lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mark Miller (JIRA)" <j...@apache.org>
Subject [jira] Commented: (LUCENE-1997) Explore performance of multi-PQ vs single-PQ sorting API
Date Fri, 23 Oct 2009 19:05:59 GMT

    [ https://issues.apache.org/jira/browse/LUCENE-1997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12769365#action_12769365
] 

Mark Miller commented on LUCENE-1997:
-------------------------------------

JAVA:
java version "1.5.0_20"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0_20-b02)
Java HotSpot(TM) 64-Bit Server VM (build 1.5.0_20-b02, mixed mode)

OS:
Linux quad-laptop 2.6.31-14-generic #48-Ubuntu SMP  x86_64 GNU/Linux


||Source||Seg size||Query||Tot hits||Sort||Top N||QPS old||QPS new||Pct change||
|wiki|log|1|317914|title|10|80.09|78.20|{color:red}-2.4%{color}|
|wiki|log|1|317914|title|25|80.12|79.51|{color:red}-0.8%{color}|
|wiki|log|1|317914|title|50|78.61|76.03|{color:red}-3.3%{color}|
|wiki|log|1|317914|title|100|77.18|75.13|{color:red}-2.7%{color}|
|wiki|log|1|317914|title|500|75.01|54.74|{color:red}-27.0%{color}|
|wiki|log|1|317914|title|1000|67.77|41.29|{color:red}-39.1%{color}|
|wiki|log|<all>|1000000|title|10|109.30|119.29|{color:green}9.1%{color}|
|wiki|log|<all>|1000000|title|25|108.34|116.02|{color:green}7.1%{color}|
|wiki|log|<all>|1000000|title|50|106.86|110.70|{color:green}3.6%{color}|
|wiki|log|<all>|1000000|title|100|94.72|101.10|{color:green}6.7%{color}|
|wiki|log|<all>|1000000|title|500|78.69|62.04|{color:red}-21.2%{color}|
|wiki|log|<all>|1000000|title|1000|71.93|43.05|{color:red}-40.2%{color}|
|random|log|<all>|1000000|rand string|10|112.81|117.80|{color:green}4.4%{color}|
|random|log|<all>|1000000|rand string|25|113.92|115.73|{color:green}1.6%{color}|
|random|log|<all>|1000000|rand string|50|113.55|110.08|{color:red}-3.1%{color}|
|random|log|<all>|1000000|rand string|100|90.30|95.35|{color:green}5.6%{color}|
|random|log|<all>|1000000|rand string|500|76.77|51.88|{color:red}-32.4%{color}|
|random|log|<all>|1000000|rand string|1000|66.78|36.93|{color:red}-44.7%{color}|
|random|log|<all>|1000000|country|10|114.26|118.72|{color:green}3.9%{color}|
|random|log|<all>|1000000|country|25|113.96|115.81|{color:green}1.6%{color}|
|random|log|<all>|1000000|country|50|113.59|109.78|{color:red}-3.4%{color}|
|random|log|<all>|1000000|country|100|91.97|94.05|{color:green}2.3%{color}|
|random|log|<all>|1000000|country|500|75.03|27.37|{color:red}-63.5%{color}|
|random|log|<all>|1000000|country|1000|66.62|34.85|{color:red}-47.7%{color}|
|random|log|<all>|1000000|rand int|10|118.06|124.42|{color:green}5.4%{color}|
|random|log|<all>|1000000|rand int|25|117.76|120.76|{color:green}2.5%{color}|
|random|log|<all>|1000000|rand int|50|117.35|115.81|{color:red}-1.3%{color}|
|random|log|<all>|1000000|rand int|100|96.35|103.60|{color:green}7.5%{color}|
|random|log|<all>|1000000|rand int|500|87.04|50.63|{color:red}-41.8%{color}|
|random|log|<all>|1000000|rand int|1000|77.59|35.47|{color:red}-54.3%{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
>
>
> 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