[18 Apr 2014] [KERNEL TESTS] [SAMMY ROMS] Kernel battery tests and benchmark results
- Thread starter It_ler
- Start date
- Status
Hiya It_ler !
Glad you finally opened a thread mate, will be easier to follow your tests
Thumbs up for this guy who spend most of his time benchmarking battery life with kernels and playing with his own battery lifetime to shares us always everyday better useful stats !
'Cheers!
Glad you finally opened a thread mate, will be easier to follow your tests
Thumbs up for this guy who spend most of his time benchmarking battery life with kernels and playing with his own battery lifetime to shares us always everyday better useful stats !
'Cheers!
CamoGeko
Retired Forum Moderator
Hi mate, just added a link to your thread in my MagicConfig thread in post 2 here - xdaforums.com/showthread.php?t=1466017
Sent from my 80GB CyanogenMOD 9 + Siyah ICS powered beast. Booya!!
Sent from my 80GB CyanogenMOD 9 + Siyah ICS powered beast. Booya!!
Great info and detailed test results!! THX!!
Out of curiosity, forgetting all the numbers, which kernel is your favorite?? Which one felt the smoothest to you?
(the reason I ask is because I just upgraded from SGS1 to SGS2 today and have to start all over again choosing roms, mods, kernels, etc, and would also like to hear your overall impressions of these different kernels)
Out of curiosity, forgetting all the numbers, which kernel is your favorite?? Which one felt the smoothest to you?
(the reason I ask is because I just upgraded from SGS1 to SGS2 today and have to start all over again choosing roms, mods, kernels, etc, and would also like to hear your overall impressions of these different kernels)
I have got no real or absolute favourite kernel.
As IMO battery life is more important, I prefer Phenomenal Extreme, which is extremely undervolted by default but stable on my phone.
For smoothness ... try latest SpeedMod kernel.
N.E.A.K. and SiyahKernel are also very good and smooth, but you can tweak a lot of system/kernel settings to optimize the kernel to your needs ... so that is something I would not recommend for "beginners".
It is your phone, bootloader is unlocked, so feel fre to flash and test any kernel which is compatible with your ROM.
You are welcome post your results/experiences here.
Thx for that. I will start with SpeedMod, and if the battery life isn't good enough, I'll try Phenomenal Extreme.
SocialHub is in my recommanded apk removing list (in my battery saving tips collection), it's sucking battery juice like hell
same about Seven engine Btw, I tried latest XXLPW CF-Root kernel but didn't found it very battery frendly in deep sleep, won't recommand it to ure future kernel-to-test list at all
SocialHub is in my recommanded apk removing list (in my battery saving tips collection), it's sucking battery juice like hell
Btw, I tried latest XXLPW CF-Root kernel but didn't found it very battery frendly in deep sleep, won't recommand it to ure future kernel-to-test list at all
Yeah, saw your tips already and have applied many of them. (See the link in the first post!)
With removing of some system apks I am very careful, as I am using GTalk, stock EMail and Google Reader app.
So I still need some of these social hub / seven apks.
But I am always confident to test new settings ;-)
Have updated my "kernels-to-test" list in post #2.
Always happy to hear about other user's experiences.
Last edited:
XXLPW is not on my current kernels-to-test list, but maybe XWLPG could be a candidate.
Want to finish my current tests first, then I decide which kernel/ROM to test next.
Thank you for the answer
LPG will be good to compare with LPF, in fact, I tested LPW and will not be back to it. Only battery under LPF a bit worst...
Will be waiting for your updates
can someone explain these numbers to me please
Which numbers do you mean?
Benchmark test numbers = scores / results of the corresponding benchmark app.
Battery test numbers:
Battery loss of x% per hour:
Battery power was a% at the beginning of the test and b% at the end, so the battery loss was a-b% during the test. This number would be only "comparable", if all the test are taking exactly the same time. As sometimes the battery test is slightly shorter and sometimes a little bit longer, I "normalize" this battery loss by dividing by the testing time t (in hours).
