Search This thread

irminsul

Senior Member
Nov 11, 2010
362
157
Almere
Re: [KERNEL][GPL][GN] franco.Kernel-r361 -> 4.2 | r300 -> 4.1

About hot reboots: having them every now and then, every two days or something. Totally irrelevant when it happens: while listening music, or just when the phone is laying on the table with the screen on. Hopefully Google fixes this, not Franco related.

Sent from my Nexus 7 using Tapatalk HD
 

LibertyMonger

Senior Member
Mar 13, 2011
6,280
2,284
Cincinnati
About hot reboots: having them every now and then, every two days or something. Totally irrelevant when it happens: while listening music, or just when the phone is laying on the table with the screen on. Hopefully Google fixes this, not Franco related.

Sent from my Nexus 7 using Tapatalk HD


I have never had a reboot or sod using Franco especially the latest one. Never while running Xylon ROM and now over a month on Baked BB7.
 

Nighthound

Senior Member
Jan 18, 2010
188
19
Great, let's put this clear for everybody

Max CPU frec: 1228
Min CPU frec: 384 (SOD avoiding)
Governor: Interactive

above_hispeed_delay: 15(k)
boost: don't touch
go_hispeed_load: 95
hispeed_freq: 0.7=729MHz
min_sample_time: 45(k)
time_rate: 15(k)
target_loads: 85

IO Scheduler: row
Read Ahead: 2048 (Can't be set through f.ku YET)

What about the max CPU freq for screen off?

Also just for completeness, what's the recommend TCP congestion avoidance algo? Is it westwood?
 

irminsul

Senior Member
Nov 11, 2010
362
157
Almere
Re: [KERNEL][GPL][GN] franco.Kernel-r361 -> 4.2 | r300 -> 4.1

I have never had a reboot or sod using Franco especially the latest one. Never while running Xylon ROM and now over a month on Baked BB7.

It isn't a hard reboot though, when you check your uptime the phone doesn't see the reboot. Just the glowing X bootscreen.
SOD doesn't happen on Franco, luckily because my phone is sensitive for it :)

Sent from my Galaxy Nexus using Tapatalk 2
 
  • Like
Reactions: ChosSimbaOne

Fenvarien

Senior Member
Dec 22, 2010
963
598
AW: [KERNEL][GPL][GN] franco.Kernel-r361 -> 4.2 | r300 -> 4.1

The final winner tunables are running perfectly stable here on the GNex of my wife and on my own GNex.
No problems with 192/1228 :good:

I used a lot of gaming, listening to music, web browsing with Chrome, Tapatalk and Repligo Reader (PDFs). Performance and battery life are just amazing, no stability issues at all. :)
 

marcocsh

Senior Member
Apr 12, 2011
295
39
Re: [KERNEL][GPL][GN] franco.Kernel-r361 -> 4.2 | r300 -> 4.1

The final winner tunables are running perfectly stable here on the GNex of my wife and on my own GNex.
No problems with 192/1228 :good:

I used a lot of gaming, listening to music, web browsing with Chrome, Tapatalk and Repligo Reader (PDFs). Performance and battery life are just amazing, no stability issues at all. :)

They are prefect except that I need to set read ahead buffer to 128 for listening to FLACs.:thumbup:

Sent from my Galaxy Nexus using xda app-developers app
 

eng.stk

Senior Member
Mar 29, 2011
7,357
61,789
Google Pixel 4a
OnePlus 9 Pro
This is for the N7 but a similar test was done for the GN a few months back:
https://docs.google.com/spreadsheet/ccc?key=0AqRFyVFtJ0vodDF2V2FrYlVSTkhseV9DWjlaRkltRmc#gid=6

The GN results were similar. IO speed increases to a point with higher read_ahead_kb buffer values, then decreases again.
If I recall correctly, 2048 was found to be the sweet spot both times.

Edit: Wrong graph.. I'll try and find the one with 3072, etc.

Here we go, it's from M-Kernel, but results are similar:
https://docs.google.com/spreadsheet/ccc?key=0AqRFyVFtJ0vodHQ0Q3dqa2NWUXBzSG4tRHNqSmwzUFE#gid=6

Yes, 1024 is the safe used by most users (and kernel devs), 2048 if I recall, as you say is "the point".
 

