | # 1622 2015-07-20 06:21:36Z heinze |
---|
| 40 | # bugfix for Mistral (lcbullhh) |
---|
| 41 | # |
---|
| 42 | # 1621 2015-07-17 11:39:33Z heinze |
---|
[1621] | 43 | # adjustments for Mistral at DKRZ Hamburg (lcbullhh) |
---|
[1200] | 44 | # |
---|
[1610] | 45 | # 1609 2015-07-03 15:37:58Z maronga |
---|
| 46 | # Modified the random number of the temporary working directory to match the |
---|
| 47 | # queuing system number on HLRN-III (lccray*). |
---|
| 48 | # |
---|
[1605] | 49 | # 1604 2015-06-23 15:48:03Z suehring |
---|
| 50 | # Enable coupled runs for lcmuk |
---|
| 51 | # |
---|
[1599] | 52 | # 1598 2015-05-29 06:45:40Z raasch |
---|
| 53 | # bugfix for creating hostfile and total number of MPI tasks if run in openmp-mode on one node |
---|
| 54 | # |
---|
[1550] | 55 | # 1549 2015-01-30 14:26:16Z suehring |
---|
| 56 | # Enable interactive debug sessions with allinea debugger |
---|
| 57 | # |
---|
[1548] | 58 | # 1547 2015-01-29 15:09:12Z witha |
---|
| 59 | # adjustments for ForWind computing cluster (lcflow) |
---|
| 60 | # |
---|
[1492] | 61 | # 1491 2014-11-12 08:12:57Z heinze |
---|
| 62 | # correct setting of batch_job allowed for ibm-machines |
---|
| 63 | # |
---|
[1469] | 64 | # 1468 2014-09-24 14:06:57Z maronga |
---|
| 65 | # Typo removed (addres->address) |
---|
| 66 | # Bugfix: returning files to IMUK via ssh did not work for lccrayh and lcycrayb |
---|
| 67 | # Added support for restart runs (remote) for lcxe6 |
---|
| 68 | # |
---|
[1443] | 69 | # 1442 2014-07-28 07:09:10Z raasch |
---|
| 70 | # default queues for HLRN-III (lccrayb/lccrayh) defined |
---|
| 71 | # |
---|
[1423] | 72 | # 1422 2014-07-02 10:01:25Z kanani |
---|
| 73 | # automatic restarts enabled for lccrayb, |
---|
| 74 | # adjustment of path to compilervars.sh at IMUK (see automatic restarts) |
---|
| 75 | # |
---|
[1421] | 76 | # 1420 2014-06-10 06:24:14Z raasch |
---|
| 77 | # -j1 option added to aprung command for lccrayf (CSC) |
---|
| 78 | # |
---|
[1403] | 79 | # 1402 2014-05-09 14:25:13Z raasch |
---|
| 80 | # batch_job added to envpar-NAMELIST |
---|
| 81 | # |
---|
[1358] | 82 | # 1357 2014-04-11 15:02:03Z witha |
---|
| 83 | # adjustments for lccrayh (automatic restarts on lccrayh outside of IMUK are now |
---|
| 84 | # possible) |
---|
| 85 | # |
---|
[1328] | 86 | # 1320 2014-03-20 08:40:49Z raasch |
---|
| 87 | # check namelist file set false by default |
---|
| 88 | # |
---|
[1305] | 89 | # 1304 2014-03-12 10:29:42Z raasch |
---|
| 90 | # ulimit option changed from -Ss to -s |
---|
| 91 | # |
---|
| 92 | # bugfix: missing "fi" in r1289 |
---|
| 93 | # |
---|
[1290] | 94 | # 1289 2014-03-04 07:12:34Z raasch |
---|
| 95 | # comments translated to English |
---|
| 96 | # necriam-, fimm-, ibmy-, and sgi-specific code removed |
---|
| 97 | # export of variables for palm and interpret_config removed |
---|
| 98 | # |
---|
[1282] | 99 | # 1281 2014-02-01 07:55:49Z raasch |
---|
| 100 | # rsync-copy restricted to Cray machines, since command is unavailable on some |
---|
| 101 | # other systems |
---|
| 102 | # |
---|
[1280] | 103 | # 1279 2014-01-28 12:10:14Z raasch |
---|
| 104 | # tasks_per_node must not be an integral divisor of numprocs any more. This was done |
---|
| 105 | # in order to remove annoying restrictions concerning the number of processors which |
---|
| 106 | # appear on machines with larger nodes (e.g. containing 24 cores). Now without this |
---|
| 107 | # restriction, one of the nodes will be filled with less than the given number of |
---|
| 108 | # tasks per node. A respective warning is given. |
---|
| 109 | # |
---|
[1275] | 110 | # 1274 2014-01-09 13:14:54Z heinze |
---|
| 111 | # adjustments for lccrayh |
---|
| 112 | # |
---|
[1273] | 113 | # 1272 2014-01-08 10:19:32Z witha |
---|
| 114 | # small adjustment for lcflow |
---|
| 115 | # |
---|
[1271] | 116 | # 1270 2013-12-16 11:05:01Z fricke |
---|
| 117 | # call of combine_plot_fields adjusted for lccrayb/lccrayh |
---|
| 118 | # |
---|
[1256] | 119 | # 1255 2013-11-07 14:43:35Z raasch |
---|
| 120 | # further adjustments for lccrayb remote access |
---|
| 121 | # |
---|
[1242] | 122 | # 1241 2013-10-30 11:36:58Z heinze |
---|
| 123 | # Enable use of nudging input and input of large scale forcing from |
---|
| 124 | # external files |
---|
| 125 | # |
---|
[1230] | 126 | # 1229 2013-09-20 06:55:19Z raasch |
---|
| 127 | # further adjustments for lccrayb |
---|
| 128 | # |
---|
[1225] | 129 | # 1224 2013-09-16 07:27:23Z raasch |
---|
| 130 | # first adjustments for lccrayb |
---|
| 131 | # |
---|
[1211] | 132 | # 1210 2013-08-14 10:58:20Z raasch |
---|
| 133 | # fftw support added |
---|
| 134 | # |
---|
[1203] | 135 | # 1201 2013-07-10 16:17:59Z witha |
---|
| 136 | # adjustments for Forwind cluster (lcflow) |
---|
| 137 | # |
---|
[1200] | 138 | # 1199 2013-07-05 14:52:22Z raasch |
---|
[1199] | 139 | # adjustments for CSC Helsinki (lccrayf), |
---|
| 140 | # executables for batch jobs can be created in advance, in order to avoid calling |
---|
| 141 | # the compiler within the batch job (only works if batch jobs are submitted on |
---|
| 142 | # local host) |
---|
[1104] | 143 | # |
---|
[1191] | 144 | # 1190 2013-06-25 09:39:21Z heinze |
---|
| 145 | # enable use of automatic restarts for ibmh |
---|
[1185] | 146 | # use of cluster/express queue enabled (ibmh) |
---|
[1104] | 147 | # |
---|
[1125] | 148 | # 1124 2013-04-09 15:46:52Z raasch |
---|
| 149 | # variable "memory" is exported via typeset option -x, because otherwise an unknown |
---|
| 150 | # side effect may lead to data loss while getopts is reading the script-option arguments |
---|
| 151 | # |
---|
[1123] | 152 | # 1122 2013-04-09 08:37:16Z heinze |
---|
| 153 | # Bugfix: change type of variable last_char |
---|
| 154 | # |
---|
[1121] | 155 | # 1119 2013-04-05 15:11:19Z raasch |
---|
| 156 | # Bugfix for setting -T option for subjob |
---|
| 157 | # |
---|
[1109] | 158 | # 1108 2013-03-05 07:03:32Z raasch |
---|
| 159 | # bugfix for coupled runs on lckyut/lckyuh |
---|
| 160 | # |
---|
[1107] | 161 | # 1106 2013-03-04 05:31:38Z raasch |
---|
| 162 | # --stdin argument for mpiexec on lckyuh |
---|
| 163 | # -y and -Y settings output to header |
---|
| 164 | # |
---|
[1104] | 165 | # 1103 2013-02-20 02:15:53Z raasch |
---|
[1103] | 166 | # default script runs again under ksh, because of unsolved problems with read |
---|
| 167 | # from stdin: when bash script is called from a ksh, message "read error: 0: |
---|
| 168 | # Resource temporarily unavailable" appears and script does not stop, |
---|
| 169 | # further bash compatibility adjustments, |
---|
| 170 | # shebang line replaced by /bin/bash when running jobs on lckyuh; no restarts |
---|
| 171 | # on lckyuh, but mrun does not terminate and issues a warning instead |
---|
[1046] | 172 | # |
---|
[1102] | 173 | # 1101 2013-02-17 10:20:21Z raasch |
---|
| 174 | # script now running under bash instead of ksh, which required small adjustments |
---|
| 175 | # (output formatting with printf instead "typeset -L/-R", print replaced by echo, |
---|
| 176 | # read from stdin), |
---|
| 177 | # cross compilername on lckyuh compute nodes replaced by real compiler name |
---|
| 178 | # |
---|
[1100] | 179 | # 1099 2013-02-10 01:47:43Z raasch |
---|
| 180 | # adjustments for Kyushu-University computing center (lckyuh - hayaka) |
---|
| 181 | # and for Forwind cluster (lcflow) |
---|
| 182 | # small further adjustments for lckyut |
---|
| 183 | # |
---|
[1096] | 184 | # 1094 2013-02-03 01:52:12Z raasch |
---|
| 185 | # explicit ssh/scp port can be set in config file with environment variable |
---|
| 186 | # scp_port. This port is handled to all ssh/scp/batch_scp calls. |
---|
| 187 | # decalpha parts (yonsei) removed |
---|
| 188 | # |
---|
[1091] | 189 | # 2013-02-02 07:06:13Z raasch |
---|
[1099] | 190 | # adjustments for Kyushu-University computing center (lckyut - tatara) |
---|
[1091] | 191 | # |
---|
[1084] | 192 | # 1083 2013-01-04 10:22:09Z maronga |
---|
| 193 | # bugfix in parameter file check (read %cpp_options was missing) |
---|
| 194 | # |
---|
[1070] | 195 | # 1069 2012-11-28 16:18:43Z maronga |
---|
| 196 | # bugfix: coupling mode was always set to mpi2, typos removed |
---|
| 197 | # |
---|
[1059] | 198 | # 1058 2012-11-21 07:00:35Z raasch |
---|
| 199 | # Intel inspector (inspxe) is given the number of PEs instead of the number of |
---|
| 200 | # nodes |
---|
| 201 | # |
---|
[1047] | 202 | # 1046 2012-11-09 14:38:45Z maronga |
---|
| 203 | # code put under GPL (PALM 3.9) |
---|
| 204 | # |
---|
[1090] | 205 | # 21/03/94 - Siggi - first version finished |
---|
| 206 | # 03/03/94 - Siggi - script development started |
---|
| 207 | # |
---|
[1059] | 208 | #--------------------------------------------------------------------------------# |
---|
[1090] | 209 | # mrun - script for running PALM jobs |
---|
| 210 | #--------------------------------------------------------------------------------# |
---|
[1] | 211 | |
---|
[1047] | 212 | |
---|
[1] | 213 | |
---|
[1289] | 214 | # DECLARATION OF VARIABLES AND THEIR DEFUALT VALUES |
---|
[1] | 215 | |
---|
[1289] | 216 | set +o allexport # SUPPRESS EXPORT OF ALL VARIABLES, SINCE IN THE PAST THIS |
---|
| 217 | # LES TO PROBLEMS IN ROUTINES CALLED BY MRUN |
---|
| 218 | # (TOO MANY ARGUMENTS - PROBLEM) |
---|
| 219 | set +o noclobber # EXISTING FILES ARE ALLOWED TO BE OVERWRITTEN |
---|
[1] | 220 | |
---|
| 221 | AddFilenames="" |
---|
| 222 | additional_conditions="" |
---|
| 223 | add_source_path="" |
---|
| 224 | afname="" |
---|
| 225 | archive_save=true |
---|
| 226 | archive_system=none |
---|
[1320] | 227 | check_namelist_files=false |
---|
[920] | 228 | combine_plot_fields=true |
---|
[1] | 229 | compiler_name="" |
---|
| 230 | cond1="" |
---|
| 231 | cond2="" |
---|
| 232 | config_file=.mrun.config |
---|
[206] | 233 | coupled_dist="" |
---|
| 234 | coupled_mode="mpi1" |
---|
[1] | 235 | cpp_opts="" |
---|
[82] | 236 | cpp_options="" |
---|
[1] | 237 | cpumax=0 |
---|
| 238 | cpurest=0 |
---|
[1199] | 239 | create_executable_for_batch=false |
---|
[1] | 240 | delete_temporary_catalog=true |
---|
| 241 | do_batch=false |
---|
| 242 | do_compile=true |
---|
| 243 | do_remote=false |
---|
| 244 | do_stagein=true |
---|
| 245 | do_stageout=true |
---|
| 246 | do_trace=false |
---|
[352] | 247 | email_notification="none" |
---|
[1] | 248 | exclude="" |
---|
| 249 | executable="" |
---|
| 250 | execution_error=false |
---|
[1210] | 251 | fftw_inc="" |
---|
| 252 | fftw_lib="" |
---|
| 253 | fftw_support=false |
---|
[1289] | 254 | fname=test |
---|
[1] | 255 | fromhost="" |
---|
[78] | 256 | global_revision="" |
---|
[122] | 257 | group_number=none |
---|
[1] | 258 | host="" |
---|
| 259 | host_file="" |
---|
| 260 | hp="" |
---|
| 261 | ignore_archive_error=false |
---|
| 262 | input_list="" |
---|
| 263 | interpreted_config_file="" |
---|
[164] | 264 | job_catalog="~/job_queue" |
---|
[1] | 265 | job_on_file="" |
---|
| 266 | keep_data_from_previous_run=false |
---|
[183] | 267 | link_local_input=false |
---|
[149] | 268 | link_local_output=false |
---|
[1] | 269 | localhost_realname=$(hostname) |
---|
[260] | 270 | local_dvrserver_running=.FALSE. |
---|
[1] | 271 | locat=normal |
---|
| 272 | mainprog="" |
---|
| 273 | makefile="" |
---|
[759] | 274 | max_par_io_str="" |
---|
[1] | 275 | mc=$0 |
---|
| 276 | while [[ $(echo $mc | grep -c "/") != 0 ]] |
---|
| 277 | do |
---|
| 278 | mc=`echo $mc | cut -f2- -d"/"` |
---|
| 279 | done |
---|
[493] | 280 | module_calls="" |
---|
[1] | 281 | mrun_script_name=$mc |
---|
[82] | 282 | netcdf_inc="" |
---|
| 283 | netcdf_lib="" |
---|
[1] | 284 | netcdf_support=false |
---|
| 285 | node_usage=default |
---|
| 286 | numprocs="" |
---|
[206] | 287 | numprocs_atmos=0 |
---|
| 288 | numprocs_ocean=0 |
---|
[1] | 289 | OOPT="" |
---|
| 290 | openmp=false |
---|
| 291 | output_list="" |
---|
| 292 | package_list="" |
---|
| 293 | queue=none |
---|
| 294 | read_from_config="" |
---|
[5] | 295 | restart_run=false |
---|
[622] | 296 | if [[ `hostname` = rte10 ]] |
---|
[1] | 297 | then |
---|
[1468] | 298 | return_address=133.5.185.60 |
---|
| 299 | echo "+++ WARNING: fixed return_address = $return_address is used !!!!!" |
---|
[693] | 300 | elif [[ `hostname` = climate0 ]] |
---|
| 301 | then |
---|
[1468] | 302 | return_address=165.132.26.68 |
---|
| 303 | echo "+++ WARNING: fixed return_address = $return_address is used !!!!!" |
---|
[1021] | 304 | elif [[ `hostname` = urban00 ]] |
---|
| 305 | then |
---|
[1468] | 306 | return_address=147.46.30.151 |
---|
| 307 | echo "+++ WARNING: fixed return_address = $return_address is used !!!!!" |
---|
[622] | 308 | else |
---|
[1468] | 309 | return_address=$(nslookup `hostname` 2>&1 | grep "Address:" | tail -1 | awk '{print $2}') |
---|
[1] | 310 | fi |
---|
| 311 | return_password="" |
---|
| 312 | return_username=$LOGNAME |
---|
| 313 | remotecall=false |
---|
| 314 | remote_username="" |
---|
[102] | 315 | run_coupled_model=false |
---|
[1] | 316 | run_mode="" |
---|
[397] | 317 | store_on_archive_system=false |
---|
[1289] | 318 | dashes=" ----------------------------------------------------------------------------" |
---|
[1] | 319 | silent=false |
---|
| 320 | source_list="" |
---|
| 321 | source_path=SOURCE |
---|
| 322 | tasks_per_node="" |
---|
| 323 | threads_per_task=1 |
---|
| 324 | tmpcreate=false |
---|
| 325 | tmp_data_catalog="" |
---|
| 326 | transfer_problems=false |
---|
| 327 | usern=$LOGNAME |
---|
[377] | 328 | use_openmp=false |
---|
[1289] | 329 | version="MRUN 2.1 Rev$Rev: 1758 $" |
---|
[1] | 330 | working_directory=`pwd` |
---|
| 331 | TOPT="" |
---|
| 332 | XOPT="" |
---|
| 333 | zeit=$( date | cut -c 12-19 ) |
---|
| 334 | |
---|
[1124] | 335 | typeset -i iec=0 iic=0 iin=0 ioc=0 iout=0 stagein_anz=0 stageout_anz=0 |
---|
[1289] | 336 | typeset -x -i memory=0 # HAS TO BE EXPORTED HERE, OTHERWISE AN UNKNOWN SIDE |
---|
| 337 | # SIDE EFFECT MAY CAUSE DATA LOSS WHEN GETOPTS IS READING THE |
---|
| 338 | # SCRIPT-OPTION ARGUMENTS |
---|
[1279] | 339 | typeset -i cputime i ii iia iii iio icycle inode ival jobges jobsek last_char_int maxcycle minuten nodes pes remaining_pes sekunden tp1 |
---|
[1] | 340 | |
---|
[1289] | 341 | # ERROR HANDLING IN CASE OF EXIT |
---|
[83] | 342 | trap 'rm -rf $working_directory/tmp_mrun |
---|
[1289] | 343 | if [[ $locat != normal && $locat != control_c ]] |
---|
[1] | 344 | then |
---|
| 345 | |
---|
[1289] | 346 | # CARRY OUT ERROR-COMMANDS GIVEN IN THE CONFIGURATION FILE (EC:) |
---|
[1] | 347 | (( i = 0 )) |
---|
| 348 | while (( i < iec )) |
---|
| 349 | do |
---|
| 350 | (( i = i + 1 )) |
---|
| 351 | printf "\n *** Execution of ERROR-command:\n" |
---|
| 352 | printf " >>> ${err_command[$i]}\n" |
---|
| 353 | eval ${err_command[$i]} |
---|
| 354 | done |
---|
[1101] | 355 | if [[ -n $interpreted_config_file ]] |
---|
| 356 | then |
---|
[1] | 357 | rm -rf $interpreted_config_file |
---|
| 358 | fi |
---|
[1101] | 359 | if [[ -n .mrun_environment ]] |
---|
| 360 | then |
---|
[71] | 361 | rm -rf .mrun_environment |
---|
| 362 | fi |
---|
[1] | 363 | if [[ $tmpcreate = true ]] |
---|
| 364 | then |
---|
| 365 | printf "\n *** Contents of \"$TEMPDIR\":\n" |
---|
| 366 | ls -al; cd |
---|
| 367 | [[ $delete_temporary_catalog = true ]] && rm -rf $TEMPDIR |
---|
| 368 | fi |
---|
[260] | 369 | if [[ "$dvrserver_id" != "" ]] |
---|
| 370 | then |
---|
| 371 | echo "+++ killing dvrserver_id=$dvrserver_id" |
---|
| 372 | kill $dvrserver_id |
---|
| 373 | fi |
---|
[1] | 374 | if [[ -f ~/job_queue/JOBINFO.$QSUB_REQID ]] |
---|
| 375 | then |
---|
| 376 | rm -rf ~/job_queue/JOBINFO.$QSUB_REQID |
---|
| 377 | fi |
---|
| 378 | printf "\n\n+++ MRUN killed \n\n" |
---|
| 379 | elif [[ $locat != control_c ]] |
---|
| 380 | then |
---|
| 381 | printf "\n\n --> all actions finished\n\n" |
---|
| 382 | printf " Bye, bye $usern !!\n\n" |
---|
| 383 | fi' exit |
---|
| 384 | |
---|
| 385 | |
---|
[1289] | 386 | # ACTIONS IN CASE OF TERMINAL-BREAK (CONTROL-C): |
---|
[83] | 387 | trap 'rm -rf $working_directory/tmp_mrun |
---|
[807] | 388 | rm -rf $working_directory/tmp_check_namelist_files |
---|
[69] | 389 | [[ $tmpcreate = true ]] && (cd; rm -rf $TEMPDIR) |
---|
[1] | 390 | if [[ -f ~/job_queue/JOBINFO.$QSUB_REQID ]] |
---|
| 391 | then |
---|
| 392 | rm -rf ~/job_queue/JOBINFO.$QSUB_REQID |
---|
| 393 | fi |
---|
[260] | 394 | if [[ "$dvrserver_id" != "" ]] |
---|
| 395 | then |
---|
| 396 | echo "+++ killing dvrserver_id=$dvrserver_id" |
---|
| 397 | kill $dvrserver_id |
---|
| 398 | fi |
---|
[1] | 399 | printf "\n+++ MRUN killed by \"^C\" \n\n" |
---|
| 400 | locat=control_c |
---|
| 401 | exit |
---|
| 402 | ' 2 |
---|
| 403 | |
---|
| 404 | |
---|
[22] | 405 | # CHECK IF THE PATH FOR THE PALM BINARIES (SCRIPTS+UTILITY-PROGRAMS) HAS |
---|
| 406 | # BEEN SET |
---|
| 407 | if [[ "$PALM_BIN" = "" ]] |
---|
| 408 | then |
---|
| 409 | printf "\n +++ environment variable PALM_BIN has not been set" |
---|
| 410 | printf "\n please set it to the directory where the PALM scripts are located" |
---|
| 411 | locat=palm_bin; exit |
---|
| 412 | fi |
---|
[180] | 413 | export PATH=$PALM_BIN:$PATH |
---|
[22] | 414 | |
---|
| 415 | |
---|
[1] | 416 | |
---|
[1289] | 417 | # READ SHELLSCRIPT-OPTIONS AND REBUILD THE MRUN-COMMAND STRING (MC), |
---|
| 418 | # WHICH WILL BE USED TO START RESTART-JOBS |
---|
[920] | 419 | while getopts :a:AbBc:Cd:D:Fg:G:h:H:i:IkK:m:M:n:o:O:p:P:q:r:R:s:St:T:u:U:vw:xX:yY:zZ option |
---|
[1] | 420 | do |
---|
| 421 | case $option in |
---|
| 422 | (a) afname=$OPTARG;; |
---|
[397] | 423 | (A) store_on_archive_system=true; mc="$mc -A";; |
---|
[1] | 424 | (b) do_batch=true; mc="$mc -b";; |
---|
| 425 | (B) delete_temporary_catalog=false; mc="$mc -B";; |
---|
| 426 | (c) config_file=$OPTARG; mc="$mc -c$OPTARG";; |
---|
[5] | 427 | (C) restart_run=true; mc="$mc -C";; |
---|
[1] | 428 | (d) fname=$OPTARG; mc="$mc -d$OPTARG";; |
---|
[210] | 429 | (D) cpp_opts="$cpp_opts $OPTARG"; mc="$mc -D'$OPTARG'";; |
---|
[79] | 430 | (F) job_on_file="-D"; mc="$mc -F";; |
---|
[122] | 431 | (g) group_number=$OPTARG; mc="$mc -g$OPTARG";; |
---|
[78] | 432 | (G) global_revision=$OPTARG; mc="$mc -G'$OPTARG'";; |
---|
[1] | 433 | (h) host=$OPTARG; mc="$mc -h$OPTARG";; |
---|
| 434 | (H) fromhost=$OPTARG; mc="$mc -H$OPTARG";; |
---|
| 435 | (i) input_list=$OPTARG; mc="$mc -i'$OPTARG'";; |
---|
| 436 | (I) ignore_archive_error=true; mc="$mc -I";; |
---|
| 437 | (k) keep_data_from_previous_run=true; mc="$mc -k";; |
---|
[210] | 438 | (K) additional_conditions="$OPTARG"; mc="$mc -K'$OPTARG'";; |
---|
[1] | 439 | (m) memory=$OPTARG; mc="$mc -m$OPTARG";; |
---|
| 440 | (M) makefile=$OPTARG; mc="$mc -M$OPTARG";; |
---|
| 441 | (n) node_usage=$OPTARG; mc="$mc -n$OPTARG";; |
---|
| 442 | (o) output_list=$OPTARG; mc="$mc -o'$OPTARG'";; |
---|
[622] | 443 | (O) use_openmp=true; threads_per_task=$OPTARG; mc="$mc -O$OPTARG";; |
---|
[1] | 444 | (p) package_list=$OPTARG; mc="$mc -p'$OPTARG'";; |
---|
| 445 | (P) return_password=$OPTARG; mc="$mc -P$OPTARG";; |
---|
| 446 | (q) queue=$OPTARG; mc="$mc -q$OPTARG";; |
---|
| 447 | (r) run_mode=$OPTARG; mc="$mc -r'$OPTARG'";; |
---|
[1468] | 448 | (R) remotecall=true;return_address=$OPTARG; mc="$mc -R$OPTARG";; |
---|
[1] | 449 | (s) source_list=$OPTARG; mc="$mc -s'$OPTARG'";; |
---|
| 450 | (S) read_from_config=false; mc="$mc -S";; |
---|
| 451 | (t) cpumax=$OPTARG; mc="$mc -t$OPTARG";; |
---|
[759] | 452 | (T) mrun_tasks_per_node=$OPTARG; mc="$mc -T$OPTARG";; |
---|
[1] | 453 | (u) remote_username=$OPTARG; mc="$mc -u$OPTARG";; |
---|
| 454 | (U) return_username=$OPTARG; mc="$mc -U$OPTARG";; |
---|
| 455 | (v) silent=true; mc="$mc -v";; |
---|
[759] | 456 | (w) max_par_io_str=$OPTARG; mc="$mc -w$OPTARG";; |
---|
[1] | 457 | (x) do_trace=true;set -x; mc="$mc -x";; |
---|
| 458 | (X) numprocs=$OPTARG; mc="$mc -X$OPTARG";; |
---|
[291] | 459 | (y) ocean_file_appendix=true; mc="$mc -y";; |
---|
[206] | 460 | (Y) run_coupled_model=true; coupled_dist=$OPTARG; mc="$mc -Y'$OPTARG'";; |
---|
[807] | 461 | (z) check_namelist_files=false; mc="$mc -z";; |
---|
[920] | 462 | (Z) combine_plot_fields=false; mc="$mc -Z";; |
---|
[1] | 463 | (\?) printf "\n +++ unknown option $OPTARG \n" |
---|
| 464 | printf "\n --> type \"$0 ?\" for available options \n" |
---|
| 465 | locat=parameter;exit;; |
---|
| 466 | esac |
---|
| 467 | done |
---|
| 468 | |
---|
| 469 | |
---|
[1289] | 470 | # SKIP GIVEN OPTIONS TO READ POSITIONAL PARAMETER, IF GIVEN |
---|
| 471 | # CURRENTLY ONLY PARAMETER ? (TO OUTPUT A SHORT COMMAND INFO) IS ALLOWED |
---|
[1101] | 472 | (( to_shift = $OPTIND - 1 )) |
---|
| 473 | shift $to_shift |
---|
[1] | 474 | |
---|
[1289] | 475 | # PRINT SHORT DESCRIPTION OF MRUN OPTIONS |
---|
[1] | 476 | if [[ "$1" = "?" ]] |
---|
| 477 | then |
---|
| 478 | (printf "\n *** mrun can be called as follows:\n" |
---|
[920] | 479 | printf "\n $mrun_script_name -b -c.. -d.. -D.. -f.. -F -h.. -i.. -I -K.. -m.. -o.. -p.. -r.. -R -s.. -t.. -T.. -v -x -X.. -y -Y.. -z -Z <modus> \n" |
---|
[1] | 480 | printf "\n Description of available options:\n" |
---|
| 481 | printf "\n Option Description Default-Value" |
---|
| 482 | printf "\n -a base name of input files equiv. -d" |
---|
[400] | 483 | printf "\n -A archiving when using file-attribute fl" |
---|
[1] | 484 | printf "\n -b batch-job on local machine ---" |
---|
| 485 | printf "\n -B do not delete temporary directory at end ---" |
---|
| 486 | printf "\n -c configuration file .mrun.config" |
---|
| 487 | printf "\n -d base name of files attached to program test" |
---|
| 488 | printf "\n -D preprocessor(cpp)-directives \"\" " |
---|
| 489 | printf "\n -F create remote job file only ---" |
---|
| 490 | printf "\n -h execution host $localhost_realname" |
---|
| 491 | printf "\n -i INPUT control list \"\" " |
---|
| 492 | printf "\n -I archiving errors of previous batch-jobs" |
---|
| 493 | printf "\n will be ignored" |
---|
| 494 | printf "\n -k keep data from previous run" |
---|
| 495 | printf "\n -K additional conditions for controling" |
---|
| 496 | printf "\n usage of conditional code and" |
---|
| 497 | printf "\n env-variables in configuration file \"\" " |
---|
| 498 | printf "\n -m memory demand in MB (batch-jobs) 0 MB" |
---|
| 499 | printf "\n -M Makefile name Makefile" |
---|
| 500 | printf "\n -n node usage (shared/not_shared) depending on -h" |
---|
| 501 | printf "\n -o OUTPUT control list \"\" " |
---|
[1304] | 502 | printf "\n -O threads per openMP task ---" |
---|
[1] | 503 | printf "\n -p software package list \"\" " |
---|
| 504 | printf "\n -q queue \"$queue\" " |
---|
| 505 | printf "\n -r run control list (combines -i -o) \"\" " |
---|
| 506 | printf "\n -s filenames of routines to be compiled \"\" " |
---|
| 507 | printf "\n must end with .f, .f90, .F, or .c !" |
---|
| 508 | printf "\n use \"..\" for more than one file and wildcards" |
---|
[837] | 509 | printf "\n -s LM compiles all locally modified files" |
---|
[1] | 510 | printf "\n -S config file interpreted by shellscript ---" |
---|
| 511 | printf "\n -t allowed cpu-time in seconds (batch) 0" |
---|
| 512 | printf "\n -T tasks per node depending on -h" |
---|
| 513 | printf "\n -u username on remote machine \"\" " |
---|
| 514 | printf "\n -v no prompt for confirmation ---" |
---|
[759] | 515 | printf "\n -w maximum parallel io streams as given by -X" |
---|
[1] | 516 | printf "\n -x tracing of mrun for debug purposes ---" |
---|
| 517 | printf "\n -X # of processors (on parallel machines) 1" |
---|
[291] | 518 | printf "\n -y add appendix \"_O\" to all local output" |
---|
| 519 | printf "\n files (ocean precursor runs followed by" |
---|
| 520 | printf "\n coupled atmosphere-ocean runs) ---" |
---|
[206] | 521 | printf "\n -Y run coupled model, \"#1 #2\" with" |
---|
| 522 | printf "\n #1 atmosphere and #2 ocean processors \"#/2 #/2\" depending on -X" |
---|
[920] | 523 | printf "\n -z disable a priori parameter file check ---" |
---|
| 524 | printf "\n -Z skip combine_plot_fields at the end of " |
---|
| 525 | printf "\n the simulation ---" |
---|
[1] | 526 | printf "\n " |
---|
| 527 | printf "\n Possible values of positional parameter <modus>:" |
---|
| 528 | printf "\n \"?\" - this outline \n\n") | more |
---|
| 529 | exit |
---|
| 530 | elif [[ "$1" != "" ]] |
---|
| 531 | then |
---|
| 532 | printf "\n +++ positional parameter $1 unknown \n" |
---|
| 533 | locat=parameter; exit |
---|
| 534 | fi |
---|
| 535 | |
---|
| 536 | |
---|
| 537 | |
---|
[1289] | 538 | # SHORT STARTING MESSAGE |
---|
[1] | 539 | printf "\n*** $version " |
---|
| 540 | printf "\n will be executed. Please wait ..." |
---|
| 541 | |
---|
| 542 | |
---|
[82] | 543 | |
---|
[1289] | 544 | # CHECK, IF CONFIGURATION FILE EXISTS |
---|
[82] | 545 | if [[ ! -f $config_file ]] |
---|
| 546 | then |
---|
| 547 | printf "\n\n +++ configuration file: " |
---|
| 548 | printf "\n $config_file" |
---|
| 549 | printf "\n does not exist" |
---|
| 550 | locat=connect; exit |
---|
| 551 | fi |
---|
| 552 | |
---|
| 553 | |
---|
[1289] | 554 | # DETERMINE THE HOST-IDENTIFIER (localhost) FROM THE CONFIGURATION FILE |
---|
[82] | 555 | line="" |
---|
| 556 | grep "%host_identifier" $config_file > tmp_mrun |
---|
| 557 | while read line |
---|
| 558 | do |
---|
[206] | 559 | if [[ "$line" != "" && $(echo $line | cut -c1) != "#" ]] |
---|
[82] | 560 | then |
---|
| 561 | HOSTNAME=`echo $line | cut -d" " -s -f2` |
---|
| 562 | host_identifier=`echo $line | cut -d" " -s -f3` |
---|
[83] | 563 | if [[ $localhost_realname = $HOSTNAME ]] |
---|
[82] | 564 | then |
---|
| 565 | localhost=$host_identifier |
---|
| 566 | break |
---|
| 567 | fi |
---|
| 568 | fi |
---|
| 569 | done < tmp_mrun |
---|
| 570 | |
---|
| 571 | if [[ "$localhost" = "" ]] |
---|
| 572 | then |
---|
| 573 | printf "\n\n +++ no host identifier found in configuration file \"$config_file\"" |
---|
| 574 | printf "\n for local host \"$localhost_realname\"." |
---|
| 575 | printf "\n Please add line" |
---|
| 576 | printf "\n \"\%host_identifier $localhost_realname <identifier>\"" |
---|
| 577 | printf "\n to the configuration file." |
---|
| 578 | locat=localhost; exit |
---|
| 579 | fi |
---|
| 580 | |
---|
| 581 | |
---|
[1289] | 582 | # SET HOST-SPECIFIC VARIABLES |
---|
[82] | 583 | case $localhost_realname in |
---|
[397] | 584 | (r1*|r2*|h01*|b01*) archive_system=tivoli;; |
---|
[82] | 585 | (cs*) archive_system=ut;; |
---|
| 586 | esac |
---|
| 587 | |
---|
| 588 | |
---|
[1289] | 589 | # ??? THIS SHOULD BE DECRIBED IN THE MRUN DOCUMENTATION ??? |
---|
| 590 | # SET BASENAME OF THE INPUT-FILES TO THE GENERAL BASENAME (GIVEN BY OPTION -d), |
---|
| 591 | # IF IT HAS NOT BEEN SET BY THE USER DIFFERENTLY (USING OPTION -a) |
---|
[1] | 592 | [[ "$afname" = "" ]] && afname=$fname |
---|
| 593 | |
---|
| 594 | |
---|
[1289] | 595 | # ADD RUN-FILE ACTIVATION STRINGS (GIVEN BY OPTION -r) TO THE |
---|
| 596 | # RESPECTIVE LISTS FOR ACTIVATING THE INPUT/OUTPUT FILE CONNECTION |
---|
| 597 | # STATEMENTS IN THE CONFIGURATION FILE |
---|
[1] | 598 | if [[ "$run_mode" != "" ]] |
---|
| 599 | then |
---|
| 600 | input_list="$input_list $run_mode" |
---|
| 601 | output_list="$output_list $run_mode" |
---|
| 602 | fi |
---|
| 603 | |
---|
| 604 | |
---|
[1289] | 605 | # ??? is this (and the respective option -H) really required ??? |
---|
| 606 | # STORE HOSTNAME, FROM WHICH THE JOB IS STARTED, |
---|
| 607 | # BUT ONLY IF IT HASN'T BEEN ALREADY SET BY OPTION -H |
---|
| 608 | # (MRUN IS AUTOMATICALLY SETTING THIS OPTION IN THE MRUN-CALLS WITHIN |
---|
| 609 | # THOSE BATCH-JOBS, WHICH ARE CREATED BY MRUN ITSELF) |