So I receive a battery loss x = (a-b) / t in %/hour, which is better comparable for all tests.
The other numbers are ratios in % for the different phone states/times during the test (Awake time, Screen on time, Deep sleep time, etc. divided by the testing time) - I get these ratios/values from BetterBatteryStats (BBS) app.
And in mixed usage battery test, I saw a direct relation between screen on time and the battery loss per hour, and as I do not have exactly the same screen on time in all tests, I tried to "normalize" this battery loss per hour by the mentioned formula:
loss*(1-screen) = battery loss in % (per hours) * (100% - screen on time in %)
So, it would be the battery loss in % (per hour) for the screen off time.
Tests for SiyahKernel v3.3.1 are finished:
Depp sleep in over-night use is very good, but battery life in mixed usage is not the best for SiyahKernels.
SiyahKernel v3.2.7.2 had significant better battery life on my phone for my "normal" usage - all tests on complete default settings (no UC/UC/manual tweaking).
Depp sleep in over-night use is very good, but battery life in mixed usage is not the best for SiyahKernels.
SiyahKernel v3.2.7.2 had significant better battery life on my phone for my "normal" usage - all tests on complete default settings (no UC/UC/manual tweaking).
CamoGeko
Retired Forum Moderator
Was on 3.2.8.2 till today, recently flashed 3.3.1... will see how it goes within the next few days.
Sent from my 80GB CyanogenMOD 9 + Siyah ICS powered beast. Booya!!
If you do try Phenomenal Extreme and encounter any issues with it, then you can also try Phenomenal Medium or Phenomenal Normal. These are the same as Extreme but less agressively undervolted. This can help if you find Extreme too unstable but still want good battery performance.
Sent from my GT-I9100 using Tapatalk 2
Issue report from nighly usage battery test with DormianX 3.2:
My phone is dead after the first night with DorimanX 3.2.
Battery drained completely empty from 100% to dead.
Now I have to charge the battery again instead of documenting the over-night battery usage stats.
I hope that I can find the culprit for draining my battery and preventing deep sleep ... while doing nothing.
I had wifi, GTalk, Skype running like every night.
Normal battery consumption is between 0,7 and 1,2 %/hour.
EDIT:
last_kmsg doesn't show any culprit and battery stats are empty.
Will wipe cache and Dalvik cache again and re-do the test.
My phone is dead after the first night with DorimanX 3.2.
Battery drained completely empty from 100% to dead.
Now I have to charge the battery again instead of documenting the over-night battery usage stats.
I hope that I can find the culprit for draining my battery and preventing deep sleep ... while doing nothing.
I had wifi, GTalk, Skype running like every night.
Normal battery consumption is between 0,7 and 1,2 %/hour.
EDIT:
last_kmsg doesn't show any culprit and battery stats are empty.
Will wipe cache and Dalvik cache again and re-do the test.
Last edited:
Maybe l2_hsic wakelock - if he uses n7000 sources like GM used on 3.2.8.
Yes, had the same idea.
Tested during the day with custom ref in BBS and recognized a slightly higher wakelock for Seven (Z7).
Maybe there is a social hub/Google sync app preventing my phone from deep sleep.
Still examine the battery life and stats.
Will report my findings.
Another wiping cache and Dalvik cache could fix the complete draining during the night, but still very high battery consumption while doing nothing.
I think about aborting battery test for DorimanX 3.2, as its battery life is too bad on my phone to be a daily driver.
I think about aborting battery test for DorimanX 3.2, as its battery life is too bad on my phone to be a daily driver.
- Status
Similar threads
- Locked
Top Liked Posts
-
There are no posts matching your filters.
-
230Results for Siyah / Jeboo / DorimanX / SpeedMod / Apolo / PhilZ-touch / stock kernels
As I like Samsung/stock ROMs, I try to find the newest/best Sammy ROM as base for my kernel tests.
My target is to find the best-balanced kernel/ROM combination for good performance (smoothness) and long battery life ... and that without any scripts/tweaks or manual undervolting (UV) / underclocking (UC) etc..
