[
https://issues.apache.org/jira/browse/HBASE-13307?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14374772#comment-14374772
]
Lars Hofhansl commented on HBASE-13307:
---------------------------------------
Ran my rig: 100m rows, 1col, FAST_DIFF encoded, all in block cache, everything
filtered at the server.
I did not see a measurable end-to-end improvement. :(
Might be my JDK (OpenJDK7 build 24.75-b04, mixed mode)
Lemme try some more stuff w/o FAST_DIFF and more columns.
> Making methods under ScannerV2#next smaller allows them to be inlined gaining
> us 7% more throughput
> ---------------------------------------------------------------------------------------------------
>
> Key: HBASE-13307
> URL: https://issues.apache.org/jira/browse/HBASE-13307
> Project: HBase
> Issue Type: Sub-task
> Components: Scanners
> Reporter: stack
> Assignee: stack
> Fix For: 2.0.0, 1.1.0
>
> Attachments: 13307.txt, traces.8.svg
>
>
> See parent issue for patch and evidence.
> I was looking at graphs of our scan and found that methods were 'too big' to
> be inlined (looking at jvm compilation and inlining output flags -- see
> parent for list). Changing method size helped some. Let me commit the
> resultant patch.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
