test failures on nest master@6997cae6c on Fedora 32 with Intel Prallel Studio XE 2020 update 2
Dear all,
I have compiled NEST master@6997cae6c onFedora 32 with both, 1- gcc (10.2.1) and OpenMPI (4.0.4): the default on this distro 2 - Intel Parallel Studio XE update 2 with GSL compiled from source using the same compiler using same configuration (-Dwith-mpi=ON, -Dwith-boost=ON) . On Intel, -Dwith-gsl=${GSL_DIR} was added to point to the GSL built with the Intel Compiler.
On the gcc configuration, all tests worked fine (I defined OMPI_MCA_rmaps_base_oversubscribe=1 to avoid oversubscription error as reported in issue #1758).
On the other hand, on the Intel configuration, several MPI tests have failed as described in the attached files. (To allow oversubscription the following variables were defined: I_MPI_WAIT_MODE=1 and I_MPI_THREAD_YIELD=3)
I report it here as it is probably not directly a NEST code issue, but hope that someone in the community could have the answer. If I should submit it as an issue on GitHub, please let me know.
Attached files: cmake_command_summary.txt : cmake command invocation and summary installcheck.log : make installcheck log reports.tar.gz : reports folder created by make installcheck; it also contains the instalcheck.log file above
Best, Nilton
Dear Nilton,
this is probably a deep compile configuration issue. Maybe our expert for the MPI libraries context has an idea what could cause this. I will contact him and ask if he could have a look at this.
Best, Dennis
On 06.10.20 02:18, Nilton Liuji Kamiji wrote:
Dear all,
I have compiled NEST master@6997cae6c onFedora 32 with both, 1- gcc (10.2.1) and OpenMPI (4.0.4): the default on this distro 2 - Intel Parallel Studio XE update 2 with GSL compiled from source using the same compiler using same configuration (-Dwith-mpi=ON, -Dwith-boost=ON) . On Intel, -Dwith-gsl=${GSL_DIR} was added to point to the GSL built with the Intel Compiler.
On the gcc configuration, all tests worked fine (I defined OMPI_MCA_rmaps_base_oversubscribe=1 to avoid oversubscription error as reported in issue #1758).
On the other hand, on the Intel configuration, several MPI tests have failed as described in the attached files. (To allow oversubscription the following variables were defined: I_MPI_WAIT_MODE=1 and I_MPI_THREAD_YIELD=3)
I report it here as it is probably not directly a NEST code issue, but hope that someone in the community could have the answer. If I should submit it as an issue on GitHub, please let me know.
Attached files: cmake_command_summary.txt : cmake command invocation and summary installcheck.log : make installcheck log reports.tar.gz : reports folder created by make installcheck; it also contains the instalcheck.log file above
Best, Nilton
NEST Users mailing list -- users@nest-simulator.org To unsubscribe send an email to users-leave@nest-simulator.org
Dear Dennis,
Thanks for looking into this. I will report back if I find any solution, but looking forward to hearing from the experts as well.
Best, Nilton
On Fri, Oct 16, 2020 at 4:37 AM Dennis Terhorst d.terhorst@fz-juelich.de wrote:
Dear Nilton,
this is probably a deep compile configuration issue. Maybe our expert for the MPI libraries context has an idea what could cause this. I will contact him and ask if he could have a look at this.
Best, Dennis On 06.10.20 02:18, Nilton Liuji Kamiji wrote:
Dear all,
I have compiled NEST master@6997cae6c onFedora 32 with both, 1- gcc (10.2.1) and OpenMPI (4.0.4): the default on this distro 2 - Intel Parallel Studio XE update 2 with GSL compiled from source using the same compiler using same configuration (-Dwith-mpi=ON, -Dwith-boost=ON) . On Intel, -Dwith-gsl=${GSL_DIR} was added to point to the GSL built with the Intel Compiler.
On the gcc configuration, all tests worked fine (I defined OMPI_MCA_rmaps_base_oversubscribe=1 to avoid oversubscription error as reported in issue #1758).
On the other hand, on the Intel configuration, several MPI tests have failed as described in the attached files. (To allow oversubscription the following variables were defined: I_MPI_WAIT_MODE=1 and I_MPI_THREAD_YIELD=3)
I report it here as it is probably not directly a NEST code issue, but hope that someone in the community could have the answer. If I should submit it as an issue on GitHub, please let me know.
Attached files: cmake_command_summary.txt : cmake command invocation and summary installcheck.log : make installcheck log reports.tar.gz : reports folder created by make installcheck; it also contains the instalcheck.log file above
Best, Nilton
NEST Users mailing list -- users@nest-simulator.org To unsubscribe send an email to users-leave@nest-simulator.org
-- Dipl.-Phys. Dennis Terhorst Coordinator Software Development
Institute of Neuroscience and Medicine (INM-6) Computational and Systems Neuroscience & Theoretical Neuroscience, Institute for Advanced Simulation (IAS-6) Jülich Research Centre, Member of the Helmholz Association and JARA 52425 Jülich, Germany
Building 15.22 Room 4004 Phone +49 2461 61-85062 Fax +49 2461 61- 9460d.terhorst@fz-juelich.de
Forschungszentrum Juelich GmbH 52425 Juelich Sitz der Gesellschaft: Juelich Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498 Vorsitzender des Aufsichtsrats: MinDir Volker Rieke Geschaeftsfuehrung: Prof. Dr.-Ing. Wolfgang Marquardt (Vorsitzender), Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt
NEST Users mailing list -- users@nest-simulator.org To unsubscribe send an email to users-leave@nest-simulator.org
-
Dennis Terhorst -
Nilton Liuji Kamiji