ART vs. DALVIK (4.4.2 KitKat)

Which runtime do you using?


  • Total voters
    85
  • Poll closed .
Search This thread

2hipso

Senior Member
Jan 4, 2014
294
90
Slovakia, Žilina
Hi guys, because I can't sleep, I dedicated to perform battery tests of Dalvik runtime and ART runtime to see which one is eating battery more.
I tried to do both tests the same to get the most relevant result:



TEST 1: Playing video while on airplane mode.

Results after TEST 1:
Battery on DALVIK (tested first) went from 100% to 97%, on ART (tested second) went from 100% to 97%.


TEST 1 standings:
DALVIK (100 + 97) = 197 points
ART (100 + 97) = 197 points

--------------------------------------------------------------------------------------------------------------------


TEST 2: (Now on full-brightness) Started the game for 10 minutes and then playing video while on airplane mode.

Results after TEST 2:
Battery on ART (tested first) went from 100% to 92%, on DALVIK (tested second) battery also went from 100% to 92%.

TEST 2 standings:
DALVIK (100 + 92) = 192 points
ART (100 + 92) = 192 points


TOTAL:
DALVIK = 389 points
ART = 389 points


More points = less battery usage, so it's tied !

also it seems like, with Dalvik, the screen uses more battery. Check screenshots

NOTE: Each test ran for exactly the same conditions (same battery status, same light conditions etc.)
NOTE 2: Results on your device and with your battery may vary.
 
Last edited:

MrKra

Senior Member
Dec 9, 2012
157
23
The Improvement from Dalvik to Art mostly affects the starting process.
Instead of a compilation on every start, its only done once.

While Running I do not expect any battery improvements ;)
 

debaj

Senior Member
Dec 8, 2010
780
168
Budapest
The battery consumption of the screen comes mainly from the backlight. Changing the runtime will not affect this in any way.
 

scott_doyland

Senior Member
Dec 8, 2011
493
145
The Improvement from Dalvik to Art mostly affects the starting process.
Instead of a compilation on every start, its only done once.

While Running I do not expect any battery improvements ;)

Exactly right.

ART v Dalvik makes zero difference when playing a video or once an app has started.

No offence to OP but its worth gaining an understanding of ART before running tests as can be seen from your results ;)

Also I would assume any app running or cached is therefore already in memory so its only apps NOT running/cached that benefit from ART when they first start.

I have heard people say ART means less lag when scrolling - how on earth would it do that!!
 
Last edited:

MrKra

Senior Member
Dec 9, 2012
157
23
Art is interesting for Low mem devices, which do not hold Apps in the RAM.
 

scott_doyland

Senior Member
Dec 8, 2011
493
145
Art is interesting for Low mem devices, which do not hold Apps in the RAM.

Yes, true. I guess a lot of people have just a handful of apps they use a lot though, and its more than likely, especially with RAM increasing on devices, that these few often used apps will always be in RAM already. Depends how the app is coded as well I guess.

I can see it may help for large apps, ie games, where a portion of the game is in RAM and as you progress in the game the next portion has to be loaded. It will make loading times quicker.

With the current speed of devices and the large RAM now available (and look to one or two years ahead and 1GB will be considered low) I wonder how much difference ART will really make especially to just the average user - I bet they wouldnt even notice. We may be able to see statistics telling us an app loads a few milliseconds quicker but in the real world how much difference will this make.

However Im in no way saying ART is not worth having, if it does save just a tiny bit of battery its maybe worth it as we all want our phone to last longer. Its just not as big a deal as some people seem to make out - in my opinion.
 

2hipso

Senior Member
Jan 4, 2014
294
90
Slovakia, Žilina
That test was because lot of people saying that battery has improved with ART.

I noticed that ART was little faster while loading the game (~0,5 sec) also when video comes in the end, Dalvik has little delay to show me the software buttons.

