Database Develop

Processor speed 528 or 400?

Hi all,
I notice that all the marketing data and spec sheets say the X1 has a 528MHz processor, but some people are saying it has a 400MHz. Does anyone here know what speed their device is actually running at?
-Oobly

well, 528mhz is correct.
the plugin homescreen++ showes 400mhz, but i think that this information in wrong.

The phone has a 528mhz processor, but it only runs at 400mhz - meaning that it does not run at full speed unless you tweak the settings.

Okay, so how do you tweak the settings? Does neuClockControl work to set it to 528MHz?
A processor specified to work at 528MHz running at 400MHz is still just a 400MHz processor if there is no way to make it run at 528. If this is the case then we have been royally screwed by SE/HTC.
I would like to be able to push it to 528MHz for gaming, etc (I don't mind the extra power usage if it means better performance), and change it back to 400 manually again for normal use (so it uses less battery power).
-Oobly

Oobly said:
Okay, so how do you tweak the settings? Does neuClockControl work to set it to 528MHz?
-Oobly
Click to expand...
Click to collapse
yep i use Nuecpl-ClockSpeed and the cpu work at max speed without problems....

PredatorX11 said:
yep i use Nuecpl-ClockSpeed and the cpu work at max speed without problems....
Click to expand...
Click to collapse
Cool! So does it actually show 528MHz and can you see a difference in performance?

I'm pretty sure it runs on 528mhz, not at 400. It is not slower than touch pro/hd expect graphical apps, where the video drivers count. Plus I saw the clock rate on some WinMo software as 528.something mhz, or was it 529mhz. Can't remember the software, but I couldn't care less if some old software with old DB misjudges the 7200A with 7200 processor and says that the cpu runs at 400mhz sharp. That's not reliable at all. Plus it's almost imposible to keep the cpu at 400.00mhz, just like 528.00mhz, it's usually a little bit more or less. So if it is a real measurement, you would probably see something like 401.25 or 528.84 or something Mhz. Have you done some benchmarks on SKTools for example. The cpu is somethimes on par with Samsung Omnia, which should be a Marvell PXA312 624 MHz processor. A 400mhz cpu just can not do that, even though mhz to mhz is not a proper way to compare CPU speed.
Here's a little info on what I mean: SKTools shows that the cpu speed is 400 000 000 hz, which is obviously not the real value but something predefined. Also the Integer Test on SkTools gave a value of 329.8084 units, which is on par with Touch Pro/HD and much faster than Omnia 1 with value of 253.10 on this test. The floating point is 7.552 MWIPS, again on par with the other 528mhz devices, and here it is slower than the top in list 8.63 MWIPS Omnia 624Mhz. The conclusion should be clear - either all devices are underclocked (together with X1) or X1 is running on 528Mhz cpu.

installed the software and started it up, it says my cpu is at 528mhz. where are people getting the 400mhz from?
Official R3 uk rom

TCPMP sometimes shows me 530Mhz XD

Ganondolf said:
installed the software and started it up, it says my cpu is at 528mhz. where are people getting the 400mhz from?
Official R3 uk rom
Click to expand...
Click to collapse
http://forum.xda-developers.com/showpost.php?p=4264970&postcount=3
And a few other places. Could just be reported incorrectly to some software, especially when they display exactly 400MHz. On the other hand it could be that CorePlayer or the OS somehow ups the speed to 528 when it is needed. I have heard reports of devices running faster when leaving a video open in the background. Maybe there is already some dynamic clock adjustment going on even without using neuDynamicClock?

this is what i think. the x1 is running at 538mhz but sometimes (with some app/ tasks) its been told to stay at 400mhz for some reason unknown to us (maybe to save battery or so it does not overheat).
if u look at the iphone we know for a fact that it was underclocked but the x1 nowhere has it even been said that its been underclocked apart from a few people on this forum, and i reckon that the app / software is not giving a proper / accurate reading or the app been told to run at 400mhz and people have just assumed that the cpu is underclocked. as i said this software is saying its running a 528mhz, which is the official spec for the cpu, so i cant see how people have handsets that are underclocked.