RevengeFNF

Senior Member
Dec 10, 2010
594
132
Porto
Benchmarks are one thing, real usage is another one.
With 2048 i feel my phone being sluggish in transitions. The best way to test that is scrolling the widget section very fast. Its simple a no go for me.

After that i tested some other values, and for now i found that 256 gives me the best transition in the Widget Section.
 

irminsul

Senior Member
Nov 11, 2010
362
157
Almere
Re: [KERNEL][GPL][GN] franco.Kernel-r361 -> 4.2 | r300 -> 4.1

I can confirm that, with 2048 my phone gets laggy. With 256 it's faster in transitions. It's just how you use your phone what's good for you :good:

Sent from my Galaxy Nexus using Tapatalk 2
 
  • Like
Reactions: drnejo

supernova_00

Senior Member
Feb 1, 2012
360
62
Aberdeen, MD
Most people I've seen comment say that 95/85 is slightly laggier than 90/80, so I'm with Fen, I'd like to see if we can tell the difference with:

a) 15/90/0.7/45/15/80
b) 15/90/0.7/45/30/80

Hmm.. but you're right, maybe lowering the timer_rate to 15 allows us to use 95/85... okay, so:

c) 15/95/0.7/45/15/85
I didn't know exactly which comment to quote since there have been dozens on comments with different tunables but why the just from 30 to 15? I haven't seen anything in between like 20 and 25. We might as well get it perfect instead of just dividing the numbers in half each time for a test, I've seen every other number be tested in increments of "5" except this one.
 

Fenvarien

Senior Member
Dec 22, 2010
963
598
AW: [KERNEL][GPL][GN] franco.Kernel-r361 -> 4.2 | r300 -> 4.1

I can confirm that, with 2048 my phone gets laggy. With 256 it's faster in transitions. It's just how you use your phone what's good for you :good:

Sent from my Galaxy Nexus using Tapatalk 2

That's true, it depends from personal usage. Until Read Ahead becomes configurable via f.KU I use an init.d script to set it to my taste :)
 

stevensoaj

Senior Member
Jul 2, 2012
431
334
31
Benchmarks are one thing, real usage is another one.
With 2048 i feel my phone being sluggish in transitions. The best way to test that is scrolling the widget section very fast. Its simple a no go for me.

After that i tested some other values, and for now i found that 256 gives me the best transition in the Widget Section.

I'm so glad I saw your post. I have been annoyed by the lag when scrolling vertically through the widgets on nova launcher for quite some time. 256 buffer cleared things right up! With no noticeable draw backs so far. Can someone who knows more than me explain the benefits of higher buffer sizes?
 

irminsul

Senior Member
Nov 11, 2010
362
157
Almere
Re: [KERNEL][GPL][GN] franco.Kernel-r361 -> 4.2 | r300 -> 4.1

That's true, it depends from personal usage. Until Read Ahead becomes configurable via f.KU I use an init.d script to set it to my taste :)

Although 512 is a little smoother to me when opening Google now for example. Using that value now.

Sent from my Galaxy Nexus using Tapatalk 2
 
  • Like
Reactions: Fenvarien

joaquinf

Senior Member
Sep 17, 2010
2,844
2,327
Bucaramanga
I didn't know exactly which comment to quote since there have been dozens on comments with different tunables but why the just from 30 to 15? I haven't seen anything in between like 20 and 25. We might as well get it perfect instead of just dividing the numbers in half each time for a test, I've seen every other number be tested in increments of "5" except this one.

From yellowdie's post

Franco believes that by setting the timer_rate to be the same thing as the CPU sample_rate (which is preset at 30000), then that will make the CPU more efficient at switching

So that's why we've set that value to 30 or multiple of that e.g 30/2 = 15; 30*(3/2) = 45 and so on :)
 

joaquinf

Senior Member
Sep 17, 2010
2,844
2,327
Bucaramanga
Got cha, thanks! So these tunables have been implemented in the latest fku, correct?

