Database Info

[FAQ] Adding UV/OV to your OC kernel

Hi,
I did release an OC/UV kernel for the Optimus 2x device, which come with an application ( Pimp My Cpu ) to manage/change the frequency/voltage on the fly. I had a lot of questions about adding this capability to other devices/kernels, so i tought i could write some little explanation on how it work. This could also be a way to have a common way to add OC/UV so developpers can easly work for any oc/uv kernel (by respecting the two added sysfs interfaces names/capability).
First, we will assume that you sucessfully added overclocking capability to your kernel. So now to be able to "talk" with your kernel and change the frequency/voltage on the fly, we need to use cpufreq capability.
Some of the required "sysfs" interfaces are already provided by cpufreq on all (i guess) kernel. Here is a list of used cpufreq interfaces that should already be present in any default kernel :
"/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq" :
This interface will show the maximum cpu frequency available on your device. This interface is read only, and you won't be able, in any case, to set an higher cpu frequency than this one. The Optimus 2X for example report "1000000" Khz (1Ghz) by default, after hacking the kernel to enable a higher frequency it can report "1200000" (1.2Ghz).
"/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_min_freq" :
This interface will show the minimum cpu frequency available on your device. This interface is read only, and you won't be able, in any case, to set an lower cpu frequency than this one.
"/sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq" :
This interface will show/set the maxmimum possible cpu frequency.
If you read this interface ( "cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq" ) it will return the current maxmimum possible cpu frequency.
You can write a new maximum frequency to this interface (anything higher reported by "cpuinfo_max_freq" will not work) so you cpu will never go higher.
"/sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq" :
This interface will show/set the minimum possible cpu frequency.
If you read this interface ( "cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq" ) it will return the current minimum possible cpu frequency.
You can write a new minimum frequency to this interface (anything lower reported by "cpuinfo_min_freq" will not work) so you cpu will never go lower.
That's it for the standard cpufreq interfaces. Now if you want to be able to change this values "on the fly", with "Pimp My Cpu" for example, you'll have to add two new interfaces to cpufreq :
"/sys/devices/system/cpu/cpu0/cpufreq/frequency_voltage_table" :
This interface will return a table of all the availabe cpu frequency and voltages for each steps, higher to lower and is read only. This allow some utility like "Pimp My Cpu" to know each possible available frequency/voltage your kernel have.
You'll have to hack your kernel around to store this table and be able to show it a new cpufreq sysfs interface. You can take a look at my github to see how i did achieve that.
An read of this interface on the Optimus 2x will report :
FREQ STOCK_MV CURRENT_MV
Code:
1200000 1100 1100
1000000 1000 1000
900000 950 950
750000 875 875
650000 850 850
500000 800 800
300000 770 770
216000 770 770
"/sys/devices/system/cpu/cpu0/cpufreq/UV_mV_table" :
This interface enable on the fly voltage change (in mV). You can read/write to it. It will report the current millivolts modifier ( this means : "STOCK_MV" - "UV_mV_table value" ) from higher frequency to lower.
Of course, you'll have to find where the voltage change occur in your kernel and substract this value from the original value. You'll find how i did that on the Optimus 2x kernel on my github.
If you want to undervolt the 1000000 Khz (1Ghz) frequency by 50mV (based on the previous "frequency_voltage_table" you would do it like this :
Code:
echo "0 50 0 0 0 0 0 0" > /sys/devices/system/cpu/cpu0/cpufreq/UV_mV_table
That's all for now, i'll try to add some more details as possible.

Good FAQ...thankyou!

Ummmm .... okay... but how can i add more frequencies to the Kernel.... ??? is there any specific file i have to edit or a few files i have to add to the source and then directly compile or some other way ?

[How-To] OC/UV a kernel

As I cannot find any related guide while google-ing, I want to post this question on here for the purpose of learning and probably helping other people.
[Q1] How to overclock a kernel
which files are needed -> kernel source .. ?
what files need to be edited ?
is the cpufreq driver related ?
how to determine which frequencies need higher voltage and how high exactly ?
[Q2] Undervolting
which files are needed -> kernel source .. ?
what files need to be edited ?
is the cpufreq driver related ?
how to determine which frequencies can use lowered voltage and how to calculate them ?
I would be very thankful if any developer could teach me (probably others too) and I will contribute back for sure.
I am right now fixing up bugs in TeamGummy rom, and I'm nearly finished.
Only thing I lack is a custom kernel, which I want to build myself along with the rom.
*push*
Watch out guys, master guevor is going to teach us a little of his enormous knowlegde, so better keep your eyes on this one.

