lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Michael McCandless (JIRA)" <>
Subject [jira] Commented: (LUCENE-1593) Optimizations to TopScoreDocCollector and TopFieldCollector
Date Tue, 14 Apr 2009 09:46:15 GMT


Michael McCandless commented on LUCENE-1593:

bq. I thought these were introduced as part of LUCENE-1483 for performance reasons. Are we
sure that removing these, just for the sake of what looks like minor optimizations to TSDC
and TFC (not breaking a tie on doc Id) is worth it?

Right, this was added in LUCENE-1483, and I think we should remove it, ie IndexSearcher will
always visit docIDs in order again.

The optimization was done for the StringValCollector: that collector "in general" has better
performance if you give it the biggest segments first, since the queue should converge faster
and require less fallback to by-value comparisons.  But I believe that is a minor optimization
in general since segments tend to be biggest to smallest, already, except for indexes with
many deletes on the older segments.

My feeling is we can proceed without testing it; I believe this "not breaking tie by docID"
will be a bigger win.

> Optimizations to TopScoreDocCollector and TopFieldCollector
> -----------------------------------------------------------
>                 Key: LUCENE-1593
>                 URL:
>             Project: Lucene - Java
>          Issue Type: Improvement
>          Components: Search
>            Reporter: Shai Erera
>             Fix For: 2.9
> This is a spin-off of LUCENE-1575 and proposes to optimize TSDC and TFC code to remove
unnecessary checks. The plan is:
> # Ensure that IndexSearcher returns segements in increasing doc Id order, instead of
> # Change TSDC and TFC's code to not use the doc id as a tie breaker. New docs will always
have larger ids and therefore cannot compete.
> # Pre-populate HitQueue with sentinel values in TSDC (score = Float.NEG_INF) and remove
the check if reusableSD == null.
> # Also move to use "changing top" and then call adjustTop(), in case we update the queue.
> # some methods in Sort explicitly add SortField.FIELD_DOC as a "tie breaker" for the
last SortField. But, doing so should not be necessary (since we already break ties by docID),
and is in fact less efficient (once the above optimization is in).
> # Investigate PQ - can we deprecate insert() and have only insertWithOverflow()? Add
a addDummyObjects method which will populate the queue without "arranging" it, just store
the objects in the array (this can be used to pre-populate sentinel values)?
> I will post a patch as well as some perf measurements as soon as I have them.

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

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

View raw message