[Met_help] [rt.rap.ucar.edu #60419] History for help: install MET4.0
John Halley Gotway via RT
met_help at ucar.edu
Tue Mar 26 09:59:28 MDT 2013
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Hi met_helper,
I'm trying to install MET4.0., but failed. Attached please find the output of the stage of make. Could you please inform me how to fix it and make it through the compilation?
Thank you,
Sha
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #60419] help: install MET4.0
From: John Halley Gotway
Time: Tue Feb 26 09:26:17 2013
Sha,
Thank you for sending the make log. I took a look and see that the
linker can't find the standard math library:
ld: cannot find -lm
Since, by default, MET is doing static linking, the linker is looking
for a library archive file name "libm.a", but can't find it. The
linker being unable to find the standard math library is a bad
sign and often means that the compiler is not installed/configured
well. We could locate the correct library archive file and tell the
linker where to find it, but in my experience, if the linker
can't find the math library, you'll often have a whole host of other
compilation issues arise.
Have you used the Intel C++ (icc) compiler to compile other code
successfully?
Here are some things you could consider:
(1) If you can locate "libm.a" on your system, you could tell the
linker where to find it by editing the top-level METv4.0/user_defs.mk
file. Just add "-L/my/directory" to the "CXX_LIBS" line,
where /my/directory is the directory containing the file libm.a. But
as I warned above, I suspect you'll run into other problems after
resolving this one.
(2) You could try rebuilding MET using dynamic linking rather than
static linking. Perhaps that'd work better:
- Just edit the top-level METv4.0/user_defs.mk file and remove "-
static" from the CXX_FLAGS and FC_FLAGS lines.
- Then rebuild MET, being sure to execute a "make clean" first.
(3) If you have other compilers available on your system, GNU or PGI
perhaps, you could try rebuilding MET and all it's dependent libraries
using them.
(4) You could talk to a systems administrator about the problems
you're having, and reinstall the intel compilers.
Sorry I don't have an easy answer for you. Please let me know how it
goes.
Thanks,
John Halley Gotway
met_help at ucar.edu
On 02/25/2013 05:27 PM, Feng, Sha via RT wrote:
>
> Mon Feb 25 17:27:36 2013: Request 60419 was acted upon.
> Transaction: Ticket created by sfeng at jifresse.ucla.edu
> Queue: met_help
> Subject: help: install MET4.0
> Owner: Nobody
> Requestors: sfeng at jifresse.ucla.edu
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>
>
> Hi met_helper,
>
> I'm trying to install MET4.0., but failed. Attached please find the
output of the stage of make. Could you please inform me how to fix it
and make it through the compilation?
>
> Thank you,
> Sha
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #60419] help: install MET4.0
From: Feng, Sha
Time: Tue Feb 26 10:00:46 2013
Hi John,
Yes, there has been no issue in icc by far, and all other lib required
by
MET have been compiled by icc or/and ifort successfully.
I want to go with the option (2) first, but the installation still
failed.
I list the error info below.
+++++++++++++++++++++++
Š.
/opt/intel/bin/icc -o pb2nc pb2nc.cc pb2nc_conf_info.o numpbmsg.o
openpb.o
closepb.o readpb.o dumppb.o \
-Wshadow -g -DBLOCK4 -DMET_BASE=\"/home/sfeng/src/METv4.0\" \
-I/home/sfeng/src/METv4.0/include
-I/home/sfeng/src/netcdf-3.6.3/include
-I/home/sfeng/src/gsl-1.15/include
-I/home/sfeng/src/BUFRLIBv10.1.0 \
-L/home/sfeng/src/METv4.0/lib -L/home/sfeng/src/netcdf-3.6.3/lib
-L/home/sfeng/src/gsl-1.15/lib -L/home/sfeng/src/BUFRLIBv10.1.0 \
-lvx_stat_out \
-lvx_statistics \
-lvx_shapedata \
-lvx_gsl_prob \
-lvx_analysis_util \
-lvx_data2d_factory \
-lvx_data2d_grib \
-lvx_data2d_nc_met \
-lvx_data2d_nc_pinterp \
-lvx_data2d \
-lvx_nc_util \
-lvx_grid \
-lvx_config \
-lvx_pb_util \
-lvx_cal \
-lvx_util \
-lvx_math \
-lvx_color \
-lvx_log \
-lm -lnetcdf_c++ -lnetcdf -lgsl -lgslcblas -lbufr \
-lifcore
/home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(stndrd.o): In function
`stndrd_':
stndrd.f:(.text+0x4a2): undefined reference to `restd_'
/home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(bort2.o): In function
`bort2_':
bort2.f:(.text+0x74): undefined reference to `bort_exit_'
/home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(bort.o): In function `bort_':
bort.f:(.text+0x5d): undefined reference to `bort_exit_'
/home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(reads3.o): In function
`reads3_':
reads3.f:(.text+0x4ae): undefined reference to `stseq_'
/home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(readmt.o): In function
`readmt_':
readmt.f:(.text+0x1186): undefined reference to `icvidx_'
make[4]: *** [pb2nc] Error 1
make[3]: *** [all] Error 2
make[2]: *** [all] Error 2
make[1]: *** [all] Error 2
make: *** [all] Error 2
+++++++++++++++++++++++
Thanks,
Sha
On 2/26/13 8:26 AM, "John Halley Gotway via RT" <met_help at ucar.edu>
wrote:
>Sha,
>
>Thank you for sending the make log. I took a look and see that the
>linker can't find the standard math library:
> ld: cannot find -lm
>
>Since, by default, MET is doing static linking, the linker is looking
for
>a library archive file name "libm.a", but can't find it. The linker
>being unable to find the standard math library is a bad
>sign and often means that the compiler is not installed/configured
well.
