How about this over clocking going on here? So I finally rooted, got a ROM, flashed a kernel.. Then bought SETCPU and turned it up to 1.9ghz. This seemed to work right well, but then I discovered a tool from chainfire and found that it wasn't sticking. So I was told about and edited this thermald.conf file.
But I wanna know, what temp was it that caused this to occur, and what I have just changed?
I've overclocked 3 phones now (8125/fuze/inspire) but none were this nice (dual core for example.) and this one gets warm. But I cannot tell how warm? And now I've even changed something and don't even know what. And it was related to warm.
Tl;dr how hard is are you Pushin yer phone?
The change:
sampling 5000
[pa_therm0]
sampling 5000
thresholds 70 80 90
thresholds_clr 65 75 85
actions none none none
action_info 0 0 0
[tsens_tz_sensor0]
sampling 1000
thresholds 50 65 75 80 99 102 105
thresholds_clr 40 62 72 77 96 99 102
actions cpu cpu cpu cpu cpu+lcd+flashlight+battery+vibrator+gpu cpu+lcd+flashlight+battery+vibrator+gpu cpu+lcd+flashlight+battery+vibrator+gpu+shutdown
action_info 1809000 1242000 1134000 918000 594000+255+0+0+0+200000000 384000+30+1+3+1+27000000 384000+30+1+3+1+27000000+5000
[tsens_tz_sensor1]
sampling 1000
thresholds 75
thresholds_clr 72
actions none
action_info 0
[tsens_tz_sensor2]
sampling 1000
thresholds 100
thresholds_clr 80
actions none
action_info 0
[tsens_tz_sensor3]
sampling 1000
thresholds 75 78 81 84 87 90
thresholds_clr 72 75 78 81 84 87
actions cpu cpu cpu cpu+lcd+flashlight+battery+vibrator+gpu cpu+lcd+flashlight+battery+vibrator+gpu cpu+lcd+flashlight+battery+vibrator+gpu+shutdown
action_info 1242000 1134000 918000 594000+255+0+0+0+200000000 384000+30+1+3+1+27000000 384000+30+1+3+1+27000000+5000
[tsens_tz_sensor4]
sampling 1000
thresholds 50 65 80 86
thresholds_clr 40 62 77 84
actions gpu gpu gpu gpu
action_info 400000000 300000000 200000000 27000000
Kom op nu
Is there an app that allows you to monitor the GPU? Im currently playing with the clockspeeds, voltages and thresholds and would like to be able to monitor the results.
Can only set 4 steps with the kernal I have the following settings at the moment.
160 - 837500
350 - 912500
440 - 962500
533 - 1025000
1 up - 90
2 down - 41
2 up - 90
3 down - 70
3 up - 90
4 down - 73
Can only select frequencies of 160, 266, 350, 440 and 533, which is why i've selected the ones I have. Im hoping it will sit at 160, 350 the majority of the time unless im gaming and have set the thresholds so that the over lap to hopefully prevent unessesary jumping between frequencies. Trying to balance performance and power consumption.
If I find it works well, then I'll then look to undervolt.
Ivanichs CM12 3.4 Kernel Overclocked by Little Lerroyy [IN-TESTING]
Disclaimer
Each kernel release, I will be pre testing on my device first, to ensure adequate stability. However by the use of this kernel, I claim no responsibility to any damaged devices due to the risky nature of overclocking. I recommend that you DO NOT USE THIS KERNEL if you want rock solid stability!
Credits
I won't be taking full credit for this, as a matter of fact, i will only take very little as this kernel is based 100% from Ivanichs CM12 kernel for the sensation. While I only simply transplanted the overclocking code from Sultanxda and SebatianFM kernels to Ivanichs with my touch of voltage tweaks.
Features
In kernel thermal throttling (MSM_THERMAL). Please thank javelinanddart for this amazing fix!
CPU OC to 1890MHz
GPU 3D OC to 320MHz
GPU 2D OC to 228MHz
BUS/Memory OC for overclocked frequencies from 310MHz to 400MHz
CPU Voltage interface. Select your own VCore!
Various idle underclocks and undervolts to improve batterylife
Included sultanxda's panel undervolt interface
Various I/O Scheduler & I/O read aheads and CPU Govenor tweaks for awesome snappiness.
