lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Doug Cutting <>
Subject Re: Memory Usage
Date Fri, 18 Nov 2005 00:25:37 GMT
Daniel Noll wrote:
> Doug Cutting wrote:
>> Daniel Noll wrote:
>>> I actually did throw a lot of terms in, and eventually chose "one" 
>>> for the tests because it was the slowest query to complete of them 
>>> all (hence I figured it was already spending some fairly long time in 
>>> I/O, and would be penalised the most.)  Every other query was around 
>>> 7ms before tweaking, and the tweak increased them all to somewhere 
>>> around 10ms but that's still a lot faster than "one" was even at its 
>>> fastest.
>> Different terms are affected differently by this tweak, so results for 
>> a single term don't reveal much.
> Hence why I just said: "I actually did throw a lot of terms in".

I'm confused.  I'm talking about the table of results you posted.  You 
measured only a single term in those, no?  I don't think results for a 
single term are representative.  If it is term 127 then it will be 
slower than average with an indexInterval of 128, and will be faster 
than average with an indexInterval of 16k.  So to get numbers that are 
representative of an average term at multiple indexIntervals you need to 
test with many terms.  Am I missing something?

Choosing a term that's slower to search actually de-emphasizes the 
effect of this change, since its query time is dominated not by term 
lookup but by posting processing.


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message