Overclock TP2 in WM?

[nayr1482]

Originally Posted by Holyman007

I'm having the exact same problem. Either I get a text message or phone call while I'm away and it locks itself back up into pocketshield before putting the phone to sleep which causes the SOD. Everything else works fine without SOD.

For now, I'm going to put it back to the "safe speed" (604mhz) so hopefully it'll prevent it from happening. I am using Setcpuspeed.

Here are the specs causing the SOD.
Overclock speed: 710mhz
Throttled speed: 604mhz
Polling: 200ms

somebody is going to have to give me the registry values associated with Pocketshield...

[lovingHDTV]

Increasing the frequency of the CPU will burn more power as the dynamic power of a cpu is directly related to the frequency with which it is run.

The dynamic power of a circuit is P=ACV^2F. So the power is a function of the A=activity factor (how much of the cpu is switching), C=capacitive load, V=voltage squared, this is why voltage matters so much, and F=frequency at which it is running. So the power consumption will increase linearly with the increase in frequency. A 20% overclocking will consume 20% more power.

For more details google "dynamic power equation"

Originally Posted by nayr1482

Everyone keeps talking about saving battery - I dont think underclocking and/or overclocking in this case really affect the battery life. We are not changing voltage applied to the CPU (yet ) so should not make much of a difference...

[nayr1482]

Originally Posted by lovingHDTV

Increasing the frequency of the CPU will burn more power as the dynamic power of a cpu is directly related to the frequency with which it is run.

The dynamic power of a circuit is P=ACV^2F. So the power is a function of the A=activity factor (how much of the cpu is switching), C=capacitive load, V=voltage squared, this is why voltage matters so much, and F=frequency at which it is running. So the power consumption will increase linearly with the increase in frequency. A 20% overclocking will consume 20% more power.

For more details google "dynamic power equation"

and how much Activity do you think is going on when the phone is asleep?
how much load is occurring when the phone is asleep?

how come all of us that have been working on this have not noticed any decrease in battery life?

one of the main reasons is the phone sets itself back to 528 when it goes to sleep. the other is - the OC does NOT use a direct linear factor of voltage...if that were the case my battery would be lasting about 60% of the time it used to.

also when overclocking computers the voltage is constant (or at least set as a constant in BIOS) - why would you assume a phone cpu is any different?

[lovingHDTV]

The reason that you don't see a direct 60% change in battery life is that the CPU is only a portion of the entire power consumption of the device. The screen, radios, Memory, etc all make up the entire power budget of the device. If the CPU only accounts for 5% of the total power consumption, then you wont see much difference in the battery life when you overclock.

Mobile cpus are very different from desktop cpus. Completely different market segments and power requirements. In mobile devices the use of dynamic voltage scaling and power switching is very common in an attempt to reduce both dynamic and leakage power. Desktops "typically" don't use dynamic voltage scaling, but instead rely on voltage switching (on,off) as they are more concerned about leakage power.

Has anyone looked at the current load on the battery to compare between an overclocked device and non-overclocked. I'll have to take a look at this I know that tBattery reports current draw.

BTW I'm running at 710Mhz and the weather animations are flawless for the first time.

Originally Posted by nayr1482

and how much Activity do you think is going on when the phone is asleep?
how much load is occurring when the phone is asleep?

how come all of us that have been working on this have not noticed any decrease in battery life?

one of the main reasons is the phone sets itself back to 528 when it goes to sleep. the other is - the OC does NOT use a direct linear factor of voltage...if that were the case my battery would be lasting about 60% of the time it used to.

also when overclocking computers the voltage is constant (or at least set as a constant in BIOS) - why would you assume a phone cpu is any different?

[nayr1482]

Originally Posted by lovingHDTV

The reason that you don't see a direct 60% change in battery life is that the CPU is only a portion of the entire power consumption of the device. The screen, radios, Memory, etc all make up the entire power budget of the device. If the CPU only accounts for 5% of the total power consumption, then you wont see much difference in the battery life when you overclock.

Mobile cpus are very different from desktop cpus. Completely different market segments and power requirements. In mobile devices the use of dynamic voltage scaling and power switching is very common in an attempt to reduce both dynamic and leakage power. Desktops "typically" don't use dynamic voltage scaling, but instead rely on voltage switching (on,off) as they are more concerned about leakage power.

Has anyone looked at the current load on the battery to compare between an overclocked device and non-overclocked. I'll have to take a look at this I know that tBattery reports current draw.