>We could locate the correct library archive file and tell the linker
>where to find it, but in my experience, if the linker
>can't find the math library, you'll often have a whole host of other
>compilation issues arise.
>
>Have you used the Intel C++ (icc) compiler to compile other code
>successfully?
>
>Here are some things you could consider:
>
> (1) If you can locate "libm.a" on your system, you could tell the
>linker where to find it by editing the top-level METv4.0/user_defs.mk
>file. Just add "-L/my/directory" to the "CXX_LIBS" line,
>where /my/directory is the directory containing the file libm.a. But
as
>I warned above, I suspect you'll run into other problems after
resolving
>this one.
>
> (2) You could try rebuilding MET using dynamic linking rather than
>static linking. Perhaps that'd work better:
> - Just edit the top-level METv4.0/user_defs.mk file and remove
>"-static" from the CXX_FLAGS and FC_FLAGS lines.
> - Then rebuild MET, being sure to execute a "make clean" first.
>
> (3) If you have other compilers available on your system, GNU or
PGI
>perhaps, you could try rebuilding MET and all it's dependent
libraries
>using them.
>
> (4) You could talk to a systems administrator about the problems
you're
>having, and reinstall the intel compilers.
>
>Sorry I don't have an easy answer for you. Please let me know how it
>goes.
>
>Thanks,
>John Halley Gotway
>met_help at ucar.edu
>
>
>On 02/25/2013 05:27 PM, Feng, Sha via RT wrote:
>>
>> Mon Feb 25 17:27:36 2013: Request 60419 was acted upon.
>> Transaction: Ticket created by sfeng at jifresse.ucla.edu
>> Queue: met_help
>> Subject: help: install MET4.0
>> Owner: Nobody
>> Requestors: sfeng at jifresse.ucla.edu
>> Status: new
>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>>
>>
>> Hi met_helper,
>>
>> I'm trying to install MET4.0., but failed. Attached please find the
>>output of the stage of make. Could you please inform me how to fix
it
>>and make it through the compilation?
>>
>> Thank you,
>> Sha
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #60419] help: install MET4.0
From: John Halley Gotway
Time: Tue Feb 26 10:36:05 2013
Sha,
The FORTRAN part of MET that interfaces with BUFRLIB is sometime the
source of issues. Let's try this...
- Edit the top-level METv4.0/user_defs.mk to disable the compilation
of the PB2NC tool, by setting ENABLE_PB2NC to 0:
ENABLE_PB2NC = 0
Then try rebuilding MET to see if you encounter any other errors.
While you do that, I'll look into the BUFRLIB issue more.
Please let me know what you find.
Thanks,
John
On 02/26/2013 10:00 AM, Feng, Sha via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>
> Hi John,
>
> Yes, there has been no issue in icc by far, and all other lib
required by
> MET have been compiled by icc or/and ifort successfully.
>
> I want to go with the option (2) first, but the installation still
failed.
> I list the error info below.
>
> +++++++++++++++++++++++
>
> Š.
>
> /opt/intel/bin/icc -o pb2nc pb2nc.cc pb2nc_conf_info.o numpbmsg.o
openpb.o
> closepb.o readpb.o dumppb.o \
> -Wshadow -g -DBLOCK4 -DMET_BASE=\"/home/sfeng/src/METv4.0\" \
> -I/home/sfeng/src/METv4.0/include
> -I/home/sfeng/src/netcdf-3.6.3/include
-I/home/sfeng/src/gsl-1.15/include
> -I/home/sfeng/src/BUFRLIBv10.1.0 \
> -L/home/sfeng/src/METv4.0/lib
-L/home/sfeng/src/netcdf-3.6.3/lib
> -L/home/sfeng/src/gsl-1.15/lib -L/home/sfeng/src/BUFRLIBv10.1.0 \
> -lvx_stat_out \
> -lvx_statistics \
> -lvx_shapedata \
> -lvx_gsl_prob \
> -lvx_analysis_util \
> -lvx_data2d_factory \
> -lvx_data2d_grib \
> -lvx_data2d_nc_met \
> -lvx_data2d_nc_pinterp \
> -lvx_data2d \
> -lvx_nc_util \
> -lvx_grid \
> -lvx_config \
> -lvx_pb_util \
> -lvx_cal \
> -lvx_util \
> -lvx_math \
> -lvx_color \
> -lvx_log \
> -lm -lnetcdf_c++ -lnetcdf -lgsl -lgslcblas -lbufr \
> -lifcore
> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(stndrd.o): In function
`stndrd_':
> stndrd.f:(.text+0x4a2): undefined reference to `restd_'
> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(bort2.o): In function
`bort2_':
> bort2.f:(.text+0x74): undefined reference to `bort_exit_'
> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(bort.o): In function
`bort_':
> bort.f:(.text+0x5d): undefined reference to `bort_exit_'
> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(reads3.o): In function
`reads3_':
> reads3.f:(.text+0x4ae): undefined reference to `stseq_'
> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(readmt.o): In function
`readmt_':
> readmt.f:(.text+0x1186): undefined reference to `icvidx_'
> make[4]: *** [pb2nc] Error 1
> make[3]: *** [all] Error 2
> make[2]: *** [all] Error 2
> make[1]: *** [all] Error 2
> make: *** [all] Error 2
> +++++++++++++++++++++++
>
>
> Thanks,
> Sha
>
> On 2/26/13 8:26 AM, "John Halley Gotway via RT" <met_help at ucar.edu>
wrote:
>
>> Sha,
>>
>> Thank you for sending the make log. I took a look and see that the
>> linker can't find the standard math library:
>> ld: cannot find -lm
>>
>> Since, by default, MET is doing static linking, the linker is
looking for
>> a library archive file name "libm.a", but can't find it. The
linker
>> being unable to find the standard math library is a bad
>> sign and often means that the compiler is not installed/configured
well.
