[petsc-users] FETI-DP implementation and call sequence

Sat Sep 6 08:48:59 CDT 2014

BTW: You could also instal lapack (liblapack-devel), openmpi
(openmpi-devel) packages from cygwin - and avoid
--download-f2cblaslapack and --download-mpich

Satish

On Sat, 6 Sep 2014, Satish Balay wrote:

> On Sat, 6 Sep 2014, Alp Kalpalp wrote:
> 
> > Hi,
> > 
> > I switched back to master again and PETSC_VERSION_SUBMINOR is equal to 1
> > again. I think with the merge of stefano_zampini/pcbddc-primalfixe
> > it is changing to 0.
> 
> the merge won't change PETSC_VERSION_SUBMINOR
> 
> > 
> > As you suggested I removed --with-shared-libraries option, but although
> > make all seems working make test is failed with attached errors.
> 
> I'll suggest "rm -rf /home/alp/petsc/arch-mswin-c-debug/" and then retry
> 
> or if you could clean the git repo with:
> git reset --hard
> git clean -f -d -x
> 
> And then rerun configure
> 
> Satish
> 
> > 
> > thanks
> > 
> > 
> > On Sat, Sep 6, 2014 at 2:44 AM, Satish Balay <balay at mcs.anl.gov> wrote:
> > 
> > > perhaps unrelated - but are you sure you had the latest master?
> > >
> > > >>>>
> > > PETSC_VERSION_RELEASE    0
> > > PETSC_VERSION_MAJOR      3
> > > PETSC_VERSION_MINOR      5
> > > PETSC_VERSION_SUBMINOR   0
> > > PETSC_VERSION_PATCH      0
> > > <<<<<<<
> > >
> > > PETSC_VERSION_PATCH should be 1
> > >
> > >
> > > >>>>>
> > >      CLINKER /home/alp/petsc/arch-mswin-c-debug/lib/libpetsc.so.3.05.0
> > > Warning: corrupt .drectve at end of def file
> > > Warning: corrupt .drectve at end of def file
> > > Warning: corrupt .drectve at end of def file
> > >
> > > rch-mswin-c-debug/obj/src/sys/utils/mpiu.o:(.text+0xdd2): undefined
> > > reference to `__security_cookie'
> > > arch-mswin-c-debug/obj/src/sys/utils/mpiu.o:(.text+0xdd2): relocation
> > > truncated to fit: R_X86_64_PC32 against undefined symbol `__security_cookie'
> > > <<<<<<
> > >
> > > Hm - must be related to --with-shared-libraries option you are
> > > specifying. This is untested on windows.
> > >
> > > Can you remove it and retry?
> > >
> > > Satish
> > >
> > > On Fri, 5 Sep 2014, Alp Kalpalp wrote:
> > >
> > > > Hi,
> > > >
> > > > Sorry for the late response. I tried the same sequence. But I need to say
> > > > that I use git pull before it. Error is now during make all
> > > >
> > > > you may find the logs in the attachment.
> > > >
> > > > thanks beforehand
> > > >
> > > > regards,
> > > >
> > > >
> > > > On Fri, Sep 5, 2014 at 2:51 PM, Matthew Knepley <knepley at gmail.com>
> > > wrote:
> > > >
> > > > > Please send your configure.log and make.log
> > > > >
> > > > >   Thanks,
> > > > >
> > > > >     Matt
> > > > >
> > > > >
> > > > > On Fri, Sep 5, 2014 at 4:27 AM, Alp Kalpalp <alpkalpalp at gmail.com>
> > > wrote:
> > > > >
> > > > >> As I said before, I have checked out "master" branch and merged with
> > > your
> > > > >> stefano_zampini/pcbddc-primalfixe branch. configured, compiled
> > > successfully.
> > > > >> I have used --with-pcbddc option in configure as Barry suggested.
> > > > >>
> > > > >> However, tests are failed with following reason:
> > > > >>
> > > > >> akalpalp at a-kalpalp ~/petsc
> > > > >> $ make PETSC_DIR=/home/akalpalp/petsc PETSC_ARCH=arch-mswin-c-debug
> > > test
> > > > >> Running test examples to verify correct installation
> > > > >> Using PETSC_DIR=/home/akalpalp/petsc and PETSC_ARCH=arch-mswin-c-debug
