[
https://issues.apache.org/jira/browse/LUCENE-9450?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17180860#comment-17180860
]
Gautam Worah commented on LUCENE-9450:
--------------------------------------
Benchmark command with the same localrun.py script:
{code:java}
python src/python/localrun.py -source wikimediumall -forceMerge=True
{code}
Instance details:
EC2 c4.8xlarge
CPU (/proc/cpuinfo):
{{Architecture: x86_64}}
{{CPU op-mode(s): 32-bit, 64-bit}}
{{Byte Order: Little Endian}}
{{CPU(s): 36}}
{{On-line CPU(s) list: 0-35}}
{{Thread(s) per core: 2}}
Memory:
{{MemTotal: 61833408 kB}}
{{MemFree: 1585956 kB}}
{{MemAvailable: 53673248 kB}}
Java:
{{openjdk version "11.0.8" 2020-07-14 LTS}}
{{OpenJDK Runtime Environment (build 11.0.8+10-LTS)}}
{{OpenJDK 64-Bit Server VM (build 11.0.8+10-LTS, mixed mode)}}
===
Results:
{code:java}
TaskQPS baseline StdDevQPS my_modified_version
StdDev Pct diff
OrNotHighHigh 532.96 (6.0%) 518.91 (8.1%)
-2.6% ( -15% - 12%)
IntNRQ 63.90 (2.8%) 62.51 (3.7%)
-2.2% ( -8% - 4%)
LowTerm 1354.18 (4.9%) 1325.24 (7.1%)
-2.1% ( -13% - 10%)
PKLookup 130.95 (2.3%) 128.48 (3.1%)
-1.9% ( -7% - 3%)
Fuzzy1 49.16 (6.3%) 48.28 (9.8%)
-1.8% ( -16% - 15%)
OrHighNotMed 518.20 (4.8%) 509.19 (5.4%)
-1.7% ( -11% - 8%)
Wildcard 53.04 (2.1%) 52.14 (3.6%)
-1.7% ( -7% - 4%)
OrHighLow 241.89 (3.8%) 237.82 (4.7%)
-1.7% ( -9% - 7%)
MedTerm 1002.39 (5.2%) 985.94 (5.8%)
-1.6% ( -12% - 9%)
AndHighLow 453.14 (3.3%) 445.71 (5.9%)
-1.6% ( -10% - 7%)
HighTermMonthSort 22.03 (10.5%) 21.70 (10.3%)
-1.5% ( -20% - 21%)
MedPhrase 228.73 (3.2%) 225.69 (4.6%)
-1.3% ( -8% - 6%)
OrHighNotLow 461.58 (4.8%) 456.17 (6.9%)
-1.2% ( -12% - 11%)
HighTermTitleBDVSort 64.10 (10.5%) 63.35 (9.7%)
-1.2% ( -19% - 21%)
LowPhrase 134.07 (2.1%) 132.72 (2.8%)
-1.0% ( -5% - 3%)
OrHighMed 56.83 (2.9%) 56.44 (3.7%)
-0.7% ( -7% - 6%)
Prefix3 31.60 (2.1%) 31.42 (2.4%)
-0.6% ( -4% - 4%)
AndHighMed 61.02 (2.7%) 60.72 (3.2%)
-0.5% ( -6% - 5%)
HighPhrase 46.05 (3.2%) 45.84 (4.7%)
-0.5% ( -8% - 7%)
LowSpanNear 40.71 (2.4%) 40.54 (2.2%)
-0.4% ( -4% - 4%)
Fuzzy2 37.09 (7.5%) 36.94 (6.0%)
-0.4% ( -12% - 14%)
Respell 28.71 (1.5%) 28.60 (2.0%)
-0.4% ( -3% - 3%)
OrNotHighMed 375.44 (2.9%) 374.10 (5.0%)
-0.4% ( -8% - 7%)
AndHighHigh 32.52 (3.2%) 32.46 (3.3%)
-0.2% ( -6% - 6%)
HighSpanNear 9.05 (3.3%) 9.03 (3.6%)
-0.2% ( -6% - 6%)
MedSloppyPhrase 61.05 (1.8%) 61.03 (2.5%)
-0.0% ( -4% - 4%)
BrowseDayOfYearSSDVFacets 2.76 (0.8%) 2.75 (1.0%)
-0.0% ( -1% - 1%)
HighTerm 917.59 (4.6%) 917.50 (6.4%)
-0.0% ( -10% - 11%)
HighIntervalsOrdered 4.32 (2.5%) 4.32 (2.8%)
0.0% ( -5% - 5%)
BrowseMonthSSDVFacets 2.94 (0.6%) 2.95 (0.6%)
0.1% ( -1% - 1%)
OrHighHigh 6.22 (3.0%) 6.24 (3.3%)
0.2% ( -5% - 6%)
OrHighNotHigh 425.45 (5.3%) 426.34 (6.8%)
0.2% ( -11% - 13%)
LowSloppyPhrase 7.30 (2.6%) 7.32 (2.5%)
0.4% ( -4% - 5%)
HighSloppyPhrase 7.52 (2.1%) 7.55 (2.8%)
0.4% ( -4% - 5%)
MedSpanNear 11.22 (3.1%) 11.26 (3.0%)
0.4% ( -5% - 6%)
OrNotHighLow 367.16 (3.6%) 368.76 (4.8%)
0.4% ( -7% - 9%)
HighTermDayOfYearSort 35.05 (8.3%) 35.47 (9.1%)
1.2% ( -14% - 20%)
BrowseMonthTaxoFacets 1.39 (2.6%) 1.44 (7.3%)
4.0% ( -5% - 14%)
BrowseDateTaxoFacets 1.21 (2.4%) 1.26 (6.7%)
4.2% ( -4% - 13%)
BrowseDayOfYearTaxoFacets 1.20 (2.8%) 1.26 (6.9%)
4.7% ( -4% - 14%)
{code}
> Taxonomy index should use DocValues not StoredFields
> ----------------------------------------------------
>
> Key: LUCENE-9450
> URL: https://issues.apache.org/jira/browse/LUCENE-9450
> Project: Lucene - Core
> Issue Type: Improvement
> Components: modules/facet
> Affects Versions: 8.5.2
> Reporter: Gautam Worah
> Priority: Minor
> Labels: performance
> Attachments: LUCENE-9450-localrun.py-v1, wip_taxonomy_patch
>
> Time Spent: 1.5h
> Remaining Estimate: 0h
>
> The taxonomy index that maps binning labels to ordinals was created before
> Lucene added BinaryDocValues.
> I've attached a WIP patch (does not pass tests currently)
> Issue suggested by [~mikemccand]
--
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
