MPI
/** * <Parallel programming technology for high performance computing> * * @Copyright Du Zhihui :edited * Li Sanli :review * Chen Yu Liu Peng :proofreading * * Rong Tao study note: 2017.10 * *//** * mpicc -o outfile Demo.c * mpirun -n 2 ./outfile > out.txt * */#include "mpi.h"#include <stdio.h>#include <math.h>int main(int argc,char *argv[]){ int myid, numprocs; char message[20]; MPI_Status status; MPI_Init(&argc,&argv); MPI_Comm_rank(MPI_COMM_WORLD,&myid); MPI_Comm_size(MPI_COMM_WORLD,&numprocs); if(myid == 0) { strcpy(message, "hello, process I"); MPI_Send(message, 20, MPI_CHAR, 1, 999, MPI_COMM_WORLD); } else if(myid == 1) { MPI_Recv(message, 20, MPI_CHAR, 0, 999, MPI_COMM_WORLD, &status); printf("received : %s\n",message); } MPI_Finalize();}/**MPI_INTEGER MPI_REAL MPI_DOUBLE_PRECISION MPI_COMPLEX MPI_LOGICAL MPI_CHARACTER MPI_BYTE MPI_PACKED */
编译与运行：
-bash-4.1$ mpicc -o a Demo_02_MPI_Send_MPI_Recv_01.c-bash-4.1$ mpirun -n 2 ./areceived : hello, process I-bash-4.1$再一个例子：
这个例子展示了所有节点互相通信，直至到rank号最大的节点终止
#include "mpi.h"#include <stdio.h>#include <math.h>int main(int argc,char *argv[]){ int myid, numprocs; int value; MPI_Status status; MPI_Init(&argc,&argv); MPI_Comm_rank(MPI_COMM_WORLD,&myid); MPI_Comm_size(MPI_COMM_WORLD,&numprocs); do{ if(myid == 0) { scanf("%d",&value); if(numprocs > 1) { MPI_Send(&value, 1, MPI_INT, myid+1, 0, MPI_COMM_WORLD); fprintf(stderr,"%d send (%d) to %d\n",myid, value, myid+1); } } else { MPI_Recv(&value, 1, MPI_INT, myid-1, 0, MPI_COMM_WORLD, &status); fprintf(stderr,"%d receive (%d) to %d\n",myid, value, myid+1); if(myid < numprocs-1) { MPI_Send(&value, 1, MPI_INT, myid+1, 0, MPI_COMM_WORLD); fprintf(stderr,"%d send (%d) to %d\n",myid, value, myid+1); } } MPI_Barrier(MPI_COMM_WORLD); }while(value >= 0); MPI_Barrier(MPI_COMM_WORLD); MPI_Finalize();} 编译与运行： -bash-4.1$ mpicc -o a Demo_02_MPI_Send_MPI_Recv_02.c-bash-4.1$ mpirun -n 3 ./a10 send (1) to 11 receive (1) to 21 send (1) to 22 receive (1) to 320 send (2) to 11 receive (2) to 21 send (2) to 22 receive (2) to 3-10 send (-1) to 11 receive (-1) to 21 send (-1) to 22 receive (-1) to 3-bash-4.1$
MPI_Send和MPI_Recv（初识1）