>> We could locate the correct library archive file and tell the
linker
>> where to find it, but in my experience, if the linker
>> can't find the math library, you'll often have a whole host of
other
>> compilation issues arise.
>>
>> Have you used the Intel C++ (icc) compiler to compile other code
>> successfully?
>>
>> Here are some things you could consider:
>>
>> (1) If you can locate "libm.a" on your system, you could tell the
>> linker where to find it by editing the top-level
METv4.0/user_defs.mk
>> file. Just add "-L/my/directory" to the "CXX_LIBS" line,
>> where /my/directory is the directory containing the file libm.a.
But as
>> I warned above, I suspect you'll run into other problems after
resolving
>> this one.
>>
>> (2) You could try rebuilding MET using dynamic linking rather
than
>> static linking. Perhaps that'd work better:
>> - Just edit the top-level METv4.0/user_defs.mk file and remove
>> "-static" from the CXX_FLAGS and FC_FLAGS lines.
>> - Then rebuild MET, being sure to execute a "make clean" first.
>>
>> (3) If you have other compilers available on your system, GNU or
PGI
>> perhaps, you could try rebuilding MET and all it's dependent
libraries
>> using them.
>>
>> (4) You could talk to a systems administrator about the problems
you're
>> having, and reinstall the intel compilers.
>>
>> Sorry I don't have an easy answer for you. Please let me know how
it
>> goes.
>>
>> Thanks,
>> John Halley Gotway
>> met_help at ucar.edu
>>
>>
>> On 02/25/2013 05:27 PM, Feng, Sha via RT wrote:
>>>
>>> Mon Feb 25 17:27:36 2013: Request 60419 was acted upon.
>>> Transaction: Ticket created by sfeng at jifresse.ucla.edu
>>> Queue: met_help
>>> Subject: help: install MET4.0
>>> Owner: Nobody
>>> Requestors: sfeng at jifresse.ucla.edu
>>> Status: new
>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>>>
>>>
>>> Hi met_helper,
>>>
>>> I'm trying to install MET4.0., but failed. Attached please find
the
>>> output of the stage of make. Could you please inform me how to fix
it
>>> and make it through the compilation?
>>>
>>> Thank you,
>>> Sha
>>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #60419] help: install MET4.0
From: John Halley Gotway
Time: Tue Feb 26 11:05:27 2013
Sha,
Regarding the BUFRLIB issue, please try rebuilding BUFRLIB using the
following commands:
cd /home/sfeng/src/BUFRLIBv10.1.0
icc -c -DUNDERSCORE *.c
ifort -c -DUNDERSCORE *.f *.F
ar crv libbufr.a *.o
These are taken from the MET online tutorial:
http://www.dtcenter.org/met/users/support/online_tutorial/METv4.0/compilation/req_libs.php
Then in METv4.0/user_defs.mk, turn back on the compilation of the
PB2NC tool:
ENABLE_PB2NC = 1
Then try to rebuild MET. If you encounter another BUFRLIB error,
please send it to me along with the file "libbufr_nm.out", generated
as follows:
nm /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a > libbufr_nm.out
Thanks,
John
On 02/26/2013 10:00 AM, Feng, Sha via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>
> Hi John,
>
> Yes, there has been no issue in icc by far, and all other lib
required by
> MET have been compiled by icc or/and ifort successfully.
>
> I want to go with the option (2) first, but the installation still
failed.
> I list the error info below.
>
> +++++++++++++++++++++++
>
> Š.