Sent from my XT1032 using xda app-developers app
 

eemgee

Senior Member
Dec 27, 2013
998
249
Teesside
I just tried it, couldn't feel much difference in speed, did see I chuck off my valuable 8gb of memory vanish. So I changed back!
 

hoschi-air

Member
Jan 21, 2011
35
2
Berlin
switched nearly immediately to ART after updating to KitKat, but what I really noticed to speed up was the recent apps drawer. Scrolling through windows was definitely faster. Only real downside for me is the ~10% greater storage demand... For my understanding using precompiled binaries kinda makes sense, but there are lots of really proficient threads here on the matter and most of them conclude that ART is the future.
 

hoschi-air

Member
Jan 21, 2011
35
2
Berlin
Because on this phone there is no benefit in using ART over the Qualcomm optimized Dalvik runtime we have. And storage space is far more important to those with an 8GB device.

Hmm, besides missing any evidence on the real performance of these libraries on the Moto G, I doubt they will be more than a short-term solution. >> Google pushing ART development

Feel free to correct me if I'm wrong...
 

mike21pr

Senior Member
May 16, 2008
1,333
81
San Juan
Im switching to ART as soon as I get mine. Did wonders on my old Nexus 4

Sent from my SAMSUNG-SM-N900A using Tapatalk
 

eddvys

Senior Member
Nov 27, 2013
322
91
Grantham
ART isn't an option for me yet, still like using xposed framework to much. Hopefully they will update that to ART sometime.

For xposed to work with ART some serious coding must be done. And personally me I don't think its possible at all, but its just me so lets sit back and wait for devs to show up :)) :angel:
 

BarnOwl

Senior Member
Jan 21, 2003
467
117
The Netherlands
www.parketvloeren.nl
One thing against ART: it is in early developer preview state, so it is not recommended to use it on a daily use phone. It is not in the development menu by mistake. I know, I know, you did not encounter any issues yet. :)

Well, for sure I have been using ART for a while now. Since I changed my Samsung Galaxy SIII to unofficial Kitkat, which is about a month ago.....
Now, with my Moto G the first thing I did was changing it to ART because it worked like a charm. :)
 

Top Liked Posts

  • There are no posts matching your filters.
  • 5
    Hi guys, because I can't sleep, I dedicated to perform battery tests of Dalvik runtime and ART runtime to see which one is eating battery more.
    I tried to do both tests the same to get the most relevant result:



    TEST 1: Playing video while on airplane mode.

    Results after TEST 1:
    Battery on DALVIK (tested first) went from 100% to 97%, on ART (tested second) went from 100% to 97%.


    TEST 1 standings:
    DALVIK (100 + 97) = 197 points
    ART (100 + 97) = 197 points

    --------------------------------------------------------------------------------------------------------------------


    TEST 2: (Now on full-brightness) Started the game for 10 minutes and then playing video while on airplane mode.

    Results after TEST 2:
    Battery on ART (tested first) went from 100% to 92%, on DALVIK (tested second) battery also went from 100% to 92%.

    TEST 2 standings:
    DALVIK (100 + 92) = 192 points
    ART (100 + 92) = 192 points


    TOTAL:
    DALVIK = 389 points
    ART = 389 points


    More points = less battery usage, so it's tied !

    also it seems like, with Dalvik, the screen uses more battery. Check screenshots

    NOTE: Each test ran for exactly the same conditions (same battery status, same light conditions etc.)
    NOTE 2: Results on your device and with your battery may vary.
    4
    The Improvement from Dalvik to Art mostly affects the starting process.
    Instead of a compilation on every start, its only done once.

    While Running I do not expect any battery improvements ;)

    Exactly right.

    ART v Dalvik makes zero difference when playing a video or once an app has started.

    No offence to OP but its worth gaining an understanding of ART before running tests as can be seen from your results ;)

    Also I would assume any app running or cached is therefore already in memory so its only apps NOT running/cached that benefit from ART when they first start.

    I have heard people say ART means less lag when scrolling - how on earth would it do that!!