[Dune] using metis & parmetis creates a build problem
Benedikt Oswald
benedikt.oswald at lspr.ch
Mon Apr 28 12:14:36 CEST 2014
Hello Markus, thanks,I just run a test, updating to the latest DUNE trunk.
Platform is:
dysseus:hades-pure-dune oswald$ uname -a
Darwin vpn-a-183-144.epfl.ch 13.1.0 Darwin Kernel Version 13.1.0: Wed Apr 2 23:52:02 PDT 2014; root:xnu-2422.92.1~2/RELEASE_X86_64 x86_64
odysseus:hades-pure-dune oswald$ mpicxx -v
Using built-in specs.
COLLECT_GCC=/Users/oswald/extlib/gcc/4.8.1/gcc/4.8.1/bin/g++
COLLECT_LTO_WRAPPER=/Users/oswald/extlib/gcc/4.8.1/gcc/4.8.1/libexec/gcc/x86_64-apple-darwin12.4.0/4.8.1/lto-wrapper
Target: x86_64-apple-darwin12.4.0
Configured with: ../gcc-4.8.1/configure --prefix=/Users/oswald/extlib/gcc/4.8.1/gcc/4.8.1 --with-gmp=/Users/oswald/extlib/gmp/5.1.2/gcc/4.8.1 --with-mpfr=/Users/oswald/extlib/mpfr/3.1.2/gcc/4.8.1 --with-mpc=/Users/oswald/extlib/mpc/1.0.1/gcc/4.8.1 --with-isl=/Users/oswald/extlib/isl/0.11.1/gcc/4.8.1 --with-cloog=/Users/oswald/extlib/cloog/0.18.0/gcc/4.8.1
Thread model: posix
gcc version 4.8.1 (GCC)
unfortunately, the error is still here:
In file included from /Users/oswald/gitwork/hades-pure-dune/dune-istl/dune/istl/matrixredistribute.hh:5:0,
from /Users/oswald/gitwork/hades-pure-dune/dune-istl/dune/istl/paamg/transfer.hh:7,
from /Users/oswald/gitwork/hades-pure-dune/dune-istl/dune/istl/paamg/amg.hh:10,
from amgtest.cc:41:
/Users/oswald/gitwork/hades-pure-dune/dune-istl/dune/istl/repartition.hh:14:19: fatal error: metis.h: No such file or directory
#include <metis.h>
^
compilation terminated.
the configure file for build with autotools is:
CONFIGURE_FLAGS="DUNE_LDFLAGS='-L$HOME/extlib/superludist/3.3/openmpi/1.6.5/gcc/4.8.1/lib -lsuperludist -L$HOME/extlib/parmetis/4.0.3/openmpi/1.6.5/gcc/4.8.1/lib -lparmetis -L$HOME/extlib/metis/5.1.0/gcc/4.8.1/lib -lmetis -framework vecLib '
CXX='mpicxx' CC='mpicc' LD='mpicxx' F77='mpif77'
CXXFLAGS='-DMPICH_IGNORE_CXX_SEEK -O2 -g0 --no-strict-aliasing '
CPPFLAGS='-DMPICH_IGNORE_CXX_SEEK -O2 -g0 -DHAVE_SUPERLU_DIST -I$HOME/extlib/superludist/3.3/openmpi/1.6.5/gcc/4.8.1/SRC '
CFLAGS='-DMPICH_IGNORE_CXX_SEEK -O2 -g0 -fomit-frame-pointer --no-strict-aliasing '
--enable-hades3deltd
--enable-hades3dfetd
--enable-hades3ddges
--enable-hades3ddgfd
--enable-hades3dfeqs
--enable-hades3dfefd
--disable-documentation
--enable-petsc
--enable-parallel
--enable-debug-output
--enable-fieldvector-size-is-method
--enable-experimental-grid-extensions
--without-gmp
--with-boost='$HOME/extlib/boost/1.55.0/gcc/4.8.1'
--with-metis='$HOME/extlib/metis/5.1.0/gcc/4.8.1'
--with-parmetis='$HOME/extlib/parmetis/4.0.3/openmpi/1.6.5/gcc/4.8.1'
--with-alugrid='$HOME/extlib/alugrid/1.96/openmpi/1.6.5/gcc/4.8.1'
--with-petsc='$HOME/extlib/petsc/3.4.2/openmpi/1.6.5/gcc/4.8.1'
--prefix='$HOME/hades'
"
I think that the error has to do with:
configure:36340: $? = 0
configure:36349: result: yes
configure:36628: PARMETIS: checking for /Users/oswald/extlib/parmetis/4.0.3/openmpi/1.6.5/gcc/4.8.1/include
configure:36657: checking for METIS library
configure:36673: result: yes
configure:36744: checking for METIS_PartGraphKway in -lmetis
configure:36785: mpicc -o conftest -I/Users/oswald/extlib/metis/5.1.0/gcc/4.8.1//include -DMPICH_IGNORE_CXX_SEEK -O2 -g0 -DHAVE_SUPERLU_DIST -I/Users/oswald/extlib/superludist/3.3/openmpi/1.6.5/gcc/4.8.1/SRC -L/Users/oswald/extlib/metis/5.1.0/gcc/4.8.1//lib conftest.c -lmetis -lm -lm >&5
configure:36785: $? = 0
configure:36794: result: yes
configure:36804: checking for metis.h
configure:36804: mpicc -c -I/Users/oswald/extlib/metis/5.1.0/gcc/4.8.1//include -DMPICH_IGNORE_CXX_SEEK -O2 -g0 -DHAVE_SUPERLU_DIST -I/Users/oswald/extlib/superludist/3.3/openmpi/1.6.5/gcc/4.8.1/SRC conftest.c >&5
configure:36804: $? = 0
configure:36804: result: yes
configure:36825: checking METIS in /Users/oswald/extlib/metis/5.1.0/gcc/4.8.1
configure:36834: result: ok
configure:36904: checking parmetis.h usability
