pthreads - Calling MPI functions from multiple threads -

i want implement following thing using mpi , pthreads facing error:

each processor have 2 threads. each processor's 1 thread sending data other processors , other thread receiving data other processors. when implementing it, giving segmentation fault error messages "current bytes -40, total bytes 0, remote id 5".

just testing purpose, when using 1 thread per processor , either sending or receiving data, errors not occur.

i found info "in general, there may problems if multiple threads make mpi calls. program may fail or behave unexpectedly. if mpi calls must made within thread, should made 1 thread." in following link: https://computing.llnl.gov/tutorials/pthreads/

i want use 2 threads per processor 1 thread use mpi_send function send data , other thread receive mpi_recv function receive data without using locking mechanism. has idea how implement or how use multiple threads call mpi functions without using mutex or locking mechanism?

here code:

int rank, size, msg_num;  // thread function sending messages void *send_func_for_thread(void *arg) {     int send, procnum, x;     send = rank;      for(x=0; x < msg_num; x++)     {         procnum = rand()%size;         if(procnum != rank)             mpi_send(&send, 1, mpi_int, procnum, 0, mpi_comm_world);     }  // sending special message other processors tag = 128 signal finishing of sending message      (x = 0; x < size; x++)     {         if(x != rank)             mpi_send(&send, 1, mpi_int, x, 128, mpi_comm_world);         }      pthread_exit((void *)null); }   // thread function receiving messages  void *recv_func_for_thread(void *arg) {     mpi_status status;     int recv, counter = 0;      while(counter != size - 1)     {         mpi_recv(&recv, 1, mpi_int, mpi_any_source, mpi_any_tag, mpi_comm_world, &status);         if(status.mpi_tag == 128)             counter++;     }      pthread_exit((void *)null); }   int main(int argc, char **argv) {     void *stat;     pthread_attr_t attr;     pthread_t thread[2];      mpi_init(&argc, &argv);     mpi_comm_rank(mpi_comm_world, &rank);   // rank -> rank of processor     mpi_comm_size(mpi_comm_world, &size);   // size -> total number of processors      srand((unsigned)time(null));      msg_num = atoi(argv[1]);      pthread_attr_init(&attr);     pthread_attr_setdetachstate(&attr, pthread_create_joinable);      // thread 0 sending   messages     pthread_create(&thread[0], &attr, send_func_for_thread, (void *)0);      // thread 1 receiving messages     pthread_create(&thread[1], &attr, recv_func_for_thread, (void *)1);      pthread_attr_destroy(&attr);      pthread_join(thread[0], &stat);     pthread_join(thread[1], &stat);      cout << "finished : proc " << rank << "\n";      mpi_finalize();     pthread_exit((void *)null);     return 0;    }  compile: ========  module load mvapich2/gcc;        mpicxx -lpthread -o demo demo.cpp  run: ==== mpiexec -comm mpich2-pmi demo 10000000  ran program 3 processors , got segmentation fault. 

(since haven't provided example, following speculation.)

you must initialize mpi using mpi_init_thread() instead of mpi_init(). if understand explanation correctly, "required" argument must have value mpi_thread_multiple. if mpi_init_thread() returns lower level thread support in "provided" argument, means mpi implementation doesn't support mpi_thread_multiple; in case must else. see http://www.mpi-forum.org/docs/mpi-20-html/node165.htm .