[
https://issues.apache.org/jira/browse/LUCENE-4923?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Robert Muir updated LUCENE-4923:
--------------------------------
Attachment: LUCENE-4923.patch
Here's a patch. Just cleaning house on the minShouldMatch gives about a 10%
improvement on luceneutil, but supporting pure disjunctions as a top-level
scorer from BooleanWeight is even better (20%):
{noformat}
Task QPS trunk StdDev QPS patch StdDev
Pct diff
OrHighHigh 22.70 (6.5%) 27.45 (4.0%)
20.9% ( 9% - 33%)
OrHighMed 31.17 (6.7%) 38.06 (4.6%)
22.1% ( 10% - 35%)
OrHighLow 71.52 (7.1%) 87.48 (5.5%)
22.3% ( 9% - 37%)
{noformat}
> remove minShouldMatch/speed up DisjunctionSumScorer
> ---------------------------------------------------
>
> Key: LUCENE-4923
> URL: https://issues.apache.org/jira/browse/LUCENE-4923
> Project: Lucene - Core
> Issue Type: Improvement
> Components: core/query/scoring
> Affects Versions: 4.3
> Reporter: Robert Muir
> Attachments: LUCENE-4923.patch
>
>
> LUCENE-4571 added a MinShouldMatchScorer: capable of using advance() on
> subscorers for minShouldMatch > 1.
> However, we didn't yet cleanup the (now dead) minShouldMatch logic from
> DisjunctionSumScorer: which is now only used for pure disjunctions. This can
> also give some reasonable performance improvements for in-order collection.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
