CPU toggle

[Yufice]

What is the CPU toggle for?

 

[publicanimal]

What is the CPU toggle for?

Unlit: Underclock CPU by approximately 200 MHz
Lit: Full speed ahead

 

[killjoylm1]

Really??

I don't see any changes if the CPU toggle is lit or un-lit using chevys kernel and system info to observe any differences. Would be great if I could see anything. Thanks for any help.

~G

 

[publicanimal]

Really??

I don't see any changes if the CPU toggle is lit or un-lit using chevys kernel and system info to observe any differences. Would be great if I could see anything. Thanks for any help.

~G

Changes wouldn't be in system info. Run Quadrant Standard performance tests and you will see the difference.

 

[killjoylm1]

Really??

I don't see any changes if the CPU toggle is lit or un-lit using chevys kernel and system info to observe any differences. Would be great if I could see anything. Thanks for any help.

~G

Changes wouldn't be in system info. Run Quadrant Standard performance tests and you will see the difference.

Just went ahead and ran Quadrant and the exact opposite is what I saw. My score was 1129 when lit and 1294 when unlit. I feel that it has to do with the CPU governor than lowering the CPU speed... just my 2 cents!

~G

 

[TRIKOTRET]

i thought lit means on and 200 under performance

 

[killjoylm1]

I've heard people say both, but from my experience unlit is faster. Also, it is a down arrow this would imply to me that when lit that it decreases something.

 

[bp328i]

Just went ahead and ran Quadrant and the exact opposite is what I saw. My score was 1129 when lit and 1294 when unlit. I feel that it has to do with the CPU governor than lowering the CPU speed... just my 2 cents!

~G

I had the same outcome a higher score with it unlit.

I've heard people say both, but from my experience unlit is faster. Also, it is a down arrow this would imply to me that when lit that it decreases something.

I was thinking the same thing when I saw the arrow down.

 

[publicanimal]

I tested this extensively on V2 and always had the same results. Higher quadrant scores when lit and lower quadrant scores when unlit.

I just ran 2 quadrant tests. I am using Chevy's ULV 1.2 GHz kernel.
Unlit: 1213
Lit: 1379

I'm not sure what is causing you guys to experience different results and I agree that it is a bit counter-intutive with the "down arrow". But I have tested this again and again and on my phone the results have always supported the conclusion that unlit = underclocked.

Are you guys using SetCPU? I'm pretty sure SetCPU would override the CPU toggle. I do not use SetCPU because I don't believe it is necessary for this ROM as it was designed for a phone with a 1.0 GHz processor (Nexus One). My performance is beyond stellar when not using SetCPU.

 

[killjoylm1]

Ran it again and same results. Not using SetCPU and I agree it would get in the way. There must be some way to actually get a readout from the system that will clear this up for everyone.

I think that it is interesting to look at the options under battery in the settings menu. It has an option to reduce to the CPU speed to save battery power. When my phone goes into battery saving mode the CPU toggle becomes lit. This would imply that it is saving battery power by lowering to the CPU speed.

 

[bp328i]

I tested this extensively on V2 and always had the same results. Higher quadrant scores when lit and lower quadrant scores when unlit.

I just ran 2 quadrant tests. I am using Chevy's ULV 1.2 GHz kernel.
Unlit: 1213
Lit: 1379

I'm not sure what is causing you guys to experience different results and I agree that it is a bit counter-intutive with the "down arrow". But I have tested this again and again and on my phone the results have always supported the conclusion that unlit = underclocked.

Are you guys using SetCPU? I'm pretty sure SetCPU would override the CPU toggle. I do not use SetCPU because I don't believe it is necessary for this ROM as it was designed for a phone with a 1.0 GHz processor (Nexus One). My performance is beyond stellar when not using SetCPU.

I was thinking the same thing as you utill last night when I ran Quadrant several times and it was always higher unlit. I'm not running SetCPU, I did on the very first release of MIUI and had major issues. So the next time I flashed the ROM I never downloaded it from the market.

The only other thing I can think of is I did not change the kernel.


Edit: I just loaded up Slayher's 1200 kernel then ran quadrant a few different ways.

Unlit:
1338 & 1498
Closed quadrant then reopened
Lit:
1427 & 1580

Then I left it lit and rebooted my phone. When my phone restarted I lit it again, this is the funny part. I received just the opposite.

Unlit:
1363 & 1369
Lit:
1072 & 1324

 

[publicanimal]

Ran it again and same results. Not using SetCPU and I agree it would get in the way. There must be some way to actually get a readout from the system that will clear this up for everyone.

I think that it is interesting to look at the options under battery in the settings menu. It has an option to reduce to the CPU speed to save battery power. When my phone goes into battery saving mode the CPU toggle becomes lit. This would imply that it is saving battery power by lowering to the CPU speed.

It is interesting that we are running the same experiment and experiencing opposite results. I have no idea what could account for that. I have the battery saving mode disabled as I simply swap batteries when my battery level gets low (I have 3 batteries and a standalone battery charger so I charge my batteries outside my phone).

What kernel are you running? Maybe kernels with interactive governors (i.e. Slayher's kernels) react differently?

Very curious indeed.

 

[killjoylm1]

Ran it again and same results. Not using SetCPU and I agree it would get in the way. There must be some way to actually get a readout from the system that will clear this up for everyone.

I think that it is interesting to look at the options under battery in the settings menu. It has an option to reduce to the CPU speed to save battery power. When my phone goes into battery saving mode the CPU toggle becomes lit. This would imply that it is saving battery power by lowering to the CPU speed.

It is interesting that we are running the same experiment and experiencing opposite results. I have no idea what could account for that. I have the battery saving mode disabled as I simply swap batteries when my battery level gets low (I have 3 batteries and a standalone battery charger so I charge my batteries outside my phone).

What kernel are you running? Maybe kernels with interactive governors (i.e. Slayher's kernels) react differently?

Very curious indeed.

I am using Chevy's ULV 1.1 GHz kernel. Not really sure what the deal is, but it is really responsive w/ or w/out the CPU toggle lit. Would like to know, but more happy that everybody is back on board developing MIUI!! :icon_ banana:

 

[publicanimal]

I just ran the test again, and my results are more inconclusive than ever, though for the first time I got a lower score with it lit than unlit.

Procedure:
Rebooted phone
Observed that CPU toggle was unlit
Ran quadrant test: 1236

then:
Turned CPU toggle on, rebooted phone
Observed that CPU toggle was unlit upon reboot (this must be the default)
Turned CPU toggle on so it was lit
Ran quadrant test: 1161

So it was lower, but only by about 75.

I now think you guys are right, however I can't for the life of me figure out why I was getting the results I was getting before. I guess I always ran the tests back to back, first unlit, then lit.

Weird. If anyone knows Chinese maybe they could ask the powers that be at Miui in China.

 

[bp328i]

Time to load up Slayher's 1250 kernel, install Set CPU crank it all the way up and see what I get on quadrant.

Well that went over like a fart in church.