Overclocking
I must say that all I'm going to write, is what I've gathered looking at the code and changes made by others (thanks Blades and Roach2010) (I have not had any documentation about it), so I could make mistakes or misinterpretation.
Nor do I explain everything, because if you do not have a minimal knowledge get to play this directly may be dangerous.
(Q1) Overclocking
Files to modify
arch/arm/mach-tegra/tegra2_clocks.c (All matters relating to the frequencies.)
The system uses a series of programmed pll clocks which in turn can be used as a basis to other clocks. Let's say for this case pll_x is used for the base of processor clock.
The frequency table are in tegra_pll_x_freq_table processor. Seeing the type definition clk_pll_freq_table in clock.h we know that has entries consisting of 6 values​​. (input_rate, output_rate, n, m, p, cpcon)
These values ​​allow tegra2_pll_clk_set_rate function to set the pll frequency.
The values ​​of interest are the frequency of input and output that indicates the input frequency base to be used and what frequency we want to set (it will use the frequency corresponding to the current base and the output frequency we want). m * p values ​​give us the divisor and n is the multiplier. For simplicity we say that n/(m*p) will give us the multiplier, which applied to the base frequency will give the target frequency.
So if we want to add new frequencies, we must add entries in this structure (tegra_pll_x_freq_table) so the system knows how to program a frequency we want to use.
Add 4 entries for each frequency we want to add (one for each base frequency: 12000000,13000000,19200000,26000000 in our case), with multipliers/dividers to calculate the frequency for which we are adding.
If we add higher frequencies than at present, we have to change the maximum (max_rate) in the tegra_clk_virtual_cpu and tegra_clk_cclk structures, otherwise the system will ignore it. Also have to change the limits for our type of processor in sku_limits. For this case the cpu, cclk y pll_x in the values ​​for our processor (0x04, 0x08, 0x0F).
Finally we modify the table of frequencies used by the processor (frequency list and their order). Is defined at cpufreq_tables and contains a list of lists where we can add a new list or not to touch anything and modify an existing list.
We can look at the existing list freq_table_1p2GHz or in my source I added the list freq_table_1p7GHz.
Contains a list of frequencies that the system can select. It uses the list whose maximum corresponds to the maximum indicated above.
​
arch/arm/mach-tegra/tegra2_dvfs.c (Everything related to the voltages and its relationship with the frequencies.)
This part is much more sensitive and therefore can affect much to the duration of the battery as to stability. I have no formula (unfortunately for me) on how to modify the values, so the task has been more trial and error.
The first thing to know is that one must distinguish between the core voltage and the cpu and should be the latter inferior to the first. In tegra2_dvfs_rel_vdd_cpu_vdd_core function establishes the relationship between the two, being the default 120mV core above the cpu.
Basically what we do if we have added frequencies higher than normal, is to increase the voltage so that they are stable (depend on each processor). I recommend (at first) increment in the same extent all the values ​​involved (maximum voltage).
Therefore we can modify:
Table core_millivolts, establishing the voltages that core can use.
Table cpu_millivolts, establishing the voltages that cpu can use. It contains all the steps and should take into account the difference with the core discussed earlier. If we added any frequency, we add the corresponding voltages.
Tables core_speedo_nominal_millivolts and cpu_speedo_nominal_millivolts contain a nominal voltage of both the core and the cpu.
The maximum and nominal voltages in structures tegra2_dvfs_rail_vdd_cpu, tegra2_dvfs_rail_vdd_core and tegra2_dvfs_rail_vdd_aon.
Finally we adjust the settings for dvfs_init ("cpu", 1,) according to the list of frequencies we want to use (adding if we added new frequencies or changing them if changed change some)
With these changes we should be able to overclock, although we know that although you can configure (once started) a maximum frequency below the maximum that we put on the tables, there is a problem and many times after suspending the system will use the maximum frequency of the table. To avoid this we must make some changes in cpufreq.c
​
drivers/cpufreq/cpufreq.c
The problem occurs in the second core when reconnected after being disconnected in the suspension process. The problem is solved with the next patch.
Code:
diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
index e157205..2626e2d 100644
--- a/drivers/cpufreq/cpufreq.c
+++ b/drivers/cpufreq/cpufreq.c
@@ -1048,18 +1048,27 @@ static int cpufreq_add_dev(struct sys_device *sys_dev)
/* Set governor before ->init, so that driver could check it */
#ifdef CONFIG_HOTPLUG_CPU
+ struct cpufreq_policy *cp;
for_each_online_cpu(sibling) {
- struct cpufreq_policy *cp = per_cpu(cpufreq_cpu_data, sibling);
+ cp = per_cpu(cpufreq_cpu_data, sibling);
if (cp && cp->governor &&
- (cpumask_test_cpu(cpu, cp->related_cpus))) {
+ (cpumask_test_cpu(cpu, cp->related_cpus))) {
+ dprintk("found sibling CPU, copying policy\n");
policy->governor = cp->governor;
+ policy->min = cp->min;
+ policy->max = cp->max;
+ policy->user_policy.min = cp->user_policy.min;
+ policy->user_policy.max = cp->user_policy.max;
found = 1;
break;
}
}
#endif
if (!found)
+ {
+ dprintk("failed to find sibling CPU, falling back to defaults\n");
policy->governor = CPUFREQ_DEFAULT_GOVERNOR;
+ }
/* call driver. From then on the cpufreq must be able
* to accept all calls to ->verify and ->setpolicy for this CPU
*/
@@ -1071,6 +1080,16 @@ static int cpufreq_add_dev(struct sys_device *sys_dev)
policy->user_policy.min = policy->min;
policy->user_policy.max = policy->max;
+ if (found)
+ {
+ /* Calling the driver can overwrite policy frequencies again */
+ dprintk("Overriding policy max and min with sibling settings\n");
+ policy->min = cp->min;
+ policy->max = cp->max;
+ policy->user_policy.min = cp->user_policy.min;
+ policy->user_policy.max = cp->user_policy.max;
+ }
+
blocking_notifier_call_chain(&cpufreq_policy_notifier_list,
CPUFREQ_START, policy);

