RFR: Deprecate -sc, -tc, -otss options

[Sergey Kuksenko]

On 08/21/2013 10:50 PM, Staffan Friberg wrote:
> Hi,
> 
> Agree that scaling threads was one of the less used features.
> 
> Would the recommendation for doing scaling tests be?
>     Multiple runs with different thread counts?
I am using this version.
Plans for Java API should give us full flexibility here in future.

>     Built in as part of the microbenchmark?
It is hard to built in. Too much parameters (HW dependent) to choose
proper scaling scheme.

> 
> Cheers,
> Staffan
> 
> On 2013-08-21 11:35, Sergey Kuksenko wrote:
>> Agreed.
>>
>> I used these options some time ago and realized that they are not
>> convenient for main purpose - scalability measurement. I switched to
>> customs scripts long time ago. And all code around is really awful.
>>
>>
>>
>> On 08/21/2013 09:41 PM, Aleksey Shipilev wrote:
>>> Hi,
>>>
>>> This is the heads up / review request for pending code sweep work.
>>>
>>> This time, we clean up JMH runners. The runner code was over-patched
>>> over the years, and now looks like a fur ball of scaling, thread count,
>>> and thread sub-statistics code.
>>>
>>> We don't see the seasoned users of these options. Also, in the wake of
>>> Java API for JMH which will allow to access different benchmark modes
>>> via the usual API, these options become the burden we don't see a clear
>>> value to support.
>>>
>>> Please speak up if you think otherwise.
>>>
>>> Here is the changeset:
>>>   
>>> http://cr.openjdk.java.net/~shade/jmh/deprecate-sc-tc-otss-1.changeset
>>>

-- 
Best regards,
Sergey Kuksenko