Re: Reasoning over millions of triples - any working reasoners?

On 2011-01-19, Ivan Mikhailov wrote:

> Virtuoso deals with owl:sameAs in a scalable way, so you can try. Of 
> course, a single chain 50 million connections long would cause 
> problems, but more traditional cases should work fine. Google for 
> "virtuoso owl:same-as input:inference" may be the fastest way to get 
> more hints.

To amplify my previous point... Even Knuth's bible contains the basic 
sequential union-tree algorithm, which solves the basic, underlying 
problem (e.g. http://en.wikipedia.org/wiki/Disjoint-set_data_structure 
etc.) in amortized linear time. The tree-like update algorithms for that 
also parallelize pretty well.

A mere fifty million indices mean nothing when you mind your algos, and 
especially if you really utilize the current parallel hardware we have 
to its fullest. It'd actually be a shame if you couldn't sustain 50M 
such arbitrary reductions per hour, guaranteed, full-time, on a single 
Duo, even using something much less efficient than well-hand-optimized C 
and/or assembly. Of course provided that you could get the data on and 
off the chip in time, given your chosen encoding.
-- 
Sampo Syreeni, aka decoy - decoy@iki.fi, http://decoy.iki.fi/front
+358-50-5756111, 025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2

Received on Friday, 21 January 2011 22:20:20 UTC