> > > > >> *******************Error detected during compile or
> > > > >> link!*******************
> > > > >> See http://www.mcs.anl.gov/petsc/documentation/faq.html
> > > > >> /home/akalpalp/petsc/src/snes/examples/tutorials ex19
> > > > >>
> > > > >>
> > > *********************************************************************************
> > > > >> /home/akalpalp/petsc/arch-mswin-c-debug/bin/mpicc -o ex19.o -c -Wall
> > > > >> -Wwrite-strings -Wno-strict-aliasing -Wno-unknown-pragmas -g3 -O0
> > > > >> -I/home/akalpalp/petsc/include
> > > > >> -I/home/akalpalp/petsc/arch-mswin-c-debug/include    `pwd`/ex19.c
> > > > >> /home/akalpalp/petsc/arch-mswin-c-debug/bin/mpicc -Wall
> > > -Wwrite-strings
> > > > >> -Wno-strict-aliasing -Wno-unknown-pragmas -g3 -O0  -o ex19  ex19.o
> > > > >> -L/home/akalpalp/petsc/arch-mswin-c-debug/lib  -lpetsc
> > > > >> -Wl,-rpath,/home/akalpalp/petsc/arch-mswin-c-debug/lib -lf2clapack
> > > > >> -lf2cblas -lpthread -lgdi32 -luser32 -ladvapi32 -lkernel32 -ldl
> > > > >>
> > > /home/akalpalp/petsc/arch-mswin-c-debug/lib/libpetsc.a(pcregis.o):pcregis.c:(.rdata$.refptr.PCCreate_BDDC[.refptr.PCCreate_BDDC]+0x0):
> > > > >> undefined reference to `PCCreate_BDDC'
> > > > >> collect2: error: ld returned 1 exit status
> > > > >> makefile:108: recipe for target 'ex19' failed
> > > > >> make[3]: [ex19] Error 1 (ignored)
> > > > >> /usr/bin/rm -f ex19.o
> > > > >> Completed test examples
> > > > >> =========================================
> > > > >> Now to evaluate the computer systems you plan use - do:
> > > > >> make PETSC_DIR=/home/akalpalp/petsc PETSC_ARCH=arch-mswin-c-debug
> > > streams
> > > > >> NPMAX=<number of MPI processes you intend to use>
> > > > >>
> > > > >>
> > > > >>
> > > > >> On Thu, Sep 4, 2014 at 5:37 PM, Barry Smith <bsmith at mcs.anl.gov>
> > > wrote:
> > > > >>
> > > > >>>
> > > > >>>    This is likely due to the horrible horrible fact that some of the
> > > > >>> bddc files only get compiled if ./configure is run with the option
> > > > >>> --with-pcbddc you will need to rerun ./configure and then make with
> > > that
> > > > >>> option.
> > > > >>>
> > > > >>>    I pray that someone removes that horrible confusing configure
> > > option.
> > > > >>>
> > > > >>>    Barry
> > > > >>>
> > > > >>> On Sep 4, 2014, at 8:52 AM, Alp Kalpalp <alpkalpalp at gmail.com>
> > > wrote:
> > > > >>>
> > > > >>> > Dear Stefano,
> > > > >>> >
> > > > >>> > I have checked out "master" branch and merged with your
> > > > >>> stefano_zampini/pcbddc-primalfixe branch. configured, compiled and
> > > all
> > > > >>> tests are completed successfully.
> > > > >>> > Then, I tried to compile ex59 with make ex59, it results in
> > > unresolved
> > > > >>> external error. I believe your bddc_feti files are not included in
> > > > >>> compilation.
> > > > >>> > Since I am not experienced on how to solve issues in Petsc, I need
> > > to
> > > > >>> ask several questions;
> > > > >>> >
> > > > >>> > 1-) Are there any global settings to add additonal directories to
> > > > >>> compilation  (src\ksp\pc\impls\bddc)
> > > > >>> > 2-) or should I include these files on top of ex59 (AFAIK,
> > > including
> > > > >>> .c files is not a good thing)
> > > > >>> > 3-) and finally what is the better way of helping you (creating
> > > > >>> another branch from yours or what)
> > > > >>> >
> > > > >>> > Thanks in advance
