[
https://issues.apache.org/jira/browse/IMPALA-6311?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16484652#comment-16484652
]
Jim Apple commented on IMPALA-6311:
-----------------------------------
[~mmokhtar], I can get a speedup on TPCH-Q4, scale factor 20, on a minicluster.
This looks to me like a time/space trade-off.
https://jenkins.impala.io/job/perf-AB-test/162/artifact/Impala/perf_results/latest/performance_result.txt
What are your thoughts on increasing the default filter sizes, but not raising
the ceiling?
> Evaluate smaller FPP for Bloom filters
> --------------------------------------
>
> Key: IMPALA-6311
> URL: https://issues.apache.org/jira/browse/IMPALA-6311
> Project: IMPALA
> Issue Type: Task
> Components: Perf Investigation
> Reporter: Jim Apple
> Priority: Major
>
> The Bloom filters are created by estimating the NDV and then using the FPP of
> 75% to get the right size for the filter. This is may be too high to be very
> useful - if our filters are currently filtering more than 75% out, then it is
> only because we are overestimating NDV.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
