<div dir="ltr"><div>Thanks a lot Satish,<br></div><div>I've learned a lot on Petsc build process !<br></div><div>I am away for the next two weeks and will test the patch when I come back .<br>Thanks again<br></div><div class="gmail_extra">Natacha <br></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Wed, Apr 20, 2016 at 6:43 PM, Satish Balay <span dir="ltr"><<a href="mailto:balay@mcs.anl.gov" target="_blank">balay@mcs.anl.gov</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Natacha,<br>
<br>
With <a href="https://bitbucket.org/petsc/petsc/commits/c0c82a18b4559f6e977fb5e0c503b006f23ed070" rel="noreferrer" target="_blank">https://bitbucket.org/petsc/petsc/commits/c0c82a18b4559f6e977fb5e0c503b006f23ed070</a><br>
you should be able to build mumps as you wish - and specify the dependencies in anyway<br>
[all clubed with mumps - or metis seprately - but scotch clubbed into metis]<br>
<br>
And I think --download-mumps=your-copy-of-mumps.tar.gz would be easier to use. But you<br>
might have other reasons to install mumps standalone - and specify to petsc [and this<br>
should also be fine]<br>
<br>
[for now this fix is in petsc git repo - and will be available in the next full release]<br>
<br>
thanks,<br>
Satish<br>
<div class="HOEnZb"><div class="h5"><br>
On Tue, 19 Apr 2016, Natacha BEREUX wrote:<br>
<br>
> Yes, we have a custom MUMPS version (because my company, EDF, belongs to<br>
> MUMPS consortium). Thereby, we get some patches and functionnalities in<br>
> advance with respect to the public version.<br>
><br>
> The option --download-mumps also works fine (I only have to define the path<br>
> to a local archive of sources). Our build does not use extra options.<br>
> We prefer to compile MUMPS alone, and in a second time build a compatible<br>
> PETSc library.<br>
><br>
> In fact we aim to use MUMPS in two ways :<br>
> - direct call of MUMPS driver (this is the legacy way)<br>
> - through PETSc interface (this is what I am experiencing)<br>
> Best regards<br>
> Natacha<br>
><br>
><br>
><br>
><br>
><br>
> On Tue, Apr 19, 2016 at 6:37 PM, Satish Balay <<a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>> wrote:<br>
><br>
> > I see ptscotch sources includes scotch aswell.<br>
> ><br>
> > BTW: What is the reason for installing mumps separately? Does it have<br>
> > some patches you are experimeinting with? Is it built with extra<br>
> > options that --download-mumps does not enable? Something else?<br>
> ><br>
> > thanks,<br>
> > Satish<br>
> ><br>
> > On Tue, 19 Apr 2016, Satish Balay wrote:<br>
> ><br>
> > > Natacha,<br>
> > ><br>
> > > Are you using mumps serially or parallely?<br>
> > ><br>
> > > My understanding is - MUMPS - when used parallely requires either<br>
> > > parmetis or ptscotch.<br>
> > ><br>
> > > And when used sequentially - it does not require anything [i.e<br>
> > > metis/scotch might be optional]<br>
> > ><br>
> > > Currently PETSc configure is trying to detect/use mumps parallely -<br>
> > > hence its insisting on parmetis or ptscotch.<br>
> > ><br>
> > > Assuming ptscotch doesn't conflict with scotch [I don't know if<br>
> > > ptscotch sources also includes scotch scources or not] - its fine to<br>
> > > use --download-ptscotch as a workarround. [or --download-parmetis -<br>
> > > assuming it can build with your install of metis]<br>
> > ><br>
> > > Ideally you should not need a workarround - hence my asking for<br>
