[
https://issues.apache.org/jira/browse/LUCENE-2694?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12980921#action_12980921
]
Michael McCandless commented on LUCENE-2694:
--------------------------------------------
Actually I see PK lookups faster -- 23 usec w/ patch vs 33 usec w/ trunk (per
lookup) for 20K lookups.
And good speedups on many-term MTQs when I force BQ rewrite:
||Query||QPS base||QPS termstate||Pct diff||||
|+nebraska +state|169.75|154.64|{color:red}-8.9%{color}|
|doctitle:.*[Uu]nited.*|4.26|4.11|{color:red}-3.5%{color}|
|+unit +state|11.40|11.09|{color:red}-2.7%{color}|
|spanFirst(unit, 5)|17.38|16.93|{color:red}-2.6%{color}|
|spanNear([unit, state], 10, true)|4.37|4.32|{color:red}-1.2%{color}|
|"unit state"~3|4.94|4.89|{color:red}-1.0%{color}|
|"unit state"|8.05|8.03|{color:red}-0.2%{color}|
|state|26.58|26.76|{color:green}0.7%{color}|
|unit state|11.24|11.46|{color:green}1.9%{color}|
|united~2.0|3.87|3.98|{color:green}2.8%{color}|
|doctimesecnum:[10000 TO 60000]|8.26|8.70|{color:green}5.3%{color}|
|unit~2.0|10.04|10.59|{color:green}5.4%{color}|
|united~1.0|16.84|18.13|{color:green}7.7%{color}|
|unit~1.0|10.09|10.99|{color:green}8.9%{color}|
|un*d|11.96|21.63|{color:green}80.8%{color}|
|unit*|7.60|14.23|{color:green}87.3%{color}|
|u*d|2.22|4.17|{color:green}87.8%{color}|
|uni*|1.83|3.53|{color:green}93.7%{color}|
+1 to commit!
> MTQ rewrite + weight/scorer init should be single pass
> ------------------------------------------------------
>
> Key: LUCENE-2694
> URL: https://issues.apache.org/jira/browse/LUCENE-2694
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Search
> Reporter: Michael McCandless
> Assignee: Simon Willnauer
> Fix For: 4.0
>
> Attachments: LUCENE-2694-FTE.patch, LUCENE-2694.patch,
> LUCENE-2694.patch, LUCENE-2694.patch, LUCENE-2694.patch, LUCENE-2694.patch,
> LUCENE-2694.patch, LUCENE-2694.patch, LUCENE-2694.patch, LUCENE-2694.patch,
> LUCENE-2694.patch, LUCENE-2694.patch, LUCENE-2694_hack.patch
>
>
> Spinoff of LUCENE-2690 (see the hacked patch on that issue)...
> Once we fix MTQ rewrite to be per-segment, we should take it further and make
> weight/scorer init also run in the same single pass as rewrite.
--
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: [email protected]
For additional commands, e-mail: [email protected]
