April 2011 - petsc-dev - mailman3.cels.anl.gov

Re: [petsc-dev] buildsystem package dependencies
by Lisandro Dalcin 05 Apr '11

05 Apr '11

On 5 April 2011 13:20, Satish Balay <balay(a)mcs.anl.gov> wrote: > Not sure I understand. If I say 'mpi depends on mpe' - then the link line is: > Sorry, I was thinking in terms of shared libraries. > -lhypre -lmpi -lmpe > > But thats incorrect usage - and not what I want.. The correct usage is: > -lhypre -lmpe -lmpi > > The issue comes up with -llmpe - which traps MPI_xx functions - and does some > logging - and then calls MPI_xx functions from -lmpi. i.e usage is: > If you are going to use MPE from MPICH2, "mpicc -mpe=mpilog" does the right thing: $ mpicc -mpe=mpilog -LLIBDIR -lLIBNAME -link_info gcc -I/usr/include/mpich2-i386 -m32 -O2 -Wl,-z,noexecstack -Wl,--export-dynamic -LLIBDIR -lLIBNAME -I/usr/include/mpich2-i386 -L/usr/lib/mpich2/lib -L/usr/lib/mpich2/lib -L/usr/lib/mpich2/lib -llmpe -lmpe -lmpich -lopa -lpthread -lrt -- Lisandro Dalcin --------------- CIMEC (INTEC/CONICET-UNL) Predio CONICET-Santa Fe Colectora RN 168 Km 472, Paraje El Pozo 3000 Santa Fe, Argentina Tel: +54-342-4511594 (ext 1011) Tel/Fax: +54-342-4511169

1 0

Re: [petsc-dev] buildsystem package dependencies
by Matthew Knepley 05 Apr '11

05 Apr '11

On Tue, Apr 5, 2011 at 11:20 AM, Satish Balay <balay(a)mcs.anl.gov> wrote: > Not sure I understand. If I say 'mpi depends on mpe' - then the link line > is: > > -lhypre -lmpi -lmpe > > But thats incorrect usage - and not what I want.. The correct usage is: > -lhypre -lmpe -lmpi > > The issue comes up with -llmpe - which traps MPI_xx functions - and does > some > logging - and then calls MPI_xx functions from -lmpi. i.e usage is: > > -lhypre -llmpe -lmpe -lmpi Right, but this MPE package is changing the MPI interface. All those other packages, like Hypre, depend on MPI. MPI depends on MPE, but the MPE module may modify the MPI lib to be '-lmpe -lmpi'. I do not see another way to do this. Matt > > Satish > > On Tue, 5 Apr 2011, Matthew Knepley wrote: > > > Yes, Lisandro is correct. > > > > Matt > > > > On Tue, Apr 5, 2011 at 11:05 AM, Lisandro Dalcin <dalcinl(a)gmail.com> > wrote: > > > > > On 5 April 2011 12:54, Satish Balay <balay(a)mcs.anl.gov> wrote: > > > > > > > > Matt, > > > > > > > > Is it possible to specify package dependencies such that the > following > > > happens: > > > > > > > > [all mpipackages like -lhypre] [-lmpe] -lmpi > > > > > > > > Currently I have: > > > > mpe: depends on mpi > > > > hypre: depends on mpi > > > > > > > > And configure can do: > > > > -lmpe -lhypre -lmpi > > > > > > > > But I want to say: > > > > mpe: depends on mpi, but before all other mpi packages. > > > > > > > > > > >From a library dependency POV, it is the other way around: MPI depends > > > on MPE, but the dependency is optional. > > > > > > > > > > > > -- > > > Lisandro Dalcin > > > --------------- > > > CIMEC (INTEC/CONICET-UNL) > > > Predio CONICET-Santa Fe > > > Colectora RN 168 Km 472, Paraje El Pozo > > > 3000 Santa Fe, Argentina > > > Tel: +54-342-4511594 (ext 1011) > > > Tel/Fax: +54-342-4511169 > > > > > > > > > > > > > -- 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

