[
https://issues.apache.org/jira/browse/SOLR-13132?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17154192#comment-17154192
]
Chris M. Hostetter commented on SOLR-13132:
-------------------------------------------
Hey Michael, just to sanity check one thing...
{quote}The results below are mostly for without any filterCache,...
{quote}
Are the {{MASTER}} results you reported here from testing with
"filterCacheSize=0" ... which would make them an apples to apples comparison
with your {{"SOLR-13132 sweep_collection=false, filterCacheSize=0}} " results?
(i'm guessing so based on your comment above, but i don't see any sort of log
in your tgz showing exactly what commands were used to run each test, and i see
a default size of 4096 in your solrconfig.xml, so i wanted to be certain what
you ment)
Assuming i'm understanding correctly, then what i see here is what we expected:
for the "common" case of sorting on {{relatedness()}} (high cardinality fields
– larger then filterCache size – w/ non-trivially FG sets) sweeping is big
improvement, and in the _uncommon_ case of low cardinality fields and/or small
FG sets there is only a small (negative) impact of using the new code – and
that seems to go away (numbers close enough to be noise) by specyifing
"{{sweep_collection:false}} ".
sound right?
So i think we're good to go – we just need the corrections/updates to the ref
guide. If you can please remove the incorrect stuff i mentioned in my last
comment, and replace it with a note explaining when/why people _may_ want to
consider use {{sweep_collection: false}} (ie: atypical SKG situations where
cardinality or FG set size is very low) I'll try to merge & backport ASAP.
> Improve JSON "terms" facet performance when sorted by relatedness
> ------------------------------------------------------------------
>
> Key: SOLR-13132
> URL: https://issues.apache.org/jira/browse/SOLR-13132
> Project: Solr
> Issue Type: Improvement
> Components: Facet Module
> Affects Versions: 7.4, master (9.0)
> Reporter: Michael Gibney
> Priority: Major
> Attachments: SOLR-13132-benchmarks.tgz,
> SOLR-13132-with-cache-01.patch, SOLR-13132-with-cache.patch,
> SOLR-13132.patch, SOLR-13132_testSweep.patch
>
> Time Spent: 1.5h
> Remaining Estimate: 0h
>
> When sorting buckets by {{relatedness}}, JSON "terms" facet must calculate
> {{relatedness}} for every term.
> The current implementation uses a standard uninverted approach (either
> {{docValues}} or {{UnInvertedField}}) to get facet counts over the domain
> base docSet, and then uses that initial pass as a pre-filter for a
> second-pass, inverted approach of fetching docSets for each relevant term
> (i.e., {{count > minCount}}?) and calculating intersection size of those sets
> with the domain base docSet.
> Over high-cardinality fields, the overhead of per-term docSet creation and
> set intersection operations increases request latency to the point where
> relatedness sort may not be usable in practice (for my use case, even after
> applying the patch for SOLR-13108, for a field with ~220k unique terms per
> core, QTime for high-cardinality domain docSets were, e.g.: cardinality
> 1816684=9000ms, cardinality 5032902=18000ms).
> The attached patch brings the above example QTimes down to a manageable
> ~300ms and ~250ms respectively. The approach calculates uninverted facet
> counts over domain base, foreground, and background docSets in parallel in a
> single pass. This allows us to take advantage of the efficiencies built into
> the standard uninverted {{FacetFieldProcessorByArray[DV|UIF]}}), and avoids
> the per-term docSet creation and set intersection overhead.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