>
> /opt/intel/bin/icc -o pb2nc pb2nc.cc pb2nc_conf_info.o numpbmsg.o
openpb.o
> closepb.o readpb.o dumppb.o \
> -Wshadow -g -DBLOCK4 -DMET_BASE=\"/home/sfeng/src/METv4.0\" \
> -I/home/sfeng/src/METv4.0/include
> -I/home/sfeng/src/netcdf-3.6.3/include
-I/home/sfeng/src/gsl-1.15/include
> -I/home/sfeng/src/BUFRLIBv10.1.0 \
> -L/home/sfeng/src/METv4.0/lib
-L/home/sfeng/src/netcdf-3.6.3/lib
> -L/home/sfeng/src/gsl-1.15/lib -L/home/sfeng/src/BUFRLIBv10.1.0 \
> -lvx_stat_out \
> -lvx_statistics \
> -lvx_shapedata \
> -lvx_gsl_prob \
> -lvx_analysis_util \
> -lvx_data2d_factory \
> -lvx_data2d_grib \
> -lvx_data2d_nc_met \
> -lvx_data2d_nc_pinterp \
> -lvx_data2d \
> -lvx_nc_util \
> -lvx_grid \
> -lvx_config \
> -lvx_pb_util \
> -lvx_cal \
> -lvx_util \
> -lvx_math \
> -lvx_color \
> -lvx_log \
> -lm -lnetcdf_c++ -lnetcdf -lgsl -lgslcblas -lbufr \
> -lifcore
> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(stndrd.o): In function
`stndrd_':
> stndrd.f:(.text+0x4a2): undefined reference to `restd_'
> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(bort2.o): In function
`bort2_':
> bort2.f:(.text+0x74): undefined reference to `bort_exit_'
> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(bort.o): In function
`bort_':
> bort.f:(.text+0x5d): undefined reference to `bort_exit_'
> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(reads3.o): In function
`reads3_':
> reads3.f:(.text+0x4ae): undefined reference to `stseq_'
> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(readmt.o): In function
`readmt_':
> readmt.f:(.text+0x1186): undefined reference to `icvidx_'
> make[4]: *** [pb2nc] Error 1
> make[3]: *** [all] Error 2
> make[2]: *** [all] Error 2
> make[1]: *** [all] Error 2
> make: *** [all] Error 2
> +++++++++++++++++++++++
>
>
> Thanks,
> Sha
>
> On 2/26/13 8:26 AM, "John Halley Gotway via RT" <met_help at ucar.edu>
wrote:
>
>> Sha,
>>
>> Thank you for sending the make log. I took a look and see that the
>> linker can't find the standard math library:
>> ld: cannot find -lm
>>
>> Since, by default, MET is doing static linking, the linker is
looking for
>> a library archive file name "libm.a", but can't find it. The
linker
>> being unable to find the standard math library is a bad
>> sign and often means that the compiler is not installed/configured
well.
>> We could locate the correct library archive file and tell the
linker
>> where to find it, but in my experience, if the linker
>> can't find the math library, you'll often have a whole host of
other
>> compilation issues arise.
>>
>> Have you used the Intel C++ (icc) compiler to compile other code
>> successfully?
>>
>> Here are some things you could consider:
>>
>> (1) If you can locate "libm.a" on your system, you could tell the
>> linker where to find it by editing the top-level
METv4.0/user_defs.mk
>> file. Just add "-L/my/directory" to the "CXX_LIBS" line,
>> where /my/directory is the directory containing the file libm.a.
But as
>> I warned above, I suspect you'll run into other problems after
resolving
>> this one.
>>
>> (2) You could try rebuilding MET using dynamic linking rather
than
>> static linking. Perhaps that'd work better:
>> - Just edit the top-level METv4.0/user_defs.mk file and remove
>> "-static" from the CXX_FLAGS and FC_FLAGS lines.
>> - Then rebuild MET, being sure to execute a "make clean" first.
>>
>> (3) If you have other compilers available on your system, GNU or
PGI
>> perhaps, you could try rebuilding MET and all it's dependent
libraries
>> using them.
>>
>> (4) You could talk to a systems administrator about the problems
you're
>> having, and reinstall the intel compilers.
>>
>> Sorry I don't have an easy answer for you. Please let me know how
it
>> goes.
>>
>> Thanks,
>> John Halley Gotway
>> met_help at ucar.edu
>>
>>
>> On 02/25/2013 05:27 PM, Feng, Sha via RT wrote:
>>>
>>> Mon Feb 25 17:27:36 2013: Request 60419 was acted upon.
>>> Transaction: Ticket created by sfeng at jifresse.ucla.edu
>>> Queue: met_help
>>> Subject: help: install MET4.0
>>> Owner: Nobody
>>> Requestors: sfeng at jifresse.ucla.edu
>>> Status: new
>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>>>
>>>
>>> Hi met_helper,
>>>
>>> I'm trying to install MET4.0., but failed. Attached please find
the
>>> output of the stage of make. Could you please inform me how to fix
it
>>> and make it through the compilation?
>>>
>>> Thank you,
>>> Sha
>>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #60419] help: install MET4.0
From: Feng, Sha
Time: Tue Feb 26 11:07:07 2013
Hi John,
Fifteen executables below has been created by following your
instruction.
+++++
[sfeng at dragon2 bin]$ ll
total 196M
-rwxrwxr-x. 1 sfeng sfeng 5.0M Feb 26 09:42 ascii2nc
-rwxrwxr-x. 1 sfeng sfeng 18M Feb 26 09:42 ensemble_stat
-rwxrwxr-x. 1 sfeng sfeng 12M Feb 26 09:42 gen_poly_mask
-rwxrwxr-x. 1 sfeng sfeng 17M Feb 26 09:42 grid_stat
-rwxrwxr-x. 1 sfeng sfeng 5.1M Feb 26 09:42 madis2nc
-rwxrwxr-x. 1 sfeng sfeng 21M Feb 26 09:42 mode
-rwxrwxr-x. 1 sfeng sfeng 5.4M Feb 26 09:42 mode_analysis
-rwxrwxr-x. 1 sfeng sfeng 13M Feb 26 09:42 pcp_combine
-rwxrwxr-x. 1 sfeng sfeng 17M Feb 26 09:42 plot_data_plane
-rwxrwxr-x. 1 sfeng sfeng 14M Feb 26 09:42 plot_point_obs
-rwxrwxr-x. 1 sfeng sfeng 18M Feb 26 09:42 point_stat
-rwxrwxr-x. 1 sfeng sfeng 14M Feb 26 09:42 stat_analysis
-rwxrwxr-x. 1 sfeng sfeng 21M Feb 26 09:42 wavelet_stat
-rwxrwxr-x. 1 sfeng sfeng 12M Feb 26 09:42 wwmca_plot
-rwxrwxr-x. 1 sfeng sfeng 8.8M Feb 26 09:42 wwmca_regrid
++++++++
Thank you,
Sha
On 2/26/13 9:36 AM, "John Halley Gotway via RT" <met_help at ucar.edu>
wrote:
>Sha,
>
>The FORTRAN part of MET that interfaces with BUFRLIB is sometime the
>source of issues. Let's try this...
