[Met_help] [rt.rap.ucar.edu #69783] History for Issue compiling MET V4.1

Tue Jan 6 11:09:54 MST 2015

----------------------------------------------------------------
  Initial Request
----------------------------------------------------------------

Hello,

I'm trying to compile MET V4.1 using NETCDF V3.6.3 with an intel compiler (definitions and log file attachted). I've gone through all suggestions in the appendix of the users guide but still get an error while trying to make the ensemble_stat executable. Is there a liblm.so or .a required for this build?

Thanks,
Brock

ld: cannot find -lm
make[4]: *** [ensemble_stat] Error 1

Brock Burghardt
Atmospheric Science PhD Student
Dept. of Geosciences
Texas Tech University

----------------------------------------------------------------
  Complete Ticket History
----------------------------------------------------------------

Subject: Issue compiling MET V4.1
From: Julie Prestopnik
Time: Mon Nov 17 15:34:01 2014

Hi Brock.

I see that you are having problems building METv4.1 with an intel
compiler.  Thank you for attaching your user_defs.mk and make.log
files.

First, I wanted to mention that METv5.0 was release in early
September,
which does include bug fixes for METv4.1.  You can view the release
notes
for METv5.0 here:

http://www.dtcenter.org/met/users/support/release_notes/METv5.0_release_notes.php

For now, let's try to overcome this error with METv4.1:

ld: cannot find -lm

We've had this issue come up with some other users in the past.  By
default, MET does static linking - meaning that it links in the
library
files on which it depends at compile time. The other type of linking
is
call dynamic linking - for which the library files on which it depends
are
linked in at runtime. Other users have had success compiling MET with
dynamic linking instead. Open up and edit the top-level user_defs.mk
file,
by commenting out or removing references to "-static". They show up on
the
CXX_FLAGS and FC_FLAGS lines.

Then try recompiling MET, being sure to execute a "make clean" first.

Thanks,
Julie

On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:

>
> Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> Transaction: Ticket created by brock.burghardt at ttu.edu
>        Queue: met_help
>      Subject: Issue compiling MET V4.1
>        Owner: Nobody
>   Requestors: brock.burghardt at ttu.edu
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
>
> Hello,
>
> I'm trying to compile MET V4.1 using NETCDF V3.6.3 with an intel
compiler
> (definitions and log file attachted). I've gone through all
suggestions in
> the appendix of the users guide but still get an error while trying
to make
> the ensemble_stat executable. Is there a liblm.so or .a required for
this
> build?
>
> Thanks,
> Brock
>
> ld: cannot find -lm
> make[4]: *** [ensemble_stat] Error 1
>
>
>
> Brock Burghardt
> Atmospheric Science PhD Student
> Dept. of Geosciences
> Texas Tech University
>
>
>

--
Julie Prestopnik
National Center for Atmospheric Research
Research Applications Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Burghardt, Brock
Time: Tue Nov 18 09:07:05 2014

Hi Julie,

Thanks for getting back to me so soon and explaining. Removing the
static flag did indeed help in compiling the executables I needed but
it did not completely compile (another ld: missing library error came
up in the pb2nc build). However, I took your advice and started over
with V5.0 and got that running, but the compile also failed before
making executables. The error was for not being able to open netcdf.hh
(make_install.log attached). I added all the met_<lib> arguments at
command line and checked they pointed to right directory, and the
configure step ran without error. I verified I do have read permission
for netcdf.hh, any ideas on this.

Thanks,
Brock

-----Original Message-----
From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
Sent: Monday, November 17, 2014 4:34 PM
To: Burghardt, Brock
Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1

Hi Brock.

I see that you are having problems building METv4.1 with an intel
compiler.  Thank you for attaching your user_defs.mk and make.log
files.

First, I wanted to mention that METv5.0 was release in early
September, which does include bug fixes for METv4.1.  You can view the
release notes for METv5.0 here:

http://www.dtcenter.org/met/users/support/release_notes/METv5.0_release_notes.php

For now, let's try to overcome this error with METv4.1:

ld: cannot find -lm

We've had this issue come up with some other users in the past.  By
default, MET does static linking - meaning that it links in the
library files on which it depends at compile time. The other type of
linking is call dynamic linking - for which the library files on which
it depends are linked in at runtime. Other users have had success
compiling MET with dynamic linking instead. Open up and edit the top-
level user_defs.mk file, by commenting out or removing references to
"-static". They show up on the CXX_FLAGS and FC_FLAGS lines.

Then try recompiling MET, being sure to execute a "make clean" first.

Thanks,
Julie

On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:

>
> Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> Transaction: Ticket created by brock.burghardt at ttu.edu
>        Queue: met_help
>      Subject: Issue compiling MET V4.1
>        Owner: Nobody
>   Requestors: brock.burghardt at ttu.edu
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> >
>
>
> Hello,
>
> I'm trying to compile MET V4.1 using NETCDF V3.6.3 with an intel
> compiler (definitions and log file attachted). I've gone through all
> suggestions in the appendix of the users guide but still get an
error
> while trying to make the ensemble_stat executable. Is there a
liblm.so
> or .a required for this build?
>
> Thanks,
> Brock
>
> ld: cannot find -lm
> make[4]: *** [ensemble_stat] Error 1
>
>
>
> Brock Burghardt
> Atmospheric Science PhD Student
> Dept. of Geosciences
> Texas Tech University
>
>
>

--
Julie Prestopnik
National Center for Atmospheric Research Research Applications
Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Julie Prestopnik
Time: Tue Nov 18 09:38:22 2014

Hi Brock.  Thank you for getting the latest version of MET.  I see
you're
still problems with MET finding netcdf.hh.

Could you please run "printenv | grep MET" and send me the output?

Thanks,
Julie

On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi Julie,
>
> Thanks for getting back to me so soon and explaining. Removing the
static
> flag did indeed help in compiling the executables I needed but it
did not
> completely compile (another ld: missing library error came up in the
pb2nc
> build). However, I took your advice and started over with V5.0 and
got that
> running, but the compile also failed before making executables. The
error
> was for not being able to open netcdf.hh (make_install.log
attached). I
> added all the met_<lib> arguments at command line and checked they
pointed
> to right directory, and the configure step ran without error. I
verified I
> do have read permission for netcdf.hh, any ideas on this.
>
> Thanks,
> Brock
>
> -----Original Message-----
> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> Sent: Monday, November 17, 2014 4:34 PM
> To: Burghardt, Brock
> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>
> Hi Brock.
>
> I see that you are having problems building METv4.1 with an intel
> compiler.  Thank you for attaching your user_defs.mk and make.log
files.
>
> First, I wanted to mention that METv5.0 was release in early
September,
> which does include bug fixes for METv4.1.  You can view the release
notes
> for METv5.0 here:
>
>
>
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_release_notes.php
>
> For now, let's try to overcome this error with METv4.1:
>
> ld: cannot find -lm
>
> We've had this issue come up with some other users in the past.  By
> default, MET does static linking - meaning that it links in the
library
> files on which it depends at compile time. The other type of linking
is
> call dynamic linking - for which the library files on which it
depends are
> linked in at runtime. Other users have had success compiling MET
with
> dynamic linking instead. Open up and edit the top-level user_defs.mk
> file, by commenting out or removing references to "-static". They
show up
> on the CXX_FLAGS and FC_FLAGS lines.
>
> Then try recompiling MET, being sure to execute a "make clean"
first.
>
> Thanks,
> Julie
>
> On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT <
> met_help at ucar.edu>
> wrote:
>
> >
> > Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> > Transaction: Ticket created by brock.burghardt at ttu.edu
> >        Queue: met_help
> >      Subject: Issue compiling MET V4.1
> >        Owner: Nobody
> >   Requestors: brock.burghardt at ttu.edu
> >       Status: new
> >  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > >
> >
> >
> > Hello,
> >
> > I'm trying to compile MET V4.1 using NETCDF V3.6.3 with an intel
> > compiler (definitions and log file attachted). I've gone through
all
> > suggestions in the appendix of the users guide but still get an
error
> > while trying to make the ensemble_stat executable. Is there a
liblm.so
> > or .a required for this build?
> >
> > Thanks,
> > Brock
> >
> > ld: cannot find -lm
> > make[4]: *** [ensemble_stat] Error 1
> >
> >
> >
> > Brock Burghardt
> > Atmospheric Science PhD Student
> > Dept. of Geosciences
> > Texas Tech University
> >
> >
> >
>
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research Research Applications
Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>
>
>

--
Julie Prestopnik
National Center for Atmospheric Research
Research Applications Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: RE: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
From: Burghardt, Brock
Time: Tue Nov 18 09:47:34 2014

Hi Julie, the output for the command is,

hrothgar:/met-5.0$ printenv | grep met
PWD=/home/brburgha/met-5.0

...unless you wanted this,

hrothgar:/met-5.0$ printenv | grep MET*
HOSTNAME=hrothgar.hpcc.ttu.edu
ANT_HOME=/opt/rocks
JAVA_HOME=/usr/java/latest
P4_GLOBMEMSIZE=83886080
MAGICK_HOME=/lustre/work/apps/ImageMagick
HOME=/home/brburgha
LOGNAME=brburgha
G_BROKEN_FILENAMES=1

Brock

-----Original Message-----
From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
Sent: Tuesday, November 18, 2014 10:38 AM
To: Burghardt, Brock
Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1

Hi Brock.  Thank you for getting the latest version of MET.  I see
you're still problems with MET finding netcdf.hh.

Could you please run "printenv | grep MET" and send me the output?

Thanks,
Julie

On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi Julie,
>
> Thanks for getting back to me so soon and explaining. Removing the
> static flag did indeed help in compiling the executables I needed
but
> it did not completely compile (another ld: missing library error
came
> up in the pb2nc build). However, I took your advice and started over
> with V5.0 and got that running, but the compile also failed before
> making executables. The error was for not being able to open
netcdf.hh
> (make_install.log attached). I added all the met_<lib> arguments at
> command line and checked they pointed to right directory, and the
> configure step ran without error. I verified I do have read
permission for netcdf.hh, any ideas on this.
>
> Thanks,
> Brock
>
> -----Original Message-----
> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> Sent: Monday, November 17, 2014 4:34 PM
> To: Burghardt, Brock
> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>
> Hi Brock.
>
> I see that you are having problems building METv4.1 with an intel
> compiler.  Thank you for attaching your user_defs.mk and make.log
files.
>
> First, I wanted to mention that METv5.0 was release in early
> September, which does include bug fixes for METv4.1.  You can view
the
> release notes for METv5.0 here:
>
>
>
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_releas
> e_notes.php
>
> For now, let's try to overcome this error with METv4.1:
>
> ld: cannot find -lm
>
> We've had this issue come up with some other users in the past.  By
> default, MET does static linking - meaning that it links in the
> library files on which it depends at compile time. The other type of
> linking is call dynamic linking - for which the library files on
which
> it depends are linked in at runtime. Other users have had success
> compiling MET with dynamic linking instead. Open up and edit the
> top-level user_defs.mk file, by commenting out or removing
references
> to "-static". They show up on the CXX_FLAGS and FC_FLAGS lines.
>
> Then try recompiling MET, being sure to execute a "make clean"
first.
>
> Thanks,
> Julie
>
> On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT <
> met_help at ucar.edu>
> wrote:
>
> >
> > Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> > Transaction: Ticket created by brock.burghardt at ttu.edu
> >        Queue: met_help
> >      Subject: Issue compiling MET V4.1
> >        Owner: Nobody
> >   Requestors: brock.burghardt at ttu.edu
> >       Status: new
> >  Ticket <URL:
> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > >
> >
> >
> > Hello,
> >
> > I'm trying to compile MET V4.1 using NETCDF V3.6.3 with an intel
> > compiler (definitions and log file attachted). I've gone through
all
> > suggestions in the appendix of the users guide but still get an
> > error while trying to make the ensemble_stat executable. Is there
a
> > liblm.so or .a required for this build?
> >
> > Thanks,
> > Brock
> >
> > ld: cannot find -lm
> > make[4]: *** [ensemble_stat] Error 1
> >
> >
> >
> > Brock Burghardt
> > Atmospheric Science PhD Student
> > Dept. of Geosciences
> > Texas Tech University
> >
> >
> >
>
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research Research Applications
> Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>
>
>

--
Julie Prestopnik
National Center for Atmospheric Research Research Applications
Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Julie Prestopnik
Time: Tue Nov 18 10:01:46 2014

Hi Brock.

Thanks you for sending the output.  I was looking to make sure that
the MET
environment variables, like MET_NETCDF or MET_NETCDFINC and
MET_NETCDFLIB,
were set correctly.

Please take a look at the following page and follow the instructions
on
compiling MET:
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/compilation/configure.php

This will help guide you in setting up the required environment
variables
and will hopefully resolve the problem you are experiencing with MET
not
being able to find netcdf.hh.

Please let me know if you have any questions, and please let me know
if you
get a successful build after setting up the environment variables.

Thanks,
Julie

On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi Julie, the output for the command is,
>
> hrothgar:/met-5.0$ printenv | grep met
> PWD=/home/brburgha/met-5.0
>
> ...unless you wanted this,
>
> hrothgar:/met-5.0$ printenv | grep MET*
> HOSTNAME=hrothgar.hpcc.ttu.edu
> ANT_HOME=/opt/rocks
> JAVA_HOME=/usr/java/latest
> P4_GLOBMEMSIZE=83886080
> MAGICK_HOME=/lustre/work/apps/ImageMagick
> HOME=/home/brburgha
> LOGNAME=brburgha
> G_BROKEN_FILENAMES=1
>
>
> Brock
>
>
> -----Original Message-----
> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> Sent: Tuesday, November 18, 2014 10:38 AM
> To: Burghardt, Brock
> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>
> Hi Brock.  Thank you for getting the latest version of MET.  I see
you're
> still problems with MET finding netcdf.hh.
>
> Could you please run "printenv | grep MET" and send me the output?
>
> Thanks,
> Julie
>
> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
> met_help at ucar.edu>
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >
> > Hi Julie,
> >
> > Thanks for getting back to me so soon and explaining. Removing the
> > static flag did indeed help in compiling the executables I needed
but
> > it did not completely compile (another ld: missing library error
came
> > up in the pb2nc build). However, I took your advice and started
over
> > with V5.0 and got that running, but the compile also failed before
> > making executables. The error was for not being able to open
netcdf.hh
> > (make_install.log attached). I added all the met_<lib> arguments
at
> > command line and checked they pointed to right directory, and the
> > configure step ran without error. I verified I do have read
permission
> for netcdf.hh, any ideas on this.
> >
> > Thanks,
> > Brock
> >
> > -----Original Message-----
> > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > Sent: Monday, November 17, 2014 4:34 PM
> > To: Burghardt, Brock
> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >
> > Hi Brock.
> >
> > I see that you are having problems building METv4.1 with an intel
> > compiler.  Thank you for attaching your user_defs.mk and make.log
files.
> >
> > First, I wanted to mention that METv5.0 was release in early
> > September, which does include bug fixes for METv4.1.  You can view
the
> > release notes for METv5.0 here:
> >
> >
> >
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_releas
> > e_notes.php
> >
> > For now, let's try to overcome this error with METv4.1:
> >
> > ld: cannot find -lm
> >
> > We've had this issue come up with some other users in the past.
By
> > default, MET does static linking - meaning that it links in the
> > library files on which it depends at compile time. The other type
of
> > linking is call dynamic linking - for which the library files on
which
> > it depends are linked in at runtime. Other users have had success
> > compiling MET with dynamic linking instead. Open up and edit the
> > top-level user_defs.mk file, by commenting out or removing
references
> > to "-static". They show up on the CXX_FLAGS and FC_FLAGS lines.
> >
> > Then try recompiling MET, being sure to execute a "make clean"
first.
> >
> > Thanks,
> > Julie
> >
> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > >
> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> > > Transaction: Ticket created by brock.burghardt at ttu.edu
> > >        Queue: met_help
> > >      Subject: Issue compiling MET V4.1
> > >        Owner: Nobody
> > >   Requestors: brock.burghardt at ttu.edu
> > >       Status: new
> > >  Ticket <URL:
> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > > >
> > >
> > >
> > > Hello,
> > >
> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3 with an intel
> > > compiler (definitions and log file attachted). I've gone through
all
> > > suggestions in the appendix of the users guide but still get an
> > > error while trying to make the ensemble_stat executable. Is
there a
> > > liblm.so or .a required for this build?
> > >
> > > Thanks,
> > > Brock
> > >
> > > ld: cannot find -lm
> > > make[4]: *** [ensemble_stat] Error 1
> > >
> > >
> > >
> > > Brock Burghardt
> > > Atmospheric Science PhD Student
> > > Dept. of Geosciences
> > > Texas Tech University
> > >
> > >
> > >
> >
> >
> > --
> > Julie Prestopnik
> > National Center for Atmospheric Research Research Applications
> > Laboratory
> > Phone: 303.497.8399
> > Email: jpresto at ucar.edu
> >
> >
> >
>
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research Research Applications
Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>
>
>
>

--
Julie Prestopnik
National Center for Atmospheric Research
Research Applications Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Julie Prestopnik
Time: Mon Dec 01 08:56:18 2014

Hi Brock.  I just wanted to follow up - Were able to get a successful
build
of METv5.0?

Thanks,
Julie

On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik <jpresto at ucar.edu>
wrote:

> Hi Brock.
>
> Thanks you for sending the output.  I was looking to make sure that
the
> MET environment variables, like MET_NETCDF or MET_NETCDFINC and
> MET_NETCDFLIB, were set correctly.
>
> Please take a look at the following page and follow the instructions
on
> compiling MET:
>
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/compilation/configure.php
>
> This will help guide you in setting up the required environment
variables
> and will hopefully resolve the problem you are experiencing with MET
not
> being able to find netcdf.hh.
>
> Please let me know if you have any questions, and please let me know
if
> you get a successful build after setting up the environment
variables.
>
> Thanks,
> Julie
>
> On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> met_help at ucar.edu> wrote:
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>>
>> Hi Julie, the output for the command is,
>>
>> hrothgar:/met-5.0$ printenv | grep met
>> PWD=/home/brburgha/met-5.0
>>
>> ...unless you wanted this,
>>
>> hrothgar:/met-5.0$ printenv | grep MET*
>> HOSTNAME=hrothgar.hpcc.ttu.edu
>> ANT_HOME=/opt/rocks
>> JAVA_HOME=/usr/java/latest
>> P4_GLOBMEMSIZE=83886080
>> MAGICK_HOME=/lustre/work/apps/ImageMagick
>> HOME=/home/brburgha
>> LOGNAME=brburgha
>> G_BROKEN_FILENAMES=1
>>
>>
>> Brock
>>
>>
>> -----Original Message-----
>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
>> Sent: Tuesday, November 18, 2014 10:38 AM
>> To: Burghardt, Brock
>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>>
>> Hi Brock.  Thank you for getting the latest version of MET.  I see
you're
>> still problems with MET finding netcdf.hh.
>>
>> Could you please run "printenv | grep MET" and send me the output?
>>
>> Thanks,
>> Julie
>>
>> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
>> met_help at ucar.edu>
>> wrote:
>>
>> >
>> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>> >
>> > Hi Julie,
>> >
>> > Thanks for getting back to me so soon and explaining. Removing
the
>> > static flag did indeed help in compiling the executables I needed
but
>> > it did not completely compile (another ld: missing library error
came
>> > up in the pb2nc build). However, I took your advice and started
over
>> > with V5.0 and got that running, but the compile also failed
before
>> > making executables. The error was for not being able to open
netcdf.hh
>> > (make_install.log attached). I added all the met_<lib> arguments
at
>> > command line and checked they pointed to right directory, and the
>> > configure step ran without error. I verified I do have read
permission
>> for netcdf.hh, any ideas on this.
>> >
>> > Thanks,
>> > Brock
>> >
>> > -----Original Message-----
>> > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
>> > Sent: Monday, November 17, 2014 4:34 PM
>> > To: Burghardt, Brock
>> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>> >
>> > Hi Brock.
>> >
>> > I see that you are having problems building METv4.1 with an intel
>> > compiler.  Thank you for attaching your user_defs.mk and make.log
>> files.
>> >
>> > First, I wanted to mention that METv5.0 was release in early
>> > September, which does include bug fixes for METv4.1.  You can
view the
>> > release notes for METv5.0 here:
>> >
>> >
>> >
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_releas
>> > e_notes.php
>> >
>> > For now, let's try to overcome this error with METv4.1:
>> >
>> > ld: cannot find -lm
>> >
>> > We've had this issue come up with some other users in the past.
By
>> > default, MET does static linking - meaning that it links in the
>> > library files on which it depends at compile time. The other type
of
>> > linking is call dynamic linking - for which the library files on
which
>> > it depends are linked in at runtime. Other users have had success
>> > compiling MET with dynamic linking instead. Open up and edit the
>> > top-level user_defs.mk file, by commenting out or removing
references
>> > to "-static". They show up on the CXX_FLAGS and FC_FLAGS lines.
>> >
>> > Then try recompiling MET, being sure to execute a "make clean"
first.
>> >
>> > Thanks,
>> > Julie
>> >
>> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT <
>> > met_help at ucar.edu>
>> > wrote:
>> >
>> > >
>> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
>> > > Transaction: Ticket created by brock.burghardt at ttu.edu
>> > >        Queue: met_help
>> > >      Subject: Issue compiling MET V4.1
>> > >        Owner: Nobody
>> > >   Requestors: brock.burghardt at ttu.edu
>> > >       Status: new
>> > >  Ticket <URL:
>> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>> > > >
>> > >
>> > >
>> > > Hello,
>> > >
>> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3 with an
intel
>> > > compiler (definitions and log file attachted). I've gone
through all
>> > > suggestions in the appendix of the users guide but still get an
>> > > error while trying to make the ensemble_stat executable. Is
there a
>> > > liblm.so or .a required for this build?
>> > >
>> > > Thanks,
>> > > Brock
>> > >
>> > > ld: cannot find -lm
>> > > make[4]: *** [ensemble_stat] Error 1
>> > >
>> > >
>> > >
>> > > Brock Burghardt
>> > > Atmospheric Science PhD Student
>> > > Dept. of Geosciences
>> > > Texas Tech University
>> > >
>> > >
>> > >
>> >
>> >
>> > --
>> > Julie Prestopnik
>> > National Center for Atmospheric Research Research Applications
>> > Laboratory
>> > Phone: 303.497.8399
>> > Email: jpresto at ucar.edu
>> >
>> >
>> >
>>
>>
>> --
>> Julie Prestopnik
>> National Center for Atmospheric Research Research Applications
Laboratory
>> Phone: 303.497.8399
>> Email: jpresto at ucar.edu
>>
>>
>>
>>
>
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research
> Research Applications Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>

--
Julie Prestopnik
National Center for Atmospheric Research
Research Applications Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Burghardt, Brock
Time: Tue Dec 02 09:35:53 2014

Hi Julie,

I've still been unable to compile METv5.0...independent of whether I
specify the parent or both the parent/lib parent/inc paths for netcdf,
bufrlib, and GSL, I get an error with it trying to open netcdf.hh

I noticed two paths to netcdf got set in the config.log (attached),
one I set and another on the system, removing the one I did not define
still did not give a successful build.

I noticed I had errors in the config.log for many of the compilers
where if tries to call, for instance, icpc -qversion which should be
icpc -V....I looked through all the config and configure scripts and
could not find the explicit icpc -qversion call to change to -V...do
you think this could be the issue?

Thank you for following up,
Brock

________________________________________
From: Julie Prestopnik via RT [met_help at ucar.edu]
Sent: Monday, December 01, 2014 9:56 AM
To: Burghardt, Brock
Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1

Hi Brock.  I just wanted to follow up - Were able to get a successful
build
of METv5.0?

Thanks,
Julie

On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik <jpresto at ucar.edu>
wrote:

> Hi Brock.
>
> Thanks you for sending the output.  I was looking to make sure that
the
> MET environment variables, like MET_NETCDF or MET_NETCDFINC and
> MET_NETCDFLIB, were set correctly.
>
> Please take a look at the following page and follow the instructions
on
> compiling MET:
>
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/compilation/configure.php
>
> This will help guide you in setting up the required environment
variables
> and will hopefully resolve the problem you are experiencing with MET
not
> being able to find netcdf.hh.
>
> Please let me know if you have any questions, and please let me know
if
> you get a successful build after setting up the environment
variables.
>
> Thanks,
> Julie
>
> On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> met_help at ucar.edu> wrote:
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>>
>> Hi Julie, the output for the command is,
>>
>> hrothgar:/met-5.0$ printenv | grep met
>> PWD=/home/brburgha/met-5.0
>>
>> ...unless you wanted this,
>>
>> hrothgar:/met-5.0$ printenv | grep MET*
>> HOSTNAME=hrothgar.hpcc.ttu.edu
>> ANT_HOME=/opt/rocks
>> JAVA_HOME=/usr/java/latest
>> P4_GLOBMEMSIZE=83886080
>> MAGICK_HOME=/lustre/work/apps/ImageMagick
>> HOME=/home/brburgha
>> LOGNAME=brburgha
>> G_BROKEN_FILENAMES=1
>>
>>
>> Brock
>>
>>
>> -----Original Message-----
>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
>> Sent: Tuesday, November 18, 2014 10:38 AM
>> To: Burghardt, Brock
>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>>
>> Hi Brock.  Thank you for getting the latest version of MET.  I see
you're
>> still problems with MET finding netcdf.hh.
>>
>> Could you please run "printenv | grep MET" and send me the output?
>>
>> Thanks,
>> Julie
>>
>> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
>> met_help at ucar.edu>
>> wrote:
>>
>> >
>> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>> >
>> > Hi Julie,
>> >
>> > Thanks for getting back to me so soon and explaining. Removing
the
>> > static flag did indeed help in compiling the executables I needed
but
>> > it did not completely compile (another ld: missing library error
came
>> > up in the pb2nc build). However, I took your advice and started
over
>> > with V5.0 and got that running, but the compile also failed
before
>> > making executables. The error was for not being able to open
netcdf.hh
>> > (make_install.log attached). I added all the met_<lib> arguments
at
>> > command line and checked they pointed to right directory, and the
>> > configure step ran without error. I verified I do have read
permission
>> for netcdf.hh, any ideas on this.
>> >
>> > Thanks,
>> > Brock
>> >
>> > -----Original Message-----
>> > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
>> > Sent: Monday, November 17, 2014 4:34 PM
>> > To: Burghardt, Brock
>> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>> >
>> > Hi Brock.
>> >
>> > I see that you are having problems building METv4.1 with an intel
>> > compiler.  Thank you for attaching your user_defs.mk and make.log
>> files.
>> >
>> > First, I wanted to mention that METv5.0 was release in early
>> > September, which does include bug fixes for METv4.1.  You can
view the
>> > release notes for METv5.0 here:
>> >
>> >
>> >
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_releas
>> > e_notes.php
>> >
>> > For now, let's try to overcome this error with METv4.1:
>> >
>> > ld: cannot find -lm
>> >
>> > We've had this issue come up with some other users in the past.
By
>> > default, MET does static linking - meaning that it links in the
>> > library files on which it depends at compile time. The other type
of
>> > linking is call dynamic linking - for which the library files on
which
>> > it depends are linked in at runtime. Other users have had success
>> > compiling MET with dynamic linking instead. Open up and edit the
>> > top-level user_defs.mk file, by commenting out or removing
references
>> > to "-static". They show up on the CXX_FLAGS and FC_FLAGS lines.
>> >
>> > Then try recompiling MET, being sure to execute a "make clean"
first.
>> >
>> > Thanks,
>> > Julie
>> >
>> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT <
>> > met_help at ucar.edu>
>> > wrote:
>> >
>> > >
>> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
>> > > Transaction: Ticket created by brock.burghardt at ttu.edu
>> > >        Queue: met_help
>> > >      Subject: Issue compiling MET V4.1
>> > >        Owner: Nobody
>> > >   Requestors: brock.burghardt at ttu.edu
>> > >       Status: new
>> > >  Ticket <URL:
>> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>> > > >
>> > >
>> > >
>> > > Hello,
>> > >
>> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3 with an
intel
>> > > compiler (definitions and log file attachted). I've gone
through all
>> > > suggestions in the appendix of the users guide but still get an
>> > > error while trying to make the ensemble_stat executable. Is
there a
>> > > liblm.so or .a required for this build?
>> > >
>> > > Thanks,
>> > > Brock
>> > >
>> > > ld: cannot find -lm
>> > > make[4]: *** [ensemble_stat] Error 1
>> > >
>> > >
>> > >
>> > > Brock Burghardt
>> > > Atmospheric Science PhD Student
>> > > Dept. of Geosciences
>> > > Texas Tech University
>> > >
>> > >
>> > >
>> >
>> >
>> > --
>> > Julie Prestopnik
>> > National Center for Atmospheric Research Research Applications
>> > Laboratory
>> > Phone: 303.497.8399
>> > Email: jpresto at ucar.edu
>> >
>> >
>> >
>>
>>
>> --
>> Julie Prestopnik
>> National Center for Atmospheric Research Research Applications
Laboratory
>> Phone: 303.497.8399
>> Email: jpresto at ucar.edu
>>
>>
>>
>>
>
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research
> Research Applications Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>

--
Julie Prestopnik
National Center for Atmospheric Research
Research Applications Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Julie Prestopnik
Time: Tue Dec 02 10:02:13 2014

Hi Brock.  I'm sorry to hear that you are still having problems
compiling
METv5.0.  Thank you for sending the config.log file.

Please try running the following command:

make clean
module load intel

and then run the configure command and the make install command.

Please let me know if that works.

