atris commented on issue #877: LUCENE-8978: Maximal Bottom Score Based Early Termination URL: https://github.com/apache/lucene-solr/pull/877#issuecomment-533555038 I realised I had some issues with my previous runs -- they were messing up because of a hack that I had put in luceneutil and hadn't removed it (from an experiment done on a separate patch). I did a fresh checkout of luceneutil and ran `wikimedium2m` and `wikimedium10m` with concurrent searches enabled. For *all* of the tasks run, the total number of hits visited by this patch were lesser than that visited by vanilla. The difference ranged from 100 odd hits to 1/6 number of hits e.g. for OrHighNotHigh: hit count visited by vanilla was 32112 vs this patch visited 5618 hits. No errors in difference in results were reported by luceneutil. The QPS difference range is the same : For HighPhrase, QPS doubled in the case of `wikimedium10m`. Full results at: https://issues.apache.org/jira/browse/LUCENE-8978?focusedCommentId=16934206&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-16934206
---------------------------------------------------------------- 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 With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
