[aspect-devel] Problem compiling Aspect on Tacc-Stampede2
FELIPE ORELLANA ROVIROSA
f_orellana at berkeley.edu
Wed May 2 23:46:56 PDT 2018
Hi John,
Thanks a lot for your support.
I have followed every instruction I had access to, or that's what I think.
Following Timo's advice, it seems one of the error message's missing
files is actually missing.
I guess that if you successfully installed Aspect on Stampede2, there
must be something I am missing, something I am doing wrongly.
cheers,
Felipe
On Wed, May 2, 2018 at 2:10 PM, John Naliboff <jbnaliboff at ucdavis.edu>
wrote:
> Hi all,
>
> Just a quick bit of clarification - my recommendation is to build on a
> "development" SKX node, using the "idev" command.
>
> The SKX nodes give the best scaling performance and compiling on an SKX
> development nodes means you are guaranteed to have compiled under the right
> architecture (e.g., if you submit a job to an SKX node). You also of course
> compile much faster (access to 48 cores) and I believe that either the
> deal.II or Trilinos builds on the head nodes may have run into memory
> issues.
>
> The command to run an interactive development session on an SKX
> development node is:
> idev -p skx-dev -N 1 -n 48 -m 90
> where skx-dev is the partition, N is the number of nodes, n is the number
> tasks and m is the number of minutes.
>
> More information here: https://portal.tacc.utexas.
> edu/user-guides/stampede2#running-idev
>
> Felipe - You are following the instructions I sent over, yes? I'm not sure
> what's going wrong, but if you can wait until next Monday I'll have a
> chance to try recompiling on see if anything has changed since I last
> compiled (1-2 months).
>
> Cheers,
> John
>
> On 05/02/2018 12:40 PM, FELIPE ORELLANA ROVIROSA wrote:
>
>
> Hi Timo,
>
> Gotta wait:
>
> The compute nodes on Stampede2 are down at the moment. I am
> asking them about this issue.
>
> As far as I understand it is highly recommended to be hanging
> from compute nodes while carrying out compilation.
>
> thanks,
> Felipe
>
> On Wed, May 2, 2018 at 12:29 PM, FELIPE ORELLANA ROVIROSA <
> f_orellana at berkeley.edu> wrote:
>
>>
>> Hallo Timo,
>>
>> Thanks a lot for replying.
>>
>> I am going step by step.
>>
>> I deleted the build folder, to rebuild everything again.
>>
>> I created a new build directory, went into it, and typed:
>>
>> cmake -D DEAL_II_DIR=$TACC_DEALII_DIR ..
>>
>> it seems to have worked:
>>
>> login2.stampede2(303)$ cmake -D DEAL_II_DIR=$TACC_DEALII_DIR ..
>> -- ====================================================
>> -- ============ Configuring ASPECT ====================
>> -- ====================================================
>> -- Found deal.II version 9.0.0-pre at '/home1/apps/intel17/impi17_0/
>> dealii/git20170615/lib/cmake/deal.II'
>> -- Using ASPECT_USE_PETSC = 'OFF'
>> -- The C compiler identification is Intel 17.0.4.20170411
>> -- The CXX compiler identification is Intel 17.0.4.20170411
>> -- Check for working C compiler: /opt/intel/compilers_and_libra
>> ries_2017.4.196/linux/mpi/intel64/bin/mpicc
>> -- Check for working C compiler: /opt/intel/compilers_and_libra
>> ries_2017.4.196/linux/mpi/intel64/bin/mpicc -- works
>> -- Detecting C compiler ABI info
>> -- Detecting C compiler ABI info - done
>> -- Detecting C compile features
>> -- Detecting C compile features - done
>> -- Check for working CXX compiler: /opt/intel/compilers_and_libra
>> ries_2017.4.196/linux/mpi/intel64/bin/mpicxx
>> -- Check for working CXX compiler: /opt/intel/compilers_and_libra
>> ries_2017.4.196/linux/mpi/intel64/bin/mpicxx -- works
>> -- Detecting CXX compiler ABI info
>> -- Detecting CXX compiler ABI info - done
>> -- Detecting CXX compile features
>> -- Detecting CXX compile features - done
>> -- Found Perl: /bin/perl (found version "5.16.3")
>> -- Added 1 tests.
