Github user jiangxb1987 commented on the issue:
https://github.com/apache/spark/pull/22849
What do you mean by "better performance" ? If that means we can spend less
time on `TaskSchedulerImpl.resourceOffers()` then I agree it's true, but AFAIK
it's never reported this can be a bottleneck of the whole cluster, so maybe the
perf gain is trivial here. If you expect better task distribution over existing
executors then I don't see any case can be improved by this proposed change.
Please correct me if I'm wrong.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
