[
https://issues.apache.org/jira/browse/LUCENE-8753?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16813164#comment-16813164
]
Adrien Grand commented on LUCENE-8753:
--------------------------------------
bq. BlockTree and UniformSplit had the same QPS for Term and Phrase queries. I
didn't understand why a different behavior between a small and a large index.
I think this is expected. Query processing needs to look up the term in the
terms dict and then process documents that contain this term. When the index
gets larger, postings usually grow more quickly than the terms dictionary, so
processing postings takes more time relatively compared to looking up the term
in the terms dictionary. Term dictionary lookup performance only really matters
for queries that have few matches (which you somehow simulated by running the
benchmark on wikimedium500k) and updates, which are simulated by the PKLookup
task.
> New PostingFormat - UniformSplit
> --------------------------------
>
> Key: LUCENE-8753
> URL: https://issues.apache.org/jira/browse/LUCENE-8753
> Project: Lucene - Core
> Issue Type: Improvement
> Components: core/codecs
> Affects Versions: 8.0
> Reporter: Bruno Roustant
> Priority: Major
> Attachments: Uniform Split Technique.pdf, luceneutil.benchmark.txt
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> This is a proposal to add a new PostingsFormat called "UniformSplit" with 4
> objectives:
> - Clear design and simple code.
> - Easily extensible, for both the logic and the index format.
> - Light memory usage with a very compact FST.
> - Focus on efficient TermQuery, PhraseQuery and PrefixQuery performance.
> (the pdf attached explains visually the technique in more details)
> The principle is to split the list of terms into blocks and use a FST to
> access the block, but not as a prefix trie, rather with a seek-floor pattern.
> For the selection of the blocks, there is a target average block size (number
> of terms), with an allowed delta variation (10%) to compare the terms and
> select the one with the minimal distinguishing prefix.
> There are also several optimizations inside the block to make it more
> compact and speed up the loading/scanning.
> The performance obtained is interesting with the luceneutil benchmark,
> comparing UniformSplit with BlockTree. Find it in the first comment and also
> attached for better formatting.
> Although the precise percentages vary between runs, three main points:
> - TermQuery and PhraseQuery are improved.
> - PrefixQuery and WildcardQuery are ok.
> - Fuzzy queries are clearly less performant, because BlockTree is so
> optimized for them.
> Compared to BlockTree, FST size is reduced by 15%, and segment writing time
> is reduced by 20%. So this PostingsFormat scales to lots of docs, as
> BlockTree.
> This initial version passes all Lucene tests. Use “ant test
> -Dtests.codec=UniformSplitTesting” to test with this PostingsFormat.
> Subjectively, we think we have fulfilled our goal of code simplicity. And we
> have already exercised this PostingsFormat extensibility to create a
> different flavor for our own use-case.
> Contributors: Juan Camilo Rodriguez Duran, Bruno Roustant, David Smiley
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
