overpowering CPU - causing harm

[tomr]

Hello mates, may I ask you one "simple" question, which I asked few times before and still didn´t get a clear answer?

I am new for overclocking. I have SB 2500K and I started my work...

Intel burn test was stable (20 runs) on 4.5 GHz ~ 1.305 V

However Prime95 crashed in few second. 4.5 GHz in Prime95 was stable (22h 30min of running) on 1.345 V (max temp. when testing was 56°C).

And now the question...

Some friends of mine warned me of overpowering CPU = providing CPU more voltage then is necessarily needed - because overpowering may cause shorter lifetime of CPU.

I don´t run demanding applications most of time + ASUS power management automatically reduces CPU frequency to 1.6 GHz in 95% of time! However, because of my setting in BIOS, it is still provided by 1.345 V, which is obviously TOO MUCH for it...

Well, please, would You tell me, if such "overpowering" of CPU may cause any harm or reduce its lifetime?

Thank you very much!

 

[SieB]

By overclocking the CPU you are pushing it harder than it is meant to run, but as long as you keep it cool with a good CPU cooler you shouldn't have a problem. Overclocking will lower the life of a processor but not drastically and as long as you keep it well cooled you are talking a few months lees in terms of life.

1.34v is slightly high for 4.5ghz have you tried auto tune with the Asus aI Suite II or running it from the BIOS? I got 4.6ghz @ 1.28v by doing a auto tune with th AI Suite so give that a try and see if you get a lower voltage, 1.34v isn't bad though.

 

[tinytomlogan]

By overclocking the CPU you are pushing it harder than it is meant to run, but as long as you keep it cool with a good CPU cooler you shouldn't have a problem. Overclocking will lower the life of a processor but not drastically and as long as you keep it well cooled you are talking a few months lees in terms of life.

1.34v is slightly high for 4.5ghz have you tried auto tune with the Asus aI Suite II or running it from the BIOS? I got 4.6ghz @ 1.28v by doing a auto tune with th AI Suite so give that a try and see if you get a lower voltage, 1.34v isn't bad though.

All CPU's are different remember. Guide him but dont state facts that are incorrect. Auto tune will OVervolt manual is the way forward.

Please be more carefull in future.

 

[SieB]

True, I was just going off the fact I got a lower voltage from auto tune and was just saying to see if he would aswell. I haven't done any manual overclocking on P67 boards so I can't give advice on settings, sorry, i'll remember for next time.

 

[TMaksimov]

1.34 should be fine, dude. Seems you haven't hit the jackpot of the chip lottery either.

 

[marsey99]

any overvolting will reduce the life of a cpu, it always has and always will, but they (intel+amd) say that the vid is enough for the cpu to have a long life.

at this point we dont yet know how much voltage these new intels can take before we accelerate the degredation to the point where we can see it happen so for now i would air on the side of caution and push it as far as it will go on vid. then in the coming months when others have fried theirs on x voltage you know how much is too much

 

[tomr]

Hello mates, thank you for your post!

Few points...

1) I ran CPU on 1.305. IB test was ok, however Prime95 crashed in 5 sec! So I had to rise voltage to 1.345 to make PC stable in Prime95.

Autotune is nice thing. Still I am not sure for stability on 1.28 V... Did you test it in Prime95? I worry, it won´t be very stable on such frequency... is it?

2) I don´t understand very much the thing of voltage and temperature. You see, my CPU is 30°C max all the time even on 1.28 or 1.345 - frequency doesn´t matter. ONLY if I put CPU under pressure, temperature rises... So I am not sure what did you mean by the cooling matter of CPU, when it is irrelevant most of time (but still my Noctua NH-D14 does great work - even under pressure, 22 hours of running = max temp 56°C

Maybe my misunderstanding is based on lack of information about CPU and how it works. So maybe I could keep it simple and ask

-> how exactly may high voltage itself harm CPU??? What if I would rise voltage of 2500K to 1.5 V on 4.5 GHz, but CPU would stay below 30°C? Are you saying, that everything would be OK and my CPU would be unharmed???

Well... how can voltage itself harm CPU (there will be logically some unused voltage most of time on every PC, since nobody stress CPU 24/7... This all confuses me quite a lot... :-(

Thank you for explanation!!!

 

[SieB]

Hello mates, thank you for your post!

