how reliable JMH's results are

Vicente Romero vicente.romero at oracle.com
Tue Mar 26 15:44:19 UTC 2019



On 3/26/19 11:07 AM, Maurizio Cimadamore wrote:
>
> On 26/03/2019 12:38, Vicente Romero wrote:
>> the only difference here is the benchmark, the builds used are 
>> exactly the same, and I think those original data you are referring 
>> to have a lot of noise, so I would trust more the table I published 
>> yesterday 
>
> Not sure - it seems to me that after V7 we never had another run which 
> featured similar 'high' scores for JDK13.
>
> Anyway, if we can't seem to no longer be able to reproduce these 
> numbers, I think we have to file them under 'anomaly/bogus run' 
> category. If there was a real issue, I would expect the numbers to 
> have popped out in some others of your run.

yes I think that that run was an outlier

>
> Maurizio
>



More information about the amber-dev mailing list