Re: [Boost-users] Confused with Depth-First-Searchmethod, forwardorcross edge found on a undirected graph ?

Subject: Re: [Boost-users] Confused with Depth-First-Searchmethod, forwardorcross edge found on a undirected graph ?
From: Jeremiah Willcock (jewillco_at_[hidden])
Date: 2013-06-28 14:22:12

On Fri, 28 Jun 2013, lizy10b wrote:

> Hi Jeremiah Willcock,
> Â 
> Attached is the build log of boost.mpi with mingw (zip format).
> Successful output libraries are:
> libboost_python-mgw47-d-1_53.dll
> libboost_python-mgw47-d-1_53.dll.a
> libboost_serialization-mgw47-d-1_53.dll
> libboost_serialization-mgw47-d-1_53.dll.a
> Â 
> No boost.mpi library output.
> Â 
> My environment is Win7 x64, Python 2.66 x86 (binary), MPICH2 1.32p1 x86 (binary),
> gcc 4.7.1 from CodeBlocks with MinGW x86 (binary).
> Â 
> It lools like there is some codes in mpicxx.h which are incompatible to boost.mpi building process.

Here is the actual error (for search engines):

D:\Program Files (x86)\MPICH2\include/mpicxx.h: At global scope:
D:\Program Files (x86)\MPICH2\include/mpicxx.h:2469:17: error: declaration 
does not declare anything [-fpermissive]

It appears that this is a known issue 
(http://trac.mpich.org/projects/mpich/ticket/1582) that the MPICH team has 
decided not to fix for some reason.  Could you please try adding 
MPICH_SKIP_MPICXX as a -D flag (to define it as a macro) when building 
Boost.MPI?  I don't remember how to do that, but it should be easy to 
find.  Thanks.

