Re: [BENCHMARK] 100Hz v 1000Hz with contest

From: Zwane Mwaikambo (zwane@linuxpower.ca)
Date: Mon Jun 02 2003 - 22:36:13 EST


On Tue, 3 Jun 2003, Con Kolivas wrote:

> I've attempted to answer the question does 1000Hz hurt responsiveness in 2.5
> as much as I've found in 2.4; since subjectively the difference wasn't there
> in 2.5. Using the same config with preempt enabled here are results from
> 2.5.70-mm3 set at default 1000Hz and at 100Hz (mm31):

Thanks for carrying out these =)

> ctar_load:
> Kernel [runs] Time CPU% Loads LCPU% Ratio
> 2.5.70-mm3 3 114 70.2 1.0 5.3 1.44
> 2.5.70-mm31 3 105 73.3 0.7 3.8 1.36

> dbench_load:
> Kernel [runs] Time CPU% Loads LCPU% Ratio
> 2.5.70-mm3 4 313 24.3 5.0 56.9 3.96
> 2.5.70-mm31 4 297 24.9 4.5 52.5 3.86
>
> At first glance everything looks faster at 100Hz. However it is well known
> that it will take slightly longer even with no load at 1000Hz. Taking that
> into consideration and looking more at the final ratios than the absolute
> numbers it is apparent that the difference is statistically insignificant,
> except on ctar_load.

What about dbench_load?
 
> Previously I had benchmark results on 1000Hz which showed preempt improved the
> results in a few of the loads. For my next experiment I will compare 100Hz
> with preempt to 100Hz without.

Cheers,
        Zwane

-- 
function.linuxpower.ca
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Sat Jun 07 2003 - 22:00:18 EST