Request: Browser energy consumption

asked 2015-01-26 13:39:54 +0200

florifreeman
1026 ●22 ●31 ●37

The energy consumption of the built-in Jolla browser is enormous.

I saw a lot of mobile Browsers out there beeing energy-thirsty (like IE on WP, or FF on And.) but the Jolla browser tops everything in line of using battery.

5 Minutes on Golem.de or the german PCGH.de costs 10-15% of battery life => This is WAY too much.

IMO: Surfing the net is THAT Feature that is annoying me most, because the rest of the Phone is soooo energy saving (even some Android apps consumpt less energy, like the HERE Maps, which is very hungry on WP´s).

Hope there will be some improvement, and not just to the UI.

Regards, Freeman

edit retag flag offensive close delete

Comments

What about Webcat? Did you try it? The best mobile browser for me is Opera mini or Opera Mobile, imho :)

zaharov.andrei ( 2015-01-26 13:50:42 +0200 )edit

I can understand there is power consumption from the display unit, but we need to take into account the content of the page, e.g. a mobile unfriendly written web page that floods the browser with lots of data and processing, is expected to consume power. Do you have the same consumption visiting for example www.google.com?

pmelas ( 2015-01-26 16:01:55 +0200 )edit

1

Battery % is not comparable between Jolla and WP. Jollas battery comp. is more linear, so when jollas battary is 40% it's much stronger than WP 40%. WP's battery had made that way to look better for your eyes :D (it stays on 100% longer but drops way faster in some point).

You should compare battery voltage, then you will know better. If you can get that info out from the WP.

Makeclick ( 2015-01-26 17:29:41 +0200 )edit

1

I just tried browsing golem.de for five minutes, skimming through a few articles, but I only lost 1% battery in that time.

null ( 2015-01-26 18:46:51 +0200 )edit

@Makeclick: I pointed out that the WP browser is also VERY energy consuming.

florifreeman ( 2015-01-27 14:23:25 +0200 )edit

6

answered 2015-02-19 16:02:19 +0200

dez

7104 ●78 ●94 ●94

moderator

updated 2015-02-19 16:30:33 +0200

So, I performed measurements with fake battery and laboratory power supply.

4 browsers: Jolla browser, Webcat, Firefox 35.0 for Android and Lightning for Android v.3.1.1a.

Screen brightness is set to adaptive (constant light conditions - artificial flourescent light) and maximum level.

Power supply: V=3.8V, measuring (average) current (you can easily calculate power and compare with other devices).

Used the golem.de site you talked about with the same usage pattern: load web site, switch to the Ticker page, scroll it, switch back to Home page. All browsers show almost the same power consumption: ~220mA on loaded page (mostly power is consumed by display backlight), around 500mA during page loading and 450-500mA when scrolling page. Firefox for Android consumes ~10-15% more power then other browsers, so taking into account Jolla browseris also built on the Gecko engine (need to check is engine version the same). So, I see that Jolla browser is optimized quite well :-)

When brightness is set to minimum, current during static page display is ~130mA and consumption during page loading/scrolling raises to ~400mA, again Firefox for Android shows higher power consumption numbers.

P.S. Internet was accessed through WLAN.

edit flag offensive delete publish link

Comments

So only dimming the display already showed significant change in current. qed.

Thanks a lot for clearing it up. Btw I am interested in your experimental setup. What did you use as fake battery? Wouldn't a resistor be enough?

lakutalo ( 2015-02-19 19:26:30 +0200 )edit

1

It is enough to have laboratory power supply connected directly to the battery connector (+/-) (with proper voltage) + capacitor. Disclaimer: I hope you understand that if you want to do it, do it on your own risk, device warranty is void as also it can be potentially dangerous.

dez ( 2015-02-20 11:02:38 +0200 )edit

@dez: Why not using a high voltage/220V laboratory power supply and the charger? Current measurements should be (at least relatively) the same. This way you would lower risk of damage. What do you think? Disclosure: I am fully aware of the consequences, both technical and legal. :)

lakutalo ( 2015-02-20 12:07:31 +0200 )edit

What do you mean by high voltage?

