hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From arkady borkovsky <ark...@yahoo-inc.com>
Subject Re: [jira] Created: (HADOOP-2852) Update gridmix to avoid artificially long tail
Date Tue, 19 Feb 2008 21:54:47 GMT
Does not this make the test closer to reality?
In real life, you cannot rely on the optimal order of job submissions...

-- ab
On Feb 19, 2008, at 1:24 PM, Chris Douglas (JIRA) wrote:

> Update gridmix to avoid artificially long tail
> ----------------------------------------------
>
>                  Key: HADOOP-2852
>                  URL: https://issues.apache.org/jira/browse/ 
> HADOOP-2852
>              Project: Hadoop Core
>           Issue Type: Test
>           Components: test
>     Affects Versions: 0.16.0
>             Reporter: Chris Douglas
>              Fix For: 0.16.1
>
>
> The MaxEntropy test in the gridmix benchmark is submitted late into  
> the queue, iterating past the point where the cluster is saturated.  
> This dilutes the throughput measurement- the purpose of this  
> benchmark- by making the tail overly dependent on the performance  
> of a single job.
>
> -- 
> This message is automatically generated by JIRA.
> -
> You can reply to this email to add a comment to the issue online.
>


Mime
View raw message