> - Edit the top-level METv4.0/user_defs.mk to disable the
compilation of
>the PB2NC tool, by setting ENABLE_PB2NC to 0:
> ENABLE_PB2NC = 0
>
>Then try rebuilding MET to see if you encounter any other errors.
While
>you do that, I'll look into the BUFRLIB issue more.
>
>Please let me know what you find.
>
>Thanks,
>John
>
>
>On 02/26/2013 10:00 AM, Feng, Sha via RT wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>>
>> Hi John,
>>
>> Yes, there has been no issue in icc by far, and all other lib
required
>>by
>> MET have been compiled by icc or/and ifort successfully.
>>
>> I want to go with the option (2) first, but the installation still
>>failed.
>> I list the error info below.
>>
>> +++++++++++++++++++++++
>>
>> Š.
>>
>> /opt/intel/bin/icc -o pb2nc pb2nc.cc pb2nc_conf_info.o numpbmsg.o
>>openpb.o
>> closepb.o readpb.o dumppb.o \
>> -Wshadow -g -DBLOCK4 -DMET_BASE=\"/home/sfeng/src/METv4.0\" \
>> -I/home/sfeng/src/METv4.0/include
>> -I/home/sfeng/src/netcdf-3.6.3/include
>>-I/home/sfeng/src/gsl-1.15/include
>> -I/home/sfeng/src/BUFRLIBv10.1.0 \
>> -L/home/sfeng/src/METv4.0/lib
-L/home/sfeng/src/netcdf-3.6.3/lib
>> -L/home/sfeng/src/gsl-1.15/lib -L/home/sfeng/src/BUFRLIBv10.1.0 \
>> -lvx_stat_out \
>> -lvx_statistics \
>> -lvx_shapedata \
>> -lvx_gsl_prob \
>> -lvx_analysis_util \
>> -lvx_data2d_factory \
>> -lvx_data2d_grib \
>> -lvx_data2d_nc_met \
>> -lvx_data2d_nc_pinterp \
>> -lvx_data2d \
>> -lvx_nc_util \
>> -lvx_grid \
>> -lvx_config \
>> -lvx_pb_util \
>> -lvx_cal \
>> -lvx_util \
>> -lvx_math \
>> -lvx_color \
>> -lvx_log \
>> -lm -lnetcdf_c++ -lnetcdf -lgsl -lgslcblas -lbufr \
>> -lifcore
>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(stndrd.o): In function
>>`stndrd_':
>> stndrd.f:(.text+0x4a2): undefined reference to `restd_'
>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(bort2.o): In function
`bort2_':
>> bort2.f:(.text+0x74): undefined reference to `bort_exit_'
>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(bort.o): In function
`bort_':
>> bort.f:(.text+0x5d): undefined reference to `bort_exit_'
>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(reads3.o): In function
>>`reads3_':
>> reads3.f:(.text+0x4ae): undefined reference to `stseq_'
>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(readmt.o): In function
>>`readmt_':
>> readmt.f:(.text+0x1186): undefined reference to `icvidx_'
>> make[4]: *** [pb2nc] Error 1
>> make[3]: *** [all] Error 2
>> make[2]: *** [all] Error 2
>> make[1]: *** [all] Error 2
>> make: *** [all] Error 2
>> +++++++++++++++++++++++
>>
>>
>> Thanks,
>> Sha
>>
>> On 2/26/13 8:26 AM, "John Halley Gotway via RT" <met_help at ucar.edu>
>>wrote:
>>
>>> Sha,
>>>
>>> Thank you for sending the make log. I took a look and see that
the
>>> linker can't find the standard math library:
>>> ld: cannot find -lm
>>>
>>> Since, by default, MET is doing static linking, the linker is
looking
>>>for
>>> a library archive file name "libm.a", but can't find it. The
linker
>>> being unable to find the standard math library is a bad
>>> sign and often means that the compiler is not installed/configured
>>>well.
>>> We could locate the correct library archive file and tell the
linker
>>> where to find it, but in my experience, if the linker
>>> can't find the math library, you'll often have a whole host of
other
>>> compilation issues arise.
>>>
>>> Have you used the Intel C++ (icc) compiler to compile other code
>>> successfully?
>>>
>>> Here are some things you could consider:
>>>
>>> (1) If you can locate "libm.a" on your system, you could tell
the
>>> linker where to find it by editing the top-level
METv4.0/user_defs.mk
>>> file. Just add "-L/my/directory" to the "CXX_LIBS" line,
>>> where /my/directory is the directory containing the file libm.a.
But
>>>as
>>> I warned above, I suspect you'll run into other problems after
>>>resolving
>>> this one.
>>>
>>> (2) You could try rebuilding MET using dynamic linking rather
than
>>> static linking. Perhaps that'd work better:
>>> - Just edit the top-level METv4.0/user_defs.mk file and remove
>>> "-static" from the CXX_FLAGS and FC_FLAGS lines.
>>> - Then rebuild MET, being sure to execute a "make clean"
first.
>>>
>>> (3) If you have other compilers available on your system, GNU or
PGI
>>> perhaps, you could try rebuilding MET and all it's dependent
libraries
>>> using them.