-- Jeremiah Willcock

> 
> ___________________________________________________________________________________________________________________________________________________________
> 发件人： Jeremiah Willcock
> 发送时间： 2013-06-28  00:54:20
> 收件人： boost-users
> 抄送：
> 主题： Re: [Boost-users] Confused with Depth-First-Searchmethod,forwardorcross edge found on a undirected graph ?
> On Fri, 28 Jun 2013, lizy10b wrote:
> >  Hi Jeremiah Willcock,
> > Good news, the parallal version BFS example (\graph_parallel\example\breadth_first_search.cpp) could be built now.
> > I followed your suggestion to locate the boost::detail::bfs_helper(), there are two declarartions of this function, A and B.
> >  A is in \graph\breadth_first_search.hpp for the serial version, with another definition for the parallel version of this function (B) which is wrapped
> > inside a macro (#ifdef BOOST_GRAPH_USE_MPI... #endif).
> > And the delcaration of B is in \graph\distributed\breadth_first_search.hpp.
> > I just comment out the delcaration of B and copy it to overwrite the definition inside the above macro.
> > and append "* = 0" at the end of the last parameter just like function A.
> > And then the example cpp built successfully.
> Could you please see if r84909 in the trunk (just committed) works without 
> needing any extra changes?  I added the "= 0" there but didn't do any of 
> the other changes you mention.
> > As for the build error of Boost.MPI with MinGW, I will redo it and save the ouput to a file.
> Thanks.
> -- Jeremiah Willcock
> >Â 
> >Â ________________________________________________________________________________________________________________________________________________________
> _
> >Â lizy10b
> >Â 
> >Â ________________________________________________________________________________________________________________________________________________________
> _
> > 发件人： Jeremiah Willcock
> > 发送时间： 2013-06-27  22:47:02
> > 收件人： boost-users
> > 抄送：
> > 主题： Re: [Boost-users] Confused with Depth-First-Searchmethod,forwardorcross edge found on a undirected graph ?
> > On Thu, 27 Jun 2013, lizy10b wrote:
> > > Hi Jeremiah Willcock,
> > > Thanks for your reply.
> > > Sad to say the pararallel version BFS example file (\boost_1_53_0_\libs\graph_parallel\example\breadth_first_search.cpp) can not be compiled with VS20
> >Â 10
> > > on my end. It complains an unresolved external symbol "void _cdecl boost::bfs_helper....". And I found a similar report on
> >Â >Â http://boost.2283326.n4.nabble.com/LINK-Error-while-building-PBGL-example-breadth-first-search-cpp-td3799909.html.%c2%a0but%c2%a0with%c2%a0no%c2%a0result.
> > Is it boost::bfs_helper or boost::detail::bfs_helper?  Are there any 
> > symbols in the object file for breadth_first_search that are defined and 
> > demangle to names involving bfs_helper?
> > > I also tried to build the boost with MinGW, but it seems the Boost.MPI cannot be compiled correctly upon MPICH2, so the parallel BGL library will not
> > > be built either. 
> > What is the issue with that version?
> > -- Jeremiah Willcock
> >Â >Â Â 
> >Â >Â 2013-06-27
> >Â >Â 
> >Â >Â ______________________________________________________________________________________________________________________________________________________
> >Â ___
> >Â >Â lizy10b
> >Â >Â 
> >Â >Â ______________________________________________________________________________________________________________________________________________________
> >Â ___
> > > 发件人： Jeremiah Willcock
> > > 发送时间： 2013-06-26  22:54:58
> > > 收件人： lizy10b
> > > 抄送： boost-users
> > > 主题： Re: Re: Re: [Boost-users] Confused with Depth-First-Searchmethod,forwardorcross edge found on a undirected graph ?
> > > On Wed, 26 Jun 2013, lizy10b wrote:
> > > >  
> > > > Hi Jeremiah Willcock,
> > > > Thanks for your reply.
> > > > As maybe there are some problems when invoke depth_first_visit() on an undirected graph, the undirected_dfs()
> > > > could be a very good substitution.
> > > > So I continue to test the parallel version of the DFS function -- tsin_depth_first_visit(), and the test file comes
> > > > from  \boost_1_53_0\libs\graph_parallel\test\distributed_dfs_test.cpp.
> > > > The test file invokes the tsin_depth_first_visit() on a distributed undirected graph, and the visitor print all
> > > > events,  the "forward_or_cross_edge" appears four times. So maybe there are some problems in
> > > > tsin_depth_first_visit() too. 
> > > > Does the parallel tsin_depth_first_visit() base on the serial depth_first_visit(), so they appear similiar problem
> > > > when work with undirected graph?
> > > It looks like it's similar to the directed DFS, so it would likely return 
> > > the same results.
> > > > My final target is to find a parallel version DFS works with undirected graph, in order to find all cycles
> > > > parallelly.  Could you please give me some advices on that? I am stucked.
> > > It doesn't look like there is one, but you can use breadth-first search to 
> > > find cycles as well.  Would that work for you?
> > > -- Jeremiah Willcock
> >Â >Â >Â 
> >Â >Â >Â ____________________________________________________________________________________________________________________
> >Â >Â >Â lizy10b
> >Â >Â >Â 
> >Â >Â >Â ____________________________________________________________________________________________________________________
> > > > 发件人： Jeremiah Willcock
> > > > 发送时间： 2013-06-26  03:34:37
> > > > 收件人： lizy10b
> > > > 抄送： boost-users
> > > > 主题： Re: Re: Re: [Boost-users] Confused with Depth-First-Searchmethod,forwardor cross edge found on a undirected
> > > > graph ?
> > > > On Tue, 25 Jun 2013, lizy10b wrote:
> > > > > Hi Jeremiah Willcock,
> > > > >    Can I use the predecessor_map in undirected_dfs() like in dijkstra_shortest_paths()?
> > > > > undirected_dfs(g, predecessor_map(&p[0]).root_vertex(vertex_descriptor(0)).visitor(vis)
> > > > >                  .edge_color_map(get(edge_color, g)).vertex_color_map(get(vertex_color, g)));
> > > > > It does not work on my end, all elements in predecessor remain 0 after invoke the undirected_dfs() .
> > > > If that parameter isn't in the documentation for undirected_dfs, the 
> > > > argument will just be ignored.
> > > > > In the last section (Extending Algorithms with Visitors) of http://www.boost.org/doc/libs/1_53_0/libs/graph/doc/q
> >Â >Â >Â uick_tour.html,
> > > > > an example shows how to write a predecessor_map using visitor,
> > > > > and the "event point" is edge_relaxed().
> > > > > But as the undirected_dfs document shows there is no edge_relaxed() for DFS.
> > > > > So is there any way to save predecessors in DFS?
> > > > > For example an edge (u,v) and start at vertex u
> > > > > When the discover_vertex(Vertex, Graph&) is invoked at vertex v, then 
> > > > > inside the discover_vertex() how could I know the previous discovered 
> > > > > vertex is u?
> > > > I believe you would want to hook tree_edge() instead of edge_relaxed().
> > > > -- Jeremiah Willcock
> >Â >Â >Â >Â 
> >Â >Â >Â >Â _________________________________________________________________________________________________________________
> >Â >Â >Â ________________________________________
> > > > > 发件人： Jeremiah Willcock
> > > > > 发送时间： 2013-06-25  00:44:05
> > > > > 收件人： lizy10b
> > > > > 抄送： boost-users
> > > > > 主题： Re: Re: [Boost-users] Confused with Depth-First-Search method,forwardor cross edge found on a undirected g
> > > > raph ?
> > > > > On Tue, 25 Jun 2013, lizy10b wrote:
> > > > > > Hi Jeremiah Willcock ,
> >Â >Â >Â >Â >Â ?
> > > > > > It seems undirected_dfs() works. I have not燾heck the result according to the graph, but at least there is no
> > > > > > forward edge or cross edge output.
> > > > > > Attached is the revised cpp file using undirected_dfs() .
> > > > > > My question is why edge_color_map(get(edge_color,爂))?is required when calling the undirected_dfs() .
> >Â >Â >Â >Â >Â Thanks
> > > > > It looks like what is going wrong with using directed DFS is that edges 
> > > > > are examined twice (once in each direction), and so you can have forward 
> > > > > edges even in a symmetric graph (which is how undirected graphs are 
> > > > > treated in directed-graph algorithms in BGL).  An example is the triangle 
> > > > > 0 -- 1 -- 2 -- 0: starting at 0 and processing the edges out of a vertex 
> > > > > in numerical order, 0 -> 1 is a tree edge, 1 -> 0 is a back edge, 1 -> 2 
> > > > > is a tree edge, 2 -> 0 is a back edge, 2 -> 1 is a back edge, then 0 -> 2 
> > > > > is a forward edge.  An undirected DFS would mark 2 -> 0 as a back edge 
> > > > > then ignore 0 -> 2 since it is the same as 2 -> 0.  The edge colors in 
> > > > > undirected_dfs are used to do this filtering-out of multiple directions 
> > > > > for the same edge.
> > > > > -- Jeremiah Willcock
> >Â >Â >Â >Â 
> >Â >Â >Â >
> >Â >Â >Â 
> >Â >Â >
> >Â >Â 
> >Â >
> >Â _______________________________________________
> > Boost-users mailing list
> >Â Boost-users_at_[hidden]
> >Â http://lists.boost.org/mailman/listinfo.cgi/boost-users
> >Â 
> >
> _______________________________________________
> Boost-users mailing list
> Boost-users_at_[hidden]
> http://lists.boost.org/mailman/listinfo.cgi/boost-users
> 
>