Hello Fedele, Would it be possible to build the open mpi package to use gfortran and run the test again?
Do you observe this problem if you build a Open MP (<-> MP not MPI) only version of the test case? I can't reproduce this problem using gfortran. I don't have access to an Intel compiler at the moment. Also, please send the output of ompi_info. Thanks, Howard 2015-06-25 10:37 GMT-06:00 Fedele Stabile <fedele.stab...@fis.unical.it>: > Hello to all, > I'm trying hybrid OpenMP + MPI programming, when I run the simple code > listed below I have an error: > forrtl: severe (40): recursive I/O operation, unit -1, file unknown > Image PC Routine Line > Source > aa 0000000000403D8E Unknown Unknown > Unknown > aa 0000000000403680 Unknown Unknown > Unknown > libiomp5.so 00002B705F7C5BB3 Unknown Unknown > Unknown > libiomp5.so 00002B705F79A617 Unknown Unknown > Unknown > libiomp5.so 00002B705F799D3A Unknown Unknown > Unknown > libiomp5.so 00002B705F7C5EAD Unknown Unknown > Unknown > libpthread.so.0 00002B705FA699D1 Unknown Unknown > Unknown > libc.so.6 00002B705FD688FD Unknown Unknown > Unknown > ------------------------------------------------------- > Primary job terminated normally, but 1 process returned > a non-zero exit code.. Per user-direction, the job has been aborted. > ------------------------------------------------------- > -------------------------------------------------------------------------- > mpirun detected that one or more processes exited with non-zero status, > thus causing > the job to be terminated. The first process to do so was: > > Process name: [[61634,1],0] > Exit code: 40 > > I have compiled OpenMPI using this configuration options: > ./configure --prefix=/data/apps/mpi/openmpi-1.8.4-intel > -enable-mpirun-prefix-by-default --enable-mpi-fortran > --enable-mpi-thread-multiple > --with-tm=/usr/local/torque-5.1.0-1_4048f77c/src --with-verbs > --with-openib --with-cuda=/usr/local/cuda-6.5 > > This is the listing of the simple code: > program hello > include "mpif.h" > > integer numprocs, rank, namelen, ierr > character*(MPI_MAX_PROCESSOR_NAME) processor_name > integer iam, np > integer omp_get_num_threads, omp_get_thread_num > > call MPI_Init(ierr) > call MPI_Comm_size(MPI_COMM_WORLD, numprocs, ierr) > call MPI_Comm_rank(MPI_COMM_WORLD, rank, ierr) > call MPI_Get_processor_name(processor_name, namelen, ierr) > iam = 0 > np = 1 > !$omp parallel default(shared) private(iam, np) > > np = omp_get_num_threads() > iam = omp_get_thread_num(); > write(*,*)"Hello from thread ", iam," out of ", np, > % " from process ", rank," out of ", numprocs, > % " on ", processor_name > > !$omp end parallel > call MPI_Finalize(ierr) > stop > end > > Can you help me to solve the problem? > Thank you, > Fedele Stabile > > > _______________________________________________ > users mailing list > us...@open-mpi.org > Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users > Link to this post: > http://www.open-mpi.org/community/lists/users/2015/06/27192.php >