Thanks,
Julie

On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi Julie,
>
> I've still been unable to compile METv5.0...independent of whether I
> specify the parent or both the parent/lib parent/inc paths for
netcdf,
> bufrlib, and GSL, I get an error with it trying to open netcdf.hh
>
> I noticed two paths to netcdf got set in the config.log (attached),
one I
> set and another on the system, removing the one I did not define
still did
> not give a successful build.
>
> I noticed I had errors in the config.log for many of the compilers
where
> if tries to call, for instance, icpc -qversion which should be icpc
-V....I
> looked through all the config and configure scripts and could not
find the
> explicit icpc -qversion call to change to -V...do you think this
could be
> the issue?
>
> Thank you for following up,
> Brock
>
> ________________________________________
> From: Julie Prestopnik via RT [met_help at ucar.edu]
> Sent: Monday, December 01, 2014 9:56 AM
> To: Burghardt, Brock
> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>
> Hi Brock.  I just wanted to follow up - Were able to get a
successful build
> of METv5.0?
>
> Thanks,
> Julie
>
> On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
<jpresto at ucar.edu>
> wrote:
>
> > Hi Brock.
> >
> > Thanks you for sending the output.  I was looking to make sure
that the
> > MET environment variables, like MET_NETCDF or MET_NETCDFINC and
> > MET_NETCDFLIB, were set correctly.
> >
> > Please take a look at the following page and follow the
instructions on
> > compiling MET:
> >
> >
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/compilation/configure.php
> >
> > This will help guide you in setting up the required environment
variables
> > and will hopefully resolve the problem you are experiencing with
MET not
> > being able to find netcdf.hh.
> >
> > Please let me know if you have any questions, and please let me
know if
> > you get a successful build after setting up the environment
variables.
> >
> > Thanks,
> > Julie
> >
> > On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> > met_help at ucar.edu> wrote:
> >
> >>
> >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >>
> >> Hi Julie, the output for the command is,
> >>
> >> hrothgar:/met-5.0$ printenv | grep met
> >> PWD=/home/brburgha/met-5.0
> >>
> >> ...unless you wanted this,
> >>
> >> hrothgar:/met-5.0$ printenv | grep MET*
> >> HOSTNAME=hrothgar.hpcc.ttu.edu
> >> ANT_HOME=/opt/rocks
> >> JAVA_HOME=/usr/java/latest
> >> P4_GLOBMEMSIZE=83886080
> >> MAGICK_HOME=/lustre/work/apps/ImageMagick
> >> HOME=/home/brburgha
> >> LOGNAME=brburgha
> >> G_BROKEN_FILENAMES=1
> >>
> >>
> >> Brock
> >>
> >>
> >> -----Original Message-----
> >> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> >> Sent: Tuesday, November 18, 2014 10:38 AM
> >> To: Burghardt, Brock
> >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >>
> >> Hi Brock.  Thank you for getting the latest version of MET.  I
see
> you're
> >> still problems with MET finding netcdf.hh.
> >>
> >> Could you please run "printenv | grep MET" and send me the
output?
> >>
> >> Thanks,
> >> Julie
> >>
> >> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
> >> met_help at ucar.edu>
> >> wrote:
> >>
> >> >
> >> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >> >
> >> > Hi Julie,
> >> >
> >> > Thanks for getting back to me so soon and explaining. Removing
the
> >> > static flag did indeed help in compiling the executables I
needed but
> >> > it did not completely compile (another ld: missing library
error came
> >> > up in the pb2nc build). However, I took your advice and started
over
> >> > with V5.0 and got that running, but the compile also failed
before
> >> > making executables. The error was for not being able to open
netcdf.hh
> >> > (make_install.log attached). I added all the met_<lib>
arguments at
> >> > command line and checked they pointed to right directory, and
the
> >> > configure step ran without error. I verified I do have read
permission
> >> for netcdf.hh, any ideas on this.
> >> >
> >> > Thanks,
> >> > Brock
> >> >
> >> > -----Original Message-----
> >> > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> >> > Sent: Monday, November 17, 2014 4:34 PM
> >> > To: Burghardt, Brock
> >> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >> >
> >> > Hi Brock.
> >> >
> >> > I see that you are having problems building METv4.1 with an
intel
> >> > compiler.  Thank you for attaching your user_defs.mk and
make.log
> >> files.
> >> >
> >> > First, I wanted to mention that METv5.0 was release in early
> >> > September, which does include bug fixes for METv4.1.  You can
view the
> >> > release notes for METv5.0 here:
> >> >
> >> >
> >> >
>
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_releas
> >> > e_notes.php
> >> >
> >> > For now, let's try to overcome this error with METv4.1:
> >> >
> >> > ld: cannot find -lm
> >> >
> >> > We've had this issue come up with some other users in the past.
By
> >> > default, MET does static linking - meaning that it links in the
> >> > library files on which it depends at compile time. The other
type of
> >> > linking is call dynamic linking - for which the library files
on which
> >> > it depends are linked in at runtime. Other users have had
success
> >> > compiling MET with dynamic linking instead. Open up and edit
the
> >> > top-level user_defs.mk file, by commenting out or removing
references
> >> > to "-static". They show up on the CXX_FLAGS and FC_FLAGS lines.
> >> >
> >> > Then try recompiling MET, being sure to execute a "make clean"
first.
> >> >
> >> > Thanks,
> >> > Julie
> >> >
> >> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT <
> >> > met_help at ucar.edu>
> >> > wrote:
> >> >
> >> > >
> >> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> >> > > Transaction: Ticket created by brock.burghardt at ttu.edu
> >> > >        Queue: met_help
> >> > >      Subject: Issue compiling MET V4.1
> >> > >        Owner: Nobody
> >> > >   Requestors: brock.burghardt at ttu.edu
> >> > >       Status: new
> >> > >  Ticket <URL:
> >> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> >> > > >
> >> > >
> >> > >
> >> > > Hello,
> >> > >
> >> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3 with an
intel
> >> > > compiler (definitions and log file attachted). I've gone
through all
> >> > > suggestions in the appendix of the users guide but still get
an
> >> > > error while trying to make the ensemble_stat executable. Is
there a
> >> > > liblm.so or .a required for this build?
> >> > >
> >> > > Thanks,
> >> > > Brock
> >> > >
> >> > > ld: cannot find -lm
> >> > > make[4]: *** [ensemble_stat] Error 1
> >> > >
> >> > >
> >> > >
> >> > > Brock Burghardt
> >> > > Atmospheric Science PhD Student
> >> > > Dept. of Geosciences
> >> > > Texas Tech University
> >> > >
> >> > >
> >> > >
> >> >
> >> >
> >> > --
> >> > Julie Prestopnik
> >> > National Center for Atmospheric Research Research Applications
> >> > Laboratory
> >> > Phone: 303.497.8399
> >> > Email: jpresto at ucar.edu
> >> >
> >> >
> >> >
> >>
> >>
> >> --
> >> Julie Prestopnik
> >> National Center for Atmospheric Research Research Applications
> Laboratory
> >> Phone: 303.497.8399
> >> Email: jpresto at ucar.edu
> >>
> >>
> >>
> >>
> >
> >
> > --
> > Julie Prestopnik
> > National Center for Atmospheric Research
> > Research Applications Laboratory
> > Phone: 303.497.8399
> > Email: jpresto at ucar.edu
> >
>
>
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research
> Research Applications Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>
>
>

--
Julie Prestopnik
National Center for Atmospheric Research
Research Applications Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: RE: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
From: Burghardt, Brock
Time: Thu Dec 04 17:09:18 2014

Hi Julie,

Still no luck with loading the intel environment (soft add intel in my
case) with the error developing with netcdf.hh in the make
install...any other ideas?

Thanks, I appreciate the help.

Brock

________________________________________
From: Julie Prestopnik via RT [met_help at ucar.edu]
Sent: Tuesday, December 02, 2014 11:02 AM
To: Burghardt, Brock
Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1

Hi Brock.  I'm sorry to hear that you are still having problems
compiling
METv5.0.  Thank you for sending the config.log file.

Please try running the following command:

make clean
module load intel

and then run the configure command and the make install command.

Please let me know if that works.

Thanks,
Julie

On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi Julie,
>
> I've still been unable to compile METv5.0...independent of whether I
> specify the parent or both the parent/lib parent/inc paths for
netcdf,
> bufrlib, and GSL, I get an error with it trying to open netcdf.hh
>
> I noticed two paths to netcdf got set in the config.log (attached),
one I
> set and another on the system, removing the one I did not define
still did
> not give a successful build.
>
> I noticed I had errors in the config.log for many of the compilers
where
> if tries to call, for instance, icpc -qversion which should be icpc
-V....I
> looked through all the config and configure scripts and could not
find the
> explicit icpc -qversion call to change to -V...do you think this
could be
> the issue?
>
> Thank you for following up,
> Brock
>
> ________________________________________
> From: Julie Prestopnik via RT [met_help at ucar.edu]
> Sent: Monday, December 01, 2014 9:56 AM
> To: Burghardt, Brock
> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>
> Hi Brock.  I just wanted to follow up - Were able to get a
successful build
> of METv5.0?
>
> Thanks,
> Julie
>
> On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
<jpresto at ucar.edu>
> wrote:
>
> > Hi Brock.
> >
> > Thanks you for sending the output.  I was looking to make sure
that the
> > MET environment variables, like MET_NETCDF or MET_NETCDFINC and
> > MET_NETCDFLIB, were set correctly.
> >
> > Please take a look at the following page and follow the
instructions on
> > compiling MET:
> >
> >
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/compilation/configure.php
> >
> > This will help guide you in setting up the required environment
variables
> > and will hopefully resolve the problem you are experiencing with
MET not
> > being able to find netcdf.hh.
> >
> > Please let me know if you have any questions, and please let me
know if
> > you get a successful build after setting up the environment
variables.
> >
> > Thanks,
> > Julie
> >
> > On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> > met_help at ucar.edu> wrote:
> >
> >>
> >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >>
> >> Hi Julie, the output for the command is,
> >>
> >> hrothgar:/met-5.0$ printenv | grep met
> >> PWD=/home/brburgha/met-5.0
> >>
> >> ...unless you wanted this,
> >>
> >> hrothgar:/met-5.0$ printenv | grep MET*
> >> HOSTNAME=hrothgar.hpcc.ttu.edu
> >> ANT_HOME=/opt/rocks
> >> JAVA_HOME=/usr/java/latest
> >> P4_GLOBMEMSIZE=83886080
> >> MAGICK_HOME=/lustre/work/apps/ImageMagick
> >> HOME=/home/brburgha
> >> LOGNAME=brburgha
> >> G_BROKEN_FILENAMES=1
> >>
> >>
> >> Brock
> >>
> >>
> >> -----Original Message-----
> >> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> >> Sent: Tuesday, November 18, 2014 10:38 AM
> >> To: Burghardt, Brock
> >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >>
> >> Hi Brock.  Thank you for getting the latest version of MET.  I
see
> you're
> >> still problems with MET finding netcdf.hh.
> >>
> >> Could you please run "printenv | grep MET" and send me the
output?
> >>
> >> Thanks,
> >> Julie
> >>
> >> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
> >> met_help at ucar.edu>
> >> wrote:
> >>
> >> >
> >> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >> >
> >> > Hi Julie,
> >> >
> >> > Thanks for getting back to me so soon and explaining. Removing
the
> >> > static flag did indeed help in compiling the executables I
needed but
> >> > it did not completely compile (another ld: missing library
error came
> >> > up in the pb2nc build). However, I took your advice and started
over
> >> > with V5.0 and got that running, but the compile also failed
before
> >> > making executables. The error was for not being able to open
netcdf.hh
> >> > (make_install.log attached). I added all the met_<lib>
arguments at
> >> > command line and checked they pointed to right directory, and
the
> >> > configure step ran without error. I verified I do have read
permission
> >> for netcdf.hh, any ideas on this.
> >> >
> >> > Thanks,
> >> > Brock
> >> >
> >> > -----Original Message-----
> >> > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> >> > Sent: Monday, November 17, 2014 4:34 PM
> >> > To: Burghardt, Brock
> >> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >> >
> >> > Hi Brock.
> >> >
> >> > I see that you are having problems building METv4.1 with an
intel
> >> > compiler.  Thank you for attaching your user_defs.mk and
make.log
> >> files.
> >> >
> >> > First, I wanted to mention that METv5.0 was release in early
> >> > September, which does include bug fixes for METv4.1.  You can
view the
> >> > release notes for METv5.0 here:
> >> >
> >> >
> >> >
>
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_releas
> >> > e_notes.php
> >> >
> >> > For now, let's try to overcome this error with METv4.1:
> >> >
> >> > ld: cannot find -lm
> >> >
> >> > We've had this issue come up with some other users in the past.
By
> >> > default, MET does static linking - meaning that it links in the
> >> > library files on which it depends at compile time. The other
type of
> >> > linking is call dynamic linking - for which the library files
on which
> >> > it depends are linked in at runtime. Other users have had
success
> >> > compiling MET with dynamic linking instead. Open up and edit
the
> >> > top-level user_defs.mk file, by commenting out or removing
references
> >> > to "-static". They show up on the CXX_FLAGS and FC_FLAGS lines.
> >> >
> >> > Then try recompiling MET, being sure to execute a "make clean"
first.
> >> >
> >> > Thanks,
> >> > Julie
> >> >
> >> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT <
> >> > met_help at ucar.edu>
> >> > wrote:
> >> >
> >> > >
> >> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> >> > > Transaction: Ticket created by brock.burghardt at ttu.edu
> >> > >        Queue: met_help
> >> > >      Subject: Issue compiling MET V4.1
> >> > >        Owner: Nobody
> >> > >   Requestors: brock.burghardt at ttu.edu
> >> > >       Status: new
> >> > >  Ticket <URL:
> >> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> >> > > >
> >> > >
> >> > >
> >> > > Hello,
> >> > >
> >> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3 with an
intel
> >> > > compiler (definitions and log file attachted). I've gone
through all
> >> > > suggestions in the appendix of the users guide but still get
an
> >> > > error while trying to make the ensemble_stat executable. Is
there a
> >> > > liblm.so or .a required for this build?
> >> > >
> >> > > Thanks,
> >> > > Brock
> >> > >
> >> > > ld: cannot find -lm
> >> > > make[4]: *** [ensemble_stat] Error 1
> >> > >
> >> > >
> >> > >
> >> > > Brock Burghardt
> >> > > Atmospheric Science PhD Student
> >> > > Dept. of Geosciences
> >> > > Texas Tech University
> >> > >
> >> > >
> >> > >
> >> >
> >> >
> >> > --
> >> > Julie Prestopnik
> >> > National Center for Atmospheric Research Research Applications
> >> > Laboratory
> >> > Phone: 303.497.8399
> >> > Email: jpresto at ucar.edu
> >> >
> >> >
> >> >
> >>
> >>
> >> --
> >> Julie Prestopnik
> >> National Center for Atmospheric Research Research Applications
> Laboratory
> >> Phone: 303.497.8399
> >> Email: jpresto at ucar.edu
> >>
> >>
> >>
> >>
> >
> >
> > --
> > Julie Prestopnik
> > National Center for Atmospheric Research
> > Research Applications Laboratory
> > Phone: 303.497.8399
> > Email: jpresto at ucar.edu
> >
>
>
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research
> Research Applications Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>
>
>

--
Julie Prestopnik
National Center for Atmospheric Research
Research Applications Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Julie Prestopnik
Time: Fri Dec 05 08:55:58 2014

Hi Brock.  Could you please send me your make_install.log file and
could
you please run "printenv | grep MET" again and send me the output?

Thanks,
Julie

On Thu, Dec 4, 2014 at 5:09 PM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi Julie,
>
> Still no luck with loading the intel environment (soft add intel in
my
> case) with the error developing with netcdf.hh in the make
install...any
> other ideas?
>
> Thanks, I appreciate the help.
>
> Brock
>
>
>
>
> ________________________________________
> From: Julie Prestopnik via RT [met_help at ucar.edu]
> Sent: Tuesday, December 02, 2014 11:02 AM
> To: Burghardt, Brock
> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>
> Hi Brock.  I'm sorry to hear that you are still having problems
compiling
> METv5.0.  Thank you for sending the config.log file.
>
> Please try running the following command:
>
> make clean
> module load intel
>
> and then run the configure command and the make install command.
>
> Please let me know if that works.
>
> Thanks,
> Julie
>
> On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
<met_help at ucar.edu
> >
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >
> > Hi Julie,
> >
> > I've still been unable to compile METv5.0...independent of whether
I
> > specify the parent or both the parent/lib parent/inc paths for
netcdf,
> > bufrlib, and GSL, I get an error with it trying to open netcdf.hh
> >
> > I noticed two paths to netcdf got set in the config.log
(attached), one I
> > set and another on the system, removing the one I did not define
still
> did
> > not give a successful build.
> >
> > I noticed I had errors in the config.log for many of the compilers
where
> > if tries to call, for instance, icpc -qversion which should be
icpc
> -V....I
> > looked through all the config and configure scripts and could not
find
> the
> > explicit icpc -qversion call to change to -V...do you think this
could be
> > the issue?
> >
> > Thank you for following up,
> > Brock
> >
> > ________________________________________
> > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > Sent: Monday, December 01, 2014 9:56 AM
> > To: Burghardt, Brock
> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >
> > Hi Brock.  I just wanted to follow up - Were able to get a
successful
> build
> > of METv5.0?
> >
> > Thanks,
> > Julie
> >
> > On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
<jpresto at ucar.edu>
> > wrote:
> >
> > > Hi Brock.
> > >
> > > Thanks you for sending the output.  I was looking to make sure
that the
> > > MET environment variables, like MET_NETCDF or MET_NETCDFINC and
> > > MET_NETCDFLIB, were set correctly.
> > >
> > > Please take a look at the following page and follow the
instructions on
> > > compiling MET:
> > >
> > >
> >
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/compilation/configure.php
> > >
> > > This will help guide you in setting up the required environment
> variables
> > > and will hopefully resolve the problem you are experiencing with
MET
> not
> > > being able to find netcdf.hh.
> > >
> > > Please let me know if you have any questions, and please let me
know if
> > > you get a successful build after setting up the environment
variables.
> > >
> > > Thanks,
> > > Julie
> > >
> > > On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > >>
> > >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > >>
> > >> Hi Julie, the output for the command is,
> > >>
> > >> hrothgar:/met-5.0$ printenv | grep met
> > >> PWD=/home/brburgha/met-5.0
> > >>
> > >> ...unless you wanted this,
> > >>
> > >> hrothgar:/met-5.0$ printenv | grep MET*
> > >> HOSTNAME=hrothgar.hpcc.ttu.edu
> > >> ANT_HOME=/opt/rocks
> > >> JAVA_HOME=/usr/java/latest
> > >> P4_GLOBMEMSIZE=83886080
> > >> MAGICK_HOME=/lustre/work/apps/ImageMagick
> > >> HOME=/home/brburgha
> > >> LOGNAME=brburgha
> > >> G_BROKEN_FILENAMES=1
> > >>
> > >>
> > >> Brock
> > >>
> > >>
> > >> -----Original Message-----
> > >> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > >> Sent: Tuesday, November 18, 2014 10:38 AM
> > >> To: Burghardt, Brock
> > >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > >>
> > >> Hi Brock.  Thank you for getting the latest version of MET.  I
see
> > you're
> > >> still problems with MET finding netcdf.hh.
> > >>
> > >> Could you please run "printenv | grep MET" and send me the
output?
> > >>
> > >> Thanks,
> > >> Julie
> > >>
> > >> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
> > >> met_help at ucar.edu>
> > >> wrote:
> > >>
> > >> >
> > >> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> > >> >
> > >> > Hi Julie,
> > >> >
> > >> > Thanks for getting back to me so soon and explaining.
Removing the
> > >> > static flag did indeed help in compiling the executables I
needed
> but
> > >> > it did not completely compile (another ld: missing library
error
> came
> > >> > up in the pb2nc build). However, I took your advice and
started over
> > >> > with V5.0 and got that running, but the compile also failed
before
> > >> > making executables. The error was for not being able to open
> netcdf.hh
> > >> > (make_install.log attached). I added all the met_<lib>
arguments at
> > >> > command line and checked they pointed to right directory, and
the
> > >> > configure step ran without error. I verified I do have read
> permission
> > >> for netcdf.hh, any ideas on this.
> > >> >
> > >> > Thanks,
> > >> > Brock
> > >> >
> > >> > -----Original Message-----
> > >> > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > >> > Sent: Monday, November 17, 2014 4:34 PM
> > >> > To: Burghardt, Brock
> > >> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > >> >
> > >> > Hi Brock.
> > >> >
> > >> > I see that you are having problems building METv4.1 with an
intel
> > >> > compiler.  Thank you for attaching your user_defs.mk and
make.log
> > >> files.
> > >> >
> > >> > First, I wanted to mention that METv5.0 was release in early
> > >> > September, which does include bug fixes for METv4.1.  You can
view
> the
> > >> > release notes for METv5.0 here:
> > >> >
> > >> >
> > >> >
> >
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_releas
> > >> > e_notes.php
> > >> >
> > >> > For now, let's try to overcome this error with METv4.1:
> > >> >
> > >> > ld: cannot find -lm
> > >> >
> > >> > We've had this issue come up with some other users in the
past.  By
> > >> > default, MET does static linking - meaning that it links in
the
> > >> > library files on which it depends at compile time. The other
type of
> > >> > linking is call dynamic linking - for which the library files
on
> which
> > >> > it depends are linked in at runtime. Other users have had
success
> > >> > compiling MET with dynamic linking instead. Open up and edit
the
> > >> > top-level user_defs.mk file, by commenting out or removing
> references
> > >> > to "-static". They show up on the CXX_FLAGS and FC_FLAGS
lines.
> > >> >
> > >> > Then try recompiling MET, being sure to execute a "make
clean"
> first.
> > >> >
> > >> > Thanks,
> > >> > Julie
> > >> >
> > >> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT <
> > >> > met_help at ucar.edu>
> > >> > wrote:
> > >> >
> > >> > >
> > >> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> > >> > > Transaction: Ticket created by brock.burghardt at ttu.edu
> > >> > >        Queue: met_help
> > >> > >      Subject: Issue compiling MET V4.1
> > >> > >        Owner: Nobody
> > >> > >   Requestors: brock.burghardt at ttu.edu
> > >> > >       Status: new
> > >> > >  Ticket <URL:
> > >> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > >> > > >
> > >> > >
> > >> > >
> > >> > > Hello,
> > >> > >
> > >> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3 with an
intel
> > >> > > compiler (definitions and log file attachted). I've gone
through
> all
> > >> > > suggestions in the appendix of the users guide but still
get an
> > >> > > error while trying to make the ensemble_stat executable. Is
there
> a
> > >> > > liblm.so or .a required for this build?
> > >> > >
> > >> > > Thanks,
> > >> > > Brock
> > >> > >
> > >> > > ld: cannot find -lm
> > >> > > make[4]: *** [ensemble_stat] Error 1
> > >> > >
> > >> > >
> > >> > >
> > >> > > Brock Burghardt
> > >> > > Atmospheric Science PhD Student
> > >> > > Dept. of Geosciences
> > >> > > Texas Tech University
> > >> > >
> > >> > >
> > >> > >
> > >> >
> > >> >
> > >> > --
> > >> > Julie Prestopnik
> > >> > National Center for Atmospheric Research Research
Applications
> > >> > Laboratory
> > >> > Phone: 303.497.8399
> > >> > Email: jpresto at ucar.edu
> > >> >
> > >> >
> > >> >
> > >>
> > >>
> > >> --
> > >> Julie Prestopnik
> > >> National Center for Atmospheric Research Research Applications
> > Laboratory
> > >> Phone: 303.497.8399
> > >> Email: jpresto at ucar.edu
> > >>
> > >>
> > >>
> > >>
> > >
> > >
> > > --
> > > Julie Prestopnik
> > > National Center for Atmospheric Research
> > > Research Applications Laboratory
> > > Phone: 303.497.8399
> > > Email: jpresto at ucar.edu
> > >
> >
> >
> >
> > --
> > Julie Prestopnik
> > National Center for Atmospheric Research
> > Research Applications Laboratory
> > Phone: 303.497.8399
> > Email: jpresto at ucar.edu
> >
> >
> >
>
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research
> Research Applications Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>
>
>
>

--
Julie Prestopnik
National Center for Atmospheric Research
Research Applications Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Burghardt, Brock
Time: Mon Dec 08 14:03:12 2014

Hi Julie,
Sorry for not getting back to you sooner (finals just wrapped up).
I've attached the make log.
The printenv statement does not give anything for "MET", however when
I set my MET_<LIB/INC> paths at command line I can print them, giving
the proper paths.

Thanks,
Brock

-----Original Message-----
From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
Sent: Friday, December 05, 2014 9:56 AM
To: Burghardt, Brock
Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1

Hi Brock.  Could you please send me your make_install.log file and
could you please run "printenv | grep MET" again and send me the
output?

Thanks,
Julie

On Thu, Dec 4, 2014 at 5:09 PM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi Julie,
>
> Still no luck with loading the intel environment (soft add intel in
my
> case) with the error developing with netcdf.hh in the make
> install...any other ideas?
>
> Thanks, I appreciate the help.
>
> Brock
>
>
>
>
> ________________________________________
> From: Julie Prestopnik via RT [met_help at ucar.edu]
> Sent: Tuesday, December 02, 2014 11:02 AM
> To: Burghardt, Brock
> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>
> Hi Brock.  I'm sorry to hear that you are still having problems
> compiling METv5.0.  Thank you for sending the config.log file.
>
> Please try running the following command:
>
> make clean
> module load intel
>
> and then run the configure command and the make install command.
>
> Please let me know if that works.
>
> Thanks,
> Julie
>
> On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
> <met_help at ucar.edu
> >
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >
> > Hi Julie,
> >
> > I've still been unable to compile METv5.0...independent of whether
I
> > specify the parent or both the parent/lib parent/inc paths for
> > netcdf, bufrlib, and GSL, I get an error with it trying to open
> > netcdf.hh
> >
> > I noticed two paths to netcdf got set in the config.log
(attached),
> > one I set and another on the system, removing the one I did not
> > define still
> did
> > not give a successful build.
> >
> > I noticed I had errors in the config.log for many of the compilers
> > where if tries to call, for instance, icpc -qversion which should
be
> > icpc
> -V....I
> > looked through all the config and configure scripts and could not
> > find
> the
> > explicit icpc -qversion call to change to -V...do you think this
> > could be the issue?
> >
> > Thank you for following up,
> > Brock
> >
> > ________________________________________
> > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > Sent: Monday, December 01, 2014 9:56 AM
> > To: Burghardt, Brock
> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >
> > Hi Brock.  I just wanted to follow up - Were able to get a
> > successful
> build
> > of METv5.0?
> >
> > Thanks,
> > Julie
> >
> > On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
> > <jpresto at ucar.edu>
> > wrote:
> >
> > > Hi Brock.
> > >
> > > Thanks you for sending the output.  I was looking to make sure
> > > that the MET environment variables, like MET_NETCDF or
> > > MET_NETCDFINC and MET_NETCDFLIB, were set correctly.
> > >
> > > Please take a look at the following page and follow the
> > > instructions on compiling MET:
> > >
> > >
> >
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/comp
> ilation/configure.php
> > >
> > > This will help guide you in setting up the required environment
> variables
> > > and will hopefully resolve the problem you are experiencing with
> > > MET
> not
> > > being able to find netcdf.hh.
> > >
> > > Please let me know if you have any questions, and please let me
> > > know if you get a successful build after setting up the
environment variables.
> > >
> > > Thanks,
> > > Julie
> > >
> > > On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > >>
> > >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > >>
> > >> Hi Julie, the output for the command is,
> > >>
> > >> hrothgar:/met-5.0$ printenv | grep met
> > >> PWD=/home/brburgha/met-5.0
> > >>
> > >> ...unless you wanted this,
> > >>
> > >> hrothgar:/met-5.0$ printenv | grep MET*
> > >> HOSTNAME=hrothgar.hpcc.ttu.edu ANT_HOME=/opt/rocks
> > >> JAVA_HOME=/usr/java/latest
> > >> P4_GLOBMEMSIZE=83886080
> > >> MAGICK_HOME=/lustre/work/apps/ImageMagick
> > >> HOME=/home/brburgha
> > >> LOGNAME=brburgha
> > >> G_BROKEN_FILENAMES=1
> > >>
> > >>
> > >> Brock
> > >>
> > >>
> > >> -----Original Message-----
> > >> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > >> Sent: Tuesday, November 18, 2014 10:38 AM
> > >> To: Burghardt, Brock
> > >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > >>
> > >> Hi Brock.  Thank you for getting the latest version of MET.  I
> > >> see
> > you're
> > >> still problems with MET finding netcdf.hh.
> > >>
> > >> Could you please run "printenv | grep MET" and send me the
output?
> > >>
> > >> Thanks,
> > >> Julie
> > >>
> > >> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
> > >> met_help at ucar.edu>
> > >> wrote:
> > >>
> > >> >
> > >> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> > >> >
> > >> > Hi Julie,
> > >> >
> > >> > Thanks for getting back to me so soon and explaining.
Removing
> > >> > the static flag did indeed help in compiling the executables
I
> > >> > needed
> but
> > >> > it did not completely compile (another ld: missing library
> > >> > error
> came
> > >> > up in the pb2nc build). However, I took your advice and
started
> > >> > over with V5.0 and got that running, but the compile also
> > >> > failed before making executables. The error was for not being
> > >> > able to open
> netcdf.hh
> > >> > (make_install.log attached). I added all the met_<lib>
> > >> > arguments at command line and checked they pointed to right
> > >> > directory, and the configure step ran without error. I
verified
> > >> > I do have read
> permission
> > >> for netcdf.hh, any ideas on this.
> > >> >
> > >> > Thanks,
> > >> > Brock
> > >> >
> > >> > -----Original Message-----
> > >> > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > >> > Sent: Monday, November 17, 2014 4:34 PM
> > >> > To: Burghardt, Brock
> > >> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > >> >
> > >> > Hi Brock.
> > >> >
> > >> > I see that you are having problems building METv4.1 with an
> > >> > intel compiler.  Thank you for attaching your user_defs.mk
and
> > >> > make.log
> > >> files.
> > >> >
> > >> > First, I wanted to mention that METv5.0 was release in early
> > >> > September, which does include bug fixes for METv4.1.  You can
> > >> > view
> the
> > >> > release notes for METv5.0 here:
> > >> >
> > >> >
> > >> >
> >
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_rele
> > as
> > >> > e_notes.php
> > >> >
> > >> > For now, let's try to overcome this error with METv4.1:
> > >> >
> > >> > ld: cannot find -lm
> > >> >
> > >> > We've had this issue come up with some other users in the
past.
> > >> > By default, MET does static linking - meaning that it links
in
> > >> > the library files on which it depends at compile time. The
> > >> > other type of linking is call dynamic linking - for which the
> > >> > library files on
> which
> > >> > it depends are linked in at runtime. Other users have had
> > >> > success compiling MET with dynamic linking instead. Open up
and
> > >> > edit the top-level user_defs.mk file, by commenting out or
> > >> > removing
> references
> > >> > to "-static". They show up on the CXX_FLAGS and FC_FLAGS
lines.
> > >> >
> > >> > Then try recompiling MET, being sure to execute a "make
clean"
> first.
> > >> >
> > >> > Thanks,
> > >> > Julie
> > >> >
> > >> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT <
> > >> > met_help at ucar.edu>
> > >> > wrote:
> > >> >
> > >> > >
> > >> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> > >> > > Transaction: Ticket created by brock.burghardt at ttu.edu
> > >> > >        Queue: met_help
> > >> > >      Subject: Issue compiling MET V4.1
> > >> > >        Owner: Nobody
> > >> > >   Requestors: brock.burghardt at ttu.edu
> > >> > >       Status: new
> > >> > >  Ticket <URL:
> > >> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > >> > > >
> > >> > >
> > >> > >
> > >> > > Hello,
> > >> > >
> > >> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3 with an
> > >> > > intel compiler (definitions and log file attachted). I've
> > >> > > gone through
> all
> > >> > > suggestions in the appendix of the users guide but still
get
> > >> > > an error while trying to make the ensemble_stat executable.
> > >> > > Is there
> a
> > >> > > liblm.so or .a required for this build?
> > >> > >
> > >> > > Thanks,
> > >> > > Brock
> > >> > >
> > >> > > ld: cannot find -lm
> > >> > > make[4]: *** [ensemble_stat] Error 1
> > >> > >
> > >> > >
> > >> > >
> > >> > > Brock Burghardt
> > >> > > Atmospheric Science PhD Student Dept. of Geosciences Texas
> > >> > > Tech University
> > >> > >
> > >> > >
> > >> > >
> > >> >
> > >> >
> > >> > --
> > >> > Julie Prestopnik
> > >> > National Center for Atmospheric Research Research
Applications
> > >> > Laboratory
> > >> > Phone: 303.497.8399
> > >> > Email: jpresto at ucar.edu
> > >> >
> > >> >
> > >> >
> > >>
> > >>
> > >> --
> > >> Julie Prestopnik
> > >> National Center for Atmospheric Research Research Applications
> > Laboratory
> > >> Phone: 303.497.8399
> > >> Email: jpresto at ucar.edu
> > >>
> > >>
> > >>
> > >>
> > >
> > >
> > > --
> > > Julie Prestopnik
> > > National Center for Atmospheric Research Research Applications
> > > Laboratory
> > > Phone: 303.497.8399
> > > Email: jpresto at ucar.edu
> > >
> >
> >
> >
> > --
> > Julie Prestopnik
> > National Center for Atmospheric Research Research Applications
> > Laboratory
> > Phone: 303.497.8399
> > Email: jpresto at ucar.edu
> >
> >
> >
>
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research Research Applications
> Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>
>
>
>

--
Julie Prestopnik
National Center for Atmospheric Research Research Applications
Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Julie Prestopnik
Time: Mon Dec 08 16:53:45 2014

Hi Brock.  I'm sorry to see that MET can't seem to find netcdf.hh.
Could
you please set up the environment variables on the command line, and
try to
build and install MET on the command line and then send me the
make_install.log file from that run along with the output of "printenv
|
grep MET".  I looked at this make_install.log, paying particular
attention
to what is being looked at for the include files.  I don't know where
your
netcdf include files are, so I can't check for that location.

