[petsc-users] PETSc Windows Installation

Satish Balay balay at mcs.anl.gov
Mon Oct 31 08:34:31 CDT 2022


Make sure you have cygwin openmpi installed [and cywin blas/lapack]

$ cygcheck -cd |grep openmpi
libopenmpi-devel              4.1.2-1
libopenmpi40                  4.1.2-1
libopenmpifh40                4.1.2-1
libopenmpiusef08_40           4.1.2-1
libopenmpiusetkr40            4.1.2-1
openmpi                       4.1.2-1
$ cygcheck -cd |grep lapack
liblapack-devel               3.10.1-1
liblapack0                    3.10.1-1


> ./configure --with-cc=gcc --with-cxx=0 --with-fc=0 --download-f2cblaslapack

Should be:

> > $ ./configure --download-scalapack --download-mumps

i.e [default] --with-cc=mpicc --with-cxx=mpicxx --with-fc=mpif90 [an default cygwin blas/lapack]

Satish


On Mon, 31 Oct 2022, Matthew Knepley wrote:

> On Mon, Oct 31, 2022 at 1:56 AM Mohammad Ali Yaqteen <mhyaqteen at sju.ac.kr>
> wrote:
> 
> > Dear Satish
> >
> > When I configure PETSc with (./configure --with-cc=gcc --with-cxx=0
> > --with-fc=0 --download-f2cblaslapack) it runs as I shared initially which
> > you said is not an issue anymore. But when I add  (--download-scalapack
> > --download-mumps) or configure with these later, it gives the following
> > error:
> >
> > $ ./configure --download-scalapack --download-mumps
> >
> > =============================================================================================
> >                          Configuring PETSc to compile on your system
> >
> > =============================================================================================
> > TESTING: FortranMPICheck from
> > config.packages.MPI(config/BuildSystem/config/packages/MPI.py:614)*******************************************************************************
> >          UNABLE to CONFIGURE with GIVEN OPTIONS    (see configure.log for
> > details):
> >
> > -------------------------------------------------------------------------------
> > Fortran error! mpi_init() could not be located!
> >
> > *******************************************************************************
> >
> > What could be the problem here?
> >
> 
> Without configure.log we cannot tell what went wrong. However, from the
> error message, I would guess that your MPI
> was not built with Fortran bindings. You need these for those packages.
> 
>   Thanks,
> 
>      Matt
> 
> 
> > Your help is highly appreciated.
> >
> > Thank you
> > Ali
> >
> > -----Original Message-----
> > From: Satish Balay <balay at mcs.anl.gov>
> > Sent: Saturday, October 29, 2022 2:11 PM
> > To: Mohammad Ali Yaqteen <mhyaqteen at sju.ac.kr>
> > Cc: Matthew Knepley <knepley at gmail.com>; petsc-users at mcs.anl.gov
> > Subject: Re: [petsc-users] PETSc Windows Installation
> >
> > On Sat, 29 Oct 2022, Mohammad Ali Yaqteen wrote:
> >
> > > I haven’t accessed PETSC or given any command of my own. I was just
> > installing by following the instructions. I don’t know why it is attaching
> > the debugger. Although it says “Possible error running C/C++
> > src/snes/tutorials/ex19 with 1 MPI process” which I think is indicating of
> > missing of MPI!
> >
> > The diff is not smart enough to detect the extra message from
> > cygwin/OpenMPI - hence it assumes there is a potential problem - and prints
> > the above message.
> >
> > But you can assume its installed properly - and use it.
> >
> > Satish
> > >
> > > From: Matthew Knepley <knepley at gmail.com>
> > > Sent: Friday, October 28, 2022 10:31 PM
> > > To: Mohammad Ali Yaqteen <mhyaqteen at sju.ac.kr>
> > > Cc: petsc-users at mcs.anl.gov
> > > Subject: Re: [petsc-users] PETSc Windows Installation
> > >
> > > On Fri, Oct 28, 2022 at 9:11 AM Mohammad Ali Yaqteen <
> > mhyaqteen at sju.ac.kr<mailto:mhyaqteen at sju.ac.kr>> wrote:
> > > Dear Sir,
> > >
> > > During the Installation of PETSc in windows, I installed Cygwin and the
> > required libraries as mentioned on your website:
> > > [cid:image001.png at 01D8EB93.7C17E410]
> > > However, when I install PETSc using the configure commands present on
> > the petsc website:
> > >
> > > ./configure --with-cc=gcc --with-cxx=0 --with-fc=0
> > --download-f2cblaslapack --download-mpich
> > >
> > > it gives me the following error:
> > >
> > > [cid:image002.png at 01D8EB93.7C17E410]
> > >
> > > I already installed OpenMPI using Cygwin installer but it still asks me
> > to. When I configure without “—download-mpich” and run “make check”
> > command, it gives me the following errors:
> > >
> > > [cid:image003.png at 01D8EB93.7C17E410]
> > >
> > > Could you kindly look into this and help me with this? Your prompt
> > response will highly be appreciated.
> > >
> > > The runs look fine.
> > >
> > > The test should not try to attach the debugger. Do you have that in the
> > PETSC_OPTIONS env variable?
> > >
> > >   Thanks,
> > >
> > >      Matt
> > >
> > > Thank you!
> > > Mohammad Ali
> > > Researcher, Sejong University
> > >
> > >
> > > --
> > > What most experimenters take for granted before they begin their
> > experiments is infinitely more interesting than any results to which their
> > experiments lead.
> > > -- Norbert Wiener
> > >
> > > https://www.cse.buffalo.edu/~knepley/<
> > http://www.cse.buffalo.edu/~knepley/>
> > >
> >
> 
> 
> 


More information about the petsc-users mailing list