>>>
>>> (4) You could talk to a systems administrator about the problems
>>>you're
>>> having, and reinstall the intel compilers.
>>>
>>> Sorry I don't have an easy answer for you. Please let me know how
it
>>> goes.
>>>
>>> Thanks,
>>> John Halley Gotway
>>> met_help at ucar.edu
>>>
>>>
>>> On 02/25/2013 05:27 PM, Feng, Sha via RT wrote:
>>>>
>>>> Mon Feb 25 17:27:36 2013: Request 60419 was acted upon.
>>>> Transaction: Ticket created by sfeng at jifresse.ucla.edu
>>>> Queue: met_help
>>>> Subject: help: install MET4.0
>>>> Owner: Nobody
>>>> Requestors: sfeng at jifresse.ucla.edu
>>>> Status: new
>>>> Ticket <URL:
>>>>https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>>>>
>>>>
>>>> Hi met_helper,
>>>>
>>>> I'm trying to install MET4.0., but failed. Attached please find
the
>>>> output of the stage of make. Could you please inform me how to
fix it
>>>> and make it through the compilation?
>>>>
>>>> Thank you,
>>>> Sha
>>>>
>>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #60419] help: install MET4.0
From: Feng, Sha
Time: Tue Feb 26 11:26:31 2013
Hi John,
All the executables have been created successfully without any issue.
Thank you so much,
Sha
On 2/26/13 10:05 AM, "John Halley Gotway via RT" <met_help at ucar.edu>
wrote:
>Sha,
>
>Regarding the BUFRLIB issue, please try rebuilding BUFRLIB using the
>following commands:
>
> cd /home/sfeng/src/BUFRLIBv10.1.0
> icc -c -DUNDERSCORE *.c
> ifort -c -DUNDERSCORE *.f *.F
> ar crv libbufr.a *.o
>
>These are taken from the MET online tutorial:
>
>http://www.dtcenter.org/met/users/support/online_tutorial/METv4.0/compilat
>ion/req_libs.php
>
>Then in METv4.0/user_defs.mk, turn back on the compilation of the
PB2NC
>tool:
> ENABLE_PB2NC = 1
>
>Then try to rebuild MET. If you encounter another BUFRLIB error,
please
>send it to me along with the file "libbufr_nm.out", generated as
follows:
> nm /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a > libbufr_nm.out
>
>Thanks,
>John
>
>On 02/26/2013 10:00 AM, Feng, Sha via RT wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>>
>> Hi John,
>>
>> Yes, there has been no issue in icc by far, and all other lib
required
>>by
>> MET have been compiled by icc or/and ifort successfully.
>>
>> I want to go with the option (2) first, but the installation still
>>failed.
>> I list the error info below.
>>
>> +++++++++++++++++++++++
>>
>> Š.
>>
>> /opt/intel/bin/icc -o pb2nc pb2nc.cc pb2nc_conf_info.o numpbmsg.o
>>openpb.o
>> closepb.o readpb.o dumppb.o \
>> -Wshadow -g -DBLOCK4 -DMET_BASE=\"/home/sfeng/src/METv4.0\" \
>> -I/home/sfeng/src/METv4.0/include
>> -I/home/sfeng/src/netcdf-3.6.3/include
>>-I/home/sfeng/src/gsl-1.15/include
>> -I/home/sfeng/src/BUFRLIBv10.1.0 \
>> -L/home/sfeng/src/METv4.0/lib
-L/home/sfeng/src/netcdf-3.6.3/lib
>> -L/home/sfeng/src/gsl-1.15/lib -L/home/sfeng/src/BUFRLIBv10.1.0 \
>> -lvx_stat_out \
>> -lvx_statistics \
>> -lvx_shapedata \
>> -lvx_gsl_prob \
>> -lvx_analysis_util \
>> -lvx_data2d_factory \
>> -lvx_data2d_grib \
>> -lvx_data2d_nc_met \
>> -lvx_data2d_nc_pinterp \
>> -lvx_data2d \
>> -lvx_nc_util \
>> -lvx_grid \
>> -lvx_config \
>> -lvx_pb_util \
>> -lvx_cal \
>> -lvx_util \
>> -lvx_math \
>> -lvx_color \
>> -lvx_log \
>> -lm -lnetcdf_c++ -lnetcdf -lgsl -lgslcblas -lbufr \
>> -lifcore
>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(stndrd.o): In function
>>`stndrd_':
>> stndrd.f:(.text+0x4a2): undefined reference to `restd_'
>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(bort2.o): In function
`bort2_':
>> bort2.f:(.text+0x74): undefined reference to `bort_exit_'
>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(bort.o): In function
`bort_':
>> bort.f:(.text+0x5d): undefined reference to `bort_exit_'
>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(reads3.o): In function
>>`reads3_':
>> reads3.f:(.text+0x4ae): undefined reference to `stseq_'
>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(readmt.o): In function
>>`readmt_':
>> readmt.f:(.text+0x1186): undefined reference to `icvidx_'
>> make[4]: *** [pb2nc] Error 1
>> make[3]: *** [all] Error 2
>> make[2]: *** [all] Error 2
>> make[1]: *** [all] Error 2
>> make: *** [all] Error 2
>> +++++++++++++++++++++++
>>
>>
>> Thanks,
>> Sha
>>
>> On 2/26/13 8:26 AM, "John Halley Gotway via RT" <met_help at ucar.edu>
>>wrote:
>>
>>> Sha,
>>>
>>> Thank you for sending the make log. I took a look and see that
the
>>> linker can't find the standard math library:
>>> ld: cannot find -lm
>>>
>>> Since, by default, MET is doing static linking, the linker is
looking
>>>for
>>> a library archive file name "libm.a", but can't find it. The
linker
>>> being unable to find the standard math library is a bad
>>> sign and often means that the compiler is not installed/configured
>>>well.