> > > specifics of usage - to see if our configure check for mumps needs<br>
> > > fixing.<br>
> > ><br>
> > > Wrt specifying scotch or metis - PETSc configure supports metis - but<br>
> > > not scotch. So --with-metis-lib/--with-metis-include should work -<br>
> > > but not --with-scotch-lib/--with-scotch-include<br>
> > ><br>
> > > But its fine to lump both into mumps options - as they are primarily<br>
> > > dependencies of mumps anyway.<br>
> > ><br>
> > > Satish<br>
> > ><br>
> > > On Tue, 19 Apr 2016, Natacha BEREUX wrote:<br>
> > ><br>
> > > > Hello,<br>
> > > > Thanks a lot for your explanations.<br>
> > > > We have a "home-made" install of MUMPS, built with metis and scotch<br>
> > (and -<br>
> > > > at the moment- neither with parmetis nor with ptscotch).<br>
> > > ><br>
> > > > When I use --with-mumps-lib/--with-mumps-include, I am forced to<br>
> > specify<br>
> > > > scotch and metis libs inside mumps libs (otherwise the link of the<br>
> > check<br>
> > > > program dmumps_c fails). And this occurs even if I have already<br>
> > defined the<br>
> > > > location of scotch libraries through<br>
> > > > --with-scotch-lib/--with-scotch-include and<br>
> > > > --with-metis-lib/--with-metis-include options.<br>
> > > > Defining PETSc interfaces to scotch and metis packages is (at least for<br>
> > > > me) not sufficient : I have to specify these libraries locations in<br>
> > > > --with-mumps-lib. (as shown in the configure.log).<br>
> > > ><br>
> > > > This is not really a problem and doing so fixes the configure step.<br>
> > > ><br>
> > > > I agree that duplicating scotch to define ptscotch is weird. I did so<br>
> > > > because we do not have a parallel built of scotch.<br>
> > > > From PETSc point of view, MUMPS requires either Parmetis or PTScotch :<br>
> > so I<br>
> > > > provided ptscotch through --download-ptscotch option.<br>
> > > > I do not intend to use it, that is only for configuring purpose !<br>
> > > ><br>
> > > > Natacha<br>
> > > ><br>
> > > > On Mon, Apr 18, 2016 at 5:52 PM, Satish Balay <<a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>><br>
> > wrote:<br>
> > > ><br>
> > > > > Same with parmetis.<br>
> > > > ><br>
> > > > > On Mon, 18 Apr 2016, Natacha BEREUX wrote:<br>
> > > > ><br>
> > > > > > Hello Satish,<br>
> > > > > > thank you very much for yor advices. They were very helpful !<br>
> > > > > ><br>
> > > > > > The configure step finally succeeds if I use the following<br>
> > configure<br>
> > > > > line:<br>
> > > > > > ./configure --with-cc=mpicc --with-cxx=mpicxx --with-fc=mpif90<br>
> > > > > --with-mpi=1<br>
> > > > > > --with-debugging=0 --PETSC_ARCH=linux-metis-mumps<br>
> > > > > > --with-scalapack-lib="-lscalapack-openmpi -lblacs-openmpi<br>
> > > > > > -lblacsF77init-openmpi -lblacsCinit-openmpi"<br>
> > > > > ><br>
> > > > ><br>
> > --with-metis-lib="-L/home/H03755/dev/codeaster-prerequisites/v13/prerequisites/Metis_aster-510_aster/lib<br>
> > > > > > -lmetis -lGKlib"<br>
> > > > > ><br>
> > > > ><br>
> > --with-metis-include=/home/H03755/dev/codeaster-prerequisites/v13/prerequisites/Metis_aster-510_aster/include<br>
> > > > > ><br>
> > > > ><br>
> > --with-mumps-lib="-L/home/H03755/dev/codeaster-prerequisites/v13/prerequisites/Mumps-501_consortium_aster5/MPI/lib<br>
> > > > > > -lzmumps -ldmumps -lmumps_common -lpord<br>
> > > > > ><br>
> > > > ><br>
> > -L/home/H03755/dev/codeaster-prerequisites/v13/prerequisites/Scotch_aster-604_aster1/lib<br>
