Changeset 2665 for trunk/NEMOGCM/SETTE

Timestamp:

2011-03-07T15:42:12+01:00 (13 years ago)

Author:

flavoni

Message:

add new functionality in SETTE, creating validation tree, and updated doc, see ticket #752

Location:

trunk/NEMOGCM/SETTE

Files:

• README (modified) (2 diffs)
• SETTE.pdf (modified) (previous)
• fcm_job.sh (modified) (2 diffs)
• sette.sh (modified) (3 diffs)
• sette_beginner.sh (modified) (2 diffs)

trunk/NEMOGCM/SETTE/README

@@ -1 +1 @@
 usage of SETTE package:
 
-principal script is sette.sh (usage: ./sette.sh)
+principal script is sette.sh (usage: ./sette.sh) 
 
 VERY IMPORTANT : for a verbose debug do:
@@ -18 +18 @@
           set_namelist : to set namelist values 
          (example : to set 75 time steps for one test : namelist nn_itend 75)
-         (      : to change cn_icerst_in name : namelist_ice cn_icerst_in  \"test_00101231_restart_icemod.nc\")
+      (      : to change cn_icerst_in name : namelist_ice cn_icerst_in  \"test_00101231_restart_icemod.nc\")
          (      : to activate control print : namelist ln_ctl \".false.\" )
-               fcm_job.sh needs input: input_CONFIG_NAME and TOTAL NUMBER OF PROCS used
-         (for example if you compile ORCA2_LIM with nproci=2 and nprocj=2 ===> ./fcm_job input_ORCA2_LIM.cfg 4)
+               fcm_job.sh needs input: input_CONFIG_NAME TOTAL_NUMBER_OF_PROCS_used TEST_NAME
+         (for example if you compile ORCA2_LIM with nproci=2 and nprocj=2 
+         and only run short test ===> ./fcm_job input_ORCA2_LIM.cfg 4 SHORT)
  in param.cfg :
-               EXE_DIR     : directory in which execute jobs (by default CONFIG_DIR)
-               FORCING_DIR : directory where to get input tar file
-               INPUT_DIR   : directory in which store input files
-               TMPDIR      : temporary directory NEEDED ONLY FOR IBM machines (put EXP00 directory)
-               OUTPUT_DIR  : directory in which store output files
+               EXE_DIR           : directory in which execute jobs (by default CONFIG_DIR)
+               FORCING_DIR       : directory where to get input tar file
+               INPUT_DIR         : directory in which store input files
+               TMPDIR            : temporary directory NEEDED ONLY FOR IBM machines (put EXP00 directory)
+               OUTPUT_DIR        : directory in which store output files
+          NEMO_VALIDATION_DIR : directory in which create NEMO_VALIDATION tree, and store restart and solver.stat files in
+                            tree NEMO_VALIDATION_DIR/WCONFIG_NAME/WCOMPILER_NAME/TEST_NAME/REVISION_NUMBER(or DATE)
  in fcm_job.sh :
                IMPORTANT : control that name of tar file in FORCING directory
                is the same of that in input_REF_CONFIG.cfg, if not change one.
                change in "RUN OPA" part the command line (some examples are already written)
-
+          TEST_NAME iby convention is LONG: 10 days run or SHORT: 5 days run
+          so NEMO_VALIDATION tree is coherent,
+          but user can choise a TEST_NAME that represent his own test.
 
 NOTE:

