++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Date: Mon, 14 May 2012 11:52:28 +0900
From: Ralf Greve <greve at lowtem.hokudai.ac.jp>
To: SICOPOLIS <sicopolis at ml.hokudai.ac.jp>
Subject: [SICOPOLIS:7] SICOPOLIS and Lis

Dear colleagues,

since I heard about some problems, let me remind you that for the 
shallow shelf approximation solver, the Library of Iterative Solvers for 
Linear Systems (Lis) _version 1.2.53_ is required. You can download the 
tarball at <http://www.ssisc.org/lis/download.html>.

Later versions of Lis (currently 1.2.64) don't work so far! If you have 
an idea why, please let me know... ;-)

Best regards,

Ralf

-- 
Dr. Ralf Greve
Professor, Glacier and Ice Sheet Research
Institute of Low Temperature Science
Hokkaido University
Kita-19, Nishi-8, Kita-ku
Sapporo 060-0819, Japan
Phone/Fax: (+81)-(0)11-706-6891 / -7142
E-Mail:    greve at lowtem.hokudai.ac.jp
Web:       http://wwwice.lowtem.hokudai.ac.jp/~greve/
_______________________________________________
SICOPOLIS mailing list
SICOPOLIS at ml.hokudai.ac.jp
https://ml.hines.hokudai.ac.jp/mailman/listinfo/sicopolis

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Date: Sun, 13 May 2012 16:58:14 -1000
From: Michelle Tigchelaar <mtigch at hawaii.edu>
To: sicopolis at ml.hokudai.ac.jp
Subject: [SICOPOLIS:8] Re: SICOPOLIS and Lis

Dear Ralf,
I just wanted to add to this that for me to get the model to compile on
my system I had to make some small adjustments to the configure.in file.
I've attached my version of the file to this email. I saw that you also
made some recent changes to this file, but I haven't checked to compare.
Hope this helps.
Best,
Michelle

On 5/13/12 4:52 PM, Ralf Greve wrote:
> Dear colleagues,
>
> since I heard about some problems, let me remind you that for the
> shallow shelf approximation solver, the Library of Iterative Solvers for
> Linear Systems (Lis) _version 1.2.53_ is required. You can download the
> tarball at <http://www.ssisc.org/lis/download.html>.
>
> Later versions of Lis (currently 1.2.64) don't work so far! If you have
> an idea why, please let me know... ;-)
>
> Best regards,
>
> Ralf

---------- configure.in.mt ----------

################################################################################
#
#  configure.in
#
#  Author: Malte Thoma, Thomas Goelles
#
################################################################################

# change the version number here
AC_INIT(src/sicopolis.F90, 3.0, sicohelp at greveweb.net )

AM_INIT_AUTOMAKE
AC_CONFIG_HEADER(config.h)
AC_ISC_POSIX
AC_PROG_RANLIB

################################################################################
# Compiler check & options
# storing the original FCFLAGS, this is needed because AC_PROG_FC also sets
# the flags to the default -g -O2
OLD_FCFLAGS=$FCFLAGS

#ignoring the FFLAGS which are for F77 compilers
if test "$FFLAGS" != "" ; then
   AC_MSG_WARN([ your FFLAGS are ignored only using FCFLAGS ])
fi

#ignoring the F90FLAGS
if test "$F90FLAGS" != "" ; then
   AC_MSG_WARN([ your F90FLAGS are ignored only using FCFLAGS ])
fi

AC_MSG_NOTICE(searching a Fortran 90 compiler ... )

AC_PROG_FC(ifort ifc gfortran)
if test "$FC" != "" ; then
   AC_MSG_RESULT([using Fortran Compiler >$FC<])
   F77=$FC         # use the same compiler for f77
else
   AC_MSG_ERROR([
    *** no Fortran Compiler $FC found, must exit 
********************************
    *** please add your compiler in >configure.in< and run autogen.sh 
again. **])
fi

################################################################################
# setting compiler flags for all compilers
################################################################################
FCFLAGS="${OLD_FCFLAGS} -I../../ -I../ " # location of config.h
FCFLAGS="${FCFLAGS} -O3 " # default Optimisation options
MODULE_INCLUDE_PATH_FLAG="-I"
################################################################################
# setting compiler flags
# for specific known compiler
###############################################################################
case $FC in
    ifc|ifort)
       AC_MSG_NOTICE(assuming this is an intel-compiler ... )
        #flags opt for speed
        # -u means implicit none
        # -no-prec-div means a less precise but faster division
        # default line length in free from is 7200 characters
       FCFLAGS="${FCFLAGS} -u -no-prec-div -fpp"
       OPENMPFLAG="-openmp"
       LARGEFILEFLAG="-mcmodel=medium -shared-intel"
       ;;
    gfortran*)
       AC_MSG_NOTICE(assuming this is a gnu-compiler ...)
       FCFLAGS="${FCFLAGS} -ffree-line-length-none -fimplicit-none "
       # NOTE: no gfortran equivalent to -no-prec-div found!
       LARGEFILEFLAG='-mcmodel=medium '
       OPENMPFLAG="-fopenmp"
       AC_DEFINE([OLD_GFORTRAN_COMPILER], [],
        [because of a bug in GNU Fortran 95 (GCC 4.0.2)])
       ;;
    *) AC_MSG_ERROR([
       *** unknown compiler flags for $FC, please add flags to 
configure.in ***
       *** run autogen.sh again. 
*******************************************]) ;;
esac

