[ROOT] Using SetCPU + Perflock Disabler to Save Battery, Underclock

[xorbe]

I don't think nandroid backed up everything ... wasn't there a row about the wifi partition not backed up after the 4G key was corrupted?

 

[pingpongboss]

wifi partition?! 4g key?! Either you know something I don't, or you have very bad memory Wifi partition certainly doesn't make sense to me..

 

[Loogie]

What is the App you are running in the first post to show the historical performance data?

 

[pingpongboss]

What is the App you are running in the first post to show the historical performance data?

SystemPanel on the market.

 

[Loogie]

Thanks!

Ok, here's a question for the EE types out there (which I am not).

If you have a processor running at 1ghz and it completes a task in 1 second and then you underclock it to 250mhz and it completes the task in 4 seconds have you saved any power? I'm not sure if the power consumption curve is linear or not. I seem to recall way back in my brain that it is linear which would make me think that you would just be breaking even in the above scenario. I'm sure that there are a bunch of other factors that I have not taken into consideration here (like system overhead processing) but I'm just interested to figure out why the underclocking seems to work from a technical perspective. Can anyone shed any light on this?

 

[Ky1e.YC]

That's something awesome...

 

[statona]

I'm using the Incredible... and it all seems to work great... of course, I have no idea what I'm doing - so thanks to PPM for his great and simple guidance.

However, am I supposed to set the main screen max back to 998 after I have set the profiles, enabled the profiles and disabled the perflock?

 

[pingpongboss]

Thanks!

Ok, here's a question for the EE types out there (which I am not).

If you have a processor running at 1ghz and it completes a task in 1 second and then you underclock it to 250mhz and it completes the task in 4 seconds have you saved any power? I'm not sure if the power consumption curve is linear or not. I seem to recall way back in my brain that it is linear which would make me think that you would just be breaking even in the above scenario. I'm sure that there are a bunch of other factors that I have not taken into consideration here (like system overhead processing) but I'm just interested to figure out why the underclocking seems to work from a technical perspective. Can anyone shed any light on this?

Im also not fully sure on the details, but i'm almost 100 percent sure that an underclock will save battery, and an overclock will wreck your battery.

I'm using the Incredible... and it all seems to work great... of course, I have no idea what I'm doing - so thanks to PPM for his great and simple guidance.

However, am I supposed to set the main screen max back to 998 after I have set the profiles, enabled the profiles and disabled the perflock?

From what i've seen, the main tab just seems to reflect whatever profile the system has currently made active. So changing the main tab wont do anything.

 

[KSmithInNY]

From what i've seen, the main tab just seems to reflect whatever profile the system has currently made active. So changing the main tab wont do anything.

I think this is true in certain scenarios but not always. An example of when you don't need to set the main screen is how you (pingpongboss) have your profiles setup. Because your profiles basically cover every aspect of the phone from charging, to screen off, to screen on .... your profiles take precedence.

An example of when the main screen actually does control the CPU is situations where the user does not have a profile set for a situation:

Screen off - min 245 max 245
Charging - min 998 max 998

You can see though that in the example I didn't have a profile setup for when the screen is on, and not on the charger. This is where whatever you set on the main screen now takes precedence. So in the example above, if i set my main screen at min 245 max 576, when the screen is on and not on the charger the CPU will max out at 576.

Keep up the great work. I love this thread.

 

[UTnick]

i'm just gonna move this up a little bit.

 

[d3xn2o]

i'm just gonna move this up a little bit.

Dont you mean....

BUMP!

 

[imex99]

I'm running damage control, what would my best setcpu settings be?

 

[g_b_h]

Question for the OP..

Are you still getting what you consider better battery like with SetCPU after you updated to Froyo? Also, do you still run that Task Killer? Isn't it possible your improved battery life is due to the TK?

A better control would be to remove the TK, then to try the handset with/without SetCPU.

My experience is that SetCPU makes like or no difference to Froyo's sucky battery life. Froyo is noticeably worse than 2.1, at least for me.

 

[Kevin OQuinn]

I'm running damage control, what would my best setcpu settings be?

Read the thread for your ROM over on XDA. I'm not sure but I think damage control has a lot of tweaks for battery life already built in. That being said, if the kernel that installs with that ROM has perflock disabled already and it doesn't have value already built in you would benefit from these setting in setCPU.

Question for the OP..

Are you still getting what you consider better battery like with SetCPU after you updated to Froyo? Also, do you still run that Task Killer? Isn't it possible your improved battery life is due to the TK?

A better control would be to remove the TK, then to try the handset with/without SetCPU.

My experience is that SetCPU makes like or no difference to Froyo's sucky battery life. Froyo is noticeably worse than 2.1, at least for me.