From time to time, I switch my base ROM and create a new kernel comparison list - you will find older kernel result lists in the following posts.
In this OP, I will continuously update my latest test results.
Everything is specific to my phone, my "normal" (light) usage and my home location. Most values are exact results, but some are rounded. I usually have wifi, sync (three email accounts, calendar and news), Google Talk/Hangout on – wifi and Skype running if specified.
For all custom kernels, I was running on the stock ROM specified in brackets.
For stock kernels, I was running on the corresponding stock ROM.
Always using the Samsung's standard / stock 1650mAh battery.
For all kernels, I have changed nothing - just default settings. (Except there is an explicit information for a change in brackets.)
Last added: Apolo v4.14 (Regular) on XWMS3 NeatROM v5.8 (battery tests and benchmark results)
Here are my current results:
Battery stats
Code:[B]Mixed (day and night) usage (*)[/B] [B]Short[/B] (battery loss in % per hour) Apolo v4.14 regular: 0,94 Apolo v4.13 regular: 1,01 Apolo v4.10 regular: 1,08 Apolo v4.12 regular: 1,08 Apolo v4.11 regular: 1,12 NeatROM XWLSZ: 1,13 Dorimanx v8.35: 1,14 NeatROM XWLS8: 1,23 SpeedMod K5-8: 1,24 PhilZ v5.10.0: 1,24 NeatROM XWLSS: 1,31 NeatROM XWMS1: 1,31 NeatROM XWLSW: 1,32 Siyah v6.0b5: 1,35 NeatROM XWLSD: 1,38 NeatROM XWMS1 (v5.4.2): 1,38 XWLS8: 1,39 NeatROM XWLSJ: 1,43 Jeboo v2.2: 1,47 NeatROM XWLSN: 1,47 NeatROM XWMS2: 1,49 XWLSD: 1,53 XXLSJ: 3,29 [B]Long[/B] Apolo v4.14 regular (XWMS3 NeatROM Lite v5.8): Battery loss of 0,94% per hour, Wifi: 9,7%, Awake: 10,6%, Screen: 9,6% => loss*(1-screen): 0,85% Apolo v4.13 regular (XWMS2 NeatROM Lite v5.6): Battery loss of 1,01% per hour, Wifi: 10,2%, Awake: 11,0%, Screen: 10,1% => loss*(1-screen): 0,91% Apolo v4.10 regular (XWLSW NeatROM Lite v4.8): Battery loss of 1,08% per hour, Wifi: 9,2%, Awake: 10,3%, Screen: 9,1% => loss*(1-screen): 0,98% Apolo v4.12 regular (XWMS1 NeatROM Lite v5.4.2): Battery loss of 1,08% per hour, Wifi: 9,1%, Awake: 10,2%, Screen: 9,0% => loss*(1-screen): 0,98% Apolo v4.13 regular (XWMS1 NeatROM Lite v5.4.2): Battery loss of 1,11% per hour, Wifi: 10,1%, Awake: 11,2%, Screen: 10,0% => loss*(1-screen): 1,0% Apolo v4.12 regular (XWMS1 NeatROM Lite v5.1): Battery loss of 1,11% per hour, Wifi: 9,9%, Awake: 10,8%, Screen: 9,8% => loss*(1-screen): 1,00% Apolo v4.11 regular (XWMS1 NeatROM Lite v5.1): Battery loss of 1,12% per hour, Wifi: 10,2%, Awake: 11,2%, Screen: 10,1% => loss*(1-screen): 1,01% Apolo v4.13 regular (XWMS3 NeatROM Lite v5.8): Battery loss of 1,12% per hour, Wifi: 10,0%, Awake: 10,8%, Screen: 9,9% => loss*(1-screen): 1,01% NeatROM XWLSZ: Battery loss of 1,13% per hour, Wifi: 9,9%, Awake: 10,8%, Screen: 9,8% => loss*(1-screen): 1,02% Dorimanx v8.35 (XWLSW NeatROM