gsmiller commented on PR #954: URL: https://github.com/apache/lucene/pull/954#issuecomment-1154550479
@jpountz: > plus there may be lots of users doing old-style iteration so we'll need to deprecate and maybe only clean this up in 10.0 or 11.0 Right, makes sense. > But I'm curious of the sort of speedup that this would yield Me too. I hacked up a version of this change on another branch ([here](https://github.com/gsmiller/lucene/commit/3768665ae6c173014e8288c46c27afa517c90ede)) that let calling code explicitly ask for a "fast" version of SSDV that doesn't do the ordinal check, and then relied on this new code path for loading SSDV within the faceting module. I didn't observe a significant change with our benchmark tooling, but I wonder how much we actually exercise these multi-value cases within our faceting benchmarks. I think it will be more interesting to test once we've migrated more use-cases to this new SSDV iteration style. You may have had a completely different thought in mind for testing this, so please let me know if I'm missing the mark here. Thanks! -- 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
