[petsc-users] petsc-users Digest, Vol 13, Issue 28

[Barry Smith]

On Jan 21, 2010, at 2:07 PM, Jack Lee wrote:

> Thanks for the reply. Perhaps I should have explained a little  
> further:
>
> > 1) MUMPS is a linear solver
> I am aware of that, what I meant was that instead of the default  
> iterative solver which snes employs, I'd like the option of being  
> able to use an external solver for each linear solution of the non- 
> linear loop.
>
> > 2)
> > http://www.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/docs/manualpages/PC/PCFactorSetMatSolverPackage.html
> Currently I am applying these steps:
> 1. set ksp context extracted from snes to pre only
> 2. set pc type to lu
> 3. set pc factor solver package to MUMPS, as you've explained above.
> However, having no examples to compare to, I'm not 100% sure if this  
> is the way it's supposed to be done.
>
> > 3) There is a manual section on this
> if you're talking about page 83, it doesn't spell out the case which  
> I'm trying to apply to, so thought I'd better double-check.
>
> > 4) You can use command line options in Fortran
> Not in the custom fortran application we're building, which is  
> compiled as a library. We're linking in petsc to exploit its  
> fantastic solver capabilities....

    You can also put the options into a file called .petscrc, petscrc  
or ~/.petscrc  or a file of your own choosing and pass the filename  
into PetscInitialize() or the environmental variable PETSC_OPTIONS    
The ability to change options without recompiling/relinking is what  
keeps me sane :-)


    Barry


>
> Cheers,
> Jack
>
>
>
> petsc-users-request at mcs.anl.gov wrote:
>> Send petsc-users mailing list submissions to
>> 	petsc-users at mcs.anl.gov
>> To subscribe or unsubscribe via the World Wide Web, visit
>> 	https://lists.mcs.anl.gov/mailman/listinfo/petsc-users
>> or, via email, send a message with subject or body 'help' to
>> 	petsc-users-request at mcs.anl.gov
>> You can reach the person managing the list at
>> 	petsc-users-owner at mcs.anl.gov
>> When replying, please edit your Subject line so it is more specific
>> than "Re: Contents of petsc-users digest..."
>> Today's Topics:
>>   1.  using external direct linear solvers (Jack Lee)
>>   2. Re:  using external direct linear solvers (Matthew Knepley)
>>   3.  using PETSC with visual studio 2005? (Zhifeng Sheng - EWI)
>>   4. Re:  using PETSC with visual studio 2005? (Satish Balay)
>> ----------------------------------------------------------------------
>> Message: 1
>> Date: Wed, 20 Jan 2010 22:50:47 +0000
>> From: Jack Lee <jack.lee at sjc.ox.ac.uk>
>> Subject: [petsc-users] using external direct linear solvers
>> To: petsc-users at mcs.anl.gov
>> Message-ID: <4B5788C7.20902 at sjc.ox.ac.uk>
>> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>> Hi there,
>> Apologies if this is an obvious question, but is there an example  
>> showing how to call an external solver? I am trying to get the snes  
>> solver to use MUMPS. As the calls are made from a fortran  
>> application, I cannot use the commandline options.
>> Jack
>> ------------------------------
>> Message: 2
>> Date: Wed, 20 Jan 2010 17:14:16 -0600
>> From: Matthew Knepley <knepley at gmail.com>
>> Subject: Re: [petsc-users] using external direct linear solvers
>> To: PETSc users list <petsc-users at mcs.anl.gov>
>> Message-ID:
>> 	<a9f269831001201514x473b849ci76cc9ab80d754586 at mail.gmail.com>
>> Content-Type: text/plain; charset="iso-8859-1"
>> On Wed, Jan 20, 2010 at 4:50 PM, Jack Lee <jack.lee at sjc.ox.ac.uk>  
>> wrote:
>>> Hi there,
>>>
>>> Apologies if this is an obvious question, but is there an example  
>>> showing
>>> how to call an external solver? I am trying to get the snes solver  
>>> to use
>>> MUMPS. As the calls are made from a fortran application, I cannot  
>>> use the
>>> commandline options.
>>>
>> 1) MUMPS is a linear solver
>> 2)
>> http://www.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/docs/manualpages/PC/PCFactorSetMatSolverPackage.html
>> 3) There is a manual section on this
>> 4) You can use command line options in Fortran
>>  Matt
>> Jack