Compiled with Linaro GCC Compiler optimized for Cortex A8.
Bugs (Major)
Device will reboot when entering airplane mode before screen is switched off and then back on again.
Thermal Daemon will set absolute top frequency 1.89GHz when returning from a CPU throttle due to high temperature. In this case some devices can't handle 1.89 and will either reboot eventually or instantly reboot. Hinxnz has provided and excellet workaround for this. - fixed by javelinanddart!
Notes
I recommend that you DO NOT USE THIS KERNEL if you want rock solid stability! Kernel is still currently in testing.
I would recommend only use 1.89GHz and 1.944GHz for benchmarking. I have set the voltages for those frequencies to 1.3v and a monstrous 1.375v respectivly to maintain device from instantly crashing.
My Device only stable on 1836MHz with a moderate 1.25v on the core (This is device depandant)
Source
Kernel Source (GitHub)
Download Link
Google Drive
XDA:DevDB Information
CM12 Overclocked 3.4 Kernel for HTC Sensation, Kernel for the HTC Sensation
Contributors
little_lerroyy
Source Code: https://github.com/littlelerroyy/android_kernel_htc_pyramid
Kernel Special Features: OC Up To: CPU = 1890MHz, GPU3D = 320MHz, GPU2D = 228MHz, L2 Cache = 1.6GHz
Version Information
Status: Testing
Created 2015-03-07
Last Updated 2015-08-12
Changelog
1/8/15
In kernel thermal throttling!!! No more thermal daemon issues! Please give your most thanks to javelinanddart for his amazing fix!
ramdisk file updated to work with the latest versions of CM12.1
Also up-to-date with ivanichs changes.
23/4/15
Increased CPU input boost duration to 300mhz to smooth out UI, and reduced the possibility of intelliactive ramping CPU to max frequency unnecessarily.
Tweaked intelliactive govenor to be less aggressive. It was going to full speed too often and ramping while idling. Phone runs much cooler due to this change and the above
Reduced BUS speed for upper standard CPU frequencies.
Underclocked idle frequency bus from 103MHz to 80Mhz. We dont need performance at idle! We could go lower.
Undervolted L2 cache for idle frequencies. 1.1V -> 1.075V. We could go lower.
Sultanxda panel undervolting sysfs interface added. Navigate to "/sys/module/board_pyramid_panel/parameters/panel_uv", where "panel_uv" is a value between 0 and 700mV to undervolt the panel by (stock panel voltage is 3100mV). Increments of 25mV only accepted
Very basic flashable ZIP made.
19/4/15
Reverted to old CPU Vcore table. The old Undervolted table didn't suit some devices.
Added CPU VCore interface. You can now change CPU voltages within android. Range: (1.35V - 0.7V)
Reverted to stock GPU2D BUS. Noticed no difference with it raised.
Increase Max current intake to 1250mA. Charging should be quicker with >1000mA chargers. (im not sure if I will keep this as i do notice a hotter phone while charging)
Dropped all partitions to 0KB readhead. People should notice a even more snappier phone.
6/4/15
Tested with CM12.1
Overclocked system bus from 310MHz to 400MHz on overclocked frequencies.
Undervolted vcore for standard frequencies. In theroy this should provide better battery life.
GPU Bus overclock
CPU will not switch off the second core while the screen is on (it will turn off when screen is off). I find this helps improve performance drastically without any noticble drop in battery life (i feel it could improve battery life)
ZRAM switched off by default. People can enable this at anytime if they want to.
Set CPU govenor to intelliactive by default. Also tuned for extra smoothness & snappiness.
Set I/O Scheduler to FIOPS as default. Optimised tunable scales to improve performance.
Dropped read ahead levels significantly for internal and external storage. I find this improves loading from storage by quite a bit.
8/3/15
Removed 1.944GHz selection to stop random crashing. I will try later to see if i can stabilize it.
