[SOLR-9599] DocValues performance regression with new iterator API

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 7.0
• Fix Version/s: 7.0
• Assignee: Unassigned
• Reporter: Yonik Seeley
• Created: 05/Oct/16 00:21
• Updated: 08/Jun/19 15:31

Description

I did a quick performance comparison of faceting indexed fields (i.e. docvalues are not stored) using method=dv before and after the new docvalues iterator went in (LUCENE-7407).

5M document index, 21 segments, single valued string fields w/ no missing values.

field cardinality	new_time / old_time
10	2.01
1000	2.02
10000	1.85
100000	1.56
1000000	1.31

So unfortunately, often twice as slow.

See followup messages for tests using real docvalues as well.

Issue Links

• is broken by: LUCENE-7407 Explore switching doc values to an iterator API
• is related to: LUCENE-8374 Reduce reads for sparse DocValues
• relates to: LUCENE-7462 Faster search APIs for doc values

Sub-Tasks

• Improve faceting performance with FieldCache (single-valued string counts only): Resolved
• Performance regression of numeric field uninversion time: Resolved