Friday, February 8, 2008

results

On Windows i managed to cut power down to 7W.
On Linux the results were very poor.


Highest brightness, mmc and irda stuff unloaded:

PowerTOP 1.9 (C) 2007 Intel Corporation

Collecting data for 15 seconds
Cn Avg residency
C0 (cpu running) ( 0.3%)
C1 0.0ms ( 0.0%)
C2 27.2ms (69.1%)
C3 17.6ms (30.6%)
P-states (frequencies)
1.67 Ghz 0.0%
1333 Mhz 0.0%
1000 Mhz 100.0%
Wakeups-from-idle per second : 42.8 interval: 15.0s
Power usage (ACPI estimate): 14.3W (4.8 hours)
Top causes for wakeups:
63.2% ( 28.3) : uhci_hcd:usb2, iwl3945, HDA Intel
11.2% ( 5.0) : fbcon_add_cursor_timer (cursor_timer_handler)
8.9% ( 4.0) : usb_hcd_poll_rh_status (rh_timer_func)
4.5% ( 2.0) : clocksource_register (clocksource_watchdog)
4.3% ( 1.9) wpa_supplicant : schedule_timeout (process_timeout)
2.1% ( 0.9) : queue_delayed_work_on (delayed_work_timer_fn)
1.3% ( 0.6) : uhci_hcd:usb1, ahci
1.0% ( 0.5) iwl3945 : ieee80211_sta_work (ieee80211_sta_timer)
0.7% ( 0.3) kcryptd : blk_plug_device (blk_unplug_timeout)
0.6% ( 0.3) : neigh_table_init_no_netlink (neigh_periodic_timer)
0.4% ( 0.2) init : schedule_timeout (process_timeout)
0.3% ( 0.1) iwl3945/0 : sta_info_start (sta_info_cleanup)
0.1% ( 0.1) : Rescheduling interrupts
0.1% ( 0.1) sh : start_this_handle (commit_timeout)
0.1% ( 0.1) : page_writeback_init (wb_timer_fn)
0.1% ( 0.1) : end_that_request_last (laptop_timer_fn)
0.1% ( 0.1) syslogd : do_setitimer (it_real_fn)
0.1% ( 0.1) : rs_tx_status (iwl_bg_rate_scale_flush)
0.1% ( 0.1) cron : do_nanosleep (hrtimer_wakeup)
0.1% ( 0.1) : igmp_rcv (igmp_gq_timer_expire)
0.1% ( 0.1) : ip_rt_init (delayed_work_timer_fn)


As you can see, the USB stuff sucks. i will unload uhci_hcd and ehci_hcd:

PowerTOP 1.9 (C) 2007 Intel Corporation

Collecting data for 15 seconds
Cn Avg residency
C0 (cpu running) ( 0.1%)
C1 0.0ms ( 0.0%)
C2 9.9ms ( 0.1%)
C3 45.8ms (99.8%)
P-states (frequencies)
1.67 Ghz 0.0%
1333 Mhz 0.0%
1000 Mhz 100.0%
Wakeups-from-idle per second : 21.9 interval: 15.0s
Power usage (ACPI estimate): 12.8W (5.4 hours)
Top causes for wakeups:
51.7% ( 10.3) : iwl3945, HDA Intel
25.0% ( 5.0) : fbcon_add_cursor_timer (cursor_timer_handler)
9.3% ( 1.9) wpa_supplicant : schedule_timeout (process_timeout)
4.7% ( 0.9) : queue_delayed_work_on (delayed_work_timer_fn)
2.3% ( 0.5) iwl3945 : ieee80211_sta_work (ieee80211_sta_timer)
2.0% ( 0.4) : ahci
1.3% ( 0.3) : neigh_table_init_no_netlink (neigh_periodic_timer)
1.0% ( 0.2) init : schedule_timeout (process_timeout)
0.7% ( 0.1) iwl3945/0 : sta_info_start (sta_info_cleanup)
0.7% ( 0.1) kcryptd : blk_plug_device (blk_unplug_timeout)
0.3% ( 0.1) : PS/2 keyboard/mouse/touchpad
0.3% ( 0.1) sh : start_this_handle (commit_timeout)
0.3% ( 0.1) : page_writeback_init (wb_timer_fn)
0.3% ( 0.1) : end_that_request_last (laptop_timer_fn)


wlan still takes a lot of power, so: ifdown wlan0_rename; modprobe -r iwl3945 and i even switched the hardware switch off.