Published boot.img with L2 Cache OC to 1.6GHz
7/3/15
Initial CPU OC to 1944MHz, GPU3D 320MHz, GPU2D 228MHz
Added Ivanichs ZRAM support
Hi . thanks for sharing your kernel ! but would you please re-compile it with the latest code ?! i checked your github and it doesn't contain ivanich's commits from 5th and 6th of March! because 6th of march commits enabled zRam and it must have great impact on performance ! Thanks.
francesco totti said:
Hi . thanks for sharing your kernel ! but would you please re-compile it with the latest code ?! i checked your github and it doesn't contain ivanich's commits from 5th and 6th of March! because 6th of march commits enabled zRam and it must have great impact on performance ! Thanks.
Click to expand...
Click to collapse
Thanks for the input, about a few hours ago I downloaded ivanichs changes onto my PC but havent pushed the updates yet, I will post the zRam kernel tonight
Updated to include zRAM Support
well i flashed your kernel but i always get reboots while doing heavy tasks and some times just normal tasks!! here's the last kmsg :
francesco totti said:
well i flashed your kernel but i always get reboots while doing heavy tasks and some times just normal tasks!! here's the last kmsg :
Click to expand...
Click to collapse
Thanks for sharing this!! May i ask, what CPU speed did you select and what GPU3D Freq did you select? Well just as i thought the last message in your log that it tripped over the 65deg threshold. I tested this last night and the thermal daemon isnt throttling cpu speed! I even tried editing the thermald.conf to drop to 1134MHz when over just a small 45 degrees. and it still powers on at full frequency. Im not sure if there is an issue with the thermal daemon itself, or if it has a problem with my kernel, i will be doing some more testing tonight.
The minimum cpu speed was 540 and maximum cpu speed was 1836 ! And also i hadn't overclocked the gpu !
little_lerroyy said:
Thanks for sharing this!! May i ask, what CPU speed did you select and what GPU3D Freq did you select? Well just as i thought the last message in your log that it tripped over the 65deg threshold. I tested this last night and the thermal daemon isnt throttling cpu speed! I even tried editing the thermald.conf to drop to 1134MHz when over just a small 45 degrees. and it still powers on at full frequency. Im not sure if there is an issue with the thermal daemon itself, or if it has a problem with my kernel, i will be doing some more testing tonight.
Click to expand...
Click to collapse
Haven't tried your kernel yet but will later, cheers.
Not all devices can handle over 1.7GHz and some also can't handle 320 for GPU.
Stock frequency for our CPU according to Qualcomm is 1.67GHz.
I don't think thermald is working either on ivanich's kernel, actually I think it maybe but temperature readings are incorrect when it gets hot causing thermald not to kick in. I'm not sure of this but this was my theory.
Once my phone got quite hot and I noticed that the CPU temp reading was to low for what it seemed like, when at normal temp it seems to be correct.
A dirty work around could be to lower the thresholds, even just to diagnose.
My phone doesn't normally get hot at all, so hence why I haven't reported it in ivanich's thread, this is also the first time I've seen someone else arise this too.
francesco totti said:
The minimum cpu speed was 540 and maximum cpu speed was 1836 ! And also i hadn't overclocked the gpu !
Click to expand...
Click to collapse
Okay cool, ive done some testing and i think i can mainly solve the issue by removing the 1.944GHz entry. For some reason i think it may actually call that frequency even if its not selected. Im testing it now without that entry and the phone isnt getting as hot, and havent had a crash yet, normally i do have a crash under high usage. Im going to try one last attempt to fix 1.944GHz, otherwise i will remove it and leave it for another time (if its even worth it).
hinxnz said:
Haven't tried your kernel yet but will later, cheers.
Not all devices can handle over 1.7GHz and some also can't handle 320 for GPU.
Stock frequency for our CPU according to Qualcomm is 1.67GHz.
I don't think thermald is working either on ivanich's kernel, actually I think it maybe but temperature readings are incorrect when it gets hot causing thermald not to kick in. I'm not sure of this but this was my theory.
Once my phone got quite hot and I noticed that the CPU temp reading was to low for what it seemed like, when at normal temp it seems to be correct.
A dirty work around could be to lower the thresholds, even just to diagnose.
My phone doesn't normally get hot at all, so hence why I haven't reported it in ivanich's thread, this is also the first time I've seen someone else arise this too.
Click to expand...
