zacharymorn commented on pull request #101: URL: https://github.com/apache/lucene/pull/101#issuecomment-840255508
I'm able to run wikibigall with the above configuration (and a few small changes to `localrun.py` to take in vector dictionary and file), and get the following results: wikibigall run 1 ``` TaskQPS baseline StdDevQPS my_modified_version StdDev Pct diff p-value AndMedOrHighHigh 25.47 (2.5%) 21.21 (4.1%) -16.7% ( -22% - -10%) 0.000 OrHighMed 72.05 (2.5%) 60.17 (4.1%) -16.5% ( -22% - -10%) 0.000 Fuzzy2 31.51 (8.4%) 29.27 (9.5%) -7.1% ( -23% - 11%) 0.012 Fuzzy1 59.00 (8.4%) 56.38 (9.8%) -4.4% ( -20% - 14%) 0.122 TermDayOfYearSort 48.68 (14.6%) 46.86 (8.9%) -3.7% ( -23% - 23%) 0.328 TermMonthSort 61.44 (9.3%) 60.23 (9.8%) -2.0% ( -19% - 18%) 0.514 TermTitleSort 110.13 (8.9%) 108.54 (9.2%) -1.4% ( -18% - 18%) 0.616 TermDTSort 56.74 (9.5%) 56.04 (10.2%) -1.2% ( -19% - 20%) 0.692 TermGroup10K 15.95 (2.6%) 15.75 (3.2%) -1.2% ( -6% - 4%) 0.183 TermGroup100 19.34 (3.5%) 19.11 (3.4%) -1.2% ( -7% - 5%) 0.276 TermBGroup1M 17.92 (2.7%) 17.72 (3.8%) -1.1% ( -7% - 5%) 0.280 TermGroup1M 19.06 (2.5%) 18.86 (3.1%) -1.1% ( -6% - 4%) 0.234 TermBGroup1M1P 43.93 (3.8%) 43.66 (4.4%) -0.6% ( -8% - 7%) 0.633 SpanNear 24.55 (2.5%) 24.46 (2.4%) -0.4% ( -5% - 4%) 0.620 SloppyPhrase 2.99 (8.2%) 2.98 (8.0%) -0.3% ( -15% - 17%) 0.897 IntNRQ 139.92 (1.2%) 139.61 (1.7%) -0.2% ( -3% - 2%) 0.627 IntervalsOrdered 2.02 (3.7%) 2.02 (3.6%) -0.1% ( -7% - 7%) 0.957 BrowseMonthSSDVFacets 19.28 (1.0%) 19.27 (0.9%) -0.0% ( -1% - 1%) 0.888 BrowseDayOfYearSSDVFacets 17.61 (1.9%) 17.61 (1.8%) 0.0% ( -3% - 3%) 0.989 Respell 48.40 (2.6%) 48.59 (3.1%) 0.4% ( -5% - 6%) 0.660 PKLookup 211.17 (3.4%) 212.28 (3.6%) 0.5% ( -6% - 7%) 0.631 BrowseDayOfYearTaxoFacets 7.30 (6.4%) 7.34 (5.8%) 0.6% ( -10% - 13%) 0.766 Prefix3 120.15 (6.1%) 120.95 (4.6%) 0.7% ( -9% - 12%) 0.694 Phrase 54.76 (3.6%) 55.13 (4.3%) 0.7% ( -6% - 8%) 0.586 BrowseDateTaxoFacets 7.62 (6.7%) 7.67 (6.1%) 0.7% ( -11% - 14%) 0.718 Wildcard 72.89 (2.8%) 73.46 (2.2%) 0.8% ( -4% - 5%) 0.331 BrowseMonthTaxoFacets 8.42 (7.0%) 8.49 (6.4%) 0.8% ( -11% - 15%) 0.694 TermDateFacets 8.56 (7.4%) 8.64 (7.3%) 0.9% ( -12% - 16%) 0.702 VectorSearch 954.06 (3.7%) 967.28 (2.5%) 1.4% ( -4% - 7%) 0.167 AndHighHigh 32.04 (3.1%) 32.65 (4.1%) 1.9% ( -5% - 9%) 0.100 AndHighMed 83.72 (1.8%) 85.38 (2.9%) 2.0% ( -2% - 6%) 0.009 AndHighOrMedMed 44.41 (2.2%) 45.34 (4.2%) 2.1% ( -4% - 8%) 0.049 Term 