[MOAB-dev] Error message when compiling the mbcoupler_test.cpp

Vijay S. Mahadevan vijay.m at gmail.com
Fri May 13 15:08:14 CDT 2016


Ok, looks like your MPI installation and MOAB+MPI are configured and
working fine. We will look at the tests and see why they are failing.
I haven't been able to replicate the errors locally but I think we
should be able to find the issue soon.

If "make install" works (you should configure with a
--prefix=$INSTALL_DIR option), you can start using the library now.
I'll update you once we find the root cause for the issues you found.

Vijay

On Fri, May 13, 2016 at 2:12 PM, Jie Wu <jie.voo at gmail.com> wrote:
> Hi there,
>
> Thank you for your directions! Here is the results I get after issuing make
> -k check at directory test/parallel, and I think they look good.
>
> dyn-160-39-10-173:parallel jiewu$ make -k check
> /Applications/Xcode.app/Contents/Developer/usr/bin/make  pcomm_unit scdtest
> pcomm_serial par_spatial_locator_test parallel_unit_tests scdpart
> read_nc_par ucdtrvpart mpastrvpart gcrm_par   parallel_hdf5_test
> mhdf_parallel parallel_write_test uber_parallel_test parallel_adj
> parmerge_test augment_with_ghosts  imeshp_test mbparallelcomm_test partcheck
> structured3 parmerge
>   CXX      pcomm_unit.o
>   CXXLD    pcomm_unit
>   CXX      scdtest.o
>   CXXLD    scdtest
>   CXX      pcomm_serial.o
>   CXXLD    pcomm_serial
>   CXX      par_spatial_locator_test.o
>   CXXLD    par_spatial_locator_test
>   CXX      parallel_unit_tests.o
>   CXXLD    parallel_unit_tests
>   CXX      scdpart.o
>   CXXLD    scdpart
>   CXX      ../io/read_nc.o
>   CXXLD    read_nc_par
>   CXX      ucdtrvpart.o
>   CXXLD    ucdtrvpart
>   CXX      mpastrvpart.o
>   CXXLD    mpastrvpart
>   CXX      gcrm_par.o
>   CXXLD    gcrm_par
>   CXX      parallel_hdf5_test.o
>   CXXLD    parallel_hdf5_test
>   CC       mhdf_parallel.o
>   CCLD     mhdf_parallel
>   CXX      parallel_write_test.o
>   CXXLD    parallel_write_test
>   CXX      uber_parallel_test.o
>   CXXLD    uber_parallel_test
>   CXX      ../adj_moab_test.o
>   CXXLD    parallel_adj
>   CXX      parmerge_test.o
>   CXXLD    parmerge_test
>   CXX      augment_with_ghosts.o
>   CXXLD    augment_with_ghosts
>   PPFC     imeshp_test.o
>   FCLD     imeshp_test
>   CXX      mbparallelcomm_test.o
>   CXXLD    mbparallelcomm_test
>   CXX      partcheck.o
>   CXXLD    partcheck
>   CXX      structured3.o
>   CXXLD    structured3
>   CXX      parmerge.o
>   CXXLD    parmerge
> /Applications/Xcode.app/Contents/Developer/usr/bin/make  check-TESTS
> PASS: pcomm_unit
> PASS: scdtest
> PASS: pcomm_serial
> PASS: par_spatial_locator_test
> PASS: parallel_unit_tests
> PASS: scdpart
> PASS: read_nc_par
> PASS: ucdtrvpart
> PASS: mpastrvpart
> PASS: gcrm_par
> PASS: parallel_hdf5_test
> PASS: mhdf_parallel
> PASS: parallel_write_test
> PASS: uber_parallel_test
> PASS: parallel_adj
> PASS: parmerge_test
> PASS: augment_with_ghosts
> PASS: imeshp_test
> ============================================================================
> Testsuite summary for MOAB 4.9.2pre
> ============================================================================
> # TOTAL: 18
> # PASS:  18
> # SKIP:  0
> # XFAIL: 0
> # FAIL:  0
> # XPASS: 0
> # ERROR: 0
> ============================================================================
>
> Following is an example I run with mpiexec, and I think it also goes well.