trunk/NEMOGCM/SETTE/fcm_job.sh

@@ -34 +34 @@
 # Simple job for SET TESTS for NEMO (SETTE)
 #
+# this job creates tree of validation in NEMO_VALIDATION_DIR (choosen in param.cfg) as follows : 
+#
+# NEMO_VALIDATION_DIR/WCONFIG_NAME/WCOMPILER_NAME/TEST_NAME/REVISION_NUMBER(or DATE)
+# 
+# NEMO_VALIDATION_DIR      : is choosen in param.cfg
+#
+# WCONFIG_NAME       : set by makenemo at the moment of compilation
+#
+# WCOMPILER_NAME     : set by makenemo at the moment of compilation
+#
+# TEST_NAME       : convention is LONG=10 days run, SHORT=5 days run
+#
+# REVISION_NUMBER(or DATE) : revision number by svn info, if problems with svn date is taken
 #
 # EXAMPLES
@@ -40 +53 @@
 # ::
 #
-#  $ ./fcm_job.sh INPUT_FILE_CONFIG_NAME       NUMBER_PROC TEST_NAME
-#
-#  run a job of config GYRE with 1 processor          test REPRODUCTINILITY
-#  $ ./fcm_job.sh input_GYRE.cfg                     1     SHORT
-#
-#  run a job of config ORCA2_LIM_PISCES with 8 processors test RESTARTABILITY
-#  $ ./fcm_job.sh input_ORCA2_LIM_PISCES.cfg         8      LONG
+#  $ ./fcm_job.sh INPUT_FILE_CONFIG_NAME NUMBER_PROC TEST_NAME
+#
+#  run a job of config GYRE with 1 processor test REPRODUCTINILITY
+#  $ ./fcm_job.sh input_GYRE.cfg 1 SHORT
+#
+#  run a job of config ORCA2_LIM_PISCES   with 8 processors test RESTARTABILITY
+#  $ ./fcm_job.sh input_ORCA2_LIM_PISCES.cfg 8 LONG
 #
 #

trunk/NEMOGCM/SETTE/sette.sh

@@ -52 +52 @@
 #
 #   Temporary directory (if needed) (TMPDIR)
-#
-#   Output files storing (OUTPUT_DIR)
+# 
+#   Validation directory (NEMO_VALIDATION_DIR)
 #
 #   (note: this file is the same for all configrations to be tested with sette)
 #
-#  all_functions.sh : loads functions used by sette (note: new functions can added here)
+#  all_functions.sh : loads functions used by sette (note: new functions can be added here)
 #
 #  set_namelist : function declared in all_functions that set namelist parameters for tests
@@ -63 +63 @@
 #  fcm_job.sh 
 #
-#   runs job and saves output files (ocean.output, solver.stat, grid_files, restart, ice_evolu, ftrace.out)
+#   runs job and saves output files (grid_files, restarts, ice_evolu, ftrace.out)
+#
+#   and it creates tree of VALIDATION, in which there are restart files, solver.stat and ocean.output
+#
+#   Tree of VALIDATION is:
+#
+#   NEMO_VALIDATION_DIR/WCONFIG_NAME/WCOMPILER_NAME/TEST_NAME/REVISION_NUMBER(or DATE)
 # 
-#   (note this job needs to have an input_CONFIG.cfg in which found tar input file)
+#   (note this job needs to have an input_CONFIG.cfg in which can be found input tar file)
 #
 #   (note other files can be saved adding at the end of fcm_job list of saved files)
@@ -71 +77 @@
 #  NOTE: if job is not launched for some problems you have executable ready in ${CONFIG_NAME}/EXP00 directory
 #
-#  at the end the directory ${CONFIG_NAME}/EXP00/VALIDATION is created
-#
-#  and output files, solver.stat, ocean.output are put in ${CONFIG_NAME}/EXP00/VALIDATION directory
-#
-#  and restart files and changed namelists are leaved in ${CONFIG_NAME}/EXP00 directory
+#  NOTE: the changed namelists are leaved in ${CONFIG_NAME}/EXP00 directory
 # 
 #  in ${SETTE_DIR} is created output.sette with the echo of commands run

trunk/NEMOGCM/SETTE/sette_beginner.sh

@@ -17 +17 @@
 #+
 #
-# ================
-# sette.sh
-# ================
+# ===================
+# sette_beginner.sh
+# ===================
 #
 # ----------------------------------------------
-# Set of tests for NEMO
+# Set of tests for NEMO for beginners
 # ----------------------------------------------
 #
@@ -30 +30 @@
 # ::
 #
-#  $ ./sette.sh
+#  $ ./sette_beginner.sh
 #
 # DESCRIPTION
 # ===========
-SETTE_DIR=/Users/sflod/nuovo_sette/prova/NEMOGCM/SETTE
+# First simple example of how to use SETTE: create GYRE_SHORT configuration, compile it with 1 proc, and test it for a SHORT test: 5days
+#-
+#
+SETTE_DIR=/PATH/OF/SETTE
 # Run for GYRE CONFIG
 # small test to start