[ https://issues.apache.org/jira/browse/LUCENE-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Han Jiang updated LUCENE-3892: ------------------------------ Attachment: LUCENE-3892-blockFor-with-packedints-decoder.patch Patch with the decoder interface, mentioned in LUCENE-4239. I'm afraid that the for loop of readLong() hurts the performance. Here is the comparison against last patch: {noformat} Task QPS base StdDev base QPS comp StdDev comp Pct diff AndHighHigh 21.89 0.64 22.14 0.43 -3% - 6% AndHighMed 52.23 2.34 52.94 1.74 -6% - 9% Fuzzy1 86.61 1.63 87.29 3.14 -4% - 6% Fuzzy2 30.54 0.54 30.95 1.18 -4% - 7% IntNRQ 38.00 1.23 38.14 1.04 -5% - 6% OrHighHigh 16.37 0.21 16.68 0.79 -4% - 8% OrHighMed 39.59 0.69 40.34 2.16 -5% - 9% PKLookup 111.51 1.34 112.78 1.37 -1% - 3% Phrase 4.54 0.12 4.52 0.13 -5% - 5% Prefix3 107.85 2.51 109.13 2.10 -3% - 5% Respell 123.21 2.18 125.15 5.01 -4% - 7% SloppyPhrase 6.51 0.11 6.44 0.29 -7% - 5% SpanNear 5.36 0.16 5.31 0.14 -6% - 4% Term 42.49 1.66 44.10 1.86 -4% - 12% TermBGroup1M 17.86 0.80 17.82 0.51 -7% - 7% TermBGroup1M1P 21.08 0.55 21.10 0.62 -5% - 5% TermGroup1M 19.57 0.82 19.57 0.64 -7% - 7% Wildcard 43.99 1.21 44.80 1.10 -3% - 7% {noformat} > Add a useful intblock postings format (eg, FOR, PFOR, PFORDelta, > Simple9/16/64, etc.) > ------------------------------------------------------------------------------------- > > Key: LUCENE-3892 > URL: https://issues.apache.org/jira/browse/LUCENE-3892 > Project: Lucene - Java > Issue Type: Improvement > Reporter: Michael McCandless > Labels: gsoc2012, lucene-gsoc-12 > Fix For: 4.1 > > Attachments: LUCENE-3892-BlockTermScorer.patch, > LUCENE-3892-blockFor-with-packedints-decoder.patch, > LUCENE-3892-blockFor-with-packedints.patch, > LUCENE-3892-direct-IntBuffer.patch, LUCENE-3892-for&pfor-with-javadoc.patch, > LUCENE-3892-for&pfor-with-javadoc.patch, > LUCENE-3892-for&pfor-with-javadoc.patch, > LUCENE-3892-for&pfor-with-javadoc.patch, LUCENE-3892-for&pfor.patch, > LUCENE-3892-handle_open_files.patch, > LUCENE-3892-pfor-compress-iterate-numbits.patch, > LUCENE-3892-pfor-compress-slow-estimate.patch, LUCENE-3892_for.patch, > LUCENE-3892_for_byte[].patch, LUCENE-3892_for_int[].patch, > LUCENE-3892_for_unfold_method.patch, LUCENE-3892_pfor.patch, > LUCENE-3892_pfor.patch, LUCENE-3892_pfor.patch, > LUCENE-3892_pfor_unfold_method.patch, LUCENE-3892_pulsing_support.patch, > LUCENE-3892_settings.patch, LUCENE-3892_settings.patch > > > On the flex branch we explored a number of possible intblock > encodings, but for whatever reason never brought them to completion. > There are still a number of issues opened with patches in different > states. > Initial results (based on prototype) were excellent (see > http://blog.mikemccandless.com/2010/08/lucene-performance-with-pfordelta-codec.html > ). > I think this would make a good GSoC project. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org