> > > > > > -lesmumps -lscotch -lscotcherr -lscotcherrexit<br>
> > > > > ><br>
> > > > ><br>
> > -L/home/H03755/dev/codeaster-prerequisites/v13/prerequisites/Metis_aster-510_aster/lib<br>
> > > > > > -lmetis"<br>
> > > > > ><br>
> > > > ><br>
> > --with-mumps-include=/home/H03755/dev/codeaster-prerequisites/v13/prerequisites/Mumps-501_consortium_aster5/MPI/include<br>
> > > > > > --with-blas-lapack-lib="-llapack -lopenblas"<br>
> > > > > > --download-ptscotch=/home/H03755/Librairies/scotch_6.0.3.tar.gz<br>
> > > > > LIBS=-lgomp<br>
> > > > > ><br>
> > > > > > I have to specify scotch shared libraries (-lscotch -lscotcherr<br>
> > > > > > -lscotcherrexit ) and metis shared library in --with-mumps-lib<br>
> > option<br>
> > > > > > otherwise the test (on dmump) in configure fails.<br>
> > > > > > Is it OK to do so ?<br>
> > > > ><br>
> > > > > Hm - its best to avoid duplication of libraries.<br>
> > > > ><br>
> > > > > i.e specifying scotch via mumps-libs and also via<br>
> > --download-ptscotch will<br>
> > > > > cause problems.<br>
> > > > ><br>
> > > > > Why not specify scotch with<br>
> > --with-ptscotch-include,--with-ptscotch-lib<br>
> > > > > options?<br>
> > > > ><br>
> > > > > ><br>
> > > > > > I also tried to use the following shorter line<br>
> > > > > ><br>
> > > > > > ./configure --with-cc=mpicc --with-cxx=mpicxx --with-fc=mpif90<br>
> > > > > > --with-mpi=1 --with-debugging=0 --PETSC_ARCH=linux-mumps-<br>
> > > > > > --with-scalapack-lib="-lscalapack-openmpi -lblacs-openmpi<br>
> > > > > > -lblacsF77init-openmpi -lblacsCinit-openmpi"<br>
> > > > > ><br>
> > > > ><br>
> > --with-mumps-lib="-L/home/H03755/dev/codeaster-prerequisites/v13/prerequisites/Mumps-501_consortium_aster5/MPI/lib<br>
> > > > > > -lzmumps -ldmumps -lmumps_common -lpord<br>
> > > > > ><br>
> > > > ><br>
> > -L/home/H03755/dev/codeaster-prerequisites/v13/prerequisites/Scotch_aster-604_aster1/lib<br>
> > > > > > -lesmumps -lscotch -lscotcherr -lscotcherrexit<br>
> > > > > ><br>
> > > > ><br>
> > -L/home/H03755/dev/codeaster-prerequisites/v13/prerequisites/Metis_aster-510_aster/lib<br>
> > > > > > -lmetis"<br>
> > > > > ><br>
> > > > ><br>
> > --with-mumps-include=/home/H03755/dev/codeaster-prerequisites/v13/prerequisites/Mumps-501_consortium_aster5/MPI/include<br>
> > > > > > --with-blas-lapack-lib="-llapack -lopenblas"<br>
> > > > > > --download-ptscotch=/home/H03755/Librairies/scotch_6.0.3.tar.gz<br>
> > > > > LIBS=-lgomp<br>
> > > > > ><br>
> > > > > > I do not use --with-metis-lib/--with-metis-include.<br>
> > > > > > I wonder if this is authorized : the libraries are given with the<br>
> > > > > > --with-mumps-lib, but the include are not defined.<br>
> > > > ><br>
> > > > > This is fine. [except for the dupliation of ptscotch]<br>
> > > > ><br>
> > > > > ><br>
> > > > > > What is the good practice ?<br>
> > > > ><br>
> > > > > Either is work - but its best to specify each package with its own<br>
> > > > > options listed by configure. The issues usually are:<br>
> > > > ><br>
> > > > > - Is this package primarily a depencency of an externalpakage or if<br>
> > there<br>
> > > > > is a petsc interface to it?<br>
> > > > ><br>
> > > > > For ex: PETSc has interface to mumps, parmetis - but not scalapack.<br>
> > So<br>
> > > > > if you club parmetis into mumps-libs then petsc interface to parmetis<br>
> > > > > would not be enabled. If you are not using this feature - it doesn't<br>
> > > > > matter if its not enabled.<br>
