July 2013 - petsc-dev - mailman3.cels.anl.gov

Re: [petsc-dev] many outstanding pull requests
by Karl Rupp 31 Jul '13

31 Jul '13

Hi Barry, > John Fettig avatarJohn Fettig > #56: CUDA: fix CMake builds with the Intel compiler > Comments 7 > 2013-06-13 this one is fairly tricky, as the fix for the Intel compiler will break compilation on other compilers. I'll get in touch with John to get this done. Best regards, Karli

1 0

Re: [petsc-dev] Q/A: submiminor on PETSc
by Toby D. Young 31 Jul '13

31 Jul '13

> petsc-3.3-p4 > > Now the new notation for the same is: > > petsc-3.3.4 That's ultra clear now, and I have it written down. Thank you. ------ Toby D. Young Assistant Professor Division of Computational Method Institute of Fundamental Technological Research of the Polish Academy of Sciences ul Pawinskiego 5b 02-106 Warsaw Poland www: http://www.ippt.pan.pl/~tyoung Session on the grid-based quantum many-body problem: http://www.ippt.pan.pl/~tyoung/icnaam.html

1 0

Re: [petsc-dev] Q/A: submiminor on PETSc
by Toby D. Young 31 Jul '13

31 Jul '13

> We've tried to come around to the "community standard" with the sub > minor numbering. It represents bug fixes from previous sub minors > but NO API changes. Users should be able to switch to a higher sub > minor without any changes to their code. That explanation fits in perfectly with my expectations and experience. Keyword: API. Thank you for clarifying what I already knew was correct. PS. Can I get my commit permissions back for PETSc dev please? ------ Toby D. Young Assistant Professor Division of Computational Method Institute of Fundamental Technological Research of the Polish Academy of Sciences ul Pawinskiego 5b 02-106 Warsaw Poland www: http://www.ippt.pan.pl/~tyoung Session on the grid-based quantum many-body problem: http://www.ippt.pan.pl/~tyoung/icnaam.html

1 0

Re: [petsc-dev] Q/A: submiminor on PETSc
by Satish Balay 31 Jul '13

31 Jul '13

i.e - for bug fix updates - previously we used the notation: petsc-3.3-p4 Now the new notation for the same is: petsc-3.3.4 we switched from incrementing PETSC_VERSION_PATCH to incrementing PETSC_VERSION_SUBMINOR Satish On Wed, 31 Jul 2013, Barry Smith wrote: > > Toby, > > We've tried to come around to the "community standard" with the sub minor numbering. It represents bug fixes from previous sub minors but NO API changes. Users should be able to switch to a higher sub minor without any changes to their code. > > Barry > > On Jul 31, 2013, at 6:03 AM, "Toby D. Young" <tyoung(a)ippt.pan.pl> wrote: > > > > > > > I have been out of the loop for around-about-a-half-a-year by now, so > > please forgive me while I catch up and get back on track. > > > > What is the deal with PETSc's subminor version these days? > > > > I recall a discussion ages ago, which I do not want to restart because I > > do not care *that* much. I recall that subminor was for all purposes > > dropped. (No?) > > > > I recently updated petsc-dev to discover it has 3.4.2 version number. > > Where did the "2" come from in petsc-dev and what does it mean to users? > > I completely missed "0" and "1" somehow, but like I said, I've been out of > > the loop... > > > > Best regs, > > Toby > > > > > > ------ > > > > Toby D. Young > > Assistant Professor > > > > Division of Computational Method > > Institute of Fundamental Technological Research > > of the Polish Academy of Sciences > > ul Pawinskiego 5b > > 02-106 Warsaw > > Poland > > > > www: http://www.ippt.pan.pl/~tyoung > > > > Session on grid-based quantum many-body theory: > > http://www.ippt.pan.pl/~tyoung/icnaam.html > > > > > > > > > >

1 0

Re: [petsc-dev] Q/A: submiminor on PETSc
by Barry Smith 31 Jul '13

31 Jul '13

Toby, We've tried to come around to the "community standard" with the sub minor numbering. It represents bug fixes from previous sub minors but NO API changes. Users should be able to switch to a higher sub minor without any changes to their code. Barry On Jul 31, 2013, at 6:03 AM, "Toby D. Young" <tyoung(a)ippt.pan.pl> wrote: > > > I have been out of the loop for around-about-a-half-a-year by now, so > please forgive me while I catch up and get back on track. > > What is the deal with PETSc's subminor version these days? > > I recall a discussion ages ago, which I do not want to restart because I > do not care *that* much. I recall that subminor was for all purposes > dropped. (No?) > > I recently updated petsc-dev to discover it has 3.4.2 version number. > Where did the "2" come from in petsc-dev and what does it mean to users? > I completely missed "0" and "1" somehow, but like I said, I've been out of > the loop... > > Best regs, > Toby > > > ------ > > Toby D. Young > Assistant Professor > > Division of Computational Method > Institute of Fundamental Technological Research > of the Polish Academy of Sciences > ul Pawinskiego 5b > 02-106 Warsaw > Poland > > www: http://www.ippt.pan.pl/~tyoung > > Session on grid-based quantum many-body theory: > http://www.ippt.pan.pl/~tyoung/icnaam.html > > > >