Lite v4.8): Battery loss of 1,14% per hour, Wifi: 10,1%, Awake: 11,0%, Screen: 10,0% => loss*(1-screen): 1,03% NeatROM XWLS8: Battery loss of 1,23% per hour, Wifi: 14,6%, Awake: 11,5%, Screen: 10,2% => loss*(1-screen): 1,10% SpeedMod K5-8 (XWLSW NeatROM Lite v4.8): Battery loss of 1,24% per hour, Wifi: 10,0%, Awake: 11,6%, Screen: 10,0% => loss*(1-screen): 1,12% PhilZ v5.10.0 (XWLSW NeatROM Lite v4.8): Battery loss of 1,24% per hour, Wifi: 14,0%, Awake: 11,6%, Screen: 9,9% => loss*(1-screen): 1,12% NeatROM XWLSS: Battery loss of 1,31% per hour, Wifi: 10,5%, Awake: 11,9%, Screen: 10,3% => loss*(1-screen): 1,18% NeatROM XWMS1: Battery loss of 1,31% per hour, Wifi: 10,4%, Awake: 11,4%, Screen: 10,1% => loss*(1-screen): 1,18% NeatROM XWLSW: Battery loss of 1,32% per hour, Wifi: 10,2%, Awake: 11,7%, Screen: 10,1% => loss*(1-screen): 1,19% Siyah v6.0b5 (XWLSW NeatROM Lite v4.8, Pegasusq): Battery loss of 1,35% per hour, Wifi: 10,4%, Awake: 12,3%, Screen: 10,2% => loss*(1-screen): 1,21% NeatROM XWLSD: Battery loss of 1,38% per hour, Wifi: 14,6%, Awake: 12,4%, Screen: 10,1% => loss*(1-screen): 1,24% XWLS8: Battery loss of 1,39% per hour, Wifi: 14,4%, Awake: 12,5%, Screen: 10,2% => loss*(1-screen): 1,25% NeatROM XWMS1 (v5.4.2): Battery loss of 1,38% per hour, Wifi: 8,7%, Awake: 9,7%, Screen: 8,6% => loss*(1-screen): 1,26% NeatROM XWLSJ: Battery loss of 1,43% per hour, Wifi: 16,1%, Awake: 12,4%, Screen: 9,7% => loss*(1-screen): 1,29% NeatROM XXLSJ (FD disabled, no Skype): Battery loss of 1,46% per hour, Wifi: 17,4%, Awake: 13,8%, Screen: 10,1% => loss*(1-screen): 1,31% Jeboo v2.2 (XWLSS NeatROM Lite v4.7): Battery loss of 1,47% per hour, Wifi: 10,1%, Awake: 12,9%, Screen: 10,0% => loss*(1-screen): 1,32% NeatROM XWLSN: Battery loss of 1,47% per hour, Wifi: 15,9%, Awake: 12,0%, Screen: 9,8% => loss*(1-screen): 1,33% NeatROM XWMS2: Battery loss of 1,49% per hour, Wifi: 9,8%, Awake: 10,7%, Screen: 9,7% => loss*(1-screen): 1,35% Siyah v6.0b5 (XWLSD NeatROM Lite v4.3, Pegasusq): Battery loss of 1,49% per hour, Wifi: 15,3%, Awake: 12,8%, Screen: 10,2% => loss*(1-screen): 1,34% XWLSD: Battery loss of 1,53% per hour, Wifi: 20,8%, Awake: 13,2%, Screen: 10,2% => loss*(1-screen): 1,37% NeatROM XWMS1 (v5.4): Battery loss of 1,93% per hour, Wifi: 10,1%, Awake: 11,6%, Screen: 10,1% => loss*(1-screen): 1,74% XXLSJ: Battery loss of 3,29% per hour, Wifi: 100%, Awake: 37,3%, Screen: 10,0% => loss*(1-screen): 2,96% [B](*)[/B] In order to get a value, which allows to compare test results with different screen on times for mixed usage battery tests, I tried to „normalized“ the battery loss per hour with the following formula: [U]loss*(1-screen):[/U] battery loss in % (per hours) * (100% - screen on time in %) So, it would be the battery loss in % (per hour) for the screen off time. That is how a kernel with