---|
[1] | 610 | if [[ "$fromhost" = "" ]] |
---|
| 611 | then |
---|
| 612 | fromhost=$localhost |
---|
| 613 | fi |
---|
| 614 | |
---|
| 615 | |
---|
[1289] | 616 | # CHECK, IF EXECUTION SHALL TO BE DONE ON THE REMOTE-MACHINE. |
---|
| 617 | # VALUE OF do_remote IS IMPORTANT FOR THE FILE CONNECTIONS. |
---|
| 618 | # IN CASE OF EXECUTION ON A REMOTE-MACHINE, A BATCH-JOB HAS |
---|
[1609] | 619 | # TO BE SUBMITTED (I.E. do_batch=true) |
---|
[1] | 620 | if [[ -n $host && "$host" != $localhost ]] |
---|
| 621 | then |
---|
| 622 | do_batch=true |
---|
| 623 | do_remote=true |
---|
| 624 | case $host in |
---|
[1620] | 625 | (ibm|ibmh|ibmkisti|ibmku|ibms|nech|lcbullhh|lccrayb|lccrayh|lccrayf|lcflow|lckyoto|unics|lcxe6|lcxt5m|lck|lckiaps|lckordi|lckyuh|lckyut|lcsb) true;; |
---|
[1] | 626 | (*) printf "\n" |
---|
[83] | 627 | printf "\n +++ sorry: execution of batch jobs on remote host \"$host\"" |
---|
| 628 | printf "\n is not available" |
---|
[1] | 629 | locat=nqs; (( iec = 0 )); exit;; |
---|
| 630 | esac |
---|
| 631 | else |
---|
| 632 | host=$localhost |
---|
| 633 | fi |
---|
| 634 | |
---|
| 635 | |
---|
[1289] | 636 | # EVALUATE ADDITIONAL CONDITIONS GIVEN BY OPTION -K |
---|
[1] | 637 | if [[ -n $additional_conditions ]] |
---|
| 638 | then |
---|
[69] | 639 | cond1=`echo $additional_conditions | cut -d" " -f1` |
---|
| 640 | cond2=`echo $additional_conditions | cut -d" " -s -f2` |
---|
| 641 | dummy=`echo $additional_conditions | cut -d" " -s -f3` |
---|
[1] | 642 | if [[ -n $dummy ]] |
---|
| 643 | then |
---|
| 644 | printf "\n +++ more than 2 additional conditions given for Option \"-K\"" |
---|
| 645 | locat=options; exit |
---|
| 646 | fi |
---|
[503] | 647 | block=_$cond1 |
---|
| 648 | [[ -n $cond2 ]] && block=${block}_$cond2 |
---|
[1] | 649 | fi |
---|
| 650 | |
---|
| 651 | |
---|
[1289] | 652 | # EVALUATE MODEL COUPLING FEATURES (OPTION -Y) AND DETERMINE coupled_mode |
---|
[206] | 653 | if [[ $run_coupled_model = true ]] |
---|
| 654 | then |
---|
| 655 | |
---|
[1289] | 656 | numprocs_atmos=`echo $coupled_dist | cut -d" " -s -f1` |
---|
| 657 | numprocs_ocean=`echo $coupled_dist | cut -d" " -s -f2` |
---|
| 658 | |
---|
| 659 | if (( $numprocs_ocean + $numprocs_atmos != $numprocs )) |
---|
[206] | 660 | then |
---|
| 661 | |
---|
[1289] | 662 | printf "\n +++ number of processors does not fit to specification by \"-Y\"." |
---|
| 663 | printf "\n PEs (total) : $numprocs" |
---|
| 664 | printf "\n PEs (atmosphere): $numprocs_atmos" |
---|
| 665 | printf "\n PEs (ocean) : $numprocs_ocean" |
---|
| 666 | locat=coupling; exit |
---|
[206] | 667 | |
---|
[1289] | 668 | fi |
---|
[206] | 669 | |
---|
| 670 | |
---|
| 671 | # GET coupled_mode FROM THE CONFIG FILE |
---|
| 672 | line="" |
---|
| 673 | grep "%cpp_options.*-D__mpi2.*$host" $config_file > tmp_mrun |
---|
| 674 | while read line |
---|
| 675 | do |
---|
[1069] | 676 | if [[ "$line" != "" && $(echo $line | cut -c1) != "#" && $(echo $line | cut -d" " -s -f4) = $cond1 && $(echo $line | cut -d" " -s -f4) = $cond2 ]] |
---|
[206] | 677 | then |
---|
| 678 | coupled_mode="mpi2" |
---|
| 679 | fi |
---|
| 680 | done < tmp_mrun |
---|
| 681 | |
---|
| 682 | fi |
---|
| 683 | |
---|
[1289] | 684 | # CHECK, IF FILE-ARCHIVING HAS FAILED IN PREVIOUS JOB (OF A JOB-CHAIN) |
---|
[1] | 685 | if [[ -f ~/job_queue/ARCHIVE_ERROR_$fname ]] |
---|
| 686 | then |
---|
| 687 | if [[ $ignore_archive_error = false ]] |
---|
| 688 | then |
---|
| 689 | printf "\n +++ data archiving of previous run failed" |
---|
| 690 | printf "\n see directory \~/job_queue on remote machine" |
---|
| 691 | locat=archive; exit |
---|
| 692 | else |
---|
| 693 | printf "\n +++ warning: data archiving in a previous run failed" |
---|
| 694 | printf "\n MRUN continues, trying to get backup copy" |
---|
| 695 | fi |
---|
| 696 | fi |
---|
| 697 | |
---|
| 698 | |
---|
[1289] | 699 | # SAVE VALUES OF MRUN-OPTIONS SICHERN IN ORDER TO OVERWRITE |
---|
| 700 | # THOSE VALUES GIVEN IN THE CONFIGURATION-FILE |
---|
[1099] | 701 | mrun_memory=$memory |
---|
| 702 | mrun_group_number=$group_number |
---|
| 703 | mrun_cpumax=$cpumax |
---|
| 704 | mrun_numprocs=$numprocs |
---|
[1] | 705 | |
---|
[1099] | 706 | |
---|
[1289] | 707 | # READ AND EVALUATE THE CONFIGURATION-FILE FROM WITHIN THIS SHELLSCRIPT |
---|
[1468] | 708 | # (OPTION -S). THE DEFAULT IS USING THE ROUTINE <<<< |
---|
[1] | 709 | if [[ "$read_from_config" = false ]] |
---|
| 710 | then |
---|
| 711 | [[ $silent = false ]] && printf "\n Reading the configuration file... " |
---|
[1289] | 712 | |
---|
| 713 | # READ THE CONFIGURATION FILE LINE BY LINE |
---|
[1] | 714 | while read zeile |
---|
| 715 | do |
---|
| 716 | [[ $silent = false ]] && printf "." |
---|
| 717 | |
---|
| 718 | |
---|
[1289] | 719 | # FIRST REPLACE ENVIRONMENT-VARIABLES BY THEIR RESPECTIVE VALUES |
---|
[1] | 720 | eval zeile=\"$zeile\" |
---|
| 721 | |
---|
| 722 | |
---|
[1289] | 723 | # INTERPRET THE LINE |
---|
[21] | 724 | if [[ "$(echo $zeile)" = "" ]] |
---|
[1] | 725 | then |
---|
[1289] | 726 | # EMPTY LINE, NO ACTION |
---|
[21] | 727 | continue |
---|
[1] | 728 | |
---|
[21] | 729 | elif [[ "$(echo $zeile | cut -c1)" = "#" ]] |
---|
| 730 | then |
---|
| 731 | |
---|
[1289] | 732 | # LINE IS A COMMENT LINE |
---|
[1] | 733 | true |
---|
| 734 | |
---|
| 735 | elif [[ "$(echo $zeile | cut -c1)" = "%" ]] |
---|
| 736 | then |
---|
| 737 | |
---|
[1289] | 738 | # LINE DEFINES AN ENVIRONMENT-VARIABLE |
---|
[1] | 739 | zeile=$(echo $zeile | cut -c2-) |
---|
[69] | 740 | var=`echo $zeile | cut -d" " -f1` |
---|
| 741 | value=`echo $zeile | cut -d" " -s -f2` |
---|
| 742 | for_host=`echo $zeile | cut -d" " -s -f3` |
---|
| 743 | for_cond1=`echo $zeile | cut -d" " -s -f4` |
---|
| 744 | for_cond2=`echo $zeile | cut -d" " -s -f5` |
---|
[1] | 745 | |
---|
| 746 | if [[ "$for_host" = "" || ( "$for_host" = $host && "$for_cond1" = "$cond1" && "$for_cond2" = "$cond2" ) || $(echo "$input_list$output_list"|grep -c "$for_host") != 0 ]] |
---|
| 747 | then |
---|
| 748 | |
---|
[1289] | 749 | # REPLACE ":" BY " " IN COMPILER- CPP- OR LINKER-OPTIONS, |
---|
| 750 | # "::" IS REPLACED BY ":". |
---|
[1] | 751 | value=`echo $value | sed 's/::/%DUM%/g' | sed 's/:/ /g' | sed 's/%DUM%/:/g'` |
---|
| 752 | |
---|
| 753 | |
---|
[1289] | 754 | # VALUE FROM THE CONFIGURATION-FILE IS ASSIGNED TO THE |
---|
| 755 | # ENVIRONMENT-VARIABLE, BUT ONLY IF NO VALUE HAS BEEN ALREADY |
---|
| 756 | # ASSIGNED WITHIN THIS SCRIPT (E.G. BY SCRIPT-OPTIONS). |
---|
| 757 | # NON-ASSIGNED VARIABLES HAVE VALUE "" OR 0 (IN CASE OF INTEGER). |
---|
| 758 | # HENCE THE GENERAL RULE IS: SCRIPT-OPTION OVERWRITES THE |
---|
| 759 | # CONFIGURATION-FILE. |
---|
[1] | 760 | if [[ "$(eval echo \$$var)" = "" || "$(eval echo \$$var)" = "0" ]] |
---|
| 761 | then |
---|
| 762 | eval $var=\$value |
---|
| 763 | |
---|
[1289] | 764 | # TERMINAL OUTPUT OF ENVIRONMENT-VARIABLES, IF TRACEBACK IS SWITCHED on |
---|
[1] | 765 | if [[ $do_trace = true ]] |
---|
| 766 | then |
---|
| 767 | printf "\n*** ENVIRONMENT-VARIABLE $var = $value" |
---|
| 768 | fi |
---|
| 769 | fi |
---|
| 770 | |
---|
[1289] | 771 | # IF AN ENVIRONMENT-VARIABLE DETERMINES THE HOST, THEN EVALUATE IT IMMEDIATELY: |
---|
| 772 | # VALUE OF do-remote IS REQUIRED FOR THE FILE CONNECTIONS (COPY OF I/O FILES). |
---|
| 773 | # IF EXECUTION IS SCHEDULED FOR A REMOTE-MASCHINE, A BATCH-JOB |
---|
| 774 | # MUST HAVE TO BE STARTED |
---|
[1] | 775 | if [[ $var = host ]] |
---|
| 776 | then |
---|
| 777 | if [[ -n $host && "$host" != $localhost ]] |
---|
| 778 | then |
---|
| 779 | do_batch=true |
---|
| 780 | do_remote=true |
---|
| 781 | case $host in |
---|
[1620] | 782 | (ibm|ibmh|ibmkisti|ibmku|ibms|lcbullhh|lccrayb|lccrayh|lccrayf|lcflow|lckyoto|nech|unics|lcxe6|lcxt5m|lck|lckiaps|lckordi|lckyuh|lckyut|lcsb) true;; |
---|
[83] | 783 | (*) printf "\n +++ sorry: execution of batch jobs on remote host \"$host\"" |
---|
| 784 | printf "\n is not available" |
---|
[1] | 785 | locat=nqs; exit;; |
---|
| 786 | esac |
---|
| 787 | else |
---|
| 788 | host=$localhost |
---|
| 789 | fi |
---|
| 790 | fi |
---|
| 791 | |
---|
[1289] | 792 | # USER-DEFINED ENVIRONMENT VARIABLES MUST BE EXPORTED, |
---|
| 793 | # BECAUSE THEY MAY BE REQUIRED IN OTHER SCRIPTS CALLED |
---|
| 794 | # BY MRUN (E.G. subjob) |
---|
[1] | 795 | export $var |
---|
| 796 | fi |
---|
| 797 | |
---|
| 798 | |
---|
| 799 | |
---|
| 800 | |
---|
| 801 | elif [[ "$(echo $zeile | cut -c1-3)" = "EC:" ]] |
---|
| 802 | then |
---|
| 803 | |
---|
[1289] | 804 | # LINE DEFINES ERROR-COMMAND |
---|
[1] | 805 | (( iec = iec + 1 )) |
---|
| 806 | zeile=$(echo $zeile | cut -c4-) |
---|
| 807 | err_command[$iec]="$zeile" |
---|
| 808 | |
---|
| 809 | elif [[ "$(echo $zeile | cut -c1-3)" = "IC:" ]] |
---|
| 810 | then |
---|
| 811 | |
---|
[1289] | 812 | # LINE DEFINES INPUT-COMMAND |
---|
[1] | 813 | (( iic = iic + 1 )) |
---|
| 814 | zeile=$(echo $zeile | cut -c4-) |
---|
| 815 | in_command[$iic]="$zeile" |
---|
| 816 | |
---|
| 817 | elif [[ "$(echo $zeile | cut -c1-3)" = "OC:" ]] |
---|
| 818 | then |
---|
| 819 | |
---|
[1289] | 820 | # LINE DEFINES OUTPUT-COMMAND |
---|
[1] | 821 | (( ioc = ioc + 1 )) |
---|
| 822 | zeile=$(echo $zeile | cut -c4-) |
---|
| 823 | out_command[$ioc]="$zeile" |
---|
| 824 | |
---|
| 825 | else |
---|
| 826 | |
---|
[1289] | 827 | # LINE DEFINES FILE CONNECTION. READ THE FILE ATTRIBUTES. |
---|
| 828 | # s2a: in/out - field |
---|
| 829 | # s2b: loc - field (optional) |
---|
| 830 | # s2c: tr/ar - field (optional) |
---|
[69] | 831 | s1=`echo $zeile | cut -d" " -f1` |
---|
| 832 | s2=`echo $zeile | cut -d" " -s -f2` |
---|
[1] | 833 | s2a=$(echo $s2 | cut -d":" -f1) |
---|
| 834 | if [[ $(echo $s2 | grep -c ":") = 0 ]] |
---|
| 835 | then |
---|
| 836 | s2b="" |
---|
| 837 | s2c="" |
---|
| 838 | else |
---|
[69] | 839 | s2b=`echo $s2 | cut -d":" -f2 | sed 's/:/ /g'` |
---|
| 840 | s2c=`echo $s2 | cut -d":" -s -f3 | sed 's/:/ /g'` |
---|
[1] | 841 | fi |
---|
[69] | 842 | s3=`echo $zeile | cut -d" " -f3` |
---|
| 843 | s4=`echo $zeile | cut -d" " -s -f4` |
---|
| 844 | s5=`echo $zeile | cut -d" " -s -f5` |
---|
| 845 | s6=`echo $zeile | cut -d" " -s -f6` |
---|
[1] | 846 | |
---|
| 847 | |
---|
[1289] | 848 | # STORE FILE CONNECTION, IF ACTIVATED BY ACTIVATION-STRING FROM |
---|
| 849 | # INPUT- OR OUTPUT-LIST. |
---|
| 850 | # VARIABLE S3 MAY CONTAIN A LIST OF ACTIVATION STRINGS (FIELD-SEPERATOR ":"). |
---|
| 851 | # IF EXECUTION IS SCHEDULED FOR A REMOTE-MACHINE AND THE FILE IS ONLY |
---|
| 852 | # LOCALLY REQUIRED ON THAT MACHINE (I.E. s2b = loc), THE FILE CONNECTION |
---|
| 853 | # IS NOT CHECKED AND STORED. |
---|
[1] | 854 | IFSALT="$IFS"; IFS="$IFS:" |
---|
| 855 | if [[ "$s2a" = in && ! ( $do_remote = true && ( "$s2b" = loc || "$s2b" = locopt ) ) ]] |
---|
| 856 | then |
---|
| 857 | found=false |
---|
| 858 | for actual in $input_list |
---|
| 859 | do |
---|
| 860 | for formal in $s3 |
---|
| 861 | do |
---|
| 862 | [[ $actual = $formal || "$formal" = "-" ]] && found=true |
---|
| 863 | done |
---|
| 864 | done |
---|
| 865 | if [[ $found = true ]] |
---|
| 866 | then |
---|
| 867 | (( iin = iin + 1 )) |
---|
| 868 | localin[$iin]=$s1; transin[$iin]=$s2b; actionin[$iin]=$s2c; |
---|
| 869 | typein[$iin]=$s3; pathin[$iin]=$s4; endin[$iin]=$s5; |
---|
| 870 | extin[$iin]=$s6 |
---|
| 871 | fi |
---|
| 872 | elif [[ "$s2a" = out && ! ( $do_remote = true && "$s2b" = loc ) ]] |
---|
| 873 | then |
---|
| 874 | found=false |
---|
| 875 | for actual in $output_list |
---|
| 876 | do |
---|
| 877 | for formal in $s3 |
---|
| 878 | do |
---|
| 879 | [[ $actual = $formal || "$formal" = "-" ]] && found=true |
---|
| 880 | done |
---|
| 881 | done |
---|
| 882 | if [[ $found = true ]] |
---|
| 883 | then |
---|
| 884 | (( iout = iout + 1 )) |
---|
| 885 | localout[$iout]=$s1; actionout[$iout]=$s2c; typeout[$iout]=$s3; |
---|
| 886 | pathout[$iout]=$s4; endout[$iout]=$s5; extout[$iout]=$s6 |
---|
| 887 | fi |
---|
| 888 | elif [[ "$s2a" != in && "$s2a" != out ]] |
---|
| 889 | then |
---|
| 890 | printf "\n +++ I/O-attribute in configuration file $config_file has the invalid" |
---|
| 891 | printf "\n value \"$s2\". Only \"in\" and \"out\" are allowed!" |
---|
| 892 | locat=connect; exit |
---|
| 893 | fi |
---|
| 894 | IFS="$IFSALT" |
---|
| 895 | fi |
---|
| 896 | done < $config_file |
---|
| 897 | |
---|
| 898 | else |
---|
| 899 | |
---|
[1289] | 900 | # EVALUATE THE CONFIGURATION-FILE BY FORTRAN-PROGRAM |
---|
[1] | 901 | [[ $silent = false ]] && printf "..." |
---|
[1289] | 902 | interpreted_config_file=.icf.$RANDOM |
---|
[1] | 903 | |
---|
| 904 | |
---|
[69] | 905 | |
---|
[1289] | 906 | # PROVIDE VALUES OF ENVIRONMENT-VARIABLE FOR interpret_config VIA NAMELIST-FILE |
---|
[69] | 907 | cat > .mrun_environment << %%END%% |
---|
| 908 | &mrun_environment cond1 = '$cond1', cond2 = '$cond2', |
---|
| 909 | config_file = '$config_file', do_remote = '$do_remote', |
---|
| 910 | do_trace = '$do_trace', host = '$host', |
---|
| 911 | input_list = '$input_list', icf = '$interpreted_config_file', |
---|
| 912 | localhost = '$localhost', output_list = '$output_list' / |
---|
| 913 | |
---|
| 914 | %%END%% |
---|
| 915 | |
---|
[1289] | 916 | if [[ "$host" != $localhost ]] |
---|
[66] | 917 | then |
---|
[503] | 918 | |
---|
[1289] | 919 | # REMOTE JOB FROM LOCAL HOST: JUST TAKE THE FIRST EXECUTABLE FOUND |
---|
| 920 | interpret_config_executable=`ls -1 ${PALM_BIN}/interpret_config*.x 2>/dev/null` |
---|
| 921 | if [[ $? != 0 ]] |
---|
[503] | 922 | then |
---|
[1289] | 923 | printf "\n\n +++ no interpret_config found" |
---|
| 924 | printf "\n run \"mbuild -u -h ...\" to generate utilities for this host" |
---|
| 925 | locat=interpret_config; exit |
---|
| 926 | fi |
---|
| 927 | interpret_config_executable=`echo $interpret_config_executable | cut -d" " -f1` |
---|
| 928 | $interpret_config_executable |
---|
[503] | 929 | |
---|
[1289] | 930 | else |
---|
[503] | 931 | |
---|
[1289] | 932 | # CHECK, IF THERE IS AN EXECUTABLE FOR THE BLOCK |
---|
| 933 | if [[ ! -f ${PALM_BIN}/interpret_config${block}.x ]] |
---|
| 934 | then |
---|
| 935 | printf "\n\n +++ no interpret_config found for given block \"$cond1 $cond2\"" |
---|
| 936 | printf "\n run \"mbuild -u -h ...\" to generate utilities for this block" |
---|
| 937 | locat=interpret_config; exit |
---|
[503] | 938 | else |
---|
[1289] | 939 | interpret_config${block}.x |
---|
| 940 | fi |
---|
[503] | 941 | |
---|
[66] | 942 | fi |
---|
[71] | 943 | rm .mrun_environment |
---|
[1] | 944 | |
---|
[71] | 945 | |
---|
[1289] | 946 | # EXECUTE SHELL-COMMANDS GENERATED BY interpret_config WITHIN THIS SHELL |
---|
[1] | 947 | chmod u+x $interpreted_config_file |
---|
| 948 | export PATH=$PATH:. |
---|
| 949 | . $interpreted_config_file |
---|
[71] | 950 | rm $interpreted_config_file |
---|
[1] | 951 | |
---|
[1099] | 952 | fi |
---|
[71] | 953 | |
---|
[69] | 954 | |
---|
[1289] | 955 | # VALUES OF MRUN-OPTIONS OVERWRITE THOSE FROM THE CONFIGURATION-FILE |
---|
[1099] | 956 | [[ $mrun_memory != 0 ]] && memory=$mrun_memory |
---|
| 957 | [[ "$mrun_group_number" != "none" ]] && group_number=$mrun_group_number |
---|
| 958 | [[ $mrun_cpumax != 0 ]] && cpumax=$mrun_cpumax |
---|
| 959 | [[ "$mrun_numprocs" != "" ]] && numprocs=$mrun_numprocs |
---|
| 960 | [[ "$max_par_io_str" != "" ]] && maximum_parallel_io_streams=$max_par_io_str |
---|
| 961 | [[ "$mrun_tasks_per_node" != "" ]] && tasks_per_node=$mrun_tasks_per_node |
---|
[1] | 962 | |
---|
| 963 | |
---|
[1289] | 964 | # ON THE LOCAL MACHINE, DETERMINE (FROM THE CONFIGURATION-FILE) THE PATH |
---|
| 965 | # FOR SOURCE-FILES TO BE COMPILED. |
---|
| 966 | # IN A BATCH-JOB, SOURCES (ROUTINES TO BE COMPILED) ARE COMPLETE ALREADY. |
---|
| 967 | # BEFORE MRUN IS CALLED IN THE JOB, SOURCES_COMPLETE = true IS ASSIGNED (SEE FURTHER BELOW). |
---|
[1] | 968 | if [[ "$SOURCES_COMPLETE" = "" ]] |
---|
| 969 | then |
---|
| 970 | |
---|
[1289] | 971 | # FIRST CHECK, IF A GLOBAL SOURCE PATH (TO BE USED FOR ALL HOSTS) |
---|
| 972 | # HAS BEEN DEFINED |
---|
[1] | 973 | source_path="" |
---|
| 974 | line="" |
---|
| 975 | grep "%source_path" $config_file > tmp_mrun |
---|
| 976 | while read line |
---|
| 977 | do |
---|
| 978 | if [[ "$line" != "" && $(echo $line | cut -c1) != "#" ]] |
---|
| 979 | then |
---|
| 980 | if [[ "$(echo $line | cut -d" " -f3)" = "" ]] |
---|
| 981 | then |
---|
| 982 | global_source_path=`echo $line | cut -d" " -f2` |
---|
| 983 | fi |
---|
| 984 | fi |
---|
| 985 | done < tmp_mrun |
---|
| 986 | |
---|
| 987 | line="" |
---|
| 988 | found=false |
---|
| 989 | grep " $localhost" $config_file | grep "%source_path" > tmp_mrun |
---|
| 990 | while read line |
---|
| 991 | do |
---|
| 992 | if [[ "$line" != "" && $(echo $line | cut -c1) != "#" ]] |
---|
| 993 | then |
---|
| 994 | if [[ $found = true ]] |
---|
| 995 | then |
---|
| 996 | printf "\n\n +++ more than one source path found in configuration file" |
---|
| 997 | printf "\n for local host \"$localhost\" " |
---|
| 998 | locat=source_path; exit |
---|
| 999 | fi |
---|
| 1000 | source_path=`echo $line | cut -d" " -f2` |
---|
| 1001 | found=true |
---|
| 1002 | fi |
---|
| 1003 | done < tmp_mrun |
---|
| 1004 | rm tmp_mrun |
---|
| 1005 | |
---|
| 1006 | if [[ "$source_path" = "" ]] |
---|
| 1007 | then |
---|
| 1008 | if [[ "$global_source_path" != "" ]] |
---|
| 1009 | then |
---|
| 1010 | source_path=$global_source_path |
---|
| 1011 | else |
---|
| 1012 | printf "\n\n +++ no source path found in configuration file" |
---|
| 1013 | printf "\n for local host \"$localhost\" " |
---|
| 1014 | locat=source_path; exit |
---|
| 1015 | fi |
---|
| 1016 | fi |
---|
| 1017 | eval source_path=$source_path |
---|
| 1018 | |
---|
| 1019 | if [[ ! -d $source_path ]] |
---|
| 1020 | then |
---|
| 1021 | printf "\n\n +++ source path \"$source_path\" on local host" |
---|
| 1022 | printf "\n \"$localhost\" does not exist" |
---|
| 1023 | locat=source_path; exit |
---|
| 1024 | fi |
---|
| 1025 | |
---|
| 1026 | fi |
---|
| 1027 | |
---|
| 1028 | |
---|
[1289] | 1029 | # GET THE GLOBAL REVISION-NUMBER OF THE SVN-REPOSITORY |
---|
| 1030 | # (HANDED OVER TO RESTART-RUNS USING OPTION -G) |
---|
| 1031 | if [[ "$global_revision" = "" && $host != "ibmkisti" ]] |
---|
[78] | 1032 | then |
---|
| 1033 | global_revision=`svnversion $source_path 2>/dev/null` |
---|
| 1034 | global_revision="Rev: $global_revision" |
---|
| 1035 | fi |
---|
[1] | 1036 | |
---|
[78] | 1037 | |
---|
[1289] | 1038 | # ??? das darf doch eigentlich garnicht passieren, weil optionen config-datei uebersteuern ??? |
---|
| 1039 | # CHECK AGAIN, IF EXECUTION SHALL BE DONE ON A REMOTE-HOST (BATCH-JOB). |
---|
| 1040 | # (HOST MAY HAVE CHANGED DUE TO SETTINGS IN THE CONFIGURATION-FILE) |
---|
[1] | 1041 | if [[ -n $host && "$host" != $localhost ]] |
---|
| 1042 | then |
---|
| 1043 | do_batch=true |
---|
| 1044 | do_remote=true |
---|
| 1045 | case $host in |
---|
[1620] | 1046 | (ibm|ibmh|ibmkisti|ibmku|ibms|lcbullhh|lccrayb|lccrayh|lccrayf|lcflow|lckyoto|nech|unics|lcxe6|lcxt5m|lck|lckiaps|lckordi|lckyuh|lckyut|lcsb) true;; |
---|
[1] | 1047 | (*) printf "\n" |
---|
[83] | 1048 | printf "\n +++ sorry: execution of batch jobs on remote host \"$host\"" |
---|
| 1049 | printf "\n is not available" |
---|
[1] | 1050 | locat=nqs; (( iec = 0 )); exit;; |
---|
| 1051 | esac |
---|
| 1052 | else |
---|
| 1053 | host=$localhost |
---|
| 1054 | fi |
---|
| 1055 | |
---|
| 1056 | |
---|
[1620] | 1057 | |
---|
[1289] | 1058 | # IN CASE OF PARALLEL EXECUTION, CHECK SOME SPECIFICATIONS CONCERNING PROCESSOR NUMBERS |
---|
[1] | 1059 | if [[ "$cond1" = parallel || "$cond2" = parallel ]] |
---|
| 1060 | then |
---|
| 1061 | |
---|
[1289] | 1062 | # CHECK, IF NUMBER OF CORES TO BE USED HAS BEEN GIVEN |
---|
[1] | 1063 | if [[ ! -n $numprocs ]] |
---|
| 1064 | then |
---|
| 1065 | printf "\n" |
---|
| 1066 | printf "\n +++ option \"-K parallel\" requires additional specification" |
---|
| 1067 | printf "\n of the number of processors to be used by" |
---|
| 1068 | printf "\n mrun-option \"-X\" or by environment-variable" |
---|
| 1069 | printf "\n \"numprocs\" in the configuration file" |
---|
| 1070 | locat=numprocs; (( iec = 0 )); exit |
---|
| 1071 | fi |
---|
| 1072 | |
---|
[1289] | 1073 | # CHECK, IF THE NUMBER OF CORES PER NODE HAS BEEN GIVEN UND IF IT IS AN |
---|
| 1074 | # INTEGRAL DIVISOR OF THE TOTAL NUMBER OF CORES GIVEN BY OPTION -X |
---|
[1039] | 1075 | if [[ "$tasks_per_node" = "" && $host != lcflow && $host != lcxt5m ]] |
---|
[1] | 1076 | then |
---|
[759] | 1077 | printf "\n" |
---|
| 1078 | printf "\n +++ option \"-T\" (tasks per node) is missing" |
---|
| 1079 | printf "\n set -T option or define tasks_per_node in the config file" |
---|
| 1080 | locat=tasks_per_node; (( iec = 0 )); exit |
---|
| 1081 | fi |
---|
[1279] | 1082 | |
---|
| 1083 | if (( numprocs < tasks_per_node )) |
---|
| 1084 | then |
---|
| 1085 | printf "\n" |
---|
| 1086 | printf "\n +++ tasks per node (-T) cannot exceed total number of processors (-X)" |
---|
| 1087 | printf "\n given values: -T $tasks_per_node -X $numprocs" |
---|
| 1088 | locat=tasks_per_node; (( iec = 0 )); exit |
---|
| 1089 | fi |
---|
| 1090 | |
---|
[1039] | 1091 | if [[ $host != lcflow && $host != lcxt5m ]] |
---|
[759] | 1092 | then |
---|
[1279] | 1093 | (( nodes = numprocs / ( tasks_per_node * threads_per_task ) )) |
---|
[848] | 1094 | (( ival = $tasks_per_node )) |
---|
| 1095 | (( pes = numprocs )) |
---|
| 1096 | (( ii = pes / ival )) |
---|
[1279] | 1097 | (( remaining_pes = pes - ii * ival )) |
---|
| 1098 | if (( remaining_pes > 0 )) |
---|
[848] | 1099 | then |
---|
| 1100 | printf "\n" |
---|
[1279] | 1101 | printf "\n +++ WARNING: tasks per node (option \"-T\") is not an integral" |
---|
[848] | 1102 | printf "\n divisor of the total number of processors (option \"-X\")" |
---|
| 1103 | printf "\n values of this mrun-call: \"-T $tasks_per_node\" \"-X $numprocs\"" |
---|
[1279] | 1104 | printf "\n One of the nodes is filled with $remaining_pes instead of $tasks_per_node tasks" |
---|
| 1105 | (( nodes = nodes + 1 )) |
---|
[848] | 1106 | fi |
---|
[1] | 1107 | fi |
---|
| 1108 | |
---|
[759] | 1109 | |
---|
[377] | 1110 | # SETTINGS FOR SUBJOB-COMMAND |
---|
[1] | 1111 | OOPT="-O $threads_per_task" |
---|
| 1112 | |
---|
[1289] | 1113 | # SET THE TOTAL NUMBER OF NODES, REQUIRED FOR THE SUBJOB-COMMAND (SEE FURTHER BELOW) |
---|
[1] | 1114 | if [[ "$tasks_per_node" != "" ]] |
---|
| 1115 | then |
---|
[1119] | 1116 | TOPT="-T $tasks_per_node" |
---|
[1] | 1117 | fi |
---|
| 1118 | |
---|
[1289] | 1119 | # CHECK node_usage FOR ALLOWED VALUES AND SET DEFAULT VALUE, IF NECESSARY |
---|
[1] | 1120 | if [[ $node_usage = default ]] |
---|
| 1121 | then |
---|
| 1122 | if [[ $host = ibms ]] |
---|
| 1123 | then |
---|
| 1124 | node_usage=shared |
---|
| 1125 | else |
---|
| 1126 | node_usage=not_shared |
---|
| 1127 | fi |
---|
| 1128 | fi |
---|
[181] | 1129 | |
---|
[635] | 1130 | if [[ $node_usage != shared && $node_usage != not_shared && $node_usage != singlejob && "$(echo $node_usage | cut -c1-3)" != "sla" ]] |
---|
[1] | 1131 | then |
---|
| 1132 | printf "\n" |
---|
| 1133 | printf "\n +++ node usage (option \"-n\") is only allowed to be set" |
---|
| 1134 | printf "\n \"shared\" or \"not_shared\"" |
---|
| 1135 | locat=tasks_per_node; (( iec = 0 )); exit |
---|
| 1136 | fi |
---|
| 1137 | |
---|
| 1138 | fi |
---|
| 1139 | |
---|
[1289] | 1140 | # CHECK IF HOSTFILE EXISTS |
---|
[1] | 1141 | if [[ -n $hostfile ]] |
---|
| 1142 | then |
---|
| 1143 | if [[ ! -f $hostfile ]] |
---|
| 1144 | then |
---|
| 1145 | printf "\n" |
---|
| 1146 | printf "\n +++ hostfile \"$hostfile\" does not exist" |
---|
| 1147 | locat=hostfile; exit |
---|
| 1148 | fi |
---|
| 1149 | fi |
---|
| 1150 | |
---|
| 1151 | |
---|
[1289] | 1152 | # SET DEFAULT VALUE FOR THE MAXIMUM NUMBER OF PARALLEL IO STREAMS |
---|
[759] | 1153 | if [[ "$maximum_parallel_io_streams" = "" ]] |
---|
| 1154 | then |
---|
| 1155 | maximum_parallel_io_streams=$numprocs |
---|
| 1156 | fi |
---|
[1] | 1157 | |
---|
[759] | 1158 | |
---|
[1289] | 1159 | # SET PORT NUMBER OPTION FOR CALLS OF SSH/SCP, subjob AND batch_scp SCRIPTS |
---|
[1094] | 1160 | if [[ "$scp_port" != "" ]] |
---|
| 1161 | then |
---|
| 1162 | PORTOPT="-P $scp_port" |
---|
| 1163 | SSH_PORTOPT="-p $scp_port" |
---|
| 1164 | fi |
---|
| 1165 | |
---|
| 1166 | |
---|
[1289] | 1167 | # SET DEFAULT-QUEUE, IF NOT GIVEN |
---|
[1] | 1168 | if [[ $queue = none ]] |
---|
| 1169 | then |
---|
| 1170 | case $host in |
---|
[1184] | 1171 | (ibmh) queue=cluster;; |
---|
[693] | 1172 | (ibmkisti) queue=class.32plus;; |
---|
[1620] | 1173 | (lcbullhh) queue=compute;; |
---|
[1442] | 1174 | (lccrayb) queue=mpp1q;; |
---|
| 1175 | (lccrayh) queue=mpp1q;; |
---|
[1040] | 1176 | (lckiaps) queue=normal;; |
---|
[440] | 1177 | (lckyoto) queue=ph;; |
---|
[1099] | 1178 | (lckyuh) queue=fx-single;; |
---|
| 1179 | (lckyut) queue=cx-single;; |
---|
[635] | 1180 | (lctit) queue=S;; |
---|
[1] | 1181 | (nech) queue=none;; |
---|
| 1182 | (unics) queue=unics;; |
---|
| 1183 | esac |
---|
| 1184 | fi |
---|
| 1185 | |
---|
| 1186 | |
---|
[1289] | 1187 | # GENERATE FULL FILENAMES OF INPUT-FILES, INCLUDING THEIR PATH |
---|
| 1188 | # CHECK, IF INPUT-FILES EXIST, AND DETERMINE HIGHEST CYCLE NUMBER (IF CYCLES EXIST) |
---|
[1] | 1189 | (( i = 0 )) |
---|
| 1190 | while (( i < iin )) |
---|
| 1191 | do |
---|
| 1192 | (( i = i + 1 )) |
---|
| 1193 | (( maxcycle = 0 )) |
---|
| 1194 | |
---|
[1289] | 1195 | # GENERATE PATH AND FULL FILE NAME (then-BRANCH: FIXED FULL NAME IS GIVEN, I.E. THE |
---|
| 1196 | # FILE IDENTIFIER IS NOT PART OF THE FILENAME)) |
---|
[1] | 1197 | if [[ "${actionin[$i]}" = di ]] |
---|
| 1198 | then |
---|
[1289] | 1199 | remotepathin[$i]=${pathin[$i]}/${endin[$i]} # EVALUATE REMOTE-PATH ON THE REMOTE |
---|
| 1200 | # HOST ONLY |
---|
[1] | 1201 | eval filename=${pathin[$i]}/${endin[$i]} |
---|
| 1202 | else |
---|
[1289] | 1203 | remotepathin[$i]=${pathin[$i]}/${afname}${endin[$i]} # EVALUATE REMOTE-PATH ON THE REMOTE |
---|
| 1204 | # HOST ONLY |
---|
[1] | 1205 | eval filename=${pathin[$i]}/${afname}${endin[$i]} |
---|
| 1206 | fi |
---|
[892] | 1207 | eval pathname=${pathin[$i]} |
---|
[1] | 1208 | |
---|
[1289] | 1209 | # SAVE INPUT FILE NAME FOR PARAMETER FILE CHECK |
---|
[920] | 1210 | if [[ ("${transin[$i]}" = job) && (-f $filename) ]] |
---|
| 1211 | then |
---|
| 1212 | filename_input=$filename |
---|
| 1213 | fi |
---|
[892] | 1214 | |
---|
[1289] | 1215 | # CHECK IF FILE EXISTS |
---|
[1] | 1216 | if [[ $(ls $filename* 2>&1 | grep -c "not found") = 1 || \ |
---|
[746] | 1217 | $(ls $filename* 2>&1 | grep -c "nicht gefunden") = 1 || \ |
---|
[1] | 1218 | $(ls $filename* 2>&1 | grep -c "No such file") = 1 || \ |
---|
| 1219 | $(ls $filename* 2>&1 | grep -c "does not exist") = 1 ]] |
---|
| 1220 | then |
---|
[1289] | 1221 | |
---|
| 1222 | # FILES WITH ATTRIBUTE locopt ARE OPTIONAL. NO ABORT, IF THEY DO NOT EXIST. |
---|
[1] | 1223 | if [[ "${transin[$i]}" != "locopt" ]] |
---|
| 1224 | then |
---|
| 1225 | printf "\n\n +++ INPUT-file: " |
---|
| 1226 | if [[ "${extin[$i]}" = "" ]] |
---|
| 1227 | then |
---|
| 1228 | printf "\n $filename" |
---|
| 1229 | else |
---|
| 1230 | printf "\n $filename.${extin[$i]}" |
---|
| 1231 | fi |
---|
| 1232 | printf "\n does not exist\n" |
---|
| 1233 | locat=input; exit |
---|
| 1234 | else |
---|
| 1235 | transin[$i]="unavailable" |
---|
| 1236 | fi |
---|
| 1237 | else |
---|
| 1238 | |
---|