1132.83 (5.2%) 1157.74 (5.0%) 2.2% ( -7% - 13%) 0.173 OrHighHigh 11.35 (2.9%) 16.95 (4.1%) 49.4% ( 41% - 58%) 0.000 ``` wikibigall run 2 ``` TaskQPS baseline StdDevQPS my_modified_version StdDev Pct diff p-value TermDayOfYearSort 59.31 (11.3%) 56.25 (9.5%) -5.2% ( -23% - 17%) 0.116 TermTitleSort 175.57 (10.3%) 169.31 (9.3%) -3.6% ( -21% - 17%) 0.252 TermMonthSort 61.73 (10.5%) 59.73 (9.2%) -3.2% ( -20% - 18%) 0.300 TermBGroup1M 11.87 (3.2%) 11.57 (2.6%) -2.6% ( -8% - 3%) 0.005 TermGroup10K 15.60 (2.8%) 15.26 (3.3%) -2.2% ( -8% - 4%) 0.023 TermGroup100 23.39 (2.6%) 22.94 (3.0%) -1.9% ( -7% - 3%) 0.030 TermGroup1M 12.16 (2.6%) 11.94 (2.9%) -1.8% ( -7% - 3%) 0.036 TermDTSort 78.08 (6.4%) 76.75 (5.4%) -1.7% ( -12% - 10%) 0.361 Fuzzy1 28.49 (5.6%) 28.08 (5.8%) -1.4% ( -12% - 10%) 0.419 TermBGroup1M1P 42.98 (2.9%) 42.49 (3.3%) -1.1% ( -7% - 5%) 0.247 IntNRQ 136.00 (3.4%) 134.92 (3.5%) -0.8% ( -7% - 6%) 0.473 AndHighHigh 22.60 (3.9%) 22.42 (2.4%) -0.8% ( -6% - 5%) 0.455 TermDateFacets 8.38 (7.0%) 8.32 (6.6%) -0.7% ( -13% - 13%) 0.738 Prefix3 33.76 (3.1%) 33.55 (2.8%) -0.6% ( -6% - 5%) 0.502 Wildcard 88.25 (3.0%) 87.72 (2.3%) -0.6% ( -5% - 4%) 0.479 BrowseDayOfYearSSDVFacets 17.39 (1.4%) 17.31 (1.5%) -0.4% ( -3% - 2%) 0.363 Term 1144.71 (8.2%) 1139.98 (8.5%) -0.4% ( -15% - 17%) 0.876 VectorSearch 807.48 (4.7%) 804.27 (4.9%) -0.4% ( -9% - 9%) 0.794 IntervalsOrdered 0.82 (5.3%) 0.82 (5.4%) -0.3% ( -10% - 11%) 0.872 PKLookup 207.38 (4.8%) 206.99 (5.5%) -0.2% ( -10% - 10%) 0.910 Respell 48.27 (3.6%) 48.18 (3.6%) -0.2% ( -7% - 7%) 0.878 BrowseDateTaxoFacets 7.58 (6.8%) 7.57 (6.5%) -0.2% ( -12% - 14%) 0.933 BrowseDayOfYearTaxoFacets 7.26 (6.3%) 7.24 (6.2%) -0.2% ( -11% - 13%) 0.932 SloppyPhrase 5.55 (5.7%) 5.54 (5.8%) -0.1% ( -10% - 12%) 0.946 BrowseMonthTaxoFacets 8.38 (6.7%) 8.38 (6.6%) -0.1% ( -12% - 14%) 0.969 SpanNear 2.12 (3.5%) 2.12 (3.9%) 0.0% ( -7% - 7%) 0.993 BrowseMonthSSDVFacets 19.52 (5.7%) 19.53 (5.7%) 0.0% ( -10% - 12%) 0.985 Phrase 59.48 (3.8%) 59.72 (3.9%) 0.4% ( -6% - 8%) 0.735 AndHighMed 60.30 (4.6%) 60.57 (2.3%) 0.4% ( -6% - 7%) 0.705 Fuzzy2 53.28 (11.3%) 53.93 (10.6%) 1.2% ( -18% - 26%) 0.727 AndMedOrHighHigh 27.58 (3.1%) 27.93 (3.4%) 1.3% ( -5% - 8%) 0.209 AndHighOrMedMed 24.38 (3.1%) 25.06 (3.3%) 2.8% ( -3% - 9%) 0.006 OrHighMed 45.08 (4.3%) 57.44 (6.9%) 27.4% ( 15% - 40%) 0.000 OrHighHigh 11.22 (4.8%) 16.16 (9.8%) 44.0% ( 28% - 61%) 0.000 ``` wikibigall run 3 ``` TaskQPS baseline StdDevQPS