> > > > ><br>
> > > > > - does this externalpakage also require the includes in the public<br>
> > > > > interface?<br>
> > > > ><br>
> > > > > If dmumps_c.h requires metis.h [when mumps is built with metis]<br>
> > > > > - then you might have to specify metis include also with<br>
> > > > > --with-mumps-include. Otherwise - it doesn't matter.<br>
> > > > ><br>
> > > > > - are there bugs in petsc configure that can trigger wrong error<br>
> > checks?<br>
> > > > ><br>
> > > > > Because mumps depends on scalapack, and optionally on<br>
> > > > > metis,parmetis,ptscotch [i.e only one of them is required - but not<br>
> > > > > all] - there is an error-check in configure to make sure atleast one<br>
> > > > > of them is specified for --download-mumps. Perhaps this check should<br>
> > > > > not trigger error for user built mumps.<br>
> > > > ><br>
> > > > > Did you build MUMPS with both metis, ptscotch? [and not parmetis?]<br>
> > > > ><br>
> > > > > Satish<br>
> > > > ><br>
> > > > > ><br>
> > > > > > Best regards<br>
> > > > > > Natacha<br>
> > > > > ><br>
> > > > > ><br>
> > > > > ><br>
> > > > > ><br>
> > > > > ><br>
> > > > > > On Thu, Apr 14, 2016 at 6:07 PM, Satish Balay <<a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>><br>
> > wrote:<br>
> > > > > ><br>
> > > > > > > you'll have to roll-in the --with-blacs-lib option into<br>
> > > > > > > --with-scalapack-lib option<br>
> > > > > > ><br>
> > > > > > > Satish<br>
> > > > > > ><br>
> > > > > > > On Thu, 14 Apr 2016, Natacha BEREUX wrote:<br>
> > > > > > ><br>
> > > > > > > > Sorry, do not take into account my last email.<br>
> > > > > > > > I made some progress and I am now able to configure PETSc<br>
> > with a<br>
> > > > > > > > pre-installed version of metis.<br>
> > > > > > > ><br>
> > > > > > > > Problems come when I try to configure PETSc with MUMPS<br>
> > > > > > > ><br>
> > > > > > > > My command line is<br>
> > > > > > > > ./configure --with-cc=mpicc --with-cxx=mpicxx --with-fc=mpif90<br>
> > > > > > > > --with-ssl=0 --with-mpi=1 --with-debugging=1<br>
> > > > > > > > --PETSC_ARCH=linux-metis-mumps<br>
> > > > > > > > --with-scalapack-lib=/usr/lib/libscalapack-openmpi.so<br>
> > > > > > > ><br>
> > > > > > ><br>
> > > > ><br>
> > --with-blacs-lib=[/usr/lib/libblacs-openmpi.so,/usr/lib/libblacsCinit-openmpi.so,/usr/lib/libblacsF77init-openmpi.so]<br>
> > > > > > > ><br>
> > > > > > ><br>
> > > > ><br>
> > --with-metis-lib=[${METIS_PRE}/lib/libmetis.a,${METIS_PRE}/lib/libGKlib.a]<br>
> > > > > > > > --with-metis-include=$METIS_PRE/include<br>
> > > > > > > ><br>
> > > > > > ><br>
> > > > ><br>
> > --with-mumps-lib=[$MUMPS_PRE/lib/libdmumps.a,$MUMPS_PRE/lib/libmumps_common.a,$MUMPS_PRE/lib/libpord.a]<br>
> > > > > > > > --with-mumps-include=$MUMPS_PRE/include<br>
> > > > > > > ><br>
> > > > > > > > where METIS_PRE and MUMPS_PRE are the path to the local<br>
> > installs of<br>
> > > > > metis<br>
> > > > > > > > and mumps)<br>
> > > > > > > ><br>
> > > > > > > > I get (at least) the following error<br>
> > > > > > > > /libdmumps.a(dend_driver.o): undefined reference to symbol<br>
> > > > > > > 'blacs_gridexit_'<br>
> > > > > > > > /usr/lib/libblacs-openmpi.so.1: error adding symbols: DSO<br>
> > missing<br>
> > > > > from<br>
> > > > > > > > command line<br>
> > > > > > > > collect2: error: ld returned 1 exit status<br>
> > > > > > > ><br>
> > > > > > > ><br>
> > > > > > > > Would you have any idea of its meaning ?<br>