I have noticed a gain in battery life with these settings. It also has a lot to do with the ROM you run. I'm running the AC ROM which has other things done to improve battery life. When developing the ROM the devs made sure that they weren't gonna have bad battery life. My girlfriends EVO is not rooted and has no tweaks done to it or anything to improve battery life. She has the Froyo OTA on her phone and gets great battery life. Some FROYO ROMs go for performance more than battery life so depending on which one you're running that might be the reason why.

 

[Kevin OQuinn]

I noticed that with the AC ROM you can go as low as 128mhz, but it causes some issues. Waking up was buggy, and getting phone calls when the screen was off was almost impossible. By the time it woke up enough to answer the call I missed the call. I stepped up to the next lowest (the one in the pictures from the OP) and I have had no issues at all.

 

[g_b_h]

Read the thread for your ROM over on XDA. I'm not sure but I think damage control has a lot of tweaks for battery life already built in. That being said, if the kernel that installs with that ROM has perflock disabled already and it doesn't have value already built in you would benefit from these setting in setCPU.



I have noticed a gain in battery life with these settings. It also has a lot to do with the ROM you run. I'm running the AC ROM which has other things done to improve battery life. When developing the ROM the devs made sure that they weren't gonna have bad battery life. My girlfriends EVO is not rooted and has no tweaks done to it or anything to improve battery life. She has the Froyo OTA on her phone and gets great battery life. Some FROYO ROMs go for performance more than battery life so depending on which one you're running that might be the reason why.

I'm using the Froyo OTA, but rooted with unrevoked3/unrevokedForever.

However, I think I might have to take back the assertion in my previous post (that setCPU made no difference). I double checked some settings I had made, and am now using the settings that match those of the OP. Since that change (this morning), battery life seems to have improved.

It will take a few days to better judge, but, after a few hours, it does seem better.

 

[Kevin OQuinn]

Let us know how it goes.

 

[willsurg#AC]

Hi, let me start by admitting I am new to this, but dont the screen shots show a serious flaw for proof that this actually works. I mean the top pic shows crappy battery service yes, but thats from the hours of like 12 noon till what 5pm? with a corresponding high amount of useage at the time period. The bottom pic after using set cpu is from the hours of 12 am till some thing like 6 am, yes the batter discharge is very good however, there was no use occuring, at all, durring corresponding time. It looks to me as if the pictures show what one would expect, if you dont use the phone, you don't discahrge it as much as you would if you were actually using the phone.

 

[Kevin OQuinn]

Hi, let me start by admitting I am new to this, but dont the screen shots show a serious flaw for proof that this actually works. I mean the top pic shows crappy battery service yes, but thats from the hours of like 12 noon till what 5pm? with a corresponding high amount of useage at the time period. The bottom pic after using set cpu is from the hours of 12 am till some thing like 6 am, yes the batter discharge is very good however, there was no use occuring, at all, durring corresponding time. It looks to me as if the pictures show what one would expect, if you dont use the phone, you don't discahrge it as much as you would if you were actually using the phone.

That's kinda the point. If the screen is off it uses a lot less battery life with setCPU. The screen is the biggest battery drain on your phone. You can't do much about that when you're using the phone. All you can do is slow the CPU down when you're doing less demanding things. What the OP's settings do is significantly slow the CPU down when the screen is off (with the obvious assumption being that you aren't doing anything so you don't need the CPU to be running at full speed).

Using these settings I've been getting better battery life than before I had these settings (and everything else stayed the same).

FYI with a stock kernel and ROM the CPU runs at full speed all of the time.

 

[Jerzyiroc]

Hi, let me start by admitting I am new to this, but dont the screen shots show a serious flaw for proof that this actually works. I mean the top pic shows crappy battery service yes, but thats from the hours of like 12 noon till what 5pm? with a corresponding high amount of useage at the time period. The bottom pic after using set cpu is from the hours of 12 am till some thing like 6 am, yes the batter discharge is very good however, there was no use occuring, at all, durring corresponding time. It looks to me as if the pictures show what one would expect, if you dont use the phone, you don't discahrge it as much as you would if you were actually using the phone.

I see what you mean but it's more than that. Basically when you're not using the phone, there's practically no power being used. I can go literally all night and my battery % doesn't drop 1%. Now that doesn't only help when I'm sleeping. It helps out in between usage during the day. How many times have you used your phone, picked it up 30 minutes later and it's down a percentage or two? During the course of the day that percentage that you aren't using when your screen is off REALLY adds up. If I can go all night without battery going down for 8-9 hours, can you imagine how much you save during the course of a normal day in between usage? Then there's the underclocking and having the PCU throttle up and down depending on what it's needed for. That also helps a LOT. Instead of running flat out at 1ghz for something like switching home screens, or browsing online, it'll throttle the pcu up and down to make sure you get what you need. That also helps battery a LOT. So it's not just when the phone is off. It's when you use it too with the pcu being throttled. So I completely see what you're saying. I thought the same before I started using it, but trust me it's more than that and it seriously helps out a lot. Me getting a day+ of moderate usage is normal for me now..