Mpi Send.tar.gz_mpi send_show

#include <mpi.h> #include <stdio.h> #include <malloc.h> #include <time.h> //-- A program to generate n number and storing in an array //-- and adding these no.'s in Parallel 95/07/14 //-- TO COMPILE mpicc fname.c //-- TO RUN mpirun -n pp a.out yy zz //-- pp (integer) No.of Processors //-- yy (integer) No.of input to generate //-- zz (integer) Range of input data to be generated randomly int getsum (int a[], int n, int id){ int i, sum = 0; // Add my portion Of data for(i = 0; i < n; i++) sum += a[i]; printf("Sum computed by processor %2d is: %10d\n", id, sum); return sum; } int main(int argc, char **argv){ int i, k, procs; int myid, mysum, finalsum; int *arr, *myarr, *sumarr, n, range, err; // initialize MPI_Init err = MPI_Init(&argc, &argv); if (err != MPI_SUCCESS){ printf("\nError initializing MPI.\n"); MPI_Abort(MPI_COMM_WORLD, err); } // end if // Get No. of processors MPI_Comm_size(MPI_COMM_WORLD, &procs); // Get processor id MPI_Comm_rank(MPI_COMM_WORLD, &myid); if (myid == 0) {// to print only once.... if (argc < 3) { printf("\n\tOOOPS...., INVALID No of Arguements,\n" "\tA program to generate n number and storing in an array\n" "\tand adding n numbers in parallel using Bcast and Reduce...\n" "\tTO RUN mpirun -n pp a.out nn rr\n" "\t\tpp (integer) No.of Processors...\n" "\t\tnn (integer) No. of inputs (size of array)...\n" "\t\trr (integer) Range of data to be generated randomly...\n" "\t\tNote: nn should be divisible by pp, otherwise some data will be lost...\n\n"); } // end if } // end myid == 0 if (argc < 3) {MPI_Finalize(); return 0;} // end if n = atoi(argv[1]); // get n range = atoi(argv[2]); // get range k = n / procs; // size for each processor arr = malloc(n * sizeof (int)); // allocate space for initial data myarr = malloc(k * sizeof (int)); // allocate space for each processor sumarr= malloc(procs * sizeof (int)); // allocate space for local sums // Generate Data if (myid == 0) { //srand(time(NULL)); time_t t1; time(&t1); // get system time srand(t1); // Initilize Random Seed for(i = 0; i < n; i++){// generate & print data arr[i] = rand() % range; printf("%6d ", arr[i]); if ((i+1) % 10 == 0) printf("\n"); } // end for } // end myid == 0 //MPI_Scatter automatically moves every k data from source to dest, //*source, k, MPI_INT, *dest, k, MPI_INT, 0, MPI_COMM_WORLD); MPI_Scatter(arr, k, MPI_INT, myarr, k, MPI_INT, 0, MPI_COMM_WORLD); // Add my portion Of data mysum = getsum(myarr, k, myid); // Gather sum //MPI_Gather automatically moves every k data from source to dest //*source, k, MPI_INT, *dest, k, MPI_INT, 0, MPI_COMM_WORLD); MPI_Gather(&mysum, 1, MPI_INT, sumarr, 1, MPI_INT, 0, MPI_COMM_WORLD); if (myid == 0) { finalsum = 0; // Master sholud add all the received sums for(i = 0; i < procs; i++) finalsum += sumarr[i]; // print final sum printf("The finalsum is %d.\n", finalsum); } // --free allocated spaces free (arr); free(myarr); free(sumarr); MPI_Finalize(); return 0; } // end main