> > > > > > > ><br>
> > > > > > > > The configure.log is attached<br>
> > > > > > > > Thanks a lot if you can help me !<br>
> > > > > > > > Natacha<br>
> > > > > > > ><br>
> > > > > > > > On Thu, Apr 14, 2016 at 5:19 PM, Natacha BEREUX <<br>
> > > > > > > <a href="mailto:natacha.bereux@gmail.com">natacha.bereux@gmail.com</a>><br>
> > > > > > > > wrote:<br>
> > > > > > > ><br>
> > > > > > > > > Hi Satish<br>
> > > > > > > > > thanks a lot for the answer. Unfortunately, it does not work<br>
> > yet.<br>
> > > > > > > > > More precisely :<br>
> > > > > > > > > --download-mumps works fine (and every --download-package<br>
> > option<br>
> > > > > works<br>
> > > > > > > > > perfectly). I am then able to compile a PETSc library.<br>
> > > > > > > > > --with-package-lib=/usr/lib/libscalapack-openmpi.so or more<br>
> > > > > generally<br>
> > > > > > > > > --with-package-lib=libXXXX.so also works<br>
> > > > > > > > ><br>
> > > > > > > > > But I would like to use static librairies, preinstalled on<br>
> > my<br>
> > > > > computer<br>
> > > > > > > > > ... and this fails.<br>
> > > > > > > > ><br>
> > > > > > > > > For the moment I gave up compiling with MUMPS, and I am<br>
> > instead<br>
> > > > > > > trying to<br>
> > > > > > > > > compile with Metis 5.<br>
> > > > > > > > > I have a preinstalled version in a some directory lets say<br>
> > > > > metis_dir<br>
> > > > > > > > > I try<br>
> > > > > > > > > -with-metis-lib=[metis_dir/lib/libmetis.a,<br>
> > > > > metis_dir/lib/libGKlib.a]<br>
> > > > > > > > > --with-metis-include=metis_dir/include<br>
> > > > > > > > > this fails (see the attached config.log)<br>
> > > > > > > > > -with-metis-dir=metis_dir also fails<br>
> > > > > > > > > Is there a problem with static librairies ?<br>
> > > > > > > > ><br>
> > > > > > > > > Natacha<br>
> > > > > > > > ><br>
> > > > > > > > ><br>
> > > > > > > > > On Tue, Apr 12, 2016 at 6:19 PM, Satish Balay <<br>
> > <a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>><br>
> > > > > > > wrote:<br>
> > > > > > > > ><br>
> > > > > > > > >> On Tue, 12 Apr 2016, Natacha BEREUX wrote:<br>
> > > > > > > > >><br>
> > > > > > > > >> > Hello,<br>
> > > > > > > > >> > I am trying to compile Petsc (3.6.3) with external<br>
> > packages<br>
> > > > > (MUMPS<br>
> > > > > > > and<br>
> > > > > > > > >> its<br>
> > > > > > > > >> > prerequisites).<br>
> > > > > > > > >> > More precisely I would like PETSc to use a pre-installed<br>
> > > > > version of<br>
> > > > > > > > >> MUMPS.<br>
> > > > > > > > >> ><br>
> > > > > > > > >> > Petsc downloads and compiles the prerequisites (parmetis,<br>
> > > > > scalapack<br>
> > > > > > > > >> etc) :<br>
> > > > > > > > >> > this works fine.<br>
> > > > > > > > >><br>
> > > > > > > > >> What metis/parmetis/slcalapack is this MUMPS installed with?<br>
> > > > > > > > >><br>
> > > > > > > > >> What version of MUMPS did you install?<br>
> > > > > > > > >><br>
> > > > > > > > >> Why could you not use --download-mumps?<br>
> > > > > > > > >><br>
> > > > > > > > >> Using a different metis/parmetis/slcalapack to install<br>
> > MUMPS - and<br>
> > > > > > > > >> then specifying --download-metis --download-parmetis<br>
> > > > > > > > >> --download-scalapack [i.e different versions/builds of the<br>
> > same<br>
> > > > > > > > >> libraries] can result in conflcits.<br>
> > > > > > > > >><br>
> > > > > > > > >> ><br>
> > > > > > > > >> > I define MUMPS location by --with-mumps-dir=top-directory<br>