Has anyone played with the new IO scheduler options?
Those are governor tunables and always have been available to tune on fku under frequencies/voltages > Governor control

On the newest version released today franco included I/O scheduler tuning which is a completely different thing and those options are available under custom kernel settings > io scheduler tuning

And no, i haven't played with that yet. I'm at work atm and sh*ty mondays are sh*ty so not so much spare time to play around :)
 

stevensoaj

Senior Member
Jul 2, 2012
431
334
31
Just so people know, read ahead can be checked with:

cat /sys/block/mmcblk0/queue/read_ahead_kb

and set with:

echo 2048 > /sys/block/mmcblk0/queue/read_ahead_kb

You might have to su first. :good:

Just to double check, would this be the correct init.d script to set read_ahead to 512 on boot?

Code:
#!/system/bin/sh
echo 512 > /sys/block/mmcblk0/queue/read_ahead_kb

Also, I'm not quite sure I completely understand the naming schemes of init.d scripts. Is a name of "10sdboost" ok?
 

Top Liked Posts

  • There are no posts matching your filters.
  • 1562
    I hate big OPs.

    Works on all roms. r390 and newer versions are ONLY for 4.3 or newer.

    F.A.Q:
    1. My device rebooted or crashed, how can I help?
    A: Get me /proc/last_kmsg on pastie.org.
    2. Battery sucks, my device is not entering deep sleep. FIX PLOX!
    A: Fix it yourself, it's an app waking your device up not the kernel's problem
    3. Signal is dropping since I flashed the kernel, amg u sucks!
    A: The kernel has nothing to do with gsm/cmda signal. Make sure you have the latest radios
    4. Do I need to wipe anything when flashing this kernel?
    A: No.
    5. Does this kernel has X or Y mod?
    A: Learn to read, everything you need to know is in the features list, changelog or public repo.

    Downloads:
    4.3: http://192.241.177.15/GalaxyNexus/4.3/
    4.4: http://192.241.177.15/GalaxyNexus/4.4/

    Kernel changelog:
    http://192.241.177.15/GalaxyNexus/4.3/appfiles/changelog.xml

    Source:
    https://github.com/franciscofranco/Tuna_JB_pre1/tree/nightlies-4.3

    franco.Kernel updater Free apk: http://xdaforums.com/showthread.php?t=1867127

    203
    [KERNEL][GPL][5 JUN - #Milestone 4][r200-ICS r210-JB] franco.Kernel | 4.0.3/4 |

    Terminal commands for all the options in this kernel:

    Color Multipliers:

    echo r g b > /sys/class/misc/colorcontrol/multiplier

    Replace r g and b for 2000000000 for stock color. If you want to change the colors just replace the first 3 numbers of each r/g/b with 0-400 of your choice.

    Gamma Control:

    echo r g b > /sys/class/misc/colorcontrol/v1_offset

    Replace r/g/b by -255/200 of your choice.

    Fsync:

    echo 1 > /sys/module/sync/parameters/fsync_enabled - to enable

    echo 0 > /sys/module/sync/parameters/fsync_enabled - to disable

    Sound Control:

    echo i > /sys/devices/virtual/misc/soundcontrol/volume_boost

    Replace i with 0-3 of your choice.

    Sound High Performance:

    echo 1 > /sys/devices/virtual/misc/soundcontrol/highperf_enabled - to enable

    echo 0 > /sys/devices/virtual/misc/soundcontrol/highperf_enabled - to disable

    USB Fast Charge:

    echo 1 > /sys/kernel/fast_charge/force_fast_charge - to enable

    echo 0 > /sys/kernel/fast_charge/force_fast_charge - to disable

    wifi_pm:

    echo 1 > /sys/module/bcmdhd/parameters/wifi_pm - to enable

    echo 0 > /sys/module/bcmdhd/parameters/wifi_pm - to disable

    Trinity's Contrast:

    echo i > /sys/module/panel_s6e8aa0/parameters/contrast

    Replace i with -25/16 of your choice.

    BLX:

    echo i > /sys/class/misc/batterylifeextender/charging_limit

    Replace i with 0-100 of your choice.

    OMAP Gamma interface:

    echo i > /sys/devices/platform/omapdss/manager0/gamma

    Replace i with 5-10 of your choice.

