MPI集群配置

Posted 2020-06-25

参考文档：Linux下MPI并行编程环境搭建配置

 

MPI是一种并行计算架构，MPICH是MPI的一种实现，本集群使用虚拟机安装，操作系统是ubuntu14.04，使用三台机器，用户名都是ubuntu，机器名分别是ub0, ub1, ub2

• 安装MPICH
  • 下载：http://www.mpich.org/static/downloads/3.0.4/mpich-3.0.4.tar.gz
  • 源码安装

  1. $ tar -xzvf soft/mpich-3.0.4.tar.gz
     $ cd mpich-3.0.4/
     $ ./configure --prefix=/usr/local/mpich
     $ make && sudo make install

     安装后加入环境变量到/etc/profile文件，并执行 source /etc/profile，追加内容到/etc/profile

  2. PATH=$PATH:/usr/local/mpich/bin
     MANPATH=$MANPATH:/usr/local/mpich/man
     export PATH MANPATH

• 单节点测试
  • 复制源代码包下的examples目录到安装目录下

  1. cp -r examples/ /usr/local/mpich

     执行

     mpirun -np 10 ./examples/cpi

     输出结果如下：

  2. Process 0 of 10 is on ub0
     Process 9 of 10 is on ub0
     Process 1 of 10 is on ub0
     Process 4 of 10 is on ub0
     Process 5 of 10 is on ub0
     Process 7 of 10 is on ub0
     Process 2 of 10 is on ub0
     Process 3 of 10 is on ub0
     Process 6 of 10 is on ub0
     Process 8 of 10 is on ub0
     
     pi is approximately 3.1415926544231256, Error is 0.0000000008333325
     wall clock time = 0.020644

• 集群配置
  • 需要先配置ssh免密码登录，把ub0(机器名)当作master node，也即主节点，其他是slave node，也即从节点。配置免密码ssh登录的步骤
    • 你需要把主节点的公钥分别发送给从节点，这样主节点登录从节点才是可信任的，就不用密码，以配置ub0的公钥给ub1为例
    • ub0产生公钥

    • $ ssh-keygen -t rsa

      一路enter就行了

    • 把ub0的/home/ubuntu/.ssh/id_rsa.pub 发送到ub1的/home/ubuntu/.ssh/下，如果没有/home/ubuntu/.ssh/就mkdir .ssh
    • 在ub1的/home/ubuntu/.ssh/下执行

    • $ cat id_rsa.pub >> authorized_keys

    • 尝试ub0 ssh到ub1试试，看看是否成功设置无密码登录，如果成功，就继续下一个节点吧
  • 复制编译程序到其他机器上面，这样就不用在其他机器上进行源码编译mpich，节省了时间

  1. scp -r mpich ub1:/usr/local/
     scp -r mpich ub2:/usr/local/

  • 在ub0, ub1, ub2的/etc/hosts上追加

  • 192.168.0.2 ub0
    192.168.0.3 ub1
    192.168.0.4 ub2

    注意，三台机器的/etc/hosts都要追加

  • 把主节点的/usr/local/mpich/example/cpi这个计算圆周率的可执行文件复制到/home/ubuntu目录下，并且发送到ub1和ub2的/home/ubuntu目录
  • 在主节点的/home/ubuntu目录下增加servers文件，记录集群的机器名和对应的进程数

  • ub0:2
    ub1:2
    ub2:2

  • 在ub0的/home/ubuntu目录下执行

  • $ mpiexec -n 10 -f servers ./cpi

    你就可以看到下面的结果

  • Process 0 of 10 is on ub0
    Process 1 of 10 is on ub1
    Process 4 of 10 is on ub0
    Process 5 of 10 is on ub2
    Process 6 of 10 is on ub1
    Process 7 of 10 is on ub2
    Process 8 of 10 is on ub0
    Process 9 of 10 is on ub1
    Process 2 of 10 is on ub2
    Process 3 of 10 is on ub1
    pi is approximately 3.1415926544231256, Error is 0.0000000008333325
    wall clock time = 0.018768