分布式计算之母 – MPI

大家好，欢迎来到IT知识分享网。

安装 OpenMPI

# 下载 OpenMPI 最新稳定版本 curl -O https://download.open-mpi.org/release/open-mpi/v5.0/openmpi-5.0.6.tar.gz # 解压 tar -zxvf openmpi-5.0.6.tar.gz # 编译安装 cd openmpi-5.0.6 mkdir $HOME/openmpi ./configure --prefix=$HOME/openmpi make -j 24 all # 这里指定了用24个线程并行编译 make install echo -e '\nexport LD_LIBRARY_PATH=$HOME/openmpi/lib:$LD_LIBRARY_PATH\n' >> ~/.bashrc echo -e '\nexport PATH=$HOME/openmpi/bin:$PATH\n' >> ~/.bashrc source $HOME/.bashrc # 删除原安装包/目录 rm -rf openmpi-5.0.6/

检查安装是否正确

which mpicc which mpirun mpirun --version

MPI HelloWorld

mpi_hello_world.c:

#include <mpi.h> #include <stdio.h> int main(int argc, char argv) { // Initialize the MPI environment. The two arguments to MPI Init are not // currently used by MPI implementations, but are there in case future // implementations might need the arguments. MPI_Init(NULL, NULL); // Get the number of processes int world_size; MPI_Comm_size(MPI_COMM_WORLD, &world_size); // Get the rank of the process int world_rank; MPI_Comm_rank(MPI_COMM_WORLD, &world_rank); // Get the name of the processor char processor_name[MPI_MAX_PROCESSOR_NAME]; int name_len; MPI_Get_processor_name(processor_name, &name_len); // Print off a hello world message printf("Hello world from processor %s, rank %d out of %d processors\n", processor_name, world_rank, world_size); // Finalize the MPI environment. No more MPI calls can be made after this MPI_Finalize();

编译 mpi_hello_world.c:

mpicc -o mpi_hello_world mpi_hello_world.c

在两节点上分布式并行运行 mpi_hello_world:

mpirun -np 35 --host slurm-worker1:30,slurm-worker2:30 --use-hwthread-cpus mpi_hello_world