1 0

Re: [petsc-dev] buildsystem package dependencies
by Satish Balay 05 Apr '11

05 Apr '11

Not sure I understand. If I say 'mpi depends on mpe' - then the link line is: -lhypre -lmpi -lmpe But thats incorrect usage - and not what I want.. The correct usage is: -lhypre -lmpe -lmpi The issue comes up with -llmpe - which traps MPI_xx functions - and does some logging - and then calls MPI_xx functions from -lmpi. i.e usage is: -lhypre -llmpe -lmpe -lmpi Satish On Tue, 5 Apr 2011, Matthew Knepley wrote: > Yes, Lisandro is correct. > > Matt > > On Tue, Apr 5, 2011 at 11:05 AM, Lisandro Dalcin <dalcinl(a)gmail.com> wrote: > > > On 5 April 2011 12:54, Satish Balay <balay(a)mcs.anl.gov> wrote: > > > > > > Matt, > > > > > > Is it possible to specify package dependencies such that the following > > happens: > > > > > > [all mpipackages like -lhypre] [-lmpe] -lmpi > > > > > > Currently I have: > > > mpe: depends on mpi > > > hypre: depends on mpi > > > > > > And configure can do: > > > -lmpe -lhypre -lmpi > > > > > > But I want to say: > > > mpe: depends on mpi, but before all other mpi packages. > > > > > > > >From a library dependency POV, it is the other way around: MPI depends > > on MPE, but the dependency is optional. > > > > > > > > -- > > Lisandro Dalcin > > --------------- > > CIMEC (INTEC/CONICET-UNL) > > Predio CONICET-Santa Fe > > Colectora RN 168 Km 472, Paraje El Pozo > > 3000 Santa Fe, Argentina > > Tel: +54-342-4511594 (ext 1011) > > Tel/Fax: +54-342-4511169 > > > > > >

1 0

Re: [petsc-dev] buildsystem package dependencies
by Matthew Knepley 05 Apr '11

05 Apr '11

Yes, Lisandro is correct. Matt On Tue, Apr 5, 2011 at 11:05 AM, Lisandro Dalcin <dalcinl(a)gmail.com> wrote: > On 5 April 2011 12:54, Satish Balay <balay(a)mcs.anl.gov> wrote: > > > > Matt, > > > > Is it possible to specify package dependencies such that the following > happens: > > > > [all mpipackages like -lhypre] [-lmpe] -lmpi > > > > Currently I have: > > mpe: depends on mpi > > hypre: depends on mpi > > > > And configure can do: > > -lmpe -lhypre -lmpi > > > > But I want to say: > > mpe: depends on mpi, but before all other mpi packages. > > > > >From a library dependency POV, it is the other way around: MPI depends > on MPE, but the dependency is optional. > > > > -- > Lisandro Dalcin > --------------- > CIMEC (INTEC/CONICET-UNL) > Predio CONICET-Santa Fe > Colectora RN 168 Km 472, Paraje El Pozo > 3000 Santa Fe, Argentina > Tel: +54-342-4511594 (ext 1011) > Tel/Fax: +54-342-4511169 > -- 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

1 0

Re: [petsc-dev] buildsystem package dependencies
by Lisandro Dalcin 05 Apr '11

05 Apr '11

On 5 April 2011 12:54, Satish Balay <balay(a)mcs.anl.gov> wrote: > > Matt, > > Is it possible to specify package dependencies such that the following happens: > > [all mpipackages like -lhypre] [-lmpe] -lmpi > > Currently I have: > mpe: depends on mpi > hypre: depends on mpi > > And configure can do: > -lmpe -lhypre -lmpi > > But I want to say: > mpe: depends on mpi, but before all other mpi packages. > >From a library dependency POV, it is the other way around: MPI depends on MPE, but the dependency is optional. -- Lisandro Dalcin --------------- CIMEC (INTEC/CONICET-UNL) Predio CONICET-Santa Fe Colectora RN 168 Km 472, Paraje El Pozo 3000 Santa Fe, Argentina Tel: +54-342-4511594 (ext 1011) Tel/Fax: +54-342-4511169

