Testing the Battery Usage of the Most Popular ROMs for the Samsung Galaxy Nexus

March 8th, 2012 § 14 comments

Battery Lost due to Android OS

Android OS Battery Usage Test for 9 Popular ROMs for the Samsung Galaxy Nexus (CDMA/LTE) (Toro)

The above results are those of what I call an “idle test”, where the phone was put in Airplane mode, to try to remove the negative battery effects of radio towers changing, and other reception related battery drains. Additionally, the phone was left untouched for the duration of the test, so as not to let the display or brightness of the screen be a drain on the battery. And similarly, all applications were closed prior to testing, in hopes of allowing the battery lost to be due to as few factors as possible. This, in my eyes, gave a better result for the overall battery use by the OS. Please note that, as seen in the screenshots below, each test was not exactly 7 hours, because over time, I noticed that the battery change was negligible after about 6 hours. Additionally, after running many 1 hour idle tests for many different ROMs, I also noticed the Fraction Lost due to OS did not change very much between 1 hour tests and 6-7 hour tests. So take the data as you wish, but note that I tried to be as careful as possible, and believe that the data is representative of the ROMs used.

Now with all that said, I do believe there are outliers here. Notably, AXI0M Rebirth v2.5 is a ROM that was just released a few days ago, and is still very much in the testing phase. The only reason I include it is to allow the developer the ability to see that the battery still needs some tweaking with respect to OS idling. I also chose not to test with unstable releases, or those not popular enough, or those without very much support. I do realize there are others I should probably test too, so feel free to comment which ROMs you believe to be worth comparing to these 9.

One thing I noticed with the battery lives is that typically the best battery statistics were seen with ROMs that were the “lightest”, or the ones that had the smallest .zip files. For example, Liquid Smooth had the smallest OS zip package of them all, and additionally had the best battery lost statistics, and the best fraction lost statistics. I noticed that this wasn’t a coincidence. I don’t know too much about how the OS runs and works while idling, so I encourage anyone who knows about this to a high degree to please comment about it and share the knowledge.

In addition to battery life, other characteristics of the ROMs that will influence my decision on which to go back to are the available launchers, the custom lockscreens, and obviously the overall speed and stability. After playing with almost all launchers available, my favorite is Apex Launcher, followed by Nova Launcher. I can definitely say I won’t be going back to any ROM that doesn’t have one of these launchers. As for lockscreens, most ROMs had this ability, and almost all were stable enough to not be eliminated from the comparisons. But overall, I would say the 3 ROMs that I will only be going back to would be Liquid Smooth, AOKP, and GummyNex. Liquid Smooth v1.0 is actually the final build of this ROM, and as you can see, it’s great on battery, super stable, and has every option you would ever want. That would be my first choice, but if you need constant builds, with more options, then I would probably choose AOKP or GummyNex. DroidTh3ory’s hybrid AXI0M Crossbreed has just as many options too, and should be as great on battery life, being based on Team Kang’s ROMs. Something worth noting is that every ROM I tested is worth   going back to, but after you try them all, you will start to get picky, and you will pick favorites. So with that said, I recommend you all to try most of these ROMs, and then pick your favorite, but I would start with those above that showed great battery usage stats.

Please leave a comment if you have anything to add. Thanks!

 

Edit: Liquid Smooth v1.0 is NOT the final release.
Edit: for more comments and ideas on reproducing this test, check out this thread.

Tagged , , , , , ,

  • Pingback: The 3rd Custom ROM for My Verizon Galaxy Nexus, GummyNex 0.9.0 [VIDEO]

  • AlicePalace

    My first question would be why you used AOKP Milestone 3 instead of Milestone 4?

    So I would like to correct a few things about your article. First off, as I understand it, any launcher can be used with any ROM. For example, I use Apex launcher with AOKP. All you have to do is download the launcher from the Android Market (I guess it’s called Google Play now). If you want the benefits of having the launcher installed at the system level, you can simply push it to system yourself with either adb or by using a file explorer app such as Root Explorer and fixing permissions.

    Second of all, you replied to another user asking which kernel you used for each test, and you replied that you used the kernel that came with each ROM. I believe this makes your study fundamentally flawed, as kernels have a much greater impact on battery life than do ROMs.