>
> Thank you very much for your help! Any help deal with the failing tests will
> be much appreciated.
>
> Best,
> Jie
>
> dyn-160-39-10-173:mbcoupler jiewu$ mpiexec -np 2  mbcoupler_test -meshes
> /Users/wujie/SourceCode/moab/MeshFiles/unittest/64bricks_1khex.h5m
> /Users/wujie/SourceCode/moab/MeshFiles/unittest/64bricks_12ktet.h5m -itag
> vertex_field -outfile dum.h5m
> ReadHDF5:             0.0324039
>   get set meta        0.00103188
>   partial subsets     0.000803947
>   partition time      2.14577e-06
>   get set ids         0.000556946
>   set contents        0.00107098
>   polyhedra           2.5034e-05
>   elements            0.00174189
>   nodes               0.0014832
>   node adjacency      0
>   side elements       0.000672817
>   update connectivity 0.000102043
>   adjacency           0
>   delete non_adj      0.00338697
>   recursive sets      0
>   find contain_sets   0.000369072
>   read sets           0.00147104
>   read tags           0.01667
>   store file ids      0.00191593
>   read qa records     0.000133038
> Parallel Read times:
>   0.0473361 PARALLEL READ PART
>   0.0138111 PARALLEL RESOLVE_SHARED_ENTS
>   0.0113361 PARALLEL EXCHANGE_GHOSTS
>   0.00113177 PARALLEL RESOLVE_SHARED_SETS
>   0.00648403 PARALLEL_AUGMENT_SETS_WITH_GHOSTS
>   0.079994 PARALLEL TOTAL
> ReadHDF5:             0.0714378
>   get set meta        0.000261784
>   partial subsets     0.000660181
>   partition time      9.53674e-07
>   get set ids         0.00101304
>   set contents        0.000728846
>   polyhedra           0.000128031
>   elements            0.0155201
>   nodes               0.00354886
>   node adjacency      6.19888e-06
>   side elements       0.00422001
>   update connectivity 0.00184798
>   adjacency           0
>   delete non_adj      0.0100739
>   recursive sets      0
>   find contain_sets   0.00218511
>   read sets           0.00258899
>   read tags           0.0235598
>   store file ids      0.0044961
>   read qa records     9.5129e-05
> Parallel Read times:
>   0.0779119 PARALLEL READ PART
>   0.0667319 PARALLEL RESOLVE_SHARED_ENTS
>   0.0397439 PARALLEL EXCHANGE_GHOSTS
>   0.00435901 PARALLEL RESOLVE_SHARED_SETS
>   0.0243261 PARALLEL_AUGMENT_SETS_WITH_GHOSTS
>   0.209468 PARALLEL TOTAL
> Proc 1 iface entities:
>     446 0d iface entities.
>     1080 1d iface entities.
>     Proc 0636 2d iface entities.
>     0 3d iface entities.
>   iface entities:
>     446 0d iface entities.
>     108   (446 verts adj to other iface ents)
> 0 1d iface entities.
>     636 2d iface entities.
>     0 3d iface entities.
