You are not logged in.
Hi guys, I'm trying to figure out why Arch uses around twice the power of OS X when idle (and light usage). With the laptop idle, not going to sleep or turning the screen off, I measured the battery usage over 80 mins,
Arch 20% (50%->30%)
OS X 10% (50%->40%)
I have checked other smaller intervals such as 100% to 95% or 80% to 78% at various times and it always works out to be around 4 mins to drain 1% on Linux and 8 mins to drain 1% on OS X.
It is also pretty obvious during moderate usage, if I use Arch on the train on my way to and from work it drains around double the battery compared to OS X. I have checked that the battery percentage is being reported correctly and it is almost the same between the two operating systems. 30% in Arch via the acpi command corresponds to 31% under OS X and 48% to 50%.
I have tried using i7z, i7z-git and powertop to measure the power states and clocks of the cpu. i7z reports that it is sitting at 800 MHz idle and only shows the C1 state.
Socket [0] - [physical cores=2, logical cores=4, max online cores ever=2]
TURBO ENABLED on 2 Cores, Hyper Threading ON
Max Frequency without considering Turbo 2498.96 MHz (99.96 x [25])
Max TURBO Multiplier (if Enabled) with 1/2/3/4 Cores is 29x/29x/29x/29x
Real Current Frequency 799.86 MHz [99.96 x 8.00] (Max of below)
Core [core-id] :Actual Freq (Mult.) C0% Halt(C1)% C3 % C6 % Temp VCore
Core 1 [0]: 799.86 (8.00x) 1 98.7 1 0 39 0.7793
Core 2 [1]: 799.27 (8.00x) 3.83 98.7 0 0 33 0.7856
powertop tells me that the processor is reaching the C10 state,
Package | Core | CPU 0 CPU 2
| | C0 active 0.1% 0.2%
| | POLL 0.0% 0.0 ms 0.0% 0.0 ms
| | C1E-HSW 0.0% 0.0 ms 0.0% 0.0 ms
C2 (pc2) 53.2% | |
C3 (pc3) 3.1% | C3 (cc3) 1.3% | C3-HSW 0.0% 0.0 ms 1.7% 2.0 ms
C6 (pc6) 39.7% | C6 (cc6) 0.1% | C6-HSW 0.1% 0.1 ms 0.1% 0.6 ms
C7 (pc7) 0.0% | C7 (cc7) 96.7% | C7s-HSW 0.0% 0.0 ms 1.5% 2.5 ms
C8 (pc8) 0.0% | | C8-HSW 13.9% 69.9 ms 0.6% 1.5 ms
C9 (pc9) 0.0% | | C9-HSW 2.4% 4.9 ms 9.5% 8.7 ms
C10 (pc10) 0.0% | | C10-HSW 82.8% 167.2 ms 85.9% 14.0 ms
| Core | CPU 1 CPU 3
| | C0 active 0.1% 0.1%
| | POLL 0.0% 0.0 ms 0.0% 0.0 ms
| | C1E-HSW 0.0% 0.0 ms 0.0% 0.0 ms
| |
| C3 (cc3) 2.0% | C3-HSW 2.2% 2.5 ms 0.0% 0.0 ms
| C6 (cc6) 0.0% | C6-HSW 0.0% 0.0 ms 0.0% 0.0 ms
| C7 (cc7) 97.4% | C7s-HSW 0.0% 0.0 ms 0.0% 0.0 ms
| | C8-HSW 0.0% 0.0 ms 0.6% 3.0 ms
| | C9-HSW 28.7% 96.5 ms 2.0% 10.3 ms
| | C10-HSW 68.9% 58.0 ms 97.0% 140.0 ms
| GPU |
| |
| Powered On 1.8% |
| RC6 98.2% |
| RC6p 0.0% |
| RC6pp 0.0% |
Does anyone have any ideas why it is using so much more juice under Linux?
Offline
ok a bit of an update I've gone thru all of the steps in the wiki on power saving,
https://wiki.archlinux.org/index.php/Power_saving
it has made no real difference, now it takes just under 5 mins to drain 1%, still much worse than the 8 mins in OS X.
Offline
I ask of out ignorance: are the battery scales themselves consistent across operating system? In other words, when you query the battery charge, is the software simply reading a readout from the hardware or is there some sort of calibration needed?
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
yeah there must be some sort of calibration involved because they don't match exactly but like I said I tested by rebooting between the two and they match within a one or two percent. 100% matches exactly, 48% Arch was 50% in OS X and likewise 30% was 31%.
Offline