giraph-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From KAUSHIK SARKAR <countkaus...@gmail.com>
Subject Re: Termination Conditions
Date Sat, 04 Aug 2012 05:58:07 GMT
Hi Nick,

I guess you can use MasterCompute together with BooleanAndAggregator to
solve your problem. Whenever a vertex decides to halt, it will notify the
MasterCompute with "true" value for the aggregator, otherwise it will send
"false". If all the vertices agree to halt at a superstep, then the value
of the  BooleanAndAggregator at the MasterCompute will evaluate to "true",
and the masterCompute can then call the haltComputation() method to stop
computation. Any message sent at the last superstep will have no effect
(which you want in this case). Of course this solution assumes that you are
using Giraph-0.2, since older versions do not have the MasterCompute class
for centralized decision making of this type.

Hope this helps.

Regards,
Kaushik

On Fri, Aug 3, 2012 at 11:28 AM, Nick West <nick.west@benchmarksolutions.com
> wrote:

>  Thanks for the reply.
>
>  Is there an easy modification that I can make to remove condition 2?
>  Can you point me to the code that addresses this?
>
>  The problem I am facing is the following:  At every iteration a
> non-halted vertex needs messages from all of its neighbors.  When deciding
> to send messages, a given vertex doesn't know if its neighbors will
> vote-to-halt in the current superstep, thus it must send a message to each
> of its neighbors.  In the case that all vertices have voted to stop, the
> sending of a messages by any vertex will cause the algorithm to continue,
> yet in this situation it is desirable to terminate.
>
>  I have worked out a few solutions that involve either increasing the
> amount of data a vertex saves each iteration or augmenting the messages
> sent with additional information, but I think it would be beneficial, and
> more general, to allow this type of termination instead.
>
>  Do you have any thoughts on this?
>
>  Thanks,
> Nick
>
>
>    On Aug 3, 2012, at 10:14 AM, Alessandro Presta wrote:
>
>  Hi Nick,
>
>  You are pretty much correct, except that not all vertices need to vote
> to halt at the same time: some vertices might have voted to halt at a
> previous superstep and never received any messages after then, in which
> case they are never reactivated.
>
>  In other words, I think you can rephrase that as:
>
>    1. All vertices are halted after a given superstep
>    2. No messages were sent in that superstep
>
> Hope it helps.
>
>  Alessandro
>
>   From: Nick West <nick.west@benchmarksolutions.com>
> Reply-To: "user@giraph.apache.org" <user@giraph.apache.org>
> Date: Friday, August 3, 2012 2:48 PM
> To: "user@giraph.apache.org" <user@giraph.apache.org>
> Subject: Termination Conditions
>
>   Excuse me if this is stated somewhere obvious, but I haven't been
> unable to find it.  What are the exact termination criteria for the global
> algorithm?
>
>  Reading the documentation on voteToHalt, looking at the Shortest Path
> Example code, and looking at the results of my own application, these two
> conditions must both hold for the global BSP algorithm to terminate:
>
>  1) All vertices vote to halt in a given superstep
> 2) No messages are sent in that supersetp
>
>  Is that correct?
>
>  Thanks,
>  *Nick West
> **
> *Benchmark Solutions
> 101 Park Avenue - 7th Floor
> New York, NY 10178
> Tel +1.212.220.4739 | Mobile +1.646.267.4324
> *www.benchmarksolutions.com * <http://www.benchmarksolutions.com/>
> **
> *<image001.png>
>
>
>
>    *
> **
>   <image001.png>
>
>
> *
> Nick West
> **
> *Benchmark Solutions
> 101 Park Avenue - 7th Floor
> New York, NY 10178
> Tel +1.212.220.4739 | Mobile +1.646.267.4324
> *www.benchmarksolutions.com * <http://www.benchmarksolutions.com/>
> ***
>
>
>
>    *
> **
>

Mime
View raw message