configure:36904: mpicc -c -DMPICH_IGNORE_CXX_SEEK -O2 -g0 -fomit-frame-pointer --no-strict-aliasing -DMPICH_IGNORE_CXX_SEEK -O2 -g0 -DHAVE_SUPERLU_DIST -I/Users/oswald/extlib/superludist/3.3/openmpi/1.6.5/gcc/4.8.1/SRC -I/Users/oswald/extlib/parmetis/4.0.3/openmpi/1.6.5/gcc/4.8.1/include -I/Users/oswald/extlib/metis/5.1.0/gcc/4.8.1/include -I/Users/oswald/extlib/openmpi/1.6.5/gcc/4.8.1/include -DMPIPP_H -DENABLE_MPI=1 conftest.c >&5
configure:36904: $? = 0
configure:36904: result: yes
configure:36904: checking parmetis.h presence
configure:36904: mpicc -E -DMPICH_IGNORE_CXX_SEEK -O2 -g0 -DHAVE_SUPERLU_DIST -I/Users/oswald/extlib/superludist/3.3/openmpi/1.6.5/gcc/4.8.1/SRC -I/Users/oswald/extlib/parmetis/4.0.3/openmpi/1.6.5/gcc/4.8.1/include -I/Users/oswald/extlib/metis/5.1.0/gcc/4.8.1/include -I/Users/oswald/extlib/openmpi/1.6.5/gcc/4.8.1/include -DMPIPP_H -DENABLE_MPI=1 conftest.c
configure:36904: $? = 0
configure:36904: result: yes
configure:36904: checking for parmetis.h
configure:36904: result: yes
configure:36929: checking for metis_partgraphkway in -L/Users/oswald/extlib/parmetis/4.0.3/openmpi/1.6.5/gcc/4.8.1/lib -lmetis
configure:36970: mpicc -o conftest -DMPICH_IGNORE_CXX_SEEK -O2 -g0 -fomit-frame-pointer --no-strict-aliasing -DMPICH_IGNORE_CXX_SEEK -O2 -g0 -DHAVE_SUPERLU_DIST -I/Users/oswald/extlib/superludist/3.3/openmpi/1.6.5/gcc/4.8.1/SRC -I/Users/oswald/extlib/parmetis/4.0.3/openmpi/1.6.5/gcc/4.8.1/include -I/Users/oswald/extlib/metis/5.1.0/gcc/4.8.1/include -I/Users/oswald/extlib/openmpi/1.6.5/gcc/4.8.1/include -DMPIPP_H -DENABLE_MPI=1 conftest.c -L/Users/oswald/extlib/parmetis/4.0.3/openmpi/1.6.5/gcc/4.8.1/lib -lmetis -L/Users/oswald/extlib/openmpi/1.6.5/gcc/4.8.1/lib -lmpi -lm -lm -lm >&5
ld: library not found for -lmetis
collect2: error: ld returned 1 exit status
Greetings, Benedikt
On 28 Apr 2014, at 11:43, Markus Blatt <markus at dr-blatt.de> wrote:
> Hi,
>
> I tried to fix this hickup. Please try the lastet master of dune-istl.
>
> Markus
>
> On Sun, Apr 27, 2014 at 12:31:19AM +0200, Benedikt Oswald wrote:
>> Dear Dune,
>>
>> I just stumbled over the problem that, all of a sudden, including metis and parmetis creates the problem that metis.h is
>> not found anymore:
>>
>> although using totally valid directives as:
>>
>> --with-metis='$HOME/extlib/metis/5.1.0/gcc/4.8.2'
>> --with-parmetis='$HOME/extlib/parmetis/4.0.3/openmpi/1.8/gcc/4.8.2’
>>
>> eventually, the build process says:
>>
>> /Users/oswald/gitwork/hades-pure-dune/dune-istl/dune/istl/repartition.hh:14:19: fatal error: metis.h: No such file or directory
>>
>>
>> interestingly,
>>
>> 1) this was not a problem a few days ago, I have used this literally for years
>>
>> 2) using --with-parmetis='$HOME/extlib/parmetis/4.0.3/openmpi/1.8/gcc/4.8.2’ runs through
>>
>>
>> platform is:
>>
>> vpn-b-195-224:hades-pure-dune oswald$ uname -a
>> Darwin vpn-b-195-224.epfl.ch 13.1.0 Darwin Kernel Version 13.1.0: Thu Jan 16 19:40:37 PST 2014; root:xnu-2422.90.20~2/RELEASE_X86_64 x86_64
>>
>> gcc 4.8.2
>>
>> openmpi 1.8
>>
>> very latest dune trunk from tonight
>>
>>
>> were there changes to including metis / parmetis ?
>> Or, did I include them in a wrong way ?
>>
>> Greetings, Benedikt
>>
>
>
>
>> _______________________________________________
>> Dune mailing list
>> Dune at dune-project.org
>> http://lists.dune-project.org/mailman/listinfo/dune
>
>
> --
> Do you need more support with DUNE or HPC in general?
>
> Dr. Markus Blatt - HPC-Simulation-Software & Services http://www.dr-blatt.de
> Hans-Bunte-Str. 8-10, 69123 Heidelberg, Germany
> Tel.: +49 (0) 160 97590858
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.dune-project.org/pipermail/dune/attachments/20140428/396264c1/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 841 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <https://lists.dune-project.org/pipermail/dune/attachments/20140428/396264c1/attachment.sig>
More information about the Dune
mailing list