lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jake Mannix (JIRA)" <j...@apache.org>
Subject [jira] Commented: (LUCENE-1997) Explore performance of multi-PQ vs single-PQ sorting API
Date Tue, 27 Oct 2009 23:54:59 GMT

    [ https://issues.apache.org/jira/browse/LUCENE-1997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12770708#action_12770708
] 

Jake Mannix commented on LUCENE-1997:
-------------------------------------

||Source||Seg size||Query||Tot hits||Sort||Top N||QPS old||QPS new||Pct change||
|random|balanced|<all>|5000000|rand int|10|22.02|22.99|{color:green}4.4%{color}|
|random|balanced|<all>|5000000|rand int|25|21.45|22.91|{color:green}6.8%{color}|
|random|balanced|<all>|5000000|rand int|50|21.40|22.58|{color:green}5.5%{color}|
|random|balanced|<all>|5000000|rand int|100|21.16|21.93|{color:green}3.6%{color}|
|random|balanced|<all>|5000000|rand int|500|21.27|17.45|{color:red}-18.0%{color}|
|random|balanced|<all>|5000000|rand int|1000|20.62|14.33|{color:red}-30.5%{color}|
|random|balanced|<all>|5000000|rand string|10|18.33|21.91|{color:green}19.5%{color}|
|random|balanced|<all>|5000000|rand string|25|18.37|22.20|{color:green}20.8%{color}|
|random|balanced|<all>|5000000|rand string|50|18.51|22.13|{color:green}19.6%{color}|
|random|balanced|<all>|5000000|rand string|100|20.30|21.91|{color:green}7.9%{color}|
|random|balanced|<all>|5000000|rand string|500|18.34|18.69|{color:green}1.9%{color}|
|random|balanced|<all>|5000000|rand string|1000|17.83|15.04|{color:red}-15.6%{color}|
|random|balanced|<all>|5000000|country|10|18.49|22.16|{color:green}19.8%{color}|
|random|balanced|<all>|5000000|country|25|18.62|22.28|{color:green}19.7%{color}|
|random|balanced|<all>|5000000|country|50|18.30|21.86|{color:green}19.5%{color}|
|random|balanced|<all>|5000000|country|100|20.46|22.08|{color:green}7.9%{color}|
|random|balanced|<all>|5000000|country|500|18.02|18.38|{color:green}2.0%{color}|
|random|balanced|<all>|5000000|country|1000|18.28|14.62|{color:red}-20.0%{color}|
|random|log|<all>|5000000|rand int|10|22.04|23.18|{color:green}5.2%{color}|
|random|log|<all>|5000000|rand int|25|21.48|23.02|{color:green}7.2%{color}|
|random|log|<all>|5000000|rand int|50|21.37|22.96|{color:green}7.4%{color}|
|random|log|<all>|5000000|rand int|100|21.47|22.43|{color:green}4.5%{color}|
|random|log|<all>|5000000|rand int|500|21.09|19.60|{color:red}-7.1%{color}|
|random|log|<all>|5000000|rand int|1000|20.37|17.58|{color:red}-13.7%{color}|
|random|log|<all>|5000000|rand string|10|18.54|22.49|{color:green}21.3%{color}|
|random|log|<all>|5000000|rand string|25|18.50|22.32|{color:green}20.6%{color}|
|random|log|<all>|5000000|rand string|50|18.53|22.24|{color:green}20.0%{color}|
|random|log|<all>|5000000|rand string|100|20.46|22.43|{color:green}9.6%{color}|
|random|log|<all>|5000000|rand string|500|18.59|20.83|{color:green}12.0%{color}|
|random|log|<all>|5000000|rand string|1000|18.27|18.01|{color:red}-1.4%{color}|
|random|log|<all>|5000000|country|10|18.55|22.50|{color:green}21.3%{color}|
|random|log|<all>|5000000|country|25|18.48|22.35|{color:green}20.9%{color}|
|random|log|<all>|5000000|country|50|18.48|22.24|{color:green}20.3%{color}|
|random|log|<all>|5000000|country|100|20.73|22.34|{color:green}7.8%{color}|
|random|log|<all>|5000000|country|500|18.77|20.12|{color:green}7.2%{color}|
|random|log|<all>|5000000|country|1000|17.92|18.44|{color:green}2.9%{color}|


> Explore performance of multi-PQ vs single-PQ sorting API
> --------------------------------------------------------
>
>                 Key: LUCENE-1997
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1997
>             Project: Lucene - Java
>          Issue Type: Improvement
>          Components: Search
>    Affects Versions: 2.9
>            Reporter: Michael McCandless
>            Assignee: Michael McCandless
>         Attachments: LUCENE-1997.patch, LUCENE-1997.patch, LUCENE-1997.patch, LUCENE-1997.patch,
LUCENE-1997.patch, LUCENE-1997.patch, LUCENE-1997.patch
>
>
> Spinoff from recent "lucene 2.9 sorting algorithm" thread on java-dev,
> where a simpler (non-segment-based) comparator API is proposed that
> gathers results into multiple PQs (one per segment) and then merges
> them in the end.
> I started from John's multi-PQ code and worked it into
> contrib/benchmark so that we could run perf tests.  Then I generified
> the Python script I use for running search benchmarks (in
> contrib/benchmark/sortBench.py).
> The script first creates indexes with 1M docs (based on
> SortableSingleDocSource, and based on wikipedia, if available).  Then
> it runs various combinations:
>   * Index with 20 balanced segments vs index with the "normal" log
>     segment size
>   * Queries with different numbers of hits (only for wikipedia index)
>   * Different top N
>   * Different sorts (by title, for wikipedia, and by random string,
>     random int, and country for the random index)
> For each test, 7 search rounds are run and the best QPS is kept.  The
> script runs singlePQ then multiPQ, and records the resulting best QPS
> for each and produces table (in Jira format) as output.

-- 
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: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org


Mime
View raw message