Need some calrification


zeldarocks

Member
Joined
Aug 22, 2010
Messages
108
I'm looking for some clarification with regards to the various lifespan figures thrown around with respect to overcloking. Can anybody provide me with any official or credible data detailing the expected lifespan at default speed?
 
AFAIK there have been 0 units killed due to overclocking so those statistics are unavailable. Wait 15years and maybe a few units will have died and we'll have a better idea.
 
Last edited by a moderator:
The problem is that the CPU is rated at 40000 hours of continuous use, which even at highly dramatic rates of CPU decay (for lack of a better term) would leave the lifespan at several years of continuous use, and nobody has had a pandora for that long.
 
This has been discussed many times before... 50,000 hours (from TI spec) /24 hours per day = 2083 days / 365 days per year = 5.7 years.


Keep in mind those are "continuous use" (although TI calls them "Powered On Hours") hours meaning the hours actually using the Pandora. The Pandora powered on while overclocked sitting at idle or even running a light-duty application like mp3 playback does not really count toward this time... you would need to actually make the processor work at or near peak capacity for the time to really count. This is the nature of the ARM processor.


Getting a little more technical, "50,000 hours use" is misleading since its really trying to quantify when the silicon will breakdown to the point that something internal gets short-circuited. When the processor is over-volted (OPP>3), the current through the chip increases creating more heat and a higher chance for arcing to occur. This is what actually damages the chip, not simply running faster (overclocking just lets you do the damaging stuff faster). Since these damaging effects do not occur unless the processor is actually doing something, sitting at idle does not really contribute to the chip degradation. There is still a very small factor of chip degradation due to leakage current but this factor is far outweighed by actual use. The 50,000 hours of use number is based on over-volted clock cycles and is therefore somewhat misleading since the assumption is that the processor is running at 100% capacity the entire time. Also misleading is the fact that TI assumes the processor is running at the stock clock rate of 720MHz. A more accurate way to measure this would be to specify a total number of clock cycles but that would just be confusing! The 50,000 hours number is easier to comprehend but somewhat inaccurate.


Ultimately it comes down to this... something else (like the buttons, screen, hinges, power switch, etc) on the Pandora is likely to break before the chip dies to overclocking.
 
Back
Top