Click to collapse
Yeah i think there is definitely an issue with the thermal daemon, I have tried setting the first trip temp point to 40c. which the phone operates over most of the time anyway, and i have no throttling of anykind. The system keeps powering on its top frequency. In sultans rom, the thermal daemon definitely worked when overclocking. maybe i might speak to him for some answers.
Running a game for a 5 mins, but can't get temp above 65
Code:
D/ThermalDaemon( 1412): CPU[1] frequency limited to 1134000
D/ThermalDaemon( 1412): CPU[0] frequency limited to 1512000
D/ThermalDaemon( 1412): CPU[1] frequency limited to 1512000
D/ThermalDaemon( 1412): CPU[0] frequency limited to 1134000
D/ThermalDaemon( 1412): CPU[1] frequency limited to 1134000
D/ThermalDaemon( 1412): CPU[0] frequency limited to 1512000
D/ThermalDaemon( 1412): CPU[1] frequency limited to 1512000
ivanich said:
Running a game for a 5 mins, but can't get temp above 65
Code:
D/ThermalDaemon( 1412): CPU[1] frequency limited to 1134000
D/ThermalDaemon( 1412): CPU[0] frequency limited to 1512000
D/ThermalDaemon( 1412): CPU[1] frequency limited to 1512000
D/ThermalDaemon( 1412): CPU[0] frequency limited to 1134000
D/ThermalDaemon( 1412): CPU[1] frequency limited to 1134000
D/ThermalDaemon( 1412): CPU[0] frequency limited to 1512000
D/ThermalDaemon( 1412): CPU[1] frequency limited to 1512000
Click to expand...
Click to collapse
Hmm, i see. Where is the best way to alter the settings of the daemon? Ive tried thermald.conf in /etc/ folder and also /sys/devices/virtual/thermal and it wont kick in to any value i choose.
Ive been testing the kernel without 1.944GHz entry and havent had a crash at all, even with the phone raring hot. I will upload this soon
little_lerroyy said:
Hmm, i see. Where is the best way to alter the settings of the daemon? Ive tried thermald.conf in /etc/ folder and also /sys/devices/virtual/thermal and it wont kick in to any value i choose.
Ive been testing the kernel without 1.944GHz entry and havent had a crash at all, even with the phone raring hot. I will upload this soon
Click to expand...
Click to collapse
I believe that /system/etc/thermald.conf should be used for this. Lower thresholds and thresholds_clr should do the trick.
ivanich said:
Running a game for a 5 mins, but can't get temp above 65
Code:
D/ThermalDaemon( 1412): CPU[1] frequency limited to 1134000
D/ThermalDaemon( 1412): CPU[0] frequency limited to 1512000
D/ThermalDaemon( 1412): CPU[1] frequency limited to 1512000
D/ThermalDaemon( 1412): CPU[0] frequency limited to 1134000
D/ThermalDaemon( 1412): CPU[1] frequency limited to 1134000
D/ThermalDaemon( 1412): CPU[0] frequency limited to 1512000
D/ThermalDaemon( 1412): CPU[1] frequency limited to 1512000
Click to expand...
Click to collapse
I can't seem to bring this up in logcat, I mean it's not in there. So I guess it's not working for me.
Edit: It working, just wasn't up to temp.
This is with your kernel. Do you have the same sensor? pm8058_tz
Code:
D/ThermalDaemon( 1603): No rules to monitor sensor 'pm8058_tz' disabled: 1 num_thresholds: 0
All I get in logcat is:
Code:
D/ThermalDaemon( 1603): Thermal daemon started
D/ThermalDaemon( 1603): Threshold 0 Action[0] : cpu
D/ThermalDaemon( 1603): Threshold 1 Action[0] : cpu
D/ThermalDaemon( 1603): Threshold 2 Action[0] : cpu
D/ThermalDaemon( 1603): Threshold 3 Action[0] : cpu
D/ThermalDaemon( 1603): Threshold 0 Action Info[0] : 1134000.000000
D/ThermalDaemon( 1603): Threshold 1 Action Info[0] : 972000.000000
D/ThermalDaemon( 1603): Threshold 2 Action Info[0] : 648000.000000
D/ThermalDaemon( 1603): Threshold 3 Action Info[0] : 384000.000000
D/ThermalDaemon( 1603): Sensor '(null)' config: sampling 500000
D/ThermalDaemon( 1603): Sensor 'tsens_tz_sensor0' config: sampling 3000000
D/ThermalDaemon( 1603): 66560 61440 cpu 1134000
D/ThermalDaemon( 1603): 76800 71680 cpu 972000
D/ThermalDaemon( 1603): 87040 81920 cpu 648000
D/ThermalDaemon( 1603): 88064 83968 cpu 384000
D/ThermalDaemon( 1603): Number of cpus :2
D/ThermalDaemon( 1603): Sensor setup:[tsens_tz_sensor0]
D/ThermalDaemon( 1603): No rules to monitor sensor 'pm8058_tz' disabled: 1 num_thresholds: 0
hinxnz said:
I can't seem to bring this up in logcat, I mean it's not in there. So I guess it's not working for me.