dez ( 2015-02-27 20:51:32 +0200 )edit

Not really high voltage, I meant a 220V outlet combined with an amperemeter.

lakutalo ( 2015-02-27 21:03:57 +0200 )edit

3

answered 2015-01-27 13:10:59 +0200

simosagi

1459 ●26 ●37 ●46

Instead of just stating 'energy consumption is enourmous', why don't you actually measure it?
Get one of the battery measurement apps available (e.g. Hunger meter, Battery Charger), open it side-by-side with the browser, and check the battery consumption. After that compare the browser consumption with some other application that it's using display and network.

edit flag offensive delete publish link

Comments

1

Or even more confined: Try to find a site that constantly updates itself and switch off your display. After that install another browser and repeat the procedure. Repeat the whole procedure with display on. To cancel non-linear effects of discharge repeat after full recharge to be most accurate. I would really like to know the results.

lakutalo ( 2015-01-27 13:33:20 +0200 )edit

Who said, that i didn´t?.

florifreeman ( 2015-01-27 14:22:31 +0200 )edit

Good experiments take time. :) I will have a look at it in the future. Maybe we can build a benchmark.

lakutalo ( 2015-01-27 14:26:24 +0200 )edit

More simple: Take different phones with almost the same apps and look how Long the last on a typical working day. This is what i did, Lumia, iPhone, Jolla. I know These phones very well and their behavior on energy consumption. So, the Jolla is very energy saving and lasts longer (using them my way) then the other two except when it Comes to browsing Websites. I think this must be a hint, that somethings going wrong there.

florifreeman ( 2015-01-27 14:33:00 +0200 )edit

Still you have to cancel the influence of the display to get objective results.

lakutalo ( 2015-01-27 14:43:24 +0200 )edit

1

answered 2015-01-26 13:54:49 +0200

lakutalo

20555 ●168 ●261 ●264

updated 2015-02-19 11:56:52 +0200

Possible that the code of the native browser could need more optimisation. And the drain due to the browser itself may also be influenced by the site you load. But the browser's cpu consumption is negligible compared to the display's.

I have managed to have my battery lasting for almost 6 days (excl. nights) if only using my Jolla as a phone. I did this to find out how strong the impact of using the display is on consumption.

In short, the display seems to be the weakest link in the chain of energy consumers on our phones.

edit flag offensive delete publish link

Comments

But playing Heebo does not consume the battery so much as browser.

Yes, the power consumption is very low when only using Jolla as a phone. I think thats because Sailfish is in deep sleep mode most of time.

zaharov.andrei ( 2015-01-26 14:26:28 +0200 )edit

Congrats mate (6 days is impossible imho) but iam using my phone as a Smartphone for private and Business: That means getting a lot of email, telephone and Messages. Like every application is there a Point that is called Efficiency, and that is not given on Jolla´s native browser. I love this browser because it has the unique UI.

florifreeman ( 2015-01-26 15:43:05 +0200 )edit

I just made it happen. 9d up and still 8% left. Average 2calls and 4sms a day.

lakutalo ( 2015-02-23 23:00:43 +0200 )edit

1

answered 2015-02-19 12:53:32 +0200

zaharov.andrei

692 ●13 ●25 ●27

http://www.gisi.ru/

updated 2015-02-19 12:54:07 +0200

Some measures:

image description

Green - using just as phone Red - web-browsing with wifi

edit flag offensive delete publish link

Comments

Could you add a measurement for using another app that needs display constantly on, since using phone app still means almost no display. Then we could cancel energy consumption by display from (browser + display).

lakutalo ( 2015-02-19 13:10:56 +0200 )edit

Ок. I repeat measure after next full charge. I think it's not so much display as wifi.

zaharov.andrei ( 2015-02-19 14:43:33 +0200 )edit

Request: Browser energy consumption

Comments

4 Answers

Comments

Comments

Comments

Comments

Question tools

Stats

Related questions

Request: Browser energy consumption

Comments

4 Answers

Comments

Comments

Comments

Comments

Question tools

Public thread

Stats

Related questions