[
https://issues.apache.org/jira/browse/KYLIN-2929?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16342967#comment-16342967
]
fengYu commented on KYLIN-2929:
-------------------------------
Upload new patch, please review it if you are free.
> speed up Dump file performance
> ------------------------------
>
> Key: KYLIN-2929
> URL: https://issues.apache.org/jira/browse/KYLIN-2929
> Project: Kylin
> Issue Type: Bug
> Components: Query Engine
> Affects Versions: v2.0.0
> Reporter: fengYu
> Assignee: fengYu
> Priority: Major
> Labels: Performance
> Fix For: v2.3.0
>
> Attachments:
> 0001-KYLIN-2929-speed-up-dump-performance-write-dump-file.patch
>
>
> when I work on KYLIN-2926, I find coprocessor will dump to disk once
> estimatedMemSize is bigger than spillThreshold, and found that spill data
> size is extraordinary smaller that estimatedMemSize, in my case dump file
> size is about 8MB and spillThreshold is setting to 3GB.
> So, I try to keep the spill data in memory rather than write the file to disk
> immediately, and when those in-memory spill data reach the threshold, write
> all spill files together.
> In my case, the coprocessor process cost time drop from 22s to 16s, it is
> about 30% upgrade。
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