>     (446 verts adj to other iface ents)
> rank 0 points of interest: 6155
> rank 1 points of interest: 6155
>
> Max time : 0.101124 0.0461781 0.00104499 (inst loc interp -- 2 procs)
> WriteHDF5:           0.0779669
>   gather mesh:       0.00237203
>   create file:       0.0655501
>     create nodes:    0.00201201
>     negotiate types: 2.19345e-05
>     create elem:     0.000785828
>     file id exch:    0.0234561
>     create adj:      0.000414848
>     create set:      0.019654
>       shared ids:    6.19888e-05
>       shared data:   0.014715
>       set offsets:   0.00447798
>     create tags:     0.0160849
>   coordinates:       0.00199103
>   connectivity:      0.00162005
>   sets:              0.0010829
>     set descrip:     0.00019598
>     set content:     0.000527859
>     set parent:      0.000226021
>     set child:       0.000128984
>   adjacencies:       5.96046e-06
>   tags:              0.00700402
>     dense data:      0.00127792
>     sparse data:     0.00569224
>     var-len data:    0
> Wrote dum.h5m
> mbcoupler_test complete.
>
>
>
> On May 13, 2016, at 2:58 PM, Grindeanu, Iulian R. <iulian at mcs.anl.gov>
> wrote:
>
> yes, can you try make -k check  ?
> It will not stop at the first failed test, it will try to test everything
> ________________________________________
> From: Vijay S. Mahadevan [vijay.m at gmail.com]
> Sent: Friday, May 13, 2016 1:53 PM
> To: Jie Wu
> Cc: Grindeanu, Iulian R.; moab-dev at mcs.anl.gov
> Subject: Re: [MOAB-dev] Error message when compiling the mbcoupler_test.cpp
>
> Jie,
>
> We are looking into the failing tests and will update in case there is
> a bug in the code. Our nightly tests are running cleanly and so I'm
> curious to see if other tests are failing. Can you go to test/parallel
> and run "make check" there to verify if the parallel job launch is
> working correctly ? Also, can you verify if you can run MPI programs
> with mpiexec ?
>
> Vijay
>
>
>
> Vijay
>
> On Fri, May 13, 2016 at 1:31 PM, Jie Wu <jie.voo at gmail.com> wrote:
>
> Hi Vijay,
>
> Deleting the auto-installed packages under sandbox before issue new
> configurations works good! No error message emerged until I input the
> command make check, which is as following.
>
> Do you have any idea for this problem?
>
> Best,
> Jie
>
> PASS: reorder_test
> FAIL: test_prog_opt
> PASS: coords_connect_iterate
> PASS: elem_eval_test
> PASS: spatial_locator_test
> PASS: test_boundbox
> PASS: adj_moab_test
> PASS: uref_mesh_test
> PASS: verdict_test
> PASS: test_Matrix3
> PASS: mbfacet_test
> PASS: gttool_test
> PASS: cropvol_test
> FAIL: mergemesh_test
> PASS: mbground_test
> PASS: lloyd_smoother_test
> ============================================================================
> Testsuite summary for MOAB 4.9.2pre
> ============================================================================
> # TOTAL: 36
> # PASS:  34
> # SKIP:  0
> # XFAIL: 0
> # FAIL:  2
> # XPASS: 0
> # ERROR: 0
> ============================================================================
> See test/test-suite.log
> Please report to moab-dev at mcs.anl.gov
> ============================================================================
> make[4]: *** [test-suite.log] Error 1
> make[3]: *** [check-TESTS] Error 2
> make[2]: *** [check-am] Error 2
> make[1]: *** [check-recursive] Error 1
> make: *** [check-recursive] Error 1
>
>
>
>
> On May 13, 2016, at 1:08 PM, Vijay S. Mahadevan <vijay.m at gmail.com> wrote:
>
> Jie, since you now changed the compilers used to configure the stack,
> I would advise deleting the auto-installed packages under sandbox
> directory and then re-configuring. This will force the rebuild of
> HDF5/NetCDF packages with the parallel MPI compilers and you should
> get the HDF5-parallel support enabled after that. Makes sense ?
>
> Vijay
>
> On Fri, May 13, 2016 at 11:27 AM, Jie Wu <jie.voo at gmail.com> wrote:
>
> Hi Vijay,
>
> Thank you for your reply. The new configuration do works. Here shows the
> command I inputed and the message while I input: make check.
