Ubuntu

Bug #152808
Comment #7

Comment 7 for bug 152808

Revision history for this message

Eduardo Silva (edsiper) wrote on 2008-03-07: Re: [Bug 152808] Re: g-p-m always try to set the main policy to 'ondemand'

Another comments:

if the computer is on the desk and don't being used, is the CPU being
used? , is there any big difference between ondemand and performance
on that case?

IIRC CPU scaling helps to save the battery life on the laptop cases,
but if I'm plugged to AC I would like to have my laptop working as
fast as possible. Final user needs to feel their laptop faster as they
want, this makes sense to me:

case 1: laptop plugged (AC) = use performance (we don't need to save anything)
case 2: laptop using battery as power resource = use ondemand

let's ask to the users... this is not a problem for geeks...

On Fri, Mar 7, 2008 at 10:05 AM, Eduardo Silva <email address hidden> wrote:
> > I think leaving it at "ondemand" is the solution. It will switch to full
> > performance when the system is busy, so users will get the performance
> > they expect from their hardware. Anyone who actually wants to change it
> > will already know what cpu scaling is and won't be scared of gconf, surely?
>
> I agree with that, my point is about people that don't know about cpu scaling...
>
>
> > my laptop is a 1.2ghz pentium m, so it's hardly the fastest thing in the
> > world, but I don't think there would be any benefit to not using
> > ondemand. It's not going to make the thing noticeably faster.
> > Can you demonstrate otherwise?
>
> I open this bug because I feel that ondemand doesn't works as well as
> I would like, it makes me feel my laptop a little slower... your
> laptop is working properly, that's not my case... maybe will be useful
> some feedback from another users. Here we have two points of view for
> two different cases/machines... Maybe my problem is an isolated one,
> maybe not... is there any forum where we can talk about this ?, more
> opinions would be very useful.
>
> Ed.-
>
> >
> > --
>
>
> > Chris Jones
> >
> > --
> >
> >
> > g-p-m always try to set the main policy to 'ondemand'
> > https://bugs.launchpad.net/bugs/152808
> > You received this bug notification because you are a direct subscriber
> > of the bug.
> >
>

Another comments:

if the computer is on the desk and don't being used, is the CPU being
used? , is there any big difference between ondemand and performance
on that case?

case 1: laptop plugged (AC) = use performance (we don't need to save anything)
case 2: laptop using battery as power resource =  use ondemand

let's ask to the users... this is not a problem for geeks...

On Fri, Mar 7, 2008 at 10:05 AM, Eduardo Silva <edsiper@gmail.com> wrote:
> >  I think leaving it at "ondemand" is the solution. It will switch to full
>  >  performance when the system is busy, so users will get the performance
>  >  they expect from their hardware. Anyone who actually wants to change it
>  >  will already know what cpu scaling is and won't be scared of gconf, surely?
>
>  I agree with that, my point is about people that don't know about cpu scaling...
>
>
>  >  my laptop is a 1.2ghz pentium m, so it's hardly the fastest thing in the
>  >  world, but I don't think there would be any benefit to not using
>  >  ondemand. It's not going to make the thing noticeably faster.
>  >  Can you demonstrate otherwise?
>
>  I open this bug because I feel that ondemand doesn't works as well as
>  I would like, it makes me feel my laptop a little slower... your
>  laptop is working properly, that's not my case... maybe will be useful
>  some feedback from another users. Here we have two points of view for
>  two different cases/machines... Maybe my problem is an isolated one,
>  maybe not... is there any forum where we can talk about this ?, more
>  opinions would be very useful.
>
>  Ed.-
>
>  >
>  >  --
>
>
> >  Chris Jones
>  >
>  >  --
>  >
>  >
>  > g-p-m always try to set the main policy to 'ondemand'
>  >  https://bugs.launchpad.net/bugs/152808
>  >  You received this bug notification because you are a direct subscriber
>  >  of the bug.
>  >
>