This is with your kernel. Do you have the same sensor? pm8058_tz
Code:
D/ThermalDaemon( 1603): No rules to monitor sensor 'pm8058_tz' disabled: 1 num_thresholds: 0
All I get in logcat is:
Code:
D/ThermalDaemon( 1603): Thermal daemon started
D/ThermalDaemon( 1603): Threshold 0 Action[0] : cpu
D/ThermalDaemon( 1603): Threshold 1 Action[0] : cpu
D/ThermalDaemon( 1603): Threshold 2 Action[0] : cpu
D/ThermalDaemon( 1603): Threshold 3 Action[0] : cpu
D/ThermalDaemon( 1603): Threshold 0 Action Info[0] : 1134000.000000
D/ThermalDaemon( 1603): Threshold 1 Action Info[0] : 972000.000000
D/ThermalDaemon( 1603): Threshold 2 Action Info[0] : 648000.000000
D/ThermalDaemon( 1603): Threshold 3 Action Info[0] : 384000.000000
D/ThermalDaemon( 1603): Sensor '(null)' config: sampling 500000
D/ThermalDaemon( 1603): Sensor 'tsens_tz_sensor0' config: sampling 3000000
D/ThermalDaemon( 1603): 66560 61440 cpu 1134000
D/ThermalDaemon( 1603): 76800 71680 cpu 972000
D/ThermalDaemon( 1603): 87040 81920 cpu 648000
D/ThermalDaemon( 1603): 88064 83968 cpu 384000
D/ThermalDaemon( 1603): Number of cpus :2
D/ThermalDaemon( 1603): Sensor setup:[tsens_tz_sensor0]
D/ThermalDaemon( 1603): No rules to monitor sensor 'pm8058_tz' disabled: 1 num_thresholds: 0
Click to expand...
Click to collapse
I have the same output
Code:
[email protected]:/ $ logcat|grep ThermalDaemon
D/ThermalDaemon( 1486): Thermal daemon started
D/ThermalDaemon( 1486): Threshold 0 Action[0] : cpu
D/ThermalDaemon( 1486): Threshold 1 Action[0] : cpu
D/ThermalDaemon( 1486): Threshold 2 Action[0] : cpu
D/ThermalDaemon( 1486): Threshold 3 Action[0] : cpu
D/ThermalDaemon( 1486): Threshold 0 Action Info[0] : 1134000.000000
D/ThermalDaemon( 1486): Threshold 1 Action Info[0] : 972000.000000
D/ThermalDaemon( 1486): Threshold 2 Action Info[0] : 648000.000000
D/ThermalDaemon( 1486): Threshold 3 Action Info[0] : 384000.000000
D/ThermalDaemon( 1486): Sensor '(null)' config: sampling 500000
D/ThermalDaemon( 1486): Sensor 'tsens_tz_sensor0' config: sampling 3000000
D/ThermalDaemon( 1486): 66560 61440 cpu 1134000
D/ThermalDaemon( 1486): 76800 71680 cpu 972000
D/ThermalDaemon( 1486): 87040 81920 cpu 648000
D/ThermalDaemon( 1486): 88064 83968 cpu 384000
D/ThermalDaemon( 1486): Number of cpus :2
D/ThermalDaemon( 1486): Sensor setup:[tsens_tz_sensor0]
D/ThermalDaemon( 1486): No rules to monitor sensor 'pm8058_tz' disabled: 1 num_thresholds: 0
pm8058_tz hardcoded into the thermald binary
Removed 1.944GHz entry. Kernel is much more stable now.
little_lerroyy said:
Removed 1.944GHz entry. Kernel is much more stable now.