    For example, let’s take Liquid vs. AOKP. Liquid comes with imoseyon’s Leankernel. I do not believe (someone can correct me if I’m wrong) that the ROM is “built around” this kernel; this is just the kernel that the developers chose to include because they like it, the same way that AOKP comes with Nova Launcher but can easily be used with Apex launcher instead. Now let’s take AOKP. AOKP comes with the stock kernel because the developer have expressed that they would like the allow the users to choose whatever kernel they like. Many people run a variety of kernels with it successfully— Trinity, Franco, Lean, you name it. All of these give people varying results in battery life.

    Here’s the problem—- you can’t compare accurately across ROMs with them all running different kernels if the kernel isn’t designed for the ROM. Some ROMs come with a kernel they built specifically for their ROMs (MIUI I believe is one?). Other ROMs come with the kernels the developers prefer, but did not build (Liquid). Others come with stock kernel to let the user choose, and aren’t built for any particular kernel at all (AOKP).

    For example, I’ve personally run both AOKP and Liquid. If I were to run AOKP with stock kernel, which it comes with, I would get absolutely horrid battery life compared to Liquid. If I run then both with Leankernel however (my preferred kernel right now), I actually see better battery life on AOKP. (Everyone’s milage may vary).

    I think this experiment, to be done accurately, should be done with the same kernel on every ROM to eliminate battery life variances caused by the kernel. I would recommend using the stock kernel on every ROM that is compatible with it. This would only be unfair to the few that actually build their own kernel specifically for use with their ROM, which I think would be very few on this list.

    • AlicePalace

      This would only be unfair to the few that actually build their own kernel specifically for use with their ROM, which I think would be very few on this list.—- I forgot to add, “and these few could actually just be excluded from the experiment.”

    • http://www.shayessays.com/ Shayes Says

      Thanks for the response. The ROM versions I chose all depended on when I was doing the testing. So M3 was the last stable release when I tested AOKP.

      I know that the launcher is just an app, and can be installed regardless of ROM. The point I was just trying to make is that I just appreciate it when a dev includes the best possible launcher already in the ROM. It’s one less thing to have to download from the market.

      Yes, I now know that including kernel as a testing variable would have made for more accurate results, but when I started this test, my only intention was to test “out of box” ROMs for those of us who don’t want to change kernels, and just want the ROM as compiled by the developer. You are correct in everything you say though.

      I would encourage you to test AOKP for every kernel you see fit (because AOKP seems to be the best combination of stable, light, and fast, and as you said, isn’t built with a certain kernel), then go back and test the most popular ROMs for the kernel with the best results. I would think that would be a good test.

  • timothymowens

    Any chance you could add one for stock 4.0.4?  That’s what I’m running now and it seems to do better than stock 4.0.2, but I’m wondering how it compares to custom ROMs…

  • Element4

    Appreciate the work put into compiling this information. A phone operating at idle is one thing, do you have any ideas if various ROMs may be more efficient while operating?

    • http://www.shayessays.com/ Shayes Says

      I didn’t really test for usage while operating the phone because you can’t get good results doing that, but I would assume that anything AOSP would get similar results with each other because of how light weight they are.

  • Pchoi94

    thanks for taking the time to do this. However it would have been nice to also include a stock ROM

  • Dpaine3

    As s beta tester for Liquid ROM, the 1.0 is not the final Liquid for ICS, it was just in a 5-stage beta before 1.0 is released. Liquid 1.1 has several great new features and is in the works. No ETA but team liquid is working constantly on the ROM.

  • Salk

    What kernel did you use with each ROM?

    • http://www.shayessays.com/ Shayes Says

      I made no kernel changes between ROMs, so every ROM in the test used the kernel that the dev built his ROM on.

  • Pingback: What is the best rom? - Page 8

  • John Previte

    thanks for posting this. time to try a few of these out, as i have been getting terrible battery life on stock.

  • Pingback: Did a test on the ROMs that are best on the battery usage - here are my results - Android Forums

What's this?

You are currently reading Testing the Battery Usage of the Most Popular ROMs for the Samsung Galaxy Nexus at Shayes Says.

meta