If your environment variable is set correctly, it should be listed in
the
output of "printenv | grep MET" and also in -I in the make_install.log
file.  If MET still can't find it after we check those things, we'll
have
to look into it further.  I know that certain recent versions of
NetCDF
(e.g. v4.1.3) no longer include netcdf.hh, but I believe that 3.6.3,
which
you said you are using, does have netcdf.hh.

Sending this additional info from an install attempt with the
environment
variables set, should help us better figure out what the problem is.

Thanks,
Julie

On Mon, Dec 8, 2014 at 2:03 PM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi Julie,
> Sorry for not getting back to you sooner (finals just wrapped up).
> I've attached the make log.
> The printenv statement does not give anything for "MET", however
when I
> set my MET_<LIB/INC> paths at command line I can print them, giving
the
> proper paths.
>
> Thanks,
> Brock
>
> -----Original Message-----
> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> Sent: Friday, December 05, 2014 9:56 AM
> To: Burghardt, Brock
> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>
> Hi Brock.  Could you please send me your make_install.log file and
could
> you please run "printenv | grep MET" again and send me the output?
>
> Thanks,
> Julie
>
> On Thu, Dec 4, 2014 at 5:09 PM, Burghardt, Brock via RT
<met_help at ucar.edu
> >
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >
> > Hi Julie,
> >
> > Still no luck with loading the intel environment (soft add intel
in my
> > case) with the error developing with netcdf.hh in the make
> > install...any other ideas?
> >
> > Thanks, I appreciate the help.
> >
> > Brock
> >
> >
> >
> >
> > ________________________________________
> > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > Sent: Tuesday, December 02, 2014 11:02 AM
> > To: Burghardt, Brock
> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >
> > Hi Brock.  I'm sorry to hear that you are still having problems
> > compiling METv5.0.  Thank you for sending the config.log file.
> >
> > Please try running the following command:
> >
> > make clean
> > module load intel
> >
> > and then run the configure command and the make install command.
> >
> > Please let me know if that works.
> >
> > Thanks,
> > Julie
> >
> > On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
> > <met_help at ucar.edu
> > >
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > >
> > > Hi Julie,
> > >
> > > I've still been unable to compile METv5.0...independent of
whether I
> > > specify the parent or both the parent/lib parent/inc paths for
> > > netcdf, bufrlib, and GSL, I get an error with it trying to open
> > > netcdf.hh
> > >
> > > I noticed two paths to netcdf got set in the config.log
(attached),
> > > one I set and another on the system, removing the one I did not
> > > define still
> > did
> > > not give a successful build.
> > >
> > > I noticed I had errors in the config.log for many of the
compilers
> > > where if tries to call, for instance, icpc -qversion which
should be
> > > icpc
> > -V....I
> > > looked through all the config and configure scripts and could
not
> > > find
> > the
> > > explicit icpc -qversion call to change to -V...do you think this
> > > could be the issue?
> > >
> > > Thank you for following up,
> > > Brock
> > >
> > > ________________________________________
> > > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > > Sent: Monday, December 01, 2014 9:56 AM
> > > To: Burghardt, Brock
> > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > >
> > > Hi Brock.  I just wanted to follow up - Were able to get a
> > > successful
> > build
> > > of METv5.0?
> > >
> > > Thanks,
> > > Julie
> > >
> > > On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
> > > <jpresto at ucar.edu>
> > > wrote:
> > >
> > > > Hi Brock.
> > > >
> > > > Thanks you for sending the output.  I was looking to make sure
> > > > that the MET environment variables, like MET_NETCDF or
> > > > MET_NETCDFINC and MET_NETCDFLIB, were set correctly.
> > > >
> > > > Please take a look at the following page and follow the
> > > > instructions on compiling MET:
> > > >
> > > >
> > >
> >
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/comp
> > ilation/configure.php
> > > >
> > > > This will help guide you in setting up the required
environment
> > variables
> > > > and will hopefully resolve the problem you are experiencing
with
> > > > MET
> > not
> > > > being able to find netcdf.hh.
> > > >
> > > > Please let me know if you have any questions, and please let
me
> > > > know if you get a successful build after setting up the
environment
> variables.
> > > >
> > > > Thanks,
> > > > Julie
> > > >
> > > > On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > >>
> > > >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> > > >>
> > > >> Hi Julie, the output for the command is,
> > > >>
> > > >> hrothgar:/met-5.0$ printenv | grep met
> > > >> PWD=/home/brburgha/met-5.0
> > > >>
> > > >> ...unless you wanted this,
> > > >>
> > > >> hrothgar:/met-5.0$ printenv | grep MET*
> > > >> HOSTNAME=hrothgar.hpcc.ttu.edu ANT_HOME=/opt/rocks
> > > >> JAVA_HOME=/usr/java/latest
> > > >> P4_GLOBMEMSIZE=83886080
> > > >> MAGICK_HOME=/lustre/work/apps/ImageMagick
> > > >> HOME=/home/brburgha
> > > >> LOGNAME=brburgha
> > > >> G_BROKEN_FILENAMES=1
> > > >>
> > > >>
> > > >> Brock
> > > >>
> > > >>
> > > >> -----Original Message-----
> > > >> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > > >> Sent: Tuesday, November 18, 2014 10:38 AM
> > > >> To: Burghardt, Brock
> > > >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > > >>
> > > >> Hi Brock.  Thank you for getting the latest version of MET.
I
> > > >> see
> > > you're
> > > >> still problems with MET finding netcdf.hh.
> > > >>
> > > >> Could you please run "printenv | grep MET" and send me the
output?
> > > >>
> > > >> Thanks,
> > > >> Julie
> > > >>
> > > >> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
> > > >> met_help at ucar.edu>
> > > >> wrote:
> > > >>
> > > >> >
> > > >> > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > > >> >
> > > >> > Hi Julie,
> > > >> >
> > > >> > Thanks for getting back to me so soon and explaining.
Removing
> > > >> > the static flag did indeed help in compiling the
executables I
> > > >> > needed
> > but
> > > >> > it did not completely compile (another ld: missing library
> > > >> > error
> > came
> > > >> > up in the pb2nc build). However, I took your advice and
started
> > > >> > over with V5.0 and got that running, but the compile also
> > > >> > failed before making executables. The error was for not
being
> > > >> > able to open
> > netcdf.hh
> > > >> > (make_install.log attached). I added all the met_<lib>
> > > >> > arguments at command line and checked they pointed to right
> > > >> > directory, and the configure step ran without error. I
verified
> > > >> > I do have read
> > permission
> > > >> for netcdf.hh, any ideas on this.
> > > >> >
> > > >> > Thanks,
> > > >> > Brock
> > > >> >
> > > >> > -----Original Message-----
> > > >> > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > > >> > Sent: Monday, November 17, 2014 4:34 PM
> > > >> > To: Burghardt, Brock
> > > >> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > > >> >
> > > >> > Hi Brock.
> > > >> >
> > > >> > I see that you are having problems building METv4.1 with an
> > > >> > intel compiler.  Thank you for attaching your user_defs.mk
and
> > > >> > make.log
> > > >> files.
> > > >> >
> > > >> > First, I wanted to mention that METv5.0 was release in
early
> > > >> > September, which does include bug fixes for METv4.1.  You
can
> > > >> > view
> > the
> > > >> > release notes for METv5.0 here:
> > > >> >
> > > >> >
> > > >> >
> > >
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_rele
> > > as
> > > >> > e_notes.php
> > > >> >
> > > >> > For now, let's try to overcome this error with METv4.1:
> > > >> >
> > > >> > ld: cannot find -lm
> > > >> >
> > > >> > We've had this issue come up with some other users in the
past.
> > > >> > By default, MET does static linking - meaning that it links
in
> > > >> > the library files on which it depends at compile time. The
> > > >> > other type of linking is call dynamic linking - for which
the
> > > >> > library files on
> > which
> > > >> > it depends are linked in at runtime. Other users have had
> > > >> > success compiling MET with dynamic linking instead. Open up
and
> > > >> > edit the top-level user_defs.mk file, by commenting out or
> > > >> > removing
> > references
> > > >> > to "-static". They show up on the CXX_FLAGS and FC_FLAGS
lines.
> > > >> >
> > > >> > Then try recompiling MET, being sure to execute a "make
clean"
> > first.
> > > >> >
> > > >> > Thanks,
> > > >> > Julie
> > > >> >
> > > >> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT <
> > > >> > met_help at ucar.edu>
> > > >> > wrote:
> > > >> >
> > > >> > >
> > > >> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> > > >> > > Transaction: Ticket created by brock.burghardt at ttu.edu
> > > >> > >        Queue: met_help
> > > >> > >      Subject: Issue compiling MET V4.1
> > > >> > >        Owner: Nobody
> > > >> > >   Requestors: brock.burghardt at ttu.edu
> > > >> > >       Status: new
> > > >> > >  Ticket <URL:
> > > >> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > > >> > > >
> > > >> > >
> > > >> > >
> > > >> > > Hello,
> > > >> > >
> > > >> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3 with
an
> > > >> > > intel compiler (definitions and log file attachted). I've
> > > >> > > gone through
> > all
> > > >> > > suggestions in the appendix of the users guide but still
get
> > > >> > > an error while trying to make the ensemble_stat
executable.
> > > >> > > Is there
> > a
> > > >> > > liblm.so or .a required for this build?
> > > >> > >
> > > >> > > Thanks,
> > > >> > > Brock
> > > >> > >
> > > >> > > ld: cannot find -lm
> > > >> > > make[4]: *** [ensemble_stat] Error 1
> > > >> > >
> > > >> > >
> > > >> > >
> > > >> > > Brock Burghardt
> > > >> > > Atmospheric Science PhD Student Dept. of Geosciences
Texas
> > > >> > > Tech University
> > > >> > >
> > > >> > >
> > > >> > >
> > > >> >
> > > >> >
> > > >> > --
> > > >> > Julie Prestopnik
> > > >> > National Center for Atmospheric Research Research
Applications
> > > >> > Laboratory
> > > >> > Phone: 303.497.8399
> > > >> > Email: jpresto at ucar.edu
> > > >> >
> > > >> >
> > > >> >
> > > >>
> > > >>
> > > >> --
> > > >> Julie Prestopnik
> > > >> National Center for Atmospheric Research Research
Applications
> > > Laboratory
> > > >> Phone: 303.497.8399
> > > >> Email: jpresto at ucar.edu
> > > >>
> > > >>
> > > >>
> > > >>
> > > >
> > > >
> > > > --
> > > > Julie Prestopnik
> > > > National Center for Atmospheric Research Research Applications
> > > > Laboratory
> > > > Phone: 303.497.8399
> > > > Email: jpresto at ucar.edu
> > > >
> > >
> > >
> > >
> > > --
> > > Julie Prestopnik
> > > National Center for Atmospheric Research Research Applications
> > > Laboratory
> > > Phone: 303.497.8399
> > > Email: jpresto at ucar.edu
> > >
> > >
> > >
> >
> >
> > --
> > Julie Prestopnik
> > National Center for Atmospheric Research Research Applications
> > Laboratory
> > Phone: 303.497.8399
> > Email: jpresto at ucar.edu
> >
> >
> >
> >
>
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research Research Applications
Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>
>
>

--
Julie Prestopnik
National Center for Atmospheric Research
Research Applications Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: Issue compiling MET V4.1
From: John Halley Gotway
Time: Tue Dec 09 10:32:43 2014

Hello Brock,

I work with Julie Prestopnik and I see that you're having trouble
compiling
MET.  I took a look at the log file you sent and see that you're using
the
intel c++ icpc compiler.  In the past when I've compiled MET using
intel
compilers, I've always used the "icc" c++ compiler.  I just tried
compiling
MET using icpc and run into all kinds of problems.  Is icc available
on
your system (type: which icc)?  If so, try using that by setting the
CXX
environment variable before running the "configure" command for MET:
  setenv CXX `which icc`

Also, I looked in the logfile you sent at the specific line where the
error
occurred and compared that to the same line in the output created when
I
compile MET.  None of the external library links are being passed on
the
compilation line... specifically between
"-I../../../src/libcode/vx_tc_util" and "-DBLOCK4" we should see links
for
NetCDF and other external library dependencies.

Not sure why this is happening but it may be related to the choice of
compiler.  Please let us know if you have icc available, and if so,
try
using that one.

Thanks,
John

On Mon, Dec 8, 2014 at 4:53 PM, Julie Prestopnik via RT
<met_help at ucar.edu>
wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi Brock.  I'm sorry to see that MET can't seem to find netcdf.hh.
Could
> you please set up the environment variables on the command line, and
try to
> build and install MET on the command line and then send me the
> make_install.log file from that run along with the output of
"printenv |
> grep MET".  I looked at this make_install.log, paying particular
attention
> to what is being looked at for the include files.  I don't know
where your
> netcdf include files are, so I can't check for that location.
>
> If your environment variable is set correctly, it should be listed
in the
> output of "printenv | grep MET" and also in -I in the
make_install.log
> file.  If MET still can't find it after we check those things, we'll
have
> to look into it further.  I know that certain recent versions of
NetCDF
> (e.g. v4.1.3) no longer include netcdf.hh, but I believe that 3.6.3,
which
> you said you are using, does have netcdf.hh.
>
> Sending this additional info from an install attempt with the
environment
> variables set, should help us better figure out what the problem is.
>
> Thanks,
> Julie
>
> On Mon, Dec 8, 2014 at 2:03 PM, Burghardt, Brock via RT
<met_help at ucar.edu
> >
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >
> > Hi Julie,
> > Sorry for not getting back to you sooner (finals just wrapped up).
> > I've attached the make log.
> > The printenv statement does not give anything for "MET", however
when I
> > set my MET_<LIB/INC> paths at command line I can print them,
giving the
> > proper paths.
> >
> > Thanks,
> > Brock
> >
> > -----Original Message-----
> > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > Sent: Friday, December 05, 2014 9:56 AM
> > To: Burghardt, Brock
> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >
> > Hi Brock.  Could you please send me your make_install.log file and
could
> > you please run "printenv | grep MET" again and send me the output?
> >
> > Thanks,
> > Julie
> >
> > On Thu, Dec 4, 2014 at 5:09 PM, Burghardt, Brock via RT <
> met_help at ucar.edu
> > >
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > >
> > > Hi Julie,
> > >
> > > Still no luck with loading the intel environment (soft add intel
in my
> > > case) with the error developing with netcdf.hh in the make
> > > install...any other ideas?
> > >
> > > Thanks, I appreciate the help.
> > >
> > > Brock
> > >
> > >
> > >
> > >
> > > ________________________________________
> > > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > > Sent: Tuesday, December 02, 2014 11:02 AM
> > > To: Burghardt, Brock
> > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > >
> > > Hi Brock.  I'm sorry to hear that you are still having problems
> > > compiling METv5.0.  Thank you for sending the config.log file.
> > >
> > > Please try running the following command:
> > >
> > > make clean
> > > module load intel
> > >
> > > and then run the configure command and the make install command.
> > >
> > > Please let me know if that works.
> > >
> > > Thanks,
> > > Julie
> > >
> > > On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
> > > <met_help at ucar.edu
> > > >
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> > > >
> > > > Hi Julie,
> > > >
> > > > I've still been unable to compile METv5.0...independent of
whether I
> > > > specify the parent or both the parent/lib parent/inc paths for
> > > > netcdf, bufrlib, and GSL, I get an error with it trying to
open
> > > > netcdf.hh
> > > >
> > > > I noticed two paths to netcdf got set in the config.log
(attached),
> > > > one I set and another on the system, removing the one I did
not
> > > > define still
> > > did
> > > > not give a successful build.
> > > >
> > > > I noticed I had errors in the config.log for many of the
compilers
> > > > where if tries to call, for instance, icpc -qversion which
should be
> > > > icpc
> > > -V....I
> > > > looked through all the config and configure scripts and could
not
> > > > find
> > > the
> > > > explicit icpc -qversion call to change to -V...do you think
this
> > > > could be the issue?
> > > >
> > > > Thank you for following up,
> > > > Brock
> > > >
> > > > ________________________________________
> > > > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > > > Sent: Monday, December 01, 2014 9:56 AM
> > > > To: Burghardt, Brock
> > > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > > >
> > > > Hi Brock.  I just wanted to follow up - Were able to get a
> > > > successful
> > > build
> > > > of METv5.0?
> > > >
> > > > Thanks,
> > > > Julie
> > > >
> > > > On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
> > > > <jpresto at ucar.edu>
> > > > wrote:
> > > >
> > > > > Hi Brock.
> > > > >
> > > > > Thanks you for sending the output.  I was looking to make
sure
> > > > > that the MET environment variables, like MET_NETCDF or
> > > > > MET_NETCDFINC and MET_NETCDFLIB, were set correctly.
> > > > >
> > > > > Please take a look at the following page and follow the
> > > > > instructions on compiling MET:
> > > > >
> > > > >
> > > >
> > >
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/comp
> > > ilation/configure.php
> > > > >
> > > > > This will help guide you in setting up the required
environment
> > > variables
> > > > > and will hopefully resolve the problem you are experiencing
with
> > > > > MET
> > > not
> > > > > being able to find netcdf.hh.
> > > > >
> > > > > Please let me know if you have any questions, and please let
me
> > > > > know if you get a successful build after setting up the
environment
> > variables.
> > > > >
> > > > > Thanks,
> > > > > Julie
> > > > >
> > > > > On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > >>
> > > > >> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > > > >>
> > > > >> Hi Julie, the output for the command is,
> > > > >>
> > > > >> hrothgar:/met-5.0$ printenv | grep met
> > > > >> PWD=/home/brburgha/met-5.0
> > > > >>
> > > > >> ...unless you wanted this,
> > > > >>
> > > > >> hrothgar:/met-5.0$ printenv | grep MET*
> > > > >> HOSTNAME=hrothgar.hpcc.ttu.edu ANT_HOME=/opt/rocks
> > > > >> JAVA_HOME=/usr/java/latest
> > > > >> P4_GLOBMEMSIZE=83886080
> > > > >> MAGICK_HOME=/lustre/work/apps/ImageMagick
> > > > >> HOME=/home/brburgha
> > > > >> LOGNAME=brburgha
> > > > >> G_BROKEN_FILENAMES=1
> > > > >>
> > > > >>
> > > > >> Brock
> > > > >>
> > > > >>
> > > > >> -----Original Message-----
> > > > >> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > > > >> Sent: Tuesday, November 18, 2014 10:38 AM
> > > > >> To: Burghardt, Brock
> > > > >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > > > >>
> > > > >> Hi Brock.  Thank you for getting the latest version of MET.
I
> > > > >> see
> > > > you're
> > > > >> still problems with MET finding netcdf.hh.
> > > > >>
> > > > >> Could you please run "printenv | grep MET" and send me the
output?
> > > > >>
> > > > >> Thanks,
> > > > >> Julie
> > > > >>
> > > > >> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
> > > > >> met_help at ucar.edu>
> > > > >> wrote:
> > > > >>
> > > > >> >
> > > > >> > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > > > >> >
> > > > >> > Hi Julie,
> > > > >> >
> > > > >> > Thanks for getting back to me so soon and explaining.
Removing
> > > > >> > the static flag did indeed help in compiling the
executables I
> > > > >> > needed
> > > but
> > > > >> > it did not completely compile (another ld: missing
library
> > > > >> > error
> > > came
> > > > >> > up in the pb2nc build). However, I took your advice and
started
> > > > >> > over with V5.0 and got that running, but the compile also
> > > > >> > failed before making executables. The error was for not
being
> > > > >> > able to open
> > > netcdf.hh
> > > > >> > (make_install.log attached). I added all the met_<lib>
> > > > >> > arguments at command line and checked they pointed to
right
> > > > >> > directory, and the configure step ran without error. I
verified
> > > > >> > I do have read
> > > permission
> > > > >> for netcdf.hh, any ideas on this.
> > > > >> >
> > > > >> > Thanks,
> > > > >> > Brock
> > > > >> >
> > > > >> > -----Original Message-----
> > > > >> > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > > > >> > Sent: Monday, November 17, 2014 4:34 PM
> > > > >> > To: Burghardt, Brock
> > > > >> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > > > >> >
> > > > >> > Hi Brock.
> > > > >> >
> > > > >> > I see that you are having problems building METv4.1 with
an
> > > > >> > intel compiler.  Thank you for attaching your
user_defs.mk and
> > > > >> > make.log
> > > > >> files.
> > > > >> >
> > > > >> > First, I wanted to mention that METv5.0 was release in
early
> > > > >> > September, which does include bug fixes for METv4.1.  You
can
> > > > >> > view
> > > the
> > > > >> > release notes for METv5.0 here:
> > > > >> >
> > > > >> >
> > > > >> >
> > > >
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_rele
> > > > as
> > > > >> > e_notes.php
> > > > >> >
> > > > >> > For now, let's try to overcome this error with METv4.1:
> > > > >> >
> > > > >> > ld: cannot find -lm
> > > > >> >
> > > > >> > We've had this issue come up with some other users in the
past.
> > > > >> > By default, MET does static linking - meaning that it
links in
> > > > >> > the library files on which it depends at compile time.
The
> > > > >> > other type of linking is call dynamic linking - for which
the
> > > > >> > library files on
> > > which
> > > > >> > it depends are linked in at runtime. Other users have had
> > > > >> > success compiling MET with dynamic linking instead. Open
up and
> > > > >> > edit the top-level user_defs.mk file, by commenting out
or
> > > > >> > removing
> > > references
> > > > >> > to "-static". They show up on the CXX_FLAGS and FC_FLAGS
lines.
> > > > >> >
> > > > >> > Then try recompiling MET, being sure to execute a "make
clean"
> > > first.
> > > > >> >
> > > > >> > Thanks,
> > > > >> > Julie
> > > > >> >
> > > > >> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT
<
> > > > >> > met_help at ucar.edu>
> > > > >> > wrote:
> > > > >> >
> > > > >> > >
> > > > >> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> > > > >> > > Transaction: Ticket created by brock.burghardt at ttu.edu
> > > > >> > >        Queue: met_help
> > > > >> > >      Subject: Issue compiling MET V4.1
> > > > >> > >        Owner: Nobody
> > > > >> > >   Requestors: brock.burghardt at ttu.edu
> > > > >> > >       Status: new
> > > > >> > >  Ticket <URL:
> > > > >> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > > > >> > > >
> > > > >> > >
> > > > >> > >
> > > > >> > > Hello,
> > > > >> > >
> > > > >> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3 with
an
> > > > >> > > intel compiler (definitions and log file attachted).
I've
> > > > >> > > gone through
> > > all
> > > > >> > > suggestions in the appendix of the users guide but
still get
> > > > >> > > an error while trying to make the ensemble_stat
executable.
> > > > >> > > Is there
> > > a
> > > > >> > > liblm.so or .a required for this build?
> > > > >> > >
> > > > >> > > Thanks,
> > > > >> > > Brock
> > > > >> > >
> > > > >> > > ld: cannot find -lm
> > > > >> > > make[4]: *** [ensemble_stat] Error 1
> > > > >> > >
> > > > >> > >
> > > > >> > >
> > > > >> > > Brock Burghardt
> > > > >> > > Atmospheric Science PhD Student Dept. of Geosciences
Texas
> > > > >> > > Tech University
> > > > >> > >
> > > > >> > >
> > > > >> > >
> > > > >> >
> > > > >> >
> > > > >> > --
> > > > >> > Julie Prestopnik
> > > > >> > National Center for Atmospheric Research Research
Applications
> > > > >> > Laboratory
> > > > >> > Phone: 303.497.8399
> > > > >> > Email: jpresto at ucar.edu
> > > > >> >
> > > > >> >
> > > > >> >
> > > > >>
> > > > >>
> > > > >> --
> > > > >> Julie Prestopnik
> > > > >> National Center for Atmospheric Research Research
Applications
> > > > Laboratory
> > > > >> Phone: 303.497.8399
> > > > >> Email: jpresto at ucar.edu
> > > > >>
> > > > >>
> > > > >>
> > > > >>
> > > > >
> > > > >
> > > > > --
> > > > > Julie Prestopnik
> > > > > National Center for Atmospheric Research Research
Applications
> > > > > Laboratory
> > > > > Phone: 303.497.8399
> > > > > Email: jpresto at ucar.edu
> > > > >
> > > >
> > > >
> > > >
> > > > --
> > > > Julie Prestopnik
> > > > National Center for Atmospheric Research Research Applications
> > > > Laboratory
> > > > Phone: 303.497.8399
> > > > Email: jpresto at ucar.edu
> > > >
> > > >
> > > >
> > >
> > >
> > > --
> > > Julie Prestopnik
> > > National Center for Atmospheric Research Research Applications
> > > Laboratory
> > > Phone: 303.497.8399
> > > Email: jpresto at ucar.edu
> > >
> > >
> > >
> > >
> >
> >
> > --
> > Julie Prestopnik
> > National Center for Atmospheric Research Research Applications
Laboratory
> > Phone: 303.497.8399
> > Email: jpresto at ucar.edu
> >
> >
> >
>
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research
> Research Applications Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>
>

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Burghardt, Brock
Time: Thu Dec 11 12:21:51 2014

Hi John,

So I set the CXX to the available intel c++ compiler, however I am
still receiving the same error set.
Below are the other paths I set before configure and make:
MET_NETCDF=${NETCDF}
MET_BUFR=$BUFRLIB
MET_GSL=/lustre/work/apps/gsl
CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort

Attached are both the config and make logs. I noticed that in my
config log that when it is looking for the icc version it gets an
error because it is invoking 'icc -qversion' but should be 'icc -v' if
I want to check it on the command line. Do you think this could be a
cause?

Thank you,
Brock

________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Tuesday, December 09, 2014 11:32 AM
To: Burghardt, Brock
Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1

Hello Brock,

I work with Julie Prestopnik and I see that you're having trouble
compiling
MET.  I took a look at the log file you sent and see that you're using
the
intel c++ icpc compiler.  In the past when I've compiled MET using
intel
compilers, I've always used the "icc" c++ compiler.  I just tried
compiling
MET using icpc and run into all kinds of problems.  Is icc available
on
your system (type: which icc)?  If so, try using that by setting the
CXX
environment variable before running the "configure" command for MET:
  setenv CXX `which icc`

Also, I looked in the logfile you sent at the specific line where the
error
occurred and compared that to the same line in the output created when
I
compile MET.  None of the external library links are being passed on
the
compilation line... specifically between
"-I../../../src/libcode/vx_tc_util" and "-DBLOCK4" we should see links
for
NetCDF and other external library dependencies.

Not sure why this is happening but it may be related to the choice of
compiler.  Please let us know if you have icc available, and if so,
try
using that one.