    Thermal Throttle:

    echo 1 > /sys/module/omap_temp_sensor/parameters/throttle_enabled - to enable

    echo 0 > /sys/module/omap_temp_sensor/parameters/throttle_enabled - to disable

    TCP Congestion Algorithm interface

    To check all the available options:

    sysctl net.ipv4.tcp_available_congestion_control

    To change to other option:

    sysctl -w net.ipv4.tcp_congestion_control=NAME_OF_THE_ALGORITHM

    Detailed test of all the algorithms:
    Latency - Download - Upload

    cubic:
    1st run: 15ms - 10,75Mbps - 7,82Mbps
    2nd run: 14ms - 10,84Mbps - 8,06Mbps

    reno:
    1st run: 13ms - 15,51Mbps - 6,73Mbps
    2nd run: 13ms - 14,73Mbps - 8,51Mbps

    bic:
    1st run: 12ms - 10,38Mbps - 8,61Mbps
    2nd run: 13ms - 10,78Mbps - 8,62Mbps

    westwood:
    1st run: 11ms - 17,65Mbps - 8,30Mbps
    2nd run: 13ms - 13,28Mbps - 8,29Mbps

    highspeed:
    1st run: 13ms - 10,76Mbps - 7,94Mbps
    2nd run: 16ms - 14,42Mbps - 8,52Mbps

    hybla:
    1st run: 14ms - 11,19Mbps - 7,44Mbps
    2nd run: 14ms - 13,47Mbps - 7,56Mbps

    htcp:
    1st run: 14ms - 13,24Mbps - 7,03Mbps
    2nd run: 15ms - 10,85Mbps - 8,00Mbps

    vegas:
    1st run: 14ms - 8,49Mbps - 6,62Mbps
    2nd run: 14ms - 12,00Mbps - 7,07Mbps

    veno:
    1st run: 13ms - 9,58Mbps - 8,13Mbps
    2nd run: 13ms - 8,50Mbps - 7,64Mbps

    scalable:
    1st run: 18ms - 12,01Mbps - 8,73Mbps
    2nd run: 14ms - 13,96Mbps - 8,23Mbps

    lp:
    1st run: 14ms - 14,90Mbps - 8,68Mbps
    2nd run: 14ms - 13,44Mbps - 8,72Mbps

    yeah:
    1st run: 14ms - 13,37Mbps - 8,28Mbps
    2nd run: 17ms - 13,89Mbps - 8,14Mbps

    illinois:
    1st run: 13ms - 12,93Mbps - 8,24Mbps
    2nd run: 16ms - 13,97Mbps - 6,46Mbps
    167
    Just woken up and feel like my head is going to explode already this last 5 pages is crazy

    haha. It's been a crazy two days. :) But it's been a blast.

    Now, sleep is over, time to get back to work!

    I got these from that thread:

    so it makes much sense, to make the min_sample_time as low as possible (?), but how low? what's the most appropriate sample time for battery and performance?

    for the timer_rate, franco suggested 30k to consider the CPUs latency. What has it to do with the cpu's latency?

    he also said min_sample_time doesn't have to be in multiple of timer_rate.
    in my case, all my timers are in 20k, which works fine as of now. But i must be missing some things, because I just saw somebody post these values, and no detailed explanation for having them.

    Yes and no. Here's what we're thinkin' so far.

    THIS POST WILL BE A RECAP OF THE LAST FEW PAGES OF RESEARCH! :)

    This was my original settings that I've been using for weeks:
    above_hispeed_delay: 20000
    go_hispeed_load: 50
    min_sample_time: 40000
    timer_rate: 20000
    So to make the short hand easier, we kept it in that order and just said: 20000/50/40000/20000 became 20k/50/40k/20k became 2/5/4/2. Make sense?

    Here is a breakdown of what they each mean:
    -above_hispeed_delay: Once speed is set to hispeed_freq, wait for this long before bumping speed higher in response to continued high load.
    -go_hispeed_load: The CPU load at which to ramp to the intermediate "hi speed".
    -min_sample_time: The minimum amount of time to spend at a frequency before we can ramp down.
    -timer_rate: Sample rate for reevaluating cpu load when the system is not idle.