Oobly said:
Cool! So does it actually show 528MHz and can you see a difference in performance?
Click to expand...
Click to collapse
sorry no, i have not tested the x1 at default speed.
but with this software you can change the speed from same state... and a little or medium improvement is possible.
i use the 528 mhz and level 0 for the voltage. to put the chipset to 528 mhz at 1.1 volt instead of 1.325 volt
and in stand-by i put the chipset at 122 mhz and 1.1 volt

Ganondolf said:
this is what i think. the x1 is running at 538mhz but sometimes (with some app/ tasks) its been told to stay at 400mhz for some reason unknown to us (maybe to save battery or so it does not overheat).
Click to expand...
Click to collapse
I agree with that. Even the modern CPUs run at lower speeds if not utilized enough, not to mention that this should be a must for the handheld/pocket devices.

funny thing is, that I've installed nue dynamic clock and without changing anything in the past it shows that my device clock is running 528 mhz @ 1.1V. it is all soo damn confusing...

bronx said:
funny thing is, that I've installed nue dynamic clock and without changing anything in the past it shows that my device clock is running 528 mhz @ 1.1V. it is all soo damn confusing...
Click to expand...
Click to collapse
I did this as well, with nue dynamic clock showing 528mHz.
Homescreen++ still shows 400 mHz though. Only thing I can think is that Homescreen++ looks up the processor and mistakes it with a predecessor.
In other news, xperia is darn side quicker than my Kaiser, with a much more beautiful screen bonus, but maybe the speed is related to that 'drivers' story

Anyone know what registry settings Nue alters?
I want to try to overclock.

ring-bearer said:
Anyone know what registry settings Nue alters?
I want to try to overclock.
Click to expand...
Click to collapse
You cannot do that.

Calvin H said:
I did this as well, with nue dynamic clock showing 528mHz.
Homescreen++ still shows 400 mHz though. Only thing I can think is that Homescreen++ looks up the processor and mistakes it with a predecessor.
In other news, xperia is darn side quicker than my Kaiser, with a much more beautiful screen bonus, but maybe the speed is related to that 'drivers' story
Click to expand...
Click to collapse
hey! I was a kaiser user as well and x1 is much faster indeed. what more, You musn't forget that it's resolution is more than four times higher.

orelsi said:
You cannot do that.
Click to expand...
Click to collapse
I actually did overclock my fuze, to 628MHz. There is a program in the raphael section; performance something something...
I tried it right before i received my x1, and i didn't notice any improvements or problems.
I belive the same app should work for the x1, since it has same chip.

Stalix said:
I actually did overclock my fuze, to 628MHz. There is a program in the raphael section; performance something something...
I tried it right before i received my x1, and i didn't notice any improvements or problems.
I belive the same app should work for the x1, since it has same chip.
Click to expand...
Click to collapse
I'd be all over that. I searched the forum but came up empty. If you happen to remember the name, let me know. I'll run a bunch of benchmarks and post them here.

What are your overclock settings? curious.

Hi folks,
I was just wondering, what are everyone's speeds? I shall start first I have attached a screenshot of my voltages, speeds, and governor.
WARNING : I TAKE NO RESPONSIBILITY FOR ANY DAMAGE YOUR NOOK MIGHT INCUR AS A RESULT OF TRYING ANY OF OVERCLOCK CONFIGURATIONS MENTIONED HERE.​

Quick note - i would strongly suggest backing your CPU down to a max of 1200Mhz. For some reason, any speed above that will tank your GPU performance (by as much as half); the additional 100Mhz CPU loss isn't really felt (especially with the GPU gains).

Divine_Madcat said:
Quick note - i would strongly suggest backing your CPU down to a max of 1200Mhz. For some reason, any speed above that will tank your GPU performance (by as much as half); the additional 100Mhz CPU loss isn't really felt (especially with the GPU gains).
Click to expand...
Click to collapse
Is there anyway to confirm this? How can I measure GPU performance?

I set mine at 1200MHz and didn't bother playing with cpu voltage.

PoisonWolf said:
Is there anyway to confirm this? How can I measure GPU performance?
Click to expand...
Click to collapse
It has been well confirmed in severl threads (Cm7 dev and general), but you can easily test yourself with nenamark 1. Set your clocks to 1200, and run the bench; then set your clocks to 1300, run again (and watch your framerate plummet).

Divine_Madcat said:
It has been well confirmed in severl threads (Cm7 dev and general), but you can easily test yourself with nenamark 1. Set your clocks to 1200, and run the bench; then set your clocks to 1300, run again (and watch your framerate plummet).
Click to expand...
Click to collapse
I tried your little scenario of dropping down to 1125mhz and running nenamark. I was 14.8 running 1125mhz and at 15.6 at 1300mhz. I am running chainfire plug ins.
Sent with my Nook, paid the $Buck for XDA Premium—you should 2!

And what do those numbers mean?
14.8 is better or worse compared to 15.6?

