[GitHub] [spark] JkSelf commented on pull request #29540: [SPARK-32698][SQL] Do not fall back to default parallelism if the minimum number of coalesced partitions is not set in AQE

[GitBox <gi...@apache.org>]

JkSelf commented on pull request #29540:
URL: https://github.com/apache/spark/pull/29540#issuecomment-683515925


   In our 3TB TPC-DS benchmark, there will be performance regression when the coalesced partition num < shuffle partition number without AQE.  So we set the min partition number to the default parallelism to avoid the regression.
   
   @manuzhang 
   Even the initial partition number is large, AQE will coalesce the original partition number to proper partition number not large number in [here](https://github.com/apache/spark/blob/6dacba7fa044555512f974595ed52f7316bb97f9/sql/core/src/main/scala/org/apache/spark/sql/execution/adaptive/ShufflePartitionsUtil.scala#L56). 
   
   


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org