W3C home > Mailing lists > Public > public-cwm-talk@w3.org > October to December 2007

improving cwm performance

From: Yosi Scharf <syosi@MIT.EDU>
Date: Sat, 17 Nov 2007 23:13:31 -0500
Message-ID: <473FBBEB.7040609@mit.edu>
To: public-cwm-talk@w3.org

I was researching rule engines, are I realized that I might be able to
improve cwm's performance in some cases tremendously without too much
effort, but insuring that only new triples are matched.

as an example:


(using http://www.w3.org/2000/10/swap/cwm-1.2a2.tar.gz )
syosi@YOSI:~/tarballs/cwm-1.2.0a2$ time ./cwm
http://www.w3.org/2000/10/swap/test/reason/longChain.n3 --think

real    7m44.575s
user    6m15.083s
sys     0m3.728s


compare that to the following

(using http://www.w3.org/2000/10/swap/cwm-1.2a3.tar.gz , which was made from
the above tarball by changing query.py and llyn.py) This passes all
tests that the other one passes, which make me feel confidant.

syosi@YOSI:~/tarballs/cwm-1.2.0a3$ time ./cwm
http://www.w3.org/2000/10/swap/test/reason/longChain.n3 --think

real    0m24.497s
user    0m21.757s
sys     0m0.264s


This was an extreme example, but did this change break anything? Is it
useful?


Yosi
Received on Sunday, 18 November 2007 04:13:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:11:02 GMT