Hindle1225 said:
I tried your little scenario of dropping down to 1125mhz and running nenamark. I was 14.8 running 1125mhz and at 15.6 at 1300mhz. I am running chainfire plug ins.
Sent with my Nook, paid the $Buck for XDA Premium—you should 2!
Click to expand...
Click to collapse
Odd.. i can get 21 Fps at @1200.. i would have expected the 1125 to be around the same; though the 1300 sounds right on.

I got 16.3 fps using my current setup. Lawl. What part of android uses 3d acceleration anyway?

PoisonWolf said:
I got 16.3 fps using my current setup. Lawl. What part of android uses 3d acceleration anyway?
Click to expand...
Click to collapse
UH.. android, not alot, but pretty much any game (or even some web browsers like Opera). Its easy to "lawl", but it actually DOES matter...

Divine_Madcat said:
UH.. android, not alot, but pretty much any game (or even some web browsers like Opera). Its easy to "lawl", but it actually DOES matter...
Click to expand...
Click to collapse
Good point however I don't play any games that uses 3d. I don't really play any games. Just browsing with dolphin and lots of email.
I will take it down per notch as per your suggestion once I update to a newer version though.

Also, some of the framerate difference may come from you running the 32bit framebuffer kernel, instead of the faster 16bit (but if you don't game alot, you may not really need it anyway).

Scratch that. I just tested it. I got 22.9fps. Lawl. I am now tweaking it again. What are your voltages?
Thanks for the suggestion again. 1200 MHz Max for me.
P.s. how do I check which framerate buffer am I running?

PoisonWolf said:
Good point however I don't play any games that uses 3d. I don't really play any games. Just browsing with dolphin and lots of email.
I will take it down per notch as per your suggestion once I update to a newer version though.
Click to expand...
Click to collapse
If u don't play any games, then why to bother for the 3D loss of 1300mhz?
I would prefer to use the extra speed.

RASTAVIPER said:
If u don't play any games, then why to bother for the 3D loss of 1300mhz?
I would prefer to use the extra speed.
Click to expand...
Click to collapse
Frankly, i am under the personal impression that overall the system is more responsive at 1200, and that is not just the GPU. Something about that speed is the sweet spot.
And really.. only rarely will a system that does web pages and email hit max speed anyway... why not opt for a better all around performance, rather than an immeasurable CPU gain?

I know this could be kind of off topic, but what battery temperatures do you guys get when overclocking? because the other day mine was at 36 Celsius overclocked at 1200, Does anyone knows of a safe temperature range?

Mine hovers around 34 to 36. My hd2 goes up to 43 and that is fine too. Honestly, I would think that as long as it is under 40 you should not worth.

I use 300 min and 1200 max and have the governor set to on demand. No issues with this setting.

300 - .8125
600 - 1.0
800 - 1.1875
1100 - 1.25
1200 - 1.2875
Haven't used it much recently, but about 5% drain for 24 hours in sleep. I need to put it through some heavy use to see how bad I can kill the battery.

biohazrd51 said:
1200 - .8125
1100 - 1.0
800 - 1.1875
600 - 1.25
300 - 1.2875
Haven't used it much recently, but about 5% drain for 24 hours in sleep. I need to put it through some heavy use to see how bad I can kill the battery.
Click to expand...
Click to collapse
You probably got the voltages reversed. Not possible to run 300 mhz on 1.2875. Lawl.
patruns said:
I use 300 min and 1200 max and have the governor set to on demand. No issues with this setting.
Click to expand...
Click to collapse
This was my initial setting as well (with the exception of running the max at 1300mhz). However, I realized that it wasn't the max speed that was the issue, I found out that increasing the lower-steppings was key in increasing the overall fluidity of the device as the interactive governor ramped through the steppings.

can anyone post "cpu throttling test" results for k20 pro?

app is available on playstore named "cpu throttling test", you need to keep the device idle atleast for 5-10mins before starting the test, run the test in foreground for 10 mins, dont use the phone while the test is going on. Also how are the thermals like whe playing pubgm on extreme fps on this device?

