Give me your Tweaks...


  1. drewhill77

    drewhill77 Well-Known Member

    This weekend I am going to be benchmarking various Scripts and build.prop settings to see what if any effect they have to our phones. Each script will be tested individually and cumulatively.

    Platform -
    Stock B08c / Stock Kernel
    Deodexed
    Bloatware Removed
    Rooted/SuperSU/Busybox
    Bash/Nano/Sqlite
    init.d support via debuggered

    Benchmark Tools -
    AnTuTu Benchmark
    Speedtest.net

    Categories -
    CPU/RAM
    2d/3d Graphics
    Database I/O
    SD I/O
    3G Data
    WiFi Data

    Process -
    All build.prop tweaks will be set via ADB setprop command.
    All scripts will be ran via init.d on boot and removed after testing.
    Each benchmark will consist of 3 tests (to allow for variance) and averaged.
    Phone will be rebooted (to remove setprops) and cache/dalvik_cache wiped.

    So please post your Tweaks here and state what category they belong in as well as whether they are intended to run alongside additional Tweaks.

    Thanks:)

    Advertisement
  2. drewhill77

    drewhill77 Well-Known Member

  3. PsyMega

    PsyMega Well-Known Member

    drewhill77 likes this.
  4. MyndCTRL

    MyndCTRL Well-Known Member

    It's just a suggestion but you can start with the superphone 3 build.prop tweaks. We already know they work for our phone so I'd imagine that's a good place to start. I guess you can also compare that to the v6 supercharger script also found at xda.

    This is a great idea, It will take a great deal of time. I would appreciate it as much as everyone else especially when it comes to developing roms and choosing which tweaks to include. Thank you for your work!
    drewhill77 likes this.
  5. downthemachine

    downthemachine Developer Incognito VIP Member

    Then redo the test with swap ;)
  6. drewhill77

    drewhill77 Well-Known Member


    I actually considered doing that...but

    1. I don't have that much time on my hands

    2. If I did yours then I'd have to do shins oc kernel...again time and I don't wanna make this a competition

    I figure that stock kernel will give me the best baseline and solid results everyone can agree on
  7. mercury0x000d

    mercury0x000d Digital Packrat Guide

    I'll second that :)

    ...and try a third time with the 1.6 OC kernel!
  8. drewhill77

    drewhill77 Well-Known Member


    Tell you what... once I get all the stock benchmarks done and figure out all the best tweaks I will do a fully loaded run on both oc and swap...ok?
  9. mercury0x000d

    mercury0x000d Digital Packrat Guide

    Sounds awesome, man :D


    :party:
  10. MyndCTRL

    MyndCTRL Well-Known Member

    or what you can do is grab 2 more ppl, one to test stock kernel with tweaks and post results, and the same with the 2 or 3 other kernels and compare resullts. there's no need for you to take this huge task on all by yourself :) and im sure there's willing able bodies to help you
  11. drewhill77

    drewhill77 Well-Known Member

    updated OP round 1 results for video acceleration...no major differences from one setting to the next according to AnTuTu benchmark...
  12. drewhill77

    drewhill77 Well-Known Member

    If anybody else wants to benchmark various tweaks with other kernels they're more than welcome to it :).

    I've just never been convinced that all of these tweaks actually have an effect on our phones...right now I'm trying to test those that should some benchmarkable effect. Some, like window scrolling and max windows will only have a visual effect...same with fast reboot and no bootanim.

    The biggest issue I've come against and why it's taking so long is that I've had to gather up all of the various tweaks, collate them and try to separate them into groups by what there effect is...I've mostly got that out of the way and now am in the process of testing
  13. PattF

    PattF Guest

    [​IMG]

    Here's mine.
  14. Tigatron

    Tigatron Well-Known Member

    Looks like some make a little difference on some areas.
  15. drewhill77

    drewhill77 Well-Known Member

    Except for when i ran debug.composition.type=hw...and that was an exceptionally low score...everything else was inside of a normal variance for benchmarking. I'm expecting better results from testing memory and 3g tweaks though....hopefully tomorrow night.