my_modified_version StdDev Pct diff p-value Fuzzy2 51.26 (10.6%) 47.91 (11.4%) -6.5% ( -25% - 17%) 0.060 TermDTSort 83.68 (11.6%) 79.69 (7.7%) -4.8% ( -21% - 16%) 0.126 Fuzzy1 60.65 (6.6%) 57.96 (8.5%) -4.4% ( -18% - 11%) 0.065 TermMonthSort 61.55 (15.5%) 59.70 (8.4%) -3.0% ( -23% - 24%) 0.447 TermTitleSort 88.13 (15.1%) 85.68 (8.4%) -2.8% ( -22% - 24%) 0.471 TermDayOfYearSort 47.32 (8.7%) 46.58 (10.0%) -1.5% ( -18% - 18%) 0.602 Wildcard 37.71 (9.9%) 37.32 (9.1%) -1.0% ( -18% - 19%) 0.729 Prefix3 164.10 (12.8%) 162.91 (11.5%) -0.7% ( -22% - 27%) 0.850 TermGroup100 23.62 (3.0%) 23.50 (2.9%) -0.5% ( -6% - 5%) 0.617 TermBGroup1M 11.90 (2.7%) 11.86 (3.1%) -0.3% ( -5% - 5%) 0.714 Term 1044.58 (8.0%) 1041.84 (6.5%) -0.3% ( -13% - 15%) 0.910 TermGroup10K 12.18 (2.9%) 12.15 (2.7%) -0.2% ( -5% - 5%) 0.808 IntervalsOrdered 3.91 (2.7%) 3.91 (2.7%) -0.1% ( -5% - 5%) 0.934 BrowseDayOfYearSSDVFacets 17.47 (2.3%) 17.47 (2.0%) -0.0% ( -4% - 4%) 0.994 VectorSearch 851.38 (5.2%) 851.53 (5.6%) 0.0% ( -10% - 11%) 0.992 SpanNear 24.59 (2.2%) 24.61 (2.3%) 0.1% ( -4% - 4%) 0.923 TermGroup1M 15.44 (2.7%) 15.45 (2.9%) 0.1% ( -5% - 5%) 0.936 BrowseMonthSSDVFacets 19.19 (1.4%) 19.22 (1.5%) 0.2% ( -2% - 3%) 0.746 TermBGroup1M1P 18.19 (4.7%) 18.24 (4.6%) 0.3% ( -8% - 9%) 0.858 IntNRQ 136.29 (3.4%) 136.75 (3.5%) 0.3% ( -6% - 7%) 0.755 Respell 40.87 (5.4%) 41.04 (4.8%) 0.4% ( -9% - 11%) 0.793 Phrase 24.71 (3.1%) 24.87 (2.7%) 0.7% ( -4% - 6%) 0.472 PKLookup 205.99 (5.9%) 208.11 (6.4%) 1.0% ( -10% - 14%) 0.596 BrowseDayOfYearTaxoFacets 7.04 (7.8%) 7.11 (7.5%) 1.0% ( -13% - 17%) 0.665 BrowseDateTaxoFacets 7.36 (7.9%) 7.45 (7.7%) 1.1% ( -13% - 18%) 0.646 SloppyPhrase 1.17 (10.0%) 1.19 (9.8%) 1.2% ( -16% - 23%) 0.705 AndHighMed 45.25 (3.5%) 45.82 (3.8%) 1.3% ( -5% - 8%) 0.274 BrowseMonthTaxoFacets 8.13 (7.9%) 8.25 (7.9%) 1.4% ( -13% - 18%) 0.569 AndHighHigh 23.27 (3.6%) 23.60 (4.0%) 1.4% ( -5% - 9%) 0.227 TermDateFacets 11.08 (9.9%) 11.27 (10.0%) 1.7% ( -16% - 23%) 0.582 AndHighOrMedMed 24.91 (3.8%) 25.72 (4.0%) 3.3% ( -4% - 11%) 0.008 AndMedOrHighHigh 10.00 (4.2%) 10.61 (4.2%) 6.1% ( -2% - 15%) 0.000 OrHighMed 59.17 (5.5%) 67.57 (4.5%) 14.2% ( 3% - 25%) 0.000 OrHighHigh 17.06 (4.9%) 24.26 (5.2%) 42.2% ( 30% - 55%) 0.000 ``` I'll run it for the other two implementations as well and post the results in the other PR. -- 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. 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