Ok, I'm on it. Trying my best, but I'm afraid you have to push me a little..
Facing heavy compile errors.
Question :
In your kernel source dir, which command do you use to start compiling ?
EDIT :
EDIT 2 :
@ all people trying to compile, do NOT use
Code:
make tegra_defconfig
!
Better grab an existing .config file from an already running kernel (thanks to guevor for this)
Just google "cyanogenmod building kernel from source" , it's described well over there.
If people request, I can share the kernel config and the precompiled blob tools once I get them from guevor.
First kernel compiled successfull, I have my zImage ready.
OC'd and Voltages changed.
First test kernel incoming the next days, I could need some testers.
Pm me if you want to test.

Thanks for this! I love your work and you sharing your vast knowledge of this subject! Thanks OP for giving him the means of enlightening us

back from the dead
edited
I was a little sleep deprived... my bad

Necromancy!

Say I want to just simply modify my cpu_max_freq via my stock kernel (no source availabel) and I have the boot.img and have unpacked it to boot.img-kernel. What should I do from here to get into the internals of the kernel to adjust the cpu_max_freq? I'm running an ARM Cortex A9, the same as the Tegra2 cpu per say.
Thanks

1chris89 said:
Say I want to just simply modify my cpu_max_freq via my stock kernel (no source availabel) and I have the boot.img and have unpacked it to boot.img-kernel. What should I do from here to get into the internals of the kernel to adjust the cpu_max_freq? I'm running an ARM Cortex A9, the same as the Tegra2 cpu per say.
Thanks
Click to expand...
Click to collapse
You can't do it that way. You can try to create oc modules but I couldn't tell you how to do that
Sent from my EVO using Tapatalk 2

Can you send me in the right direction?
Looks like unpacking the kernel isn't happening but I can create an overclock module which is FAR from easy but here's something I started on a little not sure if it's specific to the other device or whatever but I'll see how it goes.
http://forum.ppcgeeks.com/android-t...modules-tinboot-nand-boot-more-beginners.html

1chris89 said:
Say I want to just simply modify my cpu_max_freq via my stock kernel (no source availabel) and I have the boot.img and have unpacked it to boot.img-kernel.
Click to expand...
Click to collapse
Since the kernel is under GPL, get the sources from where you got the binary.
1chris89 said:
I'm running an ARM Cortex A9, the same as the Tegra2 cpu per say.
Click to expand...
Click to collapse
Is it a Tegra 2 or not? If not, this thread won't help you.

I don't understand this part:
The values ​​of interest are the frequency of input and output that indicates the input frequency base to be used and what frequency we want to set (it will use the frequency corresponding to the current base and the output frequency we want). m * p values ​​give us the divisor and n is the multiplier. For simplicity we say that n/(m*p) will give us the multiplier, which applied to the base frequency will give the target frequency.
So if we want to add new frequencies, we must add entries in this structure (tegra_pll_x_freq_table) so the system knows how to program a frequency we want to use.
Add 4 entries for each frequency we want to add (one for each base frequency: 12000000,13000000,19200000,26000000 in our case), with multipliers/dividers to calculate the frequency for which we are adding.
Click to expand...
Click to collapse
I want to add new frequencies but I don't get it.. also, why 4 entries for a single frequency? thanks!

Someone pleaseeee?

