Re: Influence of optimization level, preemption and scheduler on boot time

Re: Influence of optimization level, preemption and scheduler on boot time

[Arjan van de Ven]

On Mon, 21 Sep 2009 23:25:53 +1200
Johannes Buchner <buchner.johannes@gmx.at> wrote:

> Hi all.
> 
> I measured the kernel (and system) boot times while varying the
> parameters:
>  - Optimization level: -Os, -O2 and also -O3
>  - Preemptive model
>  - Scheduler: CFQ, Anticipatory, Deadline, Noop
> 
> My conclusion was that the optimization level and the preemptive model
> had no significant influence on speed. CFQ let my system boot several
> seconds faster than the other schedulers.
> 
> Graphs can be found at:
> http://johannes.jakeapp.com/blog/?p=913
> 
> This conclusion may not be true for all situations, but I found it
> interesting that the optimization level is so irrelevant.
> 

it's interesting to see that the IO scheduler mattered.. 
I would think that (s)readahead makes the IO scheduler irrelevant for
boot time...

--
Arjan van de Ven 	Intel Open Source Technology Centre
For development, discussion and tips for power savings, 
visit http://www.lesswatts.org

Influence of optimization level, preemption and scheduler on boot time

[Johannes Buchner]

[-- Attachment #1: Type: text/plain, Size: 813 bytes --]

Hi all.

I measured the kernel (and system) boot times while varying the
parameters:
 - Optimization level: -Os, -O2 and also -O3
 - Preemptive model
 - Scheduler: CFQ, Anticipatory, Deadline, Noop

My conclusion was that the optimization level and the preemptive model
had no significant influence on speed. CFQ let my system boot several
seconds faster than the other schedulers.

Graphs can be found at:
http://johannes.jakeapp.com/blog/?p=913

This conclusion may not be true for all situations, but I found it
interesting that the optimization level is so irrelevant.

Cheers,
Johannes

(I am not subscribed to the list.)
-- 
Emails können geändert, gefälscht und eingesehen werden. Signiere oder
verschüssele deine Mails mit GPG.
http://web.student.tuwien.ac.at/~e0625457/pgp.html

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: Influence of optimization level, preemption and scheduler on boot time

[Johannes Buchner]

[-- Attachment #1: Type: text/plain, Size: 1362 bytes --]

On Mon, 21 Sep 2009 13:20:59 +0200
Arjan van de Ven <arjan@infradead.org> wrote:

> On Mon, 21 Sep 2009 23:25:53 +1200
> Johannes Buchner <buchner.johannes@gmx.at> wrote:
> 
> > Hi all.
> > 
> > I measured the kernel (and system) boot times while varying the
> > parameters:
> >  - Optimization level: -Os, -O2 and also -O3
> >  - Preemptive model
> >  - Scheduler: CFQ, Anticipatory, Deadline, Noop
> > 
> > My conclusion was that the optimization level and the preemptive
> > model had no significant influence on speed. CFQ let my system boot
> > several seconds faster than the other schedulers.
> > 
> > Graphs can be found at:
> > http://johannes.jakeapp.com/blog/?p=913
> > 
> > This conclusion may not be true for all situations, but I found it
> > interesting that the optimization level is so irrelevant.
> > 
> 
> it's interesting to see that the IO scheduler mattered.. 
> I would think that (s)readahead makes the IO scheduler irrelevant for
> boot time...

I did not use (s)readahead though in this measurements. Trying
readahead-list with CFQ did not bring me any improvement. Maybe, for
the other schedulers, it brings the speed on par with CFQ.


-- 
Emails können geändert, gefälscht und eingesehen werden. Signiere oder
verschüssele deine Mails mit GPG.
http://web.student.tuwien.ac.at/~e0625457/pgp.html

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]