7mming7 commented on PR #42037: URL: https://github.com/apache/spark/pull/42037#issuecomment-1641465795
> Took a quick look mostly on the code aspects. @HyukjinKwon's comment on is this the right approach is valid. > > Btw, are you observing broadcast'ing to be such a large overhead ? Very curious about why this direction to begin with. Yes, it is found here that in the case of small queries with large concurrency, the consumption is more obvious . Tested some performance of 50 concurrency on SSB, parquet as the source data has a difference of 10%-13% -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
