Battery usage: artificial and real world test results

Products V (2017) | General Discussion
1

Hi there,

as I don’t have the time right now to write a full review and as there are already a few out there, I will focus on different aspects and try to give you as many information as I can. If I find some time, I will reformat the list below as table so it’s easier to get the changes.

If you have anything I should try out, just ask.

First thing that I got asked: battery
I tried out a few things and I will expand this list as I continue to try out more.
Everything is measured by perfmon.exe (standard Windows tool) with Discharge Rate as indicator, perfmon being in foreground (if possible) and almost nothing as background tasks; Windows battery slider to “better performance”

Note about perfmon refresh cycles

perfmon takes about 3-5 secs for the Discharge Rate to settle after a value is changed. When trying out yourself and adding some other indicators, keep in mind, that their internal refresh is quicker so that you have to move the Discharge Rate graph 3 to 5 secs to the left in mind.

Specs
Eve V, i7, 16GB RAM, 512GB SSD, Windows Version 1709

Base settings:
Wifi: on – no connection
Bluetooth: on – no connection
Brightness: 30%
Sound: no output
Keyboard: connected through pogo pins, fully charged
Discharge Rate: 3.625 mW

image

raw data as text

Changes (changed back to base settings after each paragraph):
Brightness: 100% → 5.610 mW
Brightness: minimum → 2.979 mW

WiFi: Off → 3.458 mW
WiFi: On, connected → between 3.920 mW and 4.278 mW, 4.063 mW in average

Bluetooth: off → 3.366 mW
Keyboard: disconnected (with Bluetooth off) → 3.176 mW

Windows Power Saving Mode: On → 3.161 mW

Typing in Word (regular text, just typing with spell checking enabled): 4.552 mW to 5.244 mW
Typing in Editor: 3.650 mW

Multitasking, starting apps, starting word, closing word: 8.345 mW in average, spikes to 19.195 mW

Coming home, activating WiFi, connected, causing Windows to fetch all kind of things: 10.860 mW average over 74s, spikes up to 24.820 mW

Running Guild Wars 2 with wireless mouse connected, WiFi on, connected: 14.911 mW average with spikes up to 22.027 mW (GW2 in Lions Arch with medium to low settings, at half native resolution, 24 to 43 FPS)

Multitasking like mad (real), running Guild Wars 2 in background and foreground, high access to SSD, maxing out processor, max brightness, speakers at full volume: 28.842 mW at max

Browsing Internet with Edge, writing this post (WiFi on, connected): 4.922 mW in average

45 Likes
2

Thanks for the info. It was very detail. So far things looks promising.

How does GW2 run throughout gameplay. I have the game

5 Likes
3

Funny thing it, the resulting calculated 12 hours runtime is exactly what was advertised. Only that it results when doing nothing. :yum:

4 Likes
4

This is definitely an interesting chart. Good job!

1 Like
5

Great research! Really surprised that Windows Power Saving mode is actually working :joy: Also surprised by the difference between Notepad and Word, as I thought they would both consume a negligible amount of power (Notepad is called Editor in some Windows languages). I think the keyboard also uses too much power, unless the backlight is on.

As a measure for saving some battery: Try changing the Power Limit of the WiFi card down in the Device Manager (I think the V comes with maximum power by default). Try the lowest possible setting where you can still get the wifi signal reliably.

6 Likes
6

Checked Device Manager and you’re absolutely right:
image

Changed it to Medium - Discharge Rate went down to 3.997 mW. Changed it back and Discharge Rate went up to 4.035 mW (what seems to be todays base value) again.

Interesting thou was: From the moment I opened the Device Manager till 20 secs after I closed it, Discharge Rate went up to 8.500ish mW. Processor also was used by about 10-20% during this time. It seems Device Manager is active polling something.

10 Likes
7

Hey guys! What are you doing here is invaluable! Thanks a lot. That’s where we can optimize a lot as most components and apps are set to consume 100% of power by default when they don’t need that much.

After all of the birds get Vs we will start working on battery optimization ASAP. Keep us great progrress here!

27 Likes
8

What is interesting is that according to the test, with the Brightness set to Minimum, you can reach a theoretical 16 hours of runtime on battery. I that’s …incredible if it can occur in real world situations.

10

And people thought I was a crackhead for thinking it’s possible to reach 16 hours with the V :stuck_out_tongue:

3 Likes
11

@s.auler Hey , great topic, great idea and measurements. Thanks a lot. It would b great if other owners could add some numbers with the same setup. And optimally with different applications running. So we would get something like a comparison of different devices and could see if they all behave the same.

1 Like
12

Just things to add to the list:
Difference between “default theme” and “black / white experimental theme” is about 0.100 mW.
Battery runtime after mixed use (including playing Guild Wars 2 for about 20 mins) is 9:10 with 13% left. [Edit: Final runtime was 10:47]

Something I noticed: When starting apps (Windows Store for example), the processor spikes very hard and falls off quickly afterwards. The question is, if artificially throttling the speed in which the clock rate is increased would be beneficial or not. I remember the power consumption not being linear with the clock rate so it should make an effect when changing from quick and short to slower and longer. Is there a way to change these settings and measure the effect?

@dartarian1: I still owe you an answer on how Guild Wars 2 plays out: It works and is great if you’re out of the house and want to do some running around in Open World PvE. Graphics is OK but not great. I turned down the resolution to 1440 x 900 get useful framerates and you can play almost without any lags. I would not use it at home typically (without an egpu) but it could be great to play outside in the garden during summer… :slight_smile:

7 Likes
13

on SP4 I do not see all these settings, how come? :pensive: image

14

Different WiFi card. Marvell is, lets just say below average. Its one of the biggest complaint in the Surface community like here https://www.reddit.com/r/Surface/comments/68oqgw/panos_panay_its_a_beautiful_day_in_new_york_are/dh08uec/ or here https://www.reddit.com/r/Surface/comments/6cu51q/introducing_the_new_surface_pro/dhxdx09/

4 Likes
15

I still do :stuck_out_tongue_winking_eye:

1 Like
16

In favour of which? (*Starts contemplating removing even more stuff on the experimental theme*)

4 Likes
17

This is brilliant work!

Could I be cheeky and ask for data on the effect of using the speakers? For example: off, 50% and 100%?

2 Likes
18

Thanks a lot. Would be happy to play it in the go:grin:

19

thanks for this great topic! your mixed used scenario is most useful to me, I think I can expect battery life on par with my current Atom x5 tablet

(also, I can’t help but read your name as some kind of bacteria, is that your real name?)

1 Like
20

This normal behaviour of every CPU. I can’t believe that this short spikes consum a lot of energy.
If you have short spikes consuming energy or you have to wait longer for the application to be loaded, I believe that the energy consumption is the same but you have to wait longer for the application and that you can start to work with and then V (or other device) would feel slow. I prefer the spikes rather than waiting longer.

21

@nawthor: “back and white” is more energy efficient than the standard theme. I think this has to do with the black background compared to the white background - something we have already seen at many other occasions.

@Scuff:
Todays base value while playing an local MP3 in Microsoft Groove is 4.480 mW (with WiFi connected).
Speakers at 50%: 5.087 mW
Speakers at 100%: 5.345 mW

@hiddenflaw:

  1. I concretized my post above: run time was already 9:10 and at that moment I had 13% left. Final runtime was 10:47
  2. Yep, that’s my name with s being the abbreviation for Stefan and Auler being my surname. I’m neither related nor related by marriage up to the fourth degree with any bacteria, at least as far as I know. :smiley:
16 Likes