1 0

Re: [petsc-dev] [petsc-maint #69333] Re: PetscBagSetName in F90
by Matthew Knepley 05 Apr '11

05 Apr '11

On Tue, Apr 5, 2011 at 10:03 AM, Ethan Coon <ecoon(a)lanl.gov> wrote: > Was there a problem with this patch? > There was a laziness wave in Chicago. Pushed, Matt > Ethan > > > On Wed, 2011-03-30 at 09:53 -0600, Ethan Coon wrote: > > PetscBagSetName isn't in the fortran world. The patch below adds it > > (and names a bag in the fortran example case to test). > > > > Thanks, > > > > Ethan > > > > -- > ------------------------------------ > Ethan Coon > Post-Doctoral Researcher > Applied Mathematics - T-5 > Los Alamos National Laboratory > 505-665-8289 > > http://www.ldeo.columbia.edu/~ecoon/ > ------------------------------------ > > -- 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

1 0

buildsystem package dependencies
by Satish Balay 05 Apr '11

05 Apr '11

Matt, Is it possible to specify package dependencies such that the following happens: [all mpipackages like -lhypre] [-lmpe] -lmpi Currently I have: mpe: depends on mpi hypre: depends on mpi And configure can do: -lmpe -lhypre -lmpi But I want to say: mpe: depends on mpi, but before all other mpi packages. Satish

1 0

Re: [petsc-dev] Compiling C code with C compiler
by Jed Brown 05 Apr '11

05 Apr '11

On Tue, Apr 5, 2011 at 17:29, Lisandro Dalcin <dalcinl(a)gmail.com> wrote: > Just to make sure the code is valid C++? > Fine, but what's the point? > Potential ABI issues with > some C++ compiler? > extern "C" is guaranteed to work according to the standard. Is there a known system where it fails?

1 0

Re: [petsc-dev] Compiling C code with C compiler
by Barry Smith 05 Apr '11

05 Apr '11

On Apr 5, 2011, at 10:02 AM, Jed Brown wrote: > On Tue, Apr 5, 2011 at 14:36, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > If it is as simple as you say then I am fine with adding support BUT it will require "yet another flag" If you don't have a flag to turn off using the C or C++ compiler on the "C" code then you are forcing the user into the model of C for C code and C++ for C++ code which we have never supported in 15 plus years. This is a new feature and has to have a way of turning it off. Barry If you make the C for C code feature the default that might not be the end of the world but you cannot make it be the only possibility. > > I'm not sure about this. What does PETSc need to do at compile time to expose the C++ interface? Isn't all the genuine C++ in headers anyway (e.g. logging interface in petsclog.hh, Exception and Point in petscsys.hh, almost all of Sieve, the few "*.c" files are obviously C++ only and thus should be renamed)? > > If indeed the C++ stuff (apart from Sieve) resides only in headers, we could unconditionally enable it when __cplusplus (so users could do a standard C build of PETSc and then use the "C++ bits" when calling PETSc from C++).

1 0

Re: [petsc-dev] Compiling C code with C compiler
by Lisandro Dalcin 05 Apr '11

05 Apr '11

On 5 April 2011 12:09, Jed Brown <jed(a)59a2.org> wrote: > On Tue, Apr 5, 2011 at 15:22, Matthew Knepley <knepley(a)gmail.com> wrote: >> >> I actually want the code compiled with the C++ compiler. > > Why? I'm really curious. Just to make sure the code is valid C++? Potential ABI issues with some C++ compiler? -- Lisandro Dalcin --------------- CIMEC (INTEC/CONICET-UNL) Predio CONICET-Santa Fe Colectora RN 168 Km 472, Paraje El Pozo 3000 Santa Fe, Argentina Tel: +54-342-4511594 (ext 1011) Tel/Fax: +54-342-4511169

1 0