July 2016 - petsc-dev - mailman3.cels.anl.gov

Re: [petsc-dev] petsc4py and DMDA
by gouarin 20 Jul '16

20 Jul '16

Le 20/07/2016 à 04:59, Matthew Knepley a écrit : > On Tue, Jul 19, 2016 at 11:12 PM, gouarin <loic.gouarin(a)math.u-psud.fr > <mailto:[email protected]>> wrote: > > Hi, > > I would like to use DMGetLocalVector and DMGetGlobalVector with > petsc4py but I didn't find the appropriate methods. If they are not > implemented, I can do a PR with these functions and their Restore part. > > > You are right, these are missing. > OK I will add these functions and do a PR. > > I have an other problem: I have to optimize a Python function that > take a numpy array given by the function getVecArray of a DMDA and I > would like to use Pythran to do that. The problem is that Pythran > doesn't recognize the array as a numpy array because the flags > C_CONTIGUOUS and F_CONTIGUOUS are both set to False. Is there a > reason that the C_CONTIGUOUS flag is set to False and not to True ? > > > For this, can't you just use vec.getArray()? It's ok if I use vec.getArray(). Thanks, Loic > > Thanks, > > Matt > > > Best, > Loic > > > > > -- > 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 -- Loic Gouarin Laboratoire de Mathématiques Université Paris-Sud Bâtiment 425 91405 Orsay Cedex France Tel: (+33) 1 69 15 60 14 Fax: (+33) 1 69 15 67 18

1 0

Re: [petsc-dev] petsc4py and DMDA
by Matthew Knepley 19 Jul '16

19 Jul '16

On Tue, Jul 19, 2016 at 11:12 PM, gouarin <loic.gouarin(a)math.u-psud.fr> wrote: > Hi, > > I would like to use DMGetLocalVector and DMGetGlobalVector with petsc4py > but I didn't find the appropriate methods. If they are not implemented, I > can do a PR with these functions and their Restore part. > You are right, these are missing. > I have an other problem: I have to optimize a Python function that take a > numpy array given by the function getVecArray of a DMDA and I would like to > use Pythran to do that. The problem is that Pythran doesn't recognize the > array as a numpy array because the flags C_CONTIGUOUS and F_CONTIGUOUS are > both set to False. Is there a reason that the C_CONTIGUOUS flag is set to > False and not to True ? > For this, can't you just use vec.getArray()? Thanks, Matt > Best, > Loic > -- 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

petsc4py and DMDA
by gouarin 19 Jul '16

19 Jul '16

Hi, I would like to use DMGetLocalVector and DMGetGlobalVector with petsc4py but I didn't find the appropriate methods. If they are not implemented, I can do a PR with these functions and their Restore part. I have an other problem: I have to optimize a Python function that take a numpy array given by the function getVecArray of a DMDA and I would like to use Pythran to do that. The problem is that Pythran doesn't recognize the array as a numpy array because the flags C_CONTIGUOUS and F_CONTIGUOUS are both set to False. Is there a reason that the C_CONTIGUOUS flag is set to False and not to True ? Best, Loic

1 0

Re: [petsc-dev] DMCreateDomainDecomposition in NASM
by Lawrence Mitchell 19 Jul '16

19 Jul '16

I presume this is in response to my email to the petsc-dev list about DMCreateDomainDecomposition. I have cc'd petsc-dev back in. On 18/07/16 18:35, Lulu Liu wrote: > Hi, > > DMCreateDomainDecomposition() is called in src/snes/impls/nasm/nasm.c, > and nasm->n always equals to 1. Could you fixed this problem in PETSc? > Thanks! Not being a core petsc developer, I cannot do this directly. Also, It seems to me that your DM is returning the wrong number of subdomains from DMCreateDomainDecomposition. If you're using a DMDA, I think you need to call DMDASetNumLocalSubdomains. If you're using a DMPlex, then this functionality is not currently implemented, I believe. Cheers, Lawrence

1 0

Re: [petsc-dev] Nested event logging and human friendly output
by Klaij, Christiaan 19 Jul '16

19 Jul '16

Barry, Thanks for testing! Koos is on holiday now, we'll come back to this later. Chris dr. ir. Christiaan Klaij | CFD Researcher | Research & Development MARIN | T +31 317 49 33 44 | mailto:[email protected] | http://www.marin.nl MARIN news: http://www.marin.nl/web/News/News-items/Van-VTS-naar-VPC-15-september-2016.… ________________________________________ From: Barry Smith <bsmith(a)mcs.anl.gov> Sent: Sunday, July 17, 2016 4:44 AM To: Koos Huijssen Cc: Richard Mills; petsc-dev(a)mcs.anl.gov; Klaij, Christiaan; Bas van 't Hof Subject: Re: [petsc-dev] Nested event logging and human friendly output Thank you very much for fixing the errors I introduced and improving the code. I have put it into branch barry/fix-xml-logging and merged to next for testing. I do have one concern, when I run on src/snes/examples/tutorials/ex19.c with no options (see attached result)

1 0

Re: [petsc-dev] Fortran interfaces for VecGetLocalVector and VecRestoreLocalVector
by Patrick Sanan 18 Jul '16

