August 2012 - petsc-dev - mailman3.cels.anl.gov

Re: [petsc-dev] TS fortran binding patch
by Jed Brown 17 Aug '12

17 Aug '12

Do you have any tests for this? There were some duplicate names +#define tssetifunction_ tssetifunction +#define tssetifunction_ tsgetifunction +#define tssetijacobian_ tssetijacobian +#define tssetijacobian_ tsgetijacobian and missing error checking +void PETSC_STDCALL tsgetijacobian_(TS *ts,Mat *J,Mat *M,int *func,void **ctx,PetscErrorCode *ierr) +{ + *ierr = TSGetIJacobian(*ts,J,M,0,ctx); +} which I am fixing up. On Fri, Aug 17, 2012 at 9:00 AM, Blaise Bourdin <bourdin(a)lsu.edu> wrote: > Hi, > > The attached trivial patch adds fortran bindings for a few TS functions > (TSGetIFunction, TSGetRHSFunction, TSGetIJAcobian, TSSetOptiosnPrefix, > TSAppendOptionsPrefix). Can somebody push it to petsc-3.3 if it makes sense? > > Blaise > > > -- > Department of Mathematics and Center for Computation & Technology > Louisiana State University, Baton Rouge, LA 70803, USA > Tel. +1 (225) 578 1612, Fax +1 (225) 578 4276 > http://www.math.lsu.edu/~bourdin > > > > > > > > >

1 0

TS fortran binding patch
by Blaise Bourdin 17 Aug '12

17 Aug '12

Hi, The attached trivial patch adds fortran bindings for a few TS functions (TSGetIFunction, TSGetRHSFunction, TSGetIJAcobian, TSSetOptiosnPrefix, TSAppendOptionsPrefix). Can somebody push it to petsc-3.3 if it makes sense? Blaise -- Department of Mathematics and Center for Computation & Technology Louisiana State University, Baton Rouge, LA 70803, USA Tel. +1 (225) 578 1612, Fax +1 (225) 578 4276 http://www.math.lsu.edu/~bourdin

1 0

Re: [petsc-dev] Command Line Link Utility for PETSc
by Jed Brown 15 Aug '12

15 Aug '12

On Wed, Aug 15, 2012 at 1:01 PM, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > > Yup. But sadly not likely instead of enumerating but in addition to > enumerating. > Right, it's only distributed with quite recent versions of MKL. > > Barry > > On Aug 15, 2012, at 10:53 AM, Jed Brown <jed(a)59A2.org> wrote: > > > It looks like configure can use this instead of enumerated probing of > link options. > > > > Should we search for this thing (in tools/mkl_link_tool) based on a > configure flag like > --with-blas-lapack-dir=/soft/com/packages/intel/12.0/339/mkl ? > > > > ---------- Forwarded message ---------- > > From: Guney, Murat E <murat.e.guney(a)intel.com> > > Date: Wed, Aug 15, 2012 at 1:07 AM > > Subject: Command Line Link Utility for PETSc > > To: "jed(a)59A2.org" <jed(a)59a2.org> > > Cc: "Sukharina, Julia A" <julia.a.sukharina(a)intel.com>, "Rosenquist, > Todd" <todd.rosenquist(a)intel.com> > > > > > > Hi Jed, > > > > > > > > Julia (CC’ed) is the developer of our command line link assist tool. You > could also find more information about the tool here: > http://software.intel.com/en-us/articles/mkl-command-line-link-tool/. > Please feel free to ask any questions regarding to the tool. > > > > > > > > Hi Julia, > > > > > > > > Jed Brown is a PETSc developer/designer. He’s interested in using our > command line tool with PETSc. > > > > > > > > Thank you, > > > > Efe > > > > > > > > > >

1 0

Re: [petsc-dev] Command Line Link Utility for PETSc
by Barry Smith 15 Aug '12

15 Aug '12

Yup. But sadly not likely instead of enumerating but in addition to enumerating. Barry On Aug 15, 2012, at 10:53 AM, Jed Brown <jed(a)59A2.org> wrote: > It looks like configure can use this instead of enumerated probing of link options. > > Should we search for this thing (in tools/mkl_link_tool) based on a configure flag like --with-blas-lapack-dir=/soft/com/packages/intel/12.0/339/mkl ? > > ---------- Forwarded message ---------- > From: Guney, Murat E <murat.e.guney(a)intel.com> > Date: Wed, Aug 15, 2012 at 1:07 AM > Subject: Command Line Link Utility for PETSc > To: "jed(a)59A2.org" <jed(a)59a2.org> > Cc: "Sukharina, Julia A" <julia.a.sukharina(a)intel.com>, "Rosenquist, Todd" <todd.rosenquist(a)intel.com> > > > Hi Jed, > > > > Julia (CC’ed) is the developer of our command line link assist tool. You could also find more information about the tool here: http://software.intel.com/en-us/articles/mkl-command-line-link-tool/. Please feel free to ask any questions regarding to the tool. > > > > Hi Julia, > > > > Jed Brown is a PETSc developer/designer. He’s interested in using our command line tool with PETSc. > > > > Thank you, > > Efe > > > >

1 0

Fwd: Command Line Link Utility for PETSc
by Jed Brown 15 Aug '12

15 Aug '12

It looks like configure can use this instead of enumerated probing of link options. Should we search for this thing (in tools/mkl_link_tool) based on a configure flag like --with-blas-lapack-dir=/soft/com/packages/intel/12.0/339/mkl ? ---------- Forwarded message ---------- From: Guney, Murat E <murat.e.guney(a)intel.com> Date: Wed, Aug 15, 2012 at 1:07 AM Subject: Command Line Link Utility for PETSc To: "jed(a)59A2.org" <jed(a)59a2.org> Cc: "Sukharina, Julia A" <julia.a.sukharina(a)intel.com>, "Rosenquist, Todd" < todd.rosenquist(a)intel.com> Hi Jed,**** ** ** Julia (CC’ed) is the developer of our command line link assist tool. You could also find more information about the tool here: http://software.intel.com/en-us/articles/mkl-command-line-link-tool/. Please feel free to ask any questions regarding to the tool. **** ** ** Hi Julia,**** ** ** Jed Brown is a PETSc developer/designer. He’s interested in using our command line tool with PETSc. **** ** ** Thank you,**** Efe**** ** **

1 0

Re: [petsc-dev] broken TS example related to Jacobian and coloring etc.
by Peter Brune 15 Aug '12

15 Aug '12

Ummmm. OK. We should remove custom -fd arguments from every example first and use -snes_fd_coloring instead. The example problems should work with that and if they do not then we fix it. The context should always be owned elsewhere so if we manually unset it we should be fine. - Peter On Aug 15, 2012 8:31 AM, "Jed Brown" <jedbrown(a)mcs.anl.gov> wrote: > On Wed, Aug 15, 2012 at 6:41 AM, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > >> I think it uses null in order to delay computing the matrix coloring etc >> from the time of SNESSetJacobian() to a later time. I guess what could >> happen is it creates a n "empty" fdcoloring beast and passes that into >> SNESSetJacobian() and then fills it up later when it has the information it >> needs. > > > Right, I think I would prefer that. A side-effect is that the caller > setting coloring is now responsible for ownership of the context. but it is > for everything else anyway. > > Lisandro asked for destructors for those contexts, but I'm not sure we > need a separate one for every context slot. Would one per SNES be enough? >

1 0

Re: [petsc-dev] broken TS example related to Jacobian and coloring etc.
by Jed Brown 15 Aug '12

15 Aug '12

On Wed, Aug 15, 2012 at 6:41 AM, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > I think it uses null in order to delay computing the matrix coloring etc > from the time of SNESSetJacobian() to a later time. I guess what could > happen is it creates a n "empty" fdcoloring beast and passes that into > SNESSetJacobian() and then fills it up later when it has the information it > needs. Right, I think I would prefer that. A side-effect is that the caller setting coloring is now responsible for ownership of the context. but it is for everything else anyway. Lisandro asked for destructors for those contexts, but I'm not sure we need a separate one for every context slot. Would one per SNES be enough?

1 0

Re: [petsc-dev] broken TS example related to Jacobian and coloring etc.
by Barry Smith 15 Aug '12

15 Aug '12

On Aug 15, 2012, at 1:03 AM, Jed Brown <jedbrown(a)mcs.anl.gov> wrote: > On Tue, Aug 14, 2012 at 5:03 PM, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > But you've just told me that ONLY ts is allowed to use the SNESSetJacobian() context slot (coloring and presumably users) aren't allowed to. > > Not at all. I mean that coloring as produced internally by SNES should not ascribe special meaning to that field being NULL or not. You could perhaps have -snes_fd_color create an empty coloring object and put it into the callback slowt. We should perhaps support destructors for function/Jacobian contexts. I think it uses null in order to delay computing the matrix coloring etc from the time of SNESSetJacobian() to a later time. I guess what could happen is it creates a n "empty" fdcoloring beast and passes that into SNESSetJacobian() and then fills it up later when it has the information it needs. Barry

1 0

Re: [petsc-dev] broken TS example related to Jacobian and coloring etc.
by Jed Brown 15 Aug '12

15 Aug '12

On Tue, Aug 14, 2012 at 5:03 PM, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > But you've just told me that ONLY ts is allowed to use the > SNESSetJacobian() context slot (coloring and presumably users) aren't > allowed to. Not at all. I mean that coloring as produced internally by SNES should not ascribe special meaning to that field being NULL or not. You could perhaps have -snes_fd_color create an empty coloring object and put it into the callback slowt. We should perhaps support destructors for function/Jacobian contexts.

1 0

Re: [petsc-dev] broken TS example related to Jacobian and coloring etc.
by Barry Smith 14 Aug '12

14 Aug '12

On Aug 14, 2012, at 5:59 PM, Jed Brown <jedbrown(a)mcs.anl.gov> wrote: > On Tue, Aug 14, 2012 at 4:51 PM, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > How about just having a SNESSetTS()? > > Please no. But you've just told me that ONLY ts is allowed to use the SNESSetJacobian() context slot (coloring and presumably users) aren't allowed to. So you already have in effect a SNESSetTS() ??? Barry

1 0