I hope there's a video tutorial for "adding overclocking features to your kernel", I hate to admit that I can't follow the instructions carefully due to lack of English knowledge

[KERNEL] 4.2.2 SaveEnergy K2_U/K2_UL

Code:
#include
/*
* I am not responsible for bricked devices, dead SD cards, thermonuclear
* war, or the current economic crisis caused by you following these
* directions. YOU are choosing to make these modificiations, and
* if you point your finger at me for messing up your device, I will
* laugh at you.
*/
FEATURES:
Based on 3.11.401.x stock kernel sources
New governors (Ondemand + MULTIPHASE, Smartmax, Intellidemand)
New I/O Sheduller (Deadline, ROW, FIOPS, SIO)
Fsync and cache optimizations
new CPU frequency table
AVS
cortex-a15 optimizations
1.674GHz CPU Overclocking and 487.5MHz GPU Overclocking (optional, separate kernel)
USB force fast charging
S2W/DT2W support
and more...
SaveEnergy K2_U/K2_UL 4.2.2 kernel ver. 1
Project stopped. GitHub
SaveEnergy K2_UL 4.2.2 kernel ver. 2
Project stopped. Link to download in the second message.
SaveEnergy K2_U 4.2.2 kernel ver. 3
Based on 3.11.401.106 stock kernel sources.
Project stopped. Link to download in the second message.
Credits:
@faux123, @flar2, @rmbq, @show-p1984, @Zarboz, @jrior001, @thicklizard - For his various kernel patches and features.
Special thanks to @rmbq, without whose invaluable assistance the process of appearance S2W we would not come to pass for a long time.
And all others mentioned on my github in contributors tab.

ver. 2
Download:
Kernel v.30 (CPU - 1.188 GHz, GPU - 450 MHz)
Kernel v.30 (CPU - 1.458 GHz, GPU - 450 MHz)
Kernel v.30 (CPU - 1.674 GHz, GPU - 487.5 MHz)
Full changelog
GitHub
ver. 3
Download:
Kernel v.20 (CPU - 1.188 GHz, GPU - 450 MHz)
Kernel v.20 (CPU - 1.458 GHz, GPU - 450 MHz)
Kernel v.20 (CPU - 1.674 GHz, GPU - 487.5 MHz)
Full changelog
GitHub

From which kernel sources is compiled?
xpirt

BrateloSlava said:
[*]...
[*]and more...
Click to expand...
Click to collapse
Does it include duoble tap to wake and sweep to wake?
Sent from my C5303 using Tapatalk

xpirt said:
From which kernel sources is compiled?
Click to expand...
Click to collapse
I'm sorry, but I'm just do not remember which source files from HTCDEV.COM I took as a basis. K2_UL - that's for sure.

J.ali.li said:
Does it include duoble tap to wake and sweep to wake?
Click to expand...
Click to collapse
No.

BrateloSlava said:
I'm sorry, but I'm just do not remember which source files I took as a basis. K2_UL - that's for sure.
Click to expand...
Click to collapse
Ok. That's enough
xpirt

Hi, I have a suggestion. Would be great if you can mod cpu frequency table and add 2xx and/or 3xx speed, there is to big gap between 162 and 432MHz. 162MHz have big Impact on performance.

Hello, I flashed your kernel but if I try to change the frequencies of the CPU as soon as I close the program back to those first, any ideas? Also i have sod

bataya said:
Hi, I have a suggestion. Would be great if you can mod cpu frequency table and add 2xx and/or 3xx speed, there is to big gap between 162 and 432MHz. 162MHz have big Impact on performance.
Click to expand...
Click to collapse
I never aimed to produce highly productive kernel. I have always striven to save energy. I specifically did overclock separate, assembly to save the source code in the github repository. To everyone could take advantage of this. I do not plan to release further overclocked kernel version.
Near710 said:
Hello, I flashed your kernel but if I try to change the frequencies of the CPU as soon as I close the program back to those first, any ideas? Also i have sod
Click to expand...
Click to collapse
I use the program TricksterMOD. It perfectly sets all the necessary parameters when starting the phone.

Link to a new version (v010) - the second message. Github will update tomorrow.

Ok, thanks a lot, any suggestions on which governor to use?

Near710 said:
Ok, thanks a lot, any suggestions on which governor to use?
Click to expand...
Click to collapse
Smartmax - for better energy saving. Ondemand - for better performance.
ADD.
If you view the my file version-2.txt - you can see that I experimented with other governor (badass, intellidemand, interactive, interactivex2, lagfree, lionheart). However, I have not noticed any particular advantages.
I also experimented with replacing mpdecision to auto_hotplug.

