Talk:How to reduce power consumption
How about linking to this page from the main page?
--Thinker 21:08, 17 Nov 2005 (CET)
Best battery discharge times
With your best power saving, what kind of battery power times are you guys experiencing? The best that I can do with my t43p with the standard (6cell) battery is 2h, and with the ultrabay battery is 1 hour. All of the power saving 'tricks' - hard drive spindown, even adjusting the brightness of my display, has relatively modest effect (~20% max), although CPU throttling definitely reduces power consumption.
In terms of discharge rate, I really can't get below 19,520 mW/h with the hard disk off, wifi on, cpu down to ~700 MHz, and display on minimal brightness with ATI driver power save enabled. In my "normal usage" environment (WiFi on, word processing / non cpu-intensive programming, etc.) I average 21,000 mW/h.
Of note, I have the UXGA display, which might be a huge power guzzler. I have friends with an X40 that claim 5 hours from the 9 cell...
--gsmenden 16:56, 10 Feb 2006 (EST)
ThinkPad T43, SXGA+, undervolted, fan disabled, minimum brightness, disk off, WiFi off, GPU power saving: 13 to 14W, i.e., slightly over 3 hours with a new 6-cell battery. See How to reduce power consumption.
--Thinker 22:46, 10 February 2006 (CET)
BTW, this machine has 1.5GB of RAM, which probably accounts for a large fraction of a Watt. Also, the DVD drive is plugged in.
--Thinker 08:39, 16 February 2006 (CET)
Down to 12.4W after ejecting the DVD drive.
--Thinker 07:35, 18 February 2006 (CET)
Interesting... that would imply ~5W for the bigger screen? That seems high. Could you tell me your basal rate at 700 Mhz by chance? - everything else should match my system (using your fan control v.28 (-thanks!), minimum brightness, disk off, WiFi off, GPU power saving enabled with ATI driver.)
--gsmenden 18:16, 10 Feb 2006 (EST)
The 13-14W figure is with automatic cpufreq scaling and low load, so it's effectively at 800MHz (minimum speed for this 1.86GHz processor). Are you undervolting the CPU too?
--Thinker 00:36, 11 February 2006 (CET)
To my understanding it would seem very unlikely that the higher resolution/bigger display uses a lot more power than lower resolution/smaller ones - except the case that they would have four light tubes instead of one. I'm pretty sure that X series displays have ónly one. Wyrfel 01:20, 11 February 2006 (CET)
for T42 with 14" XGA, ATI 7500 at 80/100MHz, 600MHz@0.7V, hdd off, wifi off, linux:
1) min. brightness - 7,6W 2) brightness is 4 of 7 - 8,6W 3) max. brightness - 10,1W
2200bg in battery mode consumes ~0,4W. and I bet the iron can go lower runnig windows with native drivers. the only thing I really miss is powerplay support for 7500.
--Bzzt
You mean they nearly doubled the minimal power draw when moving to the T43's Sonoma chipset and Dothan CPU? So you actually get 5-6 hours from a 6-cell battery?! Spooky.
--Thinker 01:42, 11 February 2006 (CET)
well, I'm not sure about 6 hours, but 5h doesn't look hard to achieve. 7,6W is a real bottom here. I couldn't get lower. The regular rate I feel comfort at is rather 10-11W. But even that, 5h is possible with a new 6cell battery.
PS. this is why I returned 2668-4DU back and bought 2373-FWG. Sonoma really sucks, IMHO. I hope Intel has learnt the lesson and the next lines won't so hungry.
-bzzz
Ok - the 15.1" 1600x1200 must be a real power monster, although I have not applied the undervolting patch. I'll give it a whirl when 2.6.15.4 is released.
Update - tried the undervolting patch with some nice results - my lap is noticably cooler, and the fan spends more time off. With only a few process running (no kdm and kde -> only very basic window manager) and a black background text sceen, undervolted down to 700 mV @798 MHz, fan disabled, minimum brightness, hard disk off (standby), WiFi off (ipw not loaded), GPU power saving, DVD unplugged (2nd battery in bay), I can *still* only get down to 17.5 W/h. D'oh!
Still, the screen is worth it if only for the sheer number of women who note, "my, what a nice screen you have" (just one this afternoon. :) It is a strange phenomenon. --gsmenden 21:45, 10 February 2006 (EST)
Haha, good one, gsmenden :)
On my T43 here, I have around 13300mW, when:
- cstate=C4, 800Mhz, undervolted (0.716V, rather than 0.988V)
- wifi=off, bt=off, fan=off, brightness=minimum
- X300 put into "low" mode (it's a 15" SXGA+ screen)
- pretty "dark" setup (I have basically all-black desktop)
Full charge is around 77000mWh. With, say, 14000mW one should last 5.5 hours on a 9-cell battery.