Thanks,
John

On Mon, Dec 8, 2014 at 4:53 PM, Julie Prestopnik via RT
<met_help at ucar.edu>
wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi Brock.  I'm sorry to see that MET can't seem to find netcdf.hh.
Could
> you please set up the environment variables on the command line, and
try to
> build and install MET on the command line and then send me the
> make_install.log file from that run along with the output of
"printenv |
> grep MET".  I looked at this make_install.log, paying particular
attention
> to what is being looked at for the include files.  I don't know
where your
> netcdf include files are, so I can't check for that location.
>
> If your environment variable is set correctly, it should be listed
in the
> output of "printenv | grep MET" and also in -I in the
make_install.log
> file.  If MET still can't find it after we check those things, we'll
have
> to look into it further.  I know that certain recent versions of
NetCDF
> (e.g. v4.1.3) no longer include netcdf.hh, but I believe that 3.6.3,
which
> you said you are using, does have netcdf.hh.
>
> Sending this additional info from an install attempt with the
environment
> variables set, should help us better figure out what the problem is.
>
> Thanks,
> Julie
>
> On Mon, Dec 8, 2014 at 2:03 PM, Burghardt, Brock via RT
<met_help at ucar.edu
> >
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >
> > Hi Julie,
> > Sorry for not getting back to you sooner (finals just wrapped up).
> > I've attached the make log.
> > The printenv statement does not give anything for "MET", however
when I
> > set my MET_<LIB/INC> paths at command line I can print them,
giving the
> > proper paths.
> >
> > Thanks,
> > Brock
> >
> > -----Original Message-----
> > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > Sent: Friday, December 05, 2014 9:56 AM
> > To: Burghardt, Brock
> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >
> > Hi Brock.  Could you please send me your make_install.log file and
could
> > you please run "printenv | grep MET" again and send me the output?
> >
> > Thanks,
> > Julie
> >
> > On Thu, Dec 4, 2014 at 5:09 PM, Burghardt, Brock via RT <
> met_help at ucar.edu
> > >
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > >
> > > Hi Julie,
> > >
> > > Still no luck with loading the intel environment (soft add intel
in my
> > > case) with the error developing with netcdf.hh in the make
> > > install...any other ideas?
> > >
> > > Thanks, I appreciate the help.
> > >
> > > Brock
> > >
> > >
> > >
> > >
> > > ________________________________________
> > > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > > Sent: Tuesday, December 02, 2014 11:02 AM
> > > To: Burghardt, Brock
> > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > >
> > > Hi Brock.  I'm sorry to hear that you are still having problems
> > > compiling METv5.0.  Thank you for sending the config.log file.
> > >
> > > Please try running the following command:
> > >
> > > make clean
> > > module load intel
> > >
> > > and then run the configure command and the make install command.
> > >
> > > Please let me know if that works.
> > >
> > > Thanks,
> > > Julie
> > >
> > > On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
> > > <met_help at ucar.edu
> > > >
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> > > >
> > > > Hi Julie,
> > > >
> > > > I've still been unable to compile METv5.0...independent of
whether I
> > > > specify the parent or both the parent/lib parent/inc paths for
> > > > netcdf, bufrlib, and GSL, I get an error with it trying to
open
> > > > netcdf.hh
> > > >
> > > > I noticed two paths to netcdf got set in the config.log
(attached),
> > > > one I set and another on the system, removing the one I did
not
> > > > define still
> > > did
> > > > not give a successful build.
> > > >
> > > > I noticed I had errors in the config.log for many of the
compilers
> > > > where if tries to call, for instance, icpc -qversion which
should be
> > > > icpc
> > > -V....I
> > > > looked through all the config and configure scripts and could
not
> > > > find
> > > the
> > > > explicit icpc -qversion call to change to -V...do you think
this
> > > > could be the issue?
> > > >
> > > > Thank you for following up,
> > > > Brock
> > > >
> > > > ________________________________________
> > > > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > > > Sent: Monday, December 01, 2014 9:56 AM
> > > > To: Burghardt, Brock
> > > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > > >
> > > > Hi Brock.  I just wanted to follow up - Were able to get a
> > > > successful
> > > build
> > > > of METv5.0?
> > > >
> > > > Thanks,
> > > > Julie
> > > >
> > > > On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
> > > > <jpresto at ucar.edu>
> > > > wrote:
> > > >
> > > > > Hi Brock.
> > > > >
> > > > > Thanks you for sending the output.  I was looking to make
sure
> > > > > that the MET environment variables, like MET_NETCDF or
> > > > > MET_NETCDFINC and MET_NETCDFLIB, were set correctly.
> > > > >
> > > > > Please take a look at the following page and follow the
> > > > > instructions on compiling MET:
> > > > >
> > > > >
> > > >
> > >
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/comp
> > > ilation/configure.php
> > > > >
> > > > > This will help guide you in setting up the required
environment
> > > variables
> > > > > and will hopefully resolve the problem you are experiencing
with
> > > > > MET
> > > not
> > > > > being able to find netcdf.hh.
> > > > >
> > > > > Please let me know if you have any questions, and please let
me
> > > > > know if you get a successful build after setting up the
environment
> > variables.
> > > > >
> > > > > Thanks,
> > > > > Julie
> > > > >
> > > > > On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > >>
> > > > >> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > > > >>
> > > > >> Hi Julie, the output for the command is,
> > > > >>
> > > > >> hrothgar:/met-5.0$ printenv | grep met
> > > > >> PWD=/home/brburgha/met-5.0
> > > > >>
> > > > >> ...unless you wanted this,
> > > > >>
> > > > >> hrothgar:/met-5.0$ printenv | grep MET*
> > > > >> HOSTNAME=hrothgar.hpcc.ttu.edu ANT_HOME=/opt/rocks
> > > > >> JAVA_HOME=/usr/java/latest
> > > > >> P4_GLOBMEMSIZE=83886080
> > > > >> MAGICK_HOME=/lustre/work/apps/ImageMagick
> > > > >> HOME=/home/brburgha
> > > > >> LOGNAME=brburgha
> > > > >> G_BROKEN_FILENAMES=1
> > > > >>
> > > > >>
> > > > >> Brock
> > > > >>
> > > > >>
> > > > >> -----Original Message-----
> > > > >> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > > > >> Sent: Tuesday, November 18, 2014 10:38 AM
> > > > >> To: Burghardt, Brock
> > > > >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > > > >>
> > > > >> Hi Brock.  Thank you for getting the latest version of MET.
I
> > > > >> see
> > > > you're
> > > > >> still problems with MET finding netcdf.hh.
> > > > >>
> > > > >> Could you please run "printenv | grep MET" and send me the
output?
> > > > >>
> > > > >> Thanks,
> > > > >> Julie
> > > > >>
> > > > >> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
> > > > >> met_help at ucar.edu>
> > > > >> wrote:
> > > > >>
> > > > >> >
> > > > >> > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > > > >> >
> > > > >> > Hi Julie,
> > > > >> >
> > > > >> > Thanks for getting back to me so soon and explaining.
Removing
> > > > >> > the static flag did indeed help in compiling the
executables I
> > > > >> > needed
> > > but
> > > > >> > it did not completely compile (another ld: missing
library
> > > > >> > error
> > > came
> > > > >> > up in the pb2nc build). However, I took your advice and
started
> > > > >> > over with V5.0 and got that running, but the compile also
> > > > >> > failed before making executables. The error was for not
being
> > > > >> > able to open
> > > netcdf.hh
> > > > >> > (make_install.log attached). I added all the met_<lib>
> > > > >> > arguments at command line and checked they pointed to
right
> > > > >> > directory, and the configure step ran without error. I
verified
> > > > >> > I do have read
> > > permission
> > > > >> for netcdf.hh, any ideas on this.
> > > > >> >
> > > > >> > Thanks,
> > > > >> > Brock
> > > > >> >
> > > > >> > -----Original Message-----
> > > > >> > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > > > >> > Sent: Monday, November 17, 2014 4:34 PM
> > > > >> > To: Burghardt, Brock
> > > > >> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > > > >> >
> > > > >> > Hi Brock.
> > > > >> >
> > > > >> > I see that you are having problems building METv4.1 with
an
> > > > >> > intel compiler.  Thank you for attaching your
user_defs.mk and
> > > > >> > make.log
> > > > >> files.
> > > > >> >
> > > > >> > First, I wanted to mention that METv5.0 was release in
early
> > > > >> > September, which does include bug fixes for METv4.1.  You
can
> > > > >> > view
> > > the
> > > > >> > release notes for METv5.0 here:
> > > > >> >
> > > > >> >
> > > > >> >
> > > >
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_rele
> > > > as
> > > > >> > e_notes.php
> > > > >> >
> > > > >> > For now, let's try to overcome this error with METv4.1:
> > > > >> >
> > > > >> > ld: cannot find -lm
> > > > >> >
> > > > >> > We've had this issue come up with some other users in the
past.
> > > > >> > By default, MET does static linking - meaning that it
links in
> > > > >> > the library files on which it depends at compile time.
The
> > > > >> > other type of linking is call dynamic linking - for which
the
> > > > >> > library files on
> > > which
> > > > >> > it depends are linked in at runtime. Other users have had
> > > > >> > success compiling MET with dynamic linking instead. Open
up and
> > > > >> > edit the top-level user_defs.mk file, by commenting out
or
> > > > >> > removing
> > > references
> > > > >> > to "-static". They show up on the CXX_FLAGS and FC_FLAGS
lines.
> > > > >> >
> > > > >> > Then try recompiling MET, being sure to execute a "make
clean"
> > > first.
> > > > >> >
> > > > >> > Thanks,
> > > > >> > Julie
> > > > >> >
> > > > >> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT
<
> > > > >> > met_help at ucar.edu>
> > > > >> > wrote:
> > > > >> >
> > > > >> > >
> > > > >> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> > > > >> > > Transaction: Ticket created by brock.burghardt at ttu.edu
> > > > >> > >        Queue: met_help
> > > > >> > >      Subject: Issue compiling MET V4.1
> > > > >> > >        Owner: Nobody
> > > > >> > >   Requestors: brock.burghardt at ttu.edu
> > > > >> > >       Status: new
> > > > >> > >  Ticket <URL:
> > > > >> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > > > >> > > >
> > > > >> > >
> > > > >> > >
> > > > >> > > Hello,
> > > > >> > >
> > > > >> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3 with
an
> > > > >> > > intel compiler (definitions and log file attachted).
I've
> > > > >> > > gone through
> > > all
> > > > >> > > suggestions in the appendix of the users guide but
still get
> > > > >> > > an error while trying to make the ensemble_stat
executable.
> > > > >> > > Is there
> > > a
> > > > >> > > liblm.so or .a required for this build?
> > > > >> > >
> > > > >> > > Thanks,
> > > > >> > > Brock
> > > > >> > >
> > > > >> > > ld: cannot find -lm
> > > > >> > > make[4]: *** [ensemble_stat] Error 1
> > > > >> > >
> > > > >> > >
> > > > >> > >
> > > > >> > > Brock Burghardt
> > > > >> > > Atmospheric Science PhD Student Dept. of Geosciences
Texas
> > > > >> > > Tech University
> > > > >> > >
> > > > >> > >
> > > > >> > >
> > > > >> >
> > > > >> >
> > > > >> > --
> > > > >> > Julie Prestopnik
> > > > >> > National Center for Atmospheric Research Research
Applications
> > > > >> > Laboratory
> > > > >> > Phone: 303.497.8399
> > > > >> > Email: jpresto at ucar.edu
> > > > >> >
> > > > >> >
> > > > >> >
> > > > >>
> > > > >>
> > > > >> --
> > > > >> Julie Prestopnik
> > > > >> National Center for Atmospheric Research Research
Applications
> > > > Laboratory
> > > > >> Phone: 303.497.8399
> > > > >> Email: jpresto at ucar.edu
> > > > >>
> > > > >>
> > > > >>
> > > > >>
> > > > >
> > > > >
> > > > > --
> > > > > Julie Prestopnik
> > > > > National Center for Atmospheric Research Research
Applications
> > > > > Laboratory
> > > > > Phone: 303.497.8399
> > > > > Email: jpresto at ucar.edu
> > > > >
> > > >
> > > >
> > > >
> > > > --
> > > > Julie Prestopnik
> > > > National Center for Atmospheric Research Research Applications
> > > > Laboratory
> > > > Phone: 303.497.8399
> > > > Email: jpresto at ucar.edu
> > > >
> > > >
> > > >
> > >
> > >
> > > --
> > > Julie Prestopnik
> > > National Center for Atmospheric Research Research Applications
> > > Laboratory
> > > Phone: 303.497.8399
> > > Email: jpresto at ucar.edu
> > >
> > >
> > >
> > >
> >
> >
> > --
> > Julie Prestopnik
> > National Center for Atmospheric Research Research Applications
Laboratory
> > Phone: 303.497.8399
> > Email: jpresto at ucar.edu
> >
> >
> >
>
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research
> Research Applications Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>
>

------------------------------------------------
Subject: Issue compiling MET V4.1
From: John Halley Gotway
Time: Fri Dec 12 15:11:51 2014

Brock,

Thanks for sending those log files.  Looking at line 3243 of
config.log, I
see:
   ac_cv_env_MET_NETCDF_set=
   ac_cv_env_MET_NETCDF_value=

This means that configure doesn't think that MET_NETCDF is set.
Looking at
the syntax of the commands you sent, I assume you're using the bash or
ksh
shell, rather than cshell.  That's fine, but you need to use "export"
when
setting variables to pass those definitions down to sub-processes.
Please
try the following:

export MET_NETCDF=${NETCDF}
export MET_BUFR=$BUFRLIB
export MET_GSL=/lustre/work/apps/gsl
export
CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
export
F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort

And then run "configure".  Then run the following command and make
sure
that NetCDF is actually set:
   grep ac_cv_env_MET_NETCDF_value= config.log

If so, proceed to "make install" and "make test".

Sorry for all of the headaches you've had to struggle through!

John

On Thu, Dec 11, 2014 at 12:21 PM, Burghardt, Brock via RT
<met_help at ucar.edu
> wrote:
>
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi John,
>
> So I set the CXX to the available intel c++ compiler, however I am
still
> receiving the same error set.
> Below are the other paths I set before configure and make:
> MET_NETCDF=${NETCDF}
> MET_BUFR=$BUFRLIB
> MET_GSL=/lustre/work/apps/gsl
> CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
> F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort
>
> Attached are both the config and make logs. I noticed that in my
config
> log that when it is looking for the icc version it gets an error
because it
> is invoking 'icc -qversion' but should be 'icc -v' if I want to
check it on
> the command line. Do you think this could be a cause?
>
> Thank you,
> Brock
>
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Tuesday, December 09, 2014 11:32 AM
> To: Burghardt, Brock
> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>
> Hello Brock,
>
> I work with Julie Prestopnik and I see that you're having trouble
compiling
> MET.  I took a look at the log file you sent and see that you're
using the
> intel c++ icpc compiler.  In the past when I've compiled MET using
intel
> compilers, I've always used the "icc" c++ compiler.  I just tried
compiling
> MET using icpc and run into all kinds of problems.  Is icc available
on
> your system (type: which icc)?  If so, try using that by setting the
CXX
> environment variable before running the "configure" command for MET:
>   setenv CXX `which icc`
>
> Also, I looked in the logfile you sent at the specific line where
the error
> occurred and compared that to the same line in the output created
when I
> compile MET.  None of the external library links are being passed on
the
> compilation line... specifically between
> "-I../../../src/libcode/vx_tc_util" and "-DBLOCK4" we should see
links for
> NetCDF and other external library dependencies.
>
> Not sure why this is happening but it may be related to the choice
of
> compiler.  Please let us know if you have icc available, and if so,
try
> using that one.
>
> Thanks,
> John
>
> On Mon, Dec 8, 2014 at 4:53 PM, Julie Prestopnik via RT
<met_help at ucar.edu
> >
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >
> > Hi Brock.  I'm sorry to see that MET can't seem to find netcdf.hh.
Could
> > you please set up the environment variables on the command line,
and try
> to
> > build and install MET on the command line and then send me the
> > make_install.log file from that run along with the output of
"printenv |
> > grep MET".  I looked at this make_install.log, paying particular
> attention
> > to what is being looked at for the include files.  I don't know
where
> your
> > netcdf include files are, so I can't check for that location.
> >
> > If your environment variable is set correctly, it should be listed
in the
> > output of "printenv | grep MET" and also in -I in the
make_install.log
> > file.  If MET still can't find it after we check those things,
we'll have
> > to look into it further.  I know that certain recent versions of
NetCDF
> > (e.g. v4.1.3) no longer include netcdf.hh, but I believe that
3.6.3,
> which
> > you said you are using, does have netcdf.hh.
> >
> > Sending this additional info from an install attempt with the
environment
> > variables set, should help us better figure out what the problem
is.
> >
> > Thanks,
> > Julie
> >
> > On Mon, Dec 8, 2014 at 2:03 PM, Burghardt, Brock via RT <
> met_help at ucar.edu
> > >
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > >
> > > Hi Julie,
> > > Sorry for not getting back to you sooner (finals just wrapped
up).
> > > I've attached the make log.
> > > The printenv statement does not give anything for "MET", however
when I
> > > set my MET_<LIB/INC> paths at command line I can print them,
giving the
> > > proper paths.
> > >
> > > Thanks,
> > > Brock
> > >
> > > -----Original Message-----
> > > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > > Sent: Friday, December 05, 2014 9:56 AM
> > > To: Burghardt, Brock
> > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > >
> > > Hi Brock.  Could you please send me your make_install.log file
and
> could
> > > you please run "printenv | grep MET" again and send me the
output?
> > >
> > > Thanks,
> > > Julie
> > >
> > > On Thu, Dec 4, 2014 at 5:09 PM, Burghardt, Brock via RT <
> > met_help at ucar.edu
> > > >
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> > > >
> > > > Hi Julie,
> > > >
> > > > Still no luck with loading the intel environment (soft add
intel in
> my
> > > > case) with the error developing with netcdf.hh in the make
> > > > install...any other ideas?
> > > >
> > > > Thanks, I appreciate the help.
> > > >
> > > > Brock
> > > >
> > > >
> > > >
> > > >
> > > > ________________________________________
> > > > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > > > Sent: Tuesday, December 02, 2014 11:02 AM
> > > > To: Burghardt, Brock
> > > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > > >
> > > > Hi Brock.  I'm sorry to hear that you are still having
problems
> > > > compiling METv5.0.  Thank you for sending the config.log file.
> > > >
> > > > Please try running the following command:
> > > >
> > > > make clean
> > > > module load intel
> > > >
> > > > and then run the configure command and the make install
command.
> > > >
> > > > Please let me know if that works.
> > > >
> > > > Thanks,
> > > > Julie
> > > >
> > > > On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
> > > > <met_help at ucar.edu
> > > > >
> > > > wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > > > >
> > > > > Hi Julie,
> > > > >
> > > > > I've still been unable to compile METv5.0...independent of
whether
> I
> > > > > specify the parent or both the parent/lib parent/inc paths
for
> > > > > netcdf, bufrlib, and GSL, I get an error with it trying to
open
> > > > > netcdf.hh
> > > > >
> > > > > I noticed two paths to netcdf got set in the config.log
(attached),
> > > > > one I set and another on the system, removing the one I did
not
> > > > > define still
> > > > did
> > > > > not give a successful build.
> > > > >
> > > > > I noticed I had errors in the config.log for many of the
compilers
> > > > > where if tries to call, for instance, icpc -qversion which
should
> be
> > > > > icpc
> > > > -V....I
> > > > > looked through all the config and configure scripts and
could not
> > > > > find
> > > > the
> > > > > explicit icpc -qversion call to change to -V...do you think
this
> > > > > could be the issue?
> > > > >
> > > > > Thank you for following up,
> > > > > Brock
> > > > >
> > > > > ________________________________________
> > > > > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > > > > Sent: Monday, December 01, 2014 9:56 AM
> > > > > To: Burghardt, Brock
> > > > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > > > >
> > > > > Hi Brock.  I just wanted to follow up - Were able to get a
> > > > > successful
> > > > build
> > > > > of METv5.0?
> > > > >
> > > > > Thanks,
> > > > > Julie
> > > > >
> > > > > On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
> > > > > <jpresto at ucar.edu>
> > > > > wrote:
> > > > >
> > > > > > Hi Brock.
> > > > > >
> > > > > > Thanks you for sending the output.  I was looking to make
sure
> > > > > > that the MET environment variables, like MET_NETCDF or
> > > > > > MET_NETCDFINC and MET_NETCDFLIB, were set correctly.
> > > > > >
> > > > > > Please take a look at the following page and follow the
> > > > > > instructions on compiling MET:
> > > > > >
> > > > > >
> > > > >
> > > >
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/comp
> > > > ilation/configure.php
> > > > > >
> > > > > > This will help guide you in setting up the required
environment
> > > > variables
> > > > > > and will hopefully resolve the problem you are
experiencing with
> > > > > > MET
> > > > not
> > > > > > being able to find netcdf.hh.
> > > > > >
> > > > > > Please let me know if you have any questions, and please
let me
> > > > > > know if you get a successful build after setting up the
> environment
> > > variables.
> > > > > >
> > > > > > Thanks,
> > > > > > Julie
> > > > > >
> > > > > > On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> > > > > > met_help at ucar.edu> wrote:
> > > > > >
> > > > > >>
> > > > > >> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > > > > >>
> > > > > >> Hi Julie, the output for the command is,
> > > > > >>
> > > > > >> hrothgar:/met-5.0$ printenv | grep met
> > > > > >> PWD=/home/brburgha/met-5.0
> > > > > >>
> > > > > >> ...unless you wanted this,
> > > > > >>
> > > > > >> hrothgar:/met-5.0$ printenv | grep MET*
> > > > > >> HOSTNAME=hrothgar.hpcc.ttu.edu ANT_HOME=/opt/rocks
> > > > > >> JAVA_HOME=/usr/java/latest
> > > > > >> P4_GLOBMEMSIZE=83886080
> > > > > >> MAGICK_HOME=/lustre/work/apps/ImageMagick
> > > > > >> HOME=/home/brburgha
> > > > > >> LOGNAME=brburgha
> > > > > >> G_BROKEN_FILENAMES=1
> > > > > >>
> > > > > >>
> > > > > >> Brock
> > > > > >>
> > > > > >>
> > > > > >> -----Original Message-----
> > > > > >> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > > > > >> Sent: Tuesday, November 18, 2014 10:38 AM
> > > > > >> To: Burghardt, Brock
> > > > > >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > > > > >>
> > > > > >> Hi Brock.  Thank you for getting the latest version of
MET.  I
> > > > > >> see
> > > > > you're
> > > > > >> still problems with MET finding netcdf.hh.
> > > > > >>
> > > > > >> Could you please run "printenv | grep MET" and send me
the
> output?
> > > > > >>
> > > > > >> Thanks,
> > > > > >> Julie
> > > > > >>
> > > > > >> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT
<
> > > > > >> met_help at ucar.edu>
> > > > > >> wrote:
> > > > > >>
> > > > > >> >
> > > > > >> > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> >
> > > > > >> >
> > > > > >> > Hi Julie,
> > > > > >> >
> > > > > >> > Thanks for getting back to me so soon and explaining.
Removing
> > > > > >> > the static flag did indeed help in compiling the
executables I
> > > > > >> > needed
> > > > but
> > > > > >> > it did not completely compile (another ld: missing
library
> > > > > >> > error
> > > > came
> > > > > >> > up in the pb2nc build). However, I took your advice and
> started
> > > > > >> > over with V5.0 and got that running, but the compile
also
> > > > > >> > failed before making executables. The error was for not
being
> > > > > >> > able to open
> > > > netcdf.hh
> > > > > >> > (make_install.log attached). I added all the met_<lib>
> > > > > >> > arguments at command line and checked they pointed to
right
> > > > > >> > directory, and the configure step ran without error. I
> verified
> > > > > >> > I do have read
> > > > permission
> > > > > >> for netcdf.hh, any ideas on this.
> > > > > >> >
> > > > > >> > Thanks,
> > > > > >> > Brock
> > > > > >> >
> > > > > >> > -----Original Message-----
> > > > > >> > From: Julie Prestopnik via RT
[mailto:met_help at ucar.edu]
> > > > > >> > Sent: Monday, November 17, 2014 4:34 PM
> > > > > >> > To: Burghardt, Brock
> > > > > >> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling
MET
> V4.1
> > > > > >> >
> > > > > >> > Hi Brock.
> > > > > >> >
> > > > > >> > I see that you are having problems building METv4.1
with an
> > > > > >> > intel compiler.  Thank you for attaching your
user_defs.mk
> and
> > > > > >> > make.log
> > > > > >> files.
> > > > > >> >
> > > > > >> > First, I wanted to mention that METv5.0 was release in
early
> > > > > >> > September, which does include bug fixes for METv4.1.
You can
> > > > > >> > view
> > > > the
> > > > > >> > release notes for METv5.0 here:
> > > > > >> >
> > > > > >> >
> > > > > >> >
> > > > >
> http://www.dtcenter.org/met/users/support/release_notes/METv5.0_rele
> > > > > as
> > > > > >> > e_notes.php
> > > > > >> >
> > > > > >> > For now, let's try to overcome this error with METv4.1:
> > > > > >> >
> > > > > >> > ld: cannot find -lm
> > > > > >> >
> > > > > >> > We've had this issue come up with some other users in
the
> past.
> > > > > >> > By default, MET does static linking - meaning that it
links in
> > > > > >> > the library files on which it depends at compile time.
The
> > > > > >> > other type of linking is call dynamic linking - for
which the
> > > > > >> > library files on
> > > > which
> > > > > >> > it depends are linked in at runtime. Other users have
had
> > > > > >> > success compiling MET with dynamic linking instead.
Open up
> and
> > > > > >> > edit the top-level user_defs.mk file, by commenting out
or
> > > > > >> > removing
> > > > references
> > > > > >> > to "-static". They show up on the CXX_FLAGS and
FC_FLAGS
> lines.
> > > > > >> >
> > > > > >> > Then try recompiling MET, being sure to execute a "make
clean"
> > > > first.
> > > > > >> >
> > > > > >> > Thanks,
> > > > > >> > Julie
> > > > > >> >
> > > > > >> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via
RT <
> > > > > >> > met_help at ucar.edu>
> > > > > >> > wrote:
> > > > > >> >
> > > > > >> > >
> > > > > >> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted
upon.
> > > > > >> > > Transaction: Ticket created by
brock.burghardt at ttu.edu
> > > > > >> > >        Queue: met_help
> > > > > >> > >      Subject: Issue compiling MET V4.1
> > > > > >> > >        Owner: Nobody
> > > > > >> > >   Requestors: brock.burghardt at ttu.edu
> > > > > >> > >       Status: new
> > > > > >> > >  Ticket <URL:
> > > > > >> > >
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > > > > >> > > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> > > Hello,
> > > > > >> > >
> > > > > >> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3
with an
> > > > > >> > > intel compiler (definitions and log file attachted).
I've
> > > > > >> > > gone through
> > > > all
> > > > > >> > > suggestions in the appendix of the users guide but
still get
> > > > > >> > > an error while trying to make the ensemble_stat
executable.
> > > > > >> > > Is there
> > > > a
> > > > > >> > > liblm.so or .a required for this build?
> > > > > >> > >
> > > > > >> > > Thanks,
> > > > > >> > > Brock
> > > > > >> > >
> > > > > >> > > ld: cannot find -lm
> > > > > >> > > make[4]: *** [ensemble_stat] Error 1
> > > > > >> > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> > > Brock Burghardt
> > > > > >> > > Atmospheric Science PhD Student Dept. of Geosciences
Texas
> > > > > >> > > Tech University
> > > > > >> > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> >
> > > > > >> >
> > > > > >> > --
> > > > > >> > Julie Prestopnik
> > > > > >> > National Center for Atmospheric Research Research
Applications
> > > > > >> > Laboratory
> > > > > >> > Phone: 303.497.8399
> > > > > >> > Email: jpresto at ucar.edu
> > > > > >> >
> > > > > >> >
> > > > > >> >
> > > > > >>
> > > > > >>
> > > > > >> --
> > > > > >> Julie Prestopnik
> > > > > >> National Center for Atmospheric Research Research
Applications
> > > > > Laboratory
> > > > > >> Phone: 303.497.8399
> > > > > >> Email: jpresto at ucar.edu
> > > > > >>
> > > > > >>
> > > > > >>
> > > > > >>
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Julie Prestopnik
> > > > > > National Center for Atmospheric Research Research
Applications
> > > > > > Laboratory
> > > > > > Phone: 303.497.8399
> > > > > > Email: jpresto at ucar.edu
> > > > > >
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Julie Prestopnik
> > > > > National Center for Atmospheric Research Research
Applications
> > > > > Laboratory
> > > > > Phone: 303.497.8399
> > > > > Email: jpresto at ucar.edu
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > --
> > > > Julie Prestopnik
> > > > National Center for Atmospheric Research Research Applications
> > > > Laboratory
> > > > Phone: 303.497.8399
> > > > Email: jpresto at ucar.edu
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > > --
> > > Julie Prestopnik
> > > National Center for Atmospheric Research Research Applications
> Laboratory
> > > Phone: 303.497.8399
> > > Email: jpresto at ucar.edu
> > >
> > >
> > >
> >
> >
> > --
> > Julie Prestopnik
> > National Center for Atmospheric Research
> > Research Applications Laboratory
> > Phone: 303.497.8399
> > Email: jpresto at ucar.edu
> >
> >
>
>
>

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Burghardt, Brock
Time: Mon Dec 15 14:10:49 2014

Hi John,

I completely forgot to export the set paths, sorry about that. Upon
exporting the paths I get a partially successful build with 11 of the
23 executables created.
The error appears to occur when looking for the -lbufr library. Would
this have something to do with the BUFRLIB directory?

Thanks,
Brock

-----Original Message-----
From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
Sent: Friday, December 12, 2014 4:12 PM
To: Burghardt, Brock
Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1

Brock,

Thanks for sending those log files.  Looking at line 3243 of
config.log, I
see:
   ac_cv_env_MET_NETCDF_set=
   ac_cv_env_MET_NETCDF_value=

This means that configure doesn't think that MET_NETCDF is set.
Looking at the syntax of the commands you sent, I assume you're using
the bash or ksh shell, rather than cshell.  That's fine, but you need
to use "export" when setting variables to pass those definitions down
to sub-processes.  Please try the following:

export MET_NETCDF=${NETCDF}
export MET_BUFR=$BUFRLIB
export MET_GSL=/lustre/work/apps/gsl
export
CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
export
F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort

And then run "configure".  Then run the following command and make
sure that NetCDF is actually set:
   grep ac_cv_env_MET_NETCDF_value= config.log

If so, proceed to "make install" and "make test".

Sorry for all of the headaches you've had to struggle through!