Click to expand...
Click to collapse
Your kernel doesn't have the modified fstab and init.pyramid.rc for zram.
By the way are you going to add 2D overclock?
hinxnz said:
Your kernel doesn't have the modified fstab and init.pyramid.rc for zram.
By the way are you going to add 2D overclock?
Click to expand...
Click to collapse
He added the zram commits from ivanich git i think !! How is it possible that it doesn't have zram ?! How do you check whether zram is enabled or not ?!
francesco totti said:
He added the zram commits from ivanich git i think !! How is it possible that it doesn't have zram ?! How do you check whether zram is enabled or not ?!
Click to expand...
Click to collapse
Enter in Terminal emulator:
Code:
cat /proc/swaps
Ok I just tested thermal daemon, I set cpu to 1728000 and thermald working but it causing reboots because my device can't handle 1890000 which it resets to.
Code:
D/ThermalDaemon( 1596): CPU[0] frequency limited to 1134000
D/ThermalDaemon( 1596): CPU[1] frequency limited to 1134000
D/ThermalDaemon( 1596): CPU[0] frequency limited to 1890000
D/ThermalDaemon( 1596): CPU[1] frequency limited to 1890000
Edit: I don't know what to change yet or add but something definitely in /drivers/thermal/msm_thermal.c
Edit: I tested on ivanich's kernel too and it's the same, meaning if you set max freq to 1188000 it reverts back to 1512000 afterwards.
Also Sultanxda fixed this in his kernel.
Some people don't like how warm the phonne gets. This can be tweaked with the thermal-engine.conf located in system/etc (requires root). Just edit the file with a text editor.
Note: The thermal-engine might have a different name or there might be multiple engines, just look for these entries. Since I don't own the phone, I can't verify this. Since it uses the same SOC (Snapdragon 808) as the LG G4 and Nexus 5X, it should also work similarly.
This is the throttle temperature for the big cluster
Code:
[SS-SKIN-XO-THERM-PERF]
set_point [COLOR="RoyalBlue"][B]43[/B][/COLOR]000
set_point_clr [COLOR="royalblue"][B]37[/B][/COLOR]000
Hotplug CPU4
Code:
[CPU4_HOTPLUG_MONITOR]
thresholds [COLOR="RoyalBlue"][B]42[/B][/COLOR]000
thresholds_clr [B][COLOR="royalblue"]37[/COLOR][/B]000
CPU5
Code:
[CPU5_HOTPLUG_MONITOR]
thresholds [COLOR="royalblue"][B]45[/B][/COLOR]000
thresholds_clr [B][COLOR="royalblue"]40[/COLOR][/B]000
Little cluster
Code:
[SS-SKIN-XO-THERM-POWER]
set_point [COLOR="royalblue"][B]48[/B][/COLOR]000
set_point_clr [COLOR="royalblue"][B]40[/B][/COLOR]000
GPU
Code:
[GPU_management]
thresholds [COLOR="royalblue"][B]40[/B][/COLOR]000 [COLOR="royalblue"][B]43[/B][/COLOR]000 [COLOR="royalblue"][B]47[/B][/COLOR]000 [COLOR="royalblue"][B]50[/B][/COLOR]000
thresholds_clr [COLOR="royalblue"][B]37[/B][/COLOR]000 [COLOR="royalblue"][B]40[/B][/COLOR]000 [COLOR="royalblue"][B]43[/B][/COLOR]000 [COLOR="royalblue"][B]47[/B][/COLOR]000
battery, this reduces the charging current at the specified temperature
Code:
[MONITOR_QUIET_THERM_BATTERY]
thresholds [COLOR="royalblue"][B]38[/B][/COLOR]000 [COLOR="royalblue"][B]40[/B][/COLOR]000 [COLOR="royalblue"][B]42[/B][/COLOR]000 [COLOR="royalblue"][B]44[/B][/COLOR]000
thresholds_clr [COLOR="royalblue"][B]35[/B][/COLOR]000 [COLOR="royalblue"][B]38[/B][/COLOR]000 [COLOR="royalblue"][B]40[/B][/COLOR]000 [COLOR="royalblue"][B]42[/B][/COLOR]000
All other values I would leave at standard. I'm actually not quite sure what they do. Some influence sensor polling rate others something unknown.