the same battery loss per hour will get a better/lower value, if it has more screen on time. (And the other way round.)Code:[B]Night use (mobile data, sync and Talk on)[/B] [B]Short[/B] (battery loss in % per hours) Apolo v4.14 regular: 0,17 Apolo v4.13 regular: 0,18 NeatROM XWMS1 (v5.4.2): 0,24 Apolo v4.11 regular: 0,36 NeatROM XWMS2: 0,37 Apolo v4.12 regular: 0,40 PhilZ v5.10.0: 0,40 NeatROM XWMS1: 0,41 Apolo v4.10 regular: 0,43 NeatROM XWLS8: 0,44 Dorimanx v8.35: 0,44 NeatROM XWLSS: 0,45 NeatROM XWLSZ: 0,47 SpeedMod K5-8: 0,52 XWLSD: 0,53 NeatROM XWLSD: 0,55 NeatROM XWLSN: 0,56 Siyah v6.0b5: 0,56 NeatROM XWLSW: 0,58 NeatROM XXLSJ: 0,58 Jeboo v2.2: 0,63 XWLS8: 0,65 XXLSJ: 1,63 [B]Long[/B] Apolo v4.14 regular (XWMS3 NeatROM Lite v5.8): Battery loss of 0,17% per hour, Awake: 1,0% Apolo v4.13 regular (XWMS3 NeatROM Lite v5.8): Battery loss of 0,18% per hour, Awake: 0,9% NeatROM XWMS1 (v5.4.2): Battery loss of 0,24% per hour, Awake: 1,2% Apolo v4.13 regular (XWMS2 NeatROM Lite v5.6): Battery loss of 0,35% per hour, Awake: 1,0% Apolo v4.13 regular (XWMS1 NeatROM Lite v5.4.2): Battery loss of 0,35% per hour, Awake: 1,1% Apolo v4.11 regular (XWMS1 NeatROM Lite v5.1): Battery loss of 0,36% per hour, Awake: 1,0% NeatROM XWMS2: Battery loss of 0,37% per hour, Awake: 0,8% Apolo v4.12 regular (XWMS1 NeatROM Lite v5.1): Battery loss of 0,40% per hour, Awake: 1,2% PhilZ v5.10.0 (XWLSW NeatROM Lite v4.8): Battery loss of 0,40% per hour, Awake: 1,3% NeatROM XWMS1: Battery loss of 0,41% per hour, Awake: 0,9% Apolo v4.12 regular (XWMS1 NeatROM Lite v5.4.2): Battery loss of 0,41% per hour, Awake: 1,0% Apolo v4.10 regular (XWLSW NeatROM Lite v4.8): Battery loss of 0,43% per hour, Awake: 1,0% NeatROM XWLS8: Battery loss of 0,44% per hour, Awake: 1,1% Dorimanx v8.35 (XWLSW NeatROM Lite v4.8): Battery loss of 0,44% per hour, Awake: 1,3% NeatROM XWLSS: Battery loss of 0,45% per hour, Awake: 1,5% NeatROM XWLSZ: Battery loss of 0,47% per hour, Awake: 1,3% SpeedMod K5-8 (XWLSW NeatROM Lite v4.8): Battery loss of 0,52% per hour, Awake: 1,8% XWLSD: Battery loss of 0,53% per hour, Awake: 2,0% NeatROM XWLSD: Battery loss of 0,55% per hour, Awake: 2,1% NeatROM XWLSN: Battery loss of 0,56% per hour, Awake: 2,0% Siyah v6.0b5 (XWLSW NeatROM Lite v4.8, Pegasusq): Battery loss of 0,56% per hour, Awake: 2,2% NeatROM XWLSW: Battery loss of 0,58% per hour, Awake: 1,8% NeatROM XXLSJ (wifi off, FD disabled, no Skype): Battery loss of 0,58% per hour, Awake: 2,2% NeatROM XWMS1 (v5.4): Battery loss of 0,59% per hour, Awake: 1,8% Jeboo v2.2 (XWLSS NeatROM Lite v4.7): Battery loss of 0,63% per hour, Awake: 2,0% NeatROM XWLSJ: Battery loss of 0,63% per hour, Awake: 2,1% XWLS8: Battery loss of 0,65% per hour, Awake: 2,1% XXLSJ: Battery loss of 1,63% per hour, Awake: 20,7%
Benchmarks
