Github user HeartSaVioR commented on the issue: https://github.com/apache/spark/pull/21733 @tdas Thanks for the detailed review! I'll follow up your comments and update the patch. Btw, If my memory is right, I tried out increasing "rate" while benchmarking, but rate source itself became bottleneck. Not sure c5.xlarge is not enough or I might be missed regarding option(s). Sadly I can't run benchmark often because I don't have any dedicated machine. I would avoid running benchmark in non-dedicated machine for seeing computational limit, so paying to AWS to get dedicated instance/machine. I'll try out increasing "rate" once more soon, but please guide me if you have any suggestions to the benchmark code or approach.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org