[1289] | 1239 | # DETERMINE THE FILE'S CYCLE NUMBER |
---|
[1] | 1240 | ls -1 -d $filename > filelist 2>/dev/null |
---|
| 1241 | ls -1 -d $filename.* >> filelist 2>/dev/null |
---|
| 1242 | while read zeile |
---|
| 1243 | do |
---|
[1101] | 1244 | cycle=$(echo $zeile | cut -f2 -d".") |
---|
[1] | 1245 | if [[ "$cycle" = "$zeile" ]] |
---|
| 1246 | then |
---|
| 1247 | (( icycle = 0 )) |
---|
| 1248 | elif [[ "$cycle" = "${extin[$i]}" ]] |
---|
| 1249 | then |
---|
| 1250 | (( icycle = 0 )) |
---|
| 1251 | else |
---|
| 1252 | (( icycle = $cycle )) |
---|
| 1253 | fi |
---|
| 1254 | if (( icycle > maxcycle )) |
---|
| 1255 | then |
---|
| 1256 | (( maxcycle = icycle )) |
---|
| 1257 | file_to_be_used=$zeile |
---|
| 1258 | fi |
---|
| 1259 | done <filelist |
---|
| 1260 | rm filelist |
---|
| 1261 | |
---|
[1289] | 1262 | # APPEND CYCLE NUMBER TO FILENAME |
---|
[1] | 1263 | if (( maxcycle > 0 )) |
---|
| 1264 | then |
---|
| 1265 | if [[ "${extin[$i]}" != " " && "${extin[$i]}" != "" ]] |
---|
| 1266 | then |
---|
| 1267 | filename=${filename}.$maxcycle.${extin[$i]} |
---|
| 1268 | else |
---|
| 1269 | filename=${filename}.$maxcycle |
---|
| 1270 | fi |
---|
| 1271 | else |
---|
| 1272 | if [[ "${extin[$i]}" != " " && "${extin[$i]}" != "" ]] |
---|
| 1273 | then |
---|
| 1274 | filename=${filename}.${extin[$i]} |
---|
| 1275 | fi |
---|
| 1276 | fi |
---|
| 1277 | |
---|
[1289] | 1278 | # STORE FILENAME WITHOUT PATH BUT WITH CYCLE NUMBER, |
---|
| 1279 | # BECAUSE IT MIGHT BE REQUIRED LATER TO RESTORE THE FILE FROM AN ARCHIVE-SYSTEM |
---|
[1] | 1280 | absnamein[$i]=$filename |
---|
| 1281 | if (( maxcycle > 0 )) |
---|
| 1282 | then |
---|
| 1283 | if [[ "${actionin[$i]}" = di ]] |
---|
| 1284 | then |
---|
| 1285 | frelin[$i]=${endin[$i]}.$maxcycle |
---|
| 1286 | else |
---|
| 1287 | frelin[$i]=${afname}${endin[$i]}.$maxcycle |
---|
| 1288 | fi |
---|
| 1289 | else |
---|
| 1290 | if [[ "${actionin[$i]}" = di ]] |
---|
| 1291 | then |
---|
| 1292 | frelin[$i]=${endin[$i]} |
---|
| 1293 | else |
---|
| 1294 | frelin[$i]=${afname}${endin[$i]} |
---|
| 1295 | fi |
---|
| 1296 | fi |
---|
| 1297 | |
---|
| 1298 | fi |
---|
| 1299 | done |
---|
[920] | 1300 | |
---|
[1289] | 1301 | # GENERATE FULL FILENAMES OF OUTPUT-FILES (WITHOUT $ OR ~), |
---|
| 1302 | # CHECK, IF OUTPUT-FILES EXIST, AND DETERMINE HIGHEST CYCLE NUMBER (IF CYCLES EXIST), |
---|
| 1303 | # OR, IN CASE THAT FILE DOES NOT EXIST, CHECK, IF IT CAN BE CREATED |
---|
| 1304 | # THESE ACTIONS ARE NOT CARRIED OUT, IF FILES SHALL BE TRASFERRED FROM THE REMOTE TO |
---|
| 1305 | # THE LOCAL HOST (BECAUSE THEIR IS NO DIRECT ACCESS TO THE LOCAL DIRECTORIES FROM THE |
---|
| 1306 | # REMOTE HOST) |
---|
[1] | 1307 | (( i = 0 )) |
---|
| 1308 | while (( i < iout )) |
---|
| 1309 | do |
---|
| 1310 | (( i = i + 1 )) |
---|
| 1311 | if [[ ! ( $fromhost != $localhost && ( "${actionout[$i]}" = tr || "${actionout[$i]}" = tra || "${actionout[$i]}" = trpe ) ) ]] |
---|
| 1312 | then |
---|
[260] | 1313 | if [[ "${actionout[$i]}" = tr ]] |
---|
[1] | 1314 | then |
---|
| 1315 | actionout[$i]="" |
---|
[260] | 1316 | elif [[ "${actionout[$i]}" = trpe ]] |
---|
| 1317 | then |
---|
| 1318 | actionout[$i]=pe |
---|
[1] | 1319 | elif [[ "${actionout[$i]}" = tra ]] |
---|
| 1320 | then |
---|
| 1321 | actionout[$i]=a |
---|
| 1322 | fi |
---|
| 1323 | (( maxcycle = 0 )) |
---|
| 1324 | eval filename=${pathout[$i]}/${fname}${endout[$i]} |
---|
| 1325 | eval catalogname=${pathout[$i]} |
---|
| 1326 | if [[ $(ls $filename* 2>&1 | grep -c "not found") = 1 || \ |
---|
[746] | 1327 | $(ls $filename* 2>&1 | grep -c "nicht gefunden") = 1 || \ |
---|
[1] | 1328 | $(ls $filename* 2>&1 | grep -c "No such file") = 1 || \ |
---|
| 1329 | $(ls $filename* 2>&1 | grep -c "does not exist") = 1 ]] |
---|
| 1330 | then |
---|
| 1331 | |
---|
[1289] | 1332 | # IF OUTPUT-FILE DOES NOT EXIST CHECK, IF IT CAN BE CREATED |
---|
[1] | 1333 | if cat /dev/null > $filename |
---|
| 1334 | then |
---|
| 1335 | rm $filename |
---|
| 1336 | else |
---|
| 1337 | |
---|
[1289] | 1338 | # CHECK, IF THE DIRECTORY WHERE FILE SHALL BE COPIED TO EXISTS |
---|
| 1339 | # IF IT DOES NOT EXIST, TRY TO CREATE IT |
---|
[1] | 1340 | if [[ ! -d $catalogname ]] |
---|
| 1341 | then |
---|
| 1342 | if mkdir -p $catalogname |
---|
| 1343 | then |
---|
| 1344 | printf "\n\n *** directory:" |
---|
| 1345 | printf "\n $catalogname" |
---|
| 1346 | printf "\n was created\n" |
---|
| 1347 | else |
---|
| 1348 | printf "\n\n +++ OUTPUT-file:" |
---|
| 1349 | printf "\n $filename" |
---|
| 1350 | printf "\n cannot be created, because directory does not exist" |
---|
| 1351 | printf "\n and cannot be created either" |
---|
| 1352 | printf "\n" |
---|
| 1353 | locat=output ; exit |
---|
| 1354 | fi 2>/dev/null |
---|
| 1355 | else |
---|
| 1356 | printf "\n\n +++ OUTPUT-file:" |
---|
| 1357 | printf "\n $filename" |
---|
| 1358 | printf "\n cannot be created, although directory exists" |
---|
| 1359 | printf "\n" |
---|
| 1360 | locat=output ; exit |
---|
| 1361 | fi |
---|
| 1362 | fi 2>/dev/null |
---|
| 1363 | else |
---|
| 1364 | |
---|
[1289] | 1365 | # DETERMINE THE CYCLE NUMBER |
---|
[1] | 1366 | ls -1 -d $filename > filelist 2>/dev/null |
---|
| 1367 | ls -1 -d $filename.* >> filelist 2>/dev/null |
---|
| 1368 | while read zeile |
---|
| 1369 | do |
---|
[1101] | 1370 | cycle=$(echo $zeile | cut -f2 -d".") |
---|
[1] | 1371 | if [[ "$cycle" = "$zeile" || "$cycle" = ${extout[$i]} ]] |
---|
| 1372 | then |
---|
| 1373 | (( icycle = 1 )) |
---|
| 1374 | else |
---|
| 1375 | (( icycle = $cycle + 1 )) |
---|
| 1376 | fi |
---|
| 1377 | if (( icycle > maxcycle )) |
---|
| 1378 | then |
---|
| 1379 | (( maxcycle = icycle )) |
---|
| 1380 | fi |
---|
| 1381 | done <filelist |
---|
| 1382 | rm filelist |
---|
| 1383 | fi |
---|
| 1384 | |
---|
[1289] | 1385 | # APPEND CYCLE NUMBER TO FILENAME AND CHECK, IF FILE CAN BE CREATED |
---|
| 1386 | # IN CASE OF FILE-APPEND, FILE MUST BE APPENDED TO THE ONE WITH HIGHEST CYCLE NUMBER |
---|
[1] | 1387 | if [[ "${actionout[$i]}" != a ]] |
---|
| 1388 | then |
---|
| 1389 | if (( maxcycle > 0 )) |
---|
| 1390 | then |
---|
| 1391 | filename=${filename}.$maxcycle |
---|
| 1392 | if cat /dev/null > $filename |
---|
| 1393 | then |
---|
| 1394 | rm $filename |
---|
| 1395 | else |
---|
| 1396 | printf "\n +++ OUTPUT-file:" |
---|
| 1397 | printf "\n $filename" |
---|
| 1398 | printf "\n cannot be created" |
---|
| 1399 | locat=output ; exit |
---|
| 1400 | fi |
---|
| 1401 | fi |
---|
| 1402 | else |
---|
| 1403 | (( maxcycle = maxcycle - 1 )) |
---|
| 1404 | if (( maxcycle > 0 )) |
---|
| 1405 | then |
---|
| 1406 | filename=${filename}.$maxcycle |
---|
| 1407 | fi |
---|
| 1408 | fi |
---|
| 1409 | |
---|
[1289] | 1410 | # STORE FILENAME WITHOUT PATH BUT WITH CYCLE NUMBER, |
---|
| 1411 | # BECAUSE IT MIGHT BE REQUIRED LATER TO STORE THE FILE ON AN ARCHIVE-SYSTEM |
---|
| 1412 | # OR TO PUT THIS FILENAME ON FILE OUTPUT_FILE_CONNECTIONS |
---|
[1] | 1413 | pathout[$i]=$filename |
---|
| 1414 | if (( maxcycle > 0 )) |
---|
| 1415 | then |
---|
| 1416 | frelout[$i]=${fname}${endout[$i]}.$maxcycle |
---|
| 1417 | else |
---|
| 1418 | frelout[$i]=${fname}${endout[$i]} |
---|
| 1419 | fi |
---|
| 1420 | |
---|
| 1421 | fi |
---|
| 1422 | done |
---|
| 1423 | |
---|
| 1424 | |
---|
[1289] | 1425 | # THE DVR-PACKAGE REQUIRES ITS OWN LIBRARY |
---|
[83] | 1426 | if [[ $(echo $package_list | grep -c dvrp_graphics) != 0 ]] |
---|
| 1427 | then |
---|
[260] | 1428 | if [[ "$dvr_inc" = "" ]] |
---|
[83] | 1429 | then |
---|
[260] | 1430 | printf "\n\n +++ no value for \"dvr_inc\" given in configuration file" |
---|
[83] | 1431 | printf "\n This is required for the dvrp_graphics package.\n" |
---|
[260] | 1432 | locat=dvr; exit |
---|
[83] | 1433 | fi |
---|
[260] | 1434 | if [[ "$dvr_lib" = "" ]] |
---|
[83] | 1435 | then |
---|
[260] | 1436 | printf "\n\n +++ no value for \"dvr_lib\" given in configuration file" |
---|
[83] | 1437 | printf "\n This is required for the dvrp_graphics package.\n" |
---|
[260] | 1438 | locat=dvr; exit |
---|
[83] | 1439 | fi |
---|
| 1440 | fi |
---|
| 1441 | |
---|
| 1442 | |
---|
[1289] | 1443 | # CHECK, WETHER A MAIN PROGRAM OR AN EXECUTABLE HAVE BEEN DECLARED. |
---|
| 1444 | # EXECUTABLES DO NOT NEED TO BE COMPILED. |
---|
[1] | 1445 | if [[ "$mainprog" = "" && "$executable" = "" ]] |
---|
| 1446 | then |
---|
| 1447 | printf "\n +++ neither main program nor executable defined" |
---|
| 1448 | locat=source; exit |
---|
| 1449 | elif [[ "$mainprog" != "" && "$executable" != "" ]] |
---|
| 1450 | then |
---|
| 1451 | printf "\n +++ main program as well as executable defined" |
---|
| 1452 | locat=source; exit |
---|
| 1453 | elif [[ "$mainprog" = "" && "$executable" != "" ]] |
---|
| 1454 | then |
---|
| 1455 | do_compile=false |
---|
| 1456 | fi |
---|
| 1457 | |
---|
| 1458 | |
---|
[1289] | 1459 | # CREATE SOURCE-DIRECTORY TO COLLECT ROUTINES TO BE COMPILED. |
---|
| 1460 | # LATER THE MRUN-SCRIPT AND CONFIGURATION-FILE WILL ALSO BE COPIED TO THIS DIRECTORY. |
---|
[193] | 1461 | if [[ $restart_run != true && "$SOURCES_COMPLETE" = "" ]] |
---|
[191] | 1462 | then |
---|
| 1463 | rm -rf SOURCES_FOR_RUN_$fname |
---|
| 1464 | mkdir SOURCES_FOR_RUN_$fname |
---|
| 1465 | fi |
---|
| 1466 | |
---|
| 1467 | |
---|
[1289] | 1468 | # COLLECT ALL ROUTINES TO BE COMPILED |
---|
| 1469 | # THIS IS NOT REQUIRED WITHIN BATCH-JOBS, BECAUSE ROUTINES HAVE ALREADY BEEN COLLECTED |
---|
| 1470 | # BY THE MRUN-CALL WHICH CREATED THE BATCH-JOB. |
---|
[1] | 1471 | if [[ $do_compile = true && "$SOURCES_COMPLETE" = "" ]] |
---|
| 1472 | then |
---|
| 1473 | |
---|
[5] | 1474 | [[ "$source_list" = LM ]] && source_list=LOCALLY_MODIFIED |
---|
[1] | 1475 | |
---|
[838] | 1476 | if [[ "$source_list" = LOCALLY_MODIFIED ]] |
---|
[1] | 1477 | then |
---|
| 1478 | |
---|
[1289] | 1479 | # DETERMINE MODIFIED FILES OF THE SVN WORKING COPY |
---|
[5] | 1480 | source_list="" |
---|
| 1481 | cd $source_path |
---|
| 1482 | |
---|
| 1483 | |
---|
[1289] | 1484 | # CHECK, IF DIRECTORY IS UNDER SVN CONTROL |
---|
[5] | 1485 | if [[ ! -d .svn ]] |
---|
| 1486 | then |
---|
| 1487 | printf "\n\n +++ source directory" |
---|
| 1488 | printf "\n \"$source_path\" " |
---|
| 1489 | printf "\n is not under control of \"subversion\"." |
---|
[306] | 1490 | printf "\n Please do not use mrun-option \"-s LOCALLY_MODIFIED\"\n" |
---|
[5] | 1491 | fi |
---|
| 1492 | |
---|
| 1493 | |
---|
[1289] | 1494 | # LIST ALL MODIFIED SOURCE CODE FILES |
---|
[5] | 1495 | Filenames="" |
---|
[69] | 1496 | svn status > tmp_mrun |
---|
| 1497 | while read line |
---|
[5] | 1498 | do |
---|
| 1499 | firstc=`echo $line | cut -c1` |
---|
| 1500 | if [[ $firstc = M || $firstc = "?" ]] |
---|
| 1501 | then |
---|
| 1502 | Name=`echo "$line" | cut -c8-` |
---|
| 1503 | extension=`echo $Name | cut -d. -f2` |
---|
| 1504 | if [[ "$extension" = f90 || "$extension" = F90 || "$extension" = f || "$extension" = F || "$extension" = c ]] |
---|
| 1505 | then |
---|
| 1506 | Filenames="$Filenames "$Name |
---|
| 1507 | fi |
---|
| 1508 | fi |
---|
[69] | 1509 | done < tmp_mrun |
---|
[5] | 1510 | |
---|
| 1511 | |
---|
[1289] | 1512 | # COPY FILES TO SOURCES_FOR_RUN_... |
---|
[5] | 1513 | for dateiname in $Filenames |
---|
| 1514 | do |
---|
| 1515 | cp $dateiname $working_directory/SOURCES_FOR_RUN_$fname |
---|
| 1516 | source_list=$source_list"$dateiname " |
---|
| 1517 | done |
---|
| 1518 | |
---|
| 1519 | cd - > /dev/null |
---|
| 1520 | |
---|
| 1521 | |
---|
[1289] | 1522 | # COPY FILES GIVEN BY OPTION -s TO DIRECTORY SOURCES_FOR_RUN_... |
---|
| 1523 | # AUTOMATIC RESTART RUNS JUST ACCESS THE DIRECTORY CREATED BY THE INITIAL RUN |
---|
[5] | 1524 | elif [[ "$source_list" != "" && $restart_run != true ]] |
---|
[1] | 1525 | then |
---|
| 1526 | |
---|
| 1527 | cd $source_path |
---|
| 1528 | |
---|
| 1529 | for filename in $source_list |
---|
| 1530 | do |
---|
| 1531 | |
---|
[1289] | 1532 | # SOURCE CODE FILE IS NOT ALLOWED TO INCLUDE PATH |
---|
[1103] | 1533 | if [[ $(echo $filename | grep -c "/") != 0 ]] |
---|
[1] | 1534 | then |
---|
| 1535 | printf "\n +++ source code file: $filename" |
---|
| 1536 | printf "\n must not contain (\"/\") " |
---|
| 1537 | locat=source; exit |
---|
| 1538 | fi |
---|
| 1539 | |
---|
| 1540 | if [[ ! -f $filename ]] |
---|
| 1541 | then |
---|
| 1542 | printf "\n +++ source code file: $filename" |
---|
| 1543 | printf "\n does not exist" |
---|
| 1544 | locat=source; exit |
---|
| 1545 | else |
---|
| 1546 | cp $filename $working_directory/SOURCES_FOR_RUN_$fname |
---|
| 1547 | fi |
---|
| 1548 | |
---|
| 1549 | done |
---|
| 1550 | |
---|
| 1551 | cd - > /dev/null |
---|
| 1552 | |
---|
| 1553 | fi |
---|
| 1554 | |
---|
| 1555 | |
---|
[1289] | 1556 | # CHECK, IF A MAIN PROGRAM EXISTS AND IF IT IS PART OF THE LIST OF FILES |
---|
| 1557 | # TO BE COMPILED. IF NOT, ADD IT TO THE LIST. |
---|
[5] | 1558 | if [[ $restart_run != true ]] |
---|
[1] | 1559 | then |
---|
| 1560 | |
---|
| 1561 | if [[ ! -f "$source_path/$mainprog" ]] |
---|
| 1562 | then |
---|
| 1563 | printf "\n\n +++ main program: $mainprog" |
---|
| 1564 | printf "\n does not exist in source directory" |
---|
| 1565 | printf "\n \"$source_path\"\n" |
---|
| 1566 | locat=source; exit |
---|
| 1567 | else |
---|
| 1568 | if [[ $(echo $source_list | grep -c $mainprog) = 0 ]] |
---|
| 1569 | then |
---|
| 1570 | cp $source_path/$mainprog SOURCES_FOR_RUN_$fname |
---|
| 1571 | source_list=${mainprog}" $source_list" |
---|
| 1572 | fi |
---|
| 1573 | fi |
---|
| 1574 | fi |
---|
| 1575 | |
---|
| 1576 | |
---|
[1289] | 1577 | # CHECK, IF MAKEFILE EXISTS AND COPY IT TO THE SOURCES_FOR_RUN... DIRECTORY |
---|
| 1578 | # NOT REQUIRED FOR RESTART RUNS, SOURCES_FOR_RUN... HAS BEEN CREATED BY THE INITIAL RUN |
---|
[211] | 1579 | if [[ "$restart_run" != true ]] |
---|
| 1580 | then |
---|
| 1581 | [[ "$makefile" = "" ]] && makefile=$source_path/Makefile |
---|
| 1582 | if [[ ! -f $makefile ]] |
---|
| 1583 | then |
---|
| 1584 | printf "\n +++ file \"$makefile\" does not exist" |
---|
| 1585 | locat=make; exit |
---|
| 1586 | else |
---|
| 1587 | cp $makefile SOURCES_FOR_RUN_$fname/Makefile |
---|
| 1588 | fi |
---|
| 1589 | fi |
---|
| 1590 | |
---|
| 1591 | |
---|
[1289] | 1592 | # COPY FILES FROM OPTIONAL SOURCE PATH GIVEN IN THE CONFIGURATION FILE |
---|
[5] | 1593 | if [[ $restart_run != true && "$add_source_path" != "" ]] |
---|
[1] | 1594 | then |
---|
| 1595 | |
---|
[1289] | 1596 | # DOES THE DIRECTORY EXIST? |
---|
[1] | 1597 | if [[ ! -d $add_source_path ]] |
---|
| 1598 | then |
---|
| 1599 | printf "\n\n +++ WARNING: additional source code directory" |
---|
| 1600 | printf "\n \"$add_source_path\" " |
---|
| 1601 | printf "\n does not exist or is not a directory." |
---|
| 1602 | printf "\n No source code will be used from this directory!\n" |
---|
| 1603 | add_source_path="" |
---|
[931] | 1604 | if [[ $silent == false ]] |
---|
| 1605 | then |
---|
| 1606 | sleep 3 |
---|
| 1607 | fi |
---|
[1] | 1608 | else |
---|
| 1609 | |
---|
| 1610 | cd $add_source_path |
---|
[211] | 1611 | found=false |
---|
[1] | 1612 | |
---|
| 1613 | Names=$(ls -1 *.f90 2>&1) |
---|
| 1614 | [[ $(echo $Names | grep -c '*.f90') = 0 ]] && AddFilenames="$Names" |
---|
| 1615 | Names=$(ls -1 *.F90 2>&1) |
---|
| 1616 | [[ $(echo $Names | grep -c '*.F90') = 0 ]] && AddFilenames="$AddFilenames $Names" |
---|
| 1617 | Names=$(ls -1 *.F 2>&1) |
---|
| 1618 | [[ $(echo $Names | grep -c '*.F') = 0 ]] && AddFilenames="$AddFilenames $Names" |
---|
| 1619 | Names=$(ls -1 *.f 2>&1) |
---|
| 1620 | [[ $(echo $Names | grep -c '*.f') = 0 ]] && AddFilenames="$AddFilenames $Names" |
---|
| 1621 | Names=$(ls -1 *.c 2>&1) |
---|
| 1622 | [[ $(echo $Names | grep -c '*.c') = 0 ]] && AddFilenames="$AddFilenames $Names" |
---|
| 1623 | |
---|
| 1624 | cd - > /dev/null |
---|
| 1625 | cd SOURCES_FOR_RUN_$fname |
---|
| 1626 | |
---|
[211] | 1627 | # COPY MAKEFILE IF EXISTING |
---|
| 1628 | if [[ -f $add_source_path/Makefile ]] |
---|
| 1629 | then |
---|
| 1630 | printf "\n\n *** user Makefile from directory" |
---|
| 1631 | printf "\n \"$add_source_path\" is used \n" |
---|
[931] | 1632 | if [[ $silent == false ]] |
---|
| 1633 | then |
---|
| 1634 | sleep 1 |
---|
| 1635 | fi |
---|
[211] | 1636 | cp $add_source_path/Makefile . |
---|
| 1637 | fi |
---|
| 1638 | |
---|
[1] | 1639 | for dateiname in $AddFilenames |
---|
| 1640 | do |
---|
| 1641 | if [[ -f $dateiname ]] |
---|
| 1642 | then |
---|
| 1643 | printf "\n +++ source code file \"$dateiname\" found in additional" |
---|
| 1644 | printf "\n source code directory \"$add_source_path\" " |
---|
[211] | 1645 | printf "\n but was also given with option \"-s\" which means that it should be taken" |
---|
| 1646 | printf "\n from directory \"$source_path\"." |
---|
[1] | 1647 | locat=source; exit |
---|
| 1648 | fi |
---|
| 1649 | |
---|
| 1650 | cp $add_source_path/$dateiname . |
---|
| 1651 | source_list="$source_list $dateiname" |
---|
| 1652 | |
---|
[211] | 1653 | # CHECK IF FILE IS CONTAINED IN MAKEFILE |
---|
| 1654 | if [[ $(grep -c $dateiname Makefile) = 0 ]] |
---|
| 1655 | then |
---|
| 1656 | printf "\n\n +++ user file \"$dateiname\" " |
---|
| 1657 | printf "\n is not listed in Makefile \n" |
---|
| 1658 | locat=source; exit |
---|
| 1659 | else |
---|
| 1660 | |
---|
| 1661 | if [[ $found = false ]] |
---|
[1] | 1662 | then |
---|
[211] | 1663 | found=true |
---|
| 1664 | printf "\n\n *** following user file(s) added to the" |
---|
| 1665 | printf " files to be translated:\n " |
---|
[1] | 1666 | fi |
---|
[211] | 1667 | printf "$dateiname " |
---|
[931] | 1668 | if [[ $silent == false ]] |
---|
| 1669 | then |
---|
| 1670 | sleep 0.5 |
---|
| 1671 | fi |
---|
| 1672 | |
---|
[1] | 1673 | fi |
---|
| 1674 | done |
---|
[211] | 1675 | [[ $found = true ]] && printf "\n" |
---|
[1] | 1676 | cd - > /dev/null |
---|
| 1677 | fi |
---|
| 1678 | fi |
---|
| 1679 | |
---|
| 1680 | |
---|
[1289] | 1681 | # ADD ALL ROUTINES BELONGING TO SOFTWARE PACKAGES (GIVEN BY OPTION -p) |
---|
| 1682 | # TO THE LIST OF FILES TO BE COMPILED |
---|
[5] | 1683 | if [[ $restart_run != true && -n $package_list ]] |
---|
[1] | 1684 | then |
---|
| 1685 | |
---|
| 1686 | cd $source_path |
---|
| 1687 | |
---|
| 1688 | for package in $package_list |
---|
| 1689 | do |
---|
| 1690 | |
---|
| 1691 | [[ $package = "dvrp_graphics+1PE" ]] && package=dvrp_graphics |
---|
| 1692 | |
---|
[1289] | 1693 | # DETERMINE FILES BELONGING TO THE PACKAGE |
---|
| 1694 | # ERROR MESSAGE ARE REDIRECTED TO /dev/null, BECAUSE WILDCARD (*) ALSO GIVES |
---|
| 1695 | # THE NAME OF THE DIRECTORY |
---|
[1] | 1696 | package_source_list=`grep "defined( __$package " * 2>/dev/null | cut -f1 -d:` |
---|
| 1697 | |
---|
| 1698 | |
---|
[1289] | 1699 | # ADD THESE FILES TO THE LIST OF FILES TO BE COMPILED, |
---|
| 1700 | # IF THEY ARE NOT ALREADY PART OF THE LIST |
---|
[1] | 1701 | for source_list_name in $package_source_list |
---|
| 1702 | do |
---|
| 1703 | if [[ $(echo $source_list | grep -c $source_list_name) = 0 ]] |
---|
| 1704 | then |
---|
| 1705 | |
---|
[1289] | 1706 | # ONLY TAKE FILES WITH VALID SUFFIX |
---|
[1] | 1707 | ending=`echo $source_list_name | cut -f2 -d.` |
---|
| 1708 | if [[ "$ending" = f90 || "$ending" = F90 || "$ending" = f || "$ending" = F || "$ending" = c ]] |
---|
| 1709 | then |
---|
| 1710 | cp $source_list_name $working_directory/SOURCES_FOR_RUN_$fname |
---|
| 1711 | source_list="$source_list $source_list_name" |
---|
| 1712 | fi |
---|
| 1713 | fi |
---|
| 1714 | done |
---|
| 1715 | done |
---|
| 1716 | |
---|
| 1717 | cd - > /dev/null |
---|
| 1718 | fi |
---|
| 1719 | |
---|
| 1720 | fi # do_compile=true |
---|
| 1721 | |
---|
| 1722 | |
---|
[1289] | 1723 | # IF SOURCE CODE IS TO BE COMPILED, DO SOME MORE CHECKS |
---|
| 1724 | # AND SET PRE-PROCESSOR DIRECTIVES |
---|
[1199] | 1725 | if [[ $do_compile = true || $create_executable_for_batch = true ]] |
---|
[1] | 1726 | then |
---|
| 1727 | |
---|
[1289] | 1728 | # SET PREPROCESSOR-DIRECTIVES TO SELECT OPERATING SYSTEM SPECIFIC CODE |
---|
[1] | 1729 | if [[ $(echo $localhost | cut -c1-3) = ibm ]] |
---|
| 1730 | then |
---|
[82] | 1731 | cpp_options="${cpp_options},-D__ibm=__ibm" |
---|
[1] | 1732 | elif [[ $(echo $localhost | cut -c1-3) = nec ]] |
---|
| 1733 | then |
---|
[82] | 1734 | cpp_options="$cpp_options -D__nec" |
---|
| 1735 | elif [[ $(echo $localhost | cut -c1-2) = lc ]] |
---|
| 1736 | then |
---|
| 1737 | cpp_options="$cpp_options -D__lc" |
---|
[1] | 1738 | else |
---|
[82] | 1739 | cpp_options="$cpp_options -D__$localhost" |
---|
[1] | 1740 | fi |
---|
[83] | 1741 | |
---|
[1289] | 1742 | # SET DIRECTIVES GIVEN BY OPTION -K (E.G. parallel) |
---|
[1] | 1743 | if [[ $(echo $localhost | cut -c1-3) = ibm ]] |
---|
| 1744 | then |
---|
[82] | 1745 | [[ -n $cond1 ]] && cpp_options="${cpp_options},-D__$cond1=__$cond1" |
---|
| 1746 | [[ -n $cond2 ]] && cpp_options="${cpp_options},-D__$cond2=__$cond2" |
---|
[1] | 1747 | else |
---|
[82] | 1748 | [[ -n $cond1 ]] && cpp_options="$cpp_options -D__$cond1" |
---|
| 1749 | [[ -n $cond2 ]] && cpp_options="$cpp_options -D__$cond2" |
---|
[1] | 1750 | fi |
---|
[83] | 1751 | |
---|
[1289] | 1752 | # SET DIRECTIVES FOR ACTIVATING SOFTWARE-PACKAGES (OPTION -p) |
---|
[1] | 1753 | if [[ -n $package_list ]] |
---|
| 1754 | then |
---|
| 1755 | for package in $package_list |
---|
| 1756 | do |
---|
| 1757 | if [[ $(echo $localhost | cut -c1-3) = ibm ]] |
---|
| 1758 | then |
---|
| 1759 | if [[ $package != "dvrp_graphics+1PE" ]] |
---|
| 1760 | then |
---|
[82] | 1761 | cpp_options="${cpp_options},-D__$package=__$package" |
---|
[1] | 1762 | else |
---|
[82] | 1763 | cpp_options="${cpp_options},-D__dvrp_graphics=__dvrp_graphics" |
---|
[1] | 1764 | export use_seperate_pe_for_dvrp_output=true |
---|
| 1765 | fi |
---|
| 1766 | else |
---|
| 1767 | if [[ $package != "dvrp_graphics+1PE" ]] |
---|
| 1768 | then |
---|
[82] | 1769 | cpp_options="$cpp_options -D__$package" |
---|
[1] | 1770 | else |
---|
[82] | 1771 | cpp_options="$cpp_options -D__dvrp_graphics" |
---|
[1] | 1772 | export use_seperate_pe_for_dvrp_output=true |
---|
| 1773 | fi |
---|
| 1774 | fi |
---|
| 1775 | done |
---|
| 1776 | fi |
---|
[83] | 1777 | |
---|
[1289] | 1778 | # SET DIRECTIVES GIVEN BY OPTION -D |
---|
[1] | 1779 | if [[ -n $cpp_opts ]] |
---|
| 1780 | then |
---|
| 1781 | for popts in $cpp_opts |
---|
| 1782 | do |
---|
[82] | 1783 | if [[ $(echo $localhost | cut -c1-3) = ibm ]] |
---|
| 1784 | then |
---|
| 1785 | cpp_options="${cpp_options},-D__$popts=__$popts" |
---|
| 1786 | else |
---|
| 1787 | cpp_options="$cpp_options -D__$popts" |
---|
| 1788 | fi |
---|
[1] | 1789 | done |
---|
| 1790 | fi |
---|
| 1791 | |
---|
| 1792 | else |
---|
| 1793 | |
---|
| 1794 | |
---|
[1289] | 1795 | # FOR LOCAL RUNS CHECK AGAIN, IF EXECUTABLE EXISTS |
---|
[1] | 1796 | if [[ $do_remote = false ]] |
---|
| 1797 | then |
---|
| 1798 | if [[ ! -f $executable ]] |
---|
| 1799 | then |
---|
| 1800 | printf "\n +++ executable file: $executable" |
---|
| 1801 | printf "\n does not exist" |
---|
| 1802 | locat=executable; exit |
---|
| 1803 | fi |
---|
| 1804 | fi |
---|
| 1805 | fi |
---|
| 1806 | |
---|
| 1807 | |
---|
[1289] | 1808 | # DETERMINE THE JOB MODE |
---|
[1491] | 1809 | if [[ $(echo $localhost | cut -c1-3) = ibm ]] |
---|
[1] | 1810 | then |
---|
[1491] | 1811 | if [[ "$LOADLBATCH" = yes ]] |
---|
| 1812 | then |
---|
| 1813 | batch_job=.TRUE. |
---|
| 1814 | jobmo=BATCH |
---|
| 1815 | else |
---|
| 1816 | batch_job=.FALSE. |
---|
| 1817 | jobmo=INTERACTIVE |
---|
| 1818 | fi |
---|
[1] | 1819 | else |
---|
[1491] | 1820 | if [[ "$ENVIRONMENT" = BATCH ]] |
---|
| 1821 | then |
---|
| 1822 | batch_job=.TRUE. |
---|
| 1823 | jobmo=BATCH |
---|
| 1824 | else |
---|
| 1825 | batch_job=.FALSE. |
---|
| 1826 | jobmo=INTERACTIVE |
---|
| 1827 | fi |
---|
[1] | 1828 | fi |
---|
| 1829 | |
---|
[1289] | 1830 | # NO INTERACTIVE RUNS ALLOWED ON LCTIT |
---|
[635] | 1831 | if [[ $host = lctit && $jobmo = INTERACTIVE && $do_batch = false ]] |
---|
| 1832 | then |
---|
| 1833 | printf "\n +++ no interactive runs allowed on host \"$host\" " |
---|
| 1834 | printf "\n please submit batch job using mrun option \"-b\" \n" |
---|
| 1835 | locat=normal; exit |
---|
| 1836 | fi |
---|
| 1837 | |
---|
| 1838 | |
---|
[1289] | 1839 | # CHECK, IF USER DEFINED A COMPILER |
---|
[1] | 1840 | if [[ "$compiler_name" = "" ]] |
---|
| 1841 | then |
---|
[82] | 1842 | |
---|
[220] | 1843 | printf "\n +++ no compiler specified for \"$host $cond1 $cond2\"" |
---|
[82] | 1844 | locat=compiler_name; exit |
---|
| 1845 | |
---|
[1] | 1846 | fi |
---|
| 1847 | |
---|
| 1848 | |
---|
[1289] | 1849 | # DETERMINE THE NAME OF MRUN'S TEMPORARY WORKING DIRECTORY |
---|
[1609] | 1850 | |
---|
| 1851 | # ON HLRN-III, USE THE QUEUING NAME. OTHERWISE USE USERNAME AND RANDOM NUMBER |
---|
| 1852 | if [[ $do_batch = false && $(echo $host | cut -c1-6) = lccray ]] |
---|
| 1853 | then |
---|
| 1854 | kennung=$(checkjob $PBS_JOBID | grep Reservation | cut -d" " -s -f2 | cut -d"." -s -f2 | sed "s/['\"]//g") |
---|
[1755] | 1855 | if [[ "$kennung" = "" ]] |
---|
[1753] | 1856 | then |
---|
| 1857 | kennung=$RANDOM |
---|
| 1858 | fi |
---|
[1609] | 1859 | else |
---|
| 1860 | kennung=$RANDOM |
---|
| 1861 | fi |
---|
| 1862 | |
---|
[1] | 1863 | if [[ "$tmp_user_catalog" = "" ]] |
---|
| 1864 | then |
---|
[312] | 1865 | if [[ $localhost = ibmh ]] |
---|
[1] | 1866 | then |
---|
[312] | 1867 | tmp_user_catalog=$SCRATCH |
---|
[1] | 1868 | elif [[ $localhost = nech ]] |
---|
| 1869 | then |
---|
| 1870 | tmp_user_catalog=$WRKSHR |
---|
| 1871 | else |
---|
| 1872 | tmp_user_catalog=/tmp |
---|
| 1873 | fi |
---|
| 1874 | fi |
---|
| 1875 | TEMPDIR=$tmp_user_catalog/${usern}.$kennung |
---|
| 1876 | |
---|
| 1877 | |
---|
[1289] | 1878 | # DETERMINE THE NAME OF THE DIRECTORY WHICH IS USED TO TEMPORARILY STORE DATA FOR RESTART RUNS |
---|
[1] | 1879 | if [[ "$tmp_data_catalog" = "" ]] |
---|
| 1880 | then |
---|
[622] | 1881 | if [[ $localhost = nech ]] |
---|
[1] | 1882 | then |
---|
| 1883 | tmp_data_catalog=$WRKSHR/mrun_restart_data |
---|
| 1884 | else |
---|
| 1885 | tmp_data_catalog=/tmp/mrun_restart_data |
---|
| 1886 | fi |
---|
| 1887 | fi |
---|
| 1888 | |
---|
| 1889 | |
---|
[1289] | 1890 | # IN CASE OF LOCAL RUNS REPLACE ENVIRONMENT VARIABLES BY THEIR VALUES |
---|
[1199] | 1891 | if [[ $do_remote = false && $do_compile = true || $create_executable_for_batch = true ]] |
---|
[1] | 1892 | then |
---|
| 1893 | eval fopts=\"$fopts\" |
---|
| 1894 | eval lopts=\"$lopts\" |
---|
| 1895 | fi |
---|
| 1896 | |
---|
| 1897 | |
---|
| 1898 | |
---|
[1289] | 1899 | # DETERMINE COMPILE- AND LINK-OPTIONS |
---|
[1210] | 1900 | fopts="$fopts $netcdf_inc $fftw_inc $dvr_inc" |
---|
| 1901 | lopts="$lopts $netcdf_lib $fftw_lib $dvr_lib" |
---|
[1] | 1902 | ROPTS="$ropts" |
---|
[1289] | 1903 | # if [[ ( $(echo $host | cut -c1-3) = nec || $(echo $host | cut -c1-3) = ibm || $host = lckyoto || $host = lctit || $host = lcflow || $host = lcxe6 || $host = lcxt5m || $host = lck || $host = lckiaps || $host = lckordi || $host = lcsb || $host ) && -n $numprocs ]] |
---|
[1090] | 1904 | # then |
---|
[1] | 1905 | XOPT="-X $numprocs" |
---|
[1090] | 1906 | # fi |
---|
[1] | 1907 | |
---|
| 1908 | |
---|
| 1909 | |
---|
[1289] | 1910 | # CHECK THE CPU-LIMIT. IT MUST BE GIVEN FOR BATCH-JOBS AND IS COMMUNICATED TO THE |
---|
| 1911 | # EXECUTABLE VIA NAMELIST-PARAMETER cputime |
---|
[1] | 1912 | done=false |
---|
| 1913 | while [[ $done = false ]] |
---|