    Sadly GB/stock kernel doesn't allow much in the way of hardware acceleration...best we've got is the fact that we do have a gpu and eventually may be able to overclock it ;) As far as I know neither shins or dtms kernels have touched on graphics...
  16. Tigatron

    Tigatron Well-Known Member

    Yea I don't see that one in the build.prop from noobV (probably a good thing)

    I'm running noobV (removed swap) and running shins 1.6 oc
    My scores are great. CPUi = 1140 CPUpf = 461 2D = 301 3D = 1197
  17. MyndCTRL

    MyndCTRL Well-Known Member

    You show me what to do and how to do it (or point me in the direction o find out) I will be more than happy to help you out sir!
  18. drewhill77

    drewhill77 Well-Known Member

    Here is a list of the build.prop tweaks I'm testing prop test

    Start out with stock build.prop in /system/.
    Using ADB shell type "getprop | grep <property>"
    -replace <property> with whatever property you're checking to see if it exists


    Code:
    getprop | grep debug.composition.type
    note whether or not that property already exisisted
    then type "setprop <property> <setting>"
    -replace <property> and <setting>

    Code:
    setprop debug.composition.type gpu
    run "getprop | grep <property>" again to confirm the setting

    run antutu 3 times and average the results...

    reboot your phone after the tests and the property will revert to system standard or disappear if it didn't exist prior to your setting it...except for properties starting with "persisit". Properties starting with "persist" will carry through on a reboot if they don't already have a setting in init.rc or build.prop. Persistent props are generally true/fale (1/0) so just change it to the opposite of what you were testing...or reinstall your rom to wipe it out entirely.
  19. MyndCTRL

    MyndCTRL Well-Known Member

    Oh awesome, I'll get started then. Now I guess what I need is a copy of shin's or downs kernel on a copy of stock, rooted rom. is that correct?
  20. drewhill77

    drewhill77 Well-Known Member

    Lol As close to stock as you can get...obviously oc needs to be oc'd and swap needs an active swapfile since those are the primary aspects of each... I don't know if additional properties were coded in either kernel.
  21. downthemachine

    downthemachine Developer Incognito VIP Member

    Problem is with mine, there are changes. V1 was just swap, v1.10 removed ext2 and ext3 support for a little more free memory, v1.11-2 not released. V1.13 adds ramzswap support(Google it). So there are many different versions that may affect testing. Noob is running the latest public release. There are other builds not quite ready, or hush hush. And as far as versioning, there is not much of a way to tell being that any visible changes can't be titled because they will break wifi. Blah blah blah, I'm done rambling on, good night :cool:
  22. drewhill77

    drewhill77 Well-Known Member

    OK...running through other settings I'm not seeing any results from AnTuTu outside of normal variance. Got to the point where I would change a setting and only test once. This is not to say that these settings don't have an effect...they may simply not be benchmarkable. At this point I am skipping ahead to the 3g/Data tweaks...which I know have an effect. I will test each setting 5 times and post the hi/low range of ping, down and up speeds.
    Tigatron likes this.
  23. Tigatron

    Tigatron Well-Known Member

    Sounds good I have more interest in the speed test results.

    What prl will you be using?
  24. drewhill77

    drewhill77 Well-Known Member

    I'm on PRL Version 61009 in Indiana...testing on PowerNet Global in Cincinnati, OH.

    My Stock Ranges over 5 tests are:
    ping 132 - 201
    down Mbps 0.54 - 0.96
    up Mbps 0.30 - 0.75
  25. Tigatron

    Tigatron Well-Known Member

    OK.

    I swapped out to the newest sprint prl (60690) and that seemed to make a big difference for my 3G speed over the Boost prl

Share This Page