jpountz commented on PR #12457:
URL: https://github.com/apache/lucene/pull/12457#issuecomment-1649684864
Another run that used a different set of queries showed a speedup on
`Fuzzy1`:
```
TaskQPS baseline StdDevQPS
my_modified_version StdDev Pct diff p-value
BrowseDateSSDVFacets 5.60 (5.8%) 5.39
(7.9%) -3.8% ( -16% - 10%) 0.082
OrHighHigh 49.87 (9.9%) 48.16
(8.8%) -3.4% ( -20% - 16%) 0.248
BrowseDateTaxoFacets 42.70 (3.9%) 42.01
(12.5%) -1.6% ( -17% - 15%) 0.580
OrHighLow 421.47 (6.8%) 415.23
(5.7%) -1.5% ( -13% - 11%) 0.455
BrowseDayOfYearTaxoFacets 43.24 (4.7%) 42.63
(12.9%) -1.4% ( -18% - 17%) 0.649
HighSloppyPhrase 60.09 (4.8%) 59.44
(6.4%) -1.1% ( -11% - 10%) 0.548
BrowseMonthTaxoFacets 28.42 (10.8%) 28.15
(12.8%) -1.0% ( -22% - 25%) 0.799
BrowseRandomLabelTaxoFacets 35.58 (7.1%) 35.24
(12.2%) -0.9% ( -18% - 19%) 0.766
HighTermTitleSort 136.77 (7.9%) 135.49
(5.5%) -0.9% ( -13% - 13%) 0.664
OrHighMed 107.12 (6.1%) 106.39
(6.1%) -0.7% ( -12% - 12%) 0.723
HighTermMonthSort 3320.49 (5.6%) 3300.19
(4.6%) -0.6% ( -10% - 10%) 0.706
HighTerm 507.30 (9.6%) 504.23
(9.1%) -0.6% ( -17% - 20%) 0.838
AndHighHigh 119.96 (5.8%) 119.28
(5.3%) -0.6% ( -10% - 11%) 0.745
BrowseRandomLabelSSDVFacets 14.70 (7.8%) 14.62
(5.6%) -0.6% ( -12% - 13%) 0.794
PKLookup 245.79 (3.0%) 244.68
(3.7%) -0.5% ( -7% - 6%) 0.673
Respell 73.53 (1.3%) 73.26
(1.4%) -0.4% ( -2% - 2%) 0.397
OrHighNotLow 361.98 (9.6%) 360.81
(9.8%) -0.3% ( -17% - 21%) 0.916
HighTermDayOfYearSort 435.58 (3.2%) 434.35
(2.5%) -0.3% ( -5% - 5%) 0.754
TermDTSort 179.69 (2.7%) 179.21
(1.4%) -0.3% ( -4% - 3%) 0.690
BrowseDayOfYearSSDVFacets 19.83 (8.2%) 19.78
(8.7%) -0.3% ( -15% - 18%) 0.924
AndHighMed 305.45 (5.2%) 304.82
(4.8%) -0.2% ( -9% - 10%) 0.896
MedSloppyPhrase 115.76 (4.3%) 115.56
(4.6%) -0.2% ( -8% - 9%) 0.902
OrNotHighMed 440.27 (5.4%) 439.70
(5.8%) -0.1% ( -10% - 11%) 0.942
IntNRQ 124.96 (10.6%) 124.83
(10.9%) -0.1% ( -19% - 23%) 0.976
MedTerm 687.29 (8.9%) 687.53
(9.4%) 0.0% ( -16% - 20%) 0.991
Wildcard 105.26 (2.6%) 105.32
(2.6%) 0.0% ( -4% - 5%) 0.952
OrNotHighHigh 473.56 (7.0%) 473.83
(7.0%) 0.1% ( -12% - 15%) 0.979
OrHighNotMed 450.91 (9.5%) 451.18
(9.0%) 0.1% ( -16% - 20%) 0.984
Fuzzy2 99.96 (1.9%) 100.14
(1.8%) 0.2% ( -3% - 3%) 0.762
LowSpanNear 21.33 (3.2%) 21.39
(2.8%) 0.3% ( -5% - 6%) 0.781
MedSpanNear 8.98 (3.5%) 9.00
(3.1%) 0.3% ( -6% - 7%) 0.790
OrHighNotHigh 383.77 (7.3%) 384.86
(6.9%) 0.3% ( -12% - 15%) 0.899
OrHighMedDayTaxoFacets 18.70 (6.0%) 18.76
(5.2%) 0.4% ( -10% - 12%) 0.842
MedIntervalsOrdered 80.23 (4.9%) 80.54
(4.2%) 0.4% ( -8% - 9%) 0.793
LowSloppyPhrase 10.69 (3.4%) 10.73
(3.0%) 0.4% ( -5% - 7%) 0.698
AndHighLow 1445.17 (4.2%) 1450.94
(3.8%) 0.4% ( -7% - 8%) 0.751
MedTermDayTaxoFacets 74.54 (4.1%) 74.86
(4.5%) 0.4% ( -7% - 9%) 0.750
LowIntervalsOrdered 44.83 (4.7%) 45.07
(4.1%) 0.5% ( -7% - 9%) 0.700
HighPhrase 105.75 (3.9%) 106.36
(4.9%) 0.6% ( -7% - 9%) 0.682
AndHighMedDayTaxoFacets 83.20 (2.1%) 83.69
(1.8%) 0.6% ( -3% - 4%) 0.352
LowPhrase 17.49 (3.2%) 17.59
(3.8%) 0.6% ( -6% - 7%) 0.591
HighIntervalsOrdered 17.39 (4.4%) 17.52
(3.4%) 0.8% ( -6% - 9%) 0.548
AndHighHighDayTaxoFacets 26.40 (3.2%) 26.63
(2.9%) 0.9% ( -5% - 7%) 0.371
HighSpanNear 46.05 (6.6%) 46.47
(4.9%) 0.9% ( -9% - 13%) 0.625
OrNotHighLow 1075.97 (3.1%) 1085.82
(3.0%) 0.9% ( -5% - 7%) 0.348
Prefix3 269.21 (3.9%) 271.85
(4.7%) 1.0% ( -7% - 9%) 0.472
LowTerm 987.41 (5.0%) 998.09
(6.9%) 1.1% ( -10% - 13%) 0.571
MedPhrase 158.62 (3.4%) 160.46
(4.0%) 1.2% ( -6% - 8%) 0.327
HighTermTitleBDVSort 21.64 (3.0%) 21.96
(3.0%) 1.5% ( -4% - 7%) 0.120
BrowseMonthSSDVFacets 20.71 (7.3%) 21.06
(10.9%) 1.7% ( -15% - 21%) 0.566
Fuzzy1 81.69 (4.7%) 100.83
(3.2%) 23.4% ( 14% - 32%) 0.000
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]