Re: cpufreq: shouldn't scaling_min_freq be lower?

[Valdis.Kletnieks@vt.edu]

[-- Attachment #1: Type: text/plain, Size: 1810 bytes --]

On Mon, 06 Dec 2004 13:34:47 GMT, Alan Cox said:
> On Sul, 2004-12-05 at 23:29, Thomas Bettler wrote:
> > > Windows tends to use a combination of CPU scaling and throttling to get
> > > the processor that slow. Take a look at
> > > /proc/acpi/processor/*/throttling
> > 
> > Is there a throttle daemon for Linux? It would be great to use this.
> 
> There are a few. cpuspeed for example

Sorry for the late follow-up.  The 'cpuspeed' shipped by Fedora Core as
part of kernel-utils doesn't seem to understand throttling, it only does
frequency stepping.  So although I *have*:

[/proc/acpi/processor/CPU0]2 cat power 
active state:            C2
default state:           C1
bus master activity:     00000000
states:
    C1:                  promotion[C2] demotion[--] latency[000] usage[00000010]
   *C2:                  promotion[--] demotion[C1] latency[050] usage[03754116]
    C3:                  <not supported>
[/proc/acpi/processor/CPU0]2 cat throttling 
state count:             8
active state:            T0
states:
   *T0:                  00%
    T1:                  12%
    T2:                  25%
    T3:                  37%
    T4:                  50%
    T5:                  62%
    T6:                  75%
    T7:                  87%

I can basically only tell the CPU to go at 1.2Ghz or 1.6Ghz.

Right now, I use:
cpuspeed -d -p 10 50 -t /proc/acpi/thermal_zone/THM/temperature 85 -a /proc/acpi/ac_adapter/AC/state
which is *almost* OK, thought if further power savings were to be obtained by
dropping it to somewhere in the T5-T7 range if we're at low speed and *still*
not using much CPU, that would be nice (modulo the usual considerations about
latency for T5->T0 transitions for a "Intel(R) Pentium(R) 4 Mobile CPU 1.60GHz"-
I realize other chipsets will have other latencies..)


[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]