[petsc-users] make test freeze

[Matthew Knepley]

On Fri, Jan 6, 2017 at 2:39 AM, Patrick Begou <
Patrick.Begou at legi.grenoble-inp.fr> wrote:

> Hi Matthew,
>
> Launching manualy ex19 shows only one process consuming cpu time, after
> 952mn I've killed the job this morning.
>
> [begou at kareline tutorials]$ make ex19
> mpicc -o ex19.o -c -Wall -Wwrite-strings -Wno-strict-aliasing
> -Wno-unknown-pragmas -fvisibility=hidden -g3
> -I/kareline/data/begou/YALES2_1.0.0/PREREQUIS/petsc-git/include
> -I/kareline/data/begou/YALES2_1.0.0/PREREQUIS/petsc-git/GCC48/include
> `pwd`/ex19.c
> mpicc -Wall -Wwrite-strings -Wno-strict-aliasing -Wno-unknown-pragmas
> -fvisibility=hidden -g3  -o ex19  ex19.o -L/kareline/data/begou/YALES2_
> 1.0.0/PREREQUIS/petsc-git/GCC48/lib  -lpetsc -llapack -lblas -lX11
> -lhwloc -lssl -lcrypto -L/opt/openmpi173-GCC48-node/lib
> -L/opt/GCC48c/lib/gcc/x86_64-unknown-linux-gnu/4.8.1 -L/opt/GCC48c/lib64
> -L/opt/GCC48c/lib -lmpi_usempi -lmpi_mpifh -lgfortran -lm -lgfortran -lm
> -lquadmath -lm -lmpi_cxx -lstdc++ -L/opt/openmpi173-GCC48-node/lib
> -L/opt/GCC48c/lib/gcc/x86_64-unknown-linux-gnu/4.8.1 -L/opt/GCC48c/lib64
> -L/opt/GCC48c/lib -ldl -lmpi -lgcc_s -lpthread -ldl
> /bin/rm -f ex19.o
> [begou at kareline tutorials]$ mpiexec -n 2 ./ex19 -snes_monitor
>
> top command shows:
>
>   PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+
> COMMAND
> 32184 begou     20   0  249m 7152 5132 R 99.8  0.0 952:15.97
> ex19
> 32183 begou     20   0 71676 3508 2264 S  0.0  0.0   0:00.04
> mpiexec
> 32185 begou     20   0  185m 7132 5124 S  0.0  0.0   0:00.04
> ex19
>
> looks like the first process waiting for something that never occur in MPI
> communication....
>

1000s of people run this every day, so I am skeptical of that explanation.

However, this could happen if the 'mpiexec' in your path does not match the
MPI libraries that PETSc is linked to.

   Matt


> Patrick
>
> Matthew Knepley a écrit :
>
> On Thu, Jan 5, 2017 at 6:31 AM, Patrick Begou <
> <Patrick.Begou at legi.grenoble-inp.fr>Patrick.Begou at legi.grenoble-inp.fr>
> wrote:
>
>> I am unable to run any test on petsc. It looks like if the ex19 run
>> freeze on the server as it do not use any cpu time and pstree shows
>>
>> sshd---bash-+-gedit
>>             `-make---sh-+-gmake---sh---gmake---sh---mpiexec---ex19
>>                         `-tee
>> I've tested petsc-3.7.5.tar.gz and  the latest sources on the Git
>> repository.
>>
>
> All make is doing is running ex19, which you can do by hand. What do you
> get for
>
>   cd $PETSC_DIR
>   cd src/snes/examples/tutorials
>   make ex19
>   mpiexec -n 2 ./ex19 -snes_monitor
>
>   Thanks,
>
>     Matt
>
>
>> Setup from the Git repo:
>> ./configure --prefix=/kareline/data/begou/YALES2_1.0.0/PREREQUIS/petsc-git-binaries
>> \
>> --PETSC_ARCH=GCC48 \
>> --PETSC_DIR=/kareline/data/begou/YALES2_1.0.0/PREREQUIS/petsc-git \
>> --with-shared-libraries=0 \
>> --with-fortran-interfaces=1 \
>> --with-fortran-kernels=1 \
>> --with-cc=mpicc \
>> --with-fc=mpif90 \
>> --with-cxx=mpicxx
>>
>>  make PETSC_DIR=/kareline/data/begou/YALES2_1.0.0/PREREQUIS/petsc-git
>> PETSC_ARCH=GCC48 all
>>
>>  make PETSC_DIR=/kareline/data/begou/YALES2_1.0.0/PREREQUIS/petsc-git
>> PETSC_ARCH=GCC48 install
>>
>>  make PETSC_DIR=/kareline/data/begou/YALES2_1.0.0/PREREQUIS/petsc-git-binaries
>> PETSC_ARCH="" test
>>
>>
>> In the log file I've just:
>>
>> Running test examples to verify correct installation
>> Using PETSC_DIR=/kareline/data/begou/YALES2_1.0.0/PREREQUIS/petsc-git-binaries
>> and PETSC_ARCH=
>>
>> I'm using:
>> gcc version 4.8.1
>> Open MPI: 1.7.3 (build with gcc 4.8.1)
>> (This environment is in production for a while for many local software
>> and works fine)
>>
>> Any suggestion is welcome
>>
>> Patrick
>>
>> --
>> ===================================================================
>> |  Equipe M.O.S.T.         |                                      |
>> |  Patrick BEGOU           | mailto:Patrick.Begou at grenoble-inp.fr |
>> |  LEGI                    |                                      |
>> |  BP 53 X                 | Tel 04 76 82 51 35                   |
>> |  38041 GRENOBLE CEDEX    | Fax 04 76 82 52 71                   |
>> ===================================================================
>>
>>
>
>
> --
> 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
>
>
>
> --
> ===================================================================
> |  Equipe M.O.S.T.         |                                      |
> |  Patrick BEGOU           | mailto:Patrick.Begou at grenoble-inp.fr <Patrick.Begou at grenoble-inp.fr> |
> |  LEGI                    |                                      |
> |  BP 53 X                 | Tel 04 76 82 51 35                   |
> |  38041 GRENOBLE CEDEX    | Fax 04 76 82 52 71                   |
> ===================================================================
>
>


-- 
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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20170106/ce88b92f/attachment.html>