Few points...

1) I ran CPU on 1.305. IB test was ok, however Prime95 crashed in 5 sec! So I had to rise voltage to 1.345 to make PC stable in Prime95.

Autotune is nice thing. Still I am not sure for stability on 1.28 V... Did you test it in Prime95? I worry, it won´t be very stable on such frequency... is it?

2) I don´t understand very much the thing of voltage and temperature. You see, my CPU is 30°C max all the time even on 1.28 or 1.345 - frequency doesn´t matter. ONLY if I put CPU under pressure, temperature rises... So I am not sure what did you mean by the cooling matter of CPU, when it is irrelevant most of time (but still my Noctua NH-D14 does great work - even under pressure, 22 hours of running = max temp 56°C

Maybe my misunderstanding is based on lack of information about CPU and how it works. So maybe I could keep it simple and ask

-> how exactly may high voltage itself harm CPU??? What if I would rise voltage of 2500K to 1.5 V on 4.5 GHz, but CPU would stay below 30°C? Are you saying, that everything would be OK and my CPU would be unharmed???

Well... how can voltage itself harm CPU (there will be logically some unused voltage most of time on every PC, since nobody stress CPU 24/7... This all confuses me quite a lot... :-(

Thank you for explanation!!!

I ran Prime 95 for 10 mins but I couldn't do it for any longer because I only have the stock cooler and temps were in the high 70s, I did also do a couple of video encodes and a couple of runs in Cinebench and they all ran stable. But I did another auto tune last night and I got 4.5ghz @ 1.32v but CPU-Z only showed 1.27v and 3.4ghz but the Asus AI Suite showed 4.5ghz and CHM showed 1.32v, when I got the 1.28v I used CPU-Z to see the voltage and Asus AI Suite to see the frequency so i'm wondering if the first time with 1.28v was actually wrong.

As for the higher voltage doing damage, to my best knowledge (sorry if i'm wrong) is you are putting the CPU under more stress and heat than the CPU is meant for pushing it to it's limits, which over time would lead it to degrade faster.

 

[tomr]

Well, that´s nice to hear Still I´ve been told, that to be sure of Prime95 stability, it takes at least 8 hours of non-stop running. But I know, that such extreme stressing of CPU is not probable to happen... ever

However, would anybody have any clear "experienced" answer for me? Can overpowering of CPU ITSELF cause any damage or reduce lifetime of CPU??

-> how exactly may high voltage itself harm CPU??? What if I would rise voltage of 2500K (just for example) to 1.5 V on 4.5 GHz, but CPU would stay below 30°C? Are you saying, that everything will be OK and my CPU will be unharmed in long time basis??? = you see, ASUS automatically drop frequency to 1.6 GHz most of time, but voltage sustain on 1.345 V ALL THE TIME (which is much more above its real needs)! Is this problem? And if yes, then why (if we consider temperature will be under 30°C all the time)??

Thank you for possible explanation!

 

[TMaksimov]

Well it's simple. Think of it this way - the more water that flows through a pipe the quicker it's eroded. I'd say 1.5v is definitely unsafe and I doubt that your CPU will stay below 30*C.. Actually I know why you have such low ambients - The chip's frequency is automatically decreased to save power. I believe Tom covered that in one of the Sandy Bridge vids? Anyway, I doubt anybody yet knows how much volts it takes to ruin the Sandy Bridge processors overtime as they're so new. Just try to keep it under 1.35 and you should be fine in the long term

 

[tomr]

Thank you Todd. Well, I will keep it on 1.345 hoping my CPU will be fine even if it drops to 1.6 GHz most of time

 

[Diablo]

No, he is saying hypothetically if the voltage is high but temperatures low then will the chip degrade. Yes it will, the higher voltage will cause damage to the semi conductors, however the difference between 1.3 and 1.35V is not something to be worried about. If you were putting crazy voltage (1.45V+) through the old 45nm chips, you probably would have problems, however 1.35V should be safe.

Additionally, it is the heat that tends to kill components. The voltage you are putting through will still give a pretty long life (longer than you would want to keep your CPU)

 

[tomr]

Thank you very much for insuring me, that 1.345 V is safe for my CPU! Of course I wouldn´t rise it to 1.4 or even 1.5 However it´s good to know, that it is not ok to rise voltage higher then it is necessarily needed.