A distributed computing system consists of multiple autonomous processors that do not share primary memory, but cooperate by communications over network.
 

Message Passing Interface.

A library which can be called from C, C++ and Fortran 77, used to implement message-passing model on distributed system, hopefully with efficiency, portability, and functionality.
 

Runing a C program on our Department Falcon clusters:

      Add /usr/local/mpi/bin to your path by adding the following line to either .cshrc or
      .login file:
        set path=($path . /usr/local/bin  /usr/local/mpi/bin)
 
      Create a .rhosts file in your home directory with the following content:

        falcon2 your_user_name
        falcon3 your_user_name
        .....
        falcon20 your_user_name

      Then excute: chmod 600 .rhosts

      Commands to run a program (Example: your_source_code.c)

        mpicc your_source_code.c -lm
        mpirun -np  4 (# of processors) a.out
 

   MPI_Init                      Initialize MPI
   MPI_Comm_size         Find out how many processes there are
   MPI_Comm_rank        Find out which processor I am
   MPI_Bcast                   Send data to all other processors
   MPI_Reduce                Collect and further process data from all processors
   MPI_Send                    Send a message
   MPI_Recv                    Recieve a message
   MPI_Wtime                 Give the timing of a process
   MPI_Finalize               Terminate MPI
 
   MPI_Init( &argc, &argv )

   MPI_Comm_size(MPI_COMM_WORLD, &numtasks)

   MPI_Comm_rank(MPI_COMM_WORLD, &rank)

   MPI_Bcast(void* buf, int count, MPI_Datatype datatype, int root, MPI_Comm
   comm)

   MPI_Reduce(void* sendbuf, void* recvbuf, int count, MPI_Datatype datatype,
   MPI_Op op, int root, MPI_Comm comm)

   MPI_Send(void* buf, int count, MPI_Datatype datatype, int dest, int tag,
   MPI_Comm comm)

   MPI_Recv(void* buf, int count, MPI_Datatype datatype, int source, int tag,
   MPI_Comm comm, MPI_Status status)

   starttime = MPI_Wtime()

   MPI_Finalize()
 
 

int main(int argc, cgar argv[] )
{

      int n, myid, numprocs, i;
      double myq q, h, sum, x;
       MPI_Init(&argc, &argv);
       MPI_Comm_size(MPI_COMM_WORLD, &numprocs);
       MPI_Comm_rank(MPI_COMM_WORLD, &&myid);
      while (1) {
            if (myid==0){
              printf("Enter the number of intervals: (0 quits) ");
               scanf("%d", &n);
             }
             MPI_Bcast(&n, 1, MPI_INT, 0, MPI_COMM_WORLD);
             if (n==0)
                break;
             else {
                 h = 1.0 / (double) n;
                 sum = 0.0
                 for (i= myid + 1; i <= n; i += numprocs) {
                     x = h * ((double)i - 0.5);
                     sum +=  sqrt( 1.0 - x*x)
                  }
              myq = h * sum
              MPI_Reduce(&myq, &q, 1, MPI_DOUBLE, MPI_SUM, 0,   MPI_COMM_WORLD);
               if (myid == 0)
                   printf("q is approximately %.16f, q);
      }
  }
   MPI_Finalize();
    return 0;
}