BrateloSlava said:
I never aimed to produce highly productive kernel. I have always striven to save energy. I specifically did overclock separate, assembly to save the source code in the github repository. To everyone could take advantage of this. I do not plan to release further overclocked kernel version.
Click to expand...
Click to collapse
You misunderstood me. I want to say that 1xx MHz is very low speed, and it is OK for low cpu task. But if you play music, or running some other task in background cpu will run on next higher speed 4xx MHz, because 1xx MHz is not enough for such task so it will use more power than if it is run on 2xx or 3xx. So if the is some cpu speed between 1xx and 4xx MHz cpu should use less power.

bataya said:
You misunderstood me. I want to say that 1xx MHz is very low speed, and it is OK for low cpu task. But if you play music, or running some other task in background cpu will run on next higher speed 4xx MHz, because 1xx MHz is not enough for such task so it will use more power than if it is run on 2xx or 3xx. So if the is some cpu speed between 1xx and 4xx MHz cpu should use less power.
Click to expand...
Click to collapse
Yeah, maybe you're right. I will try to add "a couple" of frequencies in the coming days. Although I doubt that when playing music, processor speed will be lower than 432.

bataya said:
... But if you play music, or running some other task in background cpu will run on next higher speed 4xx MHz, because 1xx MHz is not enough for such task so it will use more power than if it is run on 2xx or 3xx. So if the is some cpu speed between 1xx and 4xx MHz cpu should use less power.
Click to expand...
Click to collapse
Link to a new version (v011) - the second message.
If you are not hard - please, collect usage statistics processor frequencies when listening to music, etc.
Thanx in advance.

^Is cpu states log from BetterBatteryStates ok, or you want log from some other app?

bataya said:
^Is cpu states log from BetterBatteryStates ok, or you want log from some other app?
Click to expand...
Click to collapse
Charge the phone. Reboot. Use the usual way (listen to music, make calls, etc.). At the end of the day - to see the statistics for frequency processor. Any program, that is able to display. I'm interested to see duration of use of each frequency.

Oki, I will post cpu states at the end of the day.

BrateloSlava said:
Link to a new version (v011) - the second message.
If you are not hard - please, collect usage statistics processor frequencies when listening to music, etc.
Thanx in advance.
Click to expand...
Click to collapse
Listening to music with Spotify over WiFi CPU mostly stay on the 384MHz.
Sent from my HTC One SV using Tapatalk

Undervolt settings

Hello, I'm new on this forum then I'm sorry if i post in a bad section for my questions... But I think I should post here.
I've got several questions:
- Is there really no risk about undervolting CPU, I asked me about this because no one forum tell risk but compagny thaty produces phones don't set the minimal voltage required for each frequencies.
- Is there an app to force CPU usage at 100% to test undervolt options.
Example: I undervolt 2265MHz with -50mV, I launch app, click on 100% uage then it use at 100% my CPU, if it crash undervolt is to high. If it don't crash undervolt is good, maybe continue tu undervolt.
- Is there an automatic app to undervolt ?
Example: It decrease 1mV by 1mV with always the same frequencie of CPU and 100% CPU usage then when phone crash it restore the last value of voltage, then it do it with another frequency etc etc ...
- Which voltage do you use ? And are there voltages the same for all nexus 5 or they will change depending of each processor.
To give me an opinion about what to use
- If I found a minimum voltage, will it be the same for all kernel ?
Example: I'm on Kernel A, my min for 300MHz is 725mV, I switch to kernel B, Will it work with 725mV for 300MHz??
More generally it depend only of the processor or also of the kernel, governor, I/O scheduler ??
I Think it's all maybe other questions will come
I Thank you for all your constructive response. Don't hesitate to ask me to develop certain question and tell me mistakes .

Theres an undervolting thread. Should be linked in the "Sticky roll-up thread" which you can get to via my signature

Okay, thank you
Okay, I will copy/paste my message
Thak you for the information.