>> -- Performing Test HAVE_DLOPEN
>> -- Performing Test HAVE_DLOPEN - Success
>> -- Enabling dynamic loading of plugins from the input file
>> -- Looking for include file link.h
>> -- Looking for include file link.h - found
>> -- Enabling checking of compatible deal.II library when loading plugins
>> -- Autopilot invoked
>> ###
>> #
>> # Project aspect set up with deal.II-9.0.0-pre found at
>> # /home1/apps/intel17/impi17_0/dealii/git20170615
>> #
>> # CMAKE_BUILD_TYPE: Debug
>> #
>> # You can now run
>> # $ make - to compile and link the program
>> # $ make run - to (compile, link and) run the program
>> #
>> # $ make debug - to switch the build type to 'Debug'
>> # $ make release - to switch the build type to 'Release'
>> #
>> # $ make edit_cache - to change (cached) configuration variables
>> # and rerun the configure and generate
>> phases of CMake
>> #
>> # $ make strip_comments - to strip the source files in this
>> # directory off the documentation comments
>> # $ make clean - to remove the generated executable as
>> well as
>> # all intermediate compilation files
>> # $ make runclean - to remove all output generated by the
>> program
>> # $ make distclean - to clean the directory from _all_
>> generated
>> # files (includes clean, runclean and the
>> removal
>> # of the generated build system)
>> # $ make info - to view this message again
>> #
>> # Have a nice day!
>> #
>> ###
>> -- Performing Test HAVE_FP_EXCEPTIONS
>> -- Performing Test HAVE_FP_EXCEPTIONS - Failed
>> -- No support for feenableexcept(), disabling runtime floating point
>> exception checks.
>> -- Writing config into detailed.log...
>> -- Configuring done
>> -- Generating done
>> -- Build files have been written to: /home1/04020/unfelipe/aspect-1
>> .5/build
>> login2.stampede2(304)$
>>
>> I will proceed with compiling, the optimized version (make release)
>>
>> Felipe
>>
>>
>> On Wed, May 2, 2018 at 12:09 PM, Timo Heister <heister at clemson.edu>
>> wrote:
>>
>>> How are you configuring ASPECT? It should be something like
>>>
>>> cmake -D DEAL_II_DIR=$TACC_DEALII_DIR /your/path/to/your/aspect/sour
>>> cedir
>>>
>>> Can you post your detailed.log from the ASPECT build directory?
>>>
>>> I am also not sure who built "11) dealii/git20170615" and whether it
>>> will work for you.
>>>
>>> Lastly,
>>>
>>> > /home1/apps/gcc7_1/impi17_0/dealii/git20170615/include/deal.II/base/utilities.h(28):
>>> catastrophic error: cannot open source file "Epetra_Comm.h"
>>>
>>> This looks like some include paths are missing?!
>>>
>>>
>>>
>>>
>>> On Wed, May 2, 2018 at 1:49 PM, FELIPE ORELLANA ROVIROSA
>>> <f_orellana at berkeley.edu> wrote:
>>> >
>>> > Hi all,
>>> >
>>> > I am trying to install ASPECT on TACC Stampede2.
>>> >
>>> > I have modified my bashrc file so I can use the modules that are
>>> already
>>> > there installed, so I check the following:
>>> >
>>> > login2.stampede2(282)$ module list
>>> >
>>> > Currently Loaded Modules:
>>> > 1) autotools/1.1 4) git/2.9.0 7) impi/17.0.3 10)
>>> p4est/2.0
>>> > 2) xalt/1.7.7 5) cmake/3.10.2 8) python/2.7.13 11)
>>> > dealii/git20170615
>>> > 3) TACC 6) intel/17.0.4 9) trilinos/12.10.1
>>> >
>>> > Inactive Modules:
>>> > 1) mkl
>>> >
>>> >
>>> > Then, I have tried installing two versions of ASPECT (2.0.0-pre and
>>> 1.5)
>>> > with no success.
>>> > The generating succeeds, it produces the dependencies, etc.
>>> >
>>> > But, the problem arises at the moment of compilation. It seemingly
>>> fails
>>> > when trying to use dealii or trilinos libraries. I am pasting a
>>> fraction of
>>> > the error message at the end.
>>> >
>>> > A guy on Stampede tech support wrote me these messages:
>>> > #######################
>>> > ''
>>> > When you build ASPECT you have to tell it where trilinos is. I have no
>>> > experience with aspect so I don't know whether this is done with a
>>> cmake
>>> > option, a configure option, or by editing a make include file.