################################################################################
# adding the mandatory NETCDF library
################################################################################
AC_ARG_WITH([netcdf],
     [ --with-netcdf=<dir>],
     [NCPATHSET=1],
     [NCPATHSET=0])

# if optional parameter used, extend path flags for compiler and linker
if test $NCPATHSET = 1 ; then

    NETCDF_DIR=$withval
    FCFLAGS="${FCFLAGS} $MODULE_INCLUDE_PATH_FLAG$NETCDF_DIR/include"
    LDFLAGS="${LDFLAGS} -L$NETCDF_DIR/lib -lnetcdf "
    #checking the library
    AC_CHECK_FILE($NETCDF_DIR/include/netcdf.mod,[],
                  AC_MSG_ERROR([[coudn't find 
$NETCDF_DIR/include/netcdf.mod,
                     use --with-netcdf=<dir>]]))

    #since netcdf 4.1.1 a seperate netcdf library for fortran is needed
    # additional -lnetcdff flag
    AC_CHECK_FILE($NETCDF_DIR/lib/libnetcdff.a,
                 [LDFLAGS="${LDFLAGS} -lnetcdff"],
                  AC_MSG_NOTICE([[looks like you have an old version of 
netcdf]]))

    AC_CHECK_LIB([netcdf],[main],[],
                 [AC_MSG_ERROR([unable to find netcdf-library,
                    please define --with-netcdf=<dir> $LDFLAGS])],[]) 


   AC_MSG_NOTICE([using the netcdf library])
else
    AC_MSG_WARN([did not include the netcdf Library which is 
recommended])
fi

################################################################################
# adding the optional LIS library
################################################################################
AC_ARG_WITH([LIS],
     [ --with-LIS=<dir>],
     [LISPATHSET=1],
     [LISPATHSET=0])
	
# if optional parameter used, extend path flags for compliler and linker
if test $LISPATHSET = 1 ; then
     AC_MSG_NOTICE([using the LIS library])
     LIS_DIR=$withval
     LDFLAGS="${LDFLAGS} -L$LIS_DIR/lib"
     LIBS="${LIBS} -llis"
     #checking for some file
     AC_CHECK_FILE($LIS_DIR/lib/liblis.a,[],
      AC_MSG_ERROR([[coudn't find $LIS_DIR/lib/liblis.a, use 
--with-LIS=<dir>]]))
     FCFLAGS="${FCFLAGS} $MODULE_INCLUDE_PATH_FLAG$LIS_DIR/include "
     LDFLAGS="${LDFLAGS} -L$LIS_DIR/lib -llis"
fi
################################################################################
# enable openmp
################################################################################
AC_MSG_CHECKING(whether to enable openmp)
openmp_default="no"
AC_ARG_ENABLE(openmp, [  --enable-openmp=[no/yes] to turn on openmp  ],
                        [ enable_openmp=${enableval} ],
                        [ enable_openmp=$openmp_default ])
if test "x$enable_openmp" = "xyes"; then
    FCFLAGS="${FCFLAGS} ${OPENMPFLAG}"
    AC_MSG_RESULT(yes)
else
    AC_MSG_RESULT(no)
fi
################################################################################
# enable Large File support
################################################################################
AC_MSG_CHECKING(whether to enable large_file)
large_file_default="no"
AC_ARG_ENABLE(large_file,
           [  --enable-large_file=[no/yes] to turn on large file support  ],
           [ enable_large_file=${enableval} ],
           [ enable_large_file=$large_file_default ])
if test "x$enable_large_file" = "xyes"; then
    FCFLAGS="${LARGEFILEFLAG} ${FCFLAGS}"
    AC_MSG_RESULT(yes)
else
    AC_MSG_RESULT(no)
fi
################################################################################
FFLAGS=$FCFLAGS

AC_MSG_NOTICE([using this flags: $FCFLAGS ])

#definition of Language and source code extension
AC_LANG([Fortran])

#this sets the variables in all makefiles
AC_SUBST(LDFLAGS)
AC_SUBST(FCFLAGS)
AC_SUBST(FFLAGS)
AC_SUBST(MODULE_INCLUDE_PATH_FLAG)

