Re: Review on OVERCLOCKED n900 AND RISKS of oc
 

Well, yea. Any desktop CPU - say a Phenom II or Intel Core i7 - is the same chip. The different models are simply ones A, set to run at a particular speed, and B, guaranteed to run at that speed stably for X years.
It may work quite a bit better than that, or fail if you go just a little bit higher.

In the case of mobile processors, though, battery life is king and so a number of phone CPUs are significantly "underclocked" from their actual limits to produce the right batterylife/performance point.
Provided the chips don't overheat, and there is enough battery power, you may(depending on the chip) be able to safely overclock it a bit, at the expense of battery life.

In our case, I'm not sure. Our chip is -supposed- to be rated for 500mhz, and was "factory overclocked" to 600. But that doesn't mean that the hardware won't safely support more than that, nor that doing so will end up killing the chip quickly. The real problem is, we don't know what the -actual- average lifespan is at any given frequency; all we know is the -rated-/minimum lifespan, IIRC.

Re: Review on OVERCLOCKED n900 AND RISKS of oc
 

Speak about comparing apples with oranges.

Any modern desktop cpu is designed for efficient heat dissipation, the majority even have small heat-spreaders already built on. ARM cpus are a completely different beast. They're optimized for space and power efficiency. Ever heard about PoP-Design (package on package)? Modern ARM processors consist of multiple thin chips slapped on top of each other (random googled image). Such a design isn't meant for efficient heat dissipation. Your knowledge about desktop bios OC settings is meaningless here!

That being out of the way - you guys seriously planning on using your n900 for more than 5 years? I'd guess most of the n900 owners here would have it replaced after 3 years or so. It will still be a fine little chap, but modern "smartphones" will blow it out of the water. So why sweat it? Enjoy the openness of the platform and experiment with it.

Re: Review on OVERCLOCKED n900 AND RISKS of oc
 

don't get it wrong, you know for sure more about this than i do. my understanding was that for the cortex 1GHz should be ok. but wasn't it the for more fragile supprtchips and not the cortext that a lower clock speed was chosen?

Re: Review on OVERCLOCKED n900 AND RISKS of oc
 

Wasn't this topic already covered by another thread?
http://talk.maemo.org/showthread.php?t=51811

Re: Review on OVERCLOCKED n900 AND RISKS of oc
 

yes and in many others. but there is no real answer.

the "hurray, my n900 runs at xy hz and no damage up to now"
as in that threat are completely useless to answer the question.

destruction from overclocking/heat happens not necessarily at once(need to overclock a lot and.. ). a damaged chip might seemingly run fine for quite a while and then (sometimes slowly, sometimes out of a suddon) fail .
whats more once you damaged a chip, you started a process of accelerated selfdestruction. downclocking doesn't stop the process once started. the destruction goes on and finaly it fails.

and what you will see once that happens, would be something like: the xy app doesn't work anymore or out of a suddon xy doesn't work anymore. out of a sudden many applications often crash(with no clue to why ). so you have no clue where it comes from.
seems a lot of users having such kind of problems and we will never know if it came from once overclocking the device.

imo it stays a interessting question where the limit is. how far can we go? wich voltage...?

in short damage by overclocking is like pressing a selfdestruct button that starts an unstoppable countdown for selfdestruction.

Re: Review on OVERCLOCKED n900 AND RISKS of oc
 

Agreed. Even if an overclocked device fails, that doesn't definitively mean overclocking caused it. It could be the CPU, but it might be the memory or just a simple capacitor or other component. Maybe the CPU or capacitor or whatever would've failed even if the phone was left totally stock. It could be from overclocking - or not - but even examination by skilled technicians may not determine the exact cause of failure.

My gut feeling is that if voltages are lowered from stock settings to reduce heat, thereby offsetting some inevitable increased heat from overclocking, that it basically becomes a wash in 'normal' i.e. 'average' use and the phone isn't really significantly stressed. Of course, some users will use their phone in ways that stress it more than other users. So how a unit is used may have more to do with potential failure than overclocking will, per se.

As for me, FWIW I've been running at 1Ghz + lv profile for months with zero problems. IMHO the odds of device failure from me dropping it are far far higher than from overclocking. IOW I ain't worried about overclocking one bit.

Re: Review on OVERCLOCKED n900 AND RISKS of oc
 

many users will have a new device in one or two years and won't see effects from damage.. and now winter comes so it gets cooled down...

i raised a different question. mantan(and others for that matter) stated the cortex a8 can run at 1ghz. looking at the specs that seems to be true. so there seems to be no issue(well depends on the environment ... ).
what i asked was: whats about the other chips for i think they are more likely to be an issue if there is one.

whereever you look, they talk about the cortex but not about the rest. and not beeing an expert in this i am curious whats about the other stuff related to overclocking.

Re: Review on OVERCLOCKED n900 AND RISKS of oc
 

I hope you weren't serious when saying that.

Re: Review on OVERCLOCKED n900 AND RISKS of oc
 

sure i was. usual contracs last two years at most. after that the folks make a new contract and have a new device... but we get sidetracked.

is anyone here who realy knows about the effects of overclocking on the device(on the overall not only on the cortex)? i would love to hear some expert opinion on this. would be great.

Re: Review on OVERCLOCKED n900 AND RISKS of oc
 

Just wondering...

Since the pr1.3 was released, I see more efficient CPU usage and the overall response on the device seems more fluid and efficient.

Of course by overclocking the CPU, the device performs faster in general but there are many variables that cannot be calculated by the users or individual developers in this forum i.e. the lifespan of the cpu, overheating on the small device vis a vis a desktop cpu and other things.

Since overclocking seems to be so popular among the users here, why is it that the updated firmware not address it a little bit? Is it possible that there are somehow some minor repercussions? It might not be very severe but there could be reasons why it is set at stock value to run at 250 - 600 Mhz.

After reading a couple of forums here and there, I can see that one of the major reasons why the manufacturer would seriously underclock the device is due to battery life i.e. Iphone clocked at about 533Mhz thus Apple seems to boast LOOOONG battery life... Maybe the Nokia manufacturers think that the general public would prefer using their devices for a certain period of time and therefore decided to stock clock it at that value which allows the phone to be run at a certain length of time? No?

If not, if the device can in fact, run MUCH smoother, at NO costs whatsoever to the device, then why won't they set it at THAT value? I'm sure they have tested it to a certain degree during R&D to find out the best CPU speed to run perfectly for a lengthy period.

Besides, I would like to think that if they feel that this open-sourced device can run BETTER without much hassle (as what we all can see, individuals at home able to modify their CPU speeds), why not put it in the next firmware update? Unless there is somewhat a MARKETING PROPAGANDA and refusing to back down in what they have set it originally...

I think if Nokia realises that it could run better at a different CPU speed, they could easily do that in the next update.

In the meantime, I run my pr1.3 stock speed just fine without much complaints.

What do you guys think?