>>> We could locate the correct library archive file and tell the
linker
>>> where to find it, but in my experience, if the linker
>>> can't find the math library, you'll often have a whole host of
other
>>> compilation issues arise.
>>>
>>> Have you used the Intel C++ (icc) compiler to compile other code
>>> successfully?
>>>
>>> Here are some things you could consider:
>>>
>>> (1) If you can locate "libm.a" on your system, you could tell
the
>>> linker where to find it by editing the top-level
METv4.0/user_defs.mk
>>> file. Just add "-L/my/directory" to the "CXX_LIBS" line,
>>> where /my/directory is the directory containing the file libm.a.
But
>>>as
>>> I warned above, I suspect you'll run into other problems after
>>>resolving
>>> this one.
>>>
>>> (2) You could try rebuilding MET using dynamic linking rather
than
>>> static linking. Perhaps that'd work better:
>>> - Just edit the top-level METv4.0/user_defs.mk file and remove
>>> "-static" from the CXX_FLAGS and FC_FLAGS lines.
>>> - Then rebuild MET, being sure to execute a "make clean"
first.
>>>
>>> (3) If you have other compilers available on your system, GNU or
PGI
>>> perhaps, you could try rebuilding MET and all it's dependent
libraries
>>> using them.
>>>
>>> (4) You could talk to a systems administrator about the problems
>>>you're
>>> having, and reinstall the intel compilers.
>>>
>>> Sorry I don't have an easy answer for you. Please let me know how
it
>>> goes.
>>>
>>> Thanks,
>>> John Halley Gotway
>>> met_help at ucar.edu
>>>
>>>
>>> On 02/25/2013 05:27 PM, Feng, Sha via RT wrote:
>>>>
>>>> Mon Feb 25 17:27:36 2013: Request 60419 was acted upon.
>>>> Transaction: Ticket created by sfeng at jifresse.ucla.edu
>>>> Queue: met_help
>>>> Subject: help: install MET4.0
>>>> Owner: Nobody
>>>> Requestors: sfeng at jifresse.ucla.edu
>>>> Status: new
>>>> Ticket <URL:
>>>>https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>>>>
>>>>
>>>> Hi met_helper,
>>>>
>>>> I'm trying to install MET4.0., but failed. Attached please find
the
>>>> output of the stage of make. Could you please inform me how to
fix it
>>>> and make it through the compilation?
>>>>
>>>> Thank you,
>>>> Sha
>>>>
>>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #60419] help: install MET4.0
From: John Halley Gotway
Time: Tue Feb 26 14:21:10 2013
Sha,
Great, glad that did the trick. Feel free to write us if any more
issues/questions arise in your use of MET.
Also, if you have not already done so, be sure to download and compile
the
latest set of patches for METv4.0. Just follow the instructions at
the
top of this page:
http://www.dtcenter.org/met/users/support/known_issues/METv4.0/index.php
Thanks,
John
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>
> Hi John,
>
> All the executables have been created successfully without any
issue.
>
> Thank you so much,
> Sha
>
> On 2/26/13 10:05 AM, "John Halley Gotway via RT" <met_help at ucar.edu>
> wrote:
>
>>Sha,
>>
>>Regarding the BUFRLIB issue, please try rebuilding BUFRLIB using the
>>following commands:
>>
>> cd /home/sfeng/src/BUFRLIBv10.1.0
>> icc -c -DUNDERSCORE *.c
>> ifort -c -DUNDERSCORE *.f *.F
>> ar crv libbufr.a *.o
>>
>>These are taken from the MET online tutorial:
>>
>>http://www.dtcenter.org/met/users/support/online_tutorial/METv4.0/compilat
>>ion/req_libs.php
>>
>>Then in METv4.0/user_defs.mk, turn back on the compilation of the
PB2NC
>>tool:
>> ENABLE_PB2NC = 1
>>
>>Then try to rebuild MET. If you encounter another BUFRLIB error,
please
>>send it to me along with the file "libbufr_nm.out", generated as
follows:
>> nm /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a > libbufr_nm.out
>>
>>Thanks,
>>John
>>
>>On 02/26/2013 10:00 AM, Feng, Sha via RT wrote:
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>>>
>>> Hi John,
>>>
>>> Yes, there has been no issue in icc by far, and all other lib
required
>>>by
>>> MET have been compiled by icc or/and ifort successfully.
>>>
>>> I want to go with the option (2) first, but the installation still
>>>failed.
>>> I list the error info below.
>>>
>>> +++++++++++++++++++++++
>>>
>>> Å .
