original-brownbear opened a new pull request, #13906:
URL: https://github.com/apache/lucene/pull/13906
No need to slice when it's a clone that is to be used with random access, we
already enforce thread access rules anyway. Also, no point in copying the
memory segment instance via noop slicing.
This doesn't really have much of a runtime impact (looking at wikimedium):
```
BrowseDayOfYearSSDVFacets 6.18 (8.5%) 6.07
(8.7%) -1.8% ( -17% - 16%) 0.347
IntNRQ 21.39 (3.3%) 21.13
(3.8%) -1.2% ( -8% - 6%) 0.115
HighIntervalsOrdered 9.76 (4.3%) 9.69
(5.7%) -0.7% ( -10% - 9%) 0.511
HighTermDayOfYearSort 156.54 (2.4%) 155.52
(2.9%) -0.7% ( -5% - 4%) 0.279
LowSpanNear 61.87 (2.3%) 61.48
(2.8%) -0.6% ( -5% - 4%) 0.273
TermDTSort 53.99 (1.3%) 53.71
(1.6%) -0.5% ( -3% - 2%) 0.108
HighSpanNear 6.69 (4.4%) 6.66
(4.3%) -0.4% ( -8% - 8%) 0.697
OrHighLow 479.99 (2.1%) 479.02
(2.2%) -0.2% ( -4% - 4%) 0.670
Fuzzy1 62.30 (1.2%) 62.17
(1.3%) -0.2% ( -2% - 2%) 0.480
HighTermMonthSort 556.56 (1.3%) 555.47
(1.6%) -0.2% ( -3% - 2%) 0.546
AndHighHighDayTaxoFacets 5.60 (4.7%) 5.59
(3.7%) -0.2% ( -8% - 8%) 0.844
Fuzzy2 61.63 (1.3%) 61.54
(1.6%) -0.1% ( -3% - 2%) 0.651
Respell 34.27 (1.5%) 34.22
(1.4%) -0.1% ( -3% - 2%) 0.702
LowSloppyPhrase 19.76 (3.9%) 19.74
(3.5%) -0.1% ( -7% - 7%) 0.880
MedIntervalsOrdered 19.12 (2.6%) 19.10
(2.6%) -0.1% ( -5% - 5%) 0.872
Wildcard 115.59 (2.8%) 115.51
(2.8%) -0.1% ( -5% - 5%) 0.905
Prefix3 318.40 (1.4%) 318.82
(1.9%) 0.1% ( -3% - 3%) 0.727
AndHighMedDayTaxoFacets 52.44 (2.5%) 52.51
(2.3%) 0.1% ( -4% - 5%) 0.794
BrowseDateSSDVFacets 1.32 (6.3%) 1.33
(4.8%) 0.2% ( -10% - 11%) 0.895
PKLookup 157.21 (3.4%) 157.48
(3.8%) 0.2% ( -6% - 7%) 0.829
HighPhrase 95.48 (2.3%) 95.69
(2.5%) 0.2% ( -4% - 5%) 0.682
BrowseMonthTaxoFacets 4.89 (0.4%) 4.90
(0.4%) 0.2% ( 0% - 1%) 0.013
MedSloppyPhrase 13.62 (2.8%) 13.66
(2.5%) 0.3% ( -4% - 5%) 0.615
MedPhrase 23.07 (2.1%) 23.15
(2.4%) 0.3% ( -4% - 4%) 0.495
HighSloppyPhrase 18.02 (5.3%) 18.09
(4.0%) 0.4% ( -8% - 10%) 0.713
MedSpanNear 8.17 (2.9%) 8.20
(2.9%) 0.4% ( -5% - 6%) 0.552
LowPhrase 112.30 (2.0%) 112.75
(2.0%) 0.4% ( -3% - 4%) 0.373
AndHighLow 607.00 (5.8%) 609.50
(5.4%) 0.4% ( -10% - 12%) 0.740
LowTerm 204.43 (2.0%) 205.31
(2.7%) 0.4% ( -4% - 5%) 0.425
HighTermTitleSort 17.83 (1.4%) 17.91
(1.4%) 0.4% ( -2% - 3%) 0.159
OrHighMedDayTaxoFacets 3.09 (3.9%) 3.11
(3.8%) 0.5% ( -6% - 8%) 0.599
LowIntervalsOrdered 20.57 (3.1%) 20.67
(3.9%) 0.5% ( -6% - 7%) 0.550
HighTermTitleBDVSort 5.50 (3.4%) 5.53
(3.7%) 0.6% ( -6% - 7%) 0.468
OrHighMed 89.37 (4.0%) 89.89
(5.4%) 0.6% ( -8% - 10%) 0.583
AndHighMed 71.17 (4.3%) 71.61
(5.2%) 0.6% ( -8% - 10%) 0.555
MedTermDayTaxoFacets 20.66 (4.0%) 20.81
(4.4%) 0.7% ( -7% - 9%) 0.458
OrHighHigh 32.44 (6.3%) 32.70
(8.0%) 0.8% ( -12% - 16%) 0.625
BrowseMonthSSDVFacets 6.29 (7.4%) 6.34
(10.5%) 0.8% ( -15% - 20%) 0.684
OrNotHighLow 581.89 (2.8%) 586.83
(2.6%) 0.8% ( -4% - 6%) 0.164
OrNotHighHigh 208.03 (5.2%) 209.97
(6.0%) 0.9% ( -9% - 12%) 0.457
MedTerm 281.51 (5.0%) 284.37
(6.7%) 1.0% ( -10% - 13%) 0.442
OrNotHighMed 231.06 (4.0%) 233.53
(4.1%) 1.1% ( -6% - 9%) 0.240
OrHighNotMed 280.55 (4.6%) 283.63
(5.7%) 1.1% ( -8% - 11%) 0.339
OrHighNotHigh 141.67 (5.3%) 143.24
(6.5%) 1.1% ( -10% - 13%) 0.403
AndHighHigh 8.71 (5.8%) 8.82
(7.4%) 1.2% ( -11% - 15%) 0.419
OrHighNotLow 247.33 (5.3%) 250.51
(6.5%) 1.3% ( -9% - 13%) 0.332
HighTerm 202.30 (5.5%) 205.03
(7.4%) 1.4% ( -10% - 15%) 0.355
BrowseRandomLabelSSDVFacets 4.16 (3.6%) 4.27
(7.1%) 2.4% ( -7% - 13%) 0.054
BrowseRandomLabelTaxoFacets 4.69 (5.5%) 4.97
(22.8%) 5.9% ( -21% - 36%) 0.109
BrowseDateTaxoFacets 5.44 (12.0%) 5.91
(27.5%) 8.6% ( -27% - 54%) 0.069
BrowseDayOfYearTaxoFacets 5.54 (12.4%) 6.04
(28.9%) 8.9% ( -28% - 57%) 0.072
```
but it does save about 1% of all allocations overall (effectively
`jdk.internal.foreign.MappedMemorySegmentImpl#dup()` disappears from the
profile).
--
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]
