[Boost-users] [graph] Example parallel BGL

Subject: [Boost-users] [graph] Example parallel BGL
From: Sensei (senseiwa_at_[hidden])
Date: 2015-07-30 09:51:09

Dear all,

I'd like to play with Parallel BGL (with MPI), and I'd like to see some 
example code. I've followed the Overview and the Adjacency List example, 
but I cannot compile my code.

Right now, I'd like to create a graph, and then fill in with some nodes. 
Then, access the nodes belonging to a single process, but I wish I had 
an error there

My error concerns the definition of the graph: No template named 
'bsp_process_group' in namespace 'boost::parallel'; did you mean 
'boost::process_group'? The code follows.

What am I doing wrong here? I may have forgotten some header... Is there 
any distributed (complete) example code that I might read?

Thanks & Cheers!

#include <iostream>
#include <string>
#include <vector>

#include <boost/mpi.hpp>
#include <boost/graph/use_mpi.hpp>
#include <boost/graph/distributed/mpi_process_group.hpp>
#include <boost/graph/distributed/adjacency_list.hpp>

int main(int argc, const char * argv[])
{
     boost::mpi::environment  env;
     boost::mpi::communicator comm;

     typedef boost::adjacency_list<boost::vecS,

boost::distributedS<boost::parallel::bsp_process_group, boost::vecS>,
                                   boost::directedS> graph;
     typedef boost::erdos_renyi_iterator<boost::minstd_rand, graph> 
generator;

     if (comm.rank() == 0)
     {
         boost::add_edge(1, 2);
         boost::add_edge(2, 2);
         boost::add_edge(2, 3);
         boost::add_edge(3, 4);
         boost::add_edge(4, 5);
         boost::add_edge(5, 1);

         // access ***only rank 0 nodes***
     }
     return 0;
}