>>>
>>> /opt/intel/bin/icc -o pb2nc pb2nc.cc pb2nc_conf_info.o numpbmsg.o
>>>openpb.o
>>> closepb.o readpb.o dumppb.o \
>>> -Wshadow -g -DBLOCK4 -DMET_BASE=\"/home/sfeng/src/METv4.0\" \
>>> -I/home/sfeng/src/METv4.0/include
>>> -I/home/sfeng/src/netcdf-3.6.3/include
>>>-I/home/sfeng/src/gsl-1.15/include
>>> -I/home/sfeng/src/BUFRLIBv10.1.0 \
>>> -L/home/sfeng/src/METv4.0/lib
-L/home/sfeng/src/netcdf-3.6.3/lib
>>> -L/home/sfeng/src/gsl-1.15/lib -L/home/sfeng/src/BUFRLIBv10.1.0
\
>>> -lvx_stat_out \
>>> -lvx_statistics \
>>> -lvx_shapedata \
>>> -lvx_gsl_prob \
>>> -lvx_analysis_util \
>>> -lvx_data2d_factory \
>>> -lvx_data2d_grib \
>>> -lvx_data2d_nc_met \
>>> -lvx_data2d_nc_pinterp \
>>> -lvx_data2d \
>>> -lvx_nc_util \
>>> -lvx_grid \
>>> -lvx_config \
>>> -lvx_pb_util \
>>> -lvx_cal \
>>> -lvx_util \
>>> -lvx_math \
>>> -lvx_color \
>>> -lvx_log \
>>> -lm -lnetcdf_c++ -lnetcdf -lgsl -lgslcblas -lbufr \
>>> -lifcore
>>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(stndrd.o): In function
>>>`stndrd_':
>>> stndrd.f:(.text+0x4a2): undefined reference to `restd_'
>>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(bort2.o): In function
>>> `bort2_':
>>> bort2.f:(.text+0x74): undefined reference to `bort_exit_'
>>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(bort.o): In function
`bort_':
>>> bort.f:(.text+0x5d): undefined reference to `bort_exit_'
>>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(reads3.o): In function
>>>`reads3_':
>>> reads3.f:(.text+0x4ae): undefined reference to `stseq_'
>>> /home/sfeng/src/BUFRLIBv10.1.0/libbufr.a(readmt.o): In function
>>>`readmt_':
>>> readmt.f:(.text+0x1186): undefined reference to `icvidx_'
>>> make[4]: *** [pb2nc] Error 1
>>> make[3]: *** [all] Error 2
>>> make[2]: *** [all] Error 2
>>> make[1]: *** [all] Error 2
>>> make: *** [all] Error 2
>>> +++++++++++++++++++++++
>>>
>>>
>>> Thanks,
>>> Sha
>>>
>>> On 2/26/13 8:26 AM, "John Halley Gotway via RT"
<met_help at ucar.edu>
>>>wrote:
>>>
>>>> Sha,
>>>>
>>>> Thank you for sending the make log. I took a look and see that
the
>>>> linker can't find the standard math library:
>>>> ld: cannot find -lm
>>>>
>>>> Since, by default, MET is doing static linking, the linker is
looking
>>>>for
>>>> a library archive file name "libm.a", but can't find it. The
linker
>>>> being unable to find the standard math library is a bad
>>>> sign and often means that the compiler is not
installed/configured
>>>>well.
>>>> We could locate the correct library archive file and tell the
linker
>>>> where to find it, but in my experience, if the linker
>>>> can't find the math library, you'll often have a whole host of
other
>>>> compilation issues arise.
>>>>
>>>> Have you used the Intel C++ (icc) compiler to compile other code
>>>> successfully?
>>>>
>>>> Here are some things you could consider:
>>>>
>>>> (1) If you can locate "libm.a" on your system, you could tell
the
>>>> linker where to find it by editing the top-level
METv4.0/user_defs.mk
>>>> file. Just add "-L/my/directory" to the "CXX_LIBS" line,
>>>> where /my/directory is the directory containing the file libm.a.
But
>>>>as
>>>> I warned above, I suspect you'll run into other problems after
>>>>resolving
>>>> this one.
>>>>
>>>> (2) You could try rebuilding MET using dynamic linking rather
than
>>>> static linking. Perhaps that'd work better:
>>>> - Just edit the top-level METv4.0/user_defs.mk file and
remove
>>>> "-static" from the CXX_FLAGS and FC_FLAGS lines.
>>>> - Then rebuild MET, being sure to execute a "make clean"
first.
>>>>
>>>> (3) If you have other compilers available on your system, GNU
or PGI
>>>> perhaps, you could try rebuilding MET and all it's dependent
libraries
>>>> using them.
>>>>
>>>> (4) You could talk to a systems administrator about the
problems
>>>>you're
>>>> having, and reinstall the intel compilers.
>>>>
>>>> Sorry I don't have an easy answer for you. Please let me know
how it
>>>> goes.
>>>>
>>>> Thanks,
>>>> John Halley Gotway
>>>> met_help at ucar.edu
>>>>
>>>>
>>>> On 02/25/2013 05:27 PM, Feng, Sha via RT wrote:
>>>>>
>>>>> Mon Feb 25 17:27:36 2013: Request 60419 was acted upon.
>>>>> Transaction: Ticket created by sfeng at jifresse.ucla.edu
>>>>> Queue: met_help
>>>>> Subject: help: install MET4.0
>>>>> Owner: Nobody
>>>>> Requestors: sfeng at jifresse.ucla.edu
>>>>> Status: new
>>>>> Ticket <URL:
>>>>>https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=60419 >
>>>>>
>>>>>
>>>>> Hi met_helper,
>>>>>
>>>>> I'm trying to install MET4.0., but failed. Attached please find
the
>>>>> output of the stage of make. Could you please inform me how to
fix it
>>>>> and make it through the compilation?
>>>>>
>>>>> Thank you,
>>>>> Sha
>>>>>
>>>>
>>>
>>
>
>
------------------------------------------------
More information about the Met_help
mailing list