Hi.
I am having a (not so) funny kind of issue. The issue seems that my CPU seems
to enter into a bad 'state', where even in idle it is in C0 for 2/3rds of the
time. In this state, the interrupts per second go to more than 20K per second!!
This is in contrast to the usual nice state, where I have about 153 per second
and its in C3 for more than 95% of the time. (When I start the notebook, its in
the 'good state', and it goes into this bad state arbitrarily and never
returns. (The transition happens anywhere between an half an hour to 12 hours
from start.) Although the P-states seem to be fine (almost 90 to 100 % in
lowest state of 800 MHz). See the attachments for detailed logs.
That said, I kind of like the bad state, as the system seems more responsive,
especially the hard drive ;-).
About power usage - in the bad state, my power usage is at an estimated ~75W
and in good state, about 55W. BTW, its a 6600 mAh, 14800 mV battery ( cat
/proc/acpi/battery/BAT1/info), i.e., 98 Watt hours, and at a rate of 65W, it
lasts for just about an hour and a half. The battery runs for less than 2 hrs
in windows and just about an hour and a half in linux. (This is with all
devices running, low brightness and browsing with firefox with mp3 music
running in background.)
About the system - its a hp dv5000t notebook with intel centrino duo t2050 @
1.6 GHz. There is a Nvidia GeForce 7400go, and ipw3945 wireless.
Whew, wonder how people manage less than 20W power consumption on their
notebooks with less than 30 interrupts a second. Incidentally, I also have a
desktop with AMD athlon 64X2 3800+ and a nvidia geforce 7300 gs, and it does
manage 90 interrupts per second inside kde and less than 30 interrupts per
second when really_idle. I better had those beasts in my notebook ;-).
Hoping for some workarounds soon.
Thanks,
Hirakendu.
---------------------------------
Boardwalk for $500? In 2007? Ha!
Play Monopoly Here and Now (it's updated for today's economy) at Yahoo! Games. PowerTOP version 1.7 (C) 2007 Intel Corporation
Cn Avg residency (10s) P-states (frequencies)
C0 (cpu running) (67.6%)
C1 0.0ms ( 0.0%) 1.60 Ghz 0.0%
C2 0.0ms (23.0%) 1333 Mhz 0.0%
C3 0.0ms ( 9.4%) 1067 Mhz 0.0%
800 Mhz 100.0%
Wakeups-from-idle per second : 22560.7
Power usage (5 minute ACPI estimate) : 76.9 W (0.9 hours left)
Top causes for wakeups:
47.1% ( 60.9) <interrupt> : uhci_hcd:usb5, nvidia
8.8% ( 11.4) <interrupt> : extra timer interrupt
8.7% ( 11.3) <interrupt> : uhci_hcd:usb4, ipw3945, yenta
6.2% ( 8.0) <kernel core> : usb_hcd_poll_rh_status (rh_timer_func)
5.8% ( 7.5) <interrupt> : libata
4.6% ( 6.0) cpufreq-set : queue_delayed_work_on
(delayed_work_timer_fn)
3.7% ( 4.8) <interrupt> : acpi
1.6% ( 2.1) kicker : schedule_timeout (process_timeout)
1.5% ( 2.0) <kernel core> : queue_delayed_work_on
(delayed_work_timer_fn)
1.4% ( 1.8) X : do_setitimer (it_real_fn)
0.8% ( 1.0) dhcdbd : schedule_timeout (process_timeout)
0.8% ( 1.0) klipper : schedule_timeout (process_timeout)
0.8% ( 1.0) wpa_supplicant : schedule_timeout (process_timeout)
0.8% ( 1.0) kwrapper : do_nanosleep (hrtimer_wakeup)
0.8% ( 1.0) X : nv_start_rc_timer (nv_kern_rc_timer)
0.8% ( 1.0) kded : schedule_timeout (process_timeout)
0.5% ( 0.7) NetworkManager : schedule_timeout (process_timeout)
0.5% ( 0.6) knetworkmanager : schedule_timeout (process_timeout)
0.4% ( 0.5) <interrupt> : sdhci:slot0, eth0
0.4% ( 0.5) hald-addon-stor : schedule_timeout (process_timeout)
0.4% ( 0.5) NetworkManager : e100_up (e100_watchdog)
0.4% ( 0.5) ipw3945/0 : ipw_bg_link_up (delayed_work_timer_fn)
0.4% ( 0.5) gpm : do_nanosleep (hrtimer_wakeup)
0.3% ( 0.4) ipw3945/0 : ipw_post_associate (ipw_kickoff_roaming)
0.2% ( 0.3) <interrupt> : ahci, uhci_hcd:usb3, ohci1394
0.2% ( 0.3) ipw3945/0 : ipw_net_hard_start_xmit
(delayed_work_timer_fn)
PowerTOP version 1.7 (C) 2007 Intel Corporation
Cn Avg residency (15s) P-states (frequencies)
C0 (cpu running) ( 2.6%)
C1 0.0ms ( 0.0%) 1.60 Ghz 0.9%
C2 1.6ms ( 1.0%) 1333 Mhz 0.0%
C3 6.1ms (96.5%) 1067 Mhz 0.0%
800 Mhz 99.1%
Wakeups-from-idle per second : 163.2
Power usage (5 minute ACPI estimate) : 51.7 W (1.5 hours left)
Top causes for wakeups:
45.2% ( 60.9) <interrupt> : uhci_hcd:usb5, nvidia
15.7% ( 21.2) <interrupt> : uhci_hcd:usb4, ipw3945, yenta
7.6% ( 10.3) <interrupt> : extra timer interrupt
5.9% ( 8.0) <interrupt> : libata
5.9% ( 8.0) <kernel core> : usb_hcd_poll_rh_status (rh_timer_func)
4.4% ( 6.0) cpufreq-set : queue_delayed_work_on
(delayed_work_timer_fn)
1.7% ( 2.3) <kernel core> : queue_delayed_work_on
(delayed_work_timer_fn)
1.5% ( 2.0) kicker : schedule_timeout (process_timeout)
1.4% ( 1.9) X : do_setitimer (it_real_fn)
1.0% ( 1.4) ipw3945/0 : ipw_net_hard_start_xmit
(delayed_work_timer_fn)
1.0% ( 1.3) <interrupt> : acpi
0.7% ( 1.0) wpa_supplicant : schedule_timeout (process_timeout)
0.7% ( 1.0) kwrapper : do_nanosleep (hrtimer_wakeup)
0.7% ( 1.0) dhcdbd : schedule_timeout (process_timeout)
0.7% ( 1.0) klipper : schedule_timeout (process_timeout)
0.7% ( 1.0) X : nv_start_rc_timer (nv_kern_rc_timer)
0.6% ( 0.9) knetworkmanager : schedule_timeout (process_timeout)
0.4% ( 0.6) kded : schedule_timeout (process_timeout)
0.4% ( 0.5) hald-addon-stor : schedule_timeout (process_timeout)
0.4% ( 0.5) NetworkManager : schedule_timeout (process_timeout)
0.3% ( 0.5) <interrupt> : sdhci:slot0, eth0
0.3% ( 0.5) NetworkManager : e100_up (e100_watchdog)
0.3% ( 0.5) gpm : do_nanosleep (hrtimer_wakeup)
0.3% ( 0.5) ipw3945/0 : ipw_bg_link_up (delayed_work_timer_fn)
0.3% ( 0.4) kdesktop : schedule_timeout (process_timeout)
0.2% ( 0.3) avahi-daemon : schedule_timeout (process_timeout)
_______________________________________________
Power mailing list
[email protected]
http://www.bughost.org/mailman/listinfo/power