| 1914 | do |
---|
| 1915 | cputime=$cpumax |
---|
| 1916 | if (( $cputime == 0 )) |
---|
| 1917 | then |
---|
[246] | 1918 | if [[ $do_batch = true ]] |
---|
| 1919 | then |
---|
| 1920 | printf "\n +++ cpu-time is undefined" |
---|
| 1921 | printf "\n >>> Please type CPU-time in seconds as INTEGER:" |
---|
| 1922 | printf "\n >>> " |
---|
| 1923 | read cputime 1>/dev/null 2>&1 |
---|
| 1924 | else |
---|
| 1925 | cputime=10000000 # NO CPU LIMIT FOR INTERACTIVE RUNS |
---|
| 1926 | fi |
---|
[1] | 1927 | else |
---|
| 1928 | done=true |
---|
| 1929 | fi |
---|
| 1930 | cpumax=$cputime |
---|
| 1931 | done |
---|
| 1932 | |
---|
| 1933 | (( minuten = cputime / 60 )) |
---|
| 1934 | (( sekunden = cputime - minuten * 60 )) |
---|
| 1935 | |
---|
| 1936 | |
---|
[1289] | 1937 | # CHECK THE MEMORY DEMAND |
---|
[1] | 1938 | if [[ $do_batch = true ]] |
---|
| 1939 | then |
---|
| 1940 | done=false |
---|
| 1941 | while [[ $done = false ]] |
---|
| 1942 | do |
---|
| 1943 | if (( memory == 0 )) |
---|
| 1944 | then |
---|
| 1945 | printf "\n +++ memory demand is undefined" |
---|
| 1946 | printf "\n >>> Please type memory in MByte per process as INTEGER:" |
---|
| 1947 | printf "\n >>> " |
---|
| 1948 | read memory 1>/dev/null 2>&1 |
---|
| 1949 | else |
---|
| 1950 | done=true |
---|
| 1951 | fi |
---|
| 1952 | done |
---|
| 1953 | fi |
---|
| 1954 | |
---|
| 1955 | |
---|
[1289] | 1956 | # IN CASE OF REMOTE-JOBS CHECK, IF A USERNAME FOR THE REMOTE HOST IS GIVEN |
---|
[1] | 1957 | if [[ $do_remote = true && -z $remote_username ]] |
---|
| 1958 | then |
---|
| 1959 | while [[ -z $remote_username ]] |
---|
| 1960 | do |
---|
| 1961 | printf "\n +++ username on remote host \"$host\" is undefined" |
---|
| 1962 | printf "\n >>> Please type username:" |
---|
| 1963 | printf "\n >>> " |
---|
| 1964 | read remote_username |
---|
| 1965 | done |
---|
| 1966 | mc="$mc -u$remote_username" |
---|
| 1967 | fi |
---|
| 1968 | |
---|
[892] | 1969 | # CHECK FOR INITIAL COMMANDS AFTER LOGIN |
---|
| 1970 | if [[ "$login_init_cmd" != "" ]] |
---|
| 1971 | then |
---|
| 1972 | export init_cmds="${login_init_cmd};" |
---|
| 1973 | fi |
---|
[221] | 1974 | |
---|
[892] | 1975 | |
---|
[1289] | 1976 | # SET THE MODULE-LOAD COMMAD AND EXPORT IT FOR subjob |
---|
[493] | 1977 | if [[ "$modules" != "" ]] |
---|
| 1978 | then |
---|
[678] | 1979 | if [[ $host = lctit ]] |
---|
| 1980 | then |
---|
| 1981 | export module_calls=". $modules" |
---|
| 1982 | else |
---|
| 1983 | export module_calls="module load ${modules};" |
---|
| 1984 | fi |
---|
[493] | 1985 | fi |
---|
| 1986 | |
---|
[1289] | 1987 | # OUTPUT OF THE MRUN-HEADER |
---|
[1] | 1988 | calltime=$(date) |
---|
| 1989 | printf "\n" |
---|
[1101] | 1990 | printf "#------------------------------------------------------------------------# \n" |
---|
| 1991 | printf "| %-35s%35s | \n" "$version" "$calltime" |
---|
| 1992 | printf "| | \n" |
---|
[1] | 1993 | spalte1="called on:"; spalte2=$localhost_realname |
---|
[1101] | 1994 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[1289] | 1995 | if [[ $do_remote = true ]] |
---|
[1] | 1996 | then |
---|
[1289] | 1997 | spalte1="execution on:"; spalte2="$host (username: $remote_username)" |
---|
[1] | 1998 | else |
---|
[1289] | 1999 | spalte1="execution on:"; spalte2="$host ($localhost_realname)" |
---|
[1] | 2000 | fi |
---|
[1101] | 2001 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[1] | 2002 | if [[ -n $numprocs ]] |
---|
| 2003 | then |
---|
[1106] | 2004 | if [[ $run_coupled_model = false ]] |
---|
| 2005 | then |
---|
| 2006 | spalte1="number of PEs:"; spalte2=$numprocs |
---|
| 2007 | else |
---|
| 2008 | spalte1="number of PEs:"; spalte2="$numprocs (atmosphere: $numprocs_atmos, ocean: $numprocs_ocean)" |
---|
| 2009 | fi |
---|
[1101] | 2010 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[1] | 2011 | fi |
---|
| 2012 | if [[ -n $tasks_per_node ]] |
---|
| 2013 | then |
---|
| 2014 | spalte1="tasks per node:"; spalte2="$tasks_per_node (number of nodes: $nodes)" |
---|
[1101] | 2015 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[1279] | 2016 | if (( remaining_pes > 0 )) |
---|
| 2017 | then |
---|
| 2018 | spalte1=" "; spalte2="one of the nodes only filled with $remaining_pes tasks" |
---|
| 2019 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
| 2020 | fi |
---|
[1] | 2021 | fi |
---|
[759] | 2022 | if [[ $maximum_parallel_io_streams != $numprocs ]] |
---|
| 2023 | then |
---|
| 2024 | spalte1="max par io streams:"; spalte2="$maximum_parallel_io_streams" |
---|
[1101] | 2025 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[759] | 2026 | fi |
---|
[622] | 2027 | if [[ $use_openmp = true ]] |
---|
[1] | 2028 | then |
---|
| 2029 | spalte1="threads per task:"; spalte2="$threads_per_task" |
---|
[1101] | 2030 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[1] | 2031 | fi |
---|
[1101] | 2032 | printf "| | \n" |
---|
[1] | 2033 | if [[ $do_compile = true ]] |
---|
| 2034 | then |
---|
[475] | 2035 | if [[ "$mopts" != "" ]] |
---|
| 2036 | then |
---|
[1101] | 2037 | spalte1="make options:"; spalte2=$(echo "$mopts" | cut -c-45) |
---|
| 2038 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
| 2039 | zeile=$(echo "$mopts" | cut -c46-) |
---|
[475] | 2040 | while [[ "$zeile" != "" ]] |
---|
| 2041 | do |
---|
| 2042 | spalte1="" |
---|
[1101] | 2043 | spalte2=$(echo "$zeile" | cut -c-45) |
---|
| 2044 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
| 2045 | zeile=$(echo "$zeile" | cut -c46-) |
---|
[475] | 2046 | done |
---|
| 2047 | fi |
---|
| 2048 | |
---|
[1101] | 2049 | spalte1="cpp directives:"; spalte2=$(echo "$cpp_options" | cut -c-45) |
---|
| 2050 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
| 2051 | zeile=$(echo "$cpp_options" | cut -c46-) |
---|
[1] | 2052 | while [[ "$zeile" != "" ]] |
---|
| 2053 | do |
---|
| 2054 | spalte1="" |
---|
[1101] | 2055 | spalte2=$(echo "$zeile" | cut -c-45) |
---|
| 2056 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
| 2057 | zeile=$(echo "$zeile" | cut -c46-) |
---|
[1] | 2058 | done |
---|
| 2059 | |
---|
[1101] | 2060 | spalte1="compiler options:"; spalte2=$(echo "$fopts" | cut -c-45) |
---|
| 2061 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
| 2062 | zeile=$(echo "$fopts" | cut -c46-) |
---|
[1] | 2063 | while [[ "$zeile" != "" ]] |
---|
| 2064 | do |
---|
| 2065 | spalte1="" |
---|
[1101] | 2066 | spalte2=$(echo "$zeile" | cut -c-45) |
---|
| 2067 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
| 2068 | zeile=$(echo "$zeile" | cut -c46-) |
---|
[1] | 2069 | done |
---|
| 2070 | |
---|
[1101] | 2071 | spalte1="linker options:"; spalte2=$(echo "$lopts" | cut -c-45) |
---|
| 2072 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
| 2073 | zeile=$(echo "$lopts" | cut -c46-) |
---|
[1] | 2074 | while [[ "$zeile" != "" ]] |
---|
| 2075 | do |
---|
| 2076 | spalte1="" |
---|
[1101] | 2077 | spalte2=$(echo "$zeile" | cut -c-45) |
---|
| 2078 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
| 2079 | zeile=$(echo "$zeile" | cut -c46-) |
---|
[1] | 2080 | done |
---|
| 2081 | |
---|
[1101] | 2082 | spalte1="modules to be load:"; spalte2=$(echo "$modules" | cut -c-45) |
---|
| 2083 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
| 2084 | zeile=$(echo "$modules" | cut -c46-) |
---|
[493] | 2085 | while [[ "$zeile" != "" ]] |
---|
| 2086 | do |
---|
| 2087 | spalte1="" |
---|
[1101] | 2088 | spalte2=$(echo "$zeile" | cut -c-45) |
---|
| 2089 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
| 2090 | zeile=$(echo "$zeile" | cut -c46-) |
---|
[493] | 2091 | done |
---|
| 2092 | |
---|
[1101] | 2093 | spalte1="main program:"; spalte2=$mainprog |
---|
| 2094 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[1] | 2095 | else |
---|
| 2096 | spalte1=executable:; spalte2=$executable |
---|
[1101] | 2097 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[1] | 2098 | fi |
---|
[1101] | 2099 | printf "| | \n" |
---|
| 2100 | spalte1="base name of files:"; spalte2=$fname |
---|
| 2101 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[1] | 2102 | if [[ $fname != $afname ]] |
---|
| 2103 | then |
---|
[1101] | 2104 | spalte1="base name of input files:"; spalte2=$afname |
---|
| 2105 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[1] | 2106 | fi |
---|
[1101] | 2107 | spalte1="INPUT control list:"; spalte2=$(echo $input_list) |
---|
| 2108 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
| 2109 | spalte1="OUTPUT control list:"; spalte2=$(echo $output_list) |
---|
| 2110 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[1] | 2111 | |
---|
[1106] | 2112 | if [[ "$ocean_file_appendix" = true ]] |
---|
| 2113 | then |
---|
| 2114 | printf "| %-35s%-35s | \n" "suffix \"_O\" is added to local files" " " |
---|
| 2115 | fi |
---|
| 2116 | |
---|
[1] | 2117 | if [[ $do_batch = true || "$LOADLBATCH" = yes ]] |
---|
| 2118 | then |
---|
| 2119 | spalte1="memory demand / PE":; spalte2="$memory MB" |
---|
[1101] | 2120 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[1] | 2121 | spalte1=CPU-time:; spalte2="$minuten:$sekunden" |
---|
[1101] | 2122 | printf "| %-25s%-45s | \n" "$spalte1" "$spalte2" |
---|
[1] | 2123 | fi |
---|
| 2124 | |
---|
| 2125 | if [[ $do_compile = true ]] |
---|
| 2126 | then |
---|
[1101] | 2127 | printf "| | \n" |
---|
| 2128 | printf "| Files to be compiled: | \n" |
---|
[1] | 2129 | zeile=$source_list |
---|
| 2130 | while [[ "$zeile" != "" ]] |
---|
| 2131 | do |
---|
[1101] | 2132 | linestart=$(echo $zeile | cut -c-70) |
---|
| 2133 | printf "| %-70s | \n" "$linestart" |
---|
| 2134 | zeile=$(echo "$zeile" | cut -c71-) |
---|
[1] | 2135 | done |
---|
| 2136 | fi |
---|
[1101] | 2137 | printf "#------------------------------------------------------------------------#" |
---|
[1] | 2138 | |
---|
| 2139 | |
---|
| 2140 | |
---|
[1289] | 2141 | # OUTPUT OF FILE CONNECTIONS IN CASE OF TRACEBACK |
---|
[1] | 2142 | if [[ $do_trace = true ]] |
---|
| 2143 | then |
---|
| 2144 | (( i = 0 )) |
---|
| 2145 | while (( i < iin )) |
---|
| 2146 | do |
---|
| 2147 | (( i = i + 1 )) |
---|
| 2148 | if (( i == 1 )) |
---|
| 2149 | then |
---|
| 2150 | printf "\n\n >>> INPUT-file assignments:\n" |
---|
| 2151 | fi |
---|
| 2152 | printf "\n ${localin[$i]} : ${absnamein[$i]}" |
---|
| 2153 | done |
---|
| 2154 | (( i = 0 )) |
---|
| 2155 | while (( i < iout )) |
---|
| 2156 | do |
---|
| 2157 | (( i = i + 1 )) |
---|
| 2158 | if (( i == 1 )) |
---|
| 2159 | then |
---|
| 2160 | printf "\n\n >>> OUTPUT-file assignments:\n" |
---|
| 2161 | fi |
---|
| 2162 | printf "\n ${localout[$i]} : ${pathout[$i]}" |
---|
| 2163 | done |
---|
| 2164 | (( i = 0 )) |
---|
| 2165 | while (( i < iic )) |
---|
| 2166 | do |
---|
| 2167 | (( i = i + 1 )) |
---|
| 2168 | if (( i == 1 )) |
---|
| 2169 | then |
---|
| 2170 | printf "\n\n >>> INPUT-commands:\n" |
---|
| 2171 | fi |
---|
| 2172 | printf "\n ${in_command[$i]}" |
---|
| 2173 | done |
---|
| 2174 | (( i = 0 )) |
---|
| 2175 | while (( i < ioc )) |
---|
| 2176 | do |
---|
| 2177 | (( i = i + 1 )) |
---|
| 2178 | if (( i == 1 )) |
---|
| 2179 | then |
---|
| 2180 | printf "\n\n >>> OUTPUT-commands:\n" |
---|
| 2181 | fi |
---|
| 2182 | printf "\n ${out_command[$i]}" |
---|
| 2183 | done |
---|
| 2184 | fi |
---|
| 2185 | |
---|
[1289] | 2186 | # QUERY FOR CONTINUE (ON LOCAL MACHINES ONLY) |
---|
[1] | 2187 | if [[ $remotecall = false && $silent = false && $jobmo != BATCH ]] |
---|
| 2188 | then |
---|
| 2189 | antwort=dummy |
---|
| 2190 | printf "\n\n" |
---|
[1101] | 2191 | printf " >>> everything o.k. (y/n) ? " |
---|
| 2192 | while read antwort |
---|
[1] | 2193 | do |
---|
[1101] | 2194 | if [[ "$antwort" != y && "$antwort" != Y && "$antwort" != n && "$antwort" != N ]] |
---|
| 2195 | then |
---|
| 2196 | printf " >>> everything o.k. (y/n) ? " |
---|
| 2197 | else |
---|
| 2198 | break |
---|
| 2199 | fi |
---|
[1] | 2200 | done |
---|
| 2201 | if [[ $antwort = n || $antwort = N ]] |
---|
| 2202 | then |
---|
| 2203 | locat=user_abort; (( iec = 0 )); exit |
---|
| 2204 | fi |
---|
| 2205 | if [[ $do_batch = true ]] |
---|
| 2206 | then |
---|
| 2207 | printf " >>> batch-job will be created and submitted" |
---|
| 2208 | else |
---|
[1289] | 2209 | printf " >>> MRUN will now continue to execute on this machine" |
---|
[1] | 2210 | fi |
---|
| 2211 | fi |
---|
| 2212 | |
---|
[841] | 2213 | |
---|
[1101] | 2214 | # PERFORM PARAMETER FILE CHECK (COUPLED RUNS ARE NOT SUPPORTED YET) |
---|
| 2215 | # DEFINE VARIABLES FOR FREQUENTLY USED DIRECTORIES |
---|
[841] | 2216 | check_depository="${working_directory}/trunk/UTIL" |
---|
| 2217 | check_sources="${working_directory}/tmp_check_namelist_files" |
---|
[818] | 2218 | skip_check=false |
---|
[841] | 2219 | |
---|
[1101] | 2220 | # CHECK IF NAMELIST_FILE_CHECK HAS BEEN COMPILED SUCCESSFULLY |
---|
[920] | 2221 | if [[ ! -f $check_depository/check_namelist_files.tar ]] |
---|
[841] | 2222 | then |
---|
[1101] | 2223 | skip_check=true |
---|
| 2224 | reason="run on remote host or parameter file check has not been compiled." |
---|
[920] | 2225 | fi |
---|
| 2226 | if [[ ! -f $PALM_BIN/check_namelist_files.x ]] |
---|
| 2227 | then |
---|
[1101] | 2228 | skip_check=true |
---|
| 2229 | reason="parameter file check has not been compiled." |
---|
[892] | 2230 | fi |
---|
[841] | 2231 | |
---|
[1101] | 2232 | # CHECK FOR PARALLEL RUN; OTHERWISE SKIP CHECK |
---|
[892] | 2233 | if [[ "$cond1" != "parallel" && "$cond2" != "parallel" ]] |
---|
| 2234 | then |
---|
[1101] | 2235 | skip_check=true |
---|
| 2236 | reason="serial run." |
---|
[841] | 2237 | fi |
---|
| 2238 | |
---|
[1101] | 2239 | # ONLY PERFORM CHECK IF -z OPTION IS NOT SET, NO RESTART RUN IS CARRIED OUT |
---|
| 2240 | # AND IF THE EXECUTION HOST IS THE LOCAL HOST |
---|
| 2241 | # ATTENTION: THIS ROUTINE DOES NOT WORK IF THE COMPILER ON THE LOCAL HOST |
---|
| 2242 | # DIFFERS FROM THE COMPILER ON THE REMOTE HOST |
---|
[892] | 2243 | if [[ $check_namelist_files == false ]] |
---|
[807] | 2244 | then |
---|
[1101] | 2245 | skip_check=true |
---|
| 2246 | reason="-z option set." |
---|
[815] | 2247 | fi |
---|
[892] | 2248 | |
---|
| 2249 | if [[ $fromhost != $localhost ]] |
---|
[815] | 2250 | then |
---|
[1101] | 2251 | skip_check=true |
---|
| 2252 | reason="submitting host is local host." |
---|
[892] | 2253 | fi |
---|
[1] | 2254 | |
---|
[892] | 2255 | if [[ $run_coupled_model == true ]] |
---|
| 2256 | then |
---|
| 2257 | skip_check=true |
---|
| 2258 | reason="coupled run." |
---|
| 2259 | fi |
---|
[1] | 2260 | |
---|
[892] | 2261 | if [[ $restart_run == true ]] |
---|
| 2262 | then |
---|
| 2263 | skip_check=true |
---|
| 2264 | reason="restart run." |
---|
| 2265 | fi |
---|
[807] | 2266 | |
---|
[1101] | 2267 | # SKIP CHECK IN CASE OF RESTART RUN: CHECK WHETHER THE LAST CHAR IS "f" IN PARIN |
---|
[1122] | 2268 | (( last_char_int = `echo $filename_input | wc -c` - 1 )) |
---|
| 2269 | last_char=`echo $filename_input | cut -c $last_char_int` |
---|
[892] | 2270 | if [[ "$last_char" == "f" ]] |
---|
| 2271 | then |
---|
| 2272 | skip_check=true |
---|
[1069] | 2273 | reason="restart run." |
---|
[892] | 2274 | fi |
---|
[807] | 2275 | |
---|
[892] | 2276 | if [[ $skip_check == false ]] |
---|
| 2277 | then |
---|
| 2278 | tmp_check=${working_directory}/tmp_check |
---|
[841] | 2279 | |
---|
[1241] | 2280 | # GET TOPOGRAPHY PARAMETER FILE SUFFIX (USUALLY "_TOPO"). THIS FILE IS |
---|
[1101] | 2281 | # NOT NECESSARILY REQUIRED |
---|
[818] | 2282 | line="" |
---|
| 2283 | found=false |
---|
[892] | 2284 | grep "TOPOGRAPHY_DATA" $config_file > $tmp_check |
---|
[818] | 2285 | while read line1 |
---|
| 2286 | do |
---|
| 2287 | line="$line1" |
---|
| 2288 | if [[ "$line" != "" && $(echo $line | cut -c1) != "#" ]] |
---|
| 2289 | then |
---|
| 2290 | topo_suffix=`echo $line | tr -s " " | cut -d" " -s -f5` |
---|
| 2291 | found=true |
---|
| 2292 | fi |
---|
| 2293 | done < $tmp_check |
---|
[807] | 2294 | |
---|
[818] | 2295 | if [[ $found = false ]] |
---|
| 2296 | then |
---|
| 2297 | printf "\n +++ no TOPOGRAPHY list entry found in the configuration file." |
---|
| 2298 | fi |
---|
| 2299 | |
---|
[1241] | 2300 | # GET NUDGING PARAMETER FILE SUFFIX (USUALLY "_NUDGE"). THIS FILE IS |
---|
| 2301 | # NOT NECESSARILY REQUIRED |
---|
| 2302 | line="" |
---|
| 2303 | found=false |
---|
| 2304 | grep "NUDGING_DATA" $config_file > $tmp_check |
---|
| 2305 | while read line1 |
---|
| 2306 | do |
---|
| 2307 | line="$line1" |
---|
| 2308 | if [[ "$line" != "" && $(echo $line | cut -c1) != "#" ]] |
---|
| 2309 | then |
---|
| 2310 | nudge_suffix=`echo $line | tr -s " " | cut -d" " -s -f5` |
---|
| 2311 | found=true |
---|
| 2312 | fi |
---|
| 2313 | done < $tmp_check |
---|
| 2314 | |
---|
| 2315 | if [[ $found = false ]] |
---|
| 2316 | then |
---|
| 2317 | printf "\n +++ no NUDGING list entry found in the configuration file." |
---|
| 2318 | fi |
---|
| 2319 | |
---|
| 2320 | # GET LARGE SCALE FORCING PARAMETER FILE SUFFIX (USUALLY "_LSF"). THIS FILE IS |
---|
| 2321 | # NOT NECESSARILY REQUIRED |
---|
| 2322 | line="" |
---|
| 2323 | found=false |
---|
| 2324 | grep "LSF_DATA" $config_file > $tmp_check |
---|
| 2325 | while read line1 |
---|
| 2326 | do |
---|
| 2327 | line="$line1" |
---|
| 2328 | if [[ "$line" != "" && $(echo $line | cut -c1) != "#" ]] |
---|
| 2329 | then |
---|
| 2330 | lsf_suffix=`echo $line | tr -s " " | cut -d" " -s -f5` |
---|
| 2331 | found=true |
---|
| 2332 | fi |
---|
| 2333 | done < $tmp_check |
---|
| 2334 | |
---|
| 2335 | if [[ $found = false ]] |
---|
| 2336 | then |
---|
| 2337 | printf "\n +++ no LSF list entry found in the configuration file." |
---|
| 2338 | fi |
---|
| 2339 | |
---|
[807] | 2340 | rm -rf ${working_directory}/tmp_check |
---|
| 2341 | |
---|
[1241] | 2342 | # CHECK IF THE P3DF FILE MUST BE CHECKED |
---|
[807] | 2343 | for item in $input_list |
---|
| 2344 | do |
---|
[892] | 2345 | if [[ "$item" == "restart" ]] |
---|
[807] | 2346 | then |
---|
| 2347 | check_restart=1 |
---|
| 2348 | else |
---|
| 2349 | check_restart=0 |
---|
| 2350 | fi |
---|
| 2351 | done |
---|
| 2352 | |
---|
[1101] | 2353 | # ERROR IF NO PARAMETER FILE WAS FOUND, OTHERWISE PROCEED |
---|
| 2354 | if [[ (! -f $filename_input) && ( "$filename_input" != "" ) ]] |
---|
[807] | 2355 | then |
---|
[920] | 2356 | printf "\n\n +++ ERROR: parameter file ($filename_input) not found." |
---|
[807] | 2357 | locat=check_namelist; exit |
---|
| 2358 | fi |
---|
| 2359 | |
---|
[1101] | 2360 | # CHECK IF THE RESTART PARAMETER FILE EXISTS (IF NECESSARY) |
---|
[807] | 2361 | if [[ $check_restart == 1 ]] |
---|
| 2362 | then |
---|
[892] | 2363 | |
---|
[920] | 2364 | filenamef="${filename_input}f" |
---|
[892] | 2365 | if [[ ! -f $filenamef ]] |
---|
[807] | 2366 | then |
---|
[892] | 2367 | printf "\n\n +++ WARNING: restart parameter file ($filenamef) is missing." |
---|
[807] | 2368 | check_restart=0 |
---|
| 2369 | answer=dummy |
---|
| 2370 | printf "\n\n" |
---|
| 2371 | |
---|
[818] | 2372 | if [[ $silent == false ]] |
---|
[807] | 2373 | then |
---|
[818] | 2374 | while [[ "$answer" != c && "$answer" != C && "$answer" != a && "$answer" != A ]] |
---|
| 2375 | do |
---|
| 2376 | printf " >>> continue anyway (c(ontinue)/a(bort)) ? " |
---|
| 2377 | read answer |
---|
| 2378 | done |
---|
| 2379 | if [[ $answer = a || $answer = A ]] |
---|
| 2380 | then |
---|
| 2381 | printf "\n +++ Aborting...." |
---|
| 2382 | locat=normal; exit |
---|
| 2383 | fi |
---|
[807] | 2384 | fi |
---|
| 2385 | else |
---|
| 2386 | check_restart=1 |
---|
| 2387 | fi |
---|
| 2388 | fi |
---|
| 2389 | |
---|
[1101] | 2390 | # CREATE TEMPORARY SOURCES_FOR_CHECK PATH |
---|
[807] | 2391 | mkdir $check_sources |
---|
| 2392 | cd $check_sources |
---|
| 2393 | |
---|
[892] | 2394 | |
---|
[1101] | 2395 | # CHECK FOR USER CODE, OTHERWISE USE THE PRECOMPILED CHECK_NAMELIST_FILES.X |
---|
[892] | 2396 | if [[ -d $add_source_path ]] |
---|
[807] | 2397 | then |
---|
| 2398 | printf "\n\n *** copying files from $check_depository" |
---|
| 2399 | cp $check_depository/check_namelist_files.tar ./ |
---|
| 2400 | |
---|
| 2401 | printf "\n\n *** untar of makefile and source files in $check_sources" |
---|
| 2402 | tar -xf check_namelist_files.tar > /dev/null 2>&1 |
---|
| 2403 | |
---|
| 2404 | printf "\n\n *** adding user code." |
---|
[892] | 2405 | cp $add_source_path/* ./ |
---|
[807] | 2406 | touch check_namelist_files.f90 |
---|
| 2407 | |
---|
[1101] | 2408 | # GET COMPILER OPTIONS AND PERFORM MAKE |
---|
[807] | 2409 | printf "\n\n *** compiling code if necessary...\n" |
---|
[1045] | 2410 | |
---|
[1101] | 2411 | # workaround for batch jobs on local machine (lcxe6) |
---|
[1045] | 2412 | if [[ $do_batch == true && $do_remote == false ]] |
---|
| 2413 | then |
---|
| 2414 | eval $init_cmds |
---|
| 2415 | fi |
---|
| 2416 | |
---|
[1101] | 2417 | # GET CHECK OPTIONS |
---|
[1083] | 2418 | line="" |
---|
| 2419 | found=false |
---|
| 2420 | |
---|
| 2421 | grep "$localhost" ${base_directory}/${config_file} | grep "%cpp_options" > $tmp_check |
---|
| 2422 | |
---|
| 2423 | while read line1 |
---|
| 2424 | do |
---|
| 2425 | |
---|
| 2426 | if [[ $(echo $line1 | cut -d" " -s -f3-) = "$localhost" ]] |
---|
| 2427 | then |
---|
| 2428 | line="$line1" |
---|
| 2429 | fi |
---|
| 2430 | |
---|
| 2431 | if [[ "$line" != "" && $(echo $line | cut -c1) != "#" ]] |
---|
| 2432 | then |
---|
[1289] | 2433 | # REMOVE COLONS FROM OPTION-STRING, REMOVE ALL -D OPTIONS |
---|
[1101] | 2434 | line="$line " |
---|
[1083] | 2435 | copts_check=`echo $line | cut -d" " -s -f2 | sed 's/::/%DUM%/g' | sed 's/:/ /g' | sed 's/%DUM%/:/g' | sed 's/-D[^ ]* //g' | sed 's/ -D.*//g'` |
---|
| 2436 | found=true |
---|
| 2437 | fi |
---|
| 2438 | |
---|
| 2439 | done < $tmp_check |
---|
| 2440 | copts_check="$copts_check -D__check -D__parallel" |
---|
| 2441 | |
---|
[807] | 2442 | make -f Makefile_check F90=$compiler_name_ser COPT="$copts_check" |
---|
| 2443 | |
---|
[1101] | 2444 | # GET MAKE OUTPUT |
---|
[807] | 2445 | if [[ $? != 0 ]] |
---|
| 2446 | then |
---|
[815] | 2447 | printf "\n +++ error during make." |
---|
| 2448 | answer=dummy |
---|
| 2449 | printf "\n\n" |
---|
[818] | 2450 | if [[ $silent == false ]] |
---|
[815] | 2451 | then |
---|
[818] | 2452 | while [[ "$answer" != c && "$answer" != C && "$answer" != a && "$answer" != A ]] |
---|
| 2453 | do |
---|
| 2454 | printf " >>> continue anyway (c(ontinue)/a(bort)) ? " |
---|
| 2455 | read answer |
---|
| 2456 | done |
---|
| 2457 | if [[ $answer = a || $answer = A ]] |
---|
| 2458 | then |
---|
| 2459 | printf "\n +++ Aborting..." |
---|
| 2460 | rm -rf $check_sources |
---|
| 2461 | locat=normal; exit |
---|
| 2462 | else |
---|
| 2463 | skip_check=true |
---|
| 2464 | fi |
---|
[815] | 2465 | else |
---|
| 2466 | skip_check=true |
---|
| 2467 | fi |
---|
[807] | 2468 | fi |
---|
| 2469 | else |
---|
| 2470 | cp $PALM_BIN/check_namelist_files.x ./ |
---|
| 2471 | fi |
---|
| 2472 | |
---|
[920] | 2473 | cp $filename_input ./PARIN |
---|
[807] | 2474 | if [[ $check_restart == 1 ]] |
---|
| 2475 | then |
---|
[892] | 2476 | cp $filenamef ./PARINF |
---|
[807] | 2477 | fi |
---|
| 2478 | |
---|
[892] | 2479 | if [[ -f ${pathname}/${fname}${topo_suffix} && $skip_check == false ]] |
---|
[818] | 2480 | then |
---|
| 2481 | printf "\n *** adding topography data" |
---|
[892] | 2482 | cp ${pathname}/${fname}${topo_suffix} ./TOPOGRAPHY_DATA |
---|
[818] | 2483 | |
---|
[1101] | 2484 | # IN CASE OF TOPOGRAPHY AND HIGH GRID POINT NUMBERS, THE STACK SIZE |
---|
| 2485 | # MUST BE INCREASED. THIS IS DUE TO THE ARRAY nzb_local AND topo_height, |
---|
| 2486 | # WHICH REQUIRE SUFFICIENT MEMORY |
---|
[818] | 2487 | ulimit -s unlimited |
---|
| 2488 | fi |
---|
| 2489 | |
---|
[1241] | 2490 | if [[ -f ${pathname}/${fname}${nudge_suffix} && $skip_check == false ]] |
---|
| 2491 | then |
---|
| 2492 | printf "\n *** adding nudging data" |
---|
| 2493 | cp ${pathname}/${fname}${nudge_suffix} ./NUDGING_DATA |
---|
| 2494 | fi |
---|
| 2495 | |
---|
| 2496 | if [[ -f ${pathname}/${fname}${lsf_suffix} && $skip_check == false ]] |
---|
| 2497 | then |
---|
| 2498 | printf "\n *** adding large scale forcing data" |
---|
| 2499 | cp ${pathname}/${fname}${lsf_suffix} ./LSF_DATA |
---|
| 2500 | fi |
---|
| 2501 | |
---|
[1101] | 2502 | # CREATE ENVPAR FILE, WHICH IS NEEDED BY CHECK_NAMELIST_FILES.X |
---|
[807] | 2503 | cat > ENVPAR << %%END%% |
---|
[823] | 2504 | &envpar run_identifier = '$fname', host = '$host', |
---|
[807] | 2505 | write_binary = '$write_binary', tasks_per_node = $tasks_per_node, |
---|
| 2506 | maximum_parallel_io_streams = $maximum_parallel_io_streams, |
---|
| 2507 | maximum_cpu_time_allowed = ${cpumax}., |
---|
| 2508 | revision = '$global_revision', |
---|
| 2509 | local_dvrserver_running = $local_dvrserver_running / |
---|
| 2510 | |
---|
| 2511 | %%END%% |
---|
| 2512 | |
---|
[1101] | 2513 | # SAFETY CHECK: ONLY PROCEED IF THE PARAMETER CHECK PROGRAM WAS PROPERLY COMPILED |
---|
| 2514 | if [[ ! -f check_namelist_files.x && $skip_check == false ]] |
---|
| 2515 | then |
---|
[807] | 2516 | printf "\n +++ WARNING: check_namelist_files.x not found." |
---|
| 2517 | answer=dummy |
---|
| 2518 | printf "\n\n" |
---|
[818] | 2519 | |
---|
| 2520 | if [[ $silent == false ]] |
---|
[807] | 2521 | then |
---|
[818] | 2522 | while [[ "$answer" != c && "$answer" != C && "$answer" != a && "$answer" != A ]] |
---|
| 2523 | do |
---|
| 2524 | printf " >>> continue anyway (c(ontinue)/a(bort)) ? " |
---|
| 2525 | read answer |
---|
| 2526 | done |
---|
| 2527 | if [[ $answer = a || $answer = A ]] |
---|
| 2528 | then |
---|
| 2529 | printf "\n +++ Aborting..." |
---|
| 2530 | rm -rf $check_sources |
---|
| 2531 | locat=normal; exit |
---|
| 2532 | else |
---|
[1101] | 2533 | printf "\n *** skipping parameter file check." |
---|
[818] | 2534 | fi |
---|
| 2535 | fi |
---|
| 2536 | |
---|
[815] | 2537 | elif [[ $skip_check == false ]] |
---|
| 2538 | then |
---|
[1101] | 2539 | # STARTING THE PARAMETER FILE CHECK |
---|
[807] | 2540 | printf "\n\n *** starting parameter file check..." |
---|
| 2541 | |
---|
[1101] | 2542 | # CHECKING THE P3D FILE |
---|
[920] | 2543 | printf "\n\n (1) checking $filename_input" |
---|
[931] | 2544 | echo "$numprocs 0 0" > VARIN |
---|
[807] | 2545 | errors=`./check_namelist_files.x < VARIN 2>&1` |
---|
| 2546 | |
---|
| 2547 | check_error=false |
---|
| 2548 | if [[ "$errors" == "" ]] |
---|
| 2549 | then |
---|
| 2550 | printf " --> o.k." |
---|
| 2551 | else |
---|
| 2552 | printf " --> failed." |
---|
| 2553 | check_error=true |
---|
| 2554 | printf "\n\n $errors" |
---|
| 2555 | fi |
---|
| 2556 | |
---|
[1101] | 2557 | # CHECKING THE PD3F FILE IF NECESSARY |
---|
[807] | 2558 | if [[ $check_restart == 1 && $check_error == false ]] |
---|
| 2559 | then |
---|
[892] | 2560 | printf "\n\n (2) checking $filenamef" |
---|
[807] | 2561 | |
---|
[1101] | 2562 | # FIRST CHECK IF INITIALIZING_ACTIONS="READ_RESTART_DATA" IS SET |
---|
| 2563 | # IN &INIPAR LIST |
---|
[807] | 2564 | found=false |
---|
[1101] | 2565 | cat PARINF | while read line |
---|
[807] | 2566 | do |
---|
| 2567 | line=$(echo $line|sed 's/ //g') |
---|
| 2568 | if [[ $line == *"&inipar"* ]] |
---|
| 2569 | then |
---|
| 2570 | start_search=true |
---|
| 2571 | fi |
---|
| 2572 | |
---|
| 2573 | if [[ $start_search == true ]] |
---|
| 2574 | then |
---|
| 2575 | if [[ $line == *"initializing_actions='read_restart_data'"* ]] |
---|
| 2576 | then |
---|
| 2577 | found=true |
---|
| 2578 | break |
---|
| 2579 | fi |
---|
| 2580 | fi |
---|
| 2581 | |