    This is a good explanation that I wrote back on page 3038:
    -above_hispeed_delay: higher = better battery, lower = better performance. (100k is default)
    -go_hispeed_load:.......higher = better battery, lower = better performance. (50 is default)
    -min_sample_time:......lower = better battery, higher = better performance. (60k is default)
    -timer_rate:.................higher = better battery, lower = better performance. (20k is default)
    So Google's default is 10/5/6/2. Lower numbers are all better for performance except min_sample_time (there higher is faster). So our goal is to find a sweet spot.

    The default 10 is for "Once speed is set to hispeed_freq, wait for this long before bumping speed higher in response to continued high load." So we think 10 is too high, but if you go too low, then you'll be using the higher freqs a lot more than you need and it will hurt the battery. So we are leaning towards 6 (60000) for above_hispeed_delay.

    The default 5 is for "when the CPU hits X% amount of load, then jump to the hispeed_freq." Again if this one is too low then it will cause the higher freqs to be used more often then they need, so we actually turn go_hispeed_load up a little bit to 7 (70).

    The default 6 is for "how long do I wait before lowering the clock speed from what it's currently at." So the lower we put this, the better battery will be. We're still trying to decide between 3 (30000) and 5 (50000). Osm0sis is getting more lag at lower levels, and finds the best performance mark at 5. So we turn min_sample_time down a little from stock to help with the battery.

    The default 2 is for "wait this long before changing the clockspeeds from what it's at now." While technically 2 sounds better because it's changing more often, Franco believes that by setting the timer_rate to be the same thing as the CPU sample_rate (which is preset at 30000), then that will make the CPU more efficient at switching. So we increased it from 2 (20000) to 3 (30000).

    So TO RECAP: Using the stuff from above, Google's defaults for these settings are 10/5/6/2 and we are changing them to 6/7/3/3 or 6/7/5/3 (again, still testing that third number for the min_sample_time).

    Does that make sense for everyone interested in following along? Any questions? Feel free to try out these settings yourself (the easiest way is with Franco's app or something like Trickster). We want as much feedback as possible on this.
    The next kernel release will have the totally tweaked settings for everyone to test without having to change their own stuff.

    :):D:laugh::good:
    111
    Kernel Emergency Settings Reset Zip

    Alright so here's something that's been requested for awhile.

    I wrote it up recently with Franco's stamp of approval, so I'm posting it here now. Please direct people back to this post from other threads.

    This can/should be everyone's go-to when experiencing problems after updating to a new kernel build, you get stuck in a bootloop, or you think there's a conflict between ROM and kernel (or kernel tweaking app) you can't track down, since likely some leftover settings (voltages, etc.) are at fault. It's also useful if you just want to make sure you're running clean defaults. This should work on any device franco.Kernel and f.Ku support, as well as a variety of other devices, kernels and control apps: eXperience (Free/Pro), GLaDOS Control, TricksterMod, Trinity Kernel Toolbox, ROM Toolbox (Lite/Pro), SetCPU, Faux123 Kernel Enhancement Project, Performance Control and Synapse.

    Flashing this via custom recovery of your choice will delete the kernel app settings file(s), disable init.d and userinit.d scripts by moving them to a subfolder (/system/etc/init.d/off/ and /data/local/userinit.d/off/ respectively) and wipe cache and dalvik-cache for good measure.

    Hopefully it helps!


    Note: If your ROM has a Performance/Tweaking App built-in you should also manually disable any Set On Boot options it has. For example, I couldn't include clearing CM's Advanced Settings since they're built into the main com.android.settings/Settings.apk along with a lot of other Android OS settings (APNs, Developer Options, etc.), so to run clean and without conflicts on CM you need to unset them yourself.

    You should also avoid having more than one control app installed at a time. Conflicts have been reported even with everything "unset" in two apps.

    I also chose not to disable anything your ROM may have in /data/cron or /system/addon.d since some rely on it heavily, so you can choose to review those directories' contents and whether you want to leave them enabled on your own. I also left out disabling sysctl.conf since that's done by f.K and the Franco's Dev Team scripts, and also seemed a bit heavy-handed for this zip script.



