hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "SammiChen (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-8410) Add computation time metrics to datanode for ECWorker
Date Thu, 20 Oct 2016 05:03:58 GMT

    [ https://issues.apache.org/jira/browse/HDFS-8410?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15590805#comment-15590805

SammiChen commented on HDFS-8410:

It costs about 2 milliseconds in my desktop to decode a strip group with 6 blocks, each block
is 64k. This decoding time merely depends on how fast the CPU is.  It's "Intel(R) Core(TM)
i5-4460 CPU @ 3.20GHz" with 4  cores in my desktop, not the leading edge CPU model. Given
that CPU is becoming more and more powerful, I think it is not safe to use millisecond granularity
to record one time decoding time. We can choose between nanosecond or microsecond. I would
prefer nanosecond for one reason. it can be directly get through {{System.nanoTime()}}. If
microsecond is used, there is one extra division of 1000. That's not good from performance
point of view. And a {{long}} number can host nanoseconds, representing hundreds of years.
So there is not going to a overflow quickly. 

> Add computation time metrics to datanode for ECWorker
> -----------------------------------------------------
>                 Key: HDFS-8410
>                 URL: https://issues.apache.org/jira/browse/HDFS-8410
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>            Reporter: Li Bo
>            Assignee: SammiChen
>         Attachments: HDFS-8410-001.patch, HDFS-8410-002.patch, HDFS-8410-003.patch, HDFS-8410-004.patch
> This is a sub task of HDFS-7674. It adds time metric for ec decode work.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org

View raw message