MPI_ALLreduce函数

最新推荐文章于 2024-09-02 10:45:57 发布

zhuliting

最新推荐文章于 2024-09-02 10:45:57 发布

阅读量1w

点赞数

分类专栏： MPI FORTRAN 文章标签： integer include c

本文链接：https://blog.csdn.net/zhuliting/article/details/6411091

版权

MPI 同时被 2 个专栏收录

29 篇文章 1 订阅

订阅专栏

FORTRAN

21 篇文章 2 订阅

订阅专栏

此函数和MPI_Reduce的意思相同，只是最后结果在所有的进程中。相当于先做MPI_Reduce，然后再做MPI_Scatter.

program main implicit none include "mpif.h" integer myid, ierr real(kind=8) val, max0 call MPI_INIT(ierr) call MPI_COMM_RANK(MPI_COMM_WORLD, myid, ierr) val = myid + 5.d0 print *, 'value:',val call MPI_ALLreduce(val,max0,1,MPI_DOUBLE_PRECISION,MPI_MAX, & MPI_COMM_WORLD,IERR) if(myid.eq.0) print *, 'max:',max0 call MPI_FINALIZE(ierr) stop end program main

result:

[root@c0109 ~]# mpif90 hello.F -o hello [root@c0109 ~]# mpiexec -n 10 ./hello value: 5.00000000000000 max: 14.0000000000000 value: 6.00000000000000 value: 7.00000000000000 value: 11.0000000000000 value: 10.0000000000000 value: 8.00000000000000 value: 13.0000000000000 value: 14.0000000000000 value: 9.00000000000000 value: 12.0000000000000 [root@c0109 ~]#