> > of<br>
> > > > > MUMPS<br>
> > > > > > > > >> install,<br>
> > > > > > > > >> > but the<br>
> > > > > > > > >> > configure step fails with the followiing message:<br>
> > > > > > > > >> ><br>
> > > > > > > > >> > UNABLE to CONFIGURE with GIVEN OPTIONS (see<br>
> > configure.log for<br>
> > > > > > > > >> details):<br>
> > > > > > > > >> > --with-mumps-dir=/home/H03755/Librairies/Mumps_MPI did<br>
> > not work<br>
> > > > > > > > >> ><br>
> > > > > > > > >> > I do not understand what is wrong.<br>
> > > > > > > > >> > I have attached the configure.log file.<br>
> > > > > > > > >> ><br>
> > > > > > > > >> > Any hint would be greatly appreciated !<br>
> > > > > > > > >><br>
> > > > > > > > >> >>>><br>
> > > > > > > > >> Executing: mpicc -show<br>
> > > > > > > > >> stdout: gcc -I/usr/lib/openmpi/include<br>
> > > > > > > -I/usr/lib/openmpi/include/openmpi<br>
> > > > > > > > >> -pthread -L/usr//lib -L/usr/lib/openmpi/lib -lmpi -ldl<br>
> > -lhwloc<br>
> > > > > > > > >> Defined make macro "MPICC_SHOW" to "gcc<br>
> > > > > > > > >> -I/usr/lib/openmpi/include<br>
> > -I/usr/lib/openmpi/include/openmpi<br>
> > > > > -pthread<br>
> > > > > > > > >> -L/usr//lib -L/usr/lib/openmpi/lib -lmpi -ldl -lhwloc"<br>
> > > > > > > > >> <<<<<br>
> > > > > > > > >> Ok - so you are using system openmpi with gcc.<br>
> > > > > > > > >><br>
> > > > > > > > >><br>
> > > > > > > > >> >>>><br>
> > > > > > > > >> Executing: mpicc -o<br>
> > /tmp/petsc-0u_4WI/config.libraries/conftest<br>
> > > > > > > -fPIC<br>
> > > > > > > > >> -Wall -Wwrite-strings -Wno-strict-aliasing<br>
> > -Wno-unknown-pragmas -O<br>
> > > > > > > > >> /tmp/petsc-0u_4WI/config.libraries/conftest.o<br>
> > > > > > > > >> -Wl,-rpath,/home/H03755/Librairies/Mumps_MPI/lib<br>
> > > > > > > > >> -L/home/H03755/Librairies/Mumps_MPI/lib -lcmumps -ldmumps<br>
> > -lsmumps<br>
> > > > > > > -lzmumps<br>
> > > > > > > > >> -lmumps_common -lpord<br>
> > > > > > > > >><br>
> > > > > > ><br>
> > > > ><br>
> > -Wl,-rpath,/home/H03755/Librairies/petsc-3.6.3/linux-debug-mumps-ext/lib<br>
> > > > > > > > >><br>
> > -L/home/H03755/Librairies/petsc-3.6.3/linux-debug-mumps-ext/lib<br>
> > > > > > > -lscalapack<br>
> > > > > > > > >> -llapack -lblas -Wl,-rpath,/usr/lib/openmpi/lib<br>
> > > > > -L/usr/lib/openmpi/lib<br>
> > > > > > > > >> -Wl,-rpath,/usr/lib/gcc/x86_64-linux-gnu/4.9<br>
> > > > > > > > >> -L/usr/lib/gcc/x86_64-linux-gnu/4.9<br>
> > > > > > > -Wl,-rpath,/usr/lib/x86_64-linux-gnu<br>
> > > > > > > > >> -L/usr/lib/x86_64-linux-gnu -Wl,-rpath,/lib/x86_64-linux-gnu<br>
> > > > > > > > >> -L/lib/x86_64-linux-gnu -lmpi_f90 -lmpi_f77 -lgfortran -lm<br>
> > > > > > > > >> -Wl,-rpath,/usr/lib/openmpi/lib<br>
> > > > > > > > >> -Wl,-rpath,/usr/lib/gcc/x86_64-linux-gnu/4.9<br>
> > > > > > > > >> -Wl,-rpath,/usr/lib/x86_64-linux-gnu<br>
> > > > > -Wl,-rpath,/lib/x86_64-linux-gnu<br>
> > > > > > > > >> -lgfortran -lm -lquadmath -lm -llapack -lblas<br>
> > > > > > > > >> -Wl,-rpath,/usr/lib/openmpi/lib -L/usr/lib/openmpi/lib<br>
> > > > > > > > >> -Wl,-rpath,/usr/lib/gcc/x86_64-linux-gnu/4.9<br>
> > > > > > > > >> -L/usr/lib/gcc/x86_64-linux-gnu/4.9<br>
> > > > > > > -Wl,-rpath,/usr/lib/x86_64-linux-gnu<br>
> > > > > > > > >> -L/usr/lib/x86_64-linux-gnu -Wl,-rpath,/lib/x86_64-linux-gnu<br>
> > > > > > > > >> -L/lib/x86_64-linux-gnu -lmpi_f90 -lmpi_f77 -lgfortran -lm<br>
> > > > > > > > >> -Wl,-rpath,/usr/lib/openmpi/lib<br>