BTW I'm running at 710Mhz and the weather animations are flawless for the first time.

that would be sweet if you can look at the battery load. from our "testing" (lol) of just using our phones i dont notice any difference in life - and i guess at the end of the day that is all that is important to me. if we ever figure out a way to underclock while asleep...my guess is this will result in an additional 5 minutes of battery life =)


anyway - it would be great if we can figure out a dynamic OC and improve battery life - but I am not holding my breath

[MixIsMobile]

dont know if this is due to the overclocking or not, but this has happened to me twice now

when i get a call while talking the screen shuts off, if i try to turn the screen on while talking it will not turn on, if i unplug the phone the amber light stays lit like it's still plugged in, i can still talk, but i have to reset my phone to be able to get back into it

this happen to anyone else, doesn't happen unless i'm talking on the phone, not sure if the 1st time it happened if i had it plugged in or not

[Boominsvx]

Originally Posted by nayr1482

that would be sweet if you can look at the battery load. from our "testing" (lol) of just using our phones i dont notice any difference in life - and i guess at the end of the day that is all that is important to me. if we ever figure out a way to underclock while asleep...my guess is this will result in an additional 5 minutes of battery life =)


anyway - it would be great if we can figure out a dynamic OC and improve battery life - but I am not holding my breath

I don't see why this is even an issue. Rom to rom, we can negate the phones battery performance three fold. The transitions in sense have never been so fluent on my phone as they are overclocked, and even if it does consume a slight bit of battery, I would GLADLY hand it over any day for the positive outcome this software/script has given me. Not that it's irrelevent, but still... is it worth putting so much thought into? My phone is unbelievable these days and I run a full day without charging on heavy use. At between 16 and 20 hours use, I still have 30 to 50% life left... and I'm on ENERGYROM!! lol.

Originally Posted by MixIsMobile

dont know if this is due to the overclocking or not, but this has happened to me twice now

when i get a call while talking the screen shuts off, if i try to turn the screen on while talking it will not turn on, if i unplug the phone the amber light stays lit like it's still plugged in, i can still talk, but i have to reset my phone to be able to get back into it

this happen to anyone else, doesn't happen unless i'm talking on the phone, not sure if the 1st time it happened if i had it plugged in or not

My phone was doing this same thing with the setcpuspeed program. Yesterday morning, I hard reset for cleanup purposes, and installed the script rather than the program. The phone has been running great since. Are you using the program, or the 4.1 script?

[MixIsMobile]

Originally Posted by Boominsvx

I don't see why this is even an issue. Rom to rom, we can negate the phones battery performance three fold. The transitions in sense have never been so fluent on my phone as they are overclocked, and even if it does consume a slight bit of battery, I would GLADLY hand it over any day for the positive outcome this software/script has given me. Not that it's irrelevent, but still... is it worth putting so much thought into? My phone is unbelievable these days and I run a full day without charging on heavy use. At between 16 and 20 hours use, I still have 30 to 50% life left... and I'm on ENERGYROM!! lol.



My phone was doing this same thing with the setcpuspeed program. Yesterday morning, I hard reset for cleanup purposes, and installed the script rather than the program. The phone has been running great since. Are you using the program, or the 4.1 script?

ya i'm using the program, should i uninstall that and just use the script???

[evertec]

Originally Posted by lovingHDTV

The reason that you don't see a direct 60% change in battery life is that the CPU is only a portion of the entire power consumption of the device. The screen, radios, Memory, etc all make up the entire power budget of the device. If the CPU only accounts for 5% of the total power consumption, then you wont see much difference in the battery life when you overclock.
.

Another reason why you don't see much of a change in overclocking is that the CPU isn't necessarily using 20% more power on a 20% overclock, because the time that it takes to complete a task and return to idle is shorter when the processor is running at a higher frequency. Of course, it depends on what you're doing but for example, loading a web page might be 15% faster when you overclock 20% higher, thus the 20% more power draw is over a 15% lower time period, so the overall effect is negligible.

[Boominsvx]

Give it a try. Like I said, it worked on mine. The script you need can be found HERE from nayr's post number 175. I'm actually using v4, not v4.1, but the difference is only a failsafe timeout at startup, and I'd rather my phone start faster. It can be disabled in the gui, so go for v4.1. msm7kcpuspeed will need to be installed as usual for this to work, but does not need to be opened for anything other than checking the cpu speed to be sure the script is running. Scripts go in the storage card. shortcut to OC_auto4 goes in the startup folder in your windows directory, and remember to run OC_auto_GUI to select the desired overclock speed. Soft reset for good measure. Any problems, pull the SD card and soft reset.

Don't forget to hit the thanks buttons if you use their stuff. nayr, 240 rep? I'ma try and fix that. Anyone else showing recent dedication to this, chime in and claim your praise as well.