[
https://issues.apache.org/jira/browse/HBASE-20716?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16526516#comment-16526516
]
stack commented on HBASE-20716:
-------------------------------
Looks great.
You might change this to an abstract class rather than Interface when you get
to JMH'ing: "interface Converter {" (See "If you are choosing between
interfaces and abstract classes, interfaces should not be your choice." at end
of https://shipilev.net/blog/2015/black-magic-method-dispatch/ -- needs
measurement, might not matter if JIT decides this monomorphic).
nit: lets come up w/ better name than 'Converter'... will do for now.
... but looks right direction to me. Thanks.
> Unsafe access cleanup
> ---------------------
>
> Key: HBASE-20716
> URL: https://issues.apache.org/jira/browse/HBASE-20716
> Project: HBase
> Issue Type: Improvement
> Components: Performance
> Reporter: stack
> Assignee: Sahil Aggarwal
> Priority: Critical
> Labels: beginner
> Attachments: HBASE-20716.master.001.patch, Screen Shot 2018-06-26 at
> 11.37.49 AM.png
>
>
> We have two means of getting at unsafe; UnsafeAccess and then internal to the
> Bytes class. They are effectively doing the same thing. We should have one
> avenue to Unsafe only.
> Many of our paths to Unsafe via UnsafeAccess traverse flags to check if
> access is available, if it is aligned and the order in which words are
> written on the machine. Each check costs -- especially if done millions of
> times a second -- and on occasion adds bloat in hot code paths. The unsafe
> access inside Bytes checks on startup what the machine is capable off and
> then does a static assign of the appropriate class-to-use from there on out.
> UnsafeAccess does not do this running the checks everytime. Would be good to
> have the Bytes behavior pervasive.
> The benefit of one access to Unsafe only is plain. The benefits we gain
> removing checks will be harder to measure though should be plain when you
> disassemble a hot-path; in a (very) rare case, the saved byte codes could be
> the difference between inlining or not.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
