[jira] [Commented] (LUCENE-9063) Speed up computation of impacts

["ASF subversion and git services (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/LUCENE-9063?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16982348#comment-16982348 ] 

ASF subversion and git services commented on LUCENE-9063:
---------------------------------------------------------

Commit ded8efa82aa30f850f552872daea6036df8d1220 in lucene-solr's branch refs/heads/master from Adrien Grand
[ https://gitbox.apache.org/repos/asf?p=lucene-solr.git;h=ded8efa ]

LUCENE-9063: Speed up computation of impacts. (#1038)

The current design of CompetitiveImpactAccumulator treats norms in -128..127
as a special case that should be optimized. This commit goes a bit further by
treating it as the normal case, and only ever adding impacts to the TreeSet if
the norm is outside of the byte range. It avoids a number of operations on
TreeSets like adding impacts or removing redundant impacts.


> Speed up computation of impacts
> -------------------------------
>
>                 Key: LUCENE-9063
>                 URL: https://issues.apache.org/jira/browse/LUCENE-9063
>             Project: Lucene - Core
>          Issue Type: Improvement
>            Reporter: Adrien Grand
>            Priority: Minor
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> We have a regression in indexing speed since January 23th because we started computing and indexing impacts: http://people.apache.org/~mikemccand/lucenebench/indexing.html. While the search benefits are probably worth the slow down, maybe we can improve the way impacts are computed?



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

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