Code:[B]AnTuTu[/B] Apolo v4.14 (XWMS3 NeatROM Lite v5.8): 11159 Apolo v4.13 (XWMS2 NeatROM Lite v5.6): 11117 Apolo v4.12 (XWMS1 NeatROM Lite v5.4.2): 10990 Apolo v4.13 (XWMS1 NeatROM Lite v5.4.2): 10983 Apolo v4.13 (XWMS3 NeatROM Lite v5.8): 10933 Apolo v4.11 (XWMS1 NeatROM Lite v5.1): 10848 Apolo v4.12 (XWMS1 NeatROM Lite v5.1): 10824 NeatROM XWMS2: 10564 NeatROM XWMS1: 10527 NeatROM XWLSZ: 10212 Dorimanx v8.35 (XWLSW NeatROM Lite v4.8): 10064 NeatROM XXLSJ: 10064 Siyah v6.0b5 (XWLSN NeatROM Lite v4.6.1): 9962 NeatROM XWLSS: 9946 PhilZ v5.10.0 (XWLSW NeatROM Lite v4.8): 9929 SpeedMod K5-8 (XWLSW NeatROM Lite v4.8): 9887 Jeboo v2.2 (XWLSS NeatROM Lite v4.7): 9767 NeatROM XWLSN: 9711 NeatROM XWLSW: 9437 NeatROM XWLSD: 9399 Stock XWLS8: 9399 NeatROM XWLS8: 9248 NeatROM XWLSJ: 8799
My credits go to:
- Intratech for his excellent Official Firmware thread and his support
- Faryaab for his Stock Official Firmwares thread for recent ICS and JB ROM versions (kind of cooperation with SamMobile)
- hardcore for his great SpeedMod kernel
- fenomeno83 for his superb idea to undervolt SpeedMod and create Phenomenal kernel, and to geroxx for his hiroX kernel.
- gokhanmoral for his "father" SiyahKernel and his technical descriptions
- simone201 for his "new spirit" N.E.A.K. and his collaboration
- pedestre for his fantastic Apolo kernel
- droidphile for his very helpful and understandable descriptions and clues
- Entropy512 for his detailed and technical way of pointing into the right direction
- geko95gek for his British humor and support
- Chainfire for his CF-Root kernel, Mobile Odin, TriangleAway, CF-Bench, etc, etc., etc.
- AndroidGX for his nice battery saving tips
- AMDPOWERFIST for his inspiration, tips and trick for maximum battery life
- chamonix for his very useful BetterBatteryStats (BBS) app
- ram3n for his easy and simple but working method for battery calibration
- dorimanx for his support, team working and super-highly configurable DorimanX kernel
- _Sale_ for his excellent NeatROM, never-ending support and outstanding friendliness
- ApriliaM3 for his dev-host collection of various needful files, his helping way ... all over xda
- THE RED BLUR for his cave with useful information and instructions for clean flashing ROMs/kernels and his motivated contributions in various threads
- Many40 for his awesome guide for best battery life on NeatROM with SiyahKernel (check out the whole thread for all of his shared configs and support) and his polite way to help where possible
- hawkerpaul for his handy scripts, besides this he made some really nice mods for NeatROM (just check ApriliaM3's dev-host files)
- All kernel developers for their work, support and providing their kernels for free
- XDA for this wonderful platform to share knowledge and information about development, smartphones, battery stuff, settings, apps, etc.