[Guide][CPU Tweak] Advance Interactive Governor Tweak for Whyred

Hi Guys, this is a continuation from initial discussion (and my promise) on InsigniuX kernel thread HERE
UPDATE 1 : After further testing, incorporate input boost to CPU 1 and 5 actually help the performance without affecting battery, as long as we set it within the optimal frequency. I also put fine tuning in hispeed_delay value, please use latest Profile File if you are using PRESET, or check the new values bellow if you do manual config
NOTE : If you are too lazy too read all the technical stuff, simply download the PRESET KERNEL ADIUTOR PROFILE HERE, and apply it to your Kernel Adiutor (choose "profile" from side bar, hit the (+) icon, and choose "import" and choose the downloaded .json file)
as we all know and aware, we always look and try to find the most optimum configuration to achieve best battery life without sacrificing performance. And while of course Qualcomm and Xiaomi have the most talented developers in their team, sometimes the implementation between both in their mass product might left a room for improvement. And in that exact room, we as community here play our role,
so, in an attempt to get a better fluidness out of this device, without sacrificing battery life, i remember one of the guide made by @soniCron i used to stumbled upon few years ago. This Guide give a very detail insight on how to optimize the interactive governor on almost any device/any kernel/any rom (as long as you have root), and thats what i gonna try to apply to our device - if you want to check the guide yourself : HERE
so i take a look into Whyred Kernel Sources to see the voltage values being used by our processor for each frequency step. And here's the voltage/speed table in case you want to see :
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
from that table, we can see which frequencies are using most power, and where is the most jump in voltage usage happen when switching between frequency.
Higher voltage jump will cost more power, means less battery life.
in conclusion, i found few frequencies which are less favorable, which is
LITTLE CPU :
1136Mhz - step by step Speed Gain is fine, but when compared to straight jump to 1401Mhz, the Speed to Power Ratio is become less favourable
1536Mhz - Smallest Speed Gain compared to Power Usage
and i also found some which might be the best/favourite frequencies :
LITTLE & BIG CPU :
900Mhz - Best contender for base speed, as it use almost same power with 633Mhz, but with plenty of speed gain
1401Mhz - Good Speed to Power Ratio
and
1747Mhz - Good Speed to Power Ratio for BIG CPU
_____________________________________________________________________________________________________________________________________________________
Now we take into account of the minimum frequency needed to ensure smooth task (if you dont know what am talking about, read the GUIDE i mention in my opening paragraph) :fingers-crossed:
For Whyred, i've found the best frequency is as following :
Idle = 633MHz (Lowest we can go at the moment)
Scrolling = 900MHz (Use Chrome browser to scroll Facebook in desktop mode)
Video = 1401MHz (Play 1080p*60fps videos in Youtube app)
App load = 1747MHz (Can use any app really)
High load = 1843MHz (Max out just in case)
Using the formula take from soniCron guide, i tried calculate optimum CPU load (this will be used as target load config) config for each frequencies
If you want to see the formulas :
Code:
We want to determine 2 values for every available clock rate: the maximal efficient load and the minimal efficient load. To make this determination, we need to bust out our calculators. (Or spreadsheets!)
For the maximal efficient load, we want to correlate a load value no higher than 90% of a given clock rate before it would be more efficient to jump to the next clock rate–to avoid overwhelming a particular rate while avoiding premature jumps to the next. For this value, we calculate it as:
(clock rate * 0.9) / next highest clock rate
For example, the maximal efficient load for 600Mhz on the Nexus 5X would be calculated as:
(600000 * 0.9) / 672000 = 80.36% (rounded and normalized: 80)
For the minimal efficient load, we want to correlate a load value at which anything higher would be better served by a higher clock rate. To calculate this:
(1 - next highest clock rate / clock rate) * -1
For example, the minimal efficient load for 600Mhz on the Nexus 5X would be calculated as:
(1 - 672000 / 600000) * -1 = 12.00% (rounded and normalized: 12)
with this config, logically speaking we want to make the Governor to favour the "best" frequencies above, and minimize the usage of unfavourable freqs.
LITTLE
Code:
633Mhz : 63
900Mhz : 71
1136Mhz : 23
1401Mhz : 82
1536Mhz : 4
1612Mhz : 83
BIG
Code:
1136Mhz : 73
1401Mhz : 9
1747Mhz : 85
1843Mhz : 90
Now that we already get the optimum number, time to apply it
Use your favorite Kernel Manager, in my case, am using Kernel Adiutor, and go to CPU Config - CPU Governor Tunables and input these value (am using Hawktail base profile from soniCron thread, as it seems it work best for most of devices, and i already do trial & error with some other value like timer rate as well ) :
(LITTLE)
Code:
align_windows:1
go_hispeed_load: 95
above_hispeed_delay: 0 633600:60000 902400:75000 1401600:100000
timer_rate: 80000
hispeed_freq: 902400
timer_slack: 200000
target_loads: 70 633600:63 92400:71 1113600:23 1401600:82 1536600:4 1612800:83
min_sample_time: 35000
boost: 0
boostpulse_duration: 0
(BIG)
Code:
align_windows:1
go_hispeed_load: 95
above_hispeed_delay: 32000 1136000:60000 1401600:75000 1747200:80000
timer_rate: 60000
hispeed_freq: 1747200
timer_slack: 100000
target_loads: 98 1113600:23 1401600:9 1747200:85 1804200:94
min_sample_time: 20000
boost: 0
boostpulse_duration: 0
Other necessary adjustment :
Boost : ON For CPU 1 at 902400 and CPU 5 at 1401600 both for 100ms
Min Big CPU Hot Plug : 0
Disable all Toggle in Thermal Section
Run Terminal and command :
Code:
su
stop perfd
or Install this MAGISK MODULE to disable Stock Thermal & Hotplug Control (Courtesy of @Mr.Nguyen0504)
Now you can test it. Do full charge and use it normally, see whether you can see the improvement or not,
be mind that this config is not really adjusted towards battery life, but rather overall fluidness/performance, yet without sacrificing too much power
Hopefully it helps you as it seems to help me (you can expect no less than 7-8Hrs SoT, am quite heavy user myself, with 2 WhatsApp account and 1 LINE account constantly active. YouTube & Gaming at least hour/day as well). Discussion is more than welcome here, as these are considered an initial calculations that still yet to furtherly fine tuned for our CPU.