18 Jul '16

Thanks! On Mon, Jul 18, 2016 at 9:19 PM, Satish Balay <balay(a)mcs.anl.gov> wrote: > pushed this fix to maint. > > Satish > > On Mon, 18 Jul 2016, Satish Balay wrote: > >> yes - these 2 functions can have fortran stubs autogenerated. >> >> Removing the 'C' - and then running configure [which does an >> equivalent of 'make allfortranstubs'] - and 'make all' would be the >> steps.. >> >> Satih >> >> On Mon, 18 Jul 2016, Patrick Sanan wrote: >> >> > I'd like to use VecGetLocalVector and VecRestoreLocalVector from Fortran. >> > >> > It appears that these are not covered by the Fortran interface, based >> > on the fact that they don't appear in my include/petsc/finclude >> > directory anywhere, even though other related functions do. >> > >> > This seems to be confirmed by the fact that this function has the >> > "comment-only" Sowing block in src/vec/vec/interface/rvector.c and >> > does not appear in the ftn-custom (or f90-custom) directory there. >> > >> > I'm new to using Fortran directly with PETSc. I can delete the "C" >> > from the Sowing blocks, rebuild the library, and link, but I'm not >> > confident enough to say that this worked yet. Should that be all that >> > is required to make these functions available? >> > >> >> >

1 0

Re: [petsc-dev] Fortran interfaces for VecGetLocalVector and VecRestoreLocalVector
by Satish Balay 18 Jul '16

18 Jul '16

pushed this fix to maint. Satish On Mon, 18 Jul 2016, Satish Balay wrote: > yes - these 2 functions can have fortran stubs autogenerated. > > Removing the 'C' - and then running configure [which does an > equivalent of 'make allfortranstubs'] - and 'make all' would be the > steps.. > > Satih > > On Mon, 18 Jul 2016, Patrick Sanan wrote: > > > I'd like to use VecGetLocalVector and VecRestoreLocalVector from Fortran. > > > > It appears that these are not covered by the Fortran interface, based > > on the fact that they don't appear in my include/petsc/finclude > > directory anywhere, even though other related functions do. > > > > This seems to be confirmed by the fact that this function has the > > "comment-only" Sowing block in src/vec/vec/interface/rvector.c and > > does not appear in the ftn-custom (or f90-custom) directory there. > > > > I'm new to using Fortran directly with PETSc. I can delete the "C" > > from the Sowing blocks, rebuild the library, and link, but I'm not > > confident enough to say that this worked yet. Should that be all that > > is required to make these functions available? > > > >

1 0

Re: [petsc-dev] Fortran interfaces for VecGetLocalVector and VecRestoreLocalVector
by Satish Balay 18 Jul '16

18 Jul '16

yes - these 2 functions can have fortran stubs autogenerated. Removing the 'C' - and then running configure [which does an equivalent of 'make allfortranstubs'] - and 'make all' would be the steps.. Satih On Mon, 18 Jul 2016, Patrick Sanan wrote: > I'd like to use VecGetLocalVector and VecRestoreLocalVector from Fortran. > > It appears that these are not covered by the Fortran interface, based > on the fact that they don't appear in my include/petsc/finclude > directory anywhere, even though other related functions do. > > This seems to be confirmed by the fact that this function has the > "comment-only" Sowing block in src/vec/vec/interface/rvector.c and > does not appear in the ftn-custom (or f90-custom) directory there. > > I'm new to using Fortran directly with PETSc. I can delete the "C" > from the Sowing blocks, rebuild the library, and link, but I'm not > confident enough to say that this worked yet. Should that be all that > is required to make these functions available? >

1 0

Fortran interfaces for VecGetLocalVector and VecRestoreLocalVector
by Patrick Sanan 18 Jul '16

18 Jul '16

I'd like to use VecGetLocalVector and VecRestoreLocalVector from Fortran. It appears that these are not covered by the Fortran interface, based on the fact that they don't appear in my include/petsc/finclude directory anywhere, even though other related functions do. This seems to be confirmed by the fact that this function has the "comment-only" Sowing block in src/vec/vec/interface/rvector.c and does not appear in the ftn-custom (or f90-custom) directory there. I'm new to using Fortran directly with PETSc. I can delete the "C" from the Sowing blocks, rebuild the library, and link, but I'm not confident enough to say that this worked yet. Should that be all that is required to make these functions available?

1 0

Re: [petsc-dev] Question about DMCreateDomainDecomposition()
by Lawrence Mitchell 18 Jul '16

18 Jul '16

On 18/07/16 15:49, Lulu Liu wrote: > Hi, > > I add the following lines in the example > /petsc/src/snes/examples/tutorials/ex19. When I run this example using > different numbers of processors, n always equals to 1. However, > according to the manual, n is the number of subproblems in the domain > decomposition. I feel confused why n does not return the number of > processors, could you explain it? Thanks! > > PetscInt n; > DM *subdms; > ierr > =DMCreateDomainDecomposition(da,&n,NULL,NULL,NULL,&subdms);CHKERRQ(ierr); > PetscPrintf(PETSC_COMM_WORLD,"### n==%d\n", n); > You need to do -da_local_subdomains nsubdomain_per_process or else call DMDASetNumLocalSubdomains Cheers, Lawrence

1 0