You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-dev@hadoop.apache.org by "Edward J. Yoon" <ed...@udanax.org> on 2008/04/21 06:15:26 UTC
2D blocking technique
Hi, communities.
Basically, current Map/Reduce provides parallel computation by row
splitting. Have anyone thought about 2D blocking technique? If so,
let's discuss together.
--
B. Regards,
Edward J. Yoon
Re: 2D blocking technique
Posted by "Edward J. Yoon" <ed...@udanax.org>.
Hmm, yes. It is just such another.
Where is that located?
On Tue, Apr 22, 2008 at 12:16 AM, arkady borkovsky <ar...@yahoo-inc.com> wrote:
> What exactly do you mean by 2D blocking?
>
> I have a Python class that runs N*M copies of a program with to inputs, with
> each task getting 1/N of one input, and 1/M of another.
> A "program" can be either Python or any executable.
>
> -- ab
>
>
>
> On Apr 20, 2008, at 9:15 PM, Edward J. Yoon wrote:
>
> > Hi, communities.
> >
> > Basically, current Map/Reduce provides parallel computation by row
> > splitting. Have anyone thought about 2D blocking technique? If so,
> > let's discuss together.
> >
> >
> > --
> > B. Regards,
> > Edward J. Yoon
> >
>
>
--
B. Regards,
Edward J. Yoon
Re: 2D blocking technique
Posted by arkady borkovsky <ar...@yahoo-inc.com>.
What exactly do you mean by 2D blocking?
I have a Python class that runs N*M copies of a program with to
inputs, with each task getting 1/N of one input, and 1/M of another.
A "program" can be either Python or any executable.
-- ab
On Apr 20, 2008, at 9:15 PM, Edward J. Yoon wrote:
> Hi, communities.
>
> Basically, current Map/Reduce provides parallel computation by row
> splitting. Have anyone thought about 2D blocking technique? If so,
> let's discuss together.
>
>
> --
> B. Regards,
> Edward J. Yoon
Re: 2D blocking technique
Posted by arkady borkovsky <ar...@yahoo-inc.com>.
What exactly do you mean by 2D blocking?
I have a Python class that runs N*M copies of a program with to
inputs, with each task getting 1/N of one input, and 1/M of another.
A "program" can be either Python or any executable.
-- ab
On Apr 20, 2008, at 9:15 PM, Edward J. Yoon wrote:
> Hi, communities.
>
> Basically, current Map/Reduce provides parallel computation by row
> splitting. Have anyone thought about 2D blocking technique? If so,
> let's discuss together.
>
>
> --
> B. Regards,
> Edward J. Yoon