>
> dyn-160-39-10-173:moab jiewu$ ./configure --with-mpi=/usr/local
> CC=/usr/local/bin/mpicc CXX=/usr/local/bin/mpic++ FC=/usr/local/bin/mpif90
> F77=/usr/local/bin/mpif77 --download-metis --download-hdf5 --download-netcdf
> --enable-docs --with-doxygen=/Applications/Doxygen.app/Contents/Resources/
>
> The results are
>
> configure: WARNING:
> *************************************************************************
> *        MOAB has been configured with parallel and HDF5 support
> *     but the configured HDF5 library does not support parallel IO.
> *            Some parallel IO capabilities will be disabled.
> *************************************************************************
>
> After that, I inputted the command, make -j4, and then command make check.
> Here are the results it shows
>
> /Applications/Xcode.app/Contents/Developer/usr/bin/make  check-TESTS
> PASS: elem_util_test
> PASS: element_test
> FAIL: datacoupler_test
> FAIL: mbcoupler_test
> FAIL: parcoupler_test
> ============================================================================
> Testsuite summary for MOAB 4.9.2pre
> ============================================================================
> # TOTAL: 5
> # PASS:  2
> # SKIP:  0
> # XFAIL: 0
> # FAIL:  3
> # XPASS: 0
> # ERROR: 0
> ============================================================================
> See tools/mbcoupler/test-suite.log
> Please report to moab-dev at mcs.anl.gov
> ============================================================================
> make[4]: *** [test-suite.log] Error 1
> make[3]: *** [check-TESTS] Error 2
> make[2]: *** [check-am] Error 2
> make[1]: *** [check-recursive] Error 1
> make: *** [check-recursive] Error 1
>
> Then I was trying to run the mbcouple_test with command
> dyn-160-39-10-173:mbcoupler jiewu$    mpiexec -np 2  mbcoupler_test -meshes
> /Users/wujie/SourceCode/moab/MeshFiles/unittest/64bricks_1khex.h5m
> /Users/wujie/SourceCode/moab/MeshFiles/unittest/64bricks_12ktet.h5m -itag
> vertex_field -outfile dum.h5m
> Now the results look like,
>
> [0]MOAB ERROR: --------------------- Error Message
> ------------------------------------
> [0]MOAB ERROR: MOAB not configured with parallel HDF5 support!
> [0]MOAB ERROR: set_up_read() line 355 in src/io/ReadHDF5.cpp
> [0]MOAB ERROR: --------------------- Error Message
> ------------------------------------
> [0]MOAB ERROR: NULL file handle.!
> [0]MOAB ERROR: is_error() line 133 in src/io/ReadHDF5.hpp
> [1]MOAB ERROR: --------------------- Error Message
> ------------------------------------
> [1]MOAB ERROR: MOAB not configured with parallel HDF5 support!
> [1]MOAB ERROR: set_up_read() line 355 in src/io/ReadHDF5.cpp
> [1]MOAB ERROR: --------------------- Error Message
> ------------------------------------
> [1]MOAB ERROR: NULL file handle.!
> [1]MOAB ERROR: is_error() line 133 in src/io/ReadHDF5.hpp
> [0]MOAB ERROR: --------------------- Error Message
> ------------------------------------
> [0]MOAB ERROR: Failed to load file after trying all possible readers!
> [0]MOAB ERROR: serial_load_file() line 635 in src/Core.cpp
> [0]MOAB ERROR: --------------------- Error Message
> ------------------------------------
> [0]MOAB ERROR: Failed in step PARALLEL READ PART!
> [0]MOAB ERROR: load_file() line 553 in src/parallel/ReadParallel.cpp
> [0]MOAB ERROR: load_file() line 252 in src/parallel/ReadParallel.cpp
> [0]MOAB ERROR: load_file() line 515 in src/Core.cpp
> [1]MOAB ERROR: --------------------- Error Message
> ------------------------------------
> [1]MOAB ERROR: Failed to load file after trying all possible readers!