Thanks mate, was waiting for this.
Initial thoughts = project butter:good:

xdarkstar said:
Thanks mate, was waiting for this.
Click to expand...
Click to collapse
please let me know the experience, i only test it with my personal preferences, so your desire with your device may vary
but i think it shouldnt be that far

otonieru said:
please let me know the experience, i only test it with my personal preferences, so your desire with your device may vary
but i think it shouldnt be that far
Click to expand...
Click to collapse
Excellent guide as well. Will test parameters over the weekend.

I will test this

mxilil said:
I will test this
Click to expand...
Click to collapse
yes please,
in the meantime, i also testing another config with input boost enabled, and some fine adjustment to hispeed_delay,
if it turn out to be better, i might create 2nd preset, along with custom control to disable BCL and Perfd. So we do not need to type it in terminal manually after reboot and let adiutor do that.

otonieru said:
Run Terminal and command :
Code:
su
stop perfd
Click to expand...
Click to collapse
I think "stop perf-hal-1-0" is the proper command.

@otonieru Great thread, very nice presentation of the matter with just the right info and setup example.
I have followed the same tutorial for my previous device but ended up using the tunables from my ROM maintainer as I never managed to calculate it properly, probably because I overlooked the voltage jumps.
Now, I wonder whether the minimum freqs of 300mhz for both clusters would help in battery life gains, taking into account the proper target loads are set for both and "timer_rate" and "min_sample_time" are tuned to make CPU ramp up quickly to avoid lags and stutters.
Logically, voltage is lower for lower frequencies, but in this case, 300mhz and 633mhz might be the same on little cluster or the voltage jump might be insignificant, but the voltage jump on big cluster might be bigger. And since the big cluster is on minimum frequency most of the time we might see some gains there.
Can you check the sources of some kernels with full range of frequencies (not the ones who have undervolt applied) and see the voltage tables?

Where is perfd located, or the file that contains the values for perfd daemon? It seems that the terminal command to disable it doesn't work, on any load except standby it still hits 1612mhz on little cluster, which is extremely annoying.
EDIT: There is no line about running state of perfd like this:
Code:
[init.svc.perfd]: [running]
There is only a line pointing at the presence of perf daemon I believe. Does this mean that perfd is not running on my system at all, so the "stop perfd" command doesn't do anything?
EDIT2: My bad people, I've set target loads for 1536 and 1612 to 4 and 8 respectively, by missing another number. Now I've set them to 98 and 100 as it was set on my previous device for two of the highest CPU steps, now the use of freqs seems much better and the device performs nice.

Cirra92 said:
Where is perfd located, or the file that contains the values for perfd daemon? It seems that the terminal command to disable it doesn't work, on any load except standby it still hits 1612mhz on little cluster, which is extremely annoying.
EDIT: There is no line about running state of perfd like this:
There is only a line pointing at the presence of perf daemon I believe. Does this mean that perfd is not running on my system at all, so the "stop perfd" command doesn't do anything?
EDIT2: My bad people, I've set target loads for 1536 and 1612 to 4 and 8 respectively, by missing another number. Now I've set them to 98 and 100 as it was set on my previous device for two of the highest CPU steps, now the use of freqs seems much better and the device performs nice.
Click to expand...
Click to collapse
Actually, i didnt include 300Mhz in my calculation because i try to check various whyred kernel source, and found that not too many of them make the 300Mhz available to beanually selected, but i do check the voltage used by 300Mhz and its saving is almost neglectible,
and as i do this tweak based on my current kernel set up (InsigniuX), i do it with 633 as base.
as for 1612 being used much,
i found that there's probably a bug in our device kernel code that make cpu sometimes lock in its highest frequency (1804 & 1612), it only happen after you restart the phone, (and it happen with most kernel out there, so its not literally the dev mistake, more like xiaomi messed up some codes)
the fix for now is by opening adiutor, ensuring that the freq not locked up. If its locked, change it back manually to 633 and 1133 (for small and big respectively), i found that it manage to stay until another reboot.
my cpu usage is as how i expected from my target load. so i think it play nicely for now.

