
Boost Users : 
Subject: Re: [Boostusers] [BGL] Segmentation fault with parallel BGL
From: Quaglino Alessio (alessio.quaglino_at_[hidden])
Date: 20151106 04:25:38
It looks like the error appears only when a processor ends up with an empty edge list.
That is, the following works with mpirun np 1 but gives a segmentation fault with np 2
using namespace boost;
using boost::graph::distributed::mpi_process_group;
typedef adjacency_list<vecS, distributedS<mpi_process_group, vecS>, undirectedS> Graph;
typedef iterator_property_map<std::vector<int>::iterator, property_map<Graph, vertex_index_t>::type> LocalMap;
int nV = 30000;
int num = 0;
Graph G(nV);
synchronize(G);
std::vector<int> localComponent(nV);
LocalMap components(localComponent.begin(),get(vertex_index, G));
add_edge(vertex(0,G),vertex(1,G),G);
num = connected_components_ps(G, components);
The following works with np 2 as well (the only thing I did was adding an edge)
int nV = 30000;
int num = 0;
Graph G(nV);
synchronize(G);
std::vector<int> localComponent(nV);
LocalMap components(localComponent.begin(),get(vertex_index, G));
add_edge(vertex(0,G),vertex(1,G),G);
add_edge(vertex(29998,G),vertex(29999,G),G);
num = connected_components_ps(G, components);
Unfortunately, in my application it is very likely that some processors end up with no edges. Is there a way to fix this other than adding fake edges from a vertex to itself?
Thanks,
Alessio
Boostusers list run by williamkempf at hotmail.com, kalb at libertysoft.com, bjorn.karlsson at readsoft.com, gregod at cs.rpi.edu, wekempf at cox.net