Chinese Miui 10.3.11 8/128
Big core clocked down after 2mins, then after 4:30 all cores started downclocking

This is the results I've got
http://imgur.com/a/8qMVFPZ
Don't know how accurate it is though

hyperkiller said:
Chinese Miui 10.3.11 8/128
Big core clocked down after 2mins, then after 4:30 all cores started downclocking
Click to expand...
Click to collapse
Hahaha...? With Malakas Kernel I have score around 160,000 GIPS after 15 minutes and with unthermal module I have score around 190,000 GIPS with my Pocophone F1 after the same time. So we certainly need custom Kernels for these new phones...

billybrownbear said:
This is the results I've got
http://imgur.com/a/8qMVFPZ
Don't know how accurate it is though
Click to expand...
Click to collapse
Oh very nice, even after 10 minutes barely any throttling!
Did you let it run with room temperature?

patoberli said:
Oh very nice, even after 10 minutes barely any throttling!
Did you let it run with room temperature?
Click to expand...
Click to collapse
remember to take the results with a grain of salt. judging from the reviews it's not always accurate. don't remember the exact weather conditions but i wasn't in a sauna or anything. probably was around 25°C

You haven't seen my Mi 5s which has stock Rom. It throttles by nearly 50% of max speed after ~4 minutes of use and even disables one of the two high performance cores. In other words, games start to lag after 4-5 minutes.
So I'm happy to see that this doesn't happen here

Hey guys, can someone please direct me to a resource on "unthermal module" for Pocophone? Is it a kernel that allows modification of the termal throttling values?
Thanks a lot!

I did 2 tests, one with Game Turbo On and the other Off.
It throttled when Game Turbo was Off.

RaduNastase said:
Hey guys, can someone please direct me to a resource on "unthermal module" for Pocophone? Is it a kernel that allows modification of the termal throttling values?
Thanks a lot!
Click to expand...
Click to collapse
For pocophone? You should post in poco thread instead then.

Lineage OS, can't remember if stock or immensity, probably the latter, thermal controller set to benchmark.

odssssssss said:
I did 2 tests, one with Game Turbo On and the other Off.
It throttled when Game Turbo was Off.
Click to expand...
Click to collapse
Did it again, with Game Turbo On.
Those brown shiny strips on my phone cover are copper tapes. Got them off Lazada / Shoppee for around RM10. i.e. USD 3.
I think they could help distribute heat on the back of the phone over a larger area.

odssssssss said:
Did it again, with Game Turbo On.
Those brown shiny strips on my phone cover are copper tapes. Got them off Lazada / Shoppee for around RM10. i.e. USD 3.
I think they could help distribute heat on the back of the phone over a larger area.
Click to expand...
Click to collapse
It shows throttling so how come you're saying it helps? Or did you post a wrong screenshot?

nakulp said:
It shows throttling so how come you're saying it helps? Or did you post a wrong screenshot?
Click to expand...
Click to collapse
It still throttles, but not that badly as before. You can compare my Min score with the other users.
Meanwhile could you also run the same test and post your Min Max Average scores? Would be nice if there's more data for everyone to benchmark against

odssssssss said:
It still throttles, but not that badly as before. You can compare my Min score with the other users.
Meanwhile could you also run the same test and post your Min Max Average scores? Would be nice if there's more data for everyone to benchmark against
Click to expand...
Click to collapse
Aha! I see it now! Min.181 vs much lower minimum like 165/171 etc with others..:good::laugh:
Yes, I'll upload too

Is this benchmark score normal?

I use exynos 990/6 gb version and i take this result. Is this normal

What are you using to benchmark?

mrtomaso said:
What are you using to benchmark?
Click to expand...
Click to collapse
Geekbench 5

Here's what I'm showing with snapdragon

mrtomaso said:
Here's what I'm showing with snapdragon
Click to expand...
Click to collapse
Thanks for that i guess i got a problem or something i dont know

Here's mine.
6gb/128gb SD865

Are you sure you don't have some type of battery saving enabled? Your single core is lower but your multi core is higher.

mrtomaso said:
Are you sure you don't have some type of battery saving enabled? Your single core is lower but your multi core is higher.
Click to expand...
Click to collapse
This is my multicore score. I think thats normal for exyno and i close all batery saving thing. When i write for this to samsung they said we dont care 3rd part benchmark app. So i dont know what should i do.

I've been modding pc's since the late 90's and benching has always been a thing. However, benchmarks have always been a tool that isn't foolproof. Id recommend using the phone and seeing if it chokes up at all. I've had mine now for about 6 weeks. The only issue I've had is sometimes Chrome seizes up a little, I don't blame the phone though... That's chrome.
Use it and switch around between apps. If it's seriously chokes up, then you probably have an hardware issue. Otherwise just uninstall The benchmark program and enjoy the beautiful phone.

