AI

Mashup: building a greedy scheduling algorithm to optimize session times

zacwitte — Fri, 03 Oct 2008 23:48:07 -0500

We know what sessions people are interested in attending, we know approximately when people will be at barcamp. It should be a relatively trivial problem to calculate the optimal schedule to minimize conflicts. This can probably be done using a simple greedy approach. If we can get a drupal guru we can integrate it into the site and make it auto-update every time someone updates their attendance preferences.

Graph Servers, the grey matter of social data

jdavid.net — Thu, 02 Oct 2008 12:42:44 -0500

Graph Servers are a tough thing to build. Linkedin cheats by using 64GB of ram to hold the whole graph on one machine, but at some-point this does not scale. Let's discuss the definition of a graph server and identify some of the problems.

Since everything is going social, even our data, how do we manage massively linked data architectures? How do we search and index them? How does Hadoop and HBase work? Why are RDBMS's a pain in the ass for web scale social data?

Introduction to Hobby Robotics

Royce — Mon, 18 Aug 2008 11:46:24 -0500

Thanks for attending folks!
I am planning on starting robotics club. If you are interested in joining please shoot me a note at royce.pipkins@gmail.com

Graphics, Massive Parallel Processing, Functional and Distributed behavior

jdavid.net — Sun, 17 Aug 2008 16:28:10 -0500

nVidia has been improving its general purpose programing offerings through a C-like language extension that allows you to use your graphics card for more general tasks. The latest graphics cards from nVidia have 240 stream processors to make use of. About 8 months ago we had Micheal from nVidia talking about this. AMD has also decided to adopt CUDA, and drop CTM (close to metal).

If we have the interest we should spend some time hacking on it.

UPDATE: Mac OSX 10.6 will support General Purpose GPU processing in their SDK, it may also be available for iPhone, and if the nVidia Tegra Chip gets into an Android device this stuff may be mobile too.

I have also been looking into how to best use other language patterns for GPU processing like functional patterns, and google's MapReduce. I am no expert at this stuff but if anyone else please join us. I just want to have a session to expose what we know in milwaukee and to build this community.