[ https://issues.apache.org/jira/browse/LUCENE-1410?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12804127#action_12804127 ]
Paul Elschot commented on LUCENE-1410: -------------------------------------- Zhang, 2008, see above, reports this: {quote} The poor speed of variable- byte on position data is primarily due to the fact that position values are larger and more often require 2 bytes under variable-byte; this case tends to be much slower due to a branch mispredict. {quote} Taking another look at the position data above (3 October 2008) 11.6% of prx values take 7 bits or less, and the rest fits in 15 bits. So why not encode the position data as VShort (1 bit as in VByte and 15 bits data) ? That would enlarge a typical prx file by about 6% and increase position decoding speed a lot, probably about 3 times (see Table 1 in the same paper). > PFOR implementation > ------------------- > > Key: LUCENE-1410 > URL: https://issues.apache.org/jira/browse/LUCENE-1410 > Project: Lucene - Java > Issue Type: New Feature > Components: Other > Reporter: Paul Elschot > Priority: Minor > Attachments: autogen.tgz, LUCENE-1410-codecs.tar.bz2, > LUCENE-1410b.patch, LUCENE-1410c.patch, LUCENE-1410d.patch, > LUCENE-1410e.patch, TermQueryTests.tgz, TestPFor2.java, TestPFor2.java, > TestPFor2.java > > Original Estimate: 21840h > Remaining Estimate: 21840h > > Implementation of Patched Frame of Reference. -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online. --------------------------------------------------------------------- To unsubscribe, e-mail: java-dev-unsubscr...@lucene.apache.org For additional commands, e-mail: java-dev-h...@lucene.apache.org