My geekbemch score for the s20 Fe 5g SD865 is
Single : 559
Multi : 2987
And for antutu it's 557456
Is that normal ? I saw people get way higher than this

Snap 865 6/128
View attachment 5131557

Now i get this result with exynos 990 and i dont know what change

fyi, don't compare your score to sd devices as the exynos and sd are two completely different soc.
You're not supposed to get scores that resemble sd score with exynos score.
Also, every individual chip is also different referred to silicon lottery amongst pc enthusiasts and overclockers.
Without overclocking it's not gonna really matter, but still know this - no two chips are identical.
The only take away is to get reference for how much score you're supposed to get using your device and cpu, and only that's important.
If your device is severely underperforming, maybe that is issue.
But it has been long known that exynos is highly inferior to sd counterpart and why it's a hotly debated topic.
If you want sd chip, you need to buy 5G variant of the s20 FE, which afaik only is available with sd.

Themes / Apps / Mods Magisk Module for "Underclock conversions" to various chipsets. Less heat, greater battery life, more than sufficient performance.

Update4: Added a couple of new options for you guys to test out. Personally, I'm currently testing out the 778G Gaming version. Feel free to try them out to determine which options work best for your usage habits. (FYI, the 778G Daily version is going to cause perceptible delays in photos and 4k videos recording)
UPDATE3: Finally managed to test it. Phone temps are much better. Weather over at my location is around 32-25C, with AC turned at 26C, I'm getting phone battery temps of maximum 36C during use, generally stays around 34-35C.
UPDATE2: Just received my replacement 11U, now waiting for the 7 days unlock waiting period
UPDATE1: Added a translated version of the module. Untested.
Note: Please do not post or discuss about performance benchmarks. Regular use and gaming are not affected.
My 11 Ultra got it's wifi chip fried a couple of days ago. The service centre is now processing a new replacement for me. (I work in China so there's that).
In the midst of this extremely annoying debacle. I've looked at various threads and went down the rabbit hole in Chinese forums. Apparently there's a Magisk module that underclocks the CPU and GPU which results in greatly improved thermals. It was a pain in the butt to source this magisk module since Chinese users love asking for people to PM them to share the module. Long story short, took me an hour of searching to find the module.
I've looked through the contents of the magisk module and it does indeed just underclocks the CPU and GPU. Feel free to look at the files inside the zip.
Basically this module does the following
Underclocks X1 core from 2.84ghz to 2.49ghz
Reduces performance by 13%, reduces power consumption by 32%
Underclocks A78 cores from 2.42ghz to 2.23ghz
Reduces performance by 5%, reduces power consumption by 14%
Underclocks GPU from 840mhz to 608mhz
Reduces performance by 27%, reduces power consumption by 54%
This module will result in the SD888 performing around the same level (if not better than) of the SD870 whilst being more efficient than the SD870. User reports the battery temp during regular use as reported by Miui will sit in the range of 30-34C. During gaming it will be around 42C after 1-2 hours of MOBA gaming
Once I receive my replacement device, I'll give this a try and update OP

can't believe this works, feels much smoother now.

GuyWithRootedPhone said:
can't believe this works, feels much smoother now.
Click to expand...
Click to collapse
You can feel the results immediately? Probably due to the processor not throttling as much. Do feedback on the heat generation reduction

Can I Flash it in Recovery without having installed Magisk ?

veimus said:
Can I Flash it in Recovery without having installed Magisk ?
Click to expand...
Click to collapse
No? This is a magisk module.

What are those 2 attachments?

xNAPx said:
What are those 2 attachments?
Click to expand...
Click to collapse
Magisk modules. One is the original by the author which is in Chinese. The other is translated by me.

lambstone said:
Magisk modules. One is the original by the author which is in Chinese. The other is translated by me.
Click to expand...
Click to collapse
Did you tested it? I installed few minutes ago, in average usage I don't see any drop in performance, I hope I will find a greater battery life

xNAPx said:
Did you tested it? I installed few minutes ago, in average usage I don't see any drop in performance, I hope I will find a greater battery life
Click to expand...
Click to collapse
Not trying to be an ass over here but OP is extremely detailed. And no, I've yet to try it since I just got my replacement unit.

lambstone said:
Not trying to be an ass over here but OP is extremely detailed. And no, I've yet to try it since I just got my replacement unit.
Click to expand...
Click to collapse
Pardon? OP?

xNAPx said:
Pardon? OP?
Click to expand...
Click to collapse
Opening post. I wrote down and extremely detailed description.

