[
https://issues.apache.org/jira/browse/HDFS-10351?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chris Nauroth updated HDFS-10351:
---------------------------------
Attachment: HDFS-10351-HDFS-7240.001.patch
Quoting the JavaDocs of the bulk {{OutputStream#write}} method:
{quote}
The write method of OutputStream calls the write method of one argument on each
of the bytes to be written out. Subclasses are encouraged to override this
method and provide a more efficient implementation.
{quote}
Well, they were sure right! Profiling revealed a ton of function call overhead
from sitting in the base class {{OutputStream#write}}, iterating over the
buffer, and calling the single-byte write method repeatedly. The attached
patch provides a bulk write implementation in the subclass. I have measured a
~30% clock time performance improvement in putting a key. Profiling after the
change shows that this isn't a hotspot anymore.
There are no new tests in this patch. I'm relying on the existing tests or
regression testing, and I'm relying on the performance measure/profiling I
described above as proof that the optimization worked.
> Ozone: Optimize key writes to chunks by providing a bulk write implementation
> in ChunkOutputStream.
> ---------------------------------------------------------------------------------------------------
>
> Key: HDFS-10351
> URL: https://issues.apache.org/jira/browse/HDFS-10351
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: ozone
> Reporter: Chris Nauroth
> Assignee: Chris Nauroth
> Attachments: HDFS-10351-HDFS-7240.001.patch
>
>
> HDFS-10268 introduced the {{ChunkOutputStream}} class as part of end-to-end
> integration of Ozone receiving key content and writing it to chunks in a
> container. That patch provided an implementation of the mandatory
> single-byte write method. We can improve performance by adding an
> implementation of the bulk write method too.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
