Github user c21 commented on the issue:
https://github.com/apache/spark/pull/23163
@gatorsmile:
> Any perf number?
From my employer company workload, we see >20% reserved CPU time (executor
wall clock time) reduction, and >20% disk spill size reduction, after rolling
out the change to use concurrent writers instead of sort (i.e. hash-based write
in this pr).
I am not sure whether it's the performance number you were looking for. Let
me know if anything needed. Thanks.
In addition, I updated the pr, as I found I need to change
`BasicWriteTaskStatsTracker` as well.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