---|
| 2582 | if [[ $line == *"/"* ]] |
---|
| 2583 | then |
---|
| 2584 | start_search=false |
---|
| 2585 | fi |
---|
| 2586 | |
---|
| 2587 | done |
---|
| 2588 | |
---|
| 2589 | if [[ $found = false ]] |
---|
| 2590 | then |
---|
| 2591 | printf "\n\n +++ ERROR: initializing_actions = 'read_restart_data' not found" |
---|
| 2592 | printf "\n in &inipar list in $fname$p3df_suffix." |
---|
| 2593 | rm -rf $check_sources |
---|
| 2594 | locat=check_namelist; exit |
---|
| 2595 | fi |
---|
| 2596 | |
---|
[1101] | 2597 | # READ max_user_pr FROM FILES |
---|
[931] | 2598 | if [[ -f parin_for_check ]] |
---|
| 2599 | then |
---|
| 2600 | read max_pr_user < parin_for_check |
---|
| 2601 | else |
---|
| 2602 | max_user_pr=0 |
---|
| 2603 | fi |
---|
| 2604 | |
---|
| 2605 | echo "$numprocs 1 $max_pr_user" > VARIN |
---|
[807] | 2606 | errors=`./check_namelist_files.x < VARIN 2>&1` |
---|
| 2607 | |
---|
| 2608 | if [[ "$errors" == "" ]] |
---|
| 2609 | then |
---|
| 2610 | printf " --> o.k." |
---|
| 2611 | else |
---|
| 2612 | printf " --> failed." |
---|
| 2613 | check_error=true |
---|
| 2614 | printf "\n\n $errors" |
---|
| 2615 | fi |
---|
| 2616 | fi |
---|
| 2617 | |
---|
[1101] | 2618 | # REPORT ERRORS AND CONTINUE/EXIT |
---|
[807] | 2619 | if [[ $check_error == true ]] |
---|
| 2620 | then |
---|
| 2621 | printf "\n +++ errors found in the parameter file!\n" |
---|
| 2622 | answer=dummy |
---|
| 2623 | printf "\n\n" |
---|
| 2624 | |
---|
| 2625 | while [[ "$answer" != c && "$answer" != C && "$answer" != a && "$answer" != A ]] |
---|
| 2626 | do |
---|
[1101] | 2627 | printf " >>> continue anyway (c(ontinue)/a(bort)) ? " |
---|
| 2628 | read answer |
---|
[807] | 2629 | done |
---|
| 2630 | |
---|
| 2631 | if [[ $answer = a || $answer = A ]] |
---|
| 2632 | then |
---|
| 2633 | printf "\n +++ Aborting..." |
---|
| 2634 | rm -rf $check_sources |
---|
| 2635 | locat=normal; exit |
---|
| 2636 | fi |
---|
| 2637 | else |
---|
| 2638 | printf "\n\n *** parameter file(s) seem(s) to be o.k.\n" |
---|
| 2639 | fi |
---|
| 2640 | rm -rf $check_sources |
---|
| 2641 | fi |
---|
| 2642 | else |
---|
[1103] | 2643 | printf "\n\n +++ skipping parameter file check due to following reason: $reason \n" |
---|
[807] | 2644 | fi |
---|
| 2645 | |
---|
[818] | 2646 | # DELETE TEMPORARY DIRECTORY AND FINISH NAMELIST FILE CHECK |
---|
| 2647 | rm -rf $check_sources |
---|
[807] | 2648 | cd $working_directory |
---|
| 2649 | |
---|
[1199] | 2650 | # DETERMINE PATH FOR MAKE DEPOSITORY |
---|
| 2651 | if [[ $do_batch = false || $create_executable_for_batch = true ]] |
---|
| 2652 | then |
---|
| 2653 | |
---|
| 2654 | line="" |
---|
| 2655 | grep "%depository_path" $config_file > tmp_mrun |
---|
| 2656 | while read line |
---|
| 2657 | do |
---|
| 2658 | if [[ "$line" != "" && $(echo $line | cut -c1) != "#" ]] |
---|
| 2659 | then |
---|
| 2660 | if [[ "$(echo $line | cut -d" " -s -f3)" = "" ]] |
---|
| 2661 | then |
---|
| 2662 | global_depository_path=`echo $line | cut -d" " -s -f2` |
---|
| 2663 | fi |
---|
| 2664 | fi |
---|
| 2665 | done < tmp_mrun |
---|
| 2666 | |
---|
| 2667 | line="" |
---|
| 2668 | grep " $localhost" $config_file | grep "%depository_path" > tmp_mrun |
---|
| 2669 | while read line |
---|
| 2670 | do |
---|
| 2671 | if [[ "$line" != "" && $(echo $line | cut -c1) != "#" ]] |
---|
| 2672 | then |
---|
| 2673 | if [[ "$(echo $line | cut -d" " -s -f4)" = "$cond1" && "$(echo $line | cut -d" " -s -f5)" = "$cond2" ]] |
---|
| 2674 | then |
---|
| 2675 | local_depository_path=`echo $line | cut -d" " -s -f2` |
---|
| 2676 | fi |
---|
| 2677 | fi |
---|
| 2678 | done < tmp_mrun |
---|
| 2679 | |
---|
| 2680 | if [[ "$local_depository_path" = "" ]] |
---|
| 2681 | then |
---|
| 2682 | if [[ "$global_depository_path" != "" ]] |
---|
| 2683 | then |
---|
| 2684 | local_depository_path=$global_depository_path |
---|
| 2685 | else |
---|
| 2686 | printf "\n\n +++ no depository path found in configuration file" |
---|
| 2687 | printf "\n for local host \"$localhost\" " |
---|
| 2688 | printf "\n please set \"\%depository_path\" in configuration file\n" |
---|
| 2689 | locat=config_file; exit |
---|
| 2690 | fi |
---|
| 2691 | fi |
---|
| 2692 | eval local_depository_path=$local_depository_path |
---|
| 2693 | [[ "$cond1" != "" ]] && local_depository_path=${local_depository_path}_$cond1 |
---|
| 2694 | [[ "$cond2" != "" ]] && local_depository_path=${local_depository_path}_$cond2 |
---|
| 2695 | |
---|
| 2696 | |
---|
| 2697 | basename=`echo $mainprog | cut -f1 -d"."` |
---|
| 2698 | eval make_depository=${local_depository_path}/${basename}_current_version.tar |
---|
| 2699 | if [[ ! -f $make_depository ]] |
---|
| 2700 | then |
---|
| 2701 | printf "\n" |
---|
| 2702 | printf "\n *** WARNING: make depository \"$make_depository\" not found" |
---|
| 2703 | printf "\n \"make\" will fail, if the Makefile or other source files are missing\n" |
---|
| 2704 | fi |
---|
| 2705 | |
---|
| 2706 | fi |
---|
| 2707 | |
---|
| 2708 | |
---|
[1289] | 2709 | # NOW PERFORM THOSE ACTIONS REQUIRED TO EXECUTE THE PROGRAM (PALM) ON THIS MACHINE |
---|
| 2710 | # (COMPILING/LINKING, EXECUTING, COPYING I/O FILES) |
---|
[1] | 2711 | if [[ $do_batch = false ]] |
---|
| 2712 | then |
---|
| 2713 | |
---|
| 2714 | |
---|
[1289] | 2715 | # CREATE THE TEMPORARY WORKING DIRECTORY |
---|
[1] | 2716 | mkdir -p $TEMPDIR |
---|
| 2717 | chmod go+rx $TEMPDIR |
---|
| 2718 | tmpcreate=true |
---|
| 2719 | |
---|
| 2720 | |
---|
[1289] | 2721 | # COPY EITHER THE COMPLETE SOURCE CODE FILES TO BE COMPILED OR THE EXECUTABLE |
---|
| 2722 | # INTO THE TEMPORARY WORKING DIRECTORY |
---|
[1] | 2723 | if [[ $do_compile = true ]] |
---|
| 2724 | then |
---|
| 2725 | |
---|
| 2726 | # ON NEC, COMPILATION IS DONE ON HOST CROSS VIA CROSS COMPILING |
---|
| 2727 | # CREATE A TEMPORARY DIRECTORY ON THAT MACHINE (HOME MOUNTED VIA NFS) |
---|
| 2728 | if [[ $localhost = nech ]] |
---|
| 2729 | then |
---|
| 2730 | TEMPDIR_COMPILE=$HOME/work/${usern}.$kennung |
---|
| 2731 | if mkdir -p $TEMPDIR_COMPILE |
---|
| 2732 | then |
---|
| 2733 | printf "\n *** \"$TEMPDIR_COMPILE\" " |
---|
| 2734 | printf "\n is generated as temporary directory for cross compiling\n" |
---|
| 2735 | else |
---|
| 2736 | printf "\n +++ creating directory \"$TEMPDIR_COMPILE\" " |
---|
| 2737 | printf "\n needed for cross compilation failed" |
---|
| 2738 | locat=compile |
---|
| 2739 | exit |
---|
| 2740 | fi |
---|
| 2741 | else |
---|
| 2742 | TEMPDIR_COMPILE=$TEMPDIR |
---|
| 2743 | fi |
---|
| 2744 | |
---|
[1199] | 2745 | cp $make_depository $TEMPDIR_COMPILE |
---|
| 2746 | cd $TEMPDIR_COMPILE |
---|
| 2747 | tar -xf $make_depository > /dev/null 2>&1 |
---|
| 2748 | cd - > /dev/null |
---|
[1] | 2749 | |
---|
[1255] | 2750 | # DUE TO UNKNOWN REASONS, COPY WITH cp COMMAND CREATES CORRUPT |
---|
| 2751 | # FILES ON CRAY XC30 SYSTEMS (HLRN III), rsync IS USED INSTEAD |
---|
[1281] | 2752 | if [[ $(echo $host | cut -c1-6) = lccray ]] |
---|
| 2753 | then |
---|
| 2754 | rsync -av -t SOURCES_FOR_RUN_$fname/* $TEMPDIR_COMPILE > /dev/null |
---|
| 2755 | else |
---|
| 2756 | cp SOURCES_FOR_RUN_$fname/* $TEMPDIR_COMPILE > /dev/null |
---|
| 2757 | fi |
---|
[82] | 2758 | |
---|
[1] | 2759 | else |
---|
[475] | 2760 | |
---|
[1] | 2761 | cp $executable ${TEMPDIR}/a.out |
---|
[475] | 2762 | |
---|
[1] | 2763 | fi |
---|
| 2764 | |
---|
| 2765 | |
---|
[1289] | 2766 | # CHANGE TO THE TEMPORARY WORKING DIRECTORY |
---|
[475] | 2767 | cd $TEMPDIR |
---|
[1] | 2768 | printf "\n *** changed to temporary directory: $TEMPDIR" |
---|
| 2769 | |
---|
| 2770 | |
---|
[1289] | 2771 | # THE FOLLOWING IS REQUIRED FPR AVS-OUTPUT WITH PALM ??? REMOVE ??? |
---|
| 2772 | # WRITE INFORMATIONS ABOUT THE OUTPUT-FILE CONNECTIONS TO A TEMPORARY FILE |
---|
| 2773 | # THIS FILE CAN LATER BE READ FROM USER-DEFINED CODE TO DETERMINE THE |
---|
| 2774 | # RELATION BETWEEN THE LOCAL TEMPORARY AND PERMANENT FILE NAMES |
---|
[1] | 2775 | (( i = 0 )) |
---|
| 2776 | while (( i < iout )) |
---|
| 2777 | do |
---|
| 2778 | (( i = i + 1 )) |
---|
| 2779 | if [[ "${actionout[$i]}" = tr || "${actionout[$i]}" = tra || "${actionout[$i]}" = trpe ]] |
---|
| 2780 | then |
---|
| 2781 | printf "${localout[$i]} ${actionout[$i]}\n${pathout[$i]}\n${localhost}_${fname}${endout[$i]}\n" >> OUTPUT_FILE_CONNECTIONS |
---|
| 2782 | else |
---|
| 2783 | printf "${localout[$i]} ${actionout[$i]}\n${pathout[$i]}\n${frelout[$i]}\n" >> OUTPUT_FILE_CONNECTIONS |
---|
| 2784 | fi |
---|
| 2785 | done |
---|
| 2786 | |
---|
| 2787 | |
---|
[1289] | 2788 | # IF REQUIRED, START WITH COMPILING |
---|
[1] | 2789 | if [[ $do_compile = true ]] |
---|
| 2790 | then |
---|
| 2791 | |
---|
| 2792 | |
---|
[1199] | 2793 | if [[ -f a.out ]] |
---|
[475] | 2794 | then |
---|
[82] | 2795 | |
---|
[1199] | 2796 | # EXECUTABLE WAS CREATED DURING INTERACTIVE CALL OF MRUN |
---|
| 2797 | printf "\n\n\n *** executable a.out found" |
---|
| 2798 | printf "\n no compilation required \n" |
---|
| 2799 | |
---|
[1] | 2800 | else |
---|
[807] | 2801 | |
---|
[1199] | 2802 | # COMPILING WITH MAKE (ON NEC COMPILER IS CALLED ON HOST CROSS) |
---|
[1289] | 2803 | printf "\n\n\n *** compilation starts \n$dashes\n" |
---|
[1199] | 2804 | printf " *** compilation with make using following options:\n" |
---|
| 2805 | printf " make depository: $make_depository" |
---|
| 2806 | if [[ "$mopts" != "" ]] |
---|
| 2807 | then |
---|
| 2808 | printf " make options: $mopts\n" |
---|
| 2809 | fi |
---|
| 2810 | printf " compilername: $compiler_name\n" |
---|
| 2811 | printf " compiler options: $fopts\n" |
---|
| 2812 | printf " preprocessor directives: $cpp_options \n" |
---|
| 2813 | printf " linker options: $lopts \n" |
---|
| 2814 | if [[ "$modules" != "" ]] |
---|
| 2815 | then |
---|
| 2816 | printf " modules to be load: $modules \n" |
---|
| 2817 | fi |
---|
| 2818 | printf " source code files: $source_list \n" |
---|
[1] | 2819 | |
---|
[1199] | 2820 | if [[ $localhost = nech ]] |
---|
| 2821 | then |
---|
| 2822 | ssh $SSH_PORTOPT 136.172.44.192 -l $usern "$init_cmds $module_calls cd \$HOME/work/${usern}.$kennung; sxmake $mopts -f Makefile PROG=a.out F90=$compiler_name COPT=\"$cpp_options\" F90FLAGS=\"$fopts\" LDFLAGS=\"$lopts\" " |
---|
| 2823 | cp $TEMPDIR_COMPILE/a.out . |
---|
| 2824 | [[ $? != 0 ]] && compile_error=true |
---|
| 2825 | rm -rf $TEMPDIR_COMPILE |
---|
| 2826 | elif [[ $localhost = ibmh ]] |
---|
| 2827 | then |
---|
| 2828 | printf " compiler is called via ssh on \"plogin1\" \n" |
---|
| 2829 | ssh $SSH_PORTOPT plogin1 -l $usern "$init_cmds export PATH=/sw/ibm/xlf/13.1.0.8/usr/bin:$PATH; $module_calls cd $TEMPDIR; make $mopts -f Makefile PROG=a.out F90=$compiler_name COPT=\"$cpp_options\" F90FLAGS=\"$fopts\" LDFLAGS=\"$lopts\" " |
---|
| 2830 | [[ ! -f a.out ]] && compile_error=true |
---|
[1289] | 2831 | continue # STATUS=1, IF a.out EXISTS |
---|
[1199] | 2832 | elif [[ $localhost = lcflow ]] |
---|
| 2833 | then |
---|
| 2834 | printf " compiler is called via ssh on \"flow\" \n" |
---|
[1201] | 2835 | /usr/bin/ssh $SSH_PORTOPT flow02.hpc.uni-oldenburg.de -l $usern "$init_cmds $module_calls cd $TEMPDIR; make $mopts -f Makefile PROG=a.out F90=$compiler_name COPT=\"$cpp_options\" F90FLAGS=\"$fopts\" LDFLAGS=\"$lopts\" " |
---|
[1272] | 2836 | check_for_file=`/usr/bin/ssh $SSH_PORTOPT flow02.hpc.uni-oldenburg.de -l $usern "ls $TEMPDIR/a.out 2> /dev/null"` |
---|
| 2837 | [[ "$check_for_file" = "" ]] && compile_error=true |
---|
[1289] | 2838 | continue # STATUS=1, IF a.out EXISTS |
---|
[1620] | 2839 | elif [[ $localhost = lcbullhh || $localhost = lccrayb || $localhost = lccrayf || $localhost = lccrayh ]] |
---|
[1199] | 2840 | then |
---|
| 2841 | make $mopts -f Makefile PROG=a.out F90=$compiler_name COPT="$cpp_options" F90FLAGS="$fopts" LDFLAGS="$lopts" |
---|
| 2842 | else |
---|
| 2843 | [[ "$init_cmds" != "" ]] && eval $init_cmds |
---|
| 2844 | [[ "$module_calls" != "" ]] && eval $module_calls |
---|
| 2845 | make $mopts -f Makefile PROG=a.out F90=$compiler_name COPT="$cpp_options" F90FLAGS="$fopts" LDFLAGS="$lopts" |
---|
| 2846 | fi |
---|
[181] | 2847 | |
---|
[1199] | 2848 | if [[ $? != 0 || "$compile_error" = true || "$module_compile_error" = true ]] |
---|
| 2849 | then |
---|
| 2850 | printf "\n +++ error occured while compiling or linking" |
---|
| 2851 | locat=compile |
---|
| 2852 | |
---|
[181] | 2853 | exit |
---|
[1199] | 2854 | else |
---|
[1289] | 2855 | printf "$dashes\n *** compilation finished \n" |
---|
[1199] | 2856 | fi |
---|
| 2857 | |
---|
[1] | 2858 | fi |
---|
[1199] | 2859 | |
---|
[1] | 2860 | fi |
---|
| 2861 | |
---|
| 2862 | |
---|
[1289] | 2863 | # PROVIDE THE INPUT FILES |
---|
| 2864 | # LOOP OVER ALL ACTIVATED FILES (LISTED IN THE CONFIGURATION FILE) |
---|
[1] | 2865 | (( i = 0 )) |
---|
| 2866 | while (( i < iin )) |
---|
| 2867 | do |
---|
| 2868 | (( i = i + 1 )) |
---|
| 2869 | if (( i == 1 )) |
---|
| 2870 | then |
---|
[1289] | 2871 | printf "\n\n *** providing INPUT-files:\n$dashes" |
---|
[1] | 2872 | fi |
---|
| 2873 | |
---|
| 2874 | |
---|
[1289] | 2875 | # SKIP OPTIONAL FILES, IF THEY DO NOT EXIST |
---|
[1] | 2876 | if [[ "${transin[$i]}" = unavailable ]] |
---|
| 2877 | then |
---|
| 2878 | if [[ "${extin[$i]}" = "" || "${extin[$i]}" = " " ]] |
---|
| 2879 | then |
---|
| 2880 | printf "\n +++ WARNING: input file \"${pathin[$i]}/${afname}${endin[$i]}\" " |
---|
| 2881 | printf "\n is not available!" |
---|
| 2882 | else |
---|
| 2883 | printf "\n +++ WARNING: input file \"${pathin[$i]}/${afname}${endin[$i]}.${extin[$i]}\" " |
---|
| 2884 | printf "\n is not available!" |
---|
| 2885 | fi |
---|
| 2886 | continue |
---|
| 2887 | fi |
---|
| 2888 | |
---|
[1289] | 2889 | # CHECK FOR SINGLE FILE (SERIAL RUN) OR DIRECTORY (ONE FILE PER CORE FOR PARELLEL EXECUTION) |
---|
[1] | 2890 | files_for_pes=false; datentyp=file |
---|
| 2891 | if [[ "${actionin[$i]}" = pe && -n $numprocs ]] |
---|
| 2892 | then |
---|
| 2893 | files_for_pes=true; datentyp=directory |
---|
| 2894 | actionin[$i]="" |
---|
| 2895 | elif [[ "${actionin[$i]}" = pe && ! -n $numprocs ]] |
---|
| 2896 | then |
---|
| 2897 | actionin[$i]="" |
---|
| 2898 | elif [[ "${actionin[$i]}" = arpe && -n $numprocs ]] |
---|
| 2899 | then |
---|
| 2900 | files_for_pes=true; datentyp=directory |
---|
| 2901 | actionin[$i]="ar" |
---|
| 2902 | elif [[ "${actionin[$i]}" = arpe && ! -n $numprocs ]] |
---|
| 2903 | then |
---|
| 2904 | actionin[$i]="ar" |
---|
| 2905 | elif [[ "${actionin[$i]}" = flpe && -n $numprocs ]] |
---|
| 2906 | then |
---|
| 2907 | files_for_pes=true; datentyp=directory |
---|
| 2908 | actionin[$i]="fl" |
---|
| 2909 | elif [[ "${actionin[$i]}" = flpe && ! -n $numprocs ]] |
---|
| 2910 | then |
---|
| 2911 | actionin[$i]="fl" |
---|
| 2912 | fi |
---|
| 2913 | |
---|
| 2914 | if [[ $files_for_pes = true ]] |
---|
| 2915 | then |
---|
| 2916 | printf "\n >>> INPUT: ${absnamein[$i]}/.... to ${localin[$i]}" |
---|
| 2917 | else |
---|
| 2918 | printf "\n >>> INPUT: ${absnamein[$i]} to ${localin[$i]}" |
---|
| 2919 | fi |
---|
| 2920 | |
---|
[1289] | 2921 | # INPUT-FILE FOR A RESTART RUN. |
---|
| 2922 | # CHECK, IF THIS FILE STILL EXISTS ON THE TEMPORARY DATA DIRECTORY. |
---|
| 2923 | # IF NOT, TRY TO GET IT FROM THE ARCHIVE SYSTEM |
---|
[1] | 2924 | if [[ "${actionin[$i]}" = fl ]] |
---|
| 2925 | then |
---|
| 2926 | printf "\n $datentyp will be fetched from temporary directory \"${tmp_data_catalog}\" !" |
---|
| 2927 | if [[ $files_for_pes = false ]] |
---|
| 2928 | then |
---|
| 2929 | if [[ -f "$tmp_data_catalog/${frelin[$i]}" ]] |
---|
| 2930 | then |
---|
| 2931 | ln $tmp_data_catalog/${frelin[$i]} ${localin[$i]} |
---|
| 2932 | got_tmp[$i]=true |
---|
| 2933 | elif [[ -f "$WORK/${frelin[$i]}" && $ignore_archive_error = true ]] |
---|
| 2934 | then |
---|
| 2935 | printf "\n +++ $datentyp not found in \"$tmp_data_catalog\" !" |
---|
| 2936 | printf "\n *** trying to use backup copy in \"$WORK\" " |
---|
| 2937 | cp $WORK/${frelin[$i]} ${localin[$i]} |
---|
| 2938 | else |
---|
| 2939 | printf "\n +++ $datentyp not found in \"$tmp_data_catalog\" " |
---|
| 2940 | printf "\n or \"$tmp_data_catalog\" does not exist!" |
---|
| 2941 | printf "\n *** trying to get copy from archive" |
---|
| 2942 | actionin[$i]=ar |
---|
| 2943 | fi |
---|
| 2944 | else |
---|
| 2945 | if [[ -d "$tmp_data_catalog/${frelin[$i]}" ]] |
---|
| 2946 | then |
---|
| 2947 | mkdir ${localin[$i]} |
---|
[149] | 2948 | cd $tmp_data_catalog/${frelin[$i]} |
---|
| 2949 | for file in $(ls *) |
---|
| 2950 | do |
---|
| 2951 | ln $file $TEMPDIR/${localin[$i]} |
---|
| 2952 | done |
---|
| 2953 | cd $TEMPDIR |
---|
[1] | 2954 | got_tmp[$i]=true |
---|
| 2955 | elif [[ -d "$WORK/${frelin[$i]}" && $ignore_archive_error = true ]] |
---|
| 2956 | then |
---|
| 2957 | printf "\n +++ $datentyp not found in \"$tmp_data_catalog\" !" |
---|
| 2958 | printf "\n *** trying to use backup copy in \"$WORK\" " |
---|
| 2959 | cp -r $WORK/${frelin[$i]} ${localin[$i]} |
---|
| 2960 | else |
---|
| 2961 | printf "\n +++ $datentyp not found in \"$tmp_data_catalog\" " |
---|
| 2962 | printf "\n or \"$tmp_data_catalog\" does not exist!" |
---|
| 2963 | printf "\n *** trying to get copy from archive" |
---|
| 2964 | actionin[$i]=ar |
---|
| 2965 | fi |
---|
| 2966 | fi |
---|
| 2967 | fi |
---|
| 2968 | |
---|
| 2969 | |
---|
[1289] | 2970 | # FILE IS STORED ON THE ARCHIVE SYSTEM |
---|
[1] | 2971 | if [[ "${actionin[$i]}" = ar ]] |
---|
| 2972 | then |
---|
| 2973 | |
---|
| 2974 | if [[ $files_for_pes = false ]] |
---|
| 2975 | then |
---|
| 2976 | printf "\n file will be restored from archive-system ($archive_system)!" |
---|
| 2977 | else |
---|
| 2978 | printf "\n directory will be restored from archive-system ($archive_system)!" |
---|
| 2979 | fi |
---|
| 2980 | |
---|
| 2981 | file_restored=false |
---|
| 2982 | |
---|
| 2983 | if [[ $archive_system = asterix ]] |
---|
| 2984 | then |
---|
| 2985 | do_stagein=true |
---|
| 2986 | (( stagein_anz = 0 )) |
---|
| 2987 | while [[ $do_stagein = true ]] |
---|
| 2988 | do |
---|
| 2989 | if [[ $files_for_pes = false ]] |
---|
| 2990 | then |
---|
| 2991 | stagein -O ${frelin[$i]} > STAGEIN_OUTPUT |
---|
| 2992 | else |
---|
| 2993 | stagein -t -O ${frelin[$i]} > STAGEIN_OUTPUT |
---|
| 2994 | fi |
---|
| 2995 | cat STAGEIN_OUTPUT |
---|
| 2996 | if [[ $(grep -c "st.msg:i24" STAGEIN_OUTPUT) != 0 ]] |
---|
| 2997 | then |
---|
| 2998 | file_restored=true |
---|
| 2999 | do_stagein=false |
---|
| 3000 | else |
---|
| 3001 | (( stagein_anz = stagein_anz + 1 )) |
---|
| 3002 | if (( stagein_anz == 10 )) |
---|
| 3003 | then |
---|
| 3004 | printf "\n +++ stagein stoped after 10 tries" |
---|
| 3005 | locat=stage |
---|
| 3006 | exit |
---|
| 3007 | fi |
---|
| 3008 | printf "\n +++ restoring from archive failed, trying again:" |
---|
| 3009 | sleep 900 |
---|
| 3010 | fi |
---|
| 3011 | done |
---|
| 3012 | elif [[ $archive_system = DMF ]] |
---|
| 3013 | then |
---|
| 3014 | if [[ $files_for_pes = false ]] |
---|
| 3015 | then |
---|
| 3016 | printf "\n +++ restoring of single files impossible with $archive_system !\n" |
---|
| 3017 | locat=DMF |
---|
| 3018 | exit |
---|
| 3019 | else |
---|
| 3020 | find $ARCHIVE/${frelin[$i]} -type m -print | dmget |
---|
| 3021 | cp -r $ARCHIVE/${frelin[$i]} $PWD |
---|
| 3022 | file_restored=true |
---|
| 3023 | fi |
---|
| 3024 | elif [[ $archive_system = tivoli ]] |
---|
| 3025 | then |
---|
| 3026 | if [[ $files_for_pes = false ]] |
---|
| 3027 | then |
---|
[1289] | 3028 | ssh $SSH_PORTOPT $usern@bicedata.hlrn.de "cp $PERM/${frelin[$i]} $PWD" |
---|
[1] | 3029 | else |
---|
| 3030 | (( inode = 0 )) |
---|
| 3031 | while (( inode < nodes )) |
---|
| 3032 | do |
---|
[1289] | 3033 | ssh $SSH_PORTOPT $usern@bicedata.hlrn.de "cd $PWD; tar xf $PERM/${frelin[$i]}/${frelin[$i]}.node_$inode.tar" |
---|
[1] | 3034 | (( inode = inode + 1 )) |
---|
| 3035 | done |
---|
| 3036 | fi |
---|
| 3037 | file_restored=true |
---|
| 3038 | elif [[ $archive_system = ut ]] |
---|
| 3039 | then |
---|
| 3040 | if [[ $files_for_pes = false ]] |
---|
| 3041 | then |
---|
| 3042 | cp $UT/${frelin[$i]} . |
---|
| 3043 | else |
---|
| 3044 | (( inode = 0 )) |
---|
| 3045 | while (( inode < nodes )) |
---|
| 3046 | do |
---|
| 3047 | tar xf $UT/${frelin[$i]}/${frelin[$i]}.node_$inode.tar |
---|
| 3048 | (( inode = inode + 1 )) |
---|
| 3049 | done |
---|
| 3050 | fi |
---|
| 3051 | file_restored=true |
---|
| 3052 | else |
---|
| 3053 | printf "\n +++ archive_system=\"$archive_system\" restore impossible!" |
---|
| 3054 | locat=rearchive |
---|
| 3055 | exit |
---|
| 3056 | fi |
---|
| 3057 | |
---|
| 3058 | if [[ $file_restored = true ]] |
---|
| 3059 | then |
---|
| 3060 | |
---|
[1289] | 3061 | # PUT FILE ON THE TEMPORARY DATA DIRECTORY TOO. |
---|
| 3062 | # THIS WILL MAKE FILE ACCESS EASIER FOR LATER RUNS |
---|
[1] | 3063 | [[ ! -d $tmp_data_catalog ]] && mkdir -p $tmp_data_catalog; chmod g+rx $tmp_data_catalog |
---|
| 3064 | if [[ $files_for_pes = false ]] |
---|
| 3065 | then |
---|
| 3066 | ln -f ${frelin[$i]} $tmp_data_catalog/${frelin[$i]} |
---|
| 3067 | else |
---|
| 3068 | mkdir $tmp_data_catalog/${frelin[$i]} |
---|
| 3069 | ln -f ${frelin[$i]}/* $tmp_data_catalog/${frelin[$i]} |
---|
| 3070 | fi |
---|
| 3071 | got_tmp[$i]=true |
---|
| 3072 | |
---|
[1289] | 3073 | # PROVIDE THE FILE UNDER ITS LOCAL NAME |
---|
[1] | 3074 | mv ${frelin[$i]} ${localin[$i]} |
---|
| 3075 | |
---|
| 3076 | fi |
---|
| 3077 | fi |
---|
| 3078 | |
---|
| 3079 | |
---|
[1289] | 3080 | # FILE IS STORED IN THE RESPECTIVE DIRECTORY GIVEN IN THE CONFIGURATION FILE |
---|
[1] | 3081 | if [[ "${actionin[$i]}" = "" || "${actionin[$i]}" = "di" || "${actionin[$i]}" = "npe" ]] |
---|
| 3082 | then |
---|
| 3083 | |
---|
| 3084 | if [[ "${actionin[$i]}" = "npe" && -n $numprocs ]] |
---|
| 3085 | then |
---|
| 3086 | |
---|
[1289] | 3087 | # FILE COPIES ARE PROVIDED FOR ALL CORES |
---|
| 3088 | # EACH FILE GETS A UNIQUE FILENAME WITH A FOUR DIGIT NUMBER |
---|
[1] | 3089 | printf "\n file will be provided for $numprocs processors" |
---|
| 3090 | mkdir ${localin[$i]} |
---|
| 3091 | ival=$numprocs |
---|
| 3092 | (( ii = 0 )) |
---|
| 3093 | while (( ii <= ival-1 )) |
---|
| 3094 | do |
---|
| 3095 | if (( ii < 10 )) |
---|
| 3096 | then |
---|
| 3097 | cp ${absnamein[$i]} ${localin[$i]}/_000$ii |
---|
| 3098 | elif (( ii < 100 )) |
---|
| 3099 | then |
---|
| 3100 | cp ${absnamein[$i]} ${localin[$i]}/_00$ii |
---|
| 3101 | elif (( ii < 1000 )) |
---|
| 3102 | then |
---|
| 3103 | cp ${absnamein[$i]} ${localin[$i]}/_0$ii |
---|
| 3104 | else |
---|
| 3105 | cp ${absnamein[$i]} ${localin[$i]}/_$ii |
---|
| 3106 | fi |
---|
| 3107 | (( ii = ii + 1 )) |
---|
| 3108 | done |
---|
| 3109 | |
---|
| 3110 | else |
---|
| 3111 | |
---|
| 3112 | if [[ $files_for_pes = true ]] |
---|
| 3113 | then |
---|
| 3114 | |
---|
[1289] | 3115 | # PROVIDE FILES FOR EACH CORE |
---|
| 3116 | # FIRST CREATE THE LOCAL DIRECTORY, THEN COPY FILES |
---|
| 3117 | # FROM THE PERMANENT DIRECTORY BY LINKING THEM TO THE LOCAL ONE |
---|
[1] | 3118 | printf "\n providing $numprocs files for the respective processors" |
---|
| 3119 | mkdir ${localin[$i]} |
---|
[183] | 3120 | if [[ $link_local_input = true ]] |
---|
| 3121 | then |
---|
| 3122 | printf " using ln -f\n" |
---|
| 3123 | cd ${absnamein[$i]} |
---|
| 3124 | for file in $(ls *) |
---|
| 3125 | do |
---|
| 3126 | ln -f $file ${localin[$i]} |
---|
| 3127 | done |
---|
| 3128 | cd $TEMPDIR |
---|
| 3129 | fi |
---|
[1289] | 3130 | |
---|
| 3131 | # IF "ln -f" FAILED OR IF "$link_local_input = false" DO A NORMAL "cp -r" |
---|
[183] | 3132 | if [[ ! -f "${localin[$i]}/_0000" ]] |
---|
| 3133 | then |
---|
| 3134 | if [[ $link_local_input = true ]] |
---|
| 3135 | then |
---|
| 3136 | printf " ln failed for .../_0000, using cp...\n" |
---|
| 3137 | fi |
---|
| 3138 | cp -r ${absnamein[$i]}/* ${localin[$i]} |
---|
| 3139 | fi |
---|
[1] | 3140 | |
---|
| 3141 | else |
---|
[1289] | 3142 | |
---|
| 3143 | # PROVIDE FILE FOR RUNS ON A SINGLE CORE |
---|
[183] | 3144 | if [[ $link_local_input = true ]] |
---|
| 3145 | then |
---|
| 3146 | printf " using ln -f\n" |
---|
| 3147 | ln -f ${absnamein[$i]} ${localin[$i]} |
---|
| 3148 | fi |
---|
| 3149 | # If "ln -f" fails of if "$link_local_input = false" do a normal "cp" |
---|
| 3150 | if [[ ! -f "${localin[$i]}" ]] |
---|
| 3151 | then |
---|
| 3152 | if [[ $link_local_input = true ]] |
---|
| 3153 | then |
---|
| 3154 | printf " ln failed, using cp...\n" |
---|
| 3155 | fi |
---|
| 3156 | cp ${absnamein[$i]} ${localin[$i]} |
---|
| 3157 | fi |
---|
[1] | 3158 | fi |
---|
| 3159 | fi |
---|
| 3160 | fi |
---|
| 3161 | |
---|
| 3162 | done |
---|
| 3163 | if (( i != 0 )) |
---|
| 3164 | then |
---|
[1289] | 3165 | printf "\n$dashes\n *** all INPUT-files provided \n" |
---|
[1] | 3166 | fi |
---|
| 3167 | |
---|
| 3168 | |
---|
[1289] | 3169 | # EXECUTE INPUT-COMMANDS GIVEN IN THE CONFIGURATION FILE |
---|
[1] | 3170 | (( i = 0 )) |
---|
| 3171 | while (( i < iic )) |
---|
| 3172 | do |
---|
| 3173 | (( i = i + 1 )) |
---|
| 3174 | if (( i == 1 )) |
---|
| 3175 | then |
---|
[1289] | 3176 | printf "\n\n *** execution of INPUT-commands:\n$dashes" |
---|
[1] | 3177 | fi |
---|
| 3178 | printf "\n >>> ${in_command[$i]}" |
---|
| 3179 | eval ${in_command[$i]} |
---|
| 3180 | if (( i == iic )) |
---|
| 3181 | then |
---|
[1289] | 3182 | printf "\n$dashes\n" |
---|
[1] | 3183 | fi |
---|
| 3184 | done |
---|
| 3185 | |
---|
| 3186 | |
---|
[1289] | 3187 | # SET THE REMAINING CPU-TIME |
---|
[1] | 3188 | cpurest=${cpumax}. |
---|
| 3189 | |
---|
| 3190 | |
---|
[260] | 3191 | # START DVR STREAMING SERVER |
---|
| 3192 | if [[ $(echo $package_list | grep -c dvrp_graphics) != 0 ]] |
---|
| 3193 | then |
---|
| 3194 | if [[ "$dvr_server" != "" ]] |
---|
| 3195 | then |
---|
| 3196 | |
---|
| 3197 | printf "\n\n *** preparing the dvr streaming server configuration file" |
---|
| 3198 | |
---|
[1289] | 3199 | # CHECK, IF A DVR SERVER IS ALREADY RUNNING |
---|
[260] | 3200 | running_dvrserver_id=`echo $(ps -edaf | grep .dvrserver.config | grep -v grep) | cut -d" " -f2` |
---|
| 3201 | if [[ "$running_dvrserver_id" != "" ]] |
---|
| 3202 | then |
---|
| 3203 | |
---|
[261] | 3204 | printf "\n\n +++ WARNING: A dvr server with id=$running_dvrserver_id is already running!" |
---|
[260] | 3205 | printf "\n This server is used instead starting a new one!" |
---|
[261] | 3206 | printf "\n If required, script \"process_dvr_output\" has to be run manually." |
---|
[260] | 3207 | |
---|
| 3208 | else |
---|
| 3209 | |
---|
| 3210 | # COPY CONFIGURATION FILE FOR STREAMING SERVER FROM REPOSITORY TO HERE |
---|
| 3211 | if [[ -f ${PALM_BIN}/.dvrserver.config ]] |
---|
| 3212 | then |
---|
| 3213 | cp ${PALM_BIN}/.dvrserver.config . |
---|
| 3214 | |
---|
[1289] | 3215 | # ENTERING THE BASEDIR, UID AND GID INTO THIS FILE |
---|
[260] | 3216 | user_id=`id -u` |
---|
| 3217 | group_id=`id -g` |
---|
[1289] | 3218 | |
---|
| 3219 | # "&" IS REQUIRED AS A SEPERATOR, BECAUSE TEMPDIR CONTAINS "/" |
---|
[260] | 3220 | sed "s&<replace by dvr data directory>&${TEMPDIR}&g" .dvrserver.config > .dvrserver.1 |
---|
| 3221 | sed "s/<replace by user id>/$user_id/g" .dvrserver.1 > .dvrserver.2 |
---|
| 3222 | sed "s/<replace by group id>/$group_id/g" .dvrserver.2 > .dvrserver.3 |
---|
| 3223 | mv .dvrserver.3 .dvrserver.config |
---|
| 3224 | rm .dvrserver.1 .dvrserver.2 |
---|
| 3225 | |
---|
[1289] | 3226 | # START DVR SERVER IN BACKGROUND, GET HIS ID AND PRINT ON TERMINAL |
---|
[260] | 3227 | $dvr_server .dvrserver.config >> DVR_LOGFILE 2>&1 & |
