SNESSetUp() and matrix-free options

2009-07-22 Thread Lisandro Dalcin
On Wed, Jul 22, 2009 at 12:26 AM, Barry Smithbsmith at mcs.anl.gov wrote: Can we at least agree on the simple rules below: 1) KSPSetFromOptions() requires a previous call to KSPSetOperators() 2) SNESSetFromOptions() requires a previous call to: ?a) SNESSetFunction(), and also

SNESSetUp() and matrix-free options

2009-07-22 Thread Barry Smith
On Jul 22, 2009, at 9:39 AM, Lisandro Dalcin wrote: On Wed, Jul 22, 2009 at 12:26 AM, Barry Smithbsmith at mcs.anl.gov wrote: Can we at least agree on the simple rules below: 1) KSPSetFromOptions() requires a previous call to KSPSetOperators() 2) SNESSetFromOptions() requires a

Fwd: SNESSetUp() and matrix-free options

2009-07-21 Thread Barry Smith
PM CDT To: For users of the development version of PETSc petsc-dev at mcs.anl.gov Subject: Re: SNESSetUp() and matrix-free options Lisandro, The checking of the options can be moved to SNESSetFromOptions() that then set some flag in the snes object; but the actual action

SNESSetUp() and matrix-free options

2009-07-21 Thread Lisandro Dalcin
of the development version of PETSc petsc-dev at mcs.anl.gov Subject: Re: SNESSetUp() and matrix-free options ?Lisandro, ? ?The checking of the options can be moved to SNESSetFromOptions() that then set some flag in the snes object; but the actual action of MatCreateSNESMF() etc must continue to take

SNESSetUp() and matrix-free options

2009-07-21 Thread Barry Smith
that does not follow the rules. Begin forwarded message: From: Barry Smith bsmith at mcs.anl.gov Date: May 6, 2009 3:25:43 PM CDT To: For users of the development version of PETSc petsc-dev at mcs.anl.gov Subject: Re: SNESSetUp() and matrix-free options Lisandro, The checking

SNESSetUp() and matrix-free options

2009-07-21 Thread Lisandro Dalcin
not follow the rules. Begin forwarded message: From: Barry Smith bsmith at mcs.anl.gov Date: May 6, 2009 3:25:43 PM CDT To: For users of the development version of PETSc petsc-dev at mcs.anl.gov Subject: Re: SNESSetUp() and matrix-free options ?Lisandro, ? The checking of the options can

SNESSetUp() and matrix-free options

2009-07-21 Thread Barry Smith
forwarded message: From: Barry Smith bsmith at mcs.anl.gov Date: May 6, 2009 3:25:43 PM CDT To: For users of the development version of PETSc petsc-dev at mcs.anl.gov Subject: Re: SNESSetUp() and matrix-free options Lisandro, The checking of the options can be moved

SNESSetUp() and matrix-free options

2009-05-11 Thread Lisandro Dalcin
On Fri, May 8, 2009 at 3:55 PM, Barry Smith bsmith at mcs.anl.gov wrote: ?Quick look is ok for me. Just a note, to put it clear. With my patch, passing both -snes_mf -snes_mf_operator is valid, and equivalent to just passing -snes_mf_operator ... ?Barry On May 8, 2009, at 1:39 PM,

SNESSetUp() and matrix-free options

2009-05-08 Thread Barry Smith
Quick look is ok for me. Barry On May 8, 2009, at 1:39 PM, Lisandro Dalcin wrote: Well, it was not so easy to change the options name... We would need to update docs, tests, tutorial examples, and at this point it is not clear if we could agree on better options names... I'm now

SNESSetUp() and matrix-free options

2009-05-07 Thread Lisandro Dalcin
On Thu, May 7, 2009 at 5:14 PM, Barry Smith bsmith at mcs.anl.gov wrote: 3) I would like to add a call to set matrix-free programatically... something like SNESSetUseMFFD(snes,...). Any signature suggetions regarding my previous discussions about the many options? ? No need to add this. One

SNESSetUp() and matrix-free options

2009-05-07 Thread Barry Smith
One could have -snes_mf override -snes_mf_operator that would make sense. There is no hierarchy of options (first, last, in files etc) in order to keep the model simple. It is always possible someone has two options that conflict with each other in the options database. This is

SNESSetUp() and matrix-free options

2009-05-07 Thread Barry Smith
I don't have a problem with (somehow) merging the two options, but I don't see it as a big improvement. -snes_mf -snes_mf_type both or operator instead of -snes_mf or -snes_mf_operator I don't see this as a big improvement. Barry On May 7, 2009, at 3:47 PM, Jed Brown wrote:

SNESSetUp() and matrix-free options

2009-05-07 Thread Lisandro Dalcin
On Thu, May 7, 2009 at 6:14 PM, Jed Brown jed at 59a2.org wrote: Barry Smith wrote: ? ?I understand this. I just don't view this as a big deal worth changing the code over. I agree. ?My comment was that *if* the options were to be changed, as Lisandro was suggesting, then it would be better

SNESSetUp() and matrix-free options

2009-05-07 Thread Barry Smith
How about? -snes_mf (means what it currently means) -snes_mf operator (means what -snes_mf_operator means) Barry On May 7, 2009, at 6:03 PM, Lisandro Dalcin wrote: On Thu, May 7, 2009 at 6:14 PM, Jed Brown jed at 59a2.org wrote: Barry Smith wrote: I understand this. I

SNESSetUp() and matrix-free options

2009-05-06 Thread Lisandro Dalcin
Currently, -snes_mf* options are handled in SNESSetUp()... I think all you will agree that the proper place for them is SNESSetFromOptions()... However, perhaps I'm missing something and things are like that for some reason... In that case I would like to know the reason and perhaps I could fix

SNESSetUp() and matrix-free options

2009-05-06 Thread Barry Smith
Lisandro, The checking of the options can be moved to SNESSetFromOptions() that then set some flag in the snes object; but the actual action of MatCreateSNESMF() etc must continue to take place in the setup. Feel free to change if you like, Barry On May 6, 2009, at

SNESSetUp() and matrix-free options

2009-05-06 Thread Barry Smith
On May 6, 2009, at 3:37 PM, Lisandro Dalcin wrote: On Wed, May 6, 2009 at 5:25 PM, Barry Smith bsmith at mcs.anl.gov wrote: Lisandro, The checking of the options can be moved to SNESSetFromOptions() that then set some flag in the snes object; but the actual action of