Thread rule(s):
1. General xda rules.
2. Please no "My is the best ROM/kernel" statements.
You are all welcome to share your experiences and test results.
And I accept if you feel that this or that kernel/ROM perfectly fits your individual needs.
But I don't want that my thread becomes a platform for "my ... is better than yours" discussions.
3. For optimizing your battery life, please use BetterBatteryStats (BBS) app and post an idle dump of about 4 hours in BBS thread.
Don't mess this thread with such specific posts containing dump files, screenshots of your battery stats, etc. just for asking what drains your battery and how to optimize it.
Thanks for your understanding !46My last Jelly Bean (JB) kernel test results => last JB kernel test results
My last Ice Cream Sandwich (ICE) kernel test results => last ICS kernel test results
My last GingerBread (GB) kernel test results => last GB kernel test results
Current activity: Testing Apolo v4.14 on XWMS3 NeatROM v6.3 Aroma.
Some FAQs:
Question: Why has one kernel such good battery test results and the other is worse?
Answer:
My "normal" usage is kind of light.
No gaming, no extensive internet browsing/chatting, nearly no phone calls, no fm radio, no streaming, etc.
That's why the difference between the kernels are not that big.
I don't have an explicit reason why the one kernel has such good battery life on my phone and the other has worse results.
In suspicious cases, I already have re-done the test, and in the second battery cycle, the same kernel can have a little bit different battery life - but in most cases, the results were nearly the same.
As there are so many influences, even I try to keep all things unchanged, there are some I can not control directly.
So, the thing is that you flash the kernel on your phone and check it under "your" conditions.
Question: Why does it take so much time to post new test results?
Answer:
As I prefer doing a kind of "real-life" tests, these battery tests take some time to be finished.
Benchmarks are quickly done.
Battery stress tests - running heavy operations for a special time and checking the battery consumption - can also be quick.
But using the phone in a "normal" like every-day way, well ... to get reliable results, this tests have to take more time IMO.
And I try to use nearly the same period of time for all my mixed usage batters tests, which is about 20 hours.
Question: What do the different numbers mean?
Answer:
Benchmark test numbers = scores / results of the corresponding benchmark app.
Battery loss of x% per hour:
Battery power was a% at the beginning of the test and b% at the end, so the battery loss was a-b% during the test. This number would be only "comparable", if all the test are taking exactly the same time. As sometimes the battery test is slightly shorter and sometimes a little bit longer, I "normalize" this battery loss by dividing by the testing time t (in hours).
So I receive a battery loss x = (a-b) / t in %/hour, which is better comparable for all tests.
The other numbers are ratios in % for the different phone states/times during the test (Awake time, Screen on time, Deep sleep time, etc. divided by the testing time) - I get these ratios/values from BetterBatteryStats (BBS) app.
And in mixed usage battery test, I saw a direct relation between screen on time and the battery loss per hour, and as I do not have exactly the same screen on time in all tests, I tried to "normalize" this battery loss per hour by the mentioned formula:
loss*(1-screen) = battery loss in % (per hours) * (100% - screen on time in %)
So, it would be the battery loss in % (per hour) for the screen off time.
Question: What is most important for you/me to "judge" a kernel/ROM?
Answer:
1. Real-life experience for stability => does the system work without any issues.
2. Real-life experience for smoothness => does the system feel smooth and responsive.
3. Real-life experience for battery life => has the system a good battery life with low battery consumption.
...
4. Customizable => is the system highly customizable for tweaking/change settings/load profiles/etc.
5. Compatibility => is the system compatible with many popular tools/apps for tweaking/changing settings/applying profiles.
...
6. Benchmarks => has the system good benchmark results with high scores.
...
Maybe I forgot some aspects and every user has a different priority.
Question: Why is your/my mixed usage such "light"?
Answer:
Because this is my "normal" / everyday / real-life usage.
Sorry about that.
