[
https://issues.apache.org/jira/browse/LUCENE-1476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12662347#action_12662347
]
Michael McCandless commented on LUCENE-1476:
--------------------------------------------
{quote}
> Under the current system, I'm not certain that the deletions checks are that
> excessive.
{quote}
I made a simple test that up-front converted the deleted docs
BitVector into an int[] containing the deleted docsIDs, and then did
the same nextDeletedDoc change to SegmentTermDocs.
This was only faster if < 10% of docs were deleted, though I didn't do
very thorough testing.
I think the problem is with this change the CPU must predict a new
branch (docNum >= nextDeletion) vs doing a no-branching lookup of the
bit.
> BitVector implement DocIdSet
> ----------------------------
>
> Key: LUCENE-1476
> URL: https://issues.apache.org/jira/browse/LUCENE-1476
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Affects Versions: 2.4
> Reporter: Jason Rutherglen
> Priority: Trivial
> Attachments: LUCENE-1476.patch, quasi_iterator_deletions.diff
>
> Original Estimate: 12h
> Remaining Estimate: 12h
>
> BitVector can implement DocIdSet. This is for making
> SegmentReader.deletedDocs pluggable.
--
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-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-dev-h...@lucene.apache.org
