- From: Sean B. Palmer <sean@mysterylights.com>
- Date: Sat, 23 Feb 2002 23:39:20 -0000
- To: "Tim Berners-Lee" <timbl@w3.org>, "Norman Walsh" <ndw@nwalsh.com>
- Cc: "Dan Connolly" <connolly@w3.org>, <www-archive+n3bugs@w3.org>
[cc.remove('rdfig'); cc.append('www-archive')]
> So it I would expect it to be slower ... but four
> times is tough.
One test that I just ran was a full eleven times slower under CWM 1.87
than it was under CWM 1.82. All of the test data is on the Web [but
the test data is RegExped from another site, so use at your own risk].
The command line:-
cwm data.n3 --filter=q.n3 > out.n3
data.n3: http://infomesh.net/2001/12/map/data.n3
q.n3: http://infomesh.net/2001/12/map/q.n3
Time with 1.87: 4:08
Time with 1.82: 0:22
The simple test was actually to benchmark other N3 parsers/inference
engines, but I was surprised that the latest CWM took so long, and
slurped the 1.82 version off of the Web. The full tests are at [1] and
[2].
[1] $ u timenow; cwm data.n3 --filter=q.n3 > q-out.n3; u timenow
20020223-231747
20020223-232155
[2] $ u timenow; python cwm/cwm.py data.n3 --filter=q.n3 > q-out.n3; u
timenow
20020223-232416
20020223-232438
--
Kindest Regards,
Sean B. Palmer
@prefix : <http://purl.org/net/swn#> .
:Sean :homepage <http://purl.org/net/sbp/> .
Received on Saturday, 23 February 2002 18:41:06 UTC