>>> >
>>> > You need to tell cmake that
>>> >
>>> > $TACC_DEALII_DIR/include
>>> >
>>> > is a valid include path
>>> > ''########################
>>> >
>>> > So, where and how and at what stage should I indicate ASPECT that I
>>> am
>>> > using the local libraries?
>>> >
>>> > warm regards,
>>> > Felipe
>>> >
>>> > error message when i compile Aspect 1.5:
>>> >
>>> > make[2]: *** [CMakeFiles/aspect.dir/source/geometry_model/chunk.cc.o]
>>> Error
>>> > 4
>>> > In file included from
>>> > /home1/04020/unfelipe/aspect-1.5/include/aspect/simulator_ac
>>> cess.h(32),
>>> > from
>>> > /home1/04020/unfelipe/aspect-1.5/include/aspect/boundary_tem
>>> perature/initial_temperature.h(26),
>>> > from
>>> > /home1/04020/unfelipe/aspect-1.5/source/boundary_temperature
>>> /initial_temperature.cc(22):
>>> > /home1/04020/unfelipe/aspect-1.5/include/aspect/global.h(30):
>>> catastrophic
>>> > error: cannot open source file "deal.II/lac/trilinos_block_vector.h"
>>> > # include <deal.II/lac/trilinos_block_vector.h>
>>> >
>>> >
>>> > error message when i compile Aspect 2.0.0-pre:
>>> >
>>> > compilation aborted for
>>> > /home1/04020/unfelipe/aspect/source/boundary_velocity/interface.cc
>>> (code 4)
>>> > make[2]: ***
>>> > [CMakeFiles/aspect.dir/source/boundary_traction/initial_lith
>>> ostatic_pressure.cc.o]
>>> > Error 4
>>> > In file included from
>>> > /home1/apps/gcc7_1/impi17_0/dealii/git20170615/include/deal.
>>> II/base/index_set.h(20),
>>> > from
>>> > /home1/apps/gcc7_1/impi17_0/dealii/git20170615/include/deal.
>>> II/lac/trilinos_vector.h(24),
>>> > from
>>> > /home1/apps/gcc7_1/impi17_0/dealii/git20170615/include/deal.
>>> II/lac/trilinos_parallel_block_vector.h(24),
>>> > from
>>> > /home1/04020/unfelipe/aspect/include/aspect/global.h(34),
>>> > from
>>> > /home1/04020/unfelipe/aspect/include/aspect/plugins.h(25),
>>> > from
>>> > /home1/04020/unfelipe/aspect/include/aspect/boundary_tractio
>>> n/interface.h(25),
>>> > from
>>> > /home1/04020/unfelipe/aspect/include/aspect/boundary_tractio
>>> n/function.h(25),
>>> > from
>>> > /home1/04020/unfelipe/aspect/source/boundary_traction/function.cc(22):
>>> > /home1/apps/gcc7_1/impi17_0/dealii/git20170615/include/deal.
>>> II/base/utilities.h(28):
>>> > catastrophic error: cannot open source file "Epetra_Comm.h"
>>> > # include <Epetra_Comm.h>
>>> >
>>> >
>>> > _______________________________________________
>>> > Aspect-devel mailing list
>>> > Aspect-devel at geodynamics.org
>>> > https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.ge
>>> odynamics.org_cgi-2Dbin_mailman_listinfo_aspect-2Ddevel&d=Dw
>>> IBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=c08Btfq
>>> 4m9QEScXN3ZQwLZzzWQE7S8CYq1IYuzKV_Zk&m=C8PprsjPtOwd6V7FRP8v4
>>> w98sgtlphLWRdU3-xQ_ZTw&s=M0J-jm1Plky7BJ-JeQdxD0zigb41014IcIcDOZWKT6s&e=
>>>
>>>
>>>
>>> --
>>> Timo Heister
>>> http://www.math.clemson.edu/~heister/
>>> _______________________________________________
>>> Aspect-devel mailing list
>>> Aspect-devel at geodynamics.org
>>> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>>
>>
>>
>
>
> _______________________________________________
> Aspect-devel mailing listAspect-devel at geodynamics.orghttp://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>
>
>
> _______________________________________________
> Aspect-devel mailing list
> Aspect-devel at geodynamics.org
> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20180502/6d84f9c8/attachment-0001.html>
More information about the Aspect-devel
mailing list