You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@hama.apache.org by wang fangfang <ic...@gmail.com> on 2012/09/21 12:18:56 UTC

pagerank in hama

hi
  There is a guestion when I study the code of pagerank in hama
example.there is a line:"
*numberVertices = vertices.size() * peer.getNumPeers();"* in
GraphJobRunner.java file.
However it‘s impossible to averagely partition the vertex so beautifully.
 I want to know what's the consideration when taking this decision.
Thank you very much.
Best wishes

Re: pagerank in hama

Posted by Yuesheng Hu <yu...@gmail.com>.

I strongly suggest you to use the hama-trunk.
This is a guide to import hama-trunk into eclipse and how to debug hama
itself. http://www.oriovo.com/cloud/hama-debug.html
Have fun!


2012/9/21 Thomas Jungblut <th...@gmail.com>

> Hi,
>
> that was a simple heuristic. It is now fixed in trunk by using two
> supersteps that exchange the number of vertices.
>
> 2012/9/21 wang fangfang <ic...@gmail.com>
>
> > hi
> >   There is a guestion when I study the code of pagerank in hama
> > example.there is a line:"
> > *numberVertices = vertices.size() * peer.getNumPeers();"* in
> > GraphJobRunner.java file.
> > However it‘s impossible to averagely partition the vertex so beautifully.
> >  I want to know what's the consideration when taking this decision.
> > Thank you very much.
> > Best wishes
> >
>

Re: pagerank in hama

Posted by Thomas Jungblut <th...@gmail.com>.

Hi,

that was a simple heuristic. It is now fixed in trunk by using two
supersteps that exchange the number of vertices.

2012/9/21 wang fangfang <ic...@gmail.com>

> hi
>   There is a guestion when I study the code of pagerank in hama
> example.there is a line:"
> *numberVertices = vertices.size() * peer.getNumPeers();"* in
> GraphJobRunner.java file.
> However it‘s impossible to averagely partition the vertex so beautifully.
>  I want to know what's the consideration when taking this decision.
> Thank you very much.
> Best wishes
>