> [1]MOAB ERROR: serial_load_file() line 635 in src/Core.cpp
> [1]MOAB ERROR: --------------------- Error Message
> ------------------------------------
> [1]MOAB ERROR: Failed in step PARALLEL READ PART!
> [1]MOAB ERROR: load_file() line 553 in src/parallel/ReadParallel.cpp
> [0]MOAB ERROR: main() line 172 in mbcoupler_test.cpp
> [1]MOAB ERROR: load_file() line 252 in src/parallel/ReadParallel.cpp
> [1]MOAB ERROR: load_file() line 515 in src/Core.cpp
> [1]MOAB ERROR: main() line 172 in mbcoupler_test.cpp
> --------------------------------------------------------------------------
> MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD
> with errorcode 16.
>
> NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
> You may or may not see output from other processes, depending on
> exactly when Open MPI kills them.
> --------------------------------------------------------------------------
> [dyn-160-39-11-172.dyn.columbia.edu:26361] 1 more process has sent help
> message help-mpi-api.txt / mpi-abort
> [dyn-160-39-11-172.dyn.columbia.edu:26361] Set MCA parameter
> "orte_base_help_aggregate" to 0 to see all help / error messages
> The other excitable file in mbcoupler folder, like ./element_test now can
> run successfully.
>
> Any direction for this was deeply appreciated!
>
> Best,
> Jie
>
>
> On May 13, 2016, at 11:54 AM, Vijay S. Mahadevan <vijay.m at gmail.com> wrote:
>
> I also forgot to mention that you can be more explicit in the
> specification of compilers during configuration, if you really want.
> This would avoid headaches since the user specified wrappers will
> directly be used. For example:
>
> ./configure --with-mpi=/usr/local CC=/usr/local/bin/mpicc
> CXX=/usr/local/bin/mpic++ FC=/usr/local/bin/mpif90
> F77=/usr/local/bin/mpif77 <OTHER_CONFIGURE_OPTIONS>
>
> Vijay
>
> On Fri, May 13, 2016 at 10:52 AM, Vijay S. Mahadevan <vijay.m at gmail.com>
> wrote:
>
> Please use --with-mpi=/usr/local as the configure option. If you have
> MPI pre-installed, either through PETSc or natively in your system,
> always try to re-use that to maintain consistency. This will help
> avoid mix-up of MPI implementations when you try to launch a parallel
> job later.
>
> Let us know if the new configuration works.
>
> Vijay
>
> On Thu, May 12, 2016 at 6:32 PM, Jie Wu <jie.voo at gmail.com> wrote:
>
> Thank you very much for your reply. I search in my src/moab/MOABConfig.h
> file, and I did not find the MOAB_HAVE_MPI. So I think the MOAB on my laptop
> cannot run in parallel yet.
>
> I have this from my terminal.
>
> dyn-160-39-10-173:moab jiewu$ which mpic++
> /usr/local/bin/mpic++
>
> But I cannot get anything by inputing which mpi. So I don’t know where the
> mpi is in my laptop.
>
> dyn-160-39-10-173:moab jiewu$ which mpi
>
> I have petsc installed which has mpich, and linking that directory to the
> MOAB configure option also did not work.
>
> Should I install mpi solely? I am afraid it might conflict with the one
> installed in petsc.
>
> Is there any method that the mpi could be installed with MOAB automatically,
> or maybe other proper manner? Thanks a lot!
>
> Best,
> Jie
>
>
> On May 12, 2016, at 6:11 PM, Vijay S. Mahadevan <vijay.m at gmail.com> wrote:
>
> Can you send the config.log. Looks like MPI is not getting enabled
> (possibly), even with --dowload-mpich. Or you can check
> src/moab/MOABConfig.h in your build directory and grep for
> MOAB_HAVE_MPI.
