mikemccand commented on issue #12527: URL: https://github.com/apache/lucene/issues/12527#issuecomment-1710234616
OK I tested the "read into scratch array" approach from [this comment](https://github.com/apache/lucene/issues/12527#issuecomment-1708857931): ``` Task QPS base StdDevQPS readLongs StdDev Pct diff p-value IntNRQ 676.26 (3.7%) 607.60 (3.2%) -10.2% ( -16% - -3%) 0.000 BrowseDayOfYearSSDVFacets 13.19 (12.2%) 12.70 (10.2%) -3.7% ( -23% - 21%) 0.296 HighTermDayOfYearSort 676.54 (1.4%) 653.90 (1.2%) -3.3% ( -5% - 0%) 0.000 TermDTSort 291.54 (1.6%) 285.07 (1.4%) -2.2% ( -5% - 0%) 0.000 BrowseDayOfYearTaxoFacets 8.47 (8.6%) 8.32 (5.0%) -1.8% ( -14% - 12%) 0.417 BrowseDateTaxoFacets 8.50 (8.4%) 8.35 (5.1%) -1.7% ( -14% - 12%) 0.428 HighSloppyPhrase 27.18 (2.5%) 26.90 (3.3%) -1.0% ( -6% - 4%) 0.259 MedPhrase 12.74 (7.0%) 12.65 (7.2%) -0.7% ( -13% - 14%) 0.747 LowTerm 820.76 (3.4%) 815.74 (3.5%) -0.6% ( -7% - 6%) 0.574 HighTermTitleBDVSort 5.39 (3.5%) 5.35 (2.5%) -0.6% ( -6% - 5%) 0.557 BrowseRandomLabelSSDVFacets 9.18 (5.4%) 9.13 (5.8%) -0.6% ( -11% - 11%) 0.753 HighPhrase 10.16 (4.9%) 10.11 (4.9%) -0.5% ( -9% - 9%) 0.739 OrHighHigh 26.04 (5.6%) 25.91 (4.4%) -0.5% ( -9% - 10%) 0.745 LowPhrase 13.05 (3.4%) 12.98 (3.4%) -0.5% ( -7% - 6%) 0.650 BrowseRandomLabelTaxoFacets 7.68 (4.8%) 7.64 (3.1%) -0.5% ( -8% - 7%) 0.703 BrowseDateSSDVFacets 2.19 (1.1%) 2.18 (1.4%) -0.4% ( -2% - 2%) 0.323 Prefix3 206.91 (5.3%) 206.16 (5.7%) -0.4% ( -10% - 11%) 0.835 AndHighLow 700.11 (2.1%) 697.60 (1.6%) -0.4% ( -3% - 3%) 0.545 LowSloppyPhrase 71.11 (1.9%) 70.87 (2.2%) -0.3% ( -4% - 3%) 0.599 OrNotHighLow 666.63 (1.5%) 664.60 (1.6%) -0.3% ( -3% - 2%) 0.539 MedSloppyPhrase 51.20 (4.6%) 51.07 (5.1%) -0.3% ( -9% - 9%) 0.869 OrHighNotHigh 356.18 (4.7%) 355.40 (4.7%) -0.2% ( -9% - 9%) 0.882 HighTermMonthSort 3590.99 (1.1%) 3584.45 (0.9%) -0.2% ( -2% - 1%) 0.576 OrHighLow 337.00 (4.3%) 336.39 (3.6%) -0.2% ( -7% - 8%) 0.886 HighTermTitleSort 193.11 (1.0%) 192.80 (0.9%) -0.2% ( -2% - 1%) 0.589 OrNotHighMed 261.20 (2.7%) 261.10 (2.6%) -0.0% ( -5% - 5%) 0.962 MedTermDayTaxoFacets 41.25 (2.1%) 41.24 (1.5%) -0.0% ( -3% - 3%) 0.958 OrHighMed 148.69 (5.2%) 148.66 (4.3%) -0.0% ( -9% - 9%) 0.988 Fuzzy1 46.00 (1.1%) 46.00 (1.3%) -0.0% ( -2% - 2%) 0.999 Wildcard 213.15 (1.6%) 213.19 (1.4%) 0.0% ( -2% - 2%) 0.967 OrNotHighHigh 455.16 (3.7%) 455.25 (3.7%) 0.0% ( -7% - 7%) 0.987 OrHighNotLow 517.26 (6.4%) 517.55 (6.6%) 0.1% ( -12% - 13%) 0.978 Respell 34.81 (0.8%) 34.84 (0.7%) 0.1% ( -1% - 1%) 0.723 HighTerm 695.15 (5.9%) 695.81 (5.8%) 0.1% ( -10% - 12%) 0.959 AndHighMed 76.71 (4.4%) 76.81 (4.3%) 0.1% ( -8% - 9%) 0.922 AndHighHigh 50.22 (3.0%) 50.29 (3.0%) 0.1% ( -5% - 6%) 0.886 MedTerm 925.46 (5.3%) 926.77 (5.4%) 0.1% ( -10% - 11%) 0.933 OrHighNotMed 499.30 (5.1%) 500.08 (5.3%) 0.2% ( -9% - 11%) 0.925 AndHighHighDayTaxoFacets 16.62 (1.9%) 16.66 (0.9%) 0.3% ( -2% - 3%) 0.587 Fuzzy2 42.50 (1.0%) 42.61 (1.0%) 0.3% ( -1% - 2%) 0.399 HighIntervalsOrdered 4.84 (4.8%) 4.85 (4.7%) 0.3% ( -8% - 10%) 0.825 PKLookup 253.83 (1.1%) 254.83 (1.0%) 0.4% ( -1% - 2%) 0.225 AndHighMedDayTaxoFacets 28.49 (1.6%) 28.61 (1.9%) 0.4% ( -3% - 4%) 0.460 LowIntervalsOrdered 7.78 (3.9%) 7.82 (4.1%) 0.5% ( -7% - 8%) 0.719 HighSpanNear 24.99 (2.6%) 25.11 (2.9%) 0.5% ( -4% - 6%) 0.585 LowSpanNear 10.42 (4.0%) 10.48 (4.0%) 0.5% ( -7% - 8%) 0.680 MedIntervalsOrdered 26.55 (5.1%) 26.70 (5.5%) 0.5% ( -9% - 11%) 0.748 BrowseMonthTaxoFacets 8.20 (1.6%) 8.25 (0.3%) 0.6% ( -1% - 2%) 0.109 OrHighMedDayTaxoFacets 5.69 (3.8%) 5.73 (2.9%) 0.6% ( -5% - 7%) 0.577 MedSpanNear 13.20 (3.6%) 13.29 (3.7%) 0.6% ( -6% - 8%) 0.582 BrowseMonthSSDVFacets 13.87 (13.0%) 14.38 (13.4%) 3.7% ( -20% - 34%) 0.380 ``` It looks like `IntNRQ` is still upset, but not as much (-18% down to -10%). Maybe next we should try 4 `readLong()` for `readInts32`? Though I wonder how often in this benchy are we really needing 32 bits to encode the docid deltas in a BKD leaf block? We need some sort of simple tool to print out how often each bit-width is uses in and index... -- 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: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
