Github user ConeyLiu commented on the issue: https://github.com/apache/spark/pull/19135 @jiangxb1987 Ok, I can test it later. The following picture is when I run kmeans and put the source data into the offheap memory, and you can see the CPU time occupied by `reserveUnrollMemoryForThisTask` is very high. ![pic](https://user-images.githubusercontent.com/12733256/30142120-a3a3dd42-9344-11e7-9ae3-1c36bedf8939.png)
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org