[Battery Showdown] Top captivate roms battle it out! Updated 3/31 w Phoenix results

[jivy26]

Test # 1 Video Streaming Over WiFi

ROMS Tested So Far:

Continuum
Serendipity
Precision
Firefly
Cognition
Phoenix
Paragon
Twingerbread
Supernova
Darky's
di11igaf
Dragon
Andromeda

**DISCLAIMER**
Keep in mind that results will vary for each phone and just because I get a certain battery life using the rom's and kernel's listed below does not necessarily mean that you will.

Tired of constantly flashing to find great battery life only to realize the rom didn't meet your expectations; then this thread is for you. Over the next few weeks I will test every rom imaginable and post real world results that will include battery stats, linpack and quadrant results.

**This is not meant to put down any particular ROM. Without our great dev community my captivate would have been returned and thank you all for your hard work.**

[Goal]
To provide the captivate community with real world results from top roms in everyday usage where battery life is the key focal point, but performance is not affected. Hopefully this will help save you time from having to flash multiple ROMS.

[Roms]
*This list will continue to grow*

Stock 2.2 kept fcing so will test it on test 2 this evening 3/28

[Usage]
Each night I will run the exact same VERY LONG documentary using VLC stream and convert and current widget.


[Specs]
*Every rom will run "out of the box" meaning Kernel, Modem, Settings all default to how it was cooked*
° Default clockspeed 1000mhz (as suggested by a smart person listed below)
° 40% Brightness/Screen Timeout Turned Off
° 100% Charge each run (0mV to make ryude get a chubby)
° Auto Sync off
° GPS off
° Zeam launcher
° VLC Steam & Convert used w/ same long documentary used each time
° Phone is placed in exact same spot every test and is unplugged as soon as I hit play


Results So Far

 

[jivy26]

Test # 2 Web Browsing

Test # 2 Web Browsing on 3g

Roms Tested So Far:
Continuum
Serendipity
Precision
Stock 2.2
Firefly
Phoenix
Paragon

Currently Running:
Twingerbread

Upcoming:
Supernova
Darky's
di11igaf
Dragon
Andromeda
Juiced
Cognition

Goal:

In previous tests I tested on wifi using a vlc stream & convert. With this way of testing I'm hoping to get more realistic real-world results that you can use as a guideline to a roms overall battery performance . Keep in mind your results may vary.

Testing:

Every night I will set tab reload to 15 seconds and it will continuously reload XDA website until the phone dies. It will be placed in the same spot in my house every night and will be ran on phone data, thank god for the unlimited plan.

Applications I will be using:
- Dolphin Browser HD
- Reload tab plug-in
- Current widget
- My Settings

[Specs]
*Every rom will run "out of the box" meaning Kernel, Modem, Settings all default to how it was cooked*
° Default clockspeed 1000mhz
° 40% Brightness/Screen Timeout Turned Off
° 100% Charge each run
° Auto Sync and GPS On (To help give near real world results)
° 3g
° Zeam launcher
° Phone is placed in exact same spot every test and is unplugged as soon as I start Reload Tab Plug-in.
**Will be running all tests on 3g as some ROMs do not have HSPDA**

Results

 

[jivy26]

Reserved for future use

 

[TheSopranos16]

Cool. Any chance you can add stock Froyo to the list as a reference point? It would be interesting to see how much custom rom's are improving over stock...

 

[jivy26]

Cool. Any chance you can add stock Froyo to the list as a reference point? It would be interesting to see how much custom rom's are improving over stock...

Great suggestion only thing is I won't be able to undervolt so numbers will not be accurate to this testing, but I might run the best two roms after I'm done testing at there defaults against stock froyo.

 

[diablo009]

Since many users are set at the standard 1000, would you consider the testing at standard 1000 speed for a more accurate battery life any user can expect. Just a suggestion

 

[ryude]

Sounds like Precision should be included, with the black theme it easily gets great battery life.

 

[jivy26]

Is this under-clocked? I am on the same ROM with speed at 1000 with no under-volting.

I get 1600+ scores on quadrant, and 14+ scores on linpack (tested out even now).

Yes I listed in the specs what they will all be running at. All are uc/uv to 600mhz.

 

[Enhanced]

I have been waiting for something like this but I assume that the devs would all be up in arms if someone rated their ROMs higher/lower.

If this thread will be allowed to continue, totally looking forward to the results.

And you should totally throw Precision into the mix as it seems like the de facto battery saving ROM as it skimps on a few services.

 

[goob1284]

I have been waiting for something like this but I assume that the devs would all be up in arms if someone rated their ROMs higher/lower.

If this thread will be allowed to continue, totally looking forward to the results.

And you should totally throw Precision into the mix as it seems like the de facto battery saving ROM as it skimps on a few services.

Personally, if I was a dev and saw that my rom wasn't too good, I would try to make something better. So who knows how it turns out. A little competition wont hurt anyone, and hopefully have better roms! Btw, good idea with this

Sent from my SGH-I897 using XDA App

 

[jivy26]

Just finished di11i on to paragon. The theme on di11i is very nice too almost a carbon fiber look.

 

[bravomail]

add Torch 1.5

Also try Torch 1.5. So far been outstanding in battery department.

 

[byteme2008]

I know this doesn't help as far as battery life, but just for a reference as far as stock 2.2:

1. Linpack - around the 14 mflops (sometime in the 13 range, sometimes in the 14)
2. Quadrant - 1050 or so
3. Nenamark1 - 51 fps (not sure if anyone uses that one much, but thought I would throw it in as I had it available lol)

I would estimate the battery life if I didn't use the phone at all (just left it sitting in standby) would be at 60hours plus, as I get up to around 40 with light usage on 2.2 with battery to spare (haven't run it dead yet to be honest on 2.2) Still a little early though as the phone hasn't been drained and recharged full thus far.

I just checked my phone, and I have had it on for over eight hours, and I am at 92% battery left. I have hardly used the phone today (just a few texts is all), but going on that I should get around 80 hours on a single charge if I just leave it sitting there.

 

[james.roddy]

Great idea. I'm a huge continuum fan

Sent from my GT-I9000 using XDA App

 

[illmatic90]

good info to know...thanks alot!

 

[Ikonomi]

This sounds like quite an undertaking. Good luck with the project. I'm looking forward to the comparisons. Thank you for your hard work to benefit Captivate owners out there.

Sent from my SGH-I897 using XDA Premium App

 

[jivy26]

Charge drain charge drain lol gonna get online now and buy a new battery just incase i destroy this one.

 

[stigFX]

LOcation matters

I am using JuicePlotter to monitor my battery usage closely. From what I see, different locations give dramatically different battery life. At home my battery drops 1% per hour. At work 20 miles away it drops 2% on captivate modems but 4-5% on i9000 modems. So the signal strength, cell tower type make big difference.

I have tried multiple Cog ROMs and Paragons ROMs, currently on Twingerbread 3.2. The best ROM in terms of battery life is cog 3.2 for me.

 

[jivy26]

Yea but most people throughout the day are on the move and all I'm trying to achieve are results based on that. The constant will be the kernel in this test that way any thing I log can be directly attributed to the rom/modem. This by no means is the end all, but it will give the community a great idea of where each rom sits.

 

[Burstaholic]

Interesting kernel choice. I've heard Suckerpunch lauded, but that ignores the fact that most of these roms flash with a different kernel by default. Is it a better test to run them all with the same kernel, or with the one they chose?

I dunno. Might be interesting to test and formally see how much kernel choice affects things.