PyHPC 2012

[Jeffrey Lund]

I just got back from the PyHPC 2012 workshop held in conjunction with SC12, where I presented our paper Mrs: MapReduce for Scientific Computing in Python. I've also attached the slides I used to present the paper. Over all we got some really good and generally very positive feedback about Mrs. Hopefully we can generate some more buzz about the project. It was also really cool to see how many people are having success with Python in high performance computation.

-Jeff

[Andrew McNabb]

On Fri, Nov 16, 2012 at 05:49:05PM -0800, Jeffrey Lund wrote:
> I just got back from the PyHPC 2012 workshop held in conjunction with SC12,
> where I presented our paper Mrs: MapReduce for Scientific Computing in

> Python<http://www.dlr.de/sc/Portaldata/15/Resources/dokumente/pyhpc2012/submissions/pyhpc2012_submission_3.pdf>.

> I've also attached the slides I used to present the paper. Over all we got
> some really good and generally very positive feedback about Mrs. Hopefully
> we can generate some more buzz about the project. It was also really cool
> to see how many people are having success with Python in high performance
> computation.

It sounds like it went really well. Thanks for sharing the slides.


--
Andrew McNabb
http://www.mcnabbs.org/andrew/
PGP Fingerprint: 8A17 B57C 6879 1863 DE55 8012 AB4D 6098 8826 6868

[Jared Forsyth]

Did anyone have any concrete suggestions?
although I guess they didn't know mrs, so it would be hard to be concrete. Any abstract suggestions?

[Jeffrey Lund]

There were a few people wondering how likely it is that tools like HBase, Hive, Mahout, etc will spring up around Mrs. I had to honestly answer that we have no current plans to go re-implement all these tools (and some of them like HBase are clearly not solving the sorts of problems we were talking about in the paper). However, that doesn't mean we wouldn't welcome contributions in this area.