Github user davies commented on the pull request: https://github.com/apache/spark/pull/2933#issuecomment-60529659 Broadcast (especially TorrentBroadcast) is designed for large object, using it to send out small shared variables just like using tank to shot a mosquitoes, it's not a good approach in the begging, which make simple things complicated. The motivation of broadcasting tasks, is to solve the performance for `BIG` closure, it should not brings any regression for other cases (small closure), the latter are more common and important in daily usage. In order to fix the regression (performance or stability), we may need to introduce even more complicated logic (just like embedded broadcast or piggy back small blocks).
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org