>
> Vijay
>
> On Thu, May 12, 2016 at 4:46 PM, Jie Wu <jie.voo at gmail.com> wrote:
>
> Hi Iulian,
>
> Thanks for your reply. I think my configure is with mpi. Here is my
> configure command
>
> dyn-160-39-10-173:moab jiewu$    ./configure --download-metis
> --download-hdf5 --download-netcdf --download-mpich --enable-docs
> --with-doxygen=/Applications/Doxygen.app/Contents/Resources/
>
> Best,
> Jie
>
> On May 12, 2016, at 5:43 PM, Grindeanu, Iulian R. <iulian at mcs.anl.gov>
> wrote:
>
> Hi Jie,
> Did you configure with mpi? What is your configure command?
>
> Iulian
> ________________________________
> From: moab-dev-bounces at mcs.anl.gov [moab-dev-bounces at mcs.anl.gov] on behalf
> of Jie Wu [jie.voo at gmail.com]
> Sent: Thursday, May 12, 2016 4:23 PM
> To: moab-dev at mcs.anl.gov
> Subject: [MOAB-dev] Error message when compiling the mbcoupler_test.cpp
>
> Hi all,
>
> My name is Jie and I am part of computational mechanics group at civil
> engineering dept. of Columbia university.
>
> I am working on large deformation problems which may lead mesh distortions
> and a re-meshing become necessary.
>
> I would like to compile mbcoupler_test.cpp to learn how it transfer the
> variables from old to the new mesh.
>
> Now I can successfully compile the codes in build/examples and they works
> good! But I cannot compile the codes in folder build/tools/mbcoupler by
> following the instructions in build/tools/readme.tools, which shows error
> message like following.
>
> Do you have any idea for this problem? Thanks a lot!
>
> Best,
> Jie
>
> DataCoupler.cpp:136:25: error: member access into incomplete
>   type 'moab::ParallelComm'
> if (myPcomm && myPcomm->size() > 1) {
>                     ^
> ./DataCoupler.hpp:34:7: note: forward declaration of
>   'moab::ParallelComm'
> class ParallelComm;
>   ^
> DataCoupler.cpp:161:12: error: member access into incomplete
>   type 'moab::ParallelComm'
> myPcomm->proc_config().crystal_router()->gs_tran...
>        ^
> ./DataCoupler.hpp:34:7: note: forward declaration of
>   'moab::ParallelComm'
> class ParallelComm;
>   ^
> DataCoupler.cpp:187:12: error: member access into incomplete
>   type 'moab::ParallelComm'
> myPcomm->proc_config().crystal_router()->gs_tran...
>        ^
> ./DataCoupler.hpp:34:7: note: forward declaration of
>   'moab::ParallelComm'
> class ParallelComm;
>   ^
> 3 errors generated.
> make[2]: *** [DataCoupler.lo] Error 1
> make[2]: *** Waiting for unfinished jobs....
> Coupler.cpp:344:45: error: no member named 'gs_transfer' in
>   'moab::gs_data::crystal_data'
> (myPc->proc_config().crystal_router())->gs_trans...
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  ^
> Coupler.cpp:388:45: error: no member named 'gs_transfer' in
>   'moab::gs_data::crystal_data'
> (myPc->proc_config().crystal_router())->gs_trans...
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  ^
> Coupler.cpp:611:45: error: no member named 'gs_transfer' in
>   'moab::gs_data::crystal_data'
> (myPc->proc_config().crystal_router())->gs_trans...
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  ^
> Coupler.cpp:638:43: error: no member named 'gs_transfer' in
>   'moab::gs_data::crystal_data'
> myPc->proc_config().crystal_router()->gs_transfe...
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  ^
> 4 errors generated.
> make[2]: *** [Coupler.lo] Error 1
> make[1]: *** [all-recursive] Error 1
> make: *** [all-recursive] Error 1
>
>
>
>
>
>
>
>
>


More information about the moab-dev mailing list