Hi, On Thu, May 14, 2009 at 07:42:15AM -0700, Arjan van de Ven wrote: > Andreas Mohr wrote: >> Hi, >> >> On Wed, May 13, 2009 at 01:37:54PM -0700, hong zhang wrote: >>> 14.9% ( 77.3) <interrupt> : iwlagn >>> >>> My question is what 77.3 means here. Does 77.3 mean power consumption or >>> wakeup latency? How does powertop provide power consumption and wakeup >>> latency for my wifi device? >> >> If I'm not entirely mistaken 77.3 means 77.3 wakeups within the last >> interval which translated into 14.9% of the total number of wakeups >> there. > > it's 77.3 *per second* actually....
Humm ok. Somehow I got tricked by the individual readings being _much_ higher than the sum. This can be seen below (moderate activity), but sometimes with much higher activity there's more distortion, too. Wakeups-from-idle per second : 42.8 interval: 15.0s no ACPI power usage estimate available Top causes for wakeups: 32.3% ( 20.8) <kernel core> : hrtimer_start_range_ns (tick_sched_timer) 9.6% ( 6.2) <interrupt> : extra timer interrupt 9.5% ( 6.1) <interrupt> : uhci_hcd:usb3, ath 9.5% ( 6.1) soffice.bin : hrtimer_start_range_ns (hrtimer_wakeup) 7.8% ( 5.0) <kernel IPI> : Rescheduling interrupts 7.2% ( 4.7) hald : schedule_timeout_uninterruptible (process_timeout) 6.1% ( 3.9) <kernel core> : hrtimer_start (tick_sched_timer) 5.4% ( 3.5) <interrupt> : acpi 5.0% ( 3.2) knotify4 : hrtimer_start_range_ns (hrtimer_wakeup) 1.6% ( 1.0) kwin : hrtimer_start_range_ns (hrtimer_wakeup) 0.8% ( 0.5) <kernel core> : add_timer (neigh_periodic_timer) 0.8% ( 0.5) schedtoold : hrtimer_start_range_ns (hrtimer_wakeup) 0.7% ( 0.5) phy3 : ieee80211_authenticate (ieee80211_sta_timer) 0.4% ( 0.3) <kernel module> : add_timer (neigh_periodic_timer) Those additional readings possibly are due to rounded timeouts fireing at the same time as another event, perhaps? Or yet another kernel bug? ;) Andreas Mohr _______________________________________________ Power mailing list [email protected] http://www.bughost.org/mailman/listinfo/power
