Hey,
I did not understand. Are you suggesting me to use hama-trunk or hbase trunk
?Also in the DenseMatrix.random(conf,row,column) function, each entry updates
the
table separately. Is there no faster way of doing it. You know collect like a
100
entries and only then update the table ?
Thank You
Abhishek Agrawal
SUNY- Buffalo
(716-435-7122)
On Sun 11/15/09 7:29 AM , "Edward J. Yoon" [email protected] sent:
> Hi,
>
> Hbase 0.19 have some problem related with scalability of column. The
> only answer is, you need more data node to store the large matrix. (or
> large table)
>
> I would recommend to use the hama-trunk.
>
> On Sun, Nov 15, 2009 at 1:14 PM, <aa...@buffa
> lo.edu> wrote:> Hi,
> > Â Once again I have some doubts
> about Hama. I am now trying to generate a 12000> * 12000 matrix in hama. I am
not using map
> reduce to generate this matrix. The> function call used is  a=
> DenseMatrix.random(conf,row,column). Now I wanted to> know, approximately how
much time will it take
> to generate this matrix. My> cluster contains 4 data nodes each having
> approximately 500MB RAM. The master> node also does the work of a data node.
> The
> master node also has approximately> 500MB RAM. Do I have to change any
> settings for
> a better performance or am I> missing some thing ?
> >
> > Thank You
> >
> > Abhishek Agrawal
> >
> > SUNY- Buffalo
> > (716-435-7122)
> >
> >
> >
>
>
>
> --
> Best Regards, Edward J. Yoon @ NHN, corp.
> edwardy
> [email protected]http://blog.udanax.org
>
>
>