John

On Thu, Dec 11, 2014 at 12:21 PM, Burghardt, Brock via RT
<met_help at ucar.edu
> wrote:
>
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi John,
>
> So I set the CXX to the available intel c++ compiler, however I am
> still receiving the same error set.
> Below are the other paths I set before configure and make:
> MET_NETCDF=${NETCDF}
> MET_BUFR=$BUFRLIB
> MET_GSL=/lustre/work/apps/gsl
> CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
> F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort
>
> Attached are both the config and make logs. I noticed that in my
> config log that when it is looking for the icc version it gets an
> error because it is invoking 'icc -qversion' but should be 'icc -v'
if
> I want to check it on the command line. Do you think this could be a
cause?
>
> Thank you,
> Brock
>
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Tuesday, December 09, 2014 11:32 AM
> To: Burghardt, Brock
> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>
> Hello Brock,
>
> I work with Julie Prestopnik and I see that you're having trouble
> compiling MET.  I took a look at the log file you sent and see that
> you're using the intel c++ icpc compiler.  In the past when I've
> compiled MET using intel compilers, I've always used the "icc" c++
> compiler.  I just tried compiling MET using icpc and run into all
> kinds of problems.  Is icc available on your system (type: which
icc)?
> If so, try using that by setting the CXX environment variable before
running the "configure" command for MET:
>   setenv CXX `which icc`
>
> Also, I looked in the logfile you sent at the specific line where
the
> error occurred and compared that to the same line in the output
> created when I compile MET.  None of the external library links are
> being passed on the compilation line... specifically between
> "-I../../../src/libcode/vx_tc_util" and "-DBLOCK4" we should see
links
> for NetCDF and other external library dependencies.
>
> Not sure why this is happening but it may be related to the choice
of
> compiler.  Please let us know if you have icc available, and if so,
> try using that one.
>
> Thanks,
> John
>
> On Mon, Dec 8, 2014 at 4:53 PM, Julie Prestopnik via RT
> <met_help at ucar.edu
> >
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >
> > Hi Brock.  I'm sorry to see that MET can't seem to find netcdf.hh.
> > Could you please set up the environment variables on the command
> > line, and try
> to
> > build and install MET on the command line and then send me the
> > make_install.log file from that run along with the output of
> > "printenv | grep MET".  I looked at this make_install.log, paying
> > particular
> attention
> > to what is being looked at for the include files.  I don't know
> > where
> your
> > netcdf include files are, so I can't check for that location.
> >
> > If your environment variable is set correctly, it should be listed
> > in the output of "printenv | grep MET" and also in -I in the
> > make_install.log file.  If MET still can't find it after we check
> > those things, we'll have to look into it further.  I know that
> > certain recent versions of NetCDF (e.g. v4.1.3) no longer include
> > netcdf.hh, but I believe that 3.6.3,
> which
> > you said you are using, does have netcdf.hh.
> >
> > Sending this additional info from an install attempt with the
> > environment variables set, should help us better figure out what
the problem is.
> >
> > Thanks,
> > Julie
> >
> > On Mon, Dec 8, 2014 at 2:03 PM, Burghardt, Brock via RT <
> met_help at ucar.edu
> > >
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > >
> > > Hi Julie,
> > > Sorry for not getting back to you sooner (finals just wrapped
up).
> > > I've attached the make log.
> > > The printenv statement does not give anything for "MET", however
> > > when I set my MET_<LIB/INC> paths at command line I can print
> > > them, giving the proper paths.
> > >
> > > Thanks,
> > > Brock
> > >
> > > -----Original Message-----
> > > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > > Sent: Friday, December 05, 2014 9:56 AM
> > > To: Burghardt, Brock
> > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > >
> > > Hi Brock.  Could you please send me your make_install.log file
and
> could
> > > you please run "printenv | grep MET" again and send me the
output?
> > >
> > > Thanks,
> > > Julie
> > >
> > > On Thu, Dec 4, 2014 at 5:09 PM, Burghardt, Brock via RT <
> > met_help at ucar.edu
> > > >
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> > > >
> > > > Hi Julie,
> > > >
> > > > Still no luck with loading the intel environment (soft add
intel
> > > > in
> my
> > > > case) with the error developing with netcdf.hh in the make
> > > > install...any other ideas?
> > > >
> > > > Thanks, I appreciate the help.
> > > >
> > > > Brock
> > > >
> > > >
> > > >
> > > >
> > > > ________________________________________
> > > > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > > > Sent: Tuesday, December 02, 2014 11:02 AM
> > > > To: Burghardt, Brock
> > > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > > >
> > > > Hi Brock.  I'm sorry to hear that you are still having
problems
> > > > compiling METv5.0.  Thank you for sending the config.log file.
> > > >
> > > > Please try running the following command:
> > > >
> > > > make clean
> > > > module load intel
> > > >
> > > > and then run the configure command and the make install
command.
> > > >
> > > > Please let me know if that works.
> > > >
> > > > Thanks,
> > > > Julie
> > > >
> > > > On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
> > > > <met_help at ucar.edu
> > > > >
> > > > wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > > > > >
> > > > >
> > > > > Hi Julie,
> > > > >
> > > > > I've still been unable to compile METv5.0...independent of
> > > > > whether
> I
> > > > > specify the parent or both the parent/lib parent/inc paths
for
> > > > > netcdf, bufrlib, and GSL, I get an error with it trying to
> > > > > open netcdf.hh
> > > > >
> > > > > I noticed two paths to netcdf got set in the config.log
> > > > > (attached), one I set and another on the system, removing
the
> > > > > one I did not define still
> > > > did
> > > > > not give a successful build.
> > > > >
> > > > > I noticed I had errors in the config.log for many of the
> > > > > compilers where if tries to call, for instance, icpc
-qversion
> > > > > which should
> be
> > > > > icpc
> > > > -V....I
> > > > > looked through all the config and configure scripts and
could
> > > > > not find
> > > > the
> > > > > explicit icpc -qversion call to change to -V...do you think
> > > > > this could be the issue?
> > > > >
> > > > > Thank you for following up,
> > > > > Brock
> > > > >
> > > > > ________________________________________
> > > > > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > > > > Sent: Monday, December 01, 2014 9:56 AM
> > > > > To: Burghardt, Brock
> > > > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > > > >
> > > > > Hi Brock.  I just wanted to follow up - Were able to get a
> > > > > successful
> > > > build
> > > > > of METv5.0?
> > > > >
> > > > > Thanks,
> > > > > Julie
> > > > >
> > > > > On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
> > > > > <jpresto at ucar.edu>
> > > > > wrote:
> > > > >
> > > > > > Hi Brock.
> > > > > >
> > > > > > Thanks you for sending the output.  I was looking to make
> > > > > > sure that the MET environment variables, like MET_NETCDF
or
> > > > > > MET_NETCDFINC and MET_NETCDFLIB, were set correctly.
> > > > > >
> > > > > > Please take a look at the following page and follow the
> > > > > > instructions on compiling MET:
> > > > > >
> > > > > >
> > > > >
> > > >
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/comp
> > > > ilation/configure.php
> > > > > >
> > > > > > This will help guide you in setting up the required
> > > > > > environment
> > > > variables
> > > > > > and will hopefully resolve the problem you are
experiencing
> > > > > > with MET
> > > > not
> > > > > > being able to find netcdf.hh.
> > > > > >
> > > > > > Please let me know if you have any questions, and please
let
> > > > > > me know if you get a successful build after setting up the
> environment
> > > variables.
> > > > > >
> > > > > > Thanks,
> > > > > > Julie
> > > > > >
> > > > > > On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> > > > > > met_help at ucar.edu> wrote:
> > > > > >
> > > > > >>
> > > > > >> <URL:
> > > > > >> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > > > > >>
> > > > > >> Hi Julie, the output for the command is,
> > > > > >>
> > > > > >> hrothgar:/met-5.0$ printenv | grep met
> > > > > >> PWD=/home/brburgha/met-5.0
> > > > > >>
> > > > > >> ...unless you wanted this,
> > > > > >>
> > > > > >> hrothgar:/met-5.0$ printenv | grep MET*
> > > > > >> HOSTNAME=hrothgar.hpcc.ttu.edu ANT_HOME=/opt/rocks
> > > > > >> JAVA_HOME=/usr/java/latest
> > > > > >> P4_GLOBMEMSIZE=83886080
> > > > > >> MAGICK_HOME=/lustre/work/apps/ImageMagick
> > > > > >> HOME=/home/brburgha
> > > > > >> LOGNAME=brburgha
> > > > > >> G_BROKEN_FILENAMES=1
> > > > > >>
> > > > > >>
> > > > > >> Brock
> > > > > >>
> > > > > >>
> > > > > >> -----Original Message-----
> > > > > >> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > > > > >> Sent: Tuesday, November 18, 2014 10:38 AM
> > > > > >> To: Burghardt, Brock
> > > > > >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
> > > > > >> V4.1
> > > > > >>
> > > > > >> Hi Brock.  Thank you for getting the latest version of
MET.
> > > > > >> I see
> > > > > you're
> > > > > >> still problems with MET finding netcdf.hh.
> > > > > >>
> > > > > >> Could you please run "printenv | grep MET" and send me
the
> output?
> > > > > >>
> > > > > >> Thanks,
> > > > > >> Julie
> > > > > >>
> > > > > >> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT
<
> > > > > >> met_help at ucar.edu>
> > > > > >> wrote:
> > > > > >>
> > > > > >> >
> > > > > >> > <URL:
> > > > > >> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> >
> > > > > >> >
> > > > > >> > Hi Julie,
> > > > > >> >
> > > > > >> > Thanks for getting back to me so soon and explaining.
> > > > > >> > Removing the static flag did indeed help in compiling
the
> > > > > >> > executables I needed
> > > > but
> > > > > >> > it did not completely compile (another ld: missing
> > > > > >> > library error
> > > > came
> > > > > >> > up in the pb2nc build). However, I took your advice and
> started
> > > > > >> > over with V5.0 and got that running, but the compile
also
> > > > > >> > failed before making executables. The error was for not
> > > > > >> > being able to open
> > > > netcdf.hh
> > > > > >> > (make_install.log attached). I added all the met_<lib>
> > > > > >> > arguments at command line and checked they pointed to
> > > > > >> > right directory, and the configure step ran without
> > > > > >> > error. I
> verified
> > > > > >> > I do have read
> > > > permission
> > > > > >> for netcdf.hh, any ideas on this.
> > > > > >> >
> > > > > >> > Thanks,
> > > > > >> > Brock
> > > > > >> >
> > > > > >> > -----Original Message-----
> > > > > >> > From: Julie Prestopnik via RT
[mailto:met_help at ucar.edu]
> > > > > >> > Sent: Monday, November 17, 2014 4:34 PM
> > > > > >> > To: Burghardt, Brock
> > > > > >> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling
MET
> V4.1
> > > > > >> >
> > > > > >> > Hi Brock.
> > > > > >> >
> > > > > >> > I see that you are having problems building METv4.1
with
> > > > > >> > an intel compiler.  Thank you for attaching your
> > > > > >> > user_defs.mk
> and
> > > > > >> > make.log
> > > > > >> files.
> > > > > >> >
> > > > > >> > First, I wanted to mention that METv5.0 was release in
> > > > > >> > early September, which does include bug fixes for
> > > > > >> > METv4.1.  You can view
> > > > the
> > > > > >> > release notes for METv5.0 here:
> > > > > >> >
> > > > > >> >
> > > > > >> >
> > > > >
> http://www.dtcenter.org/met/users/support/release_notes/METv5.0_rele
> > > > > as
> > > > > >> > e_notes.php
> > > > > >> >
> > > > > >> > For now, let's try to overcome this error with METv4.1:
> > > > > >> >
> > > > > >> > ld: cannot find -lm
> > > > > >> >
> > > > > >> > We've had this issue come up with some other users in
the
> past.
> > > > > >> > By default, MET does static linking - meaning that it
> > > > > >> > links in the library files on which it depends at
compile
> > > > > >> > time. The other type of linking is call dynamic linking
-
> > > > > >> > for which the library files on
> > > > which
> > > > > >> > it depends are linked in at runtime. Other users have
had
> > > > > >> > success compiling MET with dynamic linking instead.
Open
> > > > > >> > up
> and
> > > > > >> > edit the top-level user_defs.mk file, by commenting out
> > > > > >> > or removing
> > > > references
> > > > > >> > to "-static". They show up on the CXX_FLAGS and
FC_FLAGS
> lines.
> > > > > >> >
> > > > > >> > Then try recompiling MET, being sure to execute a "make
clean"
> > > > first.
> > > > > >> >
> > > > > >> > Thanks,
> > > > > >> > Julie
> > > > > >> >
> > > > > >> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via
RT
> > > > > >> > < met_help at ucar.edu>
> > > > > >> > wrote:
> > > > > >> >
> > > > > >> > >
> > > > > >> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted
upon.
> > > > > >> > > Transaction: Ticket created by
brock.burghardt at ttu.edu
> > > > > >> > >        Queue: met_help
> > > > > >> > >      Subject: Issue compiling MET V4.1
> > > > > >> > >        Owner: Nobody
> > > > > >> > >   Requestors: brock.burghardt at ttu.edu
> > > > > >> > >       Status: new
> > > > > >> > >  Ticket <URL:
> > > > > >> > >
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > > > > >> > > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> > > Hello,
> > > > > >> > >
> > > > > >> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3
with
> > > > > >> > > an intel compiler (definitions and log file
attachted).
> > > > > >> > > I've gone through
> > > > all
> > > > > >> > > suggestions in the appendix of the users guide but
> > > > > >> > > still get an error while trying to make the
ensemble_stat executable.
> > > > > >> > > Is there
> > > > a
> > > > > >> > > liblm.so or .a required for this build?
> > > > > >> > >
> > > > > >> > > Thanks,
> > > > > >> > > Brock
> > > > > >> > >
> > > > > >> > > ld: cannot find -lm
> > > > > >> > > make[4]: *** [ensemble_stat] Error 1
> > > > > >> > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> > > Brock Burghardt
> > > > > >> > > Atmospheric Science PhD Student Dept. of Geosciences
> > > > > >> > > Texas Tech University
> > > > > >> > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> >
> > > > > >> >
> > > > > >> > --
> > > > > >> > Julie Prestopnik
> > > > > >> > National Center for Atmospheric Research Research
> > > > > >> > Applications Laboratory
> > > > > >> > Phone: 303.497.8399
> > > > > >> > Email: jpresto at ucar.edu
> > > > > >> >
> > > > > >> >
> > > > > >> >
> > > > > >>
> > > > > >>
> > > > > >> --
> > > > > >> Julie Prestopnik
> > > > > >> National Center for Atmospheric Research Research
> > > > > >> Applications
> > > > > Laboratory
> > > > > >> Phone: 303.497.8399
> > > > > >> Email: jpresto at ucar.edu
> > > > > >>
> > > > > >>
> > > > > >>
> > > > > >>
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Julie Prestopnik
> > > > > > National Center for Atmospheric Research Research
> > > > > > Applications Laboratory
> > > > > > Phone: 303.497.8399
> > > > > > Email: jpresto at ucar.edu
> > > > > >
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Julie Prestopnik
> > > > > National Center for Atmospheric Research Research
Applications
> > > > > Laboratory
> > > > > Phone: 303.497.8399
> > > > > Email: jpresto at ucar.edu
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > --
> > > > Julie Prestopnik
> > > > National Center for Atmospheric Research Research Applications
> > > > Laboratory
> > > > Phone: 303.497.8399
> > > > Email: jpresto at ucar.edu
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > > --
> > > Julie Prestopnik
> > > National Center for Atmospheric Research Research Applications
> Laboratory
> > > Phone: 303.497.8399
> > > Email: jpresto at ucar.edu
> > >
> > >
> > >
> >
> >
> > --
> > Julie Prestopnik
> > National Center for Atmospheric Research Research Applications
> > Laboratory
> > Phone: 303.497.8399
> > Email: jpresto at ucar.edu
> >
> >
>
>
>

------------------------------------------------
Subject: Issue compiling MET V4.1
From: John Halley Gotway
Time: Mon Dec 15 14:35:25 2014

Brock,

Yep, this one is easy to solve.  Change...
FROM: export MET_BUFR=$BUFRLIB
TO:      export MET_BUFRLIB=$BUFRLIB

Why is that necessary?  Because BUFRLIB doesn't follow the standard
convention of having "include" and "lib" sub-directories.  By setting
MET_BUFR={path}, MET looks for the library files in {path}/lib, but
that
doesn't exist!  So we instead set, MET_BUFRLIB to tell it where the
BUFR
library files live.

Sorry I didn't catch that wrinkle when looking at what you sent me
before.
This is mentioned in the top-level METv5.0 README file.

Please give that a shot and let me know.

Thanks,
John

On Mon, Dec 15, 2014 at 2:10 PM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:
>
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi John,
>
> I completely forgot to export the set paths, sorry about that. Upon
> exporting the paths I get a partially successful build with 11 of
the 23
> executables created.
> The error appears to occur when looking for the -lbufr library.
Would this
> have something to do with the BUFRLIB directory?
>
> Thanks,
> Brock
>
>
>
> -----Original Message-----
> From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> Sent: Friday, December 12, 2014 4:12 PM
> To: Burghardt, Brock
> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>
> Brock,
>
> Thanks for sending those log files.  Looking at line 3243 of
config.log, I
> see:
>    ac_cv_env_MET_NETCDF_set=
>    ac_cv_env_MET_NETCDF_value=
>
> This means that configure doesn't think that MET_NETCDF is set.
Looking
> at the syntax of the commands you sent, I assume you're using the
bash or
> ksh shell, rather than cshell.  That's fine, but you need to use
"export"
> when setting variables to pass those definitions down to sub-
processes.
> Please try the following:
>
> export MET_NETCDF=${NETCDF}
> export MET_BUFR=$BUFRLIB
> export MET_GSL=/lustre/work/apps/gsl
> export
CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
> export
F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort
>
> And then run "configure".  Then run the following command and make
sure
> that NetCDF is actually set:
>    grep ac_cv_env_MET_NETCDF_value= config.log
>
> If so, proceed to "make install" and "make test".
>
> Sorry for all of the headaches you've had to struggle through!
>
> John
>
> On Thu, Dec 11, 2014 at 12:21 PM, Burghardt, Brock via RT <
> met_help at ucar.edu
> > wrote:
> >
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >
> > Hi John,
> >
> > So I set the CXX to the available intel c++ compiler, however I am
> > still receiving the same error set.
> > Below are the other paths I set before configure and make:
> > MET_NETCDF=${NETCDF}
> > MET_BUFR=$BUFRLIB
> > MET_GSL=/lustre/work/apps/gsl
> > CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
> >
F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort
> >
> > Attached are both the config and make logs. I noticed that in my
> > config log that when it is looking for the icc version it gets an
> > error because it is invoking 'icc -qversion' but should be 'icc
-v' if
> > I want to check it on the command line. Do you think this could be
a
> cause?
> >
> > Thank you,
> > Brock
> >
> > ________________________________________
> > From: John Halley Gotway via RT [met_help at ucar.edu]
> > Sent: Tuesday, December 09, 2014 11:32 AM
> > To: Burghardt, Brock
> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >
> > Hello Brock,
> >
> > I work with Julie Prestopnik and I see that you're having trouble
> > compiling MET.  I took a look at the log file you sent and see
that
> > you're using the intel c++ icpc compiler.  In the past when I've
> > compiled MET using intel compilers, I've always used the "icc" c++
> > compiler.  I just tried compiling MET using icpc and run into all
> > kinds of problems.  Is icc available on your system (type: which
icc)?
> > If so, try using that by setting the CXX environment variable
before
> running the "configure" command for MET:
> >   setenv CXX `which icc`
> >
> > Also, I looked in the logfile you sent at the specific line where
the
> > error occurred and compared that to the same line in the output
> > created when I compile MET.  None of the external library links
are
> > being passed on the compilation line... specifically between
> > "-I../../../src/libcode/vx_tc_util" and "-DBLOCK4" we should see
links
> > for NetCDF and other external library dependencies.
> >
> > Not sure why this is happening but it may be related to the choice
of
> > compiler.  Please let us know if you have icc available, and if
so,
> > try using that one.
> >
> > Thanks,
> > John
> >
> > On Mon, Dec 8, 2014 at 4:53 PM, Julie Prestopnik via RT
> > <met_help at ucar.edu
> > >
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > >
> > > Hi Brock.  I'm sorry to see that MET can't seem to find
netcdf.hh.
> > > Could you please set up the environment variables on the command
> > > line, and try
> > to
> > > build and install MET on the command line and then send me the
> > > make_install.log file from that run along with the output of
> > > "printenv | grep MET".  I looked at this make_install.log,
paying
> > > particular
> > attention
> > > to what is being looked at for the include files.  I don't know
> > > where
> > your
> > > netcdf include files are, so I can't check for that location.
> > >
> > > If your environment variable is set correctly, it should be
listed
> > > in the output of "printenv | grep MET" and also in -I in the
> > > make_install.log file.  If MET still can't find it after we
check
> > > those things, we'll have to look into it further.  I know that
> > > certain recent versions of NetCDF (e.g. v4.1.3) no longer
include
> > > netcdf.hh, but I believe that 3.6.3,
> > which
> > > you said you are using, does have netcdf.hh.
> > >
> > > Sending this additional info from an install attempt with the
> > > environment variables set, should help us better figure out what
the
> problem is.
> > >
> > > Thanks,
> > > Julie
> > >
> > > On Mon, Dec 8, 2014 at 2:03 PM, Burghardt, Brock via RT <
> > met_help at ucar.edu
> > > >
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> > > >
> > > > Hi Julie,
> > > > Sorry for not getting back to you sooner (finals just wrapped
up).
> > > > I've attached the make log.
> > > > The printenv statement does not give anything for "MET",
however
> > > > when I set my MET_<LIB/INC> paths at command line I can print
> > > > them, giving the proper paths.
> > > >
> > > > Thanks,
> > > > Brock
> > > >
> > > > -----Original Message-----
> > > > From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > > > Sent: Friday, December 05, 2014 9:56 AM
> > > > To: Burghardt, Brock
> > > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > > >
> > > > Hi Brock.  Could you please send me your make_install.log file
and
> > could
> > > > you please run "printenv | grep MET" again and send me the
output?
> > > >
> > > > Thanks,
> > > > Julie
> > > >
> > > > On Thu, Dec 4, 2014 at 5:09 PM, Burghardt, Brock via RT <
> > > met_help at ucar.edu
> > > > >
> > > > wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > > > >
> > > > > Hi Julie,
> > > > >
> > > > > Still no luck with loading the intel environment (soft add
intel
> > > > > in
> > my
> > > > > case) with the error developing with netcdf.hh in the make
> > > > > install...any other ideas?
> > > > >
> > > > > Thanks, I appreciate the help.
> > > > >
> > > > > Brock
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > ________________________________________
> > > > > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > > > > Sent: Tuesday, December 02, 2014 11:02 AM
> > > > > To: Burghardt, Brock
> > > > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > > > >
> > > > > Hi Brock.  I'm sorry to hear that you are still having
problems
> > > > > compiling METv5.0.  Thank you for sending the config.log
file.
> > > > >
> > > > > Please try running the following command:
> > > > >
> > > > > make clean
> > > > > module load intel
> > > > >
> > > > > and then run the configure command and the make install
command.
> > > > >
> > > > > Please let me know if that works.
> > > > >
> > > > > Thanks,
> > > > > Julie
> > > > >
> > > > > On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
> > > > > <met_help at ucar.edu
> > > > > >
> > > > > wrote:
> > > > >
> > > > > >
> > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > > > > > >
> > > > > >
> > > > > > Hi Julie,
> > > > > >
> > > > > > I've still been unable to compile METv5.0...independent of
> > > > > > whether
> > I
> > > > > > specify the parent or both the parent/lib parent/inc paths
for
> > > > > > netcdf, bufrlib, and GSL, I get an error with it trying to
> > > > > > open netcdf.hh
> > > > > >
> > > > > > I noticed two paths to netcdf got set in the config.log
> > > > > > (attached), one I set and another on the system, removing
the
> > > > > > one I did not define still
> > > > > did
> > > > > > not give a successful build.
> > > > > >
> > > > > > I noticed I had errors in the config.log for many of the
> > > > > > compilers where if tries to call, for instance, icpc
-qversion
> > > > > > which should
> > be
> > > > > > icpc
> > > > > -V....I
> > > > > > looked through all the config and configure scripts and
could
> > > > > > not find
> > > > > the
> > > > > > explicit icpc -qversion call to change to -V...do you
think
> > > > > > this could be the issue?
> > > > > >
> > > > > > Thank you for following up,
> > > > > > Brock
> > > > > >
> > > > > > ________________________________________
> > > > > > From: Julie Prestopnik via RT [met_help at ucar.edu]
> > > > > > Sent: Monday, December 01, 2014 9:56 AM
> > > > > > To: Burghardt, Brock
> > > > > > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > > > > >
> > > > > > Hi Brock.  I just wanted to follow up - Were able to get a
> > > > > > successful
> > > > > build
> > > > > > of METv5.0?
> > > > > >
> > > > > > Thanks,
> > > > > > Julie
> > > > > >
> > > > > > On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
> > > > > > <jpresto at ucar.edu>
> > > > > > wrote:
> > > > > >
> > > > > > > Hi Brock.
> > > > > > >
> > > > > > > Thanks you for sending the output.  I was looking to
make
> > > > > > > sure that the MET environment variables, like MET_NETCDF
or
> > > > > > > MET_NETCDFINC and MET_NETCDFLIB, were set correctly.
> > > > > > >
> > > > > > > Please take a look at the following page and follow the
> > > > > > > instructions on compiling MET:
> > > > > > >
> > > > > > >
> > > > > >
> > > > >
> >
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/comp
> > > > > ilation/configure.php
> > > > > > >
> > > > > > > This will help guide you in setting up the required
> > > > > > > environment
> > > > > variables
> > > > > > > and will hopefully resolve the problem you are
experiencing
> > > > > > > with MET
> > > > > not
> > > > > > > being able to find netcdf.hh.
> > > > > > >
> > > > > > > Please let me know if you have any questions, and please
let
> > > > > > > me know if you get a successful build after setting up
the
> > environment
> > > > variables.
> > > > > > >
> > > > > > > Thanks,
> > > > > > > Julie
> > > > > > >
> > > > > > > On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT
<
> > > > > > > met_help at ucar.edu> wrote:
> > > > > > >
> > > > > > >>
> > > > > > >> <URL:
> > > > > > >> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> > > > > > >>
> > > > > > >> Hi Julie, the output for the command is,
> > > > > > >>
> > > > > > >> hrothgar:/met-5.0$ printenv | grep met
> > > > > > >> PWD=/home/brburgha/met-5.0
> > > > > > >>
> > > > > > >> ...unless you wanted this,
> > > > > > >>
> > > > > > >> hrothgar:/met-5.0$ printenv | grep MET*
> > > > > > >> HOSTNAME=hrothgar.hpcc.ttu.edu ANT_HOME=/opt/rocks
> > > > > > >> JAVA_HOME=/usr/java/latest
> > > > > > >> P4_GLOBMEMSIZE=83886080
> > > > > > >> MAGICK_HOME=/lustre/work/apps/ImageMagick
> > > > > > >> HOME=/home/brburgha
> > > > > > >> LOGNAME=brburgha
> > > > > > >> G_BROKEN_FILENAMES=1
> > > > > > >>
> > > > > > >>
> > > > > > >> Brock
> > > > > > >>
> > > > > > >>
> > > > > > >> -----Original Message-----
> > > > > > >> From: Julie Prestopnik via RT
[mailto:met_help at ucar.edu]
> > > > > > >> Sent: Tuesday, November 18, 2014 10:38 AM
> > > > > > >> To: Burghardt, Brock
> > > > > > >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling
MET
> > > > > > >> V4.1
> > > > > > >>
> > > > > > >> Hi Brock.  Thank you for getting the latest version of
MET.
> > > > > > >> I see
> > > > > > you're
> > > > > > >> still problems with MET finding netcdf.hh.
> > > > > > >>
> > > > > > >> Could you please run "printenv | grep MET" and send me
the
> > output?
> > > > > > >>
> > > > > > >> Thanks,
> > > > > > >> Julie
> > > > > > >>
> > > > > > >> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via
RT <
> > > > > > >> met_help at ucar.edu>
> > > > > > >> wrote:
> > > > > > >>
> > > > > > >> >
> > > > > > >> > <URL:
> > > > > > >> >
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > >
> > > > > > >> >
> > > > > > >> > Hi Julie,
> > > > > > >> >
> > > > > > >> > Thanks for getting back to me so soon and explaining.
> > > > > > >> > Removing the static flag did indeed help in compiling
the
> > > > > > >> > executables I needed
> > > > > but
> > > > > > >> > it did not completely compile (another ld: missing
> > > > > > >> > library error
> > > > > came
> > > > > > >> > up in the pb2nc build). However, I took your advice
and
> > started
> > > > > > >> > over with V5.0 and got that running, but the compile
also
> > > > > > >> > failed before making executables. The error was for
not
> > > > > > >> > being able to open
> > > > > netcdf.hh
> > > > > > >> > (make_install.log attached). I added all the
met_<lib>
> > > > > > >> > arguments at command line and checked they pointed to
> > > > > > >> > right directory, and the configure step ran without
> > > > > > >> > error. I
> > verified
> > > > > > >> > I do have read
> > > > > permission
> > > > > > >> for netcdf.hh, any ideas on this.
> > > > > > >> >
> > > > > > >> > Thanks,
> > > > > > >> > Brock
> > > > > > >> >
> > > > > > >> > -----Original Message-----
> > > > > > >> > From: Julie Prestopnik via RT
[mailto:met_help at ucar.edu]
> > > > > > >> > Sent: Monday, November 17, 2014 4:34 PM
> > > > > > >> > To: Burghardt, Brock
> > > > > > >> > Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling
MET
> > V4.1
> > > > > > >> >
> > > > > > >> > Hi Brock.
> > > > > > >> >
> > > > > > >> > I see that you are having problems building METv4.1
with
> > > > > > >> > an intel compiler.  Thank you for attaching your
> > > > > > >> > user_defs.mk
> > and
> > > > > > >> > make.log
> > > > > > >> files.
> > > > > > >> >
> > > > > > >> > First, I wanted to mention that METv5.0 was release
in
> > > > > > >> > early September, which does include bug fixes for
> > > > > > >> > METv4.1.  You can view
> > > > > the
> > > > > > >> > release notes for METv5.0 here:
> > > > > > >> >
> > > > > > >> >
> > > > > > >> >
> > > > > >
> >
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_rele
> > > > > > as
> > > > > > >> > e_notes.php
> > > > > > >> >
> > > > > > >> > For now, let's try to overcome this error with
METv4.1:
> > > > > > >> >
> > > > > > >> > ld: cannot find -lm
> > > > > > >> >
> > > > > > >> > We've had this issue come up with some other users in
the
> > past.
> > > > > > >> > By default, MET does static linking - meaning that it
> > > > > > >> > links in the library files on which it depends at
compile
> > > > > > >> > time. The other type of linking is call dynamic
linking -
> > > > > > >> > for which the library files on
> > > > > which
> > > > > > >> > it depends are linked in at runtime. Other users have
had
> > > > > > >> > success compiling MET with dynamic linking instead.
Open
> > > > > > >> > up
> > and
> > > > > > >> > edit the top-level user_defs.mk file, by commenting
out
> > > > > > >> > or removing
> > > > > references
> > > > > > >> > to "-static". They show up on the CXX_FLAGS and
FC_FLAGS
> > lines.
> > > > > > >> >
> > > > > > >> > Then try recompiling MET, being sure to execute a
"make
> clean"
> > > > > first.
> > > > > > >> >
> > > > > > >> > Thanks,
> > > > > > >> > Julie
> > > > > > >> >
> > > > > > >> > On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via
RT
> > > > > > >> > < met_help at ucar.edu>
> > > > > > >> > wrote:
> > > > > > >> >
> > > > > > >> > >
> > > > > > >> > > Mon Nov 17 13:55:16 2014: Request 69783 was acted
upon.
> > > > > > >> > > Transaction: Ticket created by
brock.burghardt at ttu.edu
> > > > > > >> > >        Queue: met_help
> > > > > > >> > >      Subject: Issue compiling MET V4.1
> > > > > > >> > >        Owner: Nobody
> > > > > > >> > >   Requestors: brock.burghardt at ttu.edu
> > > > > > >> > >       Status: new
> > > > > > >> > >  Ticket <URL:
> > > > > > >> > >
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > > > > > >> > > >
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> > > Hello,
> > > > > > >> > >
> > > > > > >> > > I'm trying to compile MET V4.1 using NETCDF V3.6.3
with
> > > > > > >> > > an intel compiler (definitions and log file
attachted).
> > > > > > >> > > I've gone through
> > > > > all
> > > > > > >> > > suggestions in the appendix of the users guide but
> > > > > > >> > > still get an error while trying to make the
ensemble_stat
> executable.
> > > > > > >> > > Is there
> > > > > a
> > > > > > >> > > liblm.so or .a required for this build?
> > > > > > >> > >
> > > > > > >> > > Thanks,
> > > > > > >> > > Brock
> > > > > > >> > >
> > > > > > >> > > ld: cannot find -lm
> > > > > > >> > > make[4]: *** [ensemble_stat] Error 1
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> > > Brock Burghardt
> > > > > > >> > > Atmospheric Science PhD Student Dept. of
Geosciences
> > > > > > >> > > Texas Tech University
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> >
> > > > > > >> >
> > > > > > >> > --
> > > > > > >> > Julie Prestopnik
> > > > > > >> > National Center for Atmospheric Research Research
> > > > > > >> > Applications Laboratory
> > > > > > >> > Phone: 303.497.8399
> > > > > > >> > Email: jpresto at ucar.edu
> > > > > > >> >
> > > > > > >> >
> > > > > > >> >
> > > > > > >>
> > > > > > >>
> > > > > > >> --
> > > > > > >> Julie Prestopnik
> > > > > > >> National Center for Atmospheric Research Research
> > > > > > >> Applications
> > > > > > Laboratory
> > > > > > >> Phone: 303.497.8399
> > > > > > >> Email: jpresto at ucar.edu
> > > > > > >>
> > > > > > >>
> > > > > > >>
> > > > > > >>
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > Julie Prestopnik
> > > > > > > National Center for Atmospheric Research Research
> > > > > > > Applications Laboratory
> > > > > > > Phone: 303.497.8399
> > > > > > > Email: jpresto at ucar.edu
> > > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Julie Prestopnik
> > > > > > National Center for Atmospheric Research Research
Applications
> > > > > > Laboratory
> > > > > > Phone: 303.497.8399
> > > > > > Email: jpresto at ucar.edu
> > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Julie Prestopnik
> > > > > National Center for Atmospheric Research Research
Applications
> > > > > Laboratory
> > > > > Phone: 303.497.8399
> > > > > Email: jpresto at ucar.edu
> > > > >
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > --
> > > > Julie Prestopnik
> > > > National Center for Atmospheric Research Research Applications
> > Laboratory
> > > > Phone: 303.497.8399
> > > > Email: jpresto at ucar.edu
> > > >
> > > >
> > > >
> > >
> > >
> > > --
> > > Julie Prestopnik
> > > National Center for Atmospheric Research Research Applications
> > > Laboratory
> > > Phone: 303.497.8399
> > > Email: jpresto at ucar.edu
> > >
> > >
> >
> >
> >
>
>
>

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Burghardt, Brock
Time: Tue Dec 16 09:56:45 2014

Hi John,

I got 14 of the 23 executables built, but another crash during the
pb2nc build. There are numerous 'undefined reference to’ statements
that popped up in the make_install.log, some of which had gfortran in
the statement, so i cleaned and tried with F77 pointing to the
gfortran libraries but got the same error. Any ideas on this?

