lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Grant Ingersoll <gsing...@apache.org>
Subject Re: Payload Loading and Reloading
Date Fri, 30 Nov 2007 01:36:27 GMT
Good points, like I said, I will look more into caching in the Near  
Spans.  I need to profile them some anyway, as I am hoping there is  
some speedup to be had there.

-Grant

On Nov 29, 2007, at 6:23 PM, Michael Busch wrote:

> Grant Ingersoll wrote:
>
>>
>> As for the cost of the seeks, why can't we just document that this is
>> what is going on and discourage people from doing it?
>
> I'm just trying to keep SegmentTermPositions#getPayload() as efficient
> as possible because it's often used in the most inner loops of  
> scorers.
> Keeping track of the start position of the payload adds overhead to  
> that
> method. I didn't profile it, maybe the overhead is not significant,  
> but
> I still don't see any usecase in which caching the latest payload is  
> not
> possible/sufficient. But maybe I'm missing one?
>
>
>> I think caching it in SegTermPos. is the simplest
>
> I don't think we should do caching here. Again, SegmentTermPositions  
> is
> used in inner loops, and often caching the last payload is not needed,
> e. g. in BoostingTermQuery.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-dev-help@lucene.apache.org
>



---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org


Mime
View raw message