[
https://issues.apache.org/jira/browse/PHOENIX-1188?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14103224#comment-14103224
]
James Taylor commented on PHOENIX-1188:
---------------------------------------
If you have any insights, [~gabriel.reid], let us know. I played around with it
a bit, added some logging, and made one change that in the end I don't think
would have mattered. Mujtaba is seeing that prior to this change the
individual, parallel result iterators ran in parallel. Now it seems that, after
your change, they seem to run serially. Would be good to fix this regression
for 3.1/4.1 release if possible.
> Performance regression for non-aggregate queries
> ------------------------------------------------
>
> Key: PHOENIX-1188
> URL: https://issues.apache.org/jira/browse/PHOENIX-1188
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 3.1, 4.1
> Reporter: Mujtaba Chohan
> Assignee: James Taylor
>
> Performance for non-aggregate queries is 4X+ slower. Commit that caused this
> regression (PHOENIX-539) :
> https://git-wip-us.apache.org/repos/asf?p=phoenix.git;a=commit;h=ec1b76ab28880d548c8061a6dcb80a6fe0d97ed6
> FYI [~gabriel.reid]
> Performance is slow even if no rows are returned.
> Sample perf run:
> http://phoenix-bin.github.io/client/performance/phoenix-20140819014309.htm
> Query:
> select a,b,c,d,e,f from TABLE_1CF where B>1000 and B<2000 and f>1000 and
> f<2000
--
This message was sent by Atlassian JIRA
(v6.2#6252)