#if you add a directory please add it here
AC_CONFIG_FILES(
                 src/Makefile \
                 tools/Makefile \
                 tools/sicograph/Makefile \
          )

AC_OUTPUT

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Date: Tue, 15 May 2012 20:25:15 +0900
From: Ralf Greve <greve at lowtem.hokudai.ac.jp>
To: sicopolis at ml.hokudai.ac.jp
Subject: [SICOPOLIS:9] Re: SICOPOLIS and Lis

Dear Michelle,   (cc: SICOPOLIS ML)

On 2012-05-14 11:58, Michelle Tigchelaar wrote:
> Dear Ralf,
> I just wanted to add to this that for me to get the model to compile on
> my system I had to make some small adjustments to the configure.in file.
> I've attached my version of the file to this email. I saw that you also
> made some recent changes to this file, but I haven't checked to compare.
> Hope this helps.

thanks a lot. The change you made in configure.in is actually essential,
otherwise the auto-built SICOPOLIS won't run with Lis. I've just
committed the new version (revision 254).

To everybody, just for clarification: Building SICOPOLIS with the GNU
autotools is just an option (and still somewhat experimental...). Of
course you can do it the traditional way as described in the Quick Start
Manual.

Best regards,

Ralf

-- 
Dr. Ralf Greve
Professor, Glacier and Ice Sheet Research
Institute of Low Temperature Science
Hokkaido University
Kita-19, Nishi-8, Kita-ku
Sapporo 060-0819, Japan
Phone/Fax: (+81)-(0)11-706-6891 / -7142
E-Mail:    greve at lowtem.hokudai.ac.jp
Web:       http://wwwice.lowtem.hokudai.ac.jp/~greve/
_______________________________________________
SICOPOLIS mailing list
SICOPOLIS at ml.hokudai.ac.jp
https://ml.hines.hokudai.ac.jp/mailman/listinfo/sicopolis

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Date: Tue, 15 May 2012 09:50:23 -1000
From: Michelle Tigchelaar <mtigch at hawaii.edu>
To: sicopolis at ml.hokudai.ac.jp
Subject: [SICOPOLIS:10] Re: SICOPOLIS and Lis

Dear Ralf,

It's good to hear my contribution was of some use. I actually don't know
that much about compilers and I remember there being a lot of Google,
trial and error involved in getting SICOPOLIS to run with Lis. Glad it
worked out!

Best,
Michelle

On 5/15/12 1:25 AM, Ralf Greve wrote:
> Dear Michelle,   (cc: SICOPOLIS ML)
>
> On 2012-05-14 11:58, Michelle Tigchelaar wrote:
>> Dear Ralf,
>> I just wanted to add to this that for me to get the model to compile on
>> my system I had to make some small adjustments to the configure.in file.
>> I've attached my version of the file to this email. I saw that you also
>> made some recent changes to this file, but I haven't checked to compare.
>> Hope this helps.
> thanks a lot. The change you made in configure.in is actually essential,
> otherwise the auto-built SICOPOLIS won't run with Lis. I've just
> committed the new version (revision 254).
>
> To everybody, just for clarification: Building SICOPOLIS with the GNU
> autotools is just an option (and still somewhat experimental...). Of
> course you can do it the traditional way as described in the Quick Start
> Manual.
>
> Best regards,
>
> Ralf
>
_______________________________________________
SICOPOLIS mailing list
SICOPOLIS at ml.hokudai.ac.jp
https://ml.hines.hokudai.ac.jp/mailman/listinfo/sicopolis

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Date: Thu, 17 May 2012 09:56:03 +0900
From: Ralf Greve <greve at lowtem.hokudai.ac.jp>
To: SICOPOLIS <sicopolis at ml.hokudai.ac.jp>
Subject: [SICOPOLIS:11] Reply mails

Dear colleagues,

I have changed the setting for reply mails. From now on, by default they 
will go to the sender only, not to the whole mailing list as before. So 
if you want your reply to be seen by everybody, please make sure to add 
a cc to <sicopolis at ml.hokudai.ac.jp>.

Best regards,

Ralf

-- 
Dr. Ralf Greve
Professor, Glacier and Ice Sheet Research
Institute of Low Temperature Science
Hokkaido University
Kita-19, Nishi-8, Kita-ku
Sapporo 060-0819, Japan
Phone/Fax: (+81)-(0)11-706-6891 / -7142
E-Mail:    greve at lowtem.hokudai.ac.jp
Web:       http://wwwice.lowtem.hokudai.ac.jp/~greve/
_______________________________________________
SICOPOLIS mailing list
SICOPOLIS at ml.hokudai.ac.jp
https://ml.hines.hokudai.ac.jp/mailman/listinfo/sicopolis

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++