---|
| 3228 | dvrserver_id=`echo $(ps -edaf | grep .dvrserver.config) | cut -d" " -f2` |
---|
| 3229 | printf "\n *** streaming server with id=$dvrserver_id is started in background" |
---|
| 3230 | local_dvrserver_running=.TRUE. |
---|
| 3231 | else |
---|
| 3232 | printf "\n +++ missing file \".dvrserver.config\" in directory:" |
---|
| 3233 | printf "\n \"$PALM_BIN\" " |
---|
| 3234 | locat=dvr |
---|
| 3235 | exit |
---|
| 3236 | fi |
---|
| 3237 | |
---|
| 3238 | fi |
---|
| 3239 | |
---|
| 3240 | else |
---|
| 3241 | printf "\n\n --- INFORMATIVE: no dvr streaming server will be started" |
---|
| 3242 | fi |
---|
| 3243 | fi |
---|
| 3244 | |
---|
| 3245 | |
---|
[1289] | 3246 | # CREATE THE NAMELIST-FILE WITH VALUES OF ENVIRONMENT-VARIABLES REQUIRED BY PALM |
---|
| 3247 | # (FILE ENVPAR WILL BE READ BY PALM) |
---|
[260] | 3248 | cat > ENVPAR << %%END%% |
---|
| 3249 | &envpar run_identifier = '$fname', host = '$localhost', |
---|
| 3250 | write_binary = '$write_binary', tasks_per_node = $tasks_per_node, |
---|
[759] | 3251 | maximum_parallel_io_streams = $maximum_parallel_io_streams, |
---|
[260] | 3252 | maximum_cpu_time_allowed = ${cpumax}., |
---|
| 3253 | revision = '$global_revision', |
---|
[1402] | 3254 | local_dvrserver_running = $local_dvrserver_running, |
---|
| 3255 | batch_job = $batch_job / |
---|
[260] | 3256 | |
---|
| 3257 | %%END%% |
---|
| 3258 | |
---|
| 3259 | |
---|
[1289] | 3260 | # STARTING THE EXECUTABLE |
---|
| 3261 | printf "\n\n *** execution starts in directory\n \"`pwd`\"\n$dashes\n" |
---|
[1] | 3262 | PATH=$PATH:$TEMPDIR |
---|
| 3263 | |
---|
[1289] | 3264 | # MPI DEBUG OPTION (ARGUMENT CHECKING, SLOWS DOWN EXECUTION DUE TO INCREASED LATENCY) |
---|
[359] | 3265 | if [[ "$mpi_debug" = true ]] |
---|
| 3266 | then |
---|
| 3267 | export MPI_CHECK_ARGS=1 |
---|
| 3268 | printf "\n +++ MPI_CHECK_ARGS=$MPI_CHECK_ARGS" |
---|
| 3269 | fi |
---|
| 3270 | |
---|
[693] | 3271 | if [[ "$totalview" = true ]] |
---|
| 3272 | then |
---|
| 3273 | printf "\n *** totalview debugger will be used" |
---|
| 3274 | tv_opt="-tv" |
---|
| 3275 | else |
---|
| 3276 | tv_opt="" |
---|
| 3277 | fi |
---|
| 3278 | |
---|
[1] | 3279 | if [[ "$cond1" = debug || "$cond2" = debug ]] |
---|
| 3280 | then |
---|
[1549] | 3281 | |
---|
| 3282 | #Interactive ALLINEA DEBUG seesion |
---|
| 3283 | if [[ "$ENVIRONMENT" != BATCH ]] |
---|
| 3284 | then |
---|
| 3285 | if [[ $host = lccrayb || $host = lccrayh ]] |
---|
| 3286 | then |
---|
| 3287 | if [[ "$allinea" = true ]] |
---|
| 3288 | then |
---|
| 3289 | echo "--- aprun -n $ii -N $tasks_per_node a.out $ROPTS < runfile_atmos" |
---|
| 3290 | |
---|
| 3291 | ddt aprun -n $ii -N $tasks_per_node a.out $ROPTS |
---|
| 3292 | wait |
---|
| 3293 | fi |
---|
| 3294 | fi |
---|
| 3295 | fi |
---|
| 3296 | |
---|
[1] | 3297 | if [[ "$ENVIRONMENT" = BATCH ]] |
---|
| 3298 | then |
---|
[1289] | 3299 | continue |
---|
[1] | 3300 | fi |
---|
[1096] | 3301 | if [[ $localhost = ibmh ]] |
---|
[1] | 3302 | then |
---|
| 3303 | |
---|
| 3304 | # SETUP THE IBM MPI ENVIRONMENT |
---|
| 3305 | export MP_SHARED_MEMORY=yes |
---|
| 3306 | export AIXTHREADS_SCOPE=S |
---|
| 3307 | export OMP_NUM_THREADS=$threads_per_task |
---|
| 3308 | export AUTHSTATE=files |
---|
| 3309 | export XLFRTEOPTS="nlwidth=132:err_recovery=no" # RECORD-LENGTH OF NAMELIST-OUTPUT |
---|
| 3310 | |
---|
| 3311 | # FOLLOWING OPTIONS ARE MANDATORY FOR TOTALVIEW |
---|
| 3312 | export MP_ADAPTER_USE=shared |
---|
| 3313 | export MP_CPU_USE=multiple |
---|
| 3314 | export MP_TIMEOUT=1200 |
---|
| 3315 | |
---|
| 3316 | unset MP_TASK_AFFINITY |
---|
| 3317 | |
---|
| 3318 | # SO FAR, TOTALVIEW NEEDS HOSTFILE MECHANISM FOR EXECUTION |
---|
| 3319 | #(( ii = 1 )) |
---|
| 3320 | #while (( ii <= $numprocs )) |
---|
| 3321 | #do |
---|
| 3322 | # echo $localhost_realname >> hostfile |
---|
| 3323 | # (( ii = ii + 1 )) |
---|
| 3324 | #done |
---|
| 3325 | #export MP_HOSTFILE=hostfile |
---|
| 3326 | |
---|
| 3327 | if [[ "$LOADLBATCH" = yes ]] |
---|
| 3328 | then |
---|
| 3329 | totalview poe a.out $ROPTS |
---|
| 3330 | else |
---|
| 3331 | echo totalview poe -a a.out -procs $numprocs -rmpool 0 -nodes 1 $ROPTS |
---|
| 3332 | export TVDSVRLAUNCHCMD=ssh |
---|
| 3333 | totalview poe -a a.out -procs $numprocs -rmpool 0 -nodes 1 $ROPTS |
---|
| 3334 | fi |
---|
[622] | 3335 | |
---|
[1295] | 3336 | fi # END DEBUG MODE |
---|
| 3337 | |
---|
[1] | 3338 | else |
---|
[622] | 3339 | |
---|
[1289] | 3340 | # NORMAL EXECUTION |
---|
[1] | 3341 | if [[ -n $numprocs ]] |
---|
| 3342 | then |
---|
[1289] | 3343 | |
---|
[1] | 3344 | # RUNNING THE PROGRAM ON PARALLEL MACHINES |
---|
| 3345 | if [[ $(echo $host | cut -c1-3) = ibm ]] |
---|
| 3346 | then |
---|
[1289] | 3347 | |
---|
[1] | 3348 | # SETUP THE IBM MPI ENVIRONMENT |
---|
[693] | 3349 | if [[ $host != ibmh && $host != ibmkisti ]] |
---|
[312] | 3350 | then |
---|
| 3351 | export MP_SHARED_MEMORY=yes |
---|
| 3352 | export AIXTHREAD_SCOPE=S |
---|
| 3353 | export OMP_NUM_THREADS=$threads_per_task |
---|
| 3354 | export XLSMPOPTS="spins=0:yields=0:stack=20000000" |
---|
| 3355 | export AUTHSTATE=files |
---|
| 3356 | export XLFRTEOPTS="nlwidth=132:err_recovery=no" # RECORD-LENGTH OF NAMELIST-OUTPUT |
---|
| 3357 | # export MP_PRINTENV=yes |
---|
[1] | 3358 | |
---|
[1289] | 3359 | # TUNING-VARIABLES TO IMPROVE COMMUNICATION SPEED |
---|
| 3360 | # DO NOT SHOW SIGNIFICANT EFFECTS (SEP 04, FEDERATION-SWITCH) |
---|
[312] | 3361 | export MP_WAIT_MODE=poll |
---|
| 3362 | [[ $node_usage = not_shared ]] && export MP_SINGLE_THREAD=yes |
---|
| 3363 | fi |
---|
[693] | 3364 | |
---|
| 3365 | if [[ $host = ibmkisti ]] |
---|
| 3366 | then |
---|
| 3367 | export LANG=en_US |
---|
| 3368 | export MP_SHARED_MEMORY=yes |
---|
[696] | 3369 | if [[ $threads_per_task = 1 ]] |
---|
| 3370 | then |
---|
| 3371 | export MP_SINGLE_THREAD=yes |
---|
| 3372 | export MEMORY_AFFINITY=MCM |
---|
| 3373 | else |
---|
| 3374 | export OMP_NUM_THREADS=$threads_per_task |
---|
| 3375 | fi |
---|
[693] | 3376 | fi |
---|
| 3377 | |
---|
[1] | 3378 | if [[ "$LOADLBATCH" = yes ]] |
---|
| 3379 | then |
---|
| 3380 | printf "\n--- Control: OMP_NUM_THREADS = \"$OMP_NUM_THREADS\" \n" |
---|
| 3381 | if [[ "$cond1" = hpmcount || "$cond2" = hpmcount ]] |
---|
| 3382 | then |
---|
[102] | 3383 | /opt/optibm/HPM_2_4_1/bin/hpmcount a.out $ROPTS |
---|
[1] | 3384 | else |
---|
[343] | 3385 | if [[ $run_coupled_model = false ]] |
---|
| 3386 | then |
---|
| 3387 | if [[ "$ocean_file_appendix" = true ]] |
---|
| 3388 | then |
---|
| 3389 | echo "precursor_ocean" > runfile_atmos |
---|
| 3390 | else |
---|
| 3391 | echo "precursor_atmos" > runfile_atmos |
---|
| 3392 | fi |
---|
| 3393 | else |
---|
| 3394 | (( iia = $numprocs_atmos / $threads_per_task )) |
---|
| 3395 | (( iio = $numprocs_ocean / $threads_per_task )) |
---|
| 3396 | printf "\n coupled run ($iia atmosphere, $iio ocean)" |
---|
| 3397 | printf "\n using $coupled_mode coupling" |
---|
| 3398 | printf "\n\n" |
---|
| 3399 | echo "coupled_run $iia $iio" > runfile_atmos |
---|
| 3400 | fi |
---|
| 3401 | poe ./a.out $ROPTS < runfile_atmos |
---|
[1] | 3402 | fi |
---|
| 3403 | else |
---|
[622] | 3404 | if [[ $localhost = ibmh || $localhost = ibms ]] |
---|
[1] | 3405 | then |
---|
[102] | 3406 | poe a.out -procs $numprocs -nodes 1 -rmpool 0 $ROPTS |
---|
[1289] | 3407 | elif [[ $localhost = ibmkisti || $localhost = ibmku ]] |
---|
[1] | 3408 | then |
---|
| 3409 | if [[ -f $hostfile ]] |
---|
| 3410 | then |
---|
| 3411 | cp $hostfile hostfile |
---|
| 3412 | else |
---|
| 3413 | (( ii = 1 )) |
---|
| 3414 | while (( ii <= $numprocs )) |
---|
| 3415 | do |
---|
| 3416 | echo $localhost_realname >> hostfile |
---|
| 3417 | (( ii = ii + 1 )) |
---|
| 3418 | done |
---|
| 3419 | fi |
---|
| 3420 | export MP_HOSTFILE=hostfile |
---|
[331] | 3421 | if [[ $run_coupled_model = false ]] |
---|
| 3422 | then |
---|
| 3423 | if [[ "$ocean_file_appendix" = true ]] |
---|
| 3424 | then |
---|
| 3425 | echo "precursor_ocean" > runfile_atmos |
---|
| 3426 | else |
---|
| 3427 | echo "precursor_atmos" > runfile_atmos |
---|
| 3428 | fi |
---|
| 3429 | else |
---|
| 3430 | (( iia = $numprocs_atmos / $threads_per_task )) |
---|
| 3431 | (( iio = $numprocs_ocean / $threads_per_task )) |
---|
| 3432 | printf "\n coupled run ($iia atmosphere, $iio ocean)" |
---|
| 3433 | printf "\n using $coupled_mode coupling" |
---|
| 3434 | printf "\n\n" |
---|
| 3435 | echo "coupled_run $iia $iio" > runfile_atmos |
---|
| 3436 | fi |
---|
| 3437 | |
---|
[1289] | 3438 | poe ./a.out -procs $numprocs $ROPTS < runfile_atmos |
---|
| 3439 | |
---|
[1] | 3440 | else |
---|
| 3441 | if [[ "$host_file" = "" ]] |
---|
| 3442 | then |
---|
| 3443 | printf "\n +++ no hostfile given in configuration file" |
---|
| 3444 | locat=config_file |
---|
| 3445 | exit |
---|
| 3446 | else |
---|
| 3447 | eval host_file=$host_file |
---|
| 3448 | fi |
---|
| 3449 | export MP_HOSTFILE=$host_file |
---|
[102] | 3450 | poe a.out -procs $numprocs -tasks_per_node $numprocs $ROPTS |
---|
[1] | 3451 | fi |
---|
| 3452 | fi |
---|
[1289] | 3453 | elif [[ $host = nech ]] |
---|
[1] | 3454 | then |
---|
| 3455 | (( ii = nodes )) |
---|
| 3456 | if [[ $ii = 1 ]] |
---|
| 3457 | then |
---|
| 3458 | export F_ERRCNT=0 # acceptable number of errors before program is stopped |
---|
| 3459 | export MPIPROGINF=YES |
---|
[61] | 3460 | # export F_TRACE=YES|FMT1|FMT2 # output of ftrace informations to job protocol |
---|
[1] | 3461 | echo "*** execution on single node with mpirun" |
---|
[102] | 3462 | mpirun -np $numprocs ./a.out $ROPTS |
---|
[1] | 3463 | else |
---|
| 3464 | (( i = 0 )) |
---|
| 3465 | while (( i < ii )) |
---|
| 3466 | do |
---|
| 3467 | echo "-h $i -p $tasks_per_node -e ./mpi_exec_shell" >> multinode_config |
---|
| 3468 | (( i = i + 1 )) |
---|
| 3469 | done |
---|
| 3470 | |
---|
| 3471 | echo "#!/bin/sh" > mpi_exec_shell |
---|
| 3472 | echo " " >> mpi_exec_shell |
---|
| 3473 | echo "set -u" >> mpi_exec_shell |
---|
| 3474 | echo "F_ERRCNT=0" >> mpi_exec_shell |
---|
| 3475 | echo "MPIPROGINV=YES" >> mpi_exec_shell |
---|
| 3476 | echo "OMP_NUM_THREADS=$threads_per_task" >> mpi_exec_shell |
---|
| 3477 | echo "cpurest=$cpurest" >> mpi_exec_shell |
---|
| 3478 | echo "fname=$fname" >> mpi_exec_shell |
---|
| 3479 | echo "localhost=$localhost" >> mpi_exec_shell |
---|
[1468] | 3480 | echo "return_address=$return_address" >> mpi_exec_shell |
---|
[1] | 3481 | echo "return_username=$return_username" >> mpi_exec_shell |
---|
| 3482 | echo "tasks_per_node=$tasks_per_node" >> mpi_exec_shell |
---|
| 3483 | echo "write_binary=$write_binary" >> mpi_exec_shell |
---|
| 3484 | echo "use_seperate_pe_for_dvrp_output=$use_seperate_pe_for_dvrp_output" >> mpi_exec_shell |
---|
| 3485 | echo " " >> mpi_exec_shell |
---|
| 3486 | echo "export F_ERRCNT" >> mpi_exec_shell |
---|
| 3487 | echo "export MPIPROGINV" >> mpi_exec_shell |
---|
| 3488 | echo "export OMP_NUM_THREADS" >> mpi_exec_shell |
---|
| 3489 | echo "export cpurest" >> mpi_exec_shell |
---|
| 3490 | echo "export fname" >> mpi_exec_shell |
---|
| 3491 | echo "export localhost" >> mpi_exec_shell |
---|
[1468] | 3492 | echo "export return_address" >> mpi_exec_shell |
---|
[1] | 3493 | echo "export return_username" >> mpi_exec_shell |
---|
| 3494 | echo "export tasks_per_node" >> mpi_exec_shell |
---|
| 3495 | echo "export write_binary" >> mpi_exec_shell |
---|
| 3496 | echo "export use_seperate_pe_for_dvrp_output" >> mpi_exec_shell |
---|
| 3497 | echo " " >> mpi_exec_shell |
---|
[102] | 3498 | echo "exec ./a.out" >> mpi_exec_shell |
---|
[1] | 3499 | |
---|
| 3500 | chmod u+x mpi_exec_shell |
---|
| 3501 | export MPIPROGINF=YES |
---|
| 3502 | mpirun -f multinode_config & |
---|
| 3503 | wait |
---|
| 3504 | |
---|
| 3505 | fi |
---|
[440] | 3506 | elif [[ $(echo $host | cut -c1-2) = lc && $host != lckyoto && $host != lctit ]] |
---|
[1] | 3507 | then |
---|
| 3508 | |
---|
| 3509 | # COPY HOSTFILE FROM SOURCE DIRECTORY OR CREATE IT, IF IT |
---|
| 3510 | # DOES NOT EXIST |
---|
[1622] | 3511 | if [[ $host != lcbullhh && $host != lccrayb && $host != lccrayf && $host != lccrayh && $host != lckyuh && $host != lckyut ]] |
---|
| 3512 | |
---|
[1] | 3513 | then |
---|
[179] | 3514 | if [[ -f $hostfile ]] |
---|
| 3515 | then |
---|
| 3516 | cp $hostfile hostfile |
---|
[746] | 3517 | (( ii = $numprocs / $threads_per_task )) |
---|
[1598] | 3518 | [[ $ii = 0 ]] && (( ii = 1 )) |
---|
[179] | 3519 | else |
---|
| 3520 | (( ii = 1 )) |
---|
| 3521 | while (( ii <= $numprocs / $threads_per_task )) |
---|
| 3522 | do |
---|
| 3523 | echo $localhost_realname >> hostfile |
---|
| 3524 | (( ii = ii + 1 )) |
---|
| 3525 | done |
---|
[1598] | 3526 | if (( $numprocs / $threads_per_task == 0 )) |
---|
| 3527 | then |
---|
| 3528 | echo $localhost_realname >> hostfile |
---|
| 3529 | fi |
---|
[179] | 3530 | fi |
---|
| 3531 | eval zeile=\"`head -n $ii hostfile`\" |
---|
| 3532 | printf "\n *** running on: $zeile" |
---|
[1] | 3533 | fi |
---|
[179] | 3534 | |
---|
[1] | 3535 | (( ii = $numprocs / $threads_per_task )) |
---|
[1598] | 3536 | [[ $ii = 0 ]] && (( ii = 1 )) |
---|
[1] | 3537 | export OMP_NUM_THREADS=$threads_per_task |
---|
[1289] | 3538 | |
---|
[1] | 3539 | if [[ $threads_per_task != 1 ]] |
---|
| 3540 | then |
---|
[1289] | 3541 | # INCREASE STACK SIZE TO UNLIMITED, BECAUSE OTHERWISE LARGE RUNS |
---|
| 3542 | # MAY ABORT |
---|
[1304] | 3543 | ulimit -s unlimited |
---|
[405] | 3544 | printf "\n threads per task: $threads_per_task stacksize: unlimited" |
---|
[1] | 3545 | fi |
---|
[102] | 3546 | if [[ $run_coupled_model = false ]] |
---|
[1] | 3547 | then |
---|
[291] | 3548 | if [[ "$ocean_file_appendix" = true ]] |
---|
| 3549 | then |
---|
| 3550 | echo "precursor_ocean" > runfile_atmos |
---|
| 3551 | else |
---|
| 3552 | echo "precursor_atmos" > runfile_atmos |
---|
| 3553 | fi |
---|
[102] | 3554 | printf "\n\n" |
---|
[388] | 3555 | |
---|
[1289] | 3556 | if [[ $host = lccrayb || $host = lccrayh ]] |
---|
[1224] | 3557 | then |
---|
| 3558 | echo "--- aprun -n $ii -N $tasks_per_node a.out $ROPTS < runfile_atmos" |
---|
| 3559 | aprun -n $ii -N $tasks_per_node a.out $ROPTS < runfile_atmos |
---|
| 3560 | # aprun -n $ii -ss -r2 -j1 a.out $ROPTS < runfile_atmos |
---|
[1622] | 3561 | |
---|
| 3562 | elif [[ $host = lcbullhh ]] |
---|
| 3563 | then |
---|
| 3564 | export OMPI_MCA_pml=cm |
---|
| 3565 | export OMPI_MCA_mtl=mxm |
---|
| 3566 | export OMPI_MCA_coll=^ghc |
---|
| 3567 | export OMPI_MCA_mtl_mxm_np=0 |
---|
| 3568 | export MXM_RDMA_PORTS=mlx5_0:1 |
---|
| 3569 | export MXM_LOG_LEVEL=ERROR |
---|
| 3570 | export OMP_NUM_THREADS=$threads_per_task |
---|
| 3571 | export KMP_AFFINITY=verbose,granularity=core,compact,1 |
---|
| 3572 | export KMP_STACKSIZE=64m |
---|
| 3573 | |
---|
| 3574 | srun --nodes=$nodes --ntasks-per-node=$tasks_per_node ./a.out |
---|
| 3575 | |
---|
[1197] | 3576 | elif [[ $host = lccrayf ]] |
---|
| 3577 | then |
---|
[1420] | 3578 | aprun -j1 -n $ii -N $tasks_per_node -m ${memory}M a.out $ROPTS < runfile_atmos |
---|
[892] | 3579 | elif [[ $host = lcxe6 || $host = lcxt5m ]] |
---|
[179] | 3580 | then |
---|
[544] | 3581 | aprun -n $ii -N $tasks_per_node a.out $ROPTS < runfile_atmos |
---|
[755] | 3582 | elif [[ $host = lcflow ]] |
---|
| 3583 | then |
---|
[1547] | 3584 | mpirun -np $ii a.out < runfile_atmos $ROPTS |
---|
[1021] | 3585 | elif [[ $host = lcsb ]] |
---|
| 3586 | then |
---|
| 3587 | mpirun_rsh -hostfile $PBS_NODEFILE -np `cat $PBS_NODEFILE | wc -l` a.out < runfile_atmos $ROPTS |
---|
[1040] | 3588 | elif [[ $host = lckiaps ]] |
---|
| 3589 | then |
---|
| 3590 | mpirun -np $ii -f $PBS_NODEFILE a.out < runfile_atmos $ROPTS |
---|
[1090] | 3591 | elif [[ $host = lckyu* ]] |
---|
| 3592 | then |
---|
[1106] | 3593 | mpiexec -n $ii --stdin runfile_atmos ./a.out $ROPTS |
---|
[164] | 3594 | else |
---|
[1106] | 3595 | mpiexec -machinefile hostfile -n $ii a.out < runfile_atmos $ROPTS |
---|
[164] | 3596 | fi |
---|
[1] | 3597 | else |
---|
[164] | 3598 | |
---|
[1289] | 3599 | # CURRENTLY THERE IS NO FULL MPI-2 SUPPORT ON ICE AND XT4 |
---|
[206] | 3600 | (( iia = $numprocs_atmos / $threads_per_task )) |
---|
| 3601 | (( iio = $numprocs_ocean / $threads_per_task )) |
---|
| 3602 | printf "\n coupled run ($iia atmosphere, $iio ocean)" |
---|
| 3603 | printf "\n using $coupled_mode coupling" |
---|
[102] | 3604 | printf "\n\n" |
---|
[164] | 3605 | |
---|
[206] | 3606 | if [[ $coupled_mode = "mpi2" ]] |
---|
[164] | 3607 | then |
---|
[206] | 3608 | echo "atmosphere_to_ocean $iia $iio" > runfile_atmos |
---|
| 3609 | echo "ocean_to_atmosphere $iia $iio" > runfile_ocean |
---|
[197] | 3610 | |
---|
[1289] | 3611 | if [[ $host = lccrayf || $host = lcxe6 || $host = lcxt5m ]] |
---|
[206] | 3612 | then |
---|
[197] | 3613 | |
---|
[206] | 3614 | aprun -n $iia -N $tasks_per_node a.out < runfile_atmos $ROPTS & |
---|
| 3615 | aprun -n $iio -N $tasks_per_node a.out < runfile_ocean $ROPTS & |
---|
| 3616 | |
---|
| 3617 | else |
---|
| 3618 | # WORKAROUND BECAUSE mpiexec WITH -env option IS NOT AVAILABLE ON SOME SYSTEMS |
---|
| 3619 | mpiexec -machinefile hostfile -n $iia a.out $ROPTS < runfile_atmos & |
---|
| 3620 | mpiexec -machinefile hostfile -n $iio a.out $ROPTS < runfile_ocean & |
---|
| 3621 | # mpiexec -machinefile hostfile -n $iia -env coupling_mode atmosphere_to_ocean a.out $ROPTS & |
---|
| 3622 | # mpiexec -machinefile hostfile -n $iio -env coupling_mode ocean_to_atmosphere a.out $ROPTS & |
---|
| 3623 | fi |
---|
| 3624 | wait |
---|
| 3625 | |
---|
[164] | 3626 | else |
---|
| 3627 | |
---|
[206] | 3628 | echo "coupled_run $iia $iio" > runfile_atmos |
---|
| 3629 | |
---|
[1289] | 3630 | if [[ $host = lccrayf || $host = lcxe6 || $host = lcxt5m ]] |
---|
[206] | 3631 | then |
---|
| 3632 | |
---|
| 3633 | aprun -n $ii -N $tasks_per_node a.out < runfile_atmos $ROPTS |
---|
| 3634 | |
---|
[718] | 3635 | elif [[ $host = lck || $host = lckordi ]] |
---|
[367] | 3636 | then |
---|
[1108] | 3637 | |
---|
| 3638 | mpiexec -n $ii ./a.out $ROPTS < runfile_atmos & |
---|
| 3639 | |
---|
| 3640 | elif [[ $host = lckyu* ]] |
---|
| 3641 | then |
---|
| 3642 | |
---|
| 3643 | mpiexec -n $ii --stdin runfile_atmos ./a.out $ROPTS |
---|
| 3644 | |
---|
[1604] | 3645 | elif [[ $host = lcmuk ]] |
---|
| 3646 | then |
---|
| 3647 | |
---|
| 3648 | mpiexec -machinefile hostfile -n $ii a.out < runfile_atmos $ROPTS |
---|
| 3649 | |
---|
[206] | 3650 | fi |
---|
| 3651 | wait |
---|
[164] | 3652 | fi |
---|
[102] | 3653 | |
---|
[206] | 3654 | fi |
---|
[1289] | 3655 | |
---|
[440] | 3656 | elif [[ $host = lckyoto ]] |
---|
| 3657 | then |
---|
| 3658 | set -xv |
---|
| 3659 | export P4_RSHCOMMAND=plesh |
---|
| 3660 | echo " P4_RSHCOMMAND = $P4_RSHCOMMAND" |
---|
| 3661 | if [[ "$ENVIRONMENT" = BATCH ]] |
---|
| 3662 | then |
---|
[505] | 3663 | if [[ "$cond2" = fujitsu ]] |
---|
| 3664 | then |
---|
| 3665 | mpiexec -n $numprocs ./a.out $ROPTS # for fujitsu-compiler |
---|
| 3666 | elif [[ "cond2" = pgi ]] |
---|
| 3667 | then |
---|
| 3668 | mpirun -np $numprocs -machinefile ${QSUB_NODEINF} ./a.out $ROPTS |
---|
| 3669 | else |
---|
| 3670 | mpirun_rsh -np $numprocs -hostfile ${QSUB_NODEINF} MV2_USE_SRQ=0 ./a.out ${ROPTS} || /bin/true |
---|
| 3671 | fi |
---|
[440] | 3672 | else |
---|
[505] | 3673 | if [[ "$cond2" = "" ]] |
---|
| 3674 | then |
---|
| 3675 | mpiruni_rsh -np $numprocs ./a.out $ROPTS # for intel |
---|
| 3676 | else |
---|
| 3677 | mpirun -np $numprocs ./a.out $ROPTS |
---|
| 3678 | fi |
---|
[440] | 3679 | fi |
---|
| 3680 | set +xv |
---|
[1289] | 3681 | |
---|
[1] | 3682 | elif [[ $host = lctit ]] |
---|
| 3683 | then |
---|
[678] | 3684 | export OMP_NUM_THREADS=$threads_per_task |
---|
| 3685 | echo "OMP_NUM_THREADS=$OMP_NUM_THREADS" |
---|
| 3686 | if [[ "$threads_per_task" != 1 ]] |
---|
| 3687 | then |
---|
| 3688 | export MV2_ENABLE_AFFINITY=0 |
---|
| 3689 | fi |
---|
| 3690 | echo "----- PBS_NODEFILE content:" |
---|
| 3691 | cat $PBS_NODEFILE |
---|
| 3692 | echo "-----" |
---|
| 3693 | (( ii = $numprocs / $threads_per_task )) |
---|
| 3694 | echo "mpirun -np $ii -hostfile $PBS_NODEFILE ./a.out $ROPTS" |
---|
| 3695 | mpirun -np $ii -hostfile $PBS_NODEFILE ./a.out $ROPTS |
---|
[1620] | 3696 | |
---|
[1] | 3697 | else |
---|
[102] | 3698 | mpprun -n $numprocs a.out $ROPTS |
---|
[1] | 3699 | fi |
---|
[1620] | 3700 | else |
---|
[102] | 3701 | a.out $ROPTS |
---|
[1] | 3702 | fi |
---|
| 3703 | fi |
---|
[1289] | 3704 | |
---|
| 3705 | if [[ $? != 0 ]] |
---|
[1] | 3706 | then |
---|
| 3707 | |
---|
[1289] | 3708 | # ABORT IN CASE OF RUNTIME ERRORS |
---|
[1] | 3709 | printf "\n +++ runtime error occured" |
---|
| 3710 | locat=execution |
---|
| 3711 | exit |
---|
[1289] | 3712 | |
---|
[1] | 3713 | else |
---|
[260] | 3714 | |
---|
[1289] | 3715 | printf "\n$dashes\n *** execution finished \n" |
---|
| 3716 | |
---|
| 3717 | # STOP THE DVR STREAMING SERVER AND PROCESS THE DVR OUTPUT IN ORDER |
---|
| 3718 | # TO CREAT DVRS- AND HTML-FILES CONTAINING ALL STREAMS |
---|
[260] | 3719 | if [[ "$dvrserver_id" != "" ]] |
---|
| 3720 | then |
---|
| 3721 | kill $dvrserver_id |
---|
| 3722 | printf "\n *** dvr server with id=$dvrserver_id has been stopped" |
---|
[261] | 3723 | |
---|
[1289] | 3724 | # IF THERE IS A DIRECTORY, DATA HAVE BEEN OUTPUT BY THE |
---|
| 3725 | # STREAMING SERVER. OTHERWISE, USER HAS CHOSEN dvrp_output=local |
---|
[262] | 3726 | if [[ -d DATA_DVR ]] |
---|
| 3727 | then |
---|
[261] | 3728 | |
---|
[1289] | 3729 | # ADD THE CURRENT DVR CONFIGURATION FILE TO THE DVR OUTPUT |
---|
| 3730 | # DIRECTORY |
---|
[262] | 3731 | cp .dvrserver.config DATA_DVR |
---|
| 3732 | |
---|
[1289] | 3733 | # PROCESS THE DVR OUTPUT (OPTION -s FOR GENERATING |
---|
| 3734 | # SEQUENCE MODE DATA TOO) |
---|
[262] | 3735 | process_dvr_output -d DATA_DVR -f $fname -s |
---|
| 3736 | |
---|
| 3737 | else |
---|
| 3738 | |
---|
[1289] | 3739 | # PROCESS THE LOCAL OUTPUT |
---|
[262] | 3740 | process_dvr_output -l -d DATA_DVR -f $fname |
---|
| 3741 | |
---|
| 3742 | fi |
---|
| 3743 | |
---|
| 3744 | elif [[ $(echo $package_list | grep -c dvrp_graphics) != 0 ]] |
---|
| 3745 | then |
---|
| 3746 | |
---|
[1289] | 3747 | # PROCESS DVR OUTPUT GENERATD IN LOCAL MODE (dvrp_output=local) |
---|
[262] | 3748 | process_dvr_output -l -d DATA_DVR -f $fname |
---|
| 3749 | |
---|
[260] | 3750 | fi |
---|
[1] | 3751 | fi |
---|
| 3752 | |
---|
| 3753 | |
---|
| 3754 | |
---|
[1289] | 3755 | # CALL OF combine_plot_fields IN ORDER TO MERGE SINGLE FILES WRITTEN |
---|
| 3756 | # BY EACH CORE INTO ONE FILE |
---|
[920] | 3757 | if [[ ! -f ${PALM_BIN}/combine_plot_fields${block}.x ]] |
---|
| 3758 | then |
---|
| 3759 | printf "\n\n\n +++ WARNING: no combine_plot_fields found for given block \"$cond1 $cond2\"" |
---|
| 3760 | printf "\n 2d- and/or 3d-data may be incomplete!" |
---|
| 3761 | printf "\n Run \"mbuild -u -h $localhost\" to generate utilities for this block.\n" |
---|
| 3762 | elif [[ "$combine_plot_fields" == true ]] |
---|
| 3763 | then |
---|
[1270] | 3764 | if [[ $localhost = lccrayh || $localhost = lccrayb ]] |
---|
| 3765 | then |
---|
| 3766 | printf "\n\n\n *** post-processing: now executing \"aprun -n 1 -N 1 combine_plot_fields${block}.x\" ..." |
---|
| 3767 | aprun -n 1 -N 1 combine_plot_fields${block}.x #$ROPTS < runfile_atmos |
---|
| 3768 | else |
---|
| 3769 | printf "\n\n\n *** post-processing: now executing \"combine_plot_fields${block}.x\" ..." |
---|
| 3770 | combine_plot_fields${block}.x |
---|
| 3771 | fi |
---|
[920] | 3772 | else |
---|
[1289] | 3773 | |
---|
| 3774 | # TEMPORARY SOLUTION TO SKIP combine_plot_fields. THIS IS REQUIRED IN CASE OF HUGE AMOUNT OF |
---|
| 3775 | # DATA OUTPUT. TO DO: EXTEND THIS BRANCH BY CREATING A BATCH JOB for combine_plot_fields. |
---|
| 3776 | # ??? koennen wir das streichen ??? |
---|
[920] | 3777 | printf "\n\n\n *** post-processing: skipping combine_plot_fields (-Z option set) ..." |
---|
| 3778 | fi |
---|
[503] | 3779 | |
---|
| 3780 | |
---|
| 3781 | |
---|
[1289] | 3782 | # EXECUTE OUTPUT-COMMANDS GIVEN IN THE CONFIGURATION FILE |
---|
[1] | 3783 | (( i = 0 )) |
---|
| 3784 | while (( i < ioc )) |
---|
| 3785 | do |
---|
| 3786 | (( i = i + 1 )) |
---|
| 3787 | if (( i == 1 )) |
---|
| 3788 | then |
---|
[1289] | 3789 | printf "\n\n *** execution of OUTPUT-commands:\n$dashes" |
---|
[1] | 3790 | fi |
---|
| 3791 | printf "\n >>> ${out_command[$i]}" |
---|
| 3792 | eval ${out_command[$i]} |
---|
| 3793 | if (( i == ioc )) |
---|
| 3794 | then |
---|
[1289] | 3795 | printf "\n$dashes\n" |
---|
[1] | 3796 | fi |
---|
| 3797 | done |
---|
| 3798 | |
---|
| 3799 | |
---|
[1289] | 3800 | # IN TRACE-MODE PRINT CONTENTS OF THE CURRENT (TEMPORARY) WORKING DIRECTORY |
---|
[1] | 3801 | if [[ $do_trace = true ]] |
---|
| 3802 | then |
---|
| 3803 | printf "\n\n" |
---|
| 3804 | ls -al |
---|
| 3805 | fi |
---|
| 3806 | |
---|
| 3807 | |
---|
| 3808 | |
---|
[1289] | 3809 | # COPY LOCAL OUTPUT-FILES TO THEIR PERMANENT DESTINATIONS |
---|
[1] | 3810 | (( i = 0 )) |
---|
| 3811 | while (( i < iout )) |
---|
| 3812 | do |
---|
| 3813 | (( i = i + 1 )) |
---|
| 3814 | if (( i == 1 )) |
---|
| 3815 | then |
---|
[1289] | 3816 | printf "\n\n *** saving OUTPUT-files:\n$dashes" |
---|
[1] | 3817 | fi |
---|
| 3818 | |
---|
[1289] | 3819 | # CHECK FOR SINGLE FILE (SERIAL RUN) OR DIRECTORY (ONE FILE PER CORE FOR PARELLEL EXECUTION) |
---|
[1] | 3820 | files_for_pes=false; filetyp=file |
---|
| 3821 | if [[ "${actionout[$i]}" = pe && -n $numprocs ]] |
---|
| 3822 | then |
---|
| 3823 | files_for_pes=true; filetyp=directory |
---|
| 3824 | actionout[$i]="" |
---|
| 3825 | elif [[ "${actionout[$i]}" = pe && ! -n $numprocs ]] |
---|
| 3826 | then |
---|
| 3827 | actionout[$i]="" |
---|
| 3828 | elif [[ "${actionout[$i]}" = arpe && -n $numprocs ]] |
---|
| 3829 | then |
---|
| 3830 | files_for_pes=true; filetyp=directory |
---|
| 3831 | actionout[$i]="ar" |
---|
| 3832 | elif [[ "${actionout[$i]}" = arpe && ! -n $numprocs ]] |
---|
| 3833 | then |
---|
| 3834 | actionout[$i]="ar" |
---|
| 3835 | elif [[ "${actionout[$i]}" = flpe && -n $numprocs ]] |
---|
| 3836 | then |
---|
| 3837 | files_for_pes=true; filetyp=directory |
---|
| 3838 | actionout[$i]="fl" |
---|
| 3839 | elif [[ "${actionout[$i]}" = flpe && ! -n $numprocs ]] |
---|
| 3840 | then |
---|
| 3841 | actionout[$i]="fl" |
---|
| 3842 | elif [[ "${actionout[$i]}" = trpe && -n $numprocs ]] |
---|
| 3843 | then |
---|
| 3844 | files_for_pes=true; filetyp=directory |
---|
| 3845 | actionout[$i]="tr" |
---|
| 3846 | elif [[ "${actionout[$i]}" = trpe && ! -n $numprocs ]] |
---|
| 3847 | then |
---|
| 3848 | actionout[$i]="tr" |
---|
| 3849 | fi |
---|
| 3850 | |
---|
| 3851 | if [[ ! -f ${localout[$i]} && $files_for_pes = false ]] |
---|
| 3852 | then |
---|
| 3853 | printf "\n +++ temporary OUTPUT-file ${localout[$i]} does not exist\n" |
---|
[260] | 3854 | elif [[ ! -d ${localout[$i]} && $files_for_pes = true ]] |
---|
[1] | 3855 | then |
---|
| 3856 | printf "\n +++ temporary OUTPUT-file ${localout[$i]}/.... does not exist\n" |
---|
| 3857 | else |
---|
| 3858 | |
---|
| 3859 | |
---|
[1289] | 3860 | # COPY VIA SCP TO LOCAL HOST (ALWAYS IN BINARY MODE USING batch_scp option -m) |
---|
| 3861 | # IF TARGET DIRECTORY DOES NOT EXISTS, TRY TO CREATE IT |
---|
[1] | 3862 | if [[ "${actionout[$i]}" = tr ]] |
---|
| 3863 | then |
---|
| 3864 | if [[ $localhost != $fromhost ]] |
---|
| 3865 | then |
---|
| 3866 | if [[ $files_for_pes = false ]] |
---|
| 3867 | then |
---|
| 3868 | cps="" |
---|
| 3869 | cst="" |
---|
| 3870 | else |
---|
| 3871 | cps=-c |
---|
| 3872 | cst="/" |
---|
| 3873 | fi |
---|
[622] | 3874 | if [[ $localhost = nech ]] |
---|
[1] | 3875 | then |
---|
| 3876 | |
---|
[1289] | 3877 | # TRANSFER IN SEPERATE JOB |
---|
| 3878 | # FIRST COPY FILE TO TEMPORY DATA DIRECTORY |
---|
[1] | 3879 | [[ ! -d $tmp_data_catalog/TRANSFER ]] && mkdir -p $tmp_data_catalog/TRANSFER |
---|
| 3880 | file_to_transfer=${fname}_${localout[$i]}_to_transfer_$kennung |
---|
| 3881 | if [[ $files_for_pes = false ]] |
---|
| 3882 | then |
---|
| 3883 | ln -f ${localout[$i]} $tmp_data_catalog/TRANSFER/$file_to_transfer |
---|
| 3884 | else |
---|
| 3885 | mkdir $tmp_data_catalog/TRANSFER/$file_to_transfer |
---|
| 3886 | ln ${localout[$i]}/* $tmp_data_catalog/TRANSFER/$file_to_transfer |
---|
| 3887 | fi |
---|
| 3888 | |
---|
| 3889 | echo "set -x" > transfer_${localout[$i]} |
---|
| 3890 | echo "cd $tmp_data_catalog/TRANSFER" >> transfer_${localout[$i]} |
---|
[82] | 3891 | |
---|
| 3892 | printf "\n >>> OUTPUT: ${localout[$i]}$cst by SCP in seperate job to" |
---|
| 3893 | printf "\n ${pathout[$i]}/${localhost}_${fname}${endout[$i]}$cst" |
---|
| 3894 | printf "\n or higher cycle\n" |
---|
[1468] | 3895 | echo "batch_scp $PORTOPT $cps -b -m -u $return_username $return_address $file_to_transfer \"${pathout[$i]}\" ${localhost}_${fname}${endout[$i]} ${extout[$i]}" >> transfer_${localout[$i]} |
---|
[82] | 3896 | |
---|
[1] | 3897 | echo "[[ \$? = 0 ]] && rm $file_to_transfer" >> transfer_${localout[$i]} |
---|
| 3898 | |
---|
| 3899 | if [[ $localhost = nech ]] |
---|
| 3900 | then |
---|
[1094] | 3901 | subjob -d -c /pf/b/$usern/job_queue -v -q pp -X 0 -m 1000 -t 900 $PORTOPT transfer_${localout[$i]} |
---|
[1] | 3902 | else |
---|
| 3903 | if [[ "$LOGNAME" = b323013 ]] |
---|
| 3904 | then |
---|
[1094] | 3905 | subjob -v -q c1 -X 0 -m 1000 -t 900 -c $job_catalog $PORTOPT transfer_${localout[$i]} |
---|
[1] | 3906 | else |
---|
[1094] | 3907 | subjob -d -v -q c1 -X 0 -m 1000 -t 900 -c $job_catalog $PORTOPT transfer_${localout[$i]} |
---|
[1] | 3908 | fi |
---|
| 3909 | fi |
---|
| 3910 | |
---|
| 3911 | else |
---|
| 3912 | |
---|
[1289] | 3913 | # TRANSFER WITHIN THIS JOB |
---|
[1] | 3914 | transfer_failed=false |
---|
[82] | 3915 | printf "\n >>> OUTPUT: ${localout[$i]}$cst by SCP to" |
---|
| 3916 | printf "\n ${pathout[$i]}/${localhost}_${fname}${endout[$i]}$cst\n" |
---|
[1255] | 3917 | if [[ $localhost = lccrayb ]] |
---|
| 3918 | then |
---|
[1468] | 3919 | ssh $usern@blogin1 ". \\$HOME/.profile; cd $TEMPDIR; batch_scp $PORTOPT $cps -b -m -u $return_username $return_address ${localout[$i]} \"${pathout[$i]}\" ${localhost}_${fname}${endout[$i]} ${extout[$i]}" |
---|
[1274] | 3920 | elif [[ $localhost = lccrayh ]] |
---|
| 3921 | then |
---|
[1468] | 3922 | ssh $usern@hlogin1 ". \\$HOME/.profile; cd $TEMPDIR; batch_scp $PORTOPT $cps -b -m -u $return_username $return_address ${localout[$i]} \"${pathout[$i]}\" ${localhost}_${fname}${endout[$i]} ${extout[$i]}" |
---|
[1620] | 3923 | elif [[ $localhost = lcbullhh ]] |
---|
| 3924 | then |
---|
| 3925 | ssh $usern@mlogin101 ". \\$HOME/.profile; cd $TEMPDIR; batch_scp $PORTOPT $cps -b -m -u $return_username $return_address ${localout[$i]} \"${pathout[$i]}\" ${localhost}_${fname}${endout[$i]} ${extout[$i]}" |
---|
[1468] | 3926 | elif [[ $localhost = lcxe6 ]] |
---|
| 3927 | then |
---|
| 3928 | ssh $usern@hexagon ". \\$HOME/.profile; cd $TEMPDIR; batch_scp $PORTOPT $cps -b -m -u $return_username $return_address ${localout[$i]} \"${pathout[$i]}\" ${localhost}_${fname}${endout[$i]} ${extout[$i]}" |
---|
[1255] | 3929 | else |
---|
[1468] | 3930 | batch_scp $PORTOPT $cps -b -m -u $return_username $return_address ${localout[$i]} "${pathout[$i]}" ${localhost}_${fname}${endout[$i]} ${extout[$i]} |
---|
[1255] | 3931 | fi |
---|
[82] | 3932 | [[ $? != 0 ]] && transfer_failed=true |
---|
[1] | 3933 | |
---|
[1289] | 3934 | # IF TRANSFER FAILED, CREATE BACKUP COPY ON THIS MACHINE |
---|
[1] | 3935 | if [[ $transfer_failed = true ]] |
---|
| 3936 | then |
---|
[1289] | 3937 | printf " +++ transfer failed. Trying to save a copy on this host under:\n" |
---|
[1] | 3938 | printf " ${pathout[$i]}/${localhost}_${fname}${endout[$i]}_$kennung\n" |
---|
| 3939 | |
---|
[1289] | 3940 | # FIRST CHECK, IF DIRECTORY EXISTS, AND CREATE IT, IF NECESSARY |
---|
[1] | 3941 | eval local_catalog=${pathout[$i]} |
---|
| 3942 | if [[ ! -d $local_catalog ]] |
---|
| 3943 | then |
---|
| 3944 | printf " *** local directory does not exist. Trying to create:\n" |
---|
| 3945 | printf " $local_catalog \n" |
---|
| 3946 | mkdir -p $local_catalog |
---|
| 3947 | fi |
---|
| 3948 | eval cp ${localout[$i]} ${pathout[$i]}/${localhost}_${fname}${endout[$i]}_$kennung |
---|
| 3949 | transfer_problems=true |
---|
| 3950 | fi |
---|
| 3951 | fi |
---|
| 3952 | else |
---|
| 3953 | |
---|
[1289] | 3954 | # UNSET actionout. DUE TO THIS SETTING, FILE WILL LATER JUST BE COPIED ON THIS MACHINE |
---|
[1] | 3955 | actionout[$i]="" |
---|
| 3956 | fi |
---|
| 3957 | fi |
---|
| 3958 | |
---|
| 3959 | |
---|
[1289] | 3960 | # APPEND VIA SCP TO LOCAL HOST (ALWAYS IN BINARY MODE USING batch_scp option -m) |
---|
| 3961 | # IF TARGET DIRECTORY DOES NOT EXISTS, TRY TO CREATE IT |
---|
[1] | 3962 | if [[ "${actionout[$i]}" = tra ]] |
---|
| 3963 | then |
---|
| 3964 | if [[ $localhost != $fromhost ]] |
---|
| 3965 | then |
---|
[622] | 3966 | if [[ $localhost = ibmh || $localhost = nech ]] |
---|
[1] | 3967 | then |
---|
| 3968 | |
---|
[1289] | 3969 | # TRANSFER IN SEPERATE JOB |
---|
| 3970 | # FIRST COPY FILE TO TEMPORY DATA DIRECTORY |
---|
[1] | 3971 | [[ ! -d $tmp_data_catalog/TRANSFER ]] && mkdir -p $tmp_data_catalog/TRANSFER |
---|
| 3972 | file_to_transfer=${fname}_${localout[$i]}_to_transfer_$kennung |
---|
| 3973 | ln -f ${localout[$i]} $tmp_data_catalog/TRANSFER/$file_to_transfer |
---|
| 3974 | |
---|
| 3975 | echo "set -x" > transfer_${localout[$i]} |
---|
| 3976 | echo "cd $tmp_data_catalog/TRANSFER" >> transfer_${localout[$i]} |
---|
[82] | 3977 | |
---|
| 3978 | printf "\n >>> OUTPUT: ${localout[$i]} append by SCP in seperate job to" |
---|
| 3979 | printf "\n ${pathout[$i]}/${localhost}_${fname}${endout[$i]}" |
---|
| 3980 | printf "\n or higher cycle\n" |
---|
[1468] | 3981 | echo "batch_scp $PORTOPT -A -b -m -u $return_username $return_address $file_to_transfer \"${pathout[$i]}\" ${localhost}_${fname}${endout[$i]} ${extout[$i]}" >> transfer_${localout[$i]} |
---|
[82] | 3982 | |
---|
[1] | 3983 | echo "[[ \$? = 0 ]] && rm $file_to_transfer" >> transfer_${localout[$i]} |
---|
| 3984 | |
---|
| 3985 | if [[ $localhost = nech ]] |
---|
| 3986 | then |
---|
[1094] | 3987 | subjob -d -c /pf/b/$usern/job_queue -v -q pp -X 0 -m 1000 -t 900 $PORTOPT transfer_${localout[$i]} |
---|
[1] | 3988 | else |
---|
| 3989 | if [[ $LOGNAME = b323013 ]] |
---|
| 3990 | then |
---|
[1094] | 3991 | subjob -v -q c1 -X 0 -m 1000 -t 900 -c $job_catalog $PORTOPT transfer_${localout[$i]} |
---|
[1] | 3992 | else |
---|
[1094] | 3993 | subjob -d -v -q c1 -X 0 -m 1000 -t 900 -c $job_catalog $PORTOPT transfer_${localout[$i]} |
---|
[1] | 3994 | fi |
---|
| 3995 | fi |
---|
| 3996 | |
---|
| 3997 | else |
---|
| 3998 | |
---|
[1289] | 3999 | # TRANSFER WITHIN THIS JOB |
---|
[1] | 4000 | transfer_failed=false |
---|
[82] | 4001 | printf "\n >>> OUTPUT: ${localout[$i]} append by SCP to" |
---|
| 4002 | printf "\n ${pathout[$i]}/${localhost}_${fname}${endout[$i]}\n" |
---|
[1255] | 4003 | if [[ $localhost = lccrayb ]] |
---|
| 4004 | then |
---|
[1468] | 4005 | ssh $usern@blogin1 ". \\$HOME/.profile; cd $TEMPDIR; batch_scp $PORTOPT -A -b -m -u $return_username $return_address ${localout[$i]} \"${pathout[$i]}\" ${localhost}_${fname}${endout[$i]} ${extout[$i]}" |
---|
[1274] | 4006 | elif [[ $localhost = lccrayh ]] |
---|
| 4007 | then |
---|
[1468] | 4008 | ssh $usern@hlogin1 ". \\$HOME/.profile; cd $TEMPDIR; batch_scp $PORTOPT -A -b -m -u $return_username $return_address ${localout[$i]} \"${pathout[$i]}\" ${localhost}_${fname}${endout[$i]} ${extout[$i]}" |
---|
| 4009 | elif [[ $localhost = lcxe6 ]] |
---|
| 4010 | then |
---|
| 4011 | ssh $usern@hexagon ". \\$HOME/.profile; cd $TEMPDIR; batch_scp $PORTOPT -A -b -m -u $return_username $return_address ${localout[$i]} \"${pathout[$i]}\" ${localhost}_${fname}${endout[$i]} ${extout[$i]}" |
---|
[1255] | 4012 | else |
---|
[1468] | 4013 | batch_scp $PORTOPT -A -b -m -u $return_username $return_address ${localout[$i]} "${pathout[$i]}" ${localhost}_${fname}${endout[$i]} ${extout[$i]} |
---|
[1255] | 4014 | fi |
---|
[82] | 4015 | [[ $? != 0 ]] && transfer_failed=true |
---|
[1] | 4016 | |
---|
[1289] | 4017 | # IF TRANSFER FAILED, CREATE BACKUP COPY ON THIS MACHINE |
---|
[1] | 4018 | if [[ $transfer_failed = true ]] |
---|
| 4019 | then |
---|
[1289] | 4020 | printf " +++ transfer failed. Trying to save a copy on this host under:\n" |
---|
[1] | 4021 | printf " ${pathout[$i]}/${localhost}_${fname}${endout[$i]}_$kennung\n" |
---|
| 4022 | |
---|
[1289] | 4023 | # FIRST CHECK, IF DIRECTORY EXISTS, AND CREATE IT, IF NECESSARY |
---|
[1] | 4024 | eval local_catalog=${pathout[$i]} |
---|
| 4025 | if [[ ! -d $local_catalog ]] |
---|
| 4026 | then |
---|
| 4027 | printf " *** local directory does not exist. Trying to create:\n" |
---|
| 4028 | printf " $local_catalog \n" |
---|
| 4029 | mkdir -p $local_catalog |
---|
| 4030 | fi |
---|
| 4031 | |
---|
| 4032 | eval cp ${localout[$i]} ${pathout[$i]}/${localhost}_${fname}${endout[$i]}_$kennung |
---|
| 4033 | transfer_problems=true |
---|
| 4034 | fi |
---|
| 4035 | fi |
---|
| 4036 | else |
---|
| 4037 | |
---|
[1289] | 4038 | # RESET actionout. DUE TO THIS SETTING, FILE WILL LATER JUST BE APPENDED ON THIS MACHINE |
---|
[1] | 4039 | actionout[$i]=a |
---|
| 4040 | fi |
---|
| 4041 | fi |
---|
| 4042 | |
---|
| 4043 | |
---|
[1289] | 4044 | # OUTPUT-FILE FOR A RESTART-RUN. |
---|
| 4045 | # FILE IS MOVED TO THE TEMPORARY DATA DIRECTORY USING link COMMAND. |
---|
| 4046 | # OPTION -f IS USED TO OVERWRITE EXISTING FILES. |
---|
| 4047 | # AS LAST ACTION, ARCHIVING IS INITIATED. |
---|
[1] | 4048 | if [[ "${actionout[$i]}" = fl ]] |
---|
| 4049 | then |
---|
| 4050 | [[ ! -d $tmp_data_catalog ]] && mkdir -p $tmp_data_catalog |
---|
| 4051 | chmod g+rx $tmp_data_catalog |
---|
| 4052 | if [[ $files_for_pes = false ]] |
---|
| 4053 | then |
---|
| 4054 | printf "\n >>> OUTPUT: ${localout[$i]} to" |
---|
| 4055 | printf "\n $tmp_data_catalog/${frelout[$i]} (temporary data catalog)\n" |
---|
| 4056 | ln -f ${localout[$i]} $tmp_data_catalog/${frelout[$i]} |
---|
| 4057 | else |
---|
| 4058 | printf "\n >>> OUTPUT: ${localout[$i]}/.... to" |
---|
| 4059 | printf "\n $tmp_data_catalog/${frelout[$i]} (temporary data catalog)\n" |
---|
| 4060 | mkdir $tmp_data_catalog/${frelout[$i]} |
---|
[149] | 4061 | cd ${localout[$i]} |
---|
| 4062 | for file in $(ls *) |
---|
| 4063 | do |
---|
| 4064 | ln -f $file $tmp_data_catalog/${frelout[$i]} |
---|
| 4065 | done |
---|
| 4066 | cd $TEMPDIR |
---|
[1] | 4067 | fi |
---|
| 4068 | |
---|
| 4069 | |
---|
[1289] | 4070 | # CREATE AND SUBMIT ARCHIVING-JOB |
---|
[1] | 4071 | if [[ $store_on_archive_system = true ]] |
---|
| 4072 | then |
---|
| 4073 | |
---|
| 4074 | if [[ $archive_system = asterix ]] |
---|
| 4075 | then |
---|
| 4076 | echo "cd $tmp_data_catalog" >> archive_${frelout[$i]} |
---|
| 4077 | if [[ $files_for_pes = false ]] |
---|
| 4078 | then |
---|
| 4079 | echo "stageout ${frelout[$i]} > STAGE_OUTPUT${i}_$kennung" >> archive_${frelout[$i]} |
---|
| 4080 | else |
---|
| 4081 | echo "stageout -t ${frelout[$i]} > STAGE_OUTPUT${i}_$kennung" >> archive_${frelout[$i]} |
---|
| 4082 | fi |
---|
| 4083 | echo "cat STAGE_OUTPUT${i}_$kennung" >> archive_${frelout[$i]} |
---|
| 4084 | echo "if [[ \$(grep -c \"st.msg:150\" STAGE_OUTPUT${i}_$kennung) != 0 ]]" >> archive_${frelout[$i]} |
---|
| 4085 | echo "then" >> archive_${frelout[$i]} |
---|
| 4086 | echo " do_stageout=false" >> archive_${frelout[$i]} |
---|
| 4087 | echo "else" >> archive_${frelout[$i]} |
---|
| 4088 | echo " echo \" +++ $filetyp ${frelout[$i]} could not be stored on archive-system\" " >> archive_${frelout[$i]} |
---|
| 4089 | echo " cat /dev/null > ~/job_queue/ARCHIVE_ERROR_$fname" >> archive_${frelout[$i]} |
---|
| 4090 | echo " cat STAGE_OUTPUT${i}_$kennung > ~/job_queue/archive_${frelout[$i]}_error" >> archive_${frelout[$i]} |
---|
| 4091 | echo " echo \" *** $filetyp ${frelout[$i]} will be copied to \$WORK as backup\" " >> archive_${frelout[$i]} |
---|
| 4092 | if [[ $files_for_pes = false ]] |
---|
| 4093 | then |
---|
| 4094 | echo " cp ${frelout[$i]} \$WORK" >> archive_${frelout[$i]} |
---|
| 4095 | else |
---|
| 4096 | echo " cp -r ${frelout[$i]} \$WORK/${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4097 | fi |
---|
| 4098 | echo " echo \" *** $filetyp ${frelout[$i]} saved\" " >> archive_${frelout[$i]} |
---|
| 4099 | echo "fi" >> archive_${frelout[$i]} |
---|
| 4100 | echo "rm STAGE_OUTPUT${i}_$kennung" >> archive_${frelout[$i]} |
---|
| 4101 | elif [[ $archive_system = DMF ]] |
---|
| 4102 | then |
---|
| 4103 | echo "cd $tmp_data_catalog" >> archive_${frelout[$i]} |
---|
| 4104 | if [[ $files_for_pes = false ]] |
---|
| 4105 | then |
---|
| 4106 | printf "\n +++ archiving of single files impossible with $archive_system !\n" |
---|
| 4107 | locat=DMF |
---|
| 4108 | exit |
---|
| 4109 | else |
---|
[1289] | 4110 | # ??? nicht benötigte Archiv-Systeme entfernen ??? |
---|
[1] | 4111 | # FUER RECHNER IN JUELICH. DORT KOENNTEN AUCH |
---|
| 4112 | # EINZELNE DATEIEN GESICHERT WERDEN (SPAETER KORR.) |
---|
| 4113 | echo "rm -rf \$ARCHIVE/${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4114 | echo "cp -r ${frelout[$i]} \$ARCHIVE" >> archive_${frelout[$i]} |
---|
| 4115 | fi |
---|
| 4116 | elif [[ $archive_system = tivoli ]] |
---|
| 4117 | then |
---|
| 4118 | echo "cd $tmp_data_catalog" >> archive_${frelout[$i]} |
---|
| 4119 | if [[ $files_for_pes = false ]] |
---|
| 4120 | then |
---|
[1289] | 4121 | # REMOVE EXISTING OLD FILES FROM THE ARCHIVE |
---|
[1] | 4122 | echo "set -x" >> archive_${frelout[$i]} |
---|
| 4123 | echo "rm -rf \$PERM/${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4124 | echo "cp ${frelout[$i]} \$PERM/${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4125 | else |
---|
| 4126 | |
---|
| 4127 | echo "set -x" >> archive_${frelout[$i]} |
---|
| 4128 | echo "rm -rf \$PERM/${frelout[$i]}/*" >> archive_${frelout[$i]} |
---|
| 4129 | echo "[[ ! -d \$PERM/${frelout[$i]} ]] && mkdir $PERM/${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4130 | cd $tmp_data_catalog |
---|
| 4131 | all_files=`ls -1 ${frelout[$i]}/*` |
---|
| 4132 | cd - > /dev/null |
---|
| 4133 | (( inode = 0 )) |
---|
| 4134 | (( tp1 = tasks_per_node + 1 )) |
---|
| 4135 | while (( inode < nodes )) |
---|
| 4136 | do |
---|
| 4137 | files=`echo $all_files | cut -d" " -f1-$tasks_per_node` |
---|
| 4138 | all_files=`echo $all_files | cut -d" " -f$tp1-` |
---|
[1289] | 4139 | echo "tar cvf \$PERM/${frelout[$i]}/${frelout[$i]}.node_$inode.tar $files" >> archive_${frelout[$i]} |
---|
[1] | 4140 | (( inode = inode + 1 )) |
---|
| 4141 | done |
---|
| 4142 | fi |
---|
| 4143 | elif [[ $archive_system = ut ]] |
---|
| 4144 | then |
---|
| 4145 | echo "cd $tmp_data_catalog" >> archive_${frelout[$i]} |
---|
| 4146 | if [[ $files_for_pes = false ]] |
---|
| 4147 | then |
---|
[1289] | 4148 | # REMOVE EXISTING OLD FILES FROM THE ARCHIVE |
---|
[1] | 4149 | echo "set -x" >> archive_${frelout[$i]} |
---|
| 4150 | echo "rm -rf \$UT/${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4151 | echo "cp ${frelout[$i]} \$UT/${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4152 | else |
---|
| 4153 | |
---|
| 4154 | echo "set -x" >> archive_${frelout[$i]} |
---|
| 4155 | echo "rm -rf \$UT/${frelout[$i]}/*" >> archive_${frelout[$i]} |
---|
| 4156 | echo "[[ ! -d \$UT/${frelout[$i]} ]] && mkdir $UT/${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4157 | cd $tmp_data_catalog |
---|
| 4158 | all_files=`ls -1 ${frelout[$i]}/*` |
---|
| 4159 | cd - > /dev/null |
---|
| 4160 | (( inode = 0 )) |
---|
| 4161 | (( tp1 = tasks_per_node + 1 )) |
---|
| 4162 | while (( inode < nodes )) |
---|
| 4163 | do |
---|
| 4164 | files=`echo $all_files | cut -d" " -f1-$tasks_per_node` |
---|
| 4165 | all_files=`echo $all_files | cut -d" " -f$tp1-` |
---|
| 4166 | echo "tar cvf \$UT/${frelout[$i]}/${frelout[$i]}.node_$inode.tar $files" >> archive_${frelout[$i]} |
---|
| 4167 | (( inode = inode + 1 )) |
---|
| 4168 | done |
---|
| 4169 | fi |
---|
| 4170 | elif [[ $archive_system = none ]] |
---|
| 4171 | then |
---|
| 4172 | printf " +++ archiving on $localhost not available!\n" |
---|
| 4173 | fi |
---|
| 4174 | |
---|
| 4175 | if [[ $archive_system != none ]] |
---|
| 4176 | then |
---|
[1289] | 4177 | if [[ $localhost = nech ]] |
---|
[1] | 4178 | then |
---|
[1094] | 4179 | subjob -d -c /pf/b/$usern/job_queue -v -q pp -X 0 -m 1000 -t 7200 $PORTOPT archive_${frelout[$i]} |
---|
[1] | 4180 | fi |
---|
| 4181 | printf " Archiving of $tmp_data_catalog/${frelout[$i]} initiated (batch job submitted)\n" |
---|
| 4182 | fi |
---|
| 4183 | else |
---|
[399] | 4184 | printf " +++ caution: option -A is switched off. No archiving on $archive_system!\n" |
---|
[1] | 4185 | fi |
---|
| 4186 | |
---|
| 4187 | |
---|
[1289] | 4188 | # CREATE AN EMPTY DIRECTORY IN THE USERS PERMANENT DIRECTORY, |
---|
| 4189 | # IN ORDER TO NOTE THE RESPECTIVE CYCLE NUMBER ON THE ARCHIVE SYSTEM |
---|
| 4190 | # RESTART-JOBS ARE USING THESE EMPTY-DIRECTORIES TO GET THE CURRENT HIGHEST |
---|
| 4191 | # CYCLE NUMBER ON THE ARCHIVE SYSTEM (IN CASE THAT INFORMATIONS ARE DIFFICULT TO |
---|
| 4192 | # TO ACCESS FROM THE ARCHIVE-SYSTEM DIRECTLY)) |
---|
[1] | 4193 | if [[ $files_for_pes = false ]] |
---|
| 4194 | then |
---|
| 4195 | cat /dev/null > ${pathout[$i]} |
---|
| 4196 | else |
---|
| 4197 | mkdir -p ${pathout[$i]} |
---|
| 4198 | fi |
---|
| 4199 | |
---|
| 4200 | fi |
---|
| 4201 | |
---|
| 4202 | |
---|
[1289] | 4203 | # COPY FROM THIS HOST TO THE ARCHIVE-SYSTEM |
---|
| 4204 | # IF ANY ARCHIVING FAILS, AN ERROR-FLAG-FILE IS SET |
---|
| 4205 | # THIS FILE REMAINS TO BE SET, EVEN IF ARCHIVING OF FURTHER FILES IS SUCCESSFULL |
---|
[1] | 4206 | if [[ "${actionout[$i]}" = ar ]] |
---|
| 4207 | then |
---|
| 4208 | if [[ $files_for_pes = false ]] |
---|
| 4209 | then |
---|
| 4210 | printf "\n >>> OUTPUT: ${localout[$i]} to" |
---|
| 4211 | printf "\n ${pathout[$i]}" |
---|
| 4212 | printf "\n File will be copied to archive-system ($archive_system) !\n" |
---|
| 4213 | else |
---|
| 4214 | printf "\n >>> OUTPUT: ${localout[$i]}/_.... to" |
---|
| 4215 | printf "\n ${pathout[$i]}" |
---|
| 4216 | printf "\n Directory will be copied to archive-system ($archive_system) !\n" |
---|
| 4217 | fi |
---|
| 4218 | mv ${localout[$i]} ${frelout[$i]} |
---|
| 4219 | |
---|
| 4220 | file_saved=false |
---|
| 4221 | |
---|
| 4222 | if [[ $archive_system = asterix ]] |
---|
| 4223 | then |
---|
| 4224 | do_stageout=true |
---|
| 4225 | (( stageout_anz = 0 )) |
---|
| 4226 | while [[ $do_stageout = true ]] |
---|
| 4227 | do |
---|
| 4228 | if [[ $files_for_pes = false ]] |
---|
| 4229 | then |
---|
| 4230 | stageout ${frelout[$i]} > STAGE_OUTPUT |
---|
| 4231 | else |
---|
| 4232 | stageout -t ${frelout[$i]} > STAGE_OUTPUT |
---|
| 4233 | fi |
---|
| 4234 | cat STAGE_OUTPUT |
---|
| 4235 | if [[ $(grep -c "st.msg:150" STAGE_OUTPUT) != 0 ]] |
---|
| 4236 | then |
---|
| 4237 | file_saved=true |
---|
| 4238 | do_stageout=false |
---|
| 4239 | else |
---|
| 4240 | if [[ $files_for_pes = false ]] |
---|
| 4241 | then |
---|
| 4242 | printf "\n +++ file ${frelout[$i]} could not be saved on archive-system" |
---|
| 4243 | else |
---|
| 4244 | printf "\n +++ directory ${frelout[$i]} could not be saved on archive-system" |
---|
| 4245 | fi |
---|
| 4246 | (( stageout_anz = stageout_anz + 1 )) |
---|
| 4247 | if (( stageout_anz == 10 )) |
---|
| 4248 | then |
---|
| 4249 | printf "\n +++ stoped after 10 unsuccessful tries!" |
---|
| 4250 | archive_save=false |
---|
| 4251 | do_stageout=false |
---|
| 4252 | else |
---|
| 4253 | printf "\n *** new try to store on archive after 15 min:" |
---|
| 4254 | sleep 900 |
---|
| 4255 | fi |
---|
| 4256 | fi |
---|
| 4257 | done |
---|
| 4258 | elif [[ $archive_system = DMF ]] |
---|
| 4259 | then |
---|
| 4260 | if [[ $files_for_pes = false ]] |
---|
| 4261 | then |
---|
| 4262 | printf "\n +++ archiving of single files impossible on $archive_system!\n" |
---|
| 4263 | locat=DMF |
---|
| 4264 | exit |
---|
| 4265 | else |
---|
| 4266 | rm -rf $ARCHIVE/${frelout[$i]} |
---|
| 4267 | cp -r ${frelout[$i]} $ARCHIVE |
---|
| 4268 | fi |
---|
| 4269 | file_saved=true |
---|
| 4270 | elif [[ $archive_system = tivoli ]] |
---|
| 4271 | then |
---|
[1289] | 4272 | # ARCHIVING ONLY POSSIBLE VIA BATCH-JOB |
---|
| 4273 | # IN THE MEANTIME, FILE IS STORED IN THE TEMPORARY DATA DIRECTORY, |
---|
| 4274 | # BECAUSE MRUN'S CURRENT TEMPORARY WORKING DIRECTORY MAY ALREADY BE DELETED |
---|
| 4275 | # WHEN THE ARCHIVE-JOB IS EXECUTED |
---|
[1] | 4276 | [[ ! -d $tmp_data_catalog ]] && mkdir -p $tmp_data_catalog |
---|
| 4277 | chmod g+rx $tmp_data_catalog |
---|
| 4278 | if [[ $files_for_pes = false ]] |
---|
| 4279 | then |
---|
| 4280 | ln -f ${frelout[$i]} $tmp_data_catalog/${frelout[$i]} |
---|
| 4281 | else |
---|
| 4282 | mkdir $tmp_data_catalog/${frelout[$i]} |
---|
| 4283 | ln -f ${frelout[$i]}/* $tmp_data_catalog/${frelout[$i]} |
---|
| 4284 | fi |
---|
| 4285 | |
---|
[1289] | 4286 | # GENERATE AND SUBMIT BATCH-JOB |
---|
| 4287 | # FILE HAS TO BE DELETED FROM THE TEMPORARY DATA DIRECTORY |
---|
| 4288 | # DELETE OLD EXISTING FILES FROM THE ARCHIVE |
---|
[1] | 4289 | echo "cd $tmp_data_catalog" > archive_${frelout[$i]} |
---|
| 4290 | if [[ $files_for_pes = false ]] |
---|
| 4291 | then |
---|
| 4292 | echo "rm -rf \$PERM/${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4293 | echo "cp ${frelout[$i]} \$PERM/${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4294 | echo "rm -rf ${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4295 | else |
---|
| 4296 | echo "rm -rf \$PERM/${frelout[$i]}.tar" >> archive_${frelout[$i]} |
---|
| 4297 | echo "tar cvf \$PERM/${frelout[$i]}.tar ${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4298 | echo "rm -rf ${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4299 | fi |
---|
| 4300 | |
---|
[1094] | 4301 | subjob -v -d -q cdata -X 0 -m 1000 -t 43200 -c $job_catalog $PORTOPT archive_${frelout[$i]} |
---|
[1] | 4302 | printf " Archiving of $tmp_data_catalog/${frelout[$i]} initiated (batch job submitted)\n" |
---|
| 4303 | file_saved=true |
---|
| 4304 | |
---|
| 4305 | elif [[ $archive_system = ut ]] |
---|
| 4306 | then |
---|
[1289] | 4307 | # ARCHIVING ONLY POSSIBLE VIA BATCH-JOB |
---|
| 4308 | # IN THE MEANTIME, FILE IS STORED IN THE TEMPORARY DATA DIRECTORY, |
---|
| 4309 | # BECAUSE MRUN'S CURRENT TEMPORARY WORKING DIRECTORY MAY ALREADY BE DELETED |
---|
| 4310 | # WHEN THE ARCHIVE-JOB IS EXECUTED |
---|
[1] | 4311 | [[ ! -d $tmp_data_catalog ]] && mkdir -p $tmp_data_catalog |
---|
| 4312 | chmod g+rx $tmp_data_catalog |
---|
| 4313 | if [[ $files_for_pes = false ]] |
---|
| 4314 | then |
---|
| 4315 | ln -f ${frelout[$i]} $tmp_data_catalog/${frelout[$i]} |
---|
| 4316 | else |
---|
| 4317 | mkdir $tmp_data_catalog/${frelout[$i]} |
---|
| 4318 | ln -f ${frelout[$i]}/* $tmp_data_catalog/${frelout[$i]} |
---|
| 4319 | fi |
---|
| 4320 | |
---|
[1289] | 4321 | # GENERATE AND SUBMIT BATCH-JOB |
---|
| 4322 | # FILE HAS TO BE DELETED FROM THE TEMPORARY DATA DIRECTORY |
---|
| 4323 | # DELETE OLD EXISTING FILES FROM THE ARCHIVE |
---|
[1] | 4324 | echo "cd $tmp_data_catalog" > archive_${frelout[$i]} |
---|
| 4325 | if [[ $files_for_pes = false ]] |
---|
| 4326 | then |
---|
| 4327 | echo "rm -rf \$UT/${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4328 | echo "cp ${frelout[$i]} \$UT/${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4329 | echo "rm -rf ${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4330 | else |
---|
| 4331 | echo "rm -rf \$UT/${frelout[$i]}.tar" >> archive_${frelout[$i]} |
---|
| 4332 | echo "tar cvf \$UT/${frelout[$i]}.tar ${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4333 | echo "rm -rf ${frelout[$i]}" >> archive_${frelout[$i]} |
---|
| 4334 | fi |
---|
| 4335 | |
---|
[1094] | 4336 | subjob -v -c /pf/b/$usern/job_queue -d -q pp -X 0 -m 1000 -t 7200 $PORTOPT archive_${frelout[$i]} |
---|
[1] | 4337 | printf " Archiving of $tmp_data_catalog/${frelout[$i]} initiated (batch job submitted)\n" |
---|
| 4338 | file_saved=true |
---|
| 4339 | |
---|
| 4340 | else |
---|
| 4341 | printf "\n +++ archive_system=\"$archive_system\" archiving impossible!" |
---|
| 4342 | archive_save=false |
---|
| 4343 | fi |
---|
[1289] | 4344 | |
---|
| 4345 | # CREATE EMPTY FLAG-FILE OR -DIRECTORY |
---|
| 4346 | # IN ORDER TO NOTE THE RESPECTIVE CYCLE NUMBER ON THE ARCHIVE SYSTEM |
---|
| 4347 | # RESTART-JOBS ARE USING THESE EMPTY-DIRECTORIES TO GET THE CURRENT HIGHEST |
---|
| 4348 | # CYCLE NUMBER ON THE ARCHIVE SYSTEM (IN CASE THAT INFORMATIONS ARE DIFFICULT TO |
---|
| 4349 | # TO ACCESS FROM THE ARCHIVE-SYSTEM DIRECTLY)) |
---|
[1] | 4350 | if [[ $file_saved = true ]] |
---|
| 4351 | then |
---|
| 4352 | if [[ $files_for_pes = false ]] |
---|
| 4353 | then |
---|
| 4354 | cat /dev/null > ${pathout[$i]} |
---|
| 4355 | else |
---|
| 4356 | mkdir -p ${pathout[$i]} |
---|
| 4357 | fi |
---|
| 4358 | fi |
---|
[1289] | 4359 | |
---|
[1] | 4360 | fi |
---|
| 4361 | |
---|
[1289] | 4362 | # APPEND ON THIS MACHINE |
---|
[1] | 4363 | if [[ "${actionout[$i]}" = "a" ]] |
---|
| 4364 | then |
---|
[741] | 4365 | if [[ "${extout[$i]}" != " " && "${extout[$i]}" != "" ]] |
---|
| 4366 | then |
---|
| 4367 | printf "\n >>> OUTPUT: ${localout[$i]} append to ${pathout[$i]}.${extout[$i]}\n" |
---|
| 4368 | cat ${localout[$i]} >> ${pathout[$i]}.${extout[$i]} |
---|
| 4369 | else |
---|
| 4370 | printf "\n >>> OUTPUT: ${localout[$i]} append to ${pathout[$i]}\n" |
---|
| 4371 | cat ${localout[$i]} >> ${pathout[$i]} |
---|
| 4372 | fi |
---|
[1] | 4373 | fi |
---|
| 4374 | |
---|
[1289] | 4375 | # COPY ON THIS MACHINE |
---|
| 4376 | # COPY HAS TO BE USED, BECAUSE MOVE DOES NOT WORK IF FILE-ORIGIN AND TARGET ARE |
---|
| 4377 | # ON DIFFERENT FILE-SYSTEMS |
---|
[1] | 4378 | if [[ "${actionout[$i]}" = "" && $files_for_pes = false ]] |
---|
| 4379 | then |
---|
| 4380 | |
---|
[1289] | 4381 | # COPY IN CASE OF RUNS ON SINGLE CORES |
---|
[5] | 4382 | if [[ "${extout[$i]}" != " " && "${extout[$i]}" != "" ]] |
---|
[1] | 4383 | then |
---|
| 4384 | printf "\n >>> OUTPUT: ${localout[$i]} to ${pathout[$i]}.${extout[$i]}\n" |
---|
[149] | 4385 | if [[ $link_local_output = true ]] |
---|
| 4386 | then |
---|
| 4387 | printf " using ln -f\n" |
---|
| 4388 | ln -f ${localout[$i]} ${pathout[$i]}.${extout[$i]} |
---|
| 4389 | fi |
---|
| 4390 | # If "ln -f" fails of if "$link_local_output = false" do a normal "cp" |
---|
| 4391 | if [[ ! -f "${pathout[$i]}.${extout[$i]}" ]] |
---|
| 4392 | then |
---|
| 4393 | if [[ $link_local_output = true ]] |
---|
| 4394 | then |
---|
| 4395 | printf " ln failed, using cp...\n" |
---|
| 4396 | fi |
---|
| 4397 | cp ${localout[$i]} ${pathout[$i]}.${extout[$i]} |
---|
| 4398 | fi |
---|
[1] | 4399 | else |
---|
| 4400 | printf "\n >>> OUTPUT: ${localout[$i]} to ${pathout[$i]}\n" |
---|
[149] | 4401 | if [[ $link_local_output = true ]] |
---|
| 4402 | then |
---|
| 4403 | printf " using ln -f\n" |
---|
| 4404 | ln -f ${localout[$i]} ${pathout[$i]} |
---|
| 4405 | fi |
---|
| 4406 | # If "ln -f" fails of if "$link_local_output = false" do a normal "cp" |
---|
| 4407 | if [[ ! -f "${pathout[$i]}" ]] |
---|
| 4408 | then |
---|
| 4409 | if [[ $link_local_output = true ]] |
---|
| 4410 | then |
---|
| 4411 | printf " ln failed, using cp...\n" |
---|
| 4412 | fi |
---|
| 4413 | cp ${localout[$i]} ${pathout[$i]} |
---|
| 4414 | fi |
---|
[1] | 4415 | fi |
---|
| 4416 | |
---|
| 4417 | elif [[ "${actionout[$i]}" = "" && $files_for_pes = true ]] |
---|
| 4418 | then |
---|
| 4419 | |
---|
[1289] | 4420 | # FILES FROM THE DIFFERENT CORES ARE MOVED WITH ln-COMMAND TO THE PERMANENT DIRECTORY |
---|
| 4421 | # AS A FIRST STEP, THE PERMANENT DIRECTORY IS CREATED |
---|
[1] | 4422 | printf "\n >>> OUTPUT: ${localout[$i]}/_.... to ${pathout[$i]}\n" |
---|
[149] | 4423 | if [[ $link_local_output = true ]] |
---|
| 4424 | then |
---|
| 4425 | printf " using ln -f\n" |
---|
| 4426 | mkdir ${pathout[$i]} |
---|
| 4427 | cd ${localout[$i]} |
---|
| 4428 | for file in $(ls *) |
---|
| 4429 | do |
---|
| 4430 | ln -f $file ${pathout[$i]} |
---|
| 4431 | done |
---|
| 4432 | cd $TEMPDIR |
---|
| 4433 | fi |
---|
[1289] | 4434 | |
---|
| 4435 | # IF "ln -f" HAS FAILED OR IF "$link_local_output = false" DO A NORMAL COPY "cp -r" |
---|
[149] | 4436 | if [[ ! -f "${pathout[$i]}/_0000" ]] |
---|
| 4437 | then |
---|
| 4438 | if [[ $link_local_output = true ]] |
---|
| 4439 | then |
---|
| 4440 | printf " ln failed for .../_0000, using cp...\n" |
---|
| 4441 | fi |
---|
| 4442 | cp -r ${localout[$i]} ${pathout[$i]} |
---|
| 4443 | fi |
---|
[1] | 4444 | |
---|
| 4445 | fi |
---|
| 4446 | fi |
---|
| 4447 | done |
---|
| 4448 | if (( i != 0 )) |
---|
| 4449 | then |
---|
| 4450 | if [[ $transfer_problems = true ]] |
---|
| 4451 | then |
---|
[1289] | 4452 | printf "\n$dashes\n *** OUTPUT-files saved" |
---|
[1] | 4453 | printf "\n +++ WARNING: some data transfers failed! \n" |
---|
| 4454 | else |
---|
[1289] | 4455 | printf "\n$dashes\n *** all OUTPUT-files saved \n" |
---|
[1] | 4456 | fi |
---|
| 4457 | fi |
---|
| 4458 | |
---|
| 4459 | |
---|
[1289] | 4460 | # IF REQUIRED, START A RESTART-JOB |
---|
| 4461 | # FILE CONTINUE_RUN MUST HAVE BEEN CREATED BY THE EXECUTABLE (PALM) |
---|
[1] | 4462 | if [[ -f CONTINUE_RUN ]] |
---|
| 4463 | then |
---|
| 4464 | |
---|
| 4465 | if [[ $archive_save = true ]] |
---|
| 4466 | then |
---|
| 4467 | |
---|
[1289] | 4468 | # ADD RESTART-OPTIONS TO THE MRUN-CALL (IF THEY ARE NOT USED ALREADY): |
---|
| 4469 | # -C TELLS MRUN THAT IT IS A RESTART-RUN |
---|
| 4470 | # -v SILENT MODE WITHOUT INTERACTIVE QUERIES |
---|
| 4471 | # -n BATCH-MODE (IMPORTANT ONLY IN CASE OF BATCH JOBS ON THE LOCAL MACHINE) |
---|
[1] | 4472 | [[ $(echo $mc | grep -c "\-C") = 0 ]] && mc="$mc -C" |
---|
| 4473 | [[ $(echo $mc | grep -c "\-v") = 0 ]] && mc="$mc -v" |
---|
| 4474 | [[ $(echo $mc | grep -c "\-b") = 0 ]] && mc="$mc -b" |
---|
| 4475 | if [[ $(echo $mc | grep -c "#") != 0 ]] |
---|
| 4476 | then |
---|
| 4477 | mc=`echo $mc | sed 's/#/f/g'` |
---|
| 4478 | fi |
---|
| 4479 | |
---|
| 4480 | |
---|
[1289] | 4481 | # START THE RESTART-JOB |
---|
[1468] | 4482 | printf "\n\n *** initiating restart-run on \"$return_address\" using command:\n" |
---|
[1] | 4483 | echo " $mc" |
---|
[1289] | 4484 | printf "\n$dashes\n" |
---|
[1] | 4485 | if [[ $localhost != $fromhost ]] |
---|
| 4486 | then |
---|
| 4487 | |
---|
[1620] | 4488 | if [[ $localhost = lcbullhh || $localhost = lccrayb || $localhost = lccrayh || $localhost = nech || $localhost = ibmh || $localhost = ibmkisti || $localhost = ibmku || $localhost = ibms || $localhost = lcflow || $localhost = lckyu* || $localhost = lcxe6 ]] |
---|
[1] | 4489 | then |
---|
| 4490 | echo "*** ssh will be used to initiate restart-runs!" |
---|
[1468] | 4491 | echo " return_address=\"$return_address\" " |
---|
[1] | 4492 | echo " return_username=\"$return_username\" " |
---|
[1468] | 4493 | if [[ $(echo $return_address | grep -c "130.75.105") = 1 ]] |
---|
[1] | 4494 | then |
---|
[1190] | 4495 | if [[ $localhost = ibmh ]] |
---|
| 4496 | then |
---|
[1468] | 4497 | ssh $SSH_PORTOPT $usern@136.172.40.15 "ssh $SSH_PORTOPT $return_address -l $return_username \". \\\$HOME/.profile; module load intel-compiler hdf5 netcdf; PATH=\\\$PATH:$LOCAL_MRUN_PATH;export PALM_BIN=$LOCAL_MRUN_PATH;cd $LOCAL_PWD; $mc\" " |
---|
[1620] | 4498 | elif [[ $localhost = lcbullhh ]] |
---|
| 4499 | then |
---|
| 4500 | ssh $SSH_PORTOPT $usern@mlogin101 "ssh $SSH_PORTOPT $return_address -l $return_username \". \\\$HOME/.profile; module load intel-compiler hdf5 netcdf; PATH=\\\$PATH:$LOCAL_MRUN_PATH;export PALM_BIN=$LOCAL_MRUN_PATH;cd $LOCAL_PWD; $mc\" " |
---|
[1255] | 4501 | elif [[ $localhost = lccrayb ]] |
---|
| 4502 | then |
---|
[1468] | 4503 | ssh $usern@blogin1 "ssh $SSH_PORTOPT $return_address -l $return_username \". \\\$HOME/.profile; module load intel-compiler hdf5 netcdf; PATH=\\\$PATH:$LOCAL_MRUN_PATH; export PALM_BIN=$LOCAL_MRUN_PATH;cd $LOCAL_PWD; $mc\" " |
---|
[1274] | 4504 | elif [[ $localhost = lccrayh ]] |
---|
| 4505 | then |
---|
[1468] | 4506 | ssh $usern@hlogin1 "ssh $SSH_PORTOPT $return_address -l $return_username \". \\\$HOME/.profile; module load intel-compiler hdf5 netcdf; PATH=\\\$PATH:$LOCAL_MRUN_PATH; export PALM_BIN=$LOCAL_MRUN_PATH;cd $LOCAL_PWD; $mc\" " |
---|
| 4507 | elif [[ $localhost = lcxe6 ]] |
---|
| 4508 | then |
---|
| 4509 | ssh $usern@hexagon "ssh $SSH_PORTOPT $return_address -l $return_username \". \\\$HOME/.profile; module load intel-compiler hdf5 netcdf; PATH=\\\$PATH:$LOCAL_MRUN_PATH; export PALM_BIN=$LOCAL_MRUN_PATH;cd $LOCAL_PWD; $mc\" " |
---|
[1190] | 4510 | else |
---|
[1468] | 4511 | ssh $SSH_PORTOPT $return_address -l $return_username ". \\\$HOME/.profile; module load intel-compiler hdf5 netcdf; PATH=\$PATH:$LOCAL_MRUN_PATH;export PALM_BIN=$LOCAL_MRUN_PATH;cd $LOCAL_PWD; $mc " |
---|
[1190] | 4512 | fi |
---|
[1] | 4513 | else |
---|
[693] | 4514 | if [[ $localhost = ibmkisti ]] |
---|
| 4515 | then |
---|
[1468] | 4516 | ssh $SSH_PORTOPT $usern@gaiad "ssh $SSH_PORTOPT $return_address -l $return_username \"PATH=\\\$PATH:$LOCAL_MRUN_PATH;export PALM_BIN=$LOCAL_MRUN_PATH;cd $LOCAL_PWD; $mc\" " |
---|
[1255] | 4517 | elif [[ $localhost = lcflow ]] |
---|
| 4518 | then |
---|
[1468] | 4519 | /usr/bin/ssh $SSH_PORTOPT $return_address -l $return_username "PATH=\$PATH:$LOCAL_MRUN_PATH;export PALM_BIN=$LOCAL_MRUN_PATH;cd $LOCAL_PWD; $mc " |
---|
[1255] | 4520 | elif [[ $localhost = lccrayb ]] |
---|
| 4521 | then |
---|
[1468] | 4522 | ssh $usern@blogin1 "ssh $SSH_PORTOPT $return_address -l $return_username \"PATH=\\\$PATH:$LOCAL_MRUN_PATH;export PALM_BIN=$LOCAL_MRUN_PATH;cd $LOCAL_PWD; $mc\" " |
---|
[1274] | 4523 | elif [[ $localhost = lccrayh ]] |
---|
| 4524 | then |
---|
[1468] | 4525 | ssh $usern@hlogin1 "ssh $SSH_PORTOPT $return_address -l $return_username \"PATH=\\\$PATH:$LOCAL_MRUN_PATH;export PALM_BIN=$LOCAL_MRUN_PATH;cd $LOCAL_PWD; $mc\" " |
---|
[693] | 4526 | else |
---|
[1468] | 4527 | ssh $SSH_PORTOPT $return_address -l $return_username "PATH=\$PATH:$LOCAL_MRUN_PATH;export PALM_BIN=$LOCAL_MRUN_PATH;cd $LOCAL_PWD; $mc " |
---|
[693] | 4528 | fi |
---|
[1] | 4529 | fi |
---|
| 4530 | else |
---|
| 4531 | printf "\n +++ no restart mechanism available for host \"$localhost\" " |
---|
| 4532 | locat=restart; exit |
---|
| 4533 | fi |
---|
| 4534 | |
---|
[1289] | 4535 | # WAIT TO ALLOW THE RESTART-JOB TO BE QUEUED, BEFORE THE CURRENT JOB IS FINISHED |
---|
[399] | 4536 | if [[ $queue = special1q ]] |
---|
| 4537 | then |
---|
| 4538 | sleep 120 |
---|
| 4539 | else |
---|
| 4540 | sleep 30 |
---|
| 4541 | fi |
---|
[1] | 4542 | |
---|
| 4543 | else |
---|
| 4544 | |
---|
[1289] | 4545 | # JOBS RUNNING (AND STARTED) ON THE LOCAL MACHINE CAN DIRECTLY CALL MRUN (WITHOUT |
---|
| 4546 | # USING SSH) |
---|
[1] | 4547 | cd $LOCAL_PWD |
---|
[1289] | 4548 | if [[ $localhost = lckyuh ]] |
---|
[108] | 4549 | then |
---|
[1103] | 4550 | printf "\n +++ WARNING: no restart mechanism available for host \"$localhost\" " |
---|
| 4551 | printf "\n please restart job manually using command:\n" |
---|
| 4552 | printf "\n \"$mc\" " |
---|
[108] | 4553 | else |
---|
[1289] | 4554 | eval $mc # THE ' MUST BE EVALUATED |
---|
[108] | 4555 | fi |
---|
[1] | 4556 | cd - > /dev/null |
---|
| 4557 | fi |
---|
[1103] | 4558 | if [[ $localhost = lckyuh ]] |
---|
| 4559 | then |
---|
[1289] | 4560 | printf "\n$dashes\n *** restart-run to be initiated manually\n" |
---|
[1103] | 4561 | else |
---|
[1289] | 4562 | printf "\n$dashes\n *** restart-run initiated \n" |
---|
[1103] | 4563 | fi |
---|
[1] | 4564 | |
---|
| 4565 | |
---|
[1289] | 4566 | # DELETE INPUT-(RESTART)FILES, WHICH HAVE BEEN FETCHED FROM THE TEMPORARY DATA |
---|
| 4567 | # DIRECTORY, BACAUSE THEY ARE NOT REQUIRED BY THE RESTART-JOB. |
---|
| 4568 | # THIS IS DONE IN ORDER TO AVOID EXCEEDING DISC QUOTAS OR DISC SPACE (RESTART-FILES |
---|
| 4569 | # MAY BE VERY HUGE) |
---|
[1] | 4570 | (( i = 0 )) |
---|
| 4571 | while (( i < iin )) |
---|
| 4572 | do |
---|
| 4573 | (( i = i + 1 )) |
---|
| 4574 | if [[ "${got_tmp[$i]}" = true && $keep_data_from_previous_run = false ]] |
---|
| 4575 | then |
---|
| 4576 | rm -r $tmp_data_catalog/${frelin[$i]} |
---|
| 4577 | fi |
---|
| 4578 | done |
---|
| 4579 | |
---|
| 4580 | else |
---|
| 4581 | |
---|
| 4582 | printf "\n +++ no restart-run possible, since errors occured" |
---|
| 4583 | printf "\n during the archive process" |
---|
| 4584 | fi |
---|
| 4585 | |
---|
| 4586 | fi |
---|
| 4587 | |
---|
| 4588 | |
---|
| 4589 | |
---|
[1289] | 4590 | # SEND EMAIL NOTIFICATION ABOUT THE FINISHED RUN |
---|
[352] | 4591 | if [[ "$email_notification" != "none" ]] |
---|
[1] | 4592 | then |
---|
| 4593 | |
---|
| 4594 | if [[ $localhost != $fromhost ]] |
---|
| 4595 | then |
---|
| 4596 | if [[ -f CONTINUE_RUN ]] |
---|
| 4597 | then |
---|
| 4598 | echo "PALM restart run necessary" > email_text |
---|
| 4599 | echo "description header of actual run:" >> email_text |
---|
| 4600 | cat CONTINUE_RUN >> email_text |
---|
| 4601 | echo "mrun-command to restart:" >> email_text |
---|
| 4602 | echo "$mc" >> email_text |
---|
| 4603 | else |
---|
| 4604 | echo "PALM run with base filename \"$fname\" on host \"$localhost\" finished" > email_text |
---|
| 4605 | fi |
---|
| 4606 | mail $email_notification < email_text |
---|
| 4607 | printf "\n *** email notification sent to \"$email_notification\" " |
---|
| 4608 | fi |
---|
| 4609 | fi |
---|
| 4610 | |
---|
| 4611 | |
---|
| 4612 | |
---|
[1289] | 4613 | # ALL ACTIONS FINISHED, TEMPORARY WORKING-DIRECTORY CAN BE DELETED |
---|
[1] | 4614 | cd $HOME |
---|
| 4615 | [[ $delete_temporary_catalog = true ]] && rm -rf $TEMPDIR |
---|
| 4616 | |
---|
| 4617 | else |
---|
| 4618 | |
---|
[1289] | 4619 | # PREPARING ACTIONS, |
---|
| 4620 | # IF A BATCH-JOB IS TO BE GENERATED AND STARTED ON A REMOTE-MACHINE GERECHNET |
---|
[1] | 4621 | |
---|
[1289] | 4622 | # BUILD THE MRUN-COMMAND TO BE CALLED IN THE BATCH-JOB ON THE REMOTE-MACHINE |
---|
[1468] | 4623 | mrun_com="$mrun_script_name -a $afname -c $config_file -d $fname -h $host -H $fromhost -m $memory -t $cpumax -q $queue -R $return_address -U $return_username -u $remote_username" |
---|
[1] | 4624 | [[ "$cpp_opts" != "" ]] && mrun_com=${mrun_com}" -D \"$cpp_opts\"" |
---|
[79] | 4625 | [[ "$global_revision" != "" ]] && mrun_com=${mrun_com}" -G \"$global_revision\"" |
---|
[123] | 4626 | [[ $group_number != none ]] && mrun_com=${mrun_com}" -g $group_number" |
---|
[1] | 4627 | [[ $do_compile = true ]] && mrun_com=${mrun_com}" -s \"$source_list\"" |
---|
| 4628 | [[ "$input_list" != "" ]] && mrun_com=${mrun_com}" -i \"$input_list\"" |
---|
| 4629 | [[ $ignore_archive_error = true ]] && mrun_com=${mrun_com}" -I" |
---|
| 4630 | [[ $keep_data_from_previous_run = true ]] && mrun_com=${mrun_com}" -k" |
---|
| 4631 | [[ "$additional_conditions" != "" ]] && mrun_com=${mrun_com}" -K \"$additional_conditions\"" |
---|
| 4632 | [[ "$output_list" != "" ]] && mrun_com=${mrun_com}" -o \"$output_list\"" |
---|
| 4633 | [[ "$read_from_config" = false ]] && mrun_com=${mrun_com}" -S" |
---|
| 4634 | [[ $do_trace = true ]] && mrun_com=${mrun_com}" -x" |
---|
| 4635 | [[ "$numprocs" != "" ]] && mrun_com=${mrun_com}" -X $numprocs" |
---|
| 4636 | if [[ $use_openmp = true ]] |
---|
| 4637 | then |
---|
[622] | 4638 | mrun_com=${mrun_com}" -O $threads_per_task" |
---|
[1] | 4639 | fi |
---|
[759] | 4640 | [[ "$tasks_per_node" != "" ]] && mrun_com=${mrun_com}" -T $tasks_per_node" |
---|
[397] | 4641 | [[ $store_on_archive_system = true ]] && mrun_com=${mrun_com}" -A" |
---|
[1] | 4642 | [[ $package_list != "" ]] && mrun_com=${mrun_com}" -p \"$package_list\"" |
---|
| 4643 | [[ $return_password != "" ]] && mrun_com=${mrun_com}" -P $return_password" |
---|
| 4644 | [[ $delete_temporary_catalog = false ]] && mrun_com=${mrun_com}" -B" |
---|
[127] | 4645 | [[ $node_usage != default && "$(echo $node_usage | cut -c1-3)" != "sla" && $node_usage != novice ]] && mrun_com=${mrun_com}" -n $node_usage" |
---|
[291] | 4646 | [[ "$ocean_file_appendix" = true ]] && mrun_com=${mrun_com}" -y" |
---|
[206] | 4647 | [[ $run_coupled_model = true ]] && mrun_com=${mrun_com}" -Y \"$coupled_dist\"" |
---|
[920] | 4648 | [[ "$check_namelist_files" = false ]] && mrun_com=${mrun_com}" -z" |
---|
| 4649 | [[ "$combine_plot_fields" = false ]] && mrun_com=${mrun_com}" -Z" |
---|
[759] | 4650 | [[ "$max_par_io_str" != "" ]] && mrun_com=${mrun_com}" -w $max_par_io_str" |
---|
[1] | 4651 | if [[ $do_remote = true ]] |
---|
| 4652 | then |
---|
| 4653 | printf "\n>>>> MRUN-command on execution host:\n>>>> $mrun_com \n" |
---|
| 4654 | fi |
---|
| 4655 | |
---|
| 4656 | |
---|
[1289] | 4657 | # CREATE EXECUTABLE FOR BATCH JOB |
---|
[1199] | 4658 | if [[ $create_executable_for_batch = true && $restart_run != true ]] |
---|
| 4659 | then |
---|
| 4660 | |
---|
| 4661 | printf "\n *** creating the executable for batch job\n" |
---|
| 4662 | |
---|
[1289] | 4663 | # METHOD ONLY WORKS FOR BATCH JOBS ON LOCAL HOSTS |
---|
[1199] | 4664 | if [[ $host != $localhost ]] |
---|
| 4665 | then |
---|
| 4666 | printf "\n +++ creation of executables is only allowed for batch jobs on local hosts." |
---|
| 4667 | printf "\n Please set create_executable_for_batch = false in the config-file.\n" |
---|
| 4668 | locat=create_executable; exit |
---|
| 4669 | fi |
---|
| 4670 | |
---|
| 4671 | mkdir $working_directory/SOURCES_FOR_RUN_${fname}/TMPDIR_FOR_CREATING_EXECUTABLE |
---|
| 4672 | cd $working_directory/SOURCES_FOR_RUN_${fname}/TMPDIR_FOR_CREATING_EXECUTABLE |
---|
| 4673 | |
---|
| 4674 | cp $make_depository . |
---|
| 4675 | tar -xf $make_depository > /dev/null 2>&1 |
---|
| 4676 | cp ../* . > /dev/null 2>&1 |
---|
| 4677 | |
---|
| 4678 | make $mopts -f Makefile PROG=a.out F90=$compiler_name COPT="$cpp_options" F90FLAGS="$fopts" LDFLAGS="$lopts" |
---|
| 4679 | |
---|
| 4680 | if [[ $? != 0 || "$compile_error" = true || "$module_compile_error" = true ]] |
---|
| 4681 | then |
---|
| 4682 | printf "\n +++ error occured while compiling or linking" |
---|
| 4683 | locat=compile |
---|
| 4684 | exit |
---|
| 4685 | fi |
---|
| 4686 | |
---|
| 4687 | mv a.out .. |
---|
| 4688 | cd - > /dev/null 2>&1 |
---|
| 4689 | rm -rf $working_directory/SOURCES_FOR_RUN_${fname}/TMPDIR_FOR_CREATING_EXECUTABLE |
---|
| 4690 | |
---|
| 4691 | printf " *** executable created in \"$working_directory/SOURCES_FOR_RUN_${fname}\"\n " |
---|
| 4692 | |
---|
| 4693 | fi |
---|
| 4694 | |
---|
| 4695 | |
---|
[1289] | 4696 | # BUILD THE JOB-SCRIPTS ON FILE jobfile |
---|
[1] | 4697 | jobfile=jobfile.$RANDOM |
---|
| 4698 | |
---|
| 4699 | |
---|
[1289] | 4700 | # CREATE TEMPORARY DIRECTORY AND SWITCH TO IT |
---|
[1] | 4701 | echo "mkdir $TEMPDIR" >> $jobfile |
---|
| 4702 | echo "cd $TEMPDIR" >> $jobfile |
---|
| 4703 | |
---|
| 4704 | |
---|
[1289] | 4705 | # ACTIVATE ERROR-TRACEBACK |
---|
[1] | 4706 | if [[ $do_trace = true ]] |
---|
| 4707 | then |
---|
| 4708 | echo "set -x" >> $jobfile |
---|
| 4709 | else |
---|
| 4710 | echo "set +vx" >> $jobfile |
---|
| 4711 | fi |
---|
| 4712 | |
---|
| 4713 | |
---|
[1289] | 4714 | # INITIALIZE THE ENVIRONMENT AND LOAD MODULES |
---|
[1043] | 4715 | if [[ "$init_cmds" != "" ]] |
---|
| 4716 | then |
---|
| 4717 | echo "$init_cmds" >> $jobfile |
---|
| 4718 | fi |
---|
| 4719 | if [[ "$module_calls" != "" ]] |
---|
| 4720 | then |
---|
| 4721 | echo "$module_calls" >> $jobfile |
---|
| 4722 | fi |
---|
| 4723 | |
---|
| 4724 | |
---|
[1289] | 4725 | # PROVIDE SOURCE-CODE FILES, MRUN-SCRIPT AND CONFIGURATION-FILE FOR THE JOB |
---|
| 4726 | # then-CLAUSE: FILES ARE COLLECTED IN THE SOURCES_FOR_RUN_... DIRECTORY ON THE LOCAL HOST, |
---|
| 4727 | # WHICH IS THEN FETCHED FROM THE BATCH-JOB BY USING cp/scp |
---|
| 4728 | # THE SOURCE-CODE FILES ALREADY EXIST IN THIS DIRECTORY |
---|
| 4729 | # else-CLAUSE: FILE-CONTENTS ARE PUT INTO THE JOB-FILE AS HERE-DOCUMENTS |
---|
| 4730 | # THIS MAY CREATE A QUITE LARGE JOB-FILE, WHICH CAN CAUSE PROBLEMS WITH SOME |
---|
| 4731 | # QUEUEING-SYSTEMS |
---|
[1620] | 4732 | if [[ $host = ibmkisti || $host = lcbullhh || $host = lccrayb || $host = lccrayf || $host = lccrayh ]] |
---|
[1] | 4733 | then |
---|
| 4734 | |
---|
[1289] | 4735 | # COPY CONFIGURATION-FILE AND MRUN-SCRIPT INTO THE SOURCES_FOR_RUN... DIRECTORY |
---|
[191] | 4736 | if [[ $restart_run != true ]] |
---|
| 4737 | then |
---|
| 4738 | cp $config_file $working_directory/SOURCES_FOR_RUN_$fname |
---|
| 4739 | cp ${PALM_BIN}/$mrun_script_name $working_directory/SOURCES_FOR_RUN_$fname |
---|
| 4740 | fi |
---|
[1] | 4741 | |
---|
[1289] | 4742 | # COPY THE SOURCES_FOR_RUN... DIRECTORY FROM THE LOCAL HOST TO THE JOB VIA scp |
---|
| 4743 | # (then-CLAUSE: JOBS ON THE LOCAL HOST CAN JUST USE cp) |
---|
[193] | 4744 | echo "set -x" >> $jobfile |
---|
[693] | 4745 | if [[ $host = $localhost ]] |
---|
| 4746 | then |
---|
[1199] | 4747 | # DUE TO UNKNOWN REASONS, COPY WITH cp COMMAND CREATES CORRUPT |
---|
| 4748 | # FILES ON CRAY XC30 SYSTEMS (CSC HELSINKI), rsync IS USED INSTEAD |
---|
[1281] | 4749 | if [[ $(echo $host | cut -c1-6) = lccray ]] |
---|
| 4750 | then |
---|
| 4751 | echo "rsync -av -t $working_directory/SOURCES_FOR_RUN_$fname ." >> $jobfile |
---|
| 4752 | else |
---|
| 4753 | echo "cp -r $working_directory/SOURCES_FOR_RUN_$fname ." >> $jobfile |
---|
| 4754 | fi |
---|
[1199] | 4755 | |
---|
[693] | 4756 | else |
---|
[1255] | 4757 | if [[ $host = ibmkisti ]] |
---|
[693] | 4758 | then |
---|
[1289] | 4759 | # ON KISTI'S IBM FIREWALL IS ONLY OPENED ON INTERACTIVE NODE |
---|
[693] | 4760 | echo "localdir=\`pwd\`" >> $jobfile |
---|
[1468] | 4761 | echo "ssh $SSH_PORTOPT $remote_username@gaiad \"cd \$localdir; scp $PORTOPT -r $return_username@$return_address:$working_directory/SOURCES_FOR_RUN_$fname .\" " >> $jobfile |
---|
[1620] | 4762 | elif [[ $host = lcbullhh ]] |
---|
| 4763 | then |
---|
| 4764 | echo "localdir=\`pwd\`" >> $jobfile |
---|
| 4765 | echo "ssh $SSH_PORTOPT $remote_username@mlogin101 \"cd \$localdir; scp $PORTOPT -r $return_username@$return_address:$working_directory/SOURCES_FOR_RUN_$fname .\" " >> $jobfile |
---|
| 4766 | |
---|
[1255] | 4767 | elif [[ $host = lccrayb ]] |
---|
| 4768 | then |
---|
| 4769 | echo "localdir=\`pwd\`" >> $jobfile |
---|
[1468] | 4770 | echo "ssh $SSH_PORTOPT $remote_username@blogin1 \"cd \$localdir; scp $PORTOPT -r $return_username@$return_address:$working_directory/SOURCES_FOR_RUN_$fname .\" " >> $jobfile |
---|
[1274] | 4771 | elif [[ $host = lccrayh ]] |
---|
| 4772 | then |
---|
| 4773 | echo "localdir=\`pwd\`" >> $jobfile |
---|
[1468] | 4774 | echo "ssh $SSH_PORTOPT $remote_username@hlogin1 \"cd \$localdir; scp $PORTOPT -r $return_username@$return_address:$working_directory/SOURCES_FOR_RUN_$fname .\" " >> $jobfile |
---|
[1274] | 4775 | |
---|
[1255] | 4776 | else |
---|
[1468] | 4777 | echo "scp $PORTOPT -r $return_username@$return_address:$working_directory/SOURCES_FOR_RUN_$fname ." >> $jobfile |
---|
[693] | 4778 | fi |
---|
| 4779 | fi |
---|
[193] | 4780 | echo "export SOURCES_COMPLETE=true" >> $jobfile |
---|
[1] | 4781 | |
---|
[1289] | 4782 | # MOVE MRUN-SCRIPT AND CONFIGURATION-FILE FROM THE SOURCES_FOR_RUN... DIRECTORY TO THE |
---|
| 4783 | # WORKING DIRECTORY OF THE JOB |
---|
[193] | 4784 | echo "mv SOURCES_FOR_RUN_$fname/$config_file . " >> $jobfile |
---|
| 4785 | echo "mv SOURCES_FOR_RUN_$fname/$mrun_script_name . " >> $jobfile |
---|
[1199] | 4786 | echo "chmod u+rwx $mrun_script_name" >> $jobfile |
---|
[193] | 4787 | echo "execute_mrun=true" >> $jobfile |
---|
| 4788 | echo " " >> $jobfile |
---|
| 4789 | |
---|
[191] | 4790 | else |
---|
| 4791 | |
---|
[1289] | 4792 | # PROVIDE SOURCE-CODE FILES AND MAKEFILE AS HERE DOCUMENT |
---|
[191] | 4793 | if [[ $do_compile = true ]] |
---|
| 4794 | then |
---|
| 4795 | |
---|
| 4796 | source_catalog=SOURCES_FOR_RUN_$fname |
---|
| 4797 | |
---|
[1289] | 4798 | # CREATE SOURCES_FOR_RUN... DIRECTORY TO STORE THE SOURCE CODE FILES AND THE MAKEFILE |
---|
[191] | 4799 | echo "mkdir SOURCES_FOR_RUN_$fname" >> $jobfile |
---|
| 4800 | echo "export SOURCES_COMPLETE=true" >> $jobfile |
---|
| 4801 | echo "cd SOURCES_FOR_RUN_$fname" >> $jobfile |
---|
| 4802 | |
---|
| 4803 | for filename in $source_list |
---|
| 4804 | do |
---|
[1289] | 4805 | # BACKSLASH IS USED FOR MASKING |
---|
[191] | 4806 | echo "cat > $filename << \"%END%\"" >> $jobfile |
---|
| 4807 | cat $source_catalog/$filename >> $jobfile |
---|
| 4808 | echo " " >> $jobfile |
---|
| 4809 | echo "%END%" >> $jobfile |
---|
| 4810 | echo " " >> $jobfile |
---|
| 4811 | done |
---|
| 4812 | |
---|
[1289] | 4813 | # BACKSLASH IS USED FOR MASKING |
---|
[191] | 4814 | echo "cat > Makefile << \"%END%\"" >> $jobfile |
---|
| 4815 | cat $source_catalog/Makefile >> $jobfile |
---|
[1] | 4816 | echo " " >> $jobfile |
---|
| 4817 | echo "%END%" >> $jobfile |
---|
| 4818 | echo " " >> $jobfile |
---|
| 4819 | |
---|
[191] | 4820 | echo "cd - > /dev/null" >> $jobfile |
---|
[1] | 4821 | |
---|
[191] | 4822 | fi |
---|
[1] | 4823 | |
---|
| 4824 | |
---|
[1289] | 4825 | # PROVIDE THE CONFIGURATION-FILE AS HERE-DOCUMENT |
---|
| 4826 | # BACKSLASH IS USED FOR MASKING |
---|
| 4827 | # LINES WITH #$ IN THE CONFIGURATION-FILE, COMING FROM THE SVN KEYWORD SUBSTITUTION, |
---|
| 4828 | # ARE REMOVED FROM THE FILE IN ORDER TO AVOID PROBLEMS WITH THE SGE BATCH SYSTEM |
---|
[191] | 4829 | echo "cat > $config_file << \"%END%\"" >> $jobfile |
---|
[1101] | 4830 | if [[ $host = lckyuh ]] |
---|
| 4831 | then |
---|
[1289] | 4832 | # NO CROSS-COMPILER ON COMPUTE NODE |
---|
[1101] | 4833 | sed 's/frtpx/frt/g' $config_file >> $jobfile |
---|
| 4834 | else |
---|
| 4835 | sed 's/#$.*//g' $config_file >> $jobfile |
---|
| 4836 | fi |
---|
[191] | 4837 | echo "%END%" >> $jobfile |
---|
| 4838 | echo " " >> $jobfile |
---|
[1] | 4839 | |
---|
| 4840 | |
---|
[1289] | 4841 | # PROVIDE THE MRUN-SCRIPTS AS HERE-DOCUMENT |
---|
| 4842 | # BACKSLASH IS USED FOR MASKING |
---|
[1101] | 4843 | echo "cat > $mrun_script_name <<\"%END%\"" >> $jobfile |
---|
[1103] | 4844 | if [[ $host = lckyuh ]] |
---|
| 4845 | then |
---|
| 4846 | sed 's/\/bin\/ksh/\/bin\/bash/g' ${PALM_BIN}/$mrun_script_name >> $jobfile |
---|
| 4847 | else |
---|
| 4848 | cat ${PALM_BIN}/$mrun_script_name >> $jobfile |
---|
| 4849 | fi |
---|
[1101] | 4850 | echo "%END%" >> $jobfile |
---|
[191] | 4851 | echo "chmod u+x $mrun_script_name" >> $jobfile |
---|
| 4852 | echo "execute_mrun=true" >> $jobfile |
---|
| 4853 | echo " " >> $jobfile |
---|
[1] | 4854 | |
---|
| 4855 | fi |
---|
| 4856 | |
---|
[191] | 4857 | |
---|
[1289] | 4858 | # GET REQUIRED INPUT-FILES BY SCP OR BY SENDING THEM WITH THE JOB AS HERE-DOCUMENT |
---|
| 4859 | # PUT THESE FILES INTO THE USER'S RESPECTIVE PERMANENT DIRECTORIES ON THE REMOTE-HOST |
---|
| 4860 | # IF THE DIRECTORIES DO NOT EXIST, TRY TO CREATE THEM |
---|
[1] | 4861 | if [[ $do_remote = true ]] |
---|
| 4862 | then |
---|
| 4863 | (( i = 0 )) |
---|
| 4864 | while (( i < iin )) |
---|
| 4865 | do |
---|
| 4866 | (( i = i + 1 )) |
---|
| 4867 | echo "[[ ! -d ${pathin[$i]} ]] && mkdir -p ${pathin[$i]}" >> $jobfile |
---|
| 4868 | if [[ "${transin[$i]}" = job ]] |
---|
| 4869 | then |
---|
| 4870 | echo "cat > ${remotepathin[$i]} <<\"%END%\"" >> $jobfile |
---|
| 4871 | eval cat ${pathin[$i]}/${frelin[$i]} >> $jobfile |
---|
| 4872 | echo " " >> $jobfile |
---|
| 4873 | echo "%END%" >> $jobfile |
---|
| 4874 | else |
---|
[1468] | 4875 | echo "batch_scp $PORTOPT -b -o -g -s -u $return_username $return_address ${remotepathin[$i]} \"${pathin[$i]}\" ${frelin[$i]}" >> $jobfile |
---|
[1] | 4876 | fi |
---|
| 4877 | |
---|
[1289] | 4878 | # CHECK, IF FILE COULD BE CREATED |
---|
[1] | 4879 | echo "if [[ \$? = 1 ]]" >> $jobfile |
---|
| 4880 | echo "then" >> $jobfile |
---|
| 4881 | echo " echo \" \" " >> $jobfile |
---|
| 4882 | echo " echo \"+++ file ${remotepathin[$i]} could not be created\" " >> $jobfile |
---|
| 4883 | echo " echo \" please check, if directory exists on $host!\" " >> $jobfile |
---|
| 4884 | echo " echo \"+++ MRUN will not be continued\" " >> $jobfile |
---|
| 4885 | echo " execute_mrun=false" >> $jobfile |
---|
| 4886 | echo "fi" >> $jobfile |
---|
| 4887 | done |
---|
| 4888 | fi |
---|
| 4889 | |
---|
[1289] | 4890 | # PROVIDE NAME OF THE CURRENT WORKING-DIRECTORY ON THE LOCAL MACHINE (FROM WHERE THE JOB IS |
---|
| 4891 | # STARTED) BY SETTING AN ENVIRONMENT-VARIABLE. THIS INFORMATION IS USED IN THE JOB BY MRUN |
---|
| 4892 | # IN CASE THAT RESTART-RUNS HAVE TO BE GENERATED |
---|
[1] | 4893 | echo "LOCAL_PWD=$working_directory" >> $jobfile |
---|
| 4894 | echo "export LOCAL_PWD" >> $jobfile |
---|
| 4895 | |
---|
[1289] | 4896 | # PROVIDE THE PATH OF THE LOCAL MRUN-SCRIPT FOR THE SAME REASON |
---|
[22] | 4897 | echo "LOCAL_MRUN_PATH=$PALM_BIN" >> $jobfile |
---|
[1] | 4898 | echo "export LOCAL_MRUN_PATH" >> $jobfile |
---|
| 4899 | |
---|
[1289] | 4900 | # lcflow ALSO REQUIRES TO PROVIDE PATH FOR THE PALM-SCRIPTS |
---|
[1547] | 4901 | # if [[ $host = lcflow || $localhost = lcflow ]] |
---|
| 4902 | # then |
---|
[1099] | 4903 | echo "export PALM_BIN=$PALM_BIN" | sed -e 's:'$HOME':$HOME:' >> $jobfile |
---|
| 4904 | echo "export PATH=\$PATH:\$PALM_BIN" >> $jobfile |
---|
[1547] | 4905 | # fi |
---|
[1] | 4906 | |
---|
[1289] | 4907 | # CALL MRUN WITHIN THE JOB (SETTING QUEUE IS A WORKAROUND FOR ibmkisti) |
---|
| 4908 | # AS FINAL ACTION, REMOVE THE TEMPORARY DIRECTORY CREATED AT THE BEGINNING OF THE JOB |
---|
[1101] | 4909 | echo "set -x" >> $jobfile |
---|
| 4910 | echo "queue=$queue" >> $jobfile |
---|
[1] | 4911 | echo "[[ \$execute_mrun = true ]] && ./$mrun_com" >> $jobfile |
---|
[1101] | 4912 | echo 'ls -al; echo `pwd`' >> $jobfile |
---|
| 4913 | echo "cd \$HOME" >> $jobfile |
---|
| 4914 | echo "rm -rf $TEMPDIR" >> $jobfile |
---|
[1] | 4915 | |
---|
| 4916 | |
---|
| 4917 | |
---|
| 4918 | |
---|
[1289] | 4919 | # START THE JOB USING SUBJOB-COMMAND |
---|
[1] | 4920 | if [[ $silent = false ]] |
---|
| 4921 | then |
---|
| 4922 | printf "\n " |
---|
| 4923 | else |
---|
| 4924 | printf "\n\n" |
---|
| 4925 | fi |
---|
| 4926 | |
---|
[1094] | 4927 | subjob $job_on_file -h $host -u $remote_username -g $group_number -q $queue -m $memory -N $node_usage -t $cpumax $XOPT $TOPT $OOPT -n $fname -v -c $job_catalog -e $email_notification $PORTOPT $jobfile |
---|
[1] | 4928 | rm -rf $jobfile |
---|
| 4929 | |
---|
| 4930 | |
---|
[1289] | 4931 | fi # END OF REMOTE-PART |
---|