> > > > >>> >
> > > > >>> >
> > > > >>> >
> > > > >>> >
> > > > >>> >
> > > > >>> >
> > > > >>> > On Wed, Sep 3, 2014 at 6:18 PM, Stefano Zampini <
> > > > >>> stefano.zampini at gmail.com> wrote:
> > > > >>> > FETIDP is in PETSc as a byproduct of the BDDC preconditioner (one
> > > is
> > > > >>> the dual of the other) and it does not have its own classes so far.
> > > > >>> >
> > > > >>> > That said, you can experiment with FETI-DP only after having setup
> > > a
> > > > >>> BDDC preconditioner with the options and customization you prefer.
> > > > >>> > Use
> > > > >>>
> > > http://www.mcs.anl.gov/petsc/petsc-dev/docs/manualpages/KSP/index.html
> > > > >>> for manual pages.
> > > > >>> >
> > > > >>> > For an 'how to' with FETIDP, please see
> > > > >>> src/ksp/ksp/examples/tutorials/ex.59.c as Jed told you, specifically
> > > look
> > > > >>> at ComputeKSPFETIDP for obtaining the FETIDP matrix F (implemented
> > > > >>> Matrix-free) and the optimal FETIDP dirichlet preconditioner. Once
> > > you have
> > > > >>> F, you can use PCBDDCMatFETIDPGetRHS and PCBDDCMatFetiDPGetSolution
> > > to
> > > > >>> obtain a right-hand side for the FETIDP system and a physical
> > > solution from
> > > > >>> the solution of the FETIDP system.
> > > > >>> >
> > > > >>> > I would recommend you to use the development version of the library
> > > > >>> and either use the ‘next’ branch or the ‘master' branch after having
> > > merged
> > > > >>> in the branch stefano_zampini/pcbddc-primalfixes. Note that ‘next’
> > > also
> > > > >>> contains the new deluxe scaling operator for BDDC which is not
> > > available to
> > > > >>> use with FETI-DP.
> > > > >>> >
> > > > >>> > If you have any other questions which can be useful for other PETSc
> > > > >>> users, please use the mailing list; otherwise you can contact me
> > > personally.
> > > > >>> >
> > > > >>> > Stefano
> > > > >>> >
> > > > >>> >
> > > > >>> > On Sep 3, 2014, at 5:19 PM, Jed Brown <jed at jedbrown.org> wrote:
> > > > >>> >
> > > > >>> >> Matthew Knepley <knepley at gmail.com> writes:
> > > > >>> >>>> 1- Is it possible to complete a FETI-DP solution with the
> > > provided
> > > > >>> >>>> functions in current PetSc release?
> > > > >>> >>>>
> > > > >>> >>>
> > > > >>> >>> There is no FETI-DP in PETSc.
> > > > >>> >>
> > > > >>> >> Wrong.  There is PCBDDC, which has the same eigenvalues as
> > > FETI-DP.
> > > > >>> You
> > > > >>> >> can enable it by configuring --with-pcbddc.  This will be turned
> > > on by
> > > > >>> >> default soon.  It is fairly new, so you should use the branch
> > > 'master'
> > > > >>> >> instead of the release.  It has an option to do FETI-DP instead of
> > > > >>> BDDC.
> > > > >>> >> See src/ksp/ksp/examples/tutorials/ex59.c.
> > > > >>> >>
> > > > >>> >> For either of these methods, you have to assemble a MATIS.  If
> > > you use
> > > > >>> >> MatSetValuesLocal, most of your assembly code can stay the same.
> > > > >>> >>
> > > > >>> >> Hopefully we can get better examples before the next release.
> > > Stefano
> > > > >>> >> (the author of PCBDDC, Cc'd) tests mostly with external packages,
> > > but
> > > > >>> we
> > > > >>> >> really need more complete tests within PETSc.
> > > > >>> >
> > > > >>> >
> > > > >>>
> > > > >>>
> > > > >>
> > > > >
> > > > >
> > > > > --
> > > > > 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
> > > > >
> > > >
> > >
> > 
> 