[petsc-users] configure error on Titan with Intel

[Mark Adams]

On Fri, May 8, 2015 at 3:17 PM, Satish Balay <balay at mcs.anl.gov> wrote:

> >>>>>>
> Executing: cc -c -o /tmp/petsc-8S0Qxp/config.setCompilers/conftest.o
> -I/tmp/petsc-8S0Qxp/config.setCompilers
>  /tmp/petsc-8S0Qxp/config.setCompilers/conftest.c
> <snip>
> Executing: cc  -o /tmp/petsc-8S0Qxp/config.setCompilers/conftest
> /tmp/petsc-8S0Qxp/config.setCompilers/conftest.o -lstdc++
> Executing: /tmp/petsc-8S0Qxp/config.setCompilers/conftest
> Executing: /tmp/petsc-8S0Qxp/config.setCompilers/conftest
> ERROR while running executable: Could not execute
> "/tmp/petsc-8S0Qxp/config.setCompilers/conftest":
> <<<<<<
>
> Hm - there is no error message [or error number] listed.
>
> Perhaps you can try debugging this without PETSc. i.e compile/run a
> simple code and see if it works.
>
> You can use src/benchmarks/sizeof.c or
>
> http://trac.mpich.org/projects/mpich/export/60482cc2ca9a740d049444a9261a04a3007bc306/examples/cpi.c
>
> [perhaps runs from $HOME work - but not /tmp ? - but then you said -
> 'This workflow seemed to work for pgi & gnu programming envs' - so
> this shouldn't be the problem]
>

Yea, I tried this workflow with PGI and it worked.  I'm giving up on Intel
and see if we can make GNU work for Treb.

Thanks,


>
> Satish
>
> On Fri, 8 May 2015, Mark Adams wrote:
>
> > I added the LD path stuff and that seems to have fixed this problem ...
> but
> > there are still problems.
> >
> > On Fri, May 8, 2015 at 1:23 PM, David Trebotich <dptrebotich at lbl.gov>
> wrote:
> >
> > > This has all come up because I cannot build Chombo-crunch using gnu
> > > prgenv---cannot find mkl anymore.  So olcf consultants advised Intel
> > > prgenv.  I am not sure yet what form of libs are used, that is mkl,
> blas, I
> > > don't know.
> > >
> > > David Trebotich
> > > sent from mobile
> > > (510) 384-6868
> > > On May 8, 2015 9:22 AM, "Satish Balay" <balay at mcs.anl.gov> wrote:
> > >
> > >> On Fri, 8 May 2015, Ed D'Azevedo wrote:
> > >>
> > >> > Is petsc using ACML specific routines?
> > >> > If not, perhaps MKL provides the lapack/blas and ACML is not needed?
> > >> >
> > >> > Just a thought.
> > >>
> > >> yup PETSc can use whatever blas/lapck user provides.
> > >>
> > >> In this case PETSc was not seeking acml - but the 'module' env was
> > >> presumably setup with acml [and the compilers are attempting to use
> > >> these modules directly - irrespective of PETSc]
> > >>
> > >> Satish
> > >>
> > >
> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20150509/d012b3fc/attachment.html>