After further testing, i update the config to incorporate boost & better tuning of hispeed_delay for each frequencies,
please check main post

otonieru said:
After further testing, i update the config to incorporate boost & better tuning of hispeed_delay for each frequencies,
please check main post
Click to expand...
Click to collapse
I am still on first charge since tweaking the kernel. So far it seems that the battery life is going to be the same, but it feels to me that everything is just a bit faster now, app opening and loading times are shorter and a bit smoother, might be a placebo, but it works nice.
One thing you should fix in main post, your target load for 1612 mhz in little cluster is set to 8, but in calculations it is 83, I suppose you missed the 3 when you wrote the tunables
I am going to update the settings you added after it falls to 20% and I recharge it.
One more thing, if I use the Magisk module for thermal and hotplug (or simply turn off the Core control in Kernel manager) does it mean that the cores will go offline if there is no load?
I remember having different hotplug solutions on SD801 devices which actually turned off 3 cores and left only one to do basic functions in deep sleep, or turn on 2 of them when there is low load on the CPU.

@Cirra92 ah right ! i might hit the backspace accidently, LoL
about hotplug, honestly am still testing it myself to figure whether the config work as intended or not, since the behaviour of this chipset it quite different. And since oreo its a bit harder to really track the cpu activities, we need to run an app to see the activities, but the app itself is giving load to cpu, thus the reading might not actually accurate.

1. Can you repost the voltage table image again? i can't see or is it missing? i really need this voltage info table, I've been playing with kernel tweaks this last few weeks.
2. On your 1st post,little cpu target load arguments are as follows : target_loads: 70 633600:63 92400:71 1113600:23 1401600:82 1536600:4 1612800:83
I don't think its the right setup. the first argument is target load which is 70, is affecting all freq starting from lowest to the highest (if there are no more arguments). but on 2nd argument you write 633600:63 (assuming 633600 is our lowest freq) then the first target load (70) has no effect. cpu target load for lowest freq (633Mhz) will be 63%, at 902Mhz the target load max is 71% and so on. And your screenshot shows the setup behavior. It idles at 633 then only spend small amount of time at 902 and 1401 and go right at the max freq for almost one and a half hours of total. If you want 633600 max target load 63% then the setup as follow :
target_loads: 63 92400:71 1113600:23 1401600:82 1536600:4 1612800:83
It means max target load from the lowest freq (633Mhz) will be 63% until below 902400. at 902400 max target load is 71% until below 1113600... and so on.
CMIIW

blackbellz said:
1. Can you repost the voltage table image again? i can't see or is it missing? i really need this voltage info table, I've been playing with kernel tweaks this last few weeks.
2. On your 1st post,little cpu target load arguments are as follows : target_loads: 70 633600:63 92400:71 1113600:23 1401600:82 1536600:4 1612800:83
I don't think its the right setup. the first argument is target load which is 70, is affecting all freq starting from lowest to the highest (if there are no more arguments). but on 2nd argument you write 633600:63 (assuming 633600 is our lowest freq) then the first target load (70) has no effect. cpu target load for lowest freq (633Mhz) will be 63%, at 902Mhz the target load max is 71% and so on. And your screenshot shows the setup behavior. It idles at 633 then only spend small amount of time at 902 and 1401 and go right at the max freq for almost one and a half hours of total. If you want 633600 max target load 63% then the setup as follow :
target_loads: 63 92400:71 1113600:23 1401600:82 1536600:4 1612800:83
It means max target load from the lowest freq (633Mhz) will be 63% until below 902400. at 902400 max target load is 71% until below 1113600... and so on.
CMIIW
Click to expand...
Click to collapse
i still can see the table, probably the connection from xda server is not good, its kinda patchy nowaday,
and btw, maybe you are right about the target loads. Although as i mentioned as well, that this config is lean more towards percormance, so its kinda intended from my side. But am still testing some other value as well, and i gonna try with your value as well.
as for now, my current set up is quite satisfy me.

crap, i cant even attach 2 picture in one post

otonieru said:
After further testing, i update the config to incorporate boost & better tuning of hispeed_delay for each frequencies,
please check main post
Click to expand...
Click to collapse
Can you add an image of the voltage table? The original image is not visible^^

otonieru said:
i still can see the table, probably the connection from xda server is not good, its kinda patchy nowaday,
and btw, maybe you are right about the target loads. Although as i mentioned as well, that this config is lean more towards percormance, so its kinda intended from my side. But am still testing some other value as well, and i gonna try with your value as well.
as for now, my current set up is quite satisfy me.
Click to expand...
Click to collapse
i tought this was towards battery but still keeping performance.

Try it or die said:
Can you add an image of the voltage table? The original image is not visible^^
Click to expand...
Click to collapse
Yup, it's not visible. But OP said its visible to him...

raptorddd said:
i tought this was towards battery but still keeping performance.
Click to expand...
Click to collapse
no, its the other way round as i wrote in my main post, LoL. But i guess 9-10Hrs SoT is more thwn enough wasnt it ?