I can't really say that I see any other obvious venues to pursue to reduce power consumption.
--igorr
igorr, how do you disable wifi? did you try to eject dvd drive?
-bzzz
Disabling wifi can be accomplished by echo 1 > /sys/class/net/eth1/device/rf_kill.
Ejecting the dvd drive? Why would that affect power consumption?
--igorr
yeah, rf_kill does right. if I eject dvd drive or simple type echo eject >/proc/acpi/ibm/bay, I save about 0.4W.
Gradual battery discharge
Has anyone else noticed, that the battery keeps discharging (although very slowly) when ac? My battery charges to 95%, and then drops to 90% over the course of few hours. Then charges again and so on. I'm not at all too happy about it. Also- what is the relation of last_full to design_capacity? I have one battery, which charges to 48410 mWh, while the design capacity is 51830 mWh, the other charges to 51960 mWh with the same max. Although it used to charge to around 54000. Both are new, the first one has around 50 cycles, the second around 10.
--Rasto 14:28, 17 February 2006 (CET)
Playing with /sys/devices/*/power/state
Anyone played with /sys/devices/*/power/state yet? eg.
echo -n 2 > /sys/devices/*/power/state
It looks like it can save a lot of power ~ 0.5 - 1.5 Watt on my T43, but it is also dangerous (and API will probably change in the future). Switching off the wrong device can crash the system, for example when turning the SATA off.
--Defiant 10:46, 25 July 2006 (CEST)
I've toyed with a few devices' state and didn't notice a difference. Can you identify which devices make a big difference for you?
--Thinker 12:15, 25 July 2006 (CEST)
I havn't seen a big difference for any device, each device does probably only make a difference of 50-100mW which I'm unable to measure correctly. Also just echo a 2 to /sys/devices/*/power/state doesn't seem to be enough. It look like it needs a 2 for all the state files for the device.
--Defiant 18:26, 27 July 2006 (CEST)
It would be very valuable if you managed to identify a few especially interesting devices. By taking groups of several devices, or looking at average power draw over a long time, you can discern even the small differences.
--Thinker 02:53, 28 July 2006 (CEST)
Ok here is what I've done:
For every device (except SATA and USB) that reports "Power Management" on lcpci -vv:
- enable power saving
- wait one minute and while waiting record present rate in /proc/acpi/battery/BAT0/state every 10 seconds.
- next device
Display- and harddisk standby had been disabled. And this is the result:
| Device | power avg | power diff |
|---|---|---|
| Normal | 13187 | +/-0 |
| VGA compatible controller: Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller (rev 03) | 13159 | -28 |
| Display controller: Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller (rev 03) | 12995 | -264 |
| Ethernet controller: Broadcom Corporation NetXtreme BCM5751M Gigabit Ethernet PCI Express (rev 11) | 12977 | -18 |
| PCI bridge: Intel Corporation 82801FB/FBM/FR/FW/FRW (ICH6 Family) PCI Express Port 1 (rev 03) | 12969 | -8 |
| PCI bridge: Intel Corporation 82801FB/FBM/FR/FW/FRW (ICH6 Family) PCI Express Port 3 (rev 03) | 12966 | -3 |
| Multimedia audio controller: Intel Corporation 82801FB/FBM/FR/FW/FRW (ICH6 Family) AC'97 Audio Controller (rev 03) | 12640 | -326 |
| Modem: Intel Corporation 82801FB/FBM/FR/FW/FRW (ICH6 Family) AC'97 Modem Controller (rev 03) | 12615 | -25 |
| pnp0 | 12611 | -4 |
| platform | 12554 | -57 |
So its definitely the graphic and the sound card.
--Defiant 10:17, 29 July 2006 (CEST)
Interesting! Can we automagically power off audio when it's not in use?
--Thinker 12:13, 29 July 2006 (CEST)
I'm not familar with the ALSA API, so I don't know any. Right now I binded that to Mute on/off. I did some research and the first thing I found is that it is possible to get the state, so polling would be possible (but I don't want polling). Will see if there is some kind of interrupt.
--Defiant 22:50, 30 July 2006 (CEST)
Does it preserve all configuration (mixer volume, switches etc.) when you toggle power?
--Thinker 00:13, 31 July 2006 (CEST)
