Changeset 3294 for trunk/NEMOGCM/SETTE/README
- Timestamp:
- 2012-01-28T17:44:18+01:00 (12 years ago)
- File:
-
- 1 edited
Legend:
- Unmodified
- Added
- Removed
-
trunk/NEMOGCM/SETTE/README
r2757 r3294 11 11 COMPILER : compiler among those available in NEMOGCM/ARCH, needed for makenemo (see below) 12 12 can also be pass as an argument, e.g. ./sette.sh ifort_osx 13 BATCH_COMMAND : command for job submission in batch mode 14 MPI_INTERACT : "yes" if you want to run in interactive mode for MPI run 15 "no" if you want to run in batch mode for MPI run 16 NOTE : every job using only 1 proc is run in interactive mode 17 13 18 makenemo -m : machine (compilator); see ./makenemo -m help (in NEMOGCM/CONFIG directory) 14 19 -n : config name (in examples config_name_nproci_nprocj), see ./makenemo -n help … … 17 22 del_key : to del list of keys 18 23 and for more details and options of makenemo see ./makenemo -h 24 19 25 set_namelist : to set namelist values 20 26 (example : to set 75 time steps for one test : namelist nn_itend 75) 21 27 ( : to change cn_icerst_in name : namelist_ice cn_icerst_in \"test_00101231_restart_icemod\") 22 28 ( : to activate control print : namelist ln_ctl false ) 23 copy_original : to namelist of reference configuration in configuration directory in which tests are done 29 prepare_exe_dir.sh creates execution directory for each test 30 : NOTE for EACH JOB EXECUTION it is needed to set TEST_NAME variable 31 TEST_NAME variable gives name to execution directory (created in prepare_exe_dir.sh) 32 IT is mandatory to have different TEST_NAME for each job run if not existing files are re-written (for example namelist) 33 24 34 fcm_job.sh needs input: input_CONFIG_NAME TOTAL_NUMBER_OF_PROCS_used TEST_NAME 25 35 (for example if you compile ORCA2_LIM with key_mpp_mpi 26 36 and only run short test ===> ./fcm_job input_ORCA2_LIM.cfg 4 SHORT) 37 27 38 in param.cfg : 28 EXE_DIR : directory in which job s executed (by default CONFIG_DIR)29 39 FORCING_DIR : directory in which will be found input.tar file (same name in input_CONFIG_NAME.cfg) 30 INPUT_DIR : directory in which store input files 40 INPUT_DIR : directory in which store input files (tar file) 31 41 TMPDIR : temporary directory NEEDED ONLY FOR IBM machines (put EXP00 directory) 32 OUTPUT_DIR : directory in which store output files 33 NEMO_VALIDATION_DIR : directory in which create NEMO_VALIDATION tree, and store restart, solver.stat and ocean.output files in 42 NEMO_VALIDATION_DIR : directory in which create NEMO_VALIDATION tree, and store restart, solver.stat, tracer.stat and ocean.output files in 34 43 tree NEMO_VALIDATION_DIR/WCONFIG_NAME/WCOMPILER_NAME/TEST_NAME/REVISION_NUMBER(or DATE) 35 44 in fcm_job.sh : 36 45 IMPORTANT : control that name of tar file in FORCING directory 37 46 is the same of that in input_REF_CONFIG.cfg, if not change one. 38 change in "RUN OPA" part the command line (some examples are already written) 39 TEST_NAME by convention is LONG: 10 days run or SHORT: 5 days run 40 so NEMO_VALIDATION tree is coherent, 41 but user can choise a TEST_NAME that represent his own test. 47 48 change in "RUN OPA" if [ ${MPI_INTERACT} == "yes" ] mpirun command line (some examples are already written) 49 NOTE: if MPI_INTERACT="no" it is neede a batch_file (see below to create a new one) 42 50 43 51 NOTE: … … 62 70 1. add arch-compiler.fcm in NEMOGCM/ARCH directory 63 71 2. makenemo -m new_compiler (see makenemo -h for help) 72 73 TO ADD NEW BATCH-FILE: 74 1. see in SETTE/BATCH_TEMPLATE directory existing examples 75 2. create you own batch file like: batch-${COMPILER} 76 (note: sette.sh will copy it in job_batch_template if you run tests in MPI_INTERACT="no")
Note: See TracChangeset
for help on using the changeset viewer.