Thanks,
Brock
Brock

> On Dec 15, 2014, at 3:35 PM, John Halley Gotway via RT
<met_help at ucar.edu> wrote:
>
> Brock,
>
> Yep, this one is easy to solve.  Change...
> FROM: export MET_BUFR=$BUFRLIB
> TO:      export MET_BUFRLIB=$BUFRLIB
>
> Why is that necessary?  Because BUFRLIB doesn't follow the standard
> convention of having "include" and "lib" sub-directories.  By
setting
> MET_BUFR={path}, MET looks for the library files in {path}/lib, but
that
> doesn't exist!  So we instead set, MET_BUFRLIB to tell it where the
BUFR
> library files live.
>
> Sorry I didn't catch that wrinkle when looking at what you sent me
before.
> This is mentioned in the top-level METv5.0 README file.
>
> Please give that a shot and let me know.
>
> Thanks,
> John
>
>
>
> On Mon, Dec 15, 2014 at 2:10 PM, Burghardt, Brock via RT
<met_help at ucar.edu>
> wrote:
>>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>>
>> Hi John,
>>
>> I completely forgot to export the set paths, sorry about that. Upon
>> exporting the paths I get a partially successful build with 11 of
the 23
>> executables created.
>> The error appears to occur when looking for the -lbufr library.
Would this
>> have something to do with the BUFRLIB directory?
>>
>> Thanks,
>> Brock
>>
>>
>>
>> -----Original Message-----
>> From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
>> Sent: Friday, December 12, 2014 4:12 PM
>> To: Burghardt, Brock
>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>>
>> Brock,
>>
>> Thanks for sending those log files.  Looking at line 3243 of
config.log, I
>> see:
>>   ac_cv_env_MET_NETCDF_set=
>>   ac_cv_env_MET_NETCDF_value=
>>
>> This means that configure doesn't think that MET_NETCDF is set.
Looking
>> at the syntax of the commands you sent, I assume you're using the
bash or
>> ksh shell, rather than cshell.  That's fine, but you need to use
"export"
>> when setting variables to pass those definitions down to sub-
processes.
>> Please try the following:
>>
>> export MET_NETCDF=${NETCDF}
>> export MET_BUFR=$BUFRLIB
>> export MET_GSL=/lustre/work/apps/gsl
>> export
CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
>> export
F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort
>>
>> And then run "configure".  Then run the following command and make
sure
>> that NetCDF is actually set:
>>   grep ac_cv_env_MET_NETCDF_value= config.log
>>
>> If so, proceed to "make install" and "make test".
>>
>> Sorry for all of the headaches you've had to struggle through!
>>
>> John
>>
>> On Thu, Dec 11, 2014 at 12:21 PM, Burghardt, Brock via RT <
>> met_help at ucar.edu
>>> wrote:
>>>
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>>>
>>> Hi John,
>>>
>>> So I set the CXX to the available intel c++ compiler, however I am
>>> still receiving the same error set.
>>> Below are the other paths I set before configure and make:
>>> MET_NETCDF=${NETCDF}
>>> MET_BUFR=$BUFRLIB
>>> MET_GSL=/lustre/work/apps/gsl
>>> CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
>>>
F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort
>>>
>>> Attached are both the config and make logs. I noticed that in my
>>> config log that when it is looking for the icc version it gets an
>>> error because it is invoking 'icc -qversion' but should be 'icc
-v' if
>>> I want to check it on the command line. Do you think this could be
a
>> cause?
>>>
>>> Thank you,
>>> Brock
>>>
>>> ________________________________________
>>> From: John Halley Gotway via RT [met_help at ucar.edu]
>>> Sent: Tuesday, December 09, 2014 11:32 AM
>>> To: Burghardt, Brock
>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>>>
>>> Hello Brock,
>>>
>>> I work with Julie Prestopnik and I see that you're having trouble
>>> compiling MET.  I took a look at the log file you sent and see
that
>>> you're using the intel c++ icpc compiler.  In the past when I've
>>> compiled MET using intel compilers, I've always used the "icc" c++
>>> compiler.  I just tried compiling MET using icpc and run into all
>>> kinds of problems.  Is icc available on your system (type: which
icc)?
>>> If so, try using that by setting the CXX environment variable
before
>> running the "configure" command for MET:
>>>  setenv CXX `which icc`
>>>
>>> Also, I looked in the logfile you sent at the specific line where
the
>>> error occurred and compared that to the same line in the output
>>> created when I compile MET.  None of the external library links
are
>>> being passed on the compilation line... specifically between
>>> "-I../../../src/libcode/vx_tc_util" and "-DBLOCK4" we should see
links
>>> for NetCDF and other external library dependencies.
>>>
>>> Not sure why this is happening but it may be related to the choice
of
>>> compiler.  Please let us know if you have icc available, and if
so,
>>> try using that one.
>>>
>>> Thanks,
>>> John
>>>
>>> On Mon, Dec 8, 2014 at 4:53 PM, Julie Prestopnik via RT
>>> <met_help at ucar.edu
>>>>
>>> wrote:
>>>
>>>>
>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>>>>
>>>> Hi Brock.  I'm sorry to see that MET can't seem to find
netcdf.hh.
>>>> Could you please set up the environment variables on the command
>>>> line, and try
>>> to
>>>> build and install MET on the command line and then send me the
>>>> make_install.log file from that run along with the output of
>>>> "printenv | grep MET".  I looked at this make_install.log, paying
>>>> particular
>>> attention
>>>> to what is being looked at for the include files.  I don't know
>>>> where
>>> your
>>>> netcdf include files are, so I can't check for that location.
>>>>
>>>> If your environment variable is set correctly, it should be
listed
>>>> in the output of "printenv | grep MET" and also in -I in the
>>>> make_install.log file.  If MET still can't find it after we check
>>>> those things, we'll have to look into it further.  I know that
>>>> certain recent versions of NetCDF (e.g. v4.1.3) no longer include
>>>> netcdf.hh, but I believe that 3.6.3,
>>> which
>>>> you said you are using, does have netcdf.hh.
>>>>
>>>> Sending this additional info from an install attempt with the
>>>> environment variables set, should help us better figure out what
the
>> problem is.
>>>>
>>>> Thanks,
>>>> Julie
>>>>
>>>> On Mon, Dec 8, 2014 at 2:03 PM, Burghardt, Brock via RT <
>>> met_help at ucar.edu
>>>>>
>>>> wrote:
>>>>
>>>>>
>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>>>>>
>>>>> Hi Julie,
>>>>> Sorry for not getting back to you sooner (finals just wrapped
up).
>>>>> I've attached the make log.
>>>>> The printenv statement does not give anything for "MET", however
>>>>> when I set my MET_<LIB/INC> paths at command line I can print
>>>>> them, giving the proper paths.
>>>>>
>>>>> Thanks,
>>>>> Brock
>>>>>
>>>>> -----Original Message-----
>>>>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
>>>>> Sent: Friday, December 05, 2014 9:56 AM
>>>>> To: Burghardt, Brock
>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>>>>>
>>>>> Hi Brock.  Could you please send me your make_install.log file
and
>>> could
>>>>> you please run "printenv | grep MET" again and send me the
output?
>>>>>
>>>>> Thanks,
>>>>> Julie
>>>>>
>>>>> On Thu, Dec 4, 2014 at 5:09 PM, Burghardt, Brock via RT <
>>>> met_help at ucar.edu
>>>>>>
>>>>> wrote:
>>>>>
>>>>>>
>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>>>>>>
>>>>>> Hi Julie,
>>>>>>
>>>>>> Still no luck with loading the intel environment (soft add
intel
>>>>>> in
>>> my
>>>>>> case) with the error developing with netcdf.hh in the make
>>>>>> install...any other ideas?
>>>>>>
>>>>>> Thanks, I appreciate the help.
>>>>>>
>>>>>> Brock
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> ________________________________________
>>>>>> From: Julie Prestopnik via RT [met_help at ucar.edu]
>>>>>> Sent: Tuesday, December 02, 2014 11:02 AM
>>>>>> To: Burghardt, Brock
>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>>>>>>
>>>>>> Hi Brock.  I'm sorry to hear that you are still having problems
>>>>>> compiling METv5.0.  Thank you for sending the config.log file.
>>>>>>
>>>>>> Please try running the following command:
>>>>>>
>>>>>> make clean
>>>>>> module load intel
>>>>>>
>>>>>> and then run the configure command and the make install
command.
>>>>>>
>>>>>> Please let me know if that works.
>>>>>>
>>>>>> Thanks,
>>>>>> Julie
>>>>>>
>>>>>> On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
>>>>>> <met_help at ucar.edu
>>>>>>>
>>>>>> wrote:
>>>>>>
>>>>>>>
>>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>>>>>>>>
>>>>>>>
>>>>>>> Hi Julie,
>>>>>>>
>>>>>>> I've still been unable to compile METv5.0...independent of
>>>>>>> whether
>>> I
>>>>>>> specify the parent or both the parent/lib parent/inc paths for
>>>>>>> netcdf, bufrlib, and GSL, I get an error with it trying to
>>>>>>> open netcdf.hh
>>>>>>>
>>>>>>> I noticed two paths to netcdf got set in the config.log
>>>>>>> (attached), one I set and another on the system, removing the
>>>>>>> one I did not define still
>>>>>> did
>>>>>>> not give a successful build.
>>>>>>>
>>>>>>> I noticed I had errors in the config.log for many of the
>>>>>>> compilers where if tries to call, for instance, icpc -qversion
>>>>>>> which should
>>> be
>>>>>>> icpc
>>>>>> -V....I
>>>>>>> looked through all the config and configure scripts and could
>>>>>>> not find
>>>>>> the
>>>>>>> explicit icpc -qversion call to change to -V...do you think
>>>>>>> this could be the issue?
>>>>>>>
>>>>>>> Thank you for following up,
>>>>>>> Brock
>>>>>>>
>>>>>>> ________________________________________
>>>>>>> From: Julie Prestopnik via RT [met_help at ucar.edu]
>>>>>>> Sent: Monday, December 01, 2014 9:56 AM
>>>>>>> To: Burghardt, Brock
>>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>>>>>>>
>>>>>>> Hi Brock.  I just wanted to follow up - Were able to get a
>>>>>>> successful
>>>>>> build
>>>>>>> of METv5.0?
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Julie
>>>>>>>
>>>>>>> On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
>>>>>>> <jpresto at ucar.edu>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> Hi Brock.
>>>>>>>>
>>>>>>>> Thanks you for sending the output.  I was looking to make
>>>>>>>> sure that the MET environment variables, like MET_NETCDF or
>>>>>>>> MET_NETCDFINC and MET_NETCDFLIB, were set correctly.
>>>>>>>>
>>>>>>>> Please take a look at the following page and follow the
>>>>>>>> instructions on compiling MET:
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/comp
>>>>>> ilation/configure.php
>>>>>>>>
>>>>>>>> This will help guide you in setting up the required
>>>>>>>> environment
>>>>>> variables
>>>>>>>> and will hopefully resolve the problem you are experiencing
>>>>>>>> with MET
>>>>>> not
>>>>>>>> being able to find netcdf.hh.
>>>>>>>>
>>>>>>>> Please let me know if you have any questions, and please let
>>>>>>>> me know if you get a successful build after setting up the
>>> environment
>>>>> variables.
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>> Julie
>>>>>>>>
>>>>>>>> On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
>>>>>>>> met_help at ucar.edu> wrote:
>>>>>>>>
>>>>>>>>>
>>>>>>>>> <URL:
>>>>>>>>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>>>>>>>>>
>>>>>>>>> Hi Julie, the output for the command is,
>>>>>>>>>
>>>>>>>>> hrothgar:/met-5.0$ printenv | grep met
>>>>>>>>> PWD=/home/brburgha/met-5.0
>>>>>>>>>
>>>>>>>>> ...unless you wanted this,
>>>>>>>>>
>>>>>>>>> hrothgar:/met-5.0$ printenv | grep MET*
>>>>>>>>> HOSTNAME=hrothgar.hpcc.ttu.edu ANT_HOME=/opt/rocks
>>>>>>>>> JAVA_HOME=/usr/java/latest
>>>>>>>>> P4_GLOBMEMSIZE=83886080
>>>>>>>>> MAGICK_HOME=/lustre/work/apps/ImageMagick
>>>>>>>>> HOME=/home/brburgha
>>>>>>>>> LOGNAME=brburgha
>>>>>>>>> G_BROKEN_FILENAMES=1
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Brock
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> -----Original Message-----
>>>>>>>>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
>>>>>>>>> Sent: Tuesday, November 18, 2014 10:38 AM
>>>>>>>>> To: Burghardt, Brock
>>>>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
>>>>>>>>> V4.1
>>>>>>>>>
>>>>>>>>> Hi Brock.  Thank you for getting the latest version of MET.
>>>>>>>>> I see
>>>>>>> you're
>>>>>>>>> still problems with MET finding netcdf.hh.
>>>>>>>>>
>>>>>>>>> Could you please run "printenv | grep MET" and send me the
>>> output?
>>>>>>>>>
>>>>>>>>> Thanks,
>>>>>>>>> Julie
>>>>>>>>>
>>>>>>>>> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
>>>>>>>>> met_help at ucar.edu>
>>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> <URL:
>>>>>>>>>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>>>>
>>>>>>>>>>
>>>>>>>>>> Hi Julie,
>>>>>>>>>>
>>>>>>>>>> Thanks for getting back to me so soon and explaining.
>>>>>>>>>> Removing the static flag did indeed help in compiling the
>>>>>>>>>> executables I needed
>>>>>> but
>>>>>>>>>> it did not completely compile (another ld: missing
>>>>>>>>>> library error
>>>>>> came
>>>>>>>>>> up in the pb2nc build). However, I took your advice and
>>> started
>>>>>>>>>> over with V5.0 and got that running, but the compile also
>>>>>>>>>> failed before making executables. The error was for not
>>>>>>>>>> being able to open
>>>>>> netcdf.hh
>>>>>>>>>> (make_install.log attached). I added all the met_<lib>
>>>>>>>>>> arguments at command line and checked they pointed to
>>>>>>>>>> right directory, and the configure step ran without
>>>>>>>>>> error. I
>>> verified
>>>>>>>>>> I do have read
>>>>>> permission
>>>>>>>>> for netcdf.hh, any ideas on this.
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>> Brock
>>>>>>>>>>
>>>>>>>>>> -----Original Message-----
>>>>>>>>>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
>>>>>>>>>> Sent: Monday, November 17, 2014 4:34 PM
>>>>>>>>>> To: Burghardt, Brock
>>>>>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
>>> V4.1
>>>>>>>>>>
>>>>>>>>>> Hi Brock.
>>>>>>>>>>
>>>>>>>>>> I see that you are having problems building METv4.1 with
>>>>>>>>>> an intel compiler.  Thank you for attaching your
>>>>>>>>>> user_defs.mk
>>> and
>>>>>>>>>> make.log
>>>>>>>>> files.
>>>>>>>>>>
>>>>>>>>>> First, I wanted to mention that METv5.0 was release in
>>>>>>>>>> early September, which does include bug fixes for
>>>>>>>>>> METv4.1.  You can view
>>>>>> the
>>>>>>>>>> release notes for METv5.0 here:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>
>>>
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_rele
>>>>>>> as
>>>>>>>>>> e_notes.php
>>>>>>>>>>
>>>>>>>>>> For now, let's try to overcome this error with METv4.1:
>>>>>>>>>>
>>>>>>>>>> ld: cannot find -lm
>>>>>>>>>>
>>>>>>>>>> We've had this issue come up with some other users in the
>>> past.
>>>>>>>>>> By default, MET does static linking - meaning that it
>>>>>>>>>> links in the library files on which it depends at compile
>>>>>>>>>> time. The other type of linking is call dynamic linking -
>>>>>>>>>> for which the library files on
>>>>>> which
>>>>>>>>>> it depends are linked in at runtime. Other users have had
>>>>>>>>>> success compiling MET with dynamic linking instead. Open
>>>>>>>>>> up
>>> and
>>>>>>>>>> edit the top-level user_defs.mk file, by commenting out
>>>>>>>>>> or removing
>>>>>> references
>>>>>>>>>> to "-static". They show up on the CXX_FLAGS and FC_FLAGS
>>> lines.
>>>>>>>>>>
>>>>>>>>>> Then try recompiling MET, being sure to execute a "make
>> clean"
>>>>>> first.
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>> Julie
>>>>>>>>>>
>>>>>>>>>> On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT
>>>>>>>>>> < met_help at ucar.edu>
>>>>>>>>>> wrote:
>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
>>>>>>>>>>> Transaction: Ticket created by brock.burghardt at ttu.edu
>>>>>>>>>>>       Queue: met_help
>>>>>>>>>>>     Subject: Issue compiling MET V4.1
>>>>>>>>>>>       Owner: Nobody
>>>>>>>>>>>  Requestors: brock.burghardt at ttu.edu
>>>>>>>>>>>      Status: new
>>>>>>>>>>> Ticket <URL:
>>>>>>>>>>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Hello,
>>>>>>>>>>>
>>>>>>>>>>> I'm trying to compile MET V4.1 using NETCDF V3.6.3 with
>>>>>>>>>>> an intel compiler (definitions and log file attachted).
>>>>>>>>>>> I've gone through
>>>>>> all
>>>>>>>>>>> suggestions in the appendix of the users guide but
>>>>>>>>>>> still get an error while trying to make the ensemble_stat
>> executable.
>>>>>>>>>>> Is there
>>>>>> a
>>>>>>>>>>> liblm.so or .a required for this build?
>>>>>>>>>>>
>>>>>>>>>>> Thanks,
>>>>>>>>>>> Brock
>>>>>>>>>>>
>>>>>>>>>>> ld: cannot find -lm
>>>>>>>>>>> make[4]: *** [ensemble_stat] Error 1
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Brock Burghardt
>>>>>>>>>>> Atmospheric Science PhD Student Dept. of Geosciences
>>>>>>>>>>> Texas Tech University
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> --
>>>>>>>>>> Julie Prestopnik
>>>>>>>>>> National Center for Atmospheric Research Research
>>>>>>>>>> Applications Laboratory
>>>>>>>>>> Phone: 303.497.8399
>>>>>>>>>> Email: jpresto at ucar.edu
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>> Julie Prestopnik
>>>>>>>>> National Center for Atmospheric Research Research
>>>>>>>>> Applications
>>>>>>> Laboratory
>>>>>>>>> Phone: 303.497.8399
>>>>>>>>> Email: jpresto at ucar.edu
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> Julie Prestopnik
>>>>>>>> National Center for Atmospheric Research Research
>>>>>>>> Applications Laboratory
>>>>>>>> Phone: 303.497.8399
>>>>>>>> Email: jpresto at ucar.edu
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Julie Prestopnik
>>>>>>> National Center for Atmospheric Research Research Applications
>>>>>>> Laboratory
>>>>>>> Phone: 303.497.8399
>>>>>>> Email: jpresto at ucar.edu
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Julie Prestopnik
>>>>>> National Center for Atmospheric Research Research Applications
>>>>>> Laboratory
>>>>>> Phone: 303.497.8399
>>>>>> Email: jpresto at ucar.edu
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Julie Prestopnik
>>>>> National Center for Atmospheric Research Research Applications
>>> Laboratory
>>>>> Phone: 303.497.8399
>>>>> Email: jpresto at ucar.edu
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> Julie Prestopnik
>>>> National Center for Atmospheric Research Research Applications
>>>> Laboratory
>>>> Phone: 303.497.8399
>>>> Email: jpresto at ucar.edu
>>>>
>>>>
>>>
>>>
>>>
>>
>>
>>
>

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Julie Prestopnik
Time: Tue Dec 16 10:12:46 2014

Hi Brock.

gfortran is the GNU Fortran compiler.  When compiling MET, you should
use
the same family of compilers. While I haven't actually tried it,
mixing GNU
compilers with Intel compilers isn't a good idea.

Since you're using icc please try compiling MET using ifort, the Intel
Fortran compiler.  Please note that when you compile the required MET
libraries, you should do so using the Intel compilers as well. Here's
the
tutorial page that talks about that:
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/compilation/req_libs.php

Please let us know if you have any other questions or encounter any
problems.