1 0

Q/A: submiminor on PETSc
by Toby D. Young 31 Jul '13

31 Jul '13

I have been out of the loop for around-about-a-half-a-year by now, so please forgive me while I catch up and get back on track. What is the deal with PETSc's subminor version these days? I recall a discussion ages ago, which I do not want to restart because I do not care *that* much. I recall that subminor was for all purposes dropped. (No?) I recently updated petsc-dev to discover it has 3.4.2 version number. Where did the "2" come from in petsc-dev and what does it mean to users? I completely missed "0" and "1" somehow, but like I said, I've been out of the loop... Best regs, Toby ------ Toby D. Young Assistant Professor Division of Computational Method Institute of Fundamental Technological Research of the Polish Academy of Sciences ul Pawinskiego 5b 02-106 Warsaw Poland www: http://www.ippt.pan.pl/~tyoung Session on grid-based quantum many-body theory: http://www.ippt.pan.pl/~tyoung/icnaam.html

1 0

future of PetscSubcomm
by Barry Smith 31 Jul '13

31 Jul '13

Jed, Hong and I were discussing MatGetRedundantMatrix() etc and she mention that you don't like PetscSubcomm. I suggested removing the PetscSubcomm object from PETSc but retaining the functionality of PetscSubcomm by putting the subcomm information into an MPI attribute inside an MPI_Comm (so this new comm behaves like a MPI_Comm subclass with extra data over a regular MPI_Comm) and having something like PetscCommSublimate(comm, …., &subcomm with attribute). Then in all the places in PETSc that take PetscSubcomm they would instead take a MPI_Comm and check for the attribute and use the additional information if it is there. Managing destruction will be tricky due to possible reference loops but presumably we can figure out a way to do it. What do you think? Barry

1 0

many outstanding pull requests
by Barry Smith 30 Jul '13

30 Jul '13

We have a bunch of outstanding pull requests that have been hanging around for a while that should get resolved. How do we manage this? BarryFSmith #67: hdf5 installed without zlib should configure without neededing zlib dependency Comments 5 2013-07-23 psanan #70: Added MatSetUp calls to TS tutorials 16 and 21 2013-07-15 Emil Constantinescu #60: Replace van der Pol ODE example and new nonstiff ODE examples 2013-06-19 Emil Constantinescu #59: Renovated explicit RK Comments 2 2013-06-19 Emil Constantinescu #58: ARK stage predictors, bug fix in ARK interpolate 2013-06-18 Emil Constantinescu #57: Explicit RK, ARK stage predictors, new ODE examples Comments 17 2013-06-18 John Fettig #56: CUDA: fix CMake builds with the Intel compiler Comments 7 2013-06-13 BarryFSmith #49: Correct logging for object and its decsendants Approvals 1 2013-05-29

1 0

Deprecated PETSc functions and Fortran codes?
by Richard Tran Mills 30 Jul '13

30 Jul '13

Hi Folks, Is there a way to deprecate PETSc functions that will work with Fortran codes? In include/petscdmda.h I have PETSC_DEPRECATED("Use DMLocalToLocalBegin()") PETSC_STATIC_INLINE PetscErrorCode DMDALocalToLocalBegin(DM dm,Vec g,InsertMode mode,Vec l) {return DMLocalToLocalBegin(dm,g,mode,l);} PETSC_DEPRECATED("Use DMLocalToLocalEnd()") PETSC_STATIC_INLINE PetscErrorCode DMDALocalToLocalEnd(DM dm,Vec g,InsertMode mode,Vec l) {return DMLocalToLocalEnd(dm,g,mode,l);} And this works fine with C examples: /Users/rmills/proj/petsc-git/mac_gcc47_g/bin/mpicc -o ex7.o -c -fPIC -Wall -Wwrite-strings -Wno-strict-aliasing -Wno-unknown-pragmas -g -I/Users/rmills/proj/petsc-git/include -I/Users/rmills/proj/petsc-git/mac_gcc47_g/include -I/opt/X11/include -D__INSDIR__=src/dm/examples/tests/ ex7.c ex7.c: In function 'main': ex7.c:72:3: warning: 'DMDALocalToLocalBegin' is deprecated (declared at /Users/rmills/proj/petsc-git/include/petscdmda.h:52): Use DMLocalToLocalBegin() [-Wdeprecated-declarations] ex7.c:73:3: warning: 'DMDALocalToLocalEnd' is deprecated (declared at /Users/rmills/proj/petsc-git/include/petscdmda.h:53): Use DMLocalToLocalEnd() [-Wdeprecated-declarations] But with a Fortran code (PFLOTRAN), I just get this at link time: Undefined symbols for architecture x86_64: "_dmdalocaltolocalbegin_", referenced from: ___discretization_module_MOD_discretizationlocaltolocalbegin in discretization.o ___discretization_module_MOD_discretizationlocaltolocal in discretization.o "_dmdalocaltolocalend_", referenced from: ___discretization_module_MOD_discretizationlocaltolocalend in discretization.o ___discretization_module_MOD_discretizationlocaltolocal in discretization.o Is there any good way to get the same behavior as with the C code? Not a huge issue of course, but it would be nice to have the consistency. --Richard -- Richard Tran Mills, Ph.D. Computational Earth Scientist | Joint Assistant Professor Hydrogeochemical Dynamics Team | EECS and Earth & Planetary Sciences Oak Ridge National Laboratory | University of Tennessee, Knoxville E-mail: rmills(a)ornl.gov V: 865-241-3198 http://climate.ornl.gov/~rmills

1 0

Re: [petsc-dev] fun plots
by Karl Rupp 30 Jul '13

30 Jul '13

Hey, > Yes, because we always put the important information down below some totally irrelevant information that most people never need that takes up the whole screen. START with the overview information, who's got time to scroll on a cellphone. Actually, there is some headline which is supposed to describe what the project is about. The authors arguably did not exploit the full potential here: "hub helps you win at git. http://hub.github.com/" Following the link finally gives the desired information: "hub is a command-line wrapper for git that makes you better at GitHub." You could complain at Aron for sending a not very Barry-friendly link to the repository instead of the main webpage ;-) Best regards, Karli

1 0