lambstone said:
Opening post. I wrote down and extremely detailed description.
Click to expand...
Click to collapse
My statement has nothing to do with your opening post to be honest apart from asking you if you had a chance to try it.

xNAPx said:
My statement has nothing to do with your opening post to be honest apart from asking you if you had a chance to try it.
Click to expand...
Click to collapse
It does since you're asking a question to which the answer can be found in the opening post.

lambstone said:
It does since you're asking a question to which the answer can be found in the opening post.
Click to expand...
Click to collapse
I missed the part in which you say you are waiting for the unlock. By the way you could be more kind, thanks anyway

xNAPx said:
I missed the part in which you say you are waiting for the unlock. By the way you could be more kind, thanks anyway
Click to expand...
Click to collapse
Pretty sure I answered your question even though you couldn't be bothered to read through the post.

You did, that's why I thanked you

From my personal testing so far I can tell that battery life is improved, can't say the same for temps thought, probably a couple of degrees cooler average. 'llI reserve more time for testing the battery life as I need more battery cicles to really tell the difference if there is any. Is there any way to check the actual CPU frequency via terminal emulator?

xNAPx said:
From my personal testing so far I can tell that battery life is improved, can't say the same for temps thought, probably a couple of degrees cooler average. 'llI reserve more time for testing the battery life as I need more battery cicles to really tell the difference if there is any. Is there any way to check the actual CPU frequency via terminal emulator?
Click to expand...
Click to collapse
From what I've read in the discussion posts by the original chinese developer. They use something called perfmon. I believe you'll require both the app and the magisk module.

lambstone said:
Basically this module does the following
Underclocks X1 core from 2.84ghz to 2.49ghz
Reduces performance by 13%, reduces power consumption by 32%
Underclocks A78 cores from 2.42ghz to 2.23ghz
Reduces performance by 5%, reduces power consumption by 14%
Underclocks GPU from 840mhz to 608mhz
Reduces performance by 27%, reduces power consumption by 54%
Click to expand...
Click to collapse
Code:
sleep 30
chmod 644 /sys/devices/system/cpu/cpu4/cpufreq/scaling_max_freq
echo 2227200 > /sys/devices/system/cpu/cpu4/cpufreq/scaling_max_freq
chmod 644 /sys/devices/system/cpu/cpu7/cpufreq/scaling_max_freq
echo 2496000 > /sys/devices/system/cpu/cpu7/cpufreq/scaling_max_freq
chmod 644 /sys/class/kgsl/kgsl-3d0/max_pwrlevel
echo 4 > /sys/class/kgsl/kgsl-3d0/max_pwrlevel
exit 0
The code cap the maximum frequence for CPU 4&7 and the power for GPU.
Mostly what a kernel auditor can do.
The other way around for reference:
Xiaomi Mi 10T with Snapdragon 865 beats ROG phone 5 with GPU Overclock
Hi I have Xiaomi Mi 10T with Snapdragon 865 with Adreno 650 GPU and 8 GB LPDDR5 RAM and I used Konabess app to Overclock GPU to beat ROG Phone 5 in Vulkan performance. If you need anything help here is my new group on Telegram for support in...
forum.xda-developers.com

pl1992aw said:
Code:
sleep 30
chmod 644 /sys/devices/system/cpu/cpu4/cpufreq/scaling_max_freq
echo 2227200 > /sys/devices/system/cpu/cpu4/cpufreq/scaling_max_freq
chmod 644 /sys/devices/system/cpu/cpu7/cpufreq/scaling_max_freq
echo 2496000 > /sys/devices/system/cpu/cpu7/cpufreq/scaling_max_freq
chmod 644 /sys/class/kgsl/kgsl-3d0/max_pwrlevel
echo 4 > /sys/class/kgsl/kgsl-3d0/max_pwrlevel
exit 0
The code cap the maximum frequence for CPU 4&7 and the power for GPU.
Mostly what a kernel auditor can do.
The other way around for reference:
Xiaomi Mi 10T with Snapdragon 865 beats ROG phone 5 with GPU Overclock
Hi I have Xiaomi Mi 10T with Snapdragon 865 with Adreno 650 GPU and 8 GB LPDDR5 RAM and I used Konabess app to Overclock GPU to beat ROG Phone 5 in Vulkan performance. If you need anything help here is my new group on Telegram for support in...
forum.xda-developers.com
Click to expand...
Click to collapse
Pretty much yes. However kernel auditor set them post boot. This magisk module sets them on boot. Or you could always implement a init.d solution.
The magic of a magisk module is it's ease of use and a fire and forget solution