Doug Cutting wrote:
Byron Miller wrote:
http://www.cs.yorku.ca/~mladen/pdf/Read6_u.pisa-attardi.tera.pdf
Anyone have any further details on this?
The first author of the paper is also the founder of the company which
sells the software described, so these benchmarks should not be
considered entirely objective.
That's not to say that IXE is not faster than Lucene, it might well
be. But they do not list any JVM details, the Lucene version or any
Lucene options. Chances are, with a few informed tweaks, one could
improve Lucene's performance on this benchmark. Chances are also that
IXE was configured for optimal performance on this benchmark, since it
was performed by the authors of IXE.
Not only that, they mention that each test was run _twice_ to get an
average score. With hotspot JVMs this is ridiculous, you need to run at
least a dozen or more cycles so that the hotspots are recompiled... This
alone discredits the results in my eyes.
--
Best regards,
Andrzej Bialecki <><
___. ___ ___ ___ _ _ __________________________________
[__ || __|__/|__||\/| Information Retrieval, Semantic Web
___|||__|| \| || | Embedded Unix, System Integration
http://www.sigram.com Contact: info at sigram dot com
