hama-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Thomas Jungblut <thomas.jungb...@gmail.com>
Subject Re: How to load pre-partitioned data separately?
Date Wed, 15 Aug 2012 12:06:49 GMT
Stay tuned at:

https://issues.apache.org/jira/browse/HAMA-561

We are going to introduce this for the whole framework, not just for the
graph package.
Sorry for the inconvenience.

2012/8/15 Suraj Menon <surajsmenon@apache.org>

> Hi,
>
> Sorry, you would go through a re-partitioning at runtime today. I consulted
> Thomas and I think we can get this fixed in the next release.
> The next release would be made between end of this month and mid next
> month.
>
> Thanks,
> Suraj
>
> On Tue, Aug 14, 2012 at 4:54 AM, Bo Locker <locker.fighter@gmail.com>
> wrote:
>
> > Hello,
> > I want to use hama in large graph processing, but here comes the problem:
> > the graph data is pre-partitioned and I want each groom gets its own
> part.
> > The data looks like following:
> > 1 2
> > 1 3
> > 2 3
> > 2 4
> > 3 4
> > ...
> > Each line represents an edge in the graph. We pre-partition the graph
> into
> > several parts and each groom holds one part.
> > I wonder if this problem could be fixed. How could i deal with it?
> Should I
> > load the data first, then let each groom get its own part by
> communication,
> > or should I use the Partition class? Some suggestions must be
> appreciated.
> > Thank you...
> > ps. The graph data is extremely large, about 10 billion nodes.
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message