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

Re: [petsc-dev] config/PETSc/FEM.py
by Matthew Knepley 06 Jun '12

06 Jun '12

On Wed, Jun 6, 2012 at 8:26 PM, Jed Brown <jedbrown(a)mcs.anl.gov> wrote: > On Wed, Jun 6, 2012 at 7:19 AM, Matthew Knepley <knepley(a)gmail.com> wrote: > >> It is Python code generation. We agreed that all Python code be under >> 'config'. >> > > Really? What about bin/pythonscripts (where other, possibly unused, FEM > files exist)? > 1) That place is only for binaries, i.e. things run by the user 2) Those FEM scripts are indeed used 3) The FEM scripts import FEM.py Matt -- 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] config/PETSc/FEM.py
by Jed Brown 06 Jun '12

06 Jun '12

On Wed, Jun 6, 2012 at 7:19 AM, Matthew Knepley <knepley(a)gmail.com> wrote: > It is Python code generation. We agreed that all Python code be under > 'config'. > Really? What about bin/pythonscripts (where other, possibly unused, FEM files exist)?

1 0

Re: [petsc-dev] config/PETSc/FEM.py
by Matthew Knepley 06 Jun '12

06 Jun '12

On Wed, Jun 6, 2012 at 11:09 AM, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > > Matt > > Why is config/PETSc/FEM.py where it is. It isn't configure stuff is it? > Can it be removed or moved to somewhere proper? > It is Python code generation. We agreed that all Python code be under 'config'. It is PETSc specific, so its under PETSc, like the PETSc specific configure files. I did not see the point in creating another directory. It can't be removed, because its so amazing. I have no idea where would be more proper. > What about ./config/BuildSystem/install.old Can that be removed? > Yes, terminate with extreme prejudice. Matt > Thanks > > Barry > > -- 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

config/PETSc/FEM.py
by Barry Smith 05 Jun '12

05 Jun '12

Matt Why is config/PETSc/FEM.py where it is. It isn't configure stuff is it? Can it be removed or moved to somewhere proper? What about ./config/BuildSystem/install.old Can that be removed? Thanks Barry

1 0

Re: [petsc-dev] misleading error message in VecWAXPY
by Jed Brown 05 Jun '12

05 Jun '12

We could check PetscIsInfOrNanScalar first. On Tue, Jun 5, 2012 at 1:13 PM, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > > Suggestions for improvements for the checks? > > #define PetscValidLogicalCollectiveScalar(a,b,c) \ > do { \ > PetscErrorCode _7_ierr; \ > PetscReal b1[2],b2[2]; \ > b1[0] = -PetscRealPart(b); b1[1] = PetscRealPart(b); \ > _7_ierr = > MPI_Allreduce(b1,b2,2,MPIU_REAL,MPIU_MAX,((PetscObject)a)->comm);CHKERRQ(_7_ierr); > \ > if (-b2[0] != b2[1]) > SETERRQ1(((PetscObject)a)->comm,PETSC_ERR_ARG_WRONG,"Scalar value must be > same on all processes, argument # %d",c); \ > } while (0) > > #define PetscValidLogicalCollectiveReal(a,b,c) \ > do { \ > PetscErrorCode _7_ierr; \ > PetscReal b1[2],b2[2]; \ > b1[0] = -b; b1[1] = b; \ > _7_ierr = > MPI_Allreduce(b1,b2,2,MPIU_REAL,MPIU_MAX,((PetscObject)a)->comm);CHKERRQ(_7_ierr); > \ > if (-b2[0] != b2[1]) > SETERRQ1(((PetscObject)a)->comm,PETSC_ERR_ARG_WRONG,"Real value must be > same on all processes, argument # %d",c); \ > } while (0) > > > > On Jun 5, 2012, at 11:25 AM, Blaise Bourdin wrote: > > > Hi, > > > > It looks like when VecWAXPY is called with alpha=Nan, > PetscValidLogicalCollectiveScalar causes the message "Scalar value must be > same on all processes, argument # 2" to be printed. This is a bit > misleading, and confusing when running on only 1 processor. > > > > Is this something worth fixing? > > > > 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] misleading error message in VecWAXPY
by Barry Smith 05 Jun '12

