[jira] [Updated] (LUCENE-7101) OfflineSorter's merging is O(N^2) cost for large sorts

["Hoss Man (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/LUCENE-7101?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Hoss Man updated LUCENE-7101:
-----------------------------
    Fix Version/s: master (7.0)
                   6.1

Manually correcting fixVersion per Step #S6 of LUCENE-7271


> OfflineSorter's merging is O(N^2) cost for large sorts
> ------------------------------------------------------
>
>                 Key: LUCENE-7101
>                 URL: https://issues.apache.org/jira/browse/LUCENE-7101
>             Project: Lucene - Core
>          Issue Type: Bug
>            Reporter: Michael McCandless
>            Assignee: Michael McCandless
>            Priority: Blocker
>             Fix For: 6.0, 6.1, master (7.0)
>
>         Attachments: LUCENE-7101.patch
>
>
> Our {{OfflineSorter}} acts just like Lucene, writing small initial
> segments of sorted values (from what it was able to sort at once in
> heap), periodically merging them when there are too many, and doing a
> {{forceMerge(1)}} in the end.
> But the merge logic is too simplistic today, resulting in O(N^2)
> cost.  Smallish sorts usually won't hit it, because the default 128
> merge factor is so high, but e.g. the new 2B points tests do hit the
> N^2 behavior.  I suspect the high merge factor hurts performance (OS
> struggles to use what free RAM it has to read-ahead on 128 files), and
> also risks file descriptor exhaustion.
> I think we should implement a simple log merge policy for it, and drop
> its default merge factor to 10.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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