what's the average life of a processor that has higher voltage?

[RollerCam540]

For example, how long will an A64 winchester live (stock voltage of 1.4) if the voltage is set to 1.5 (which is a reasonable voltage). I've been trying to decide whether or not I should try going for a higher OC (which would mean raising the VCore) but I do not want to diminish chip life. I heard one place that if temps are good then the chip will live for as long as it was supposed to orginally but I read somewhere else on overclockers.com that power leaks out and chip life can diminish by 1/2 (100,000 hours to 50,000) hours. I'll try and find that link so you guys can read it.

BTW, I only skimmed through the link. I'm a busy guy and I'm only flirting with the idea of raising the VCore so I'll give me life to the article after I hear what you guys have to say.

[Tha SmokaH]

I wouldnt think it what shorten it by any at 1.5volts..that is if its cooled well.

[Zebo]

50000 is 17 yrs 8 hours a day 365 days a year... i think cpu may be obsolete by then don't u?

[freecableguy]

You'll be getting rid of that chip WELL before you see the end of it's life. Crank that fu(ker up to 2.0V.

[RollerCam540]

In other words:, OC'ing at any voltage (especially low voltages) will allow a processor to live 10+ years with no problems at that voltage? For example

a p4 1800a at 1.9v?????
a A64 3200+ at 1.5v

I'm trying to find a document at AMD's site that says what the voltages are for 3500+ and up winchesters. I hear they are 1.5v. Maybe the documentation is only refering to newcastles... I don't know... all I know is it says 1.4v for the lower A64's and 1.5v for the higher A64's... weird!?

My point being is I had a teacher tell me that there are not seperate assembly lines creating processors at different speeds. They make the processor at the highest speed and then test it. If the rate of failure is acceptable, it's stamped at that speed and sold. If not, the clock rate is reduced on the multipler (along with the proper voltage for that clock rate) and the processor is tried again. If an acceptable rate of failure is reached, they stamp the processor and sell it (after locking it and all that good stuff). If not they repeat the last few steps and try again.

If this is the case, then I can safely use 1.5v on my A64 since the higher versions of the chips use 1.5v and not have to worry as long as I get it prime95/benchmark stable.

[Zebo]

My rule is 10% on air but I'm very conservative for n00bs.
i.e.
winch = 1.55
NC/hammers = 1.65

without additional cooling you don't see huge gains above those Vcore anyways..

[RollerCam540]

I'm new here but I really wouldn't consider myself a "n00b". The only reason I'm asking this is because of my budget -- I cannot aford to purchase another chip any where in the future. This is why I'm overclocking (otherwise I would have bought a faster chip to begin with).

I really wish I could post a signature but it won't allow me. I'm currently watercooling this computer so it can handle a bump in VCore?

RollerCam540

P.S. while I'm at it, since I used the word "bump", why do people always reply to their own post with the word "bump" if no one has responded. I haven't caught onto this yet.

[Zebo]

get 3000 if on budget...only $150... 1.55 vcore you should see ~2700+mhz post week 40 CPU's. Run for years like that with stock HSF.

bump = ttt (to the top) so more people see it on first page

[Crankster]

Well my NC get's 1.7 V or more, hasn't hurt yet!

[MaxxxRacer]

anything below 1.65 and it should last longer than you will want the chip for..

Zebo you comment on the 1.55 and 2700mhz is very incorrect. not a whole lot of hte winnies hit 2.7ghz and even less of them do it on the stock hsf.. and the week doesnt make much of a diff. only the week 37's were different, and it was in a bad way...

[Holst]

I think that you wil be able to run 1.7v or more providing you keep your temps reasonable.

The problem with questions like this is that nobody can ever test a chip overclocked for very long.

AXP lasted for years on 2v+ as did other chips people overclocked.
My old P3450 dropped dead after about 4 years of pretty high voltage and overclock, but I dont know why it died.. might have been nothing to do with overclocking...

Look at it this way..
Chances are the chip will live at least a year (being conservatve) and in a years time you will be glad of the excuse to upgrade.