> > > > > > > > >> -Wl,-rpath,/usr/lib/gcc/x86_64-linux-gnu/4.9<br>
> > > > > > > > >> -Wl,-rpath,/usr/lib/x86_64-linux-gnu<br>
> > > > > -Wl,-rpath,/lib/x86_64-linux-gnu<br>
> > > > > > > > >> -lgfortran -lm -lquadmath -lm<br>
> > > > > > > > >><br>
> > > > > > ><br>
> > > > ><br>
> > -Wl,-rpath,/home/H03755/Librairies/petsc-3.6.3/linux-debug-mumps-ext/lib<br>
> > > > > > > > >><br>
> > -L/home/H03755/Librairies/petsc-3.6.3/linux-debug-mumps-ext/lib<br>
> > > > > > > -lparmetis<br>
> > > > > > > > >><br>
> > > > > > ><br>
> > > > ><br>
> > -Wl,-rpath,/home/H03755/Librairies/petsc-3.6.3/linux-debug-mumps-ext/lib<br>
> > > > > > > > >><br>
> > -L/home/H03755/Librairies/petsc-3.6.3/linux-debug-mumps-ext/lib<br>
> > > > > > > -lmetis -lm<br>
> > > > > > > > >> -lm -Wl,-rpath,/usr/lib/openmpi/lib -L/usr/lib/openmpi/lib<br>
> > > > > > > > >> -Wl,-rpath,/usr/lib/gcc/x86_64-linux-gnu/4.9<br>
> > > > > > > > >> -L/usr/lib/gcc/x86_64-linux-gnu/4.9<br>
> > > > > > > -Wl,-rpath,/usr/lib/x86_64-linux-gnu<br>
> > > > > > > > >> -L/usr/lib/x86_64-linux-gnu -Wl,-rpath,/lib/x86_64-linux-gnu<br>
> > > > > > > > >> -L/lib/x86_64-linux-gnu -Wl,-rpath,/usr/lib/x86_64-linux-gnu<br>
> > > > > > > > >> -L/usr/lib/x86_64-linux-gnu -ldl -lmpi -lhwloc -lgcc_s<br>
> > -lpthread<br>
> > > > > -ldl<br>
> > > > > > > > >> Possible ERROR while running linker: exit code 256<br>
> > > > > > > > >> stderr:<br>
> > > > > > > > >><br>
> > /home/H03755/Librairies/Mumps_MPI/lib/libdmumps.a(dlr_stats.o): In<br>
> > > > > > > > >> function<br>
> > `__dmumps_lr_stats_MOD_update_flop_stats_lrb_product':<br>
> > > > > > > > >> dlr_stats.F:(.text+0x3079): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_start'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x30fa): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_end'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x310e): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_start'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x318f): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_end'<br>
> > > > > > > > >><br>
> > /home/H03755/Librairies/Mumps_MPI/lib/libdmumps.a(dlr_stats.o): In<br>
> > > > > > > > >> function `__dmumps_lr_stats_MOD_update_flop_stats_trsm':<br>
> > > > > > > > >> dlr_stats.F:(.text+0x33a9): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_start'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x33f9): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_end'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x340a): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_start'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x345a): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_end'<br>
> > > > > > > > >><br>
> > /home/H03755/Librairies/Mumps_MPI/lib/libdmumps.a(dlr_stats.o): In<br>
> > > > > > > > >> function `__dmumps_lr_stats_MOD_update_flop_stats_panel':<br>
> > > > > > > > >> dlr_stats.F:(.text+0x3576): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_start'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x35a7): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_end'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x35b8): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_start'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x35e9): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_end'<br>
> > > > > > > > >><br>
> > /home/H03755/Librairies/Mumps_MPI/lib/libdmumps.a(dlr_stats.o): In<br>