Thanks,
Julie

On Tue, Dec 16, 2014 at 9:56 AM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:
>
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi John,
>
> I got 14 of the 23 executables built, but another crash during the
pb2nc
> build. There are numerous 'undefined reference to’ statements that
popped
> up in the make_install.log, some of which had gfortran in the
statement, so
> i cleaned and tried with F77 pointing to the gfortran libraries but
got the
> same error. Any ideas on this?
>
>
> Thanks,
> Brock
> Brock
>
> > On Dec 15, 2014, at 3:35 PM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
> >
> > Brock,
> >
> > Yep, this one is easy to solve.  Change...
> > FROM: export MET_BUFR=$BUFRLIB
> > TO:      export MET_BUFRLIB=$BUFRLIB
> >
> > Why is that necessary?  Because BUFRLIB doesn't follow the
standard
> > convention of having "include" and "lib" sub-directories.  By
setting
> > MET_BUFR={path}, MET looks for the library files in {path}/lib,
but that
> > doesn't exist!  So we instead set, MET_BUFRLIB to tell it where
the BUFR
> > library files live.
> >
> > Sorry I didn't catch that wrinkle when looking at what you sent me
> before.
> > This is mentioned in the top-level METv5.0 README file.
> >
> > Please give that a shot and let me know.
> >
> > Thanks,
> > John
> >
> >
> >
> > On Mon, Dec 15, 2014 at 2:10 PM, Burghardt, Brock via RT <
> met_help at ucar.edu>
> > wrote:
> >>
> >>
> >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >>
> >> Hi John,
> >>
> >> I completely forgot to export the set paths, sorry about that.
Upon
> >> exporting the paths I get a partially successful build with 11 of
the 23
> >> executables created.
> >> The error appears to occur when looking for the -lbufr library.
Would
> this
> >> have something to do with the BUFRLIB directory?
> >>
> >> Thanks,
> >> Brock
> >>
> >>
> >>
> >> -----Original Message-----
> >> From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> >> Sent: Friday, December 12, 2014 4:12 PM
> >> To: Burghardt, Brock
> >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >>
> >> Brock,
> >>
> >> Thanks for sending those log files.  Looking at line 3243 of
> config.log, I
> >> see:
> >>   ac_cv_env_MET_NETCDF_set=
> >>   ac_cv_env_MET_NETCDF_value=
> >>
> >> This means that configure doesn't think that MET_NETCDF is set.
Looking
> >> at the syntax of the commands you sent, I assume you're using the
bash
> or
> >> ksh shell, rather than cshell.  That's fine, but you need to use
> "export"
> >> when setting variables to pass those definitions down to sub-
processes.
> >> Please try the following:
> >>
> >> export MET_NETCDF=${NETCDF}
> >> export MET_BUFR=$BUFRLIB
> >> export MET_GSL=/lustre/work/apps/gsl
> >> export
> CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
> >> export
> F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort
> >>
> >> And then run "configure".  Then run the following command and
make sure
> >> that NetCDF is actually set:
> >>   grep ac_cv_env_MET_NETCDF_value= config.log
> >>
> >> If so, proceed to "make install" and "make test".
> >>
> >> Sorry for all of the headaches you've had to struggle through!
> >>
> >> John
> >>
> >> On Thu, Dec 11, 2014 at 12:21 PM, Burghardt, Brock via RT <
> >> met_help at ucar.edu
> >>> wrote:
> >>>
> >>>
> >>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >>>
> >>> Hi John,
> >>>
> >>> So I set the CXX to the available intel c++ compiler, however I
am
> >>> still receiving the same error set.
> >>> Below are the other paths I set before configure and make:
> >>> MET_NETCDF=${NETCDF}
> >>> MET_BUFR=$BUFRLIB
> >>> MET_GSL=/lustre/work/apps/gsl
> >>>
CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
> >>>
F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort
> >>>
> >>> Attached are both the config and make logs. I noticed that in my
> >>> config log that when it is looking for the icc version it gets
an
> >>> error because it is invoking 'icc -qversion' but should be 'icc
-v' if
> >>> I want to check it on the command line. Do you think this could
be a
> >> cause?
> >>>
> >>> Thank you,
> >>> Brock
> >>>
> >>> ________________________________________
> >>> From: John Halley Gotway via RT [met_help at ucar.edu]
> >>> Sent: Tuesday, December 09, 2014 11:32 AM
> >>> To: Burghardt, Brock
> >>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >>>
> >>> Hello Brock,
> >>>
> >>> I work with Julie Prestopnik and I see that you're having
trouble
> >>> compiling MET.  I took a look at the log file you sent and see
that
> >>> you're using the intel c++ icpc compiler.  In the past when I've
> >>> compiled MET using intel compilers, I've always used the "icc"
c++
> >>> compiler.  I just tried compiling MET using icpc and run into
all
> >>> kinds of problems.  Is icc available on your system (type: which
icc)?
> >>> If so, try using that by setting the CXX environment variable
before
> >> running the "configure" command for MET:
> >>>  setenv CXX `which icc`
> >>>
> >>> Also, I looked in the logfile you sent at the specific line
where the
> >>> error occurred and compared that to the same line in the output
> >>> created when I compile MET.  None of the external library links
are
> >>> being passed on the compilation line... specifically between
> >>> "-I../../../src/libcode/vx_tc_util" and "-DBLOCK4" we should see
links
> >>> for NetCDF and other external library dependencies.
> >>>
> >>> Not sure why this is happening but it may be related to the
choice of
> >>> compiler.  Please let us know if you have icc available, and if
so,
> >>> try using that one.
> >>>
> >>> Thanks,
> >>> John
> >>>
> >>> On Mon, Dec 8, 2014 at 4:53 PM, Julie Prestopnik via RT
> >>> <met_help at ucar.edu
> >>>>
> >>> wrote:
> >>>
> >>>>
> >>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >>>>
> >>>> Hi Brock.  I'm sorry to see that MET can't seem to find
netcdf.hh.
> >>>> Could you please set up the environment variables on the
command
> >>>> line, and try
> >>> to
> >>>> build and install MET on the command line and then send me the
> >>>> make_install.log file from that run along with the output of
> >>>> "printenv | grep MET".  I looked at this make_install.log,
paying
> >>>> particular
> >>> attention
> >>>> to what is being looked at for the include files.  I don't know
> >>>> where
> >>> your
> >>>> netcdf include files are, so I can't check for that location.
> >>>>
> >>>> If your environment variable is set correctly, it should be
listed
> >>>> in the output of "printenv | grep MET" and also in -I in the
> >>>> make_install.log file.  If MET still can't find it after we
check
> >>>> those things, we'll have to look into it further.  I know that
> >>>> certain recent versions of NetCDF (e.g. v4.1.3) no longer
include
> >>>> netcdf.hh, but I believe that 3.6.3,
> >>> which
> >>>> you said you are using, does have netcdf.hh.
> >>>>
> >>>> Sending this additional info from an install attempt with the
> >>>> environment variables set, should help us better figure out
what the
> >> problem is.
> >>>>
> >>>> Thanks,
> >>>> Julie
> >>>>
> >>>> On Mon, Dec 8, 2014 at 2:03 PM, Burghardt, Brock via RT <
> >>> met_help at ucar.edu
> >>>>>
> >>>> wrote:
> >>>>
> >>>>>
> >>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> >>>>>
> >>>>> Hi Julie,
> >>>>> Sorry for not getting back to you sooner (finals just wrapped
up).
> >>>>> I've attached the make log.
> >>>>> The printenv statement does not give anything for "MET",
however
> >>>>> when I set my MET_<LIB/INC> paths at command line I can print
> >>>>> them, giving the proper paths.
> >>>>>
> >>>>> Thanks,
> >>>>> Brock
> >>>>>
> >>>>> -----Original Message-----
> >>>>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> >>>>> Sent: Friday, December 05, 2014 9:56 AM
> >>>>> To: Burghardt, Brock
> >>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >>>>>
> >>>>> Hi Brock.  Could you please send me your make_install.log file
and
> >>> could
> >>>>> you please run "printenv | grep MET" again and send me the
output?
> >>>>>
> >>>>> Thanks,
> >>>>> Julie
> >>>>>
> >>>>> On Thu, Dec 4, 2014 at 5:09 PM, Burghardt, Brock via RT <
> >>>> met_help at ucar.edu
> >>>>>>
> >>>>> wrote:
> >>>>>
> >>>>>>
> >>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> >>>>>>
> >>>>>> Hi Julie,
> >>>>>>
> >>>>>> Still no luck with loading the intel environment (soft add
intel
> >>>>>> in
> >>> my
> >>>>>> case) with the error developing with netcdf.hh in the make
> >>>>>> install...any other ideas?
> >>>>>>
> >>>>>> Thanks, I appreciate the help.
> >>>>>>
> >>>>>> Brock
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> ________________________________________
> >>>>>> From: Julie Prestopnik via RT [met_help at ucar.edu]
> >>>>>> Sent: Tuesday, December 02, 2014 11:02 AM
> >>>>>> To: Burghardt, Brock
> >>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> >>>>>>
> >>>>>> Hi Brock.  I'm sorry to hear that you are still having
problems
> >>>>>> compiling METv5.0.  Thank you for sending the config.log
file.
> >>>>>>
> >>>>>> Please try running the following command:
> >>>>>>
> >>>>>> make clean
> >>>>>> module load intel
> >>>>>>
> >>>>>> and then run the configure command and the make install
command.
> >>>>>>
> >>>>>> Please let me know if that works.
> >>>>>>
> >>>>>> Thanks,
> >>>>>> Julie
> >>>>>>
> >>>>>> On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
> >>>>>> <met_help at ucar.edu
> >>>>>>>
> >>>>>> wrote:
> >>>>>>
> >>>>>>>
> >>>>>>> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> >>>>>>>>
> >>>>>>>
> >>>>>>> Hi Julie,
> >>>>>>>
> >>>>>>> I've still been unable to compile METv5.0...independent of
> >>>>>>> whether
> >>> I
> >>>>>>> specify the parent or both the parent/lib parent/inc paths
for
> >>>>>>> netcdf, bufrlib, and GSL, I get an error with it trying to
> >>>>>>> open netcdf.hh
> >>>>>>>
> >>>>>>> I noticed two paths to netcdf got set in the config.log
> >>>>>>> (attached), one I set and another on the system, removing
the
> >>>>>>> one I did not define still
> >>>>>> did
> >>>>>>> not give a successful build.
> >>>>>>>
> >>>>>>> I noticed I had errors in the config.log for many of the
> >>>>>>> compilers where if tries to call, for instance, icpc
-qversion
> >>>>>>> which should
> >>> be
> >>>>>>> icpc
> >>>>>> -V....I
> >>>>>>> looked through all the config and configure scripts and
could
> >>>>>>> not find
> >>>>>> the
> >>>>>>> explicit icpc -qversion call to change to -V...do you think
> >>>>>>> this could be the issue?
> >>>>>>>
> >>>>>>> Thank you for following up,
> >>>>>>> Brock
> >>>>>>>
> >>>>>>> ________________________________________
> >>>>>>> From: Julie Prestopnik via RT [met_help at ucar.edu]
> >>>>>>> Sent: Monday, December 01, 2014 9:56 AM
> >>>>>>> To: Burghardt, Brock
> >>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> >>>>>>>
> >>>>>>> Hi Brock.  I just wanted to follow up - Were able to get a
> >>>>>>> successful
> >>>>>> build
> >>>>>>> of METv5.0?
> >>>>>>>
> >>>>>>> Thanks,
> >>>>>>> Julie
> >>>>>>>
> >>>>>>> On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
> >>>>>>> <jpresto at ucar.edu>
> >>>>>>> wrote:
> >>>>>>>
> >>>>>>>> Hi Brock.
> >>>>>>>>
> >>>>>>>> Thanks you for sending the output.  I was looking to make
> >>>>>>>> sure that the MET environment variables, like MET_NETCDF or
> >>>>>>>> MET_NETCDFINC and MET_NETCDFLIB, were set correctly.
> >>>>>>>>
> >>>>>>>> Please take a look at the following page and follow the
> >>>>>>>> instructions on compiling MET:
> >>>>>>>>
> >>>>>>>>
> >>>>>>>
> >>>>>>
> >>>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/comp
> >>>>>> ilation/configure.php
> >>>>>>>>
> >>>>>>>> This will help guide you in setting up the required
> >>>>>>>> environment
> >>>>>> variables
> >>>>>>>> and will hopefully resolve the problem you are experiencing
> >>>>>>>> with MET
> >>>>>> not
> >>>>>>>> being able to find netcdf.hh.
> >>>>>>>>
> >>>>>>>> Please let me know if you have any questions, and please
let
> >>>>>>>> me know if you get a successful build after setting up the
> >>> environment
> >>>>> variables.
> >>>>>>>>
> >>>>>>>> Thanks,
> >>>>>>>> Julie
> >>>>>>>>
> >>>>>>>> On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> >>>>>>>> met_help at ucar.edu> wrote:
> >>>>>>>>
> >>>>>>>>>
> >>>>>>>>> <URL:
> >>>>>>>>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >>>>>>>>>
> >>>>>>>>> Hi Julie, the output for the command is,
> >>>>>>>>>
> >>>>>>>>> hrothgar:/met-5.0$ printenv | grep met
> >>>>>>>>> PWD=/home/brburgha/met-5.0
> >>>>>>>>>
> >>>>>>>>> ...unless you wanted this,
> >>>>>>>>>
> >>>>>>>>> hrothgar:/met-5.0$ printenv | grep MET*
> >>>>>>>>> HOSTNAME=hrothgar.hpcc.ttu.edu ANT_HOME=/opt/rocks
> >>>>>>>>> JAVA_HOME=/usr/java/latest
> >>>>>>>>> P4_GLOBMEMSIZE=83886080
> >>>>>>>>> MAGICK_HOME=/lustre/work/apps/ImageMagick
> >>>>>>>>> HOME=/home/brburgha
> >>>>>>>>> LOGNAME=brburgha
> >>>>>>>>> G_BROKEN_FILENAMES=1
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> Brock
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> -----Original Message-----
> >>>>>>>>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> >>>>>>>>> Sent: Tuesday, November 18, 2014 10:38 AM
> >>>>>>>>> To: Burghardt, Brock
> >>>>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
> >>>>>>>>> V4.1
> >>>>>>>>>
> >>>>>>>>> Hi Brock.  Thank you for getting the latest version of
MET.
> >>>>>>>>> I see
> >>>>>>> you're
> >>>>>>>>> still problems with MET finding netcdf.hh.
> >>>>>>>>>
> >>>>>>>>> Could you please run "printenv | grep MET" and send me the
> >>> output?
> >>>>>>>>>
> >>>>>>>>> Thanks,
> >>>>>>>>> Julie
> >>>>>>>>>
> >>>>>>>>> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
> >>>>>>>>> met_help at ucar.edu>
> >>>>>>>>> wrote:
> >>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>> <URL:
> >>>>>>>>>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> >>>>
> >>>>>>>>>>
> >>>>>>>>>> Hi Julie,
> >>>>>>>>>>
> >>>>>>>>>> Thanks for getting back to me so soon and explaining.
> >>>>>>>>>> Removing the static flag did indeed help in compiling the
> >>>>>>>>>> executables I needed
> >>>>>> but
> >>>>>>>>>> it did not completely compile (another ld: missing
> >>>>>>>>>> library error
> >>>>>> came
> >>>>>>>>>> up in the pb2nc build). However, I took your advice and
> >>> started
> >>>>>>>>>> over with V5.0 and got that running, but the compile also
> >>>>>>>>>> failed before making executables. The error was for not
> >>>>>>>>>> being able to open
> >>>>>> netcdf.hh
> >>>>>>>>>> (make_install.log attached). I added all the met_<lib>
> >>>>>>>>>> arguments at command line and checked they pointed to
> >>>>>>>>>> right directory, and the configure step ran without
> >>>>>>>>>> error. I
> >>> verified
> >>>>>>>>>> I do have read
> >>>>>> permission
> >>>>>>>>> for netcdf.hh, any ideas on this.
> >>>>>>>>>>
> >>>>>>>>>> Thanks,
> >>>>>>>>>> Brock
> >>>>>>>>>>
> >>>>>>>>>> -----Original Message-----
> >>>>>>>>>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> >>>>>>>>>> Sent: Monday, November 17, 2014 4:34 PM
> >>>>>>>>>> To: Burghardt, Brock
> >>>>>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
> >>> V4.1
> >>>>>>>>>>
> >>>>>>>>>> Hi Brock.
> >>>>>>>>>>
> >>>>>>>>>> I see that you are having problems building METv4.1 with
> >>>>>>>>>> an intel compiler.  Thank you for attaching your
> >>>>>>>>>> user_defs.mk
> >>> and
> >>>>>>>>>> make.log
> >>>>>>>>> files.
> >>>>>>>>>>
> >>>>>>>>>> First, I wanted to mention that METv5.0 was release in
> >>>>>>>>>> early September, which does include bug fixes for
> >>>>>>>>>> METv4.1.  You can view
> >>>>>> the
> >>>>>>>>>> release notes for METv5.0 here:
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>
> >>>
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_rele
> >>>>>>> as
> >>>>>>>>>> e_notes.php
> >>>>>>>>>>
> >>>>>>>>>> For now, let's try to overcome this error with METv4.1:
> >>>>>>>>>>
> >>>>>>>>>> ld: cannot find -lm
> >>>>>>>>>>
> >>>>>>>>>> We've had this issue come up with some other users in the
> >>> past.
> >>>>>>>>>> By default, MET does static linking - meaning that it
> >>>>>>>>>> links in the library files on which it depends at compile
> >>>>>>>>>> time. The other type of linking is call dynamic linking -
> >>>>>>>>>> for which the library files on
> >>>>>> which
> >>>>>>>>>> it depends are linked in at runtime. Other users have had
> >>>>>>>>>> success compiling MET with dynamic linking instead. Open
> >>>>>>>>>> up
> >>> and
> >>>>>>>>>> edit the top-level user_defs.mk file, by commenting out
> >>>>>>>>>> or removing
> >>>>>> references
> >>>>>>>>>> to "-static". They show up on the CXX_FLAGS and FC_FLAGS
> >>> lines.
> >>>>>>>>>>
> >>>>>>>>>> Then try recompiling MET, being sure to execute a "make
> >> clean"
> >>>>>> first.
> >>>>>>>>>>
> >>>>>>>>>> Thanks,
> >>>>>>>>>> Julie
> >>>>>>>>>>
> >>>>>>>>>> On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT
> >>>>>>>>>> < met_help at ucar.edu>
> >>>>>>>>>> wrote:
> >>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>> Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> >>>>>>>>>>> Transaction: Ticket created by brock.burghardt at ttu.edu
> >>>>>>>>>>>       Queue: met_help
> >>>>>>>>>>>     Subject: Issue compiling MET V4.1
> >>>>>>>>>>>       Owner: Nobody
> >>>>>>>>>>>  Requestors: brock.burghardt at ttu.edu
> >>>>>>>>>>>      Status: new
> >>>>>>>>>>> Ticket <URL:
> >>>>>>>>>>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> >>>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>> Hello,
> >>>>>>>>>>>
> >>>>>>>>>>> I'm trying to compile MET V4.1 using NETCDF V3.6.3 with
> >>>>>>>>>>> an intel compiler (definitions and log file attachted).
> >>>>>>>>>>> I've gone through
> >>>>>> all
> >>>>>>>>>>> suggestions in the appendix of the users guide but
> >>>>>>>>>>> still get an error while trying to make the
ensemble_stat
> >> executable.
> >>>>>>>>>>> Is there
> >>>>>> a
> >>>>>>>>>>> liblm.so or .a required for this build?
> >>>>>>>>>>>
> >>>>>>>>>>> Thanks,
> >>>>>>>>>>> Brock
> >>>>>>>>>>>
> >>>>>>>>>>> ld: cannot find -lm
> >>>>>>>>>>> make[4]: *** [ensemble_stat] Error 1
> >>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>> Brock Burghardt
> >>>>>>>>>>> Atmospheric Science PhD Student Dept. of Geosciences
> >>>>>>>>>>> Texas Tech University
> >>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>> --
> >>>>>>>>>> Julie Prestopnik
> >>>>>>>>>> National Center for Atmospheric Research Research
> >>>>>>>>>> Applications Laboratory
> >>>>>>>>>> Phone: 303.497.8399
> >>>>>>>>>> Email: jpresto at ucar.edu
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> --
> >>>>>>>>> Julie Prestopnik
> >>>>>>>>> National Center for Atmospheric Research Research
> >>>>>>>>> Applications
> >>>>>>> Laboratory
> >>>>>>>>> Phone: 303.497.8399
> >>>>>>>>> Email: jpresto at ucar.edu
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>
> >>>>>>>>
> >>>>>>>> --
> >>>>>>>> Julie Prestopnik
> >>>>>>>> National Center for Atmospheric Research Research
> >>>>>>>> Applications Laboratory
> >>>>>>>> Phone: 303.497.8399
> >>>>>>>> Email: jpresto at ucar.edu
> >>>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>> --
> >>>>>>> Julie Prestopnik
> >>>>>>> National Center for Atmospheric Research Research
Applications
> >>>>>>> Laboratory
> >>>>>>> Phone: 303.497.8399
> >>>>>>> Email: jpresto at ucar.edu
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>
> >>>>>>
> >>>>>> --
> >>>>>> Julie Prestopnik
> >>>>>> National Center for Atmospheric Research Research
Applications
> >>>>>> Laboratory
> >>>>>> Phone: 303.497.8399
> >>>>>> Email: jpresto at ucar.edu
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>
> >>>>>
> >>>>> --
> >>>>> Julie Prestopnik
> >>>>> National Center for Atmospheric Research Research Applications
> >>> Laboratory
> >>>>> Phone: 303.497.8399
> >>>>> Email: jpresto at ucar.edu
> >>>>>
> >>>>>
> >>>>>
> >>>>
> >>>>
> >>>> --
> >>>> Julie Prestopnik
> >>>> National Center for Atmospheric Research Research Applications
> >>>> Laboratory
> >>>> Phone: 303.497.8399
> >>>> Email: jpresto at ucar.edu
> >>>>
> >>>>
> >>>
> >>>
> >>>
> >>
> >>
> >>
> >
>
>
>

--
Julie Prestopnik
National Center for Atmospheric Research
Research Applications Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: Issue compiling MET V4.1
From: Burghardt, Brock
Time: Tue Dec 16 14:00:14 2014

Hi John and Julie,

I successfully built metv5.0 and received no errors in the make test
step. I did not define my CXX and F77 paths and let the autoconfig do
so, which seemed to help. I'll attach the config.log in case you are
interested but it appears autoconfig set CXX='icpc' and CC='icc', with
F77 still 'ifort'.

Thanks again for the help from both of you,

Brock

-----Original Message-----
From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
Sent: Tuesday, December 16, 2014 11:13 AM
To: Burghardt, Brock
Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1

Hi Brock.

gfortran is the GNU Fortran compiler.  When compiling MET, you should
use the same family of compilers. While I haven't actually tried it,
mixing GNU compilers with Intel compilers isn't a good idea.

Since you're using icc please try compiling MET using ifort, the Intel
Fortran compiler.  Please note that when you compile the required MET
libraries, you should do so using the Intel compilers as well. Here's
the tutorial page that talks about that:
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/compilation/req_libs.php

Please let us know if you have any other questions or encounter any
problems.

Thanks,
Julie

On Tue, Dec 16, 2014 at 9:56 AM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:
>
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi John,
>
> I got 14 of the 23 executables built, but another crash during the
> pb2nc build. There are numerous 'undefined reference to’ statements
> that popped up in the make_install.log, some of which had gfortran
in
> the statement, so i cleaned and tried with F77 pointing to the
> gfortran libraries but got the same error. Any ideas on this?
>
>
> Thanks,
> Brock
> Brock
>
> > On Dec 15, 2014, at 3:35 PM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
> >
> > Brock,
> >
> > Yep, this one is easy to solve.  Change...
> > FROM: export MET_BUFR=$BUFRLIB
> > TO:      export MET_BUFRLIB=$BUFRLIB
> >
> > Why is that necessary?  Because BUFRLIB doesn't follow the
standard
> > convention of having "include" and "lib" sub-directories.  By
> > setting MET_BUFR={path}, MET looks for the library files in
> > {path}/lib, but that doesn't exist!  So we instead set,
MET_BUFRLIB
> > to tell it where the BUFR library files live.
> >
> > Sorry I didn't catch that wrinkle when looking at what you sent me
> before.
> > This is mentioned in the top-level METv5.0 README file.
> >
> > Please give that a shot and let me know.
> >
> > Thanks,
> > John
> >
> >
> >
> > On Mon, Dec 15, 2014 at 2:10 PM, Burghardt, Brock via RT <
> met_help at ucar.edu>
> > wrote:
> >>
> >>
> >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >>
> >> Hi John,
> >>
> >> I completely forgot to export the set paths, sorry about that.
Upon
> >> exporting the paths I get a partially successful build with 11 of
> >> the 23 executables created.
> >> The error appears to occur when looking for the -lbufr library.
> >> Would
> this
> >> have something to do with the BUFRLIB directory?
> >>
> >> Thanks,
> >> Brock
> >>
> >>
> >>
> >> -----Original Message-----
> >> From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> >> Sent: Friday, December 12, 2014 4:12 PM
> >> To: Burghardt, Brock
> >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >>
> >> Brock,
> >>
> >> Thanks for sending those log files.  Looking at line 3243 of
> config.log, I
> >> see:
> >>   ac_cv_env_MET_NETCDF_set=
> >>   ac_cv_env_MET_NETCDF_value=
> >>
> >> This means that configure doesn't think that MET_NETCDF is set.
> >> Looking at the syntax of the commands you sent, I assume you're
> >> using the bash
> or
> >> ksh shell, rather than cshell.  That's fine, but you need to use
> "export"
> >> when setting variables to pass those definitions down to sub-
processes.
> >> Please try the following:
> >>
> >> export MET_NETCDF=${NETCDF}
> >> export MET_BUFR=$BUFRLIB
> >> export MET_GSL=/lustre/work/apps/gsl export
> CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
> >> export
> F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort
> >>
> >> And then run "configure".  Then run the following command and
make
> >> sure that NetCDF is actually set:
> >>   grep ac_cv_env_MET_NETCDF_value= config.log
> >>
> >> If so, proceed to "make install" and "make test".
> >>
> >> Sorry for all of the headaches you've had to struggle through!
> >>
> >> John
> >>
> >> On Thu, Dec 11, 2014 at 12:21 PM, Burghardt, Brock via RT <
> >> met_help at ucar.edu
> >>> wrote:
> >>>
> >>>
> >>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >>>
> >>> Hi John,
> >>>
> >>> So I set the CXX to the available intel c++ compiler, however I
am
> >>> still receiving the same error set.
> >>> Below are the other paths I set before configure and make:
> >>> MET_NETCDF=${NETCDF}
> >>> MET_BUFR=$BUFRLIB
> >>> MET_GSL=/lustre/work/apps/gsl
> >>>
CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
> >>>
F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifo
> >>> rt
> >>>
> >>> Attached are both the config and make logs. I noticed that in my
> >>> config log that when it is looking for the icc version it gets
an
> >>> error because it is invoking 'icc -qversion' but should be 'icc
> >>> -v' if I want to check it on the command line. Do you think this
> >>> could be a
> >> cause?
> >>>
> >>> Thank you,
> >>> Brock
> >>>
> >>> ________________________________________
> >>> From: John Halley Gotway via RT [met_help at ucar.edu]
> >>> Sent: Tuesday, December 09, 2014 11:32 AM
> >>> To: Burghardt, Brock
> >>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >>>
> >>> Hello Brock,
> >>>
> >>> I work with Julie Prestopnik and I see that you're having
trouble
> >>> compiling MET.  I took a look at the log file you sent and see
> >>> that you're using the intel c++ icpc compiler.  In the past when
> >>> I've compiled MET using intel compilers, I've always used the
> >>> "icc" c++ compiler.  I just tried compiling MET using icpc and
run
> >>> into all kinds of problems.  Is icc available on your system
(type: which icc)?
> >>> If so, try using that by setting the CXX environment variable
> >>> before
> >> running the "configure" command for MET:
> >>>  setenv CXX `which icc`
> >>>
> >>> Also, I looked in the logfile you sent at the specific line
where
> >>> the error occurred and compared that to the same line in the
> >>> output created when I compile MET.  None of the external library
> >>> links are being passed on the compilation line... specifically
> >>> between "-I../../../src/libcode/vx_tc_util" and "-DBLOCK4" we
> >>> should see links for NetCDF and other external library
dependencies.
> >>>
> >>> Not sure why this is happening but it may be related to the
choice
> >>> of compiler.  Please let us know if you have icc available, and
if
> >>> so, try using that one.
> >>>
> >>> Thanks,
> >>> John
> >>>
> >>> On Mon, Dec 8, 2014 at 4:53 PM, Julie Prestopnik via RT
> >>> <met_help at ucar.edu
> >>>>
> >>> wrote:
> >>>
> >>>>
> >>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >>>>
> >>>> Hi Brock.  I'm sorry to see that MET can't seem to find
netcdf.hh.
> >>>> Could you please set up the environment variables on the
command
> >>>> line, and try
> >>> to
> >>>> build and install MET on the command line and then send me the
> >>>> make_install.log file from that run along with the output of
> >>>> "printenv | grep MET".  I looked at this make_install.log,
paying
> >>>> particular
> >>> attention
> >>>> to what is being looked at for the include files.  I don't know
> >>>> where
> >>> your
> >>>> netcdf include files are, so I can't check for that location.
> >>>>
> >>>> If your environment variable is set correctly, it should be
> >>>> listed in the output of "printenv | grep MET" and also in -I in
> >>>> the make_install.log file.  If MET still can't find it after we
> >>>> check those things, we'll have to look into it further.  I know
> >>>> that certain recent versions of NetCDF (e.g. v4.1.3) no longer
> >>>> include netcdf.hh, but I believe that 3.6.3,
> >>> which
> >>>> you said you are using, does have netcdf.hh.
> >>>>
> >>>> Sending this additional info from an install attempt with the
> >>>> environment variables set, should help us better figure out
what
> >>>> the
> >> problem is.
> >>>>
> >>>> Thanks,
> >>>> Julie
> >>>>
> >>>> On Mon, Dec 8, 2014 at 2:03 PM, Burghardt, Brock via RT <
> >>> met_help at ucar.edu
> >>>>>
> >>>> wrote:
> >>>>
> >>>>>
> >>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> >>>>>
> >>>>> Hi Julie,
> >>>>> Sorry for not getting back to you sooner (finals just wrapped
up).
> >>>>> I've attached the make log.
> >>>>> The printenv statement does not give anything for "MET",
however
> >>>>> when I set my MET_<LIB/INC> paths at command line I can print
> >>>>> them, giving the proper paths.
> >>>>>
> >>>>> Thanks,
> >>>>> Brock
> >>>>>
> >>>>> -----Original Message-----
> >>>>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> >>>>> Sent: Friday, December 05, 2014 9:56 AM
> >>>>> To: Burghardt, Brock
> >>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> >>>>>
> >>>>> Hi Brock.  Could you please send me your make_install.log file
> >>>>> and
> >>> could
> >>>>> you please run "printenv | grep MET" again and send me the
output?
> >>>>>
> >>>>> Thanks,
> >>>>> Julie
> >>>>>
> >>>>> On Thu, Dec 4, 2014 at 5:09 PM, Burghardt, Brock via RT <
> >>>> met_help at ucar.edu
> >>>>>>
> >>>>> wrote:
> >>>>>
> >>>>>>
> >>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> >>>>>>
> >>>>>> Hi Julie,
> >>>>>>
> >>>>>> Still no luck with loading the intel environment (soft add
> >>>>>> intel in
> >>> my
> >>>>>> case) with the error developing with netcdf.hh in the make
> >>>>>> install...any other ideas?
> >>>>>>
> >>>>>> Thanks, I appreciate the help.
> >>>>>>
> >>>>>> Brock
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> ________________________________________
> >>>>>> From: Julie Prestopnik via RT [met_help at ucar.edu]
> >>>>>> Sent: Tuesday, December 02, 2014 11:02 AM
> >>>>>> To: Burghardt, Brock
> >>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> >>>>>>
> >>>>>> Hi Brock.  I'm sorry to hear that you are still having
problems
> >>>>>> compiling METv5.0.  Thank you for sending the config.log
file.
> >>>>>>
> >>>>>> Please try running the following command:
> >>>>>>
> >>>>>> make clean
> >>>>>> module load intel
> >>>>>>
> >>>>>> and then run the configure command and the make install
command.
> >>>>>>
> >>>>>> Please let me know if that works.
> >>>>>>
> >>>>>> Thanks,
> >>>>>> Julie
> >>>>>>
> >>>>>> On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
> >>>>>> <met_help at ucar.edu
> >>>>>>>
> >>>>>> wrote:
> >>>>>>
> >>>>>>>
> >>>>>>> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> >>>>>>>>
> >>>>>>>
> >>>>>>> Hi Julie,
> >>>>>>>
> >>>>>>> I've still been unable to compile METv5.0...independent of
> >>>>>>> whether
> >>> I
> >>>>>>> specify the parent or both the parent/lib parent/inc paths
for
> >>>>>>> netcdf, bufrlib, and GSL, I get an error with it trying to
> >>>>>>> open netcdf.hh
> >>>>>>>
> >>>>>>> I noticed two paths to netcdf got set in the config.log
> >>>>>>> (attached), one I set and another on the system, removing
the
> >>>>>>> one I did not define still
> >>>>>> did
> >>>>>>> not give a successful build.
> >>>>>>>
> >>>>>>> I noticed I had errors in the config.log for many of the
> >>>>>>> compilers where if tries to call, for instance, icpc
-qversion
> >>>>>>> which should
> >>> be
> >>>>>>> icpc
> >>>>>> -V....I
> >>>>>>> looked through all the config and configure scripts and
could
> >>>>>>> not find
> >>>>>> the
> >>>>>>> explicit icpc -qversion call to change to -V...do you think
> >>>>>>> this could be the issue?
> >>>>>>>
> >>>>>>> Thank you for following up,
> >>>>>>> Brock
> >>>>>>>
> >>>>>>> ________________________________________
> >>>>>>> From: Julie Prestopnik via RT [met_help at ucar.edu]
> >>>>>>> Sent: Monday, December 01, 2014 9:56 AM
> >>>>>>> To: Burghardt, Brock
> >>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> >>>>>>>
> >>>>>>> Hi Brock.  I just wanted to follow up - Were able to get a
> >>>>>>> successful
> >>>>>> build
> >>>>>>> of METv5.0?
> >>>>>>>
> >>>>>>> Thanks,
> >>>>>>> Julie
> >>>>>>>
> >>>>>>> On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
> >>>>>>> <jpresto at ucar.edu>
> >>>>>>> wrote:
> >>>>>>>
> >>>>>>>> Hi Brock.
> >>>>>>>>
> >>>>>>>> Thanks you for sending the output.  I was looking to make
> >>>>>>>> sure that the MET environment variables, like MET_NETCDF or
> >>>>>>>> MET_NETCDFINC and MET_NETCDFLIB, were set correctly.
> >>>>>>>>
> >>>>>>>> Please take a look at the following page and follow the
> >>>>>>>> instructions on compiling MET:
> >>>>>>>>
> >>>>>>>>
> >>>>>>>
> >>>>>>
> >>>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/
> >>> comp
> >>>>>> ilation/configure.php
> >>>>>>>>
> >>>>>>>> This will help guide you in setting up the required
> >>>>>>>> environment
> >>>>>> variables
> >>>>>>>> and will hopefully resolve the problem you are experiencing
> >>>>>>>> with MET
> >>>>>> not
> >>>>>>>> being able to find netcdf.hh.
> >>>>>>>>
> >>>>>>>> Please let me know if you have any questions, and please
let
> >>>>>>>> me know if you get a successful build after setting up the
> >>> environment
> >>>>> variables.
> >>>>>>>>
> >>>>>>>> Thanks,
> >>>>>>>> Julie
> >>>>>>>>
> >>>>>>>> On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT <
> >>>>>>>> met_help at ucar.edu> wrote:
> >>>>>>>>
> >>>>>>>>>
> >>>>>>>>> <URL:
> >>>>>>>>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >>>>>>>>>
> >>>>>>>>> Hi Julie, the output for the command is,
> >>>>>>>>>
> >>>>>>>>> hrothgar:/met-5.0$ printenv | grep met
> >>>>>>>>> PWD=/home/brburgha/met-5.0
> >>>>>>>>>
> >>>>>>>>> ...unless you wanted this,
> >>>>>>>>>
> >>>>>>>>> hrothgar:/met-5.0$ printenv | grep MET*
> >>>>>>>>> HOSTNAME=hrothgar.hpcc.ttu.edu ANT_HOME=/opt/rocks
> >>>>>>>>> JAVA_HOME=/usr/java/latest
> >>>>>>>>> P4_GLOBMEMSIZE=83886080
> >>>>>>>>> MAGICK_HOME=/lustre/work/apps/ImageMagick
> >>>>>>>>> HOME=/home/brburgha
> >>>>>>>>> LOGNAME=brburgha
> >>>>>>>>> G_BROKEN_FILENAMES=1
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> Brock
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> -----Original Message-----
> >>>>>>>>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> >>>>>>>>> Sent: Tuesday, November 18, 2014 10:38 AM
> >>>>>>>>> To: Burghardt, Brock
> >>>>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
> >>>>>>>>> V4.1
> >>>>>>>>>
> >>>>>>>>> Hi Brock.  Thank you for getting the latest version of
MET.
> >>>>>>>>> I see
> >>>>>>> you're
> >>>>>>>>> still problems with MET finding netcdf.hh.
> >>>>>>>>>
> >>>>>>>>> Could you please run "printenv | grep MET" and send me the
> >>> output?
> >>>>>>>>>
> >>>>>>>>> Thanks,
> >>>>>>>>> Julie
> >>>>>>>>>
> >>>>>>>>> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT <
> >>>>>>>>> met_help at ucar.edu>
> >>>>>>>>> wrote:
> >>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>> <URL:
> >>>>>>>>>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> >>>>
> >>>>>>>>>>
> >>>>>>>>>> Hi Julie,
> >>>>>>>>>>
> >>>>>>>>>> Thanks for getting back to me so soon and explaining.
> >>>>>>>>>> Removing the static flag did indeed help in compiling the
> >>>>>>>>>> executables I needed
> >>>>>> but
> >>>>>>>>>> it did not completely compile (another ld: missing
library
> >>>>>>>>>> error
> >>>>>> came
> >>>>>>>>>> up in the pb2nc build). However, I took your advice and
> >>> started
> >>>>>>>>>> over with V5.0 and got that running, but the compile also
> >>>>>>>>>> failed before making executables. The error was for not
> >>>>>>>>>> being able to open
> >>>>>> netcdf.hh
> >>>>>>>>>> (make_install.log attached). I added all the met_<lib>
> >>>>>>>>>> arguments at command line and checked they pointed to
right
> >>>>>>>>>> directory, and the configure step ran without error. I
> >>> verified
> >>>>>>>>>> I do have read
> >>>>>> permission
> >>>>>>>>> for netcdf.hh, any ideas on this.
> >>>>>>>>>>
> >>>>>>>>>> Thanks,
> >>>>>>>>>> Brock
> >>>>>>>>>>
> >>>>>>>>>> -----Original Message-----
> >>>>>>>>>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> >>>>>>>>>> Sent: Monday, November 17, 2014 4:34 PM
> >>>>>>>>>> To: Burghardt, Brock
> >>>>>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
> >>> V4.1
> >>>>>>>>>>
> >>>>>>>>>> Hi Brock.
> >>>>>>>>>>
> >>>>>>>>>> I see that you are having problems building METv4.1 with
an
> >>>>>>>>>> intel compiler.  Thank you for attaching your
user_defs.mk
> >>> and
> >>>>>>>>>> make.log
> >>>>>>>>> files.
> >>>>>>>>>>
> >>>>>>>>>> First, I wanted to mention that METv5.0 was release in
> >>>>>>>>>> early September, which does include bug fixes for
METv4.1.
> >>>>>>>>>> You can view
> >>>>>> the
> >>>>>>>>>> release notes for METv5.0 here:
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>
> >>>
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_re
> >>> le
> >>>>>>> as
> >>>>>>>>>> e_notes.php
> >>>>>>>>>>
> >>>>>>>>>> For now, let's try to overcome this error with METv4.1:
> >>>>>>>>>>
> >>>>>>>>>> ld: cannot find -lm
> >>>>>>>>>>
> >>>>>>>>>> We've had this issue come up with some other users in the
> >>> past.
> >>>>>>>>>> By default, MET does static linking - meaning that it
links
> >>>>>>>>>> in the library files on which it depends at compile time.
> >>>>>>>>>> The other type of linking is call dynamic linking - for
> >>>>>>>>>> which the library files on
> >>>>>> which
> >>>>>>>>>> it depends are linked in at runtime. Other users have had
> >>>>>>>>>> success compiling MET with dynamic linking instead. Open
up
> >>> and
> >>>>>>>>>> edit the top-level user_defs.mk file, by commenting out
or
> >>>>>>>>>> removing
> >>>>>> references
> >>>>>>>>>> to "-static". They show up on the CXX_FLAGS and FC_FLAGS
> >>> lines.
> >>>>>>>>>>
> >>>>>>>>>> Then try recompiling MET, being sure to execute a "make
> >> clean"
> >>>>>> first.
> >>>>>>>>>>
> >>>>>>>>>> Thanks,
> >>>>>>>>>> Julie
> >>>>>>>>>>
> >>>>>>>>>> On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via RT
<
> >>>>>>>>>> met_help at ucar.edu>
> >>>>>>>>>> wrote:
> >>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>> Mon Nov 17 13:55:16 2014: Request 69783 was acted upon.
> >>>>>>>>>>> Transaction: Ticket created by brock.burghardt at ttu.edu
> >>>>>>>>>>>       Queue: met_help
> >>>>>>>>>>>     Subject: Issue compiling MET V4.1
> >>>>>>>>>>>       Owner: Nobody
> >>>>>>>>>>>  Requestors: brock.burghardt at ttu.edu
> >>>>>>>>>>>      Status: new
> >>>>>>>>>>> Ticket <URL:
> >>>>>>>>>>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> >>>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>> Hello,
> >>>>>>>>>>>
> >>>>>>>>>>> I'm trying to compile MET V4.1 using NETCDF V3.6.3 with
an
> >>>>>>>>>>> intel compiler (definitions and log file attachted).
> >>>>>>>>>>> I've gone through
> >>>>>> all
> >>>>>>>>>>> suggestions in the appendix of the users guide but still
> >>>>>>>>>>> get an error while trying to make the ensemble_stat
> >> executable.
> >>>>>>>>>>> Is there
> >>>>>> a
> >>>>>>>>>>> liblm.so or .a required for this build?
> >>>>>>>>>>>
> >>>>>>>>>>> Thanks,
> >>>>>>>>>>> Brock
> >>>>>>>>>>>
> >>>>>>>>>>> ld: cannot find -lm
> >>>>>>>>>>> make[4]: *** [ensemble_stat] Error 1
> >>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>> Brock Burghardt
> >>>>>>>>>>> Atmospheric Science PhD Student Dept. of Geosciences
Texas
> >>>>>>>>>>> Tech University
> >>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>> --
> >>>>>>>>>> Julie Prestopnik
> >>>>>>>>>> National Center for Atmospheric Research Research
> >>>>>>>>>> Applications Laboratory
> >>>>>>>>>> Phone: 303.497.8399
> >>>>>>>>>> Email: jpresto at ucar.edu
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> --
> >>>>>>>>> Julie Prestopnik
> >>>>>>>>> National Center for Atmospheric Research Research
> >>>>>>>>> Applications
> >>>>>>> Laboratory
> >>>>>>>>> Phone: 303.497.8399
> >>>>>>>>> Email: jpresto at ucar.edu
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>
> >>>>>>>>
> >>>>>>>> --
> >>>>>>>> Julie Prestopnik
> >>>>>>>> National Center for Atmospheric Research Research
> >>>>>>>> Applications Laboratory
> >>>>>>>> Phone: 303.497.8399
> >>>>>>>> Email: jpresto at ucar.edu
> >>>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>> --
> >>>>>>> Julie Prestopnik
> >>>>>>> National Center for Atmospheric Research Research
Applications
> >>>>>>> Laboratory
> >>>>>>> Phone: 303.497.8399
> >>>>>>> Email: jpresto at ucar.edu
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>
> >>>>>>
> >>>>>> --
> >>>>>> Julie Prestopnik
> >>>>>> National Center for Atmospheric Research Research
Applications
> >>>>>> Laboratory
> >>>>>> Phone: 303.497.8399
> >>>>>> Email: jpresto at ucar.edu
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>
> >>>>>
> >>>>> --
> >>>>> Julie Prestopnik
> >>>>> National Center for Atmospheric Research Research Applications
> >>> Laboratory
> >>>>> Phone: 303.497.8399
> >>>>> Email: jpresto at ucar.edu
> >>>>>
> >>>>>
> >>>>>
> >>>>
> >>>>
> >>>> --
> >>>> Julie Prestopnik
> >>>> National Center for Atmospheric Research Research Applications
> >>>> Laboratory
> >>>> Phone: 303.497.8399
> >>>> Email: jpresto at ucar.edu
> >>>>
> >>>>
> >>>
> >>>
> >>>
> >>
> >>
> >>
> >
>
>
>

--
Julie Prestopnik
National Center for Atmospheric Research Research Applications
Laboratory
Phone: 303.497.8399
Email: jpresto at ucar.edu

------------------------------------------------
Subject: Issue compiling MET V4.1
From: John Halley Gotway
Time: Tue Dec 16 14:28:03 2014

Great, glad you were able to figure it out!  I'll go ahead and resolve
this
ticket.  Just let us know if any more issue or questions arise in your
use
of MET.