    Download counts for the previous versions: 790; 2678.
    105
    Sure. Happy to post my franco.Kernel settings here for anyone who wants to try them. I'll keep these up-to-date if I change them. :)

    These are my f.K settings, my DirtyV settings can be found here.
    Device: Galaxy Nexus
    Max Frequency: 1305 MHz
    Min Frequency: 384 MHz
    Governor: interactive
    Governor Tunables: a_h_d 15000 / g_h_l 95 / h_f 729600 Hz / m_s_t 45000 / t_r 15000 / i_b_f 1036800 Hz / t_l 85 / t_s 80000 / b_d 1000000
    IO Scheduler: row
    IO Scheduler Tunables: h_r_q 100 / r_r_q 75 / h_s_q 5 / r_s_q 4 / r_w_q 4 / l_r_q 3 / l_s_q 2 / r_i 15 / r_i_f 25
    Read Ahead Buffer: 512; NR Requests: 512
    TCP Congestion Avoidance Algorithm: westwood
    Screen Off Max Frequency: 537 MHz

    Color Multipliers: 230 235 340
    RGB Gamma: -4 0 5
    Trinity Contrast: -22; OMAP4 Gamma: 4; CAB: Disabled

    Sent from franco.Kernel updater app:
    https://play.google.com/store/apps/details?id=com.franco.kernel

    The following voltages are just as a reference. DO NOT enter them; your device may not be able to handle them this low, causing freezes and reboots. They are here so that people can compare their stable voltages after following the UV guides linked below, to see how our devices compare. Every device is different, so again, DO NOT enter them.
    mpu_voltages:
    1804mhz: 1425 mV
    1728mhz: 1375 mV
    1612mhz: 1325 mV
    1536mhz: 1275 mV
    1420mhz: 1225 mV
    1305mhz: 1175 mV
    1228mhz: 1125 mV
    1036mhz: 1075 mV
    729mhz: 925 mV
    537mhz: 825 mV
    384mhz: 775 mV
    192mhz: 725 mV

    iva_voltages:
    430mhz: 1125 mV
    332mhz: 1025 mV
    266mhz: 925 mV
    133mhz: 825 mV

    core_voltages:
    512mhz: 1050 mV
    384mhz: 975 mV
    153mhz: 800 mV
    There are some battery savings from UVing but they aren't extreme. If you do not know how UVing works and/or you can't/won't read to properly build your own voltage profile, then just stick with the defaults Franco has already UVed. If you do still want to learn to UV then, once again, everyone should build their own voltage profile for their own device. Read this, this, this, and this for starters on UV methods. You don't know how your device will do until you learn and try.

    As a sidenote, if anyone else wants to share their voltages for comparison the easiest way is to copy the contents of the 3 voltage files (mpu/iva/core) in /sys/class/misc/customvoltage/ - they look (almost) exactly as the sections above.
    Fsync on just in case anything ever happens I want my data protected. That said I do disable it temporarily if I ever want a bit more performance in an app or game. SR is hardcoded disabled in the latest builds.

    The other thing worth noting is I set my color settings on boot with an init.d script. Doing it that way you can end up with a completely different result from the same values. The contrast, blacks and colors are much more natural. Also it's seamless which is nice; they take effect during the Google logo. Attached is my init.d script for that. Alter it to your own color settings, remove the .txt extension, push to /system/etc/init.d/ and set the correct perms; rwxr-xr-x (chmod -R 755 /system/etc/init.d in Terminal Emulator).

    900colorsettings can be found with my other init.d scripts (from Franco Dev Team), here: https://s.basketbuild.com/filedl/devs?dev=osm0sis&dl=osm0sis/scripts/900colorsettings

    Note: On my new replacement screen, setting the color multiplier too early in the boot (any time before the boot animation) made everything look dark, oversaturated, uncontrasted and dusky; basically the opposite of the way it used to work on my old panel. Adding sleep 2; before that line in the init.d script fixed this, so experiment if your panel is more like my new one, and the init.d as-is after you put in your settings makes it look worse.

    [ New init.d script with my new colors. I need the sleep 4 because of my problem in my "Note" above, so comment it out if you don't need it, it will make it look worse. Previous download counts were 1880, 646, and 307. Thanks everyone! ]