[
https://issues.apache.org/jira/browse/LUCENE-7258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15265328#comment-15265328
]
Adrien Grand commented on LUCENE-7258:
--------------------------------------
When I said per thread, I did not mean that the pool needs to be per thread, I
just wanted to highlight that if you want N concurrent requests to all use this
cache, you need a cache of quite a significant size. Sorry for not being
encouraging but I worked on similar caches before and I was not happy with the
end result: it is sometimes hard to know when an instance can actually be
recycled, it is hard to know when the cache should give memory back to the JVM
(especially for a library), the fact that you often end up using oversized
instances encourages to use paging but then things are slower, etc. Relying on
the JVM makes things simpler.
> Tune DocIdSetBuilder allocation rate
> ------------------------------------
>
> Key: LUCENE-7258
> URL: https://issues.apache.org/jira/browse/LUCENE-7258
> Project: Lucene - Core
> Issue Type: Improvement
> Components: modules/spatial
> Reporter: Jeff Wartes
> Attachments:
> LUCENE-7258-Tune-memory-allocation-rate-for-Intersec.patch,
> LUCENE-7258-Tune-memory-allocation-rate-for-Intersec.patch,
> allocation_plot.jpg
>
>
> LUCENE-7211 converted IntersectsPrefixTreeQuery to use DocIdSetBuilder, but
> didn't actually reduce garbage generation for my Solr index.
> Since something like 40% of my garbage (by space) is now attributed to
> DocIdSetBuilder.growBuffer, I charted a few different allocation strategies
> to see if I could tune things more.
> See here: http://i.imgur.com/7sXLAYv.jpg
> The jump-then-flatline at the right would be where DocIdSetBuilder gives up
> and allocates a FixedBitSet for a 100M-doc index. (The 1M-doc index
> curve/cutoff looked similar)
> Perhaps unsurprisingly, the 1/8th growth factor in ArrayUtil.oversize is
> terrible from an allocation standpoint if you're doing a lot of expansions,
> and is especially terrible when used to build a short-lived data structure
> like this one.
> By the time it goes with the FBS, it's allocated around twice as much memory
> for the buffer as it would have needed for just the FBS.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