[kiwi]

Reasonable voltage won't shorten cpu life that much, so don't worry. :cool:

[reject]

who keeps stuf more than a year anyway?
all that will happen by heat damage it it slowly gets a lower oc, and old cpus are only used for folding/ office pcs.

[saaya]

amd is thinking about moving to a 5 year warranty for their boxed a64 cpus, that means that even the worst cpus that die faster than others live around 5 years with stock vcore...

and dont forget its not the vcore that kills the cpu!
cpus die because of electromigration, more heat=more electromigration!
so a low vcore wont help you to extend a cpus life if its passively cooled and always cooking at around 70°C ...

dont worry about this, your cpu will be too slow to run the latest apps and games waaaay before it dies from too much vcore :)

[scorp]

A little bit offtopic :
Since the conclusion is that the temp. is very important (usually more important than the vcore you feed the CPU), what temps are recommanded for a Winnie and what tempos for a NC. AMD specifies max. 65C for Winnies and max. 70C for NCs.

[saaya]

A little bit offtopic :
Since the conclusion is that the temp. is very important (usually more important than the vcore you feed the CPU), what temps are recommanded for a Winnie and what tempos for a NC. AMD specifies max. 65C for Winnies and max. 70C for NCs.

temps are not more important, its the combination of vcore and heat that slowly kills cpus. low vcore +a lot of heat is just as bad as low temps but high vcore.

if you want your cpu to live for 2 years or longer i would recommend to keep the temp below 50°C ...dont know about vcore for winnies...

this however doesnt mean that your cpu will die after 1 year or 2 years if you let it run at temps of 50°C or 60°C! lol :D

I read somewhere else on overclockers.com that power leaks out and chip life can diminish by 1/2 (100,000 hours to 50,000) hours. I'll try and find that link so you guys can read it.

BTW, I only skimmed through the link. I'm a busy guy and I'm only flirting with the idea of raising the VCore so I'll give me life to the article after I hear what you guys have to say.

thats the theory dude... just the theory ...
not even amd or intel themselves know how long their own cpus live with a certain vcore and temp ;)

listen to the guys who use cpus for years! :D
stories about cpus that "just died" after some time because of too high vcore are scarce, very scarce... most of the time when cpus die its because it was physically damaged or extremely overvolted.

a friend has an athlon thunderbird 1ghz from back in 2000 oced to 1.8ghz with watercooling. that chip has a default vcore of 1.75v and hes running his chip at 2.1v (or even higher, dont remember) for 5 years now.... and its totally fine ;)

and theres a guy on this forum who was running a pentium3 coppermine (default vcore 1.65v) at 2.4v for a year or longer and that chip is still running afaik...

and my step father still has 3 old 8088 machines that still run after 20 years :D plus he has an oced pentium1 thats running fine for 10 years...


so in a way, yes temps are more important because they rise quickly when increasing vcore... if you keep a cpu cool then you can increase its vcore quite a bit before it dies

[Zebo]

amd is thinking about moving to a 5 year warranty for their boxed a64 cpus, that means that even the worst cpus that die faster than others live around 5 years with stock vcore...

and dont forget its not the vcore that kills the cpu!
cpus die because of electromigration, more heat=more electromigration!
so a low vcore wont help you to extend a cpus life if its passively cooled and always cooking at around 70°C ...

dont worry about this, your cpu will be too slow to run the latest apps and games waaaay before it dies from too much vcore :)

They should discontinue all warranties. I hate having to pay for other peoples mistakes...processor/video card prices would plumet if warranty cost not ammortized over all thier hardware.

MAX I run 2655mhz on SHSF w. 1.55 so does my buddy Duvie@anand.. guess what? both week 37. 24/7 prime too.. Post week 40's all run higher form what i see sticked to AMD CPU forum.

CHECK OUT last 4-6 page here some @ default vcore@2700..

http://www.xtremesystems.org/forums/showthread.php?t=44570

[saaya]

tray cpus dont have an amd warranty afaik, or only a warranty of 1 year but the shops have to decide whether they replace the dead tray cpus or not.

and yepp. amd is sure pushing those winnies, they need to, they need to have chips that do 2.8ghz @stock voltage soon ;)