And for me, it is most important to test my system in real-life usage.
To be precise: I my real-life usage.
...
If you want some test results for another usage ... maybe your normal usage which seems to be "heavier", then be inspired to start testing different kernels on your phone.
You are welcome to share your results and experiences.
Question: What are your settings for better battery life?
Answer:
Have a look at this tips collection => http://xdaforums.com/showthread.php?p=16973603
Nothing more to add.
I disabled all location services, auto-sync, bluetooth, set to lowest brightness, use a black background, disabled some system apps (such as Social Hub, Games Hub, Music Hub, Samsung Hub, Google+, BBC, Zinio, etc.).
On GB and ICS, I enabled wifi all the time as I am in a strong wifi area. Since JB, I am using Better Wifi On/Off app to turn wifi off when phone is unplugged and screen off.
On GB and ICS, I had enabled fast dormancy, as my mobile provider supports it. But since JB I have disabled fast dormancy, as it has worse battery life with a massive fd wakelock.
Question: What is your xda username and what does it stand for?
Answer:
My xda username is "It_ler" = IT (information technology) + underscore + LER (German postfix for personalizing things).
Get a definition of the meaning of "IT'ler" in German => http://translate.google.de/translat...zenesprachenwiki.de/definition/itler/&act=url
Question: What apps have you disabled? What are your manual tweaks for better battery life?
Answer:
Here I posted a short summary of what I have done manually to get better battery life and a link to my disabled apps (specific for XXLSJ NeatROM Lite v4.0) => http://xdaforums.com/showpost.php?p=36229516&postcount=1359, and for XWLS8 NeatROM Lite v4.2 => http://xdaforums.com/showpost.php?p=37210595&postcount=1705. For stock XWLSD => http://xdaforums.com/showpost.php?p=37783486&postcount=1868.
Question: What autostarts have you disabled?
Answer:
I read a lot in the BBS app thread and found two very good advices how to reduce the number of alarms for Maps, GSF and Google Play Sevices. Here I linked to the instruction posts.
Usual procedure for testing:
First, I am searching for a stable, smooth and battery friendly stock ROM.
Then I try to keep as much unchanged as possible (same apps, settings, connections, Modem/Phone/Baseband, CSC, region, time for nightly use, time for mixed usage, screen on time, etc.) in order to get reliable results which can be reproduced.
After flashing a new kernel, I do a normal reboot and if I recognize any issue, I wipe cache partition and Dalvik cache in CWM recovery.
Normally I do not wipe/clear anything for switching kernels, as I do not use any kind of scripts, tweaks, manual UV/UC, etc..
I test all kernel on its completely default settings => kind of "Out-of-the-box".
After booting the kernel and having tested all important functions (internet, wifi, synching accounts, root and busybox (BB) check, scrolling, swiping, browsing, etc.), I reboot the phone again and start the benchmark tests.
Then I shut down the phone and fully charge it to 100%.
Then I unplug the phone, pull out the battery for some minutes, put it back again and turn on the phone.
Usually I wait so a while and start the battery test somewhere in the 95-99% ... as I am flashing in the evening, I start directly with the night usage battery test.
I am using chamonix's BetterBattery Stats (BBS) app to set a custom reference and log all battery stats/wakelocks/alarms/times/etc.
This will take about 8-10 hours, depending on the start and end time
This nightly usage is part of the mixed usage battery test too.
Additionally in the day usage, I try to reproduce the same usage at the same places with nearly the same screen on time of about 2 hours in a time period of about 20 hours.
Sometimes I have to finish the test some minutes earlier or later, sometimes I have a little bit more or less screen on time, but I really try to keep everything which I can influence directly the same.
If I recognize some unusual wakelocks or significant differences from previous kernel versions, I do a second battery test in a second battery cycle.
And once I had to restore the nandroid backup of an older system to come back to my "usual" battery life/consumption and repeat the kernel tests.
But this is the absolute exception.
Normally, I do not have any issues switching ROMs and kernels.