PowerTOP 1.9 (C) 2007 Intel Corporation

Collecting data for 15 seconds
Cn Avg residency
C0 (cpu running) ( 0.1%)
C1 0.0ms ( 0.0%)
C2 17.2ms ( 0.2%)
C3 112.3ms (99.6%)
P-states (frequencies)
1.67 Ghz 0.0%
1333 Mhz 0.0%
1000 Mhz 100.0%
Wakeups-from-idle per second : 9.0 interval: 15.0s
Power usage (ACPI estimate): 11.7W (5.8 hours)
Top causes for wakeups:
65.8% ( 5.0) : fbcon_add_cursor_timer (cursor_timer_handler)
12.3% ( 0.9) : queue_delayed_work_on (delayed_work_timer_fn)
7.0% ( 0.5) : ahci
3.5% ( 0.3) : Rescheduling interrupts
3.5% ( 0.3) : neigh_table_init_no_netlink (neigh_periodic_timer)
2.6% ( 0.2) kcryptd : blk_plug_device (blk_unplug_timeout)
2.6% ( 0.2) init : schedule_timeout (process_timeout)
0.9% ( 0.1) : PS/2 keyboard/mouse/touchpad
0.9% ( 0.1) : end_that_request_last (laptop_timer_fn)
0.9% ( 0.1) : page_writeback_init (wb_timer_fn)


After setting the brightness to the lowest state you get this:

PowerTOP 1.9 (C) 2007 Intel Corporation

Collecting data for 15 seconds
Cn Avg residency
C0 (cpu running) ( 0.1%)
C1 0.0ms ( 0.0%)
C2 0.0ms ( 0.0%)
C3 137.4ms (99.9%)
P-states (frequencies)
1.67 Ghz 5.3%
1333 Mhz 0.0%
1000 Mhz 94.7%
Wakeups-from-idle per second : 7.3 interval: 15.0s
Power usage (ACPI estimate): 9.5W (7.1 hours)
Top causes for wakeups:
69.4% ( 5.0) : fbcon_add_cursor_timer (cursor_timer_handler)
14.8% ( 1.1) : queue_delayed_work_on (delayed_work_timer_fn)
4.6% ( 0.3) : Rescheduling interrupts
3.7% ( 0.3) : neigh_table_init_no_netlink (neigh_periodic_timer)
2.8% ( 0.2) init : schedule_timeout (process_timeout)
0.9% ( 0.1) : PS/2 keyboard/mouse/touchpad
0.9% ( 0.1) syslogd : do_setitimer (it_real_fn)
0.9% ( 0.1) : page_writeback_init (wb_timer_fn)
0.9% ( 0.1) sh : start_this_handle (commit_timeout)
0.9% ( 0.1) : inet_initpeers (peer_check_expire)

Add X on top of it, with an xterm running in dwm and you get only 0.2W more:

PowerTOP 1.9 (C) 2007 Intel Corporation

Collecting data for 15 seconds
Cn Avg residency
C0 (cpu running) ( 0.0%)
C1 0.0ms ( 0.0%)
C2 0.0ms ( 0.0%)
C3 218.7ms (100.0%)
P-states (frequencies)
1.67 Ghz 0.0%
1333 Mhz 0.0%
1000 Mhz 100.0%
Wakeups-from-idle per second : 4.6 interval: 15.0s
Power usage (ACPI estimate): 9.7W (6.9 hours)
Top causes for wakeups:
26.7% ( 1.1) Xorg : do_setitimer (it_real_fn)
26.7% ( 1.1) Xorg : schedule_timeout (process_timeout)
23.3% ( 0.9) : queue_delayed_work_on (delayed_work_timer_fn)
6.7% ( 0.3) : neigh_table_init_no_netlink (neigh_periodic_timer)
5.0% ( 0.2) init : schedule_timeout (process_timeout)
1.7% ( 0.1) : PS/2 keyboard/mouse/touchpad
1.7% ( 0.1) : Rescheduling interrupts
1.7% ( 0.1) syslogd : do_setitimer (it_real_fn)
1.7% ( 0.1) cron : do_nanosleep (hrtimer_wakeup)
1.7% ( 0.1) sh : start_this_handle (commit_timeout)
1.7% ( 0.1) : ip_rt_init (delayed_work_timer_fn)
1.7% ( 0.1) : page_writeback_init (wb_timer_fn)

Even with these unloaded modules the x60s consumes more power than on windows. The cpu is working quite well. But the other components of the computer still consume too much. That's bad. Is there something else one can do? Something i have overseen?

No comments: