[
https://issues.apache.org/jira/browse/HIVE-18866?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16478218#comment-16478218
]
Sergey Shelukhin commented on HIVE-18866:
-----------------------------------------
I found the same issue thru some perf testing and tried this patch, it reduces
certain analyze query DAG runtime from ~1550s. to 450s. Next hotspot is also in
HLL but not related to hash, I might file a bug later.
> Semijoin: Implement a Long -> Hash64 vector fast-path
> -----------------------------------------------------
>
> Key: HIVE-18866
> URL: https://issues.apache.org/jira/browse/HIVE-18866
> Project: Hive
> Issue Type: Improvement
> Components: Vectorization
> Reporter: Gopal V
> Priority: Major
> Labels: performance
> Attachments: 0001-hash64-WIP.patch, perf-hash64-long.png
>
>
> A significant amount of CPU is wasted with JMM restrictions on byte[] arrays.
> To transform from one Long -> another Long, this goes into a byte[] array,
> which shows up as a hotspot.
> !perf-hash64-long.png!
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
