You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@lucene.apache.org by "Adrien Grand (Jira)" <ji...@apache.org> on 2021/08/25 12:29:00 UTC
[jira] [Commented] (LUCENE-10067) investigate 6/23/2021 ->
6/24/2021 drop in facets perf
[ https://issues.apache.org/jira/browse/LUCENE-10067?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17404398#comment-17404398 ]
Adrien Grand commented on LUCENE-10067:
---------------------------------------
I believe it's LUCENE-9613 too since this seems to affect sorting tasks too (in a lesser extent). I don't expect this change to have changed significantly the on-disk format, but it removed some specialization, I'll test if I can get some percents back.
> investigate 6/23/2021 -> 6/24/2021 drop in facets perf
> ------------------------------------------------------
>
> Key: LUCENE-10067
> URL: https://issues.apache.org/jira/browse/LUCENE-10067
> Project: Lucene - Core
> Issue Type: Task
> Reporter: Robert Muir
> Priority: Major
>
> Just looking at perf graph (and recent gains from LUCENE-5309), it is unclear why performance dropped on that date, and if we paved new, better performance, or instead just regained some lost-ground only in the single value case?
> Example:
> https://home.apache.org/~mikemccand/lucenebench/BrowseDayOfYearSSDVFacets.html
> after some debugging, it looks like LUCENE-9613 change may be responsible. It was the only relevant change between the 6/23 and 6/24 benchmarks.
--
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