Thanks,
John

On Tue, Dec 16, 2014 at 2:00 PM, Burghardt, Brock via RT
<met_help at ucar.edu>
wrote:
>
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
>
> Hi John and Julie,
>
> I successfully built metv5.0 and received no errors in the make test
step.
> I did not define my CXX and F77 paths and let the autoconfig do so,
which
> seemed to help. I'll attach the config.log in case you are
interested but
> it appears autoconfig set CXX='icpc' and CC='icc', with F77 still
'ifort'.
>
> Thanks again for the help from both of you,
>
> Brock
>
>
>
> -----Original Message-----
> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> Sent: Tuesday, December 16, 2014 11:13 AM
> To: Burghardt, Brock
> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
>
> Hi Brock.
>
> gfortran is the GNU Fortran compiler.  When compiling MET, you
should use
> the same family of compilers. While I haven't actually tried it,
mixing GNU
> compilers with Intel compilers isn't a good idea.
>
> Since you're using icc please try compiling MET using ifort, the
Intel
> Fortran compiler.  Please note that when you compile the required
MET
> libraries, you should do so using the Intel compilers as well.
Here's the
> tutorial page that talks about that:
>
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/compilation/req_libs.php
>
> Please let us know if you have any other questions or encounter any
> problems.
>
> Thanks,
> Julie
>
>
> On Tue, Dec 16, 2014 at 9:56 AM, Burghardt, Brock via RT <
> met_help at ucar.edu>
> wrote:
> >
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> >
> > Hi John,
> >
> > I got 14 of the 23 executables built, but another crash during the
> > pb2nc build. There are numerous 'undefined reference to’
statements
> > that popped up in the make_install.log, some of which had gfortran
in
> > the statement, so i cleaned and tried with F77 pointing to the
> > gfortran libraries but got the same error. Any ideas on this?
> >
> >
> > Thanks,
> > Brock
> > Brock
> >
> > > On Dec 15, 2014, at 3:35 PM, John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> > >
> > > Brock,
> > >
> > > Yep, this one is easy to solve.  Change...
> > > FROM: export MET_BUFR=$BUFRLIB
> > > TO:      export MET_BUFRLIB=$BUFRLIB
> > >
> > > Why is that necessary?  Because BUFRLIB doesn't follow the
standard
> > > convention of having "include" and "lib" sub-directories.  By
> > > setting MET_BUFR={path}, MET looks for the library files in
> > > {path}/lib, but that doesn't exist!  So we instead set,
MET_BUFRLIB
> > > to tell it where the BUFR library files live.
> > >
> > > Sorry I didn't catch that wrinkle when looking at what you sent
me
> > before.
> > > This is mentioned in the top-level METv5.0 README file.
> > >
> > > Please give that a shot and let me know.
> > >
> > > Thanks,
> > > John
> > >
> > >
> > >
> > > On Mon, Dec 15, 2014 at 2:10 PM, Burghardt, Brock via RT <
> > met_help at ucar.edu>
> > > wrote:
> > >>
> > >>
> > >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > >>
> > >> Hi John,
> > >>
> > >> I completely forgot to export the set paths, sorry about that.
Upon
> > >> exporting the paths I get a partially successful build with 11
of
> > >> the 23 executables created.
> > >> The error appears to occur when looking for the -lbufr library.
> > >> Would
> > this
> > >> have something to do with the BUFRLIB directory?
> > >>
> > >> Thanks,
> > >> Brock
> > >>
> > >>
> > >>
> > >> -----Original Message-----
> > >> From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> > >> Sent: Friday, December 12, 2014 4:12 PM
> > >> To: Burghardt, Brock
> > >> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > >>
> > >> Brock,
> > >>
> > >> Thanks for sending those log files.  Looking at line 3243 of
> > config.log, I
> > >> see:
> > >>   ac_cv_env_MET_NETCDF_set=
> > >>   ac_cv_env_MET_NETCDF_value=
> > >>
> > >> This means that configure doesn't think that MET_NETCDF is set.
> > >> Looking at the syntax of the commands you sent, I assume you're
> > >> using the bash
> > or
> > >> ksh shell, rather than cshell.  That's fine, but you need to
use
> > "export"
> > >> when setting variables to pass those definitions down to
> sub-processes.
> > >> Please try the following:
> > >>
> > >> export MET_NETCDF=${NETCDF}
> > >> export MET_BUFR=$BUFRLIB
> > >> export MET_GSL=/lustre/work/apps/gsl export
> > CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
> > >> export
> >
F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifort
> > >>
> > >> And then run "configure".  Then run the following command and
make
> > >> sure that NetCDF is actually set:
> > >>   grep ac_cv_env_MET_NETCDF_value= config.log
> > >>
> > >> If so, proceed to "make install" and "make test".
> > >>
> > >> Sorry for all of the headaches you've had to struggle through!
> > >>
> > >> John
> > >>
> > >> On Thu, Dec 11, 2014 at 12:21 PM, Burghardt, Brock via RT <
> > >> met_help at ucar.edu
> > >>> wrote:
> > >>>
> > >>>
> > >>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> > >>>
> > >>> Hi John,
> > >>>
> > >>> So I set the CXX to the available intel c++ compiler, however
I am
> > >>> still receiving the same error set.
> > >>> Below are the other paths I set before configure and make:
> > >>> MET_NETCDF=${NETCDF}
> > >>> MET_BUFR=$BUFRLIB
> > >>> MET_GSL=/lustre/work/apps/gsl
> > >>>
CXX=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/icc
> > >>>
F77=/lustre/work/apps/intel/composer_xe_2013.1.117/bin/intel64/ifo
> > >>> rt
> > >>>
> > >>> Attached are both the config and make logs. I noticed that in
my
> > >>> config log that when it is looking for the icc version it gets
an
> > >>> error because it is invoking 'icc -qversion' but should be
'icc
> > >>> -v' if I want to check it on the command line. Do you think
this
> > >>> could be a
> > >> cause?
> > >>>
> > >>> Thank you,
> > >>> Brock
> > >>>
> > >>> ________________________________________
> > >>> From: John Halley Gotway via RT [met_help at ucar.edu]
> > >>> Sent: Tuesday, December 09, 2014 11:32 AM
> > >>> To: Burghardt, Brock
> > >>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET V4.1
> > >>>
> > >>> Hello Brock,
> > >>>
> > >>> I work with Julie Prestopnik and I see that you're having
trouble
> > >>> compiling MET.  I took a look at the log file you sent and see
> > >>> that you're using the intel c++ icpc compiler.  In the past
when
> > >>> I've compiled MET using intel compilers, I've always used the
> > >>> "icc" c++ compiler.  I just tried compiling MET using icpc and
run
> > >>> into all kinds of problems.  Is icc available on your system
(type:
> which icc)?
> > >>> If so, try using that by setting the CXX environment variable
> > >>> before
> > >> running the "configure" command for MET:
> > >>>  setenv CXX `which icc`
> > >>>
> > >>> Also, I looked in the logfile you sent at the specific line
where
> > >>> the error occurred and compared that to the same line in the
> > >>> output created when I compile MET.  None of the external
library
> > >>> links are being passed on the compilation line... specifically
> > >>> between "-I../../../src/libcode/vx_tc_util" and "-DBLOCK4" we
> > >>> should see links for NetCDF and other external library
dependencies.
> > >>>
> > >>> Not sure why this is happening but it may be related to the
choice
> > >>> of compiler.  Please let us know if you have icc available,
and if
> > >>> so, try using that one.
> > >>>
> > >>> Thanks,
> > >>> John
> > >>>
> > >>> On Mon, Dec 8, 2014 at 4:53 PM, Julie Prestopnik via RT
> > >>> <met_help at ucar.edu
> > >>>>
> > >>> wrote:
> > >>>
> > >>>>
> > >>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> > >>>>
> > >>>> Hi Brock.  I'm sorry to see that MET can't seem to find
netcdf.hh.
> > >>>> Could you please set up the environment variables on the
command
> > >>>> line, and try
> > >>> to
> > >>>> build and install MET on the command line and then send me
the
> > >>>> make_install.log file from that run along with the output of
> > >>>> "printenv | grep MET".  I looked at this make_install.log,
paying
> > >>>> particular
> > >>> attention
> > >>>> to what is being looked at for the include files.  I don't
know
> > >>>> where
> > >>> your
> > >>>> netcdf include files are, so I can't check for that location.
> > >>>>
> > >>>> If your environment variable is set correctly, it should be
> > >>>> listed in the output of "printenv | grep MET" and also in -I
in
> > >>>> the make_install.log file.  If MET still can't find it after
we
> > >>>> check those things, we'll have to look into it further.  I
know
> > >>>> that certain recent versions of NetCDF (e.g. v4.1.3) no
longer
> > >>>> include netcdf.hh, but I believe that 3.6.3,
> > >>> which
> > >>>> you said you are using, does have netcdf.hh.
> > >>>>
> > >>>> Sending this additional info from an install attempt with the
> > >>>> environment variables set, should help us better figure out
what
> > >>>> the
> > >> problem is.
> > >>>>
> > >>>> Thanks,
> > >>>> Julie
> > >>>>
> > >>>> On Mon, Dec 8, 2014 at 2:03 PM, Burghardt, Brock via RT <
> > >>> met_help at ucar.edu
> > >>>>>
> > >>>> wrote:
> > >>>>
> > >>>>>
> > >>>>> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > >>>>>
> > >>>>> Hi Julie,
> > >>>>> Sorry for not getting back to you sooner (finals just
wrapped up).
> > >>>>> I've attached the make log.
> > >>>>> The printenv statement does not give anything for "MET",
however
> > >>>>> when I set my MET_<LIB/INC> paths at command line I can
print
> > >>>>> them, giving the proper paths.
> > >>>>>
> > >>>>> Thanks,
> > >>>>> Brock
> > >>>>>
> > >>>>> -----Original Message-----
> > >>>>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > >>>>> Sent: Friday, December 05, 2014 9:56 AM
> > >>>>> To: Burghardt, Brock
> > >>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > >>>>>
> > >>>>> Hi Brock.  Could you please send me your make_install.log
file
> > >>>>> and
> > >>> could
> > >>>>> you please run "printenv | grep MET" again and send me the
output?
> > >>>>>
> > >>>>> Thanks,
> > >>>>> Julie
> > >>>>>
> > >>>>> On Thu, Dec 4, 2014 at 5:09 PM, Burghardt, Brock via RT <
> > >>>> met_help at ucar.edu
> > >>>>>>
> > >>>>> wrote:
> > >>>>>
> > >>>>>>
> > >>>>>> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783 >
> > >>>>>>
> > >>>>>> Hi Julie,
> > >>>>>>
> > >>>>>> Still no luck with loading the intel environment (soft add
> > >>>>>> intel in
> > >>> my
> > >>>>>> case) with the error developing with netcdf.hh in the make
> > >>>>>> install...any other ideas?
> > >>>>>>
> > >>>>>> Thanks, I appreciate the help.
> > >>>>>>
> > >>>>>> Brock
> > >>>>>>
> > >>>>>>
> > >>>>>>
> > >>>>>>
> > >>>>>> ________________________________________
> > >>>>>> From: Julie Prestopnik via RT [met_help at ucar.edu]
> > >>>>>> Sent: Tuesday, December 02, 2014 11:02 AM
> > >>>>>> To: Burghardt, Brock
> > >>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > >>>>>>
> > >>>>>> Hi Brock.  I'm sorry to hear that you are still having
problems
> > >>>>>> compiling METv5.0.  Thank you for sending the config.log
file.
> > >>>>>>
> > >>>>>> Please try running the following command:
> > >>>>>>
> > >>>>>> make clean
> > >>>>>> module load intel
> > >>>>>>
> > >>>>>> and then run the configure command and the make install
command.
> > >>>>>>
> > >>>>>> Please let me know if that works.
> > >>>>>>
> > >>>>>> Thanks,
> > >>>>>> Julie
> > >>>>>>
> > >>>>>> On Tue, Dec 2, 2014 at 9:35 AM, Burghardt, Brock via RT
> > >>>>>> <met_help at ucar.edu
> > >>>>>>>
> > >>>>>> wrote:
> > >>>>>>
> > >>>>>>>
> > >>>>>>> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > >>>>>>>>
> > >>>>>>>
> > >>>>>>> Hi Julie,
> > >>>>>>>
> > >>>>>>> I've still been unable to compile METv5.0...independent of
> > >>>>>>> whether
> > >>> I
> > >>>>>>> specify the parent or both the parent/lib parent/inc paths
for
> > >>>>>>> netcdf, bufrlib, and GSL, I get an error with it trying to
> > >>>>>>> open netcdf.hh
> > >>>>>>>
> > >>>>>>> I noticed two paths to netcdf got set in the config.log
> > >>>>>>> (attached), one I set and another on the system, removing
the
> > >>>>>>> one I did not define still
> > >>>>>> did
> > >>>>>>> not give a successful build.
> > >>>>>>>
> > >>>>>>> I noticed I had errors in the config.log for many of the
> > >>>>>>> compilers where if tries to call, for instance, icpc
-qversion
> > >>>>>>> which should
> > >>> be
> > >>>>>>> icpc
> > >>>>>> -V....I
> > >>>>>>> looked through all the config and configure scripts and
could
> > >>>>>>> not find
> > >>>>>> the
> > >>>>>>> explicit icpc -qversion call to change to -V...do you
think
> > >>>>>>> this could be the issue?
> > >>>>>>>
> > >>>>>>> Thank you for following up,
> > >>>>>>> Brock
> > >>>>>>>
> > >>>>>>> ________________________________________
> > >>>>>>> From: Julie Prestopnik via RT [met_help at ucar.edu]
> > >>>>>>> Sent: Monday, December 01, 2014 9:56 AM
> > >>>>>>> To: Burghardt, Brock
> > >>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling MET
V4.1
> > >>>>>>>
> > >>>>>>> Hi Brock.  I just wanted to follow up - Were able to get a
> > >>>>>>> successful
> > >>>>>> build
> > >>>>>>> of METv5.0?
> > >>>>>>>
> > >>>>>>> Thanks,
> > >>>>>>> Julie
> > >>>>>>>
> > >>>>>>> On Tue, Nov 18, 2014 at 10:01 AM, Julie Prestopnik
> > >>>>>>> <jpresto at ucar.edu>
> > >>>>>>> wrote:
> > >>>>>>>
> > >>>>>>>> Hi Brock.
> > >>>>>>>>
> > >>>>>>>> Thanks you for sending the output.  I was looking to make
> > >>>>>>>> sure that the MET environment variables, like MET_NETCDF
or
> > >>>>>>>> MET_NETCDFINC and MET_NETCDFLIB, were set correctly.
> > >>>>>>>>
> > >>>>>>>> Please take a look at the following page and follow the
> > >>>>>>>> instructions on compiling MET:
> > >>>>>>>>
> > >>>>>>>>
> > >>>>>>>
> > >>>>>>
> > >>>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/
> > >>> comp
> > >>>>>> ilation/configure.php
> > >>>>>>>>
> > >>>>>>>> This will help guide you in setting up the required
> > >>>>>>>> environment
> > >>>>>> variables
> > >>>>>>>> and will hopefully resolve the problem you are
experiencing
> > >>>>>>>> with MET
> > >>>>>> not
> > >>>>>>>> being able to find netcdf.hh.
> > >>>>>>>>
> > >>>>>>>> Please let me know if you have any questions, and please
let
> > >>>>>>>> me know if you get a successful build after setting up
the
> > >>> environment
> > >>>>> variables.
> > >>>>>>>>
> > >>>>>>>> Thanks,
> > >>>>>>>> Julie
> > >>>>>>>>
> > >>>>>>>> On Tue, Nov 18, 2014 at 9:47 AM, Burghardt, Brock via RT
<
> > >>>>>>>> met_help at ucar.edu> wrote:
> > >>>>>>>>
> > >>>>>>>>>
> > >>>>>>>>> <URL:
> > >>>>>>>>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
>
> > >>>>>>>>>
> > >>>>>>>>> Hi Julie, the output for the command is,
> > >>>>>>>>>
> > >>>>>>>>> hrothgar:/met-5.0$ printenv | grep met
> > >>>>>>>>> PWD=/home/brburgha/met-5.0
> > >>>>>>>>>
> > >>>>>>>>> ...unless you wanted this,
> > >>>>>>>>>
> > >>>>>>>>> hrothgar:/met-5.0$ printenv | grep MET*
> > >>>>>>>>> HOSTNAME=hrothgar.hpcc.ttu.edu ANT_HOME=/opt/rocks
> > >>>>>>>>> JAVA_HOME=/usr/java/latest
> > >>>>>>>>> P4_GLOBMEMSIZE=83886080
> > >>>>>>>>> MAGICK_HOME=/lustre/work/apps/ImageMagick
> > >>>>>>>>> HOME=/home/brburgha
> > >>>>>>>>> LOGNAME=brburgha
> > >>>>>>>>> G_BROKEN_FILENAMES=1
> > >>>>>>>>>
> > >>>>>>>>>
> > >>>>>>>>> Brock
> > >>>>>>>>>
> > >>>>>>>>>
> > >>>>>>>>> -----Original Message-----
> > >>>>>>>>> From: Julie Prestopnik via RT [mailto:met_help at ucar.edu]
> > >>>>>>>>> Sent: Tuesday, November 18, 2014 10:38 AM
> > >>>>>>>>> To: Burghardt, Brock
> > >>>>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling
MET
> > >>>>>>>>> V4.1
> > >>>>>>>>>
> > >>>>>>>>> Hi Brock.  Thank you for getting the latest version of
MET.
> > >>>>>>>>> I see
> > >>>>>>> you're
> > >>>>>>>>> still problems with MET finding netcdf.hh.
> > >>>>>>>>>
> > >>>>>>>>> Could you please run "printenv | grep MET" and send me
the
> > >>> output?
> > >>>>>>>>>
> > >>>>>>>>> Thanks,
> > >>>>>>>>> Julie
> > >>>>>>>>>
> > >>>>>>>>> On Tue, Nov 18, 2014 at 9:07 AM, Burghardt, Brock via RT
<
> > >>>>>>>>> met_help at ucar.edu>
> > >>>>>>>>> wrote:
> > >>>>>>>>>
> > >>>>>>>>>>
> > >>>>>>>>>> <URL:
> > >>>>>>>>>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > >>>>
> > >>>>>>>>>>
> > >>>>>>>>>> Hi Julie,
> > >>>>>>>>>>
> > >>>>>>>>>> Thanks for getting back to me so soon and explaining.
> > >>>>>>>>>> Removing the static flag did indeed help in compiling
the
> > >>>>>>>>>> executables I needed
> > >>>>>> but
> > >>>>>>>>>> it did not completely compile (another ld: missing
library
> > >>>>>>>>>> error
> > >>>>>> came
> > >>>>>>>>>> up in the pb2nc build). However, I took your advice and
> > >>> started
> > >>>>>>>>>> over with V5.0 and got that running, but the compile
also
> > >>>>>>>>>> failed before making executables. The error was for not
> > >>>>>>>>>> being able to open
> > >>>>>> netcdf.hh
> > >>>>>>>>>> (make_install.log attached). I added all the met_<lib>
> > >>>>>>>>>> arguments at command line and checked they pointed to
right
> > >>>>>>>>>> directory, and the configure step ran without error. I
> > >>> verified
> > >>>>>>>>>> I do have read
> > >>>>>> permission
> > >>>>>>>>> for netcdf.hh, any ideas on this.
> > >>>>>>>>>>
> > >>>>>>>>>> Thanks,
> > >>>>>>>>>> Brock
> > >>>>>>>>>>
> > >>>>>>>>>> -----Original Message-----
> > >>>>>>>>>> From: Julie Prestopnik via RT
[mailto:met_help at ucar.edu]
> > >>>>>>>>>> Sent: Monday, November 17, 2014 4:34 PM
> > >>>>>>>>>> To: Burghardt, Brock
> > >>>>>>>>>> Subject: Re: [rt.rap.ucar.edu #69783] Issue compiling
MET
> > >>> V4.1
> > >>>>>>>>>>
> > >>>>>>>>>> Hi Brock.
> > >>>>>>>>>>
> > >>>>>>>>>> I see that you are having problems building METv4.1
with an
> > >>>>>>>>>> intel compiler.  Thank you for attaching your
user_defs.mk
> > >>> and
> > >>>>>>>>>> make.log
> > >>>>>>>>> files.
> > >>>>>>>>>>
> > >>>>>>>>>> First, I wanted to mention that METv5.0 was release in
> > >>>>>>>>>> early September, which does include bug fixes for
METv4.1.
> > >>>>>>>>>> You can view
> > >>>>>> the
> > >>>>>>>>>> release notes for METv5.0 here:
> > >>>>>>>>>>
> > >>>>>>>>>>
> > >>>>>>>>>>
> > >>>>>>>
> > >>>
http://www.dtcenter.org/met/users/support/release_notes/METv5.0_re
> > >>> le
> > >>>>>>> as
> > >>>>>>>>>> e_notes.php
> > >>>>>>>>>>
> > >>>>>>>>>> For now, let's try to overcome this error with METv4.1:
> > >>>>>>>>>>
> > >>>>>>>>>> ld: cannot find -lm
> > >>>>>>>>>>
> > >>>>>>>>>> We've had this issue come up with some other users in
the
> > >>> past.
> > >>>>>>>>>> By default, MET does static linking - meaning that it
links
> > >>>>>>>>>> in the library files on which it depends at compile
time.
> > >>>>>>>>>> The other type of linking is call dynamic linking - for
> > >>>>>>>>>> which the library files on
> > >>>>>> which
> > >>>>>>>>>> it depends are linked in at runtime. Other users have
had
> > >>>>>>>>>> success compiling MET with dynamic linking instead.
Open up
> > >>> and
> > >>>>>>>>>> edit the top-level user_defs.mk file, by commenting out
or
> > >>>>>>>>>> removing
> > >>>>>> references
> > >>>>>>>>>> to "-static". They show up on the CXX_FLAGS and
FC_FLAGS
> > >>> lines.
> > >>>>>>>>>>
> > >>>>>>>>>> Then try recompiling MET, being sure to execute a "make
> > >> clean"
> > >>>>>> first.
> > >>>>>>>>>>
> > >>>>>>>>>> Thanks,
> > >>>>>>>>>> Julie
> > >>>>>>>>>>
> > >>>>>>>>>> On Mon, Nov 17, 2014 at 1:55 PM, Burghardt, Brock via
RT <
> > >>>>>>>>>> met_help at ucar.edu>
> > >>>>>>>>>> wrote:
> > >>>>>>>>>>
> > >>>>>>>>>>>
> > >>>>>>>>>>> Mon Nov 17 13:55:16 2014: Request 69783 was acted
upon.
> > >>>>>>>>>>> Transaction: Ticket created by brock.burghardt at ttu.edu
> > >>>>>>>>>>>       Queue: met_help
> > >>>>>>>>>>>     Subject: Issue compiling MET V4.1
> > >>>>>>>>>>>       Owner: Nobody
> > >>>>>>>>>>>  Requestors: brock.burghardt at ttu.edu
> > >>>>>>>>>>>      Status: new
> > >>>>>>>>>>> Ticket <URL:
> > >>>>>>>>>>>
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69783
> > >>>>>>>>>>>>
> > >>>>>>>>>>>
> > >>>>>>>>>>>
> > >>>>>>>>>>> Hello,
> > >>>>>>>>>>>
> > >>>>>>>>>>> I'm trying to compile MET V4.1 using NETCDF V3.6.3
with an
> > >>>>>>>>>>> intel compiler (definitions and log file attachted).
> > >>>>>>>>>>> I've gone through
> > >>>>>> all
> > >>>>>>>>>>> suggestions in the appendix of the users guide but
still
> > >>>>>>>>>>> get an error while trying to make the ensemble_stat
> > >> executable.
> > >>>>>>>>>>> Is there
> > >>>>>> a
> > >>>>>>>>>>> liblm.so or .a required for this build?
> > >>>>>>>>>>>
> > >>>>>>>>>>> Thanks,
> > >>>>>>>>>>> Brock
> > >>>>>>>>>>>
> > >>>>>>>>>>> ld: cannot find -lm
> > >>>>>>>>>>> make[4]: *** [ensemble_stat] Error 1
> > >>>>>>>>>>>
> > >>>>>>>>>>>
> > >>>>>>>>>>>
> > >>>>>>>>>>> Brock Burghardt
> > >>>>>>>>>>> Atmospheric Science PhD Student Dept. of Geosciences
Texas
> > >>>>>>>>>>> Tech University
> > >>>>>>>>>>>
> > >>>>>>>>>>>
> > >>>>>>>>>>>
> > >>>>>>>>>>
> > >>>>>>>>>>
> > >>>>>>>>>> --
> > >>>>>>>>>> Julie Prestopnik
> > >>>>>>>>>> National Center for Atmospheric Research Research
> > >>>>>>>>>> Applications Laboratory
> > >>>>>>>>>> Phone: 303.497.8399
> > >>>>>>>>>> Email: jpresto at ucar.edu
> > >>>>>>>>>>
> > >>>>>>>>>>
> > >>>>>>>>>>
> > >>>>>>>>>
> > >>>>>>>>>
> > >>>>>>>>> --
> > >>>>>>>>> Julie Prestopnik
> > >>>>>>>>> National Center for Atmospheric Research Research
> > >>>>>>>>> Applications
> > >>>>>>> Laboratory
> > >>>>>>>>> Phone: 303.497.8399
> > >>>>>>>>> Email: jpresto at ucar.edu
> > >>>>>>>>>
> > >>>>>>>>>
> > >>>>>>>>>
> > >>>>>>>>>
> > >>>>>>>>
> > >>>>>>>>
> > >>>>>>>> --
> > >>>>>>>> Julie Prestopnik
> > >>>>>>>> National Center for Atmospheric Research Research
> > >>>>>>>> Applications Laboratory
> > >>>>>>>> Phone: 303.497.8399
> > >>>>>>>> Email: jpresto at ucar.edu
> > >>>>>>>>
> > >>>>>>>
> > >>>>>>>
> > >>>>>>>
> > >>>>>>> --
> > >>>>>>> Julie Prestopnik
> > >>>>>>> National Center for Atmospheric Research Research
Applications
> > >>>>>>> Laboratory
> > >>>>>>> Phone: 303.497.8399
> > >>>>>>> Email: jpresto at ucar.edu
> > >>>>>>>
> > >>>>>>>
> > >>>>>>>
> > >>>>>>
> > >>>>>>
> > >>>>>> --
> > >>>>>> Julie Prestopnik
> > >>>>>> National Center for Atmospheric Research Research
Applications
> > >>>>>> Laboratory
> > >>>>>> Phone: 303.497.8399
> > >>>>>> Email: jpresto at ucar.edu
> > >>>>>>
> > >>>>>>
> > >>>>>>
> > >>>>>>
> > >>>>>
> > >>>>>
> > >>>>> --
> > >>>>> Julie Prestopnik
> > >>>>> National Center for Atmospheric Research Research
Applications
> > >>> Laboratory
> > >>>>> Phone: 303.497.8399
> > >>>>> Email: jpresto at ucar.edu
> > >>>>>
> > >>>>>
> > >>>>>
> > >>>>
> > >>>>
> > >>>> --
> > >>>> Julie Prestopnik
> > >>>> National Center for Atmospheric Research Research
Applications
> > >>>> Laboratory
> > >>>> Phone: 303.497.8399
> > >>>> Email: jpresto at ucar.edu
> > >>>>
> > >>>>
> > >>>
> > >>>
> > >>>
> > >>
> > >>
> > >>
> > >
> >
> >
> >
>
> --
> Julie Prestopnik
> National Center for Atmospheric Research Research Applications
Laboratory
> Phone: 303.497.8399
> Email: jpresto at ucar.edu
>
>
>

------------------------------------------------