All of lore.kernel.org
 help / color / mirror / Atom feed
* speedstep and deep sleep on celeron-M
@ 2004-09-18  8:21 Christophe Couronne
  2004-09-18 21:58 ` Dominik Brodowski
  0 siblings, 1 reply; 5+ messages in thread
From: Christophe Couronne @ 2004-09-18  8:21 UTC (permalink / raw)
  To: cpufreq

Hi everybody !

I've just bought a new laptop based on a celeron-M processor (an
IBM thinkpad R51). According to intel, this family of processor doesn't
support speedstep, which is replaced by another technology called "deep
sleep". I can't find more information about it.

I would like to know if the speedstep functionnality can be enabled
anyway. The celeron-M processor seems to be based on a pentium-M banias
(same family, same model id, same stepping) and the p4-clockmod module
can be loaded (but not the speedstep-centrino). In fact, the celeron-M
seems to be just a pentium-M banias without speedstep (?) and with a 
512Ko L2
memory cache.

When the p4-clockmod is enabled, a new /sys entry is created. below is a
listing of the directory :

brasserie:/sys/devices/system/cpu/cpu0/cpufreq# ls -l
total 0
-r--------  1 root root    0 2004-09-18 05:17 cpuinfo_cur_freq
-r--r--r--  1 root root 4096 2004-09-18 05:16 cpuinfo_max_freq
-r--r--r--  1 root root 4096 2004-09-18 05:16 cpuinfo_min_freq
-r--r--r--  1 root root 4096 2004-09-18 03:12 scaling_available_frequencies
-r--r--r--  1 root root 4096 2004-09-18 05:22 scaling_available_governors
-r--r--r--  1 root root 4096 2004-09-18 05:22 scaling_cur_freq
-r--r--r--  1 root root 4096 2004-09-18 05:22 scaling_driver
-rw-r--r--  1 root root    0 2004-09-18 03:12 scaling_governor
-rw-r--r--  1 root root 4096 2004-09-18 05:22 scaling_max_freq
-rw-r--r--  1 root root 4096 2004-09-18 05:22 scaling_min_freq

The problem is that the scaling_governor is set to "performance" and
can't be changed, so the processor always runs at his max frequency.

Is there people which have succeeded to enable power management (deep
sleep or speedstep) on a Celeron-M based laptop? 

Thank you for any information.

Best regards.

   Christophe Couronne.

-- 
Christophe Couronne 
xof@[phplop|calvix|altern].org

* phplop.org - yet another php tribioune !
* E = MC² <=> WE = WMC² <=> Le Week-end, je moule !

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: speedstep and deep sleep on celeron-M
  2004-09-18  8:21 speedstep and deep sleep on celeron-M Christophe Couronne
@ 2004-09-18 21:58 ` Dominik Brodowski
  2004-09-20 10:47   ` Bruno Ducrot
  0 siblings, 1 reply; 5+ messages in thread
From: Dominik Brodowski @ 2004-09-18 21:58 UTC (permalink / raw)
  To: Christophe Couronne; +Cc: cpufreq

Hi,

> I've just bought a new laptop based on a celeron-M processor (an
> IBM thinkpad R51). According to intel, this family of processor doesn't
> support speedstep, which is replaced by another technology called "deep
> sleep".

Uh? SpeedStep replaced by "Deep Sleep"? Where did you find that, ehm, joke? 

> I would like to know if the speedstep functionnality can be enabled
> anyway. The celeron-M processor seems to be based on a pentium-M banias
> (same family, same model id, same stepping)

... but with SpeedStep support not present in the CPU, AFAIK. So there's no
way to enable celeron-M.

> and the p4-clockmod module can be loaded (but not the speedstep-centrino).
... which only offers frequency modulation instead of frequency scaling, so
it will only save you minimal amounts of battery power. There should be a
warning (in dmesg) when loading the p4-clockmod module which informs of 
this fact.

> The problem is that the scaling_governor is set to "performance" and
> can't be changed, so the processor always runs at his max frequency.

you can only change the CPUfreq governor if more than one is loaded. You can
see which ones are loaded in scaling_available_governors; and you can load,
for example, the module "cpufreq_userspace" in case you built the userspace
cpufreq governor as module.

> Is there people which have succeeded to enable power management (deep
> sleep or speedstep) on a Celeron-M based laptop? 

Deep Sleep might be supported (depending on the BIOS in your notebook) by
the ACPI processor subsystem. It's a form of "idling" which is enabled when
the CPU has nothing to do, and may be C-State 3. For more information on
that, check the current ACPI processor information at http://acpi.sf.net;
also bug #1958 at bugzilla.kernel.org may be of interest as it can enable
more C-States. However, it depends on the BIOS which CPU Power State (Deep
Sleep is one of them) is enabled as what C-State. It has nothing to do with
CPUfreq, though.

