Powertop output question

[Ladislav Laska]

Hello,

I have a IBM x60s notebook and have been tuning my Gentoo system for
power consumption for some time. Still I'm a little puzzled by some
outputs from powertop (powertop -d output attached). The lines that
trouble me are:

36.7% ( 58.1) <kernel core> : hrtimer_start_range_ns (tick_sched_timer)
15.3% ( 24.3) <interrupt> : extra timer interrupt

Since I have enabled tickless kernel (kernel config attached), I don't
know where these interrupts come from. Can someone please point
somewhere or give a hint? Is it possible to get more info from
powertop or do I have to dig somewhere else?

Note that iwl3945 is probably normal and goes down after some idle time.

Regards Ladislav Laska
S pozdravem Ladislav Laska
---
xmpp/jabber: ladislav.laska-***@public.gmane.org

[Florian Reitmeir]

Hi,

Post by Ladislav Laska
I have a IBM x60s notebook and have been tuning my Gentoo system for
power consumption for some time. Still I'm a little puzzled by some
outputs from powertop (powertop -d output attached). The lines that
36.7% ( 58.1) <kernel core> : hrtimer_start_range_ns (tick_sched_timer)
15.3% ( 24.3) <interrupt> : extra timer interrupt
Since I have enabled tickless kernel (kernel config attached), I don't
know where these interrupts come from. Can someone please point
somewhere or give a hint? Is it possible to get more info from
powertop or do I have to dig somewhere else?
Note that iwl3945 is probably normal and goes down after some idle time.

IMHO on my notebook (also x60s) adding a 'hpet=force' to the kernel
commandline resolved those wakeups.

[Ladislav Laska]

Hello,

I want to let you know that the problem of mine is resolved.

The problem was here:

C2 9.4ms (76.8%)
C3 3.8ms (22.0%)

That means someone was not allowing cpu to sleep deep enough and it
was not related to the interrupts. I have determined that it was kde3
window manageer (kwin) causing problems and switched to openbox-kde
session. Now cpu sleeps in c3 all the time, although almost all of the
interrupts still happen.

I have also tried measure the consumption without all the interrupts
(without kde stuff running, i.e. with openbox session only) and it
lowered power consumption by ~0.1 - tha't pretty small, considering
that getting rid of kwin saved me around 2-4W.

Now I can run at 9W without wireless and minimal brightness. That's
good enough for me :-)

Regards Ladislav Laska
S pozdravem Ladislav Laska
---
xmpp/jabber: ***@jabber.cz



On Thu, Oct 29, 2009 at 12:50 PM, Ladislav Laska

(sorry for incomplete mail, "Send" button got in my way)
... Could this be a bug in bios? In that case, I could try upgrading
bios (I have some version from 2008 and more recent versions are
available I think, but I don't want to mess with it until I'm at
pretty sure it could be it)
Regards Ladislav Laska
S pozdravem Ladislav Laska
---
On Thu, Oct 29, 2009 at 12:47 PM, Ladislav Laska

I have just tried it, but it's the same.
Could
Regards Ladislav Laska
S pozdravem Ladislav Laska
---

Post by Florian Reitmeir
Hi,

Post by Ladislav Laska
I have a IBM x60s notebook and have been tuning my Gentoo system for
power consumption for some time. Still I'm a little puzzled by some
outputs from powertop (powertop -d output attached). The lines that
 36.7% ( 58.1)     <kernel core> : hrtimer_start_range_ns (tick_sched_timer)
 15.3% ( 24.3)       <interrupt> : extra timer interrupt
Since I have enabled tickless kernel (kernel config attached), I don't
know where these interrupts come from. Can someone please point
somewhere or give a hint? Is it possible to get more info from
powertop or do I have to dig somewhere else?
Note that iwl3945 is probably normal and goes down after some idle time.

IMHO on my notebook (also x60s) adding a 'hpet=force' to the kernel
commandline resolved those wakeups.
--
Florian Reitmeir
_______________________________________________
Power mailing list
http://www.bughost.org/mailman/listinfo/power