wolfiex / dsmacc-testing Goto Github PK

add dummy, remove carriage return

sed cmd

47 export KPP_HOME=/work/data/kpp-2.2.3
48 export PATH=$PATH:$KPP_HOME/bin

Program to do this written, just not yet refined to be released.

      return ['do i = %d,%d\n'%(int(jval[0]),int(jval[-1])), 
           'X(1) = X(1) -JVS(i)*X(i)\nend do\n\n',
           'X(1) = X(1)  / (JVS(1))\n']

for linear algebra, check i and not hard coded number

remove approximation to zero

• change begin and ropatool to executables
• mechanism duplicates
• colour code error messages
• check of outputs folder exists before execution
• Makefile - check if ifort present, else use gfortran compiler settings. (something to do with tuv herre too)
• DEFIX variables not read by util by default... fix these (for geoschem)
• subset generator - multiprocess symbolic simplification
• Mech compare plot function in dsmacc library
• Rename sim_data class to something more intuitive.
• PDF in class not actually showing percentage (times by 100) - also floor this
• mechcomp not sorted by name
• ics view- remake- to latex
• continue model from last ics + optional merge of netcdfs.
• checker if model failed, do not make an nc file...
• constrain to observation
• parallel omp of constrain etc.
• rm filter **0*EXP(0/TEMP)) from mcm
• set kpp paths from install/make (export env)
• fitobs general for all cases
• Check if interactive, is so load file / check results print w/h option. (Maybe use pdb trace)
• OBS switch within ICS - triggers which command to run with in mpi script
• MECHLUMP comments too long for kpp
• Check for LON and LAT within range.
• Adjust ICS with local time to match percentage of jday
• testscript to preload modules on install
• run to check if model has been made.

When make kpp is run from the DSMACC top-level directory, the final step fails with errors of the following type:

Fatal error : FILENAME: Can't read file

where FILENAME denotes a data file required. Attached is strace output - kpp seems to look in very particular places for files, and these places are not valid search locations from the top-level DSMACC directory in which kpp is run when compiling.

kpp_strace.txt

Use the DUMMY species as #DEFFIX and remove the dummy equation
{dummy} DUMMY = 0.5DUMMY : 1 ; #

KPP generates f77 and not f90 files for non main mcm mechanisms.
@pb866 any ideas?

make update_submodule should be:
git submodule update --recursive (without --remote that gave a git error message)
git pull did not do anything, if the submodule is up to date. TUV was only downloaded with git submodule update.
This also leaves the possibility for developers to test out the submodules and then release the changes themselves by pulling the changed version and commit changes themselves.

More names than reactions in rates file only.

colour flags

compilation flags

comment out fix issue (see before)
change to eqv

change i to an A