	Dominik

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: speedstep and deep sleep on celeron-M
  2004-09-18 21:58 ` Dominik Brodowski
@ 2004-09-20 10:47   ` Bruno Ducrot
  2004-09-20 17:07     ` Dominik Brodowski
  0 siblings, 1 reply; 5+ messages in thread
From: Bruno Ducrot @ 2004-09-20 10:47 UTC (permalink / raw)
  To: Christophe Couronne, cpufreq

On Sat, Sep 18, 2004 at 11:58:39PM +0200, Dominik Brodowski wrote:
> Hi,
> 
> Deep Sleep might be supported (depending on the BIOS in your notebook) by
> the ACPI processor subsystem. It's a form of "idling" which is enabled when
> the CPU has nothing to do, and may be C-State 3. For more information on
> that, check the current ACPI processor information at http://acpi.sf.net;
> also bug #1958 at bugzilla.kernel.org may be of interest as it can enable
> more C-States. However, it depends on the BIOS which CPU Power State (Deep
> Sleep is one of them) is enabled as what C-State. It has nothing to do with
> CPUfreq, though.


Actually, deep sleep state is C3, whereas deeper sleep state is supposed
to be C4 (though it's possible to enter deeper sleep state instead of
deep sleep state IIRC).

Cheers,

-- 
Bruno Ducrot

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: speedstep and deep sleep on celeron-M
  2004-09-20 10:47   ` Bruno Ducrot
@ 2004-09-20 17:07     ` Dominik Brodowski
  2004-09-25 20:30       ` Christophe Couronne
  0 siblings, 1 reply; 5+ messages in thread
From: Dominik Brodowski @ 2004-09-20 17:07 UTC (permalink / raw)
  To: Bruno Ducrot; +Cc: cpufreq

On Mon, Sep 20, 2004 at 12:47:30PM +0200, Bruno Ducrot wrote:
> On Sat, Sep 18, 2004 at 11:58:39PM +0200, Dominik Brodowski wrote:
> > Hi,
> > 
> > Deep Sleep might be supported (depending on the BIOS in your notebook) by
> > the ACPI processor subsystem. It's a form of "idling" which is enabled when
> > the CPU has nothing to do, and may be C-State 3. For more information on
> > that, check the current ACPI processor information at http://acpi.sf.net;
> > also bug #1958 at bugzilla.kernel.org may be of interest as it can enable
> > more C-States. However, it depends on the BIOS which CPU Power State (Deep
> > Sleep is one of them) is enabled as what C-State. It has nothing to do with
> > CPUfreq, though.
> 
> 
> Actually, deep sleep state is C3, whereas deeper sleep state is supposed
> to be C4 (though it's possible to enter deeper sleep state instead of
> deep sleep state IIRC).

Excactly. And Deep Sleep is nothing new, Deeper Sleep is.

	Dominik

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: speedstep and deep sleep on celeron-M
  2004-09-20 17:07     ` Dominik Brodowski
@ 2004-09-25 20:30       ` Christophe Couronne
  0 siblings, 0 replies; 5+ messages in thread
From: Christophe Couronne @ 2004-09-25 20:30 UTC (permalink / raw)
  To: cpufreq

Dominik Brodowski a écrit :

>On Mon, Sep 20, 2004 at 12:47:30PM +0200, Bruno Ducrot wrote:
>  
>
>> Actually, deep sleep state is C3, whereas deeper sleep state is supposed
>>
>>to be C4 (though it's possible to enter deeper sleep state instead of
>>deep sleep state IIRC).
>>    
>>
>
>Excactly. And Deep Sleep is nothing new, Deeper Sleep is.
>
>	Dominik.
>
Everything is clearer now. I've succeeded to enable the CPU frequency modulation with p4-clockmod and the C3 sleep state too. Thanks a lot.

Christophe.

-- 
Christophe Couronne 
xof@[phplop|calvix|altern].org

* phplop.org - yet another php tribioune !
* E = MC² <=> WE = WMC² <=> Le Week-end, je moule !

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2004-09-25 20:30 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-09-18  8:21 speedstep and deep sleep on celeron-M Christophe Couronne
2004-09-18 21:58 ` Dominik Brodowski
2004-09-20 10:47   ` Bruno Ducrot
2004-09-20 17:07     ` Dominik Brodowski
2004-09-25 20:30       ` Christophe Couronne

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.