I think a parallel BFS from all the sources is out of any possibility. One thing one can do is do it in parallel from k sources, maybe N times (with different sources each, of course), and approximate centrality from these trees.
On Wed, Oct 16, 2013 at 8:44 AM, Silvio Di gregorio < [email protected]> wrote: > I'm also trying to implement but with poor results. my target is the > betweenness Centrality. > The only thing that i did is generate giraph code throught Green-Marl, but > this is very far from being a running code. > Update us if there are progress. > But if complexity is quadratic in the of space it's hard to do. > > > 2013/10/16 Jyoti Yadav <[email protected]> > >> Hi Silvio.. >> Thanks for your reply... >> Yes, I want to realize All-source- shortest -path..Would you please >> help me out.? >> >> >> >> >> On Tue, Oct 15, 2013 at 12:45 PM, Silvio Di gregorio < >> [email protected]> wrote: >> >>> Do you want realize all-source-shortest path? >>> Il giorno 15/ott/2013 11:21, "Jyoti Yadav" <[email protected]> >>> ha scritto: >>> >>> Hi.. >>>> If I want to run SingleSourceShortest Path example given in Giraph for >>>> all the vertices ,then what changes I need to introduce and how ?? >>>> How master node will come to know that algorithm has successfully >>>> completed for one vertex and it has to consider other vertex and >>>> reinitialize the vertex values .. >>>> >>>> Any ideas are really appreciated.... >>>> >>>> Jyoti Yadav >>>> >>> >> > -- Claudio Martella [email protected]
