[petsc-dev] [petsc-maint] compile PETSc with HYPRE

Barry Smith bsmith at mcs.anl.gov
Tue May 21 10:03:36 CDT 2013


On May 21, 2013, at 5:39 AM, Matthew Knepley <knepley at gmail.com> wrote:

> On Mon, May 20, 2013 at 8:15 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
> 
> >
> > Thank you for replying. The problem is gone after I reset the variable
> > 'C_INCLUDE_PATH' and remove the old values.
> 
>     Is there we can make PETSc configure and make not let this variable waste peoples time? (Configure I know can force it to be ignored, does it currently?)
> 
> Was it us picking this up through CMake, or was it the Hypre build?

   I assume this is a regular old environmental variable 


>    Matt
>  
>     Thanks
> 
>         Barry
> 
> >
> > Jian  Du
> >
> >
> >> Jian Du <du at math.utah.edu> writes:
> >>
> >>> Dear PETSc Developer,
> >>>
> >>> I am now trying to compile PETSc, for which I want to use PC from HYPRE.
> >>> I used configure option '--download-hypre', but always got error
> >>> messages
> >>> like ''PC_HYPRE' has no member named xxx' during make. Could you please
> >>> let me know what is the possible cause of this? I have attached make.log
> >>> file for your reference.
> >>
> >> I suspect you have a broken version of Hypre somewhere, perhaps in your
> >> environment.
> >>
> >> /u/ma/du/petsc-3.3-p7/src/ksp/pc/impls/hypre/hypre.c:19: error: expected
> >> specifier-qualifier-list before 'HYPRE_Int'
> >> /u/ma/du/petsc-3.3-p7/src/ksp/pc/impls/hypre/hypre.c: In function
> >> 'PCSetUp_HYPRE':
> >> /u/ma/du/petsc-3.3-p7/src/ksp/pc/impls/hypre/hypre.c:85: error: 'PC_HYPRE'
> >> has no member named 'hypre_type'
> >>
> >> You can use the -M compiler option to figure out where the Hypre headers
> >> are coming from.
> >>
> >>>
> >>> I have tried the compiling with the latest version of PETSc and got the
> >>> similar errors.
> >>
> >
> >
> 
> 
> 
> 
> -- 
> 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




More information about the petsc-dev mailing list