If you think your phone gets too hot you can lower all the blue values by 2-4°C. This should be a good compromise between heat and performance. Once you are done, save the file and restart.
so insted of having 43000 in some values we set it at 38000 for instance? Did you do it and does it work?
geokor said:
so insted of having 43000 in some values we set it at 38000 for instance? Did you do it and does it work?
Click to expand...
Click to collapse
Sorry for the late reply, but I don't own the phone so I can't verify this. It works the same way on the LG G4 and Nexus 5X, though.
Hi. I'm trying to optimize some computer vision code, but variable CPU core frequencies is making it extremely hard to quantify any improvements I make. Basically I want to be able to force all 4 cores to 2.26GHz so that my test environment can be consistent. I'm currently running rooted Lineage OS 14.1.
I have tried doing the following through a root shell:
Code:
stop mpdecision
echo "1" > /sys/devices/system/cpu/cpu0/online
echo "1" > /sys/devices/system/cpu/cpu1/online
echo "1" > /sys/devices/system/cpu/cpu2/online
echo "1" > /sys/devices/system/cpu/cpu3/online
echo "2265000"> /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq
echo "2265000"> /sys/devices/system/cpu/cpu1/cpufreq/scaling_max_freq
echo "2265000"> /sys/devices/system/cpu/cpu2/cpufreq/scaling_max_freq
echo "2265000"> /sys/devices/system/cpu/cpu3/cpufreq/scaling_max_freq
echo "2265000"> /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq
echo "2265000"> /sys/devices/system/cpu/cpu1/cpufreq/scaling_min_freq
echo "2265000"> /sys/devices/system/cpu/cpu2/cpufreq/scaling_min_freq
echo "2265000"> /sys/devices/system/cpu/cpu3/cpufreq/scaling_min_freq
After doing this, CPU-Z shows all 4 cores running at 2.26GHz - great! But, while running my code, the performance immediately starts to drop. Upon re-opening CPU-Z, it shows all 4 cores running at ~1500-1900MHz. What gives?
Thermal throttling.
You need to delete or modify /system/etc/thermal-engine-8974.conf
Note: This could make your device run really hot.
Kotaless said:
Thermal throttling.
You need to delete or modify /system/etc/thermal-engine-8974.conf
Note: This could make your device run really hot.
Click to expand...
Click to collapse
Ah, ok, I had a hunch it might be that. Any guidance on max safe temp? Somewhere around 85C maybe? I can always put the phone on top of an icepack while running my code hahaha
Also, which value do I need to modify?
Code:
[CPU0_MONITOR]
algo_type monitor
sensor cpu0
sampling 65
thresholds 115000
thresholds_clr 110000
actions shutdown
action_info 0
Ok so I did some testing, without modifying the thermal-engine.conf file, here's what I found:
Code:
Applying load, idling at 45C:
Runs at 2.26GHz until hitting 67C;
Runs at 1.96GHz for a little bit;
Settles at 1.57Ghz @ 66C
When removing load:
Starting at 1.57GHz @ 65C
Ramps up to 1.96GHz @ 53C
Ramps up to 2.26GHz @ 51C
It looks like throttling starts at 67C and ends at 51C, but none of those values appear in the thermal-engine file. The only value that even looks remotely close is the sampling value, but I'm assuming that's some sort of sampling interval parameter, no?
Well I found a Qualcomm PDF about the DragonBoard 410 that happened to mention
Code:
~# stop thermal-engine
and what do you know, it must apply to the SD800 too, because it worked
Good for you.
I modified this part for myself.
Code:
[SKIN_THERMAL_management_1]
algo_type monitor
sensor xo_therm_pu2
sampling 10000
thresholds 40000 42000 44000
thresholds_clr 38500 40500 42500
actions cpu+lcd cpu+lcd cpu+lcd
action_info 1958400+229 1728000+204 1574400+178