[
https://issues.apache.org/jira/browse/LUCENE-9845?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17302645#comment-17302645
]
Robert Muir commented on LUCENE-9845:
-------------------------------------
OK, but I don't think that is typical: we should optimize for typical cases. It
is of course ok to separately optimize sparse case too, same as DocValues. But
from what I see, the dense case is currently suffering (instantiating large
arrays into RAM, binary searching these arrays, arrays that shouldnt need to be
there...)
> Improve encoding of HNSW graph offsets
> --------------------------------------
>
> Key: LUCENE-9845
> URL: https://issues.apache.org/jira/browse/LUCENE-9845
> Project: Lucene - Core
> Issue Type: Wish
> Reporter: Michael Sokolov
> Priority: Major
>
> Today we use a simple {{long[]}} to encode the offset of each document's
> array of neighbors in the HNSW graph. Instead we should use a data structure
> that is optimized for encoding an increasing numeric array, like monotonic
> {{PackedInts}}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
