[Boost-users] Building a large graph using Parallel Boost Graph Library

Subject: [Boost-users] Building a large graph using Parallel Boost Graph Library
From: William Shipman (shipman.william_at_[hidden])
Date: 2013-06-15 17:19:54

Hi all,

I'm working with a graph that has about 16 million vertices and 320 million
edges.  Currently, I'm using graph-tool in Python but building the graph is
far too slow.  Before I start porting to C++ I starting reading up on the
Parallel Boost Graph Library.  From the examples in the documentation, it
seems that one must build the graph on one node in the cluster, then call
synchronize to send that graph to all nodes.  Is there a way to build parts
of the graph on each node and then synchronize those parts?

Some background: the graph actually comes from a CT scan.  The vertices are
voxels, the edges are relationships between adjacent voxels.  At the moment
I'm working on a very small test volume - 256^3 voxels - but later volumes
will have billions of vertices and edges.

Thanks for the help.