Re: Difference beetween Nokia, LV, ULV, XLV, Ideal
 

This is why I started using SR.
I use the media player for playing music a lot on a day to day basis.
I found that the media player makes the processor sit at 125Mhz Stock.
On Ideal 500Mhz my battery didn't last as long as stock. I even changed removed the avoid freqs to 250Mhz then 125Mhz, not much difference.
If SR is stable on your device (should be on KP), I find it saves a good bit of battery for my usage.

Re: Difference beetween Nokia, LV, ULV, XLV, Ideal
 

Did you keep 250Mhz disabled?

Re: Difference beetween Nokia, LV, ULV, XLV, Ideal
 

Back in the day, several tests were done to find the "peek" for where the CPU works most efficiently. Most systems can run faster or slower, but that peek is where it runs best (cycles / amp-hour). Titan did most of the testing, and found that the sweet spot was between 500 and 600, depending on the device.

There's also the "race to finish" idea, that the faster you get something done, the more time the CPU can sit in idle mode before it has to do anything else. Suppose a background process running at 250Mhz takes 2 seconds to complete (like checking crcs on files). If that same computation can be done in 1 second with a default of 500Mhz, the energy saved in the 1 second of extra idle time outweighs the extra voltage used to get to 500Mhz. (Especially when the CPU is operating more efficiently at 500Mhz than it is at 125Mhz.)

The only time this doesn't work is when you have a *constant* stream of work at a slower rate (like decoding mp3 streams and dealing with filling audio buffers by hand, a-la gstreamer). Then reguardless, you're going to run above 0, making the task eat a bit more.

FYI: The default Nokia kernel disabled 125Mhz mode, leaving idle mode at 250Mhz. 125 was removed because it was found to be unstable, causing more than it's fair share of problems (mainly write corruption to flash and dsp issues). I would personally avoid that frequency set. In fact, I run 500-900, and find my battery lasts 2 days with moderate usage if needed. (I generally charge every night, regardless of level.)

Re: Difference beetween Nokia, LV, ULV, XLV, Ideal
 

No! Didn't think of that. I could try it for a bit. Thanks for the suggestion.

Re: Difference beetween Nokia, LV, ULV, XLV, Ideal
 

Tried that today and it was flat by 14:30, normally get a day and a bit.

Will test it for a bit longer.

Re: Difference beetween Nokia, LV, ULV, XLV, Ideal
 

Now I'm testing stability of kp49+ideal+sr, limits 500-720 on normal daily usage (crashes etc.) I'll let you know as soon as I finish it - it should be on 19th Jan

Re: Difference beetween Nokia, LV, ULV, XLV, Ideal
 

Unfortunately, SmartReflex is very unstable with kp49+ideal, limits 500-720. During 12 hours I had a reboot (or two, because my setting were set to default kernel)
I strongly discourage from using SR with that profile.

Now I'll test XLV profile, kp49, limits 250-600 with SR

Re: Difference beetween Nokia, LV, ULV, XLV, Ideal
 

I shall say this only once.

Smart reflex in kernel-power v49 renders setting voltage profiles OBSOLETE.

If you are going to use smart reflex then DO NOT BOTHER changing the voltage profiles. Smart reflex automatically undervolts your CPU for you to a level appropriate for your CPU better than any of your empirical testing can indicate.

The only other thing you can really futz with is scaling governor and clock speeds (which is a whole other 1000 post thread as to what is good/bad).

Re: Difference beetween Nokia, LV, ULV, XLV, Ideal
 

And is it possible to get the stock voltages once having set for instance xlv as default?

Re: Difference beetween Nokia, LV, ULV, XLV, Ideal
 

LOL I dunno.