05 Jun '12

Suggestions for improvements for the checks? #define PetscValidLogicalCollectiveScalar(a,b,c) \ do { \ PetscErrorCode _7_ierr; \ PetscReal b1[2],b2[2]; \ b1[0] = -PetscRealPart(b); b1[1] = PetscRealPart(b); \ _7_ierr = MPI_Allreduce(b1,b2,2,MPIU_REAL,MPIU_MAX,((PetscObject)a)->comm);CHKERRQ(_7_ierr); \ if (-b2[0] != b2[1]) SETERRQ1(((PetscObject)a)->comm,PETSC_ERR_ARG_WRONG,"Scalar value must be same on all processes, argument # %d",c); \ } while (0) #define PetscValidLogicalCollectiveReal(a,b,c) \ do { \ PetscErrorCode _7_ierr; \ PetscReal b1[2],b2[2]; \ b1[0] = -b; b1[1] = b; \ _7_ierr = MPI_Allreduce(b1,b2,2,MPIU_REAL,MPIU_MAX,((PetscObject)a)->comm);CHKERRQ(_7_ierr); \ if (-b2[0] != b2[1]) SETERRQ1(((PetscObject)a)->comm,PETSC_ERR_ARG_WRONG,"Real value must be same on all processes, argument # %d",c); \ } while (0) On Jun 5, 2012, at 11:25 AM, Blaise Bourdin wrote: > Hi, > > It looks like when VecWAXPY is called with alpha=Nan, PetscValidLogicalCollectiveScalar causes the message "Scalar value must be same on all processes, argument # 2" to be printed. This is a bit misleading, and confusing when running on only 1 processor. > > Is this something worth fixing? > > 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

misleading error message in VecWAXPY
by Blaise Bourdin 05 Jun '12

05 Jun '12

Hi, It looks like when VecWAXPY is called with alpha=Nan, PetscValidLogicalCollectiveScalar causes the message "Scalar value must be same on all processes, argument # 2" to be printed. This is a bit misleading, and confusing when running on only 1 processor. Is this something worth fixing? 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] segv on Cray with 64K processes
by Jed Brown 05 Jun '12

05 Jun '12

On Tue, Jun 5, 2012 at 9:58 AM, Mark F. Adams <mark.adams(a)columbia.edu>wrote: > I suppose so. How does one do that? I don't know on Cray. There are environment variables that can be set like BG_COREDUMPBINARY=0,3,5 BG_COREDUMPONERROR=1 on Blue Gene. Presumably there is something similar on the Cray.

1 0

Re: [petsc-dev] segv on Cray with 64K processes
by Mark F. Adams 05 Jun '12

05 Jun '12

On Jun 5, 2012, at 10:49 AM, Jed Brown wrote: > On Tue, Jun 5, 2012 at 9:47 AM, Mark F. Adams <mark.adams(a)columbia.edu> wrote: > printf surrounded by barriers. > > There are a lot of system calls in PetscErrorPrintfInitialize. I've added printfs to this and submitted the job so I should have that info tomorrow. > > Can you instruct one process (or a few) to dump core? I suppose so. How does one do that?

1 0

Re: [petsc-dev] segv on Cray with 64K processes
by Jed Brown 05 Jun '12

05 Jun '12

On Tue, Jun 5, 2012 at 9:47 AM, Mark F. Adams <mark.adams(a)columbia.edu>wrote: > printf surrounded by barriers. > > There are a lot of system calls in PetscErrorPrintfInitialize. I've added > printfs to this and submitted the job so I should have that info tomorrow. > Can you instruct one process (or a few) to dump core?

1 0