> > > > > > > > >> function `__dmumps_lr_stats_MOD_update_flop_stats_demote':<br>
> > > > > > > > >> dlr_stats.F:(.text+0x36ac): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_start'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x36ce): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_end'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x36df): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_start'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x3701): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_end'<br>
> > > > > > > > >><br>
> > /home/H03755/Librairies/Mumps_MPI/lib/libdmumps.a(dlr_stats.o): In<br>
> > > > > > > > >> function<br>
> > `__dmumps_lr_stats_MOD_update_flop_stats_cb_demote':<br>
> > > > > > > > >> dlr_stats.F:(.text+0x37c1): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_start'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x37e3): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_end'<br>
> > > > > > > > >><br>
> > /home/H03755/Librairies/Mumps_MPI/lib/libdmumps.a(dlr_stats.o): In<br>
> > > > > > > > >> function<br>
> > `__dmumps_lr_stats_MOD_update_flop_stats_cb_promote':<br>
> > > > > > > > >> dlr_stats.F:(.text+0x3839): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_start'<br>
> > > > > > > > >> dlr_stats.F:(.text+0x3856): undefined reference to<br>
> > > > > > > > >> `GOMP_critical_name_end'<br>
> > > > > > > > >><br>
> > /home/H03755/Librairies/Mumps_MPI/lib/libdmumps.a(dana_lr.o): In<br>
> > > > > > > function<br>
> > > > > > > > >> `__dmumps_ana_lr_MOD_mumps_scotch_kway':<br>
> > > > > > > > >> dana_lr.F:(.text+0x115): undefined reference to<br>
> > > > > `scotchfgraphbuild_'<br>
> > > > > > > > >> dana_lr.F:(.text+0x131): undefined reference to<br>
> > > > > `scotchfstratinit_'<br>
> > > > > > > > >> dana_lr.F:(.text+0x151): undefined reference to<br>
> > > > > `scotchfgraphpart_'<br>
> > > > > > > > >> dana_lr.F:(.text+0x15e): undefined reference to<br>
> > > > > `scotchfstratexit_'<br>
> > > > > > > > >> dana_lr.F:(.text+0x16b): undefined reference to<br>
> > > > > `scotchfgraphexit_'<br>
> > > > > > > > >> <snip><br>
> > > > > > > > >><br>
> > > > > > > > >><br>
> > > > > > > > >> Looks like this MUMPS is built with ptscotch and openmp.<br>
> > > > > > > > >><br>
> > > > > > > > >> You can specify -lgomp for openmp. This can be done with the<br>
> > > > > configure<br>
> > > > > > > > >> option<br>
> > > > > > > > >> LIBS=-lgomp<br>
> > > > > > > > >><br>
> > > > > > > > >> Wrt and PTSCOTCH stuff depending upon how its installed<br>
> > we'll<br>
> > > > > have to<br>
> > > > > > > > >> figureout<br>
> > > > > > > > >> how to specify it.<br>
> > > > > > > > >><br>
> > > > > > > > >> Its best to specify all the packages [mumps and its<br>
> > dependencies]<br>
> > > > > you<br>
> > > > > > > > >> built manually with the options:<br>
> > > > > > > > >><br>
> > > > > > > > >> --with-package-include --with-package-lib<br>
> > > > > > > > >><br>
> > > > > > > > >> Satish<br>
> > > > > > > > >><br>
> > > > > > > > >> > Best regards,<br>
> > > > > > > > >> ><br>
> > > > > > > > >> > Natacha<br>
> > > > > > > > >> ><br>
> > > > > > > > >><br>
> > > > > > > > ><br>
> > > > > > > > ><br>
> > > > > > > ><br>
> > > > > > ><br>
> > > > > > ><br>
> > > > > ><br>
> > > > ><br>
> > > > ><br>
> > > ><br>
> > ><br>
> > ><br>
> ><br>
> ><br>
><br>
<br>
</div></div></blockquote></div><br></div></div>