How much quicker does overvolting kill your CPU?

[spat55]

I know this can't be a define answer, but I have a i5 3570k on stock voltage. I would like to try and get it to 4.5GHZ but I am hoping to keep it for at least another 3-5 years, so just wondering if say I put 1.3v through it would it be ok? I have had it a year so far and it does have a EVO cooler so that should be fine.

 

[Mushkins]

"Overvolting kills your CPU" is a concern, but it's important to remember that a lot of these overclocking forums and resources get a lot into the theoretical and put blinders in the practical.

Yes, pumping more voltage through the chip will reduce its lifespan. Yes, theres a fancy curve and some math that will estimate theoretically by how much under perfect circumstances if its run 24/7/356. Some of those results are big and scary, but need to be taken into context. If a .02v overvolt will reduce the life of your chip by 25%, and that lifespan was say, 20 years (note that these numbers are all made up), the overvolt is knocking five years off the life of the chip, but it'll still last for another 15 years.

As long as your not overvolting to the point of instantly frying the chip, odds are you're going to replace the whole system anyway before the chip itself dies because of the increased stress. There's plenty of cases just on this forum of people with highly overclocked chips of all shapes and sizes eventually hitting a little instability after *years* of use and having to lower the overclock a little to keep stable, but they still work just fine.

 

[spat55]

"Overvolting kills your CPU" is a concern, but it's important to remember that a lot of these overclocking forums and resources get a lot into the theoretical and put blinders in the practical.

Yes, pumping more voltage through the chip will reduce its lifespan. Yes, theres a fancy curve and some math that will estimate theoretically by how much under perfect circumstances if its run 24/7/356. Some of those results are big and scary, but need to be taken into context. If a .02v overvolt will reduce the life of your chip by 25%, and that lifespan was say, 20 years (note that these numbers are all made up), the overvolt is knocking five years off the life of the chip, but it'll still last for another 15 years.

As long as your not overvolting to the point of instantly frying the chip, odds are you're going to replace the whole system anyway before the chip itself dies because of the increased stress. There's plenty of cases just on this forum of people with highly overclocked chips of all shapes and sizes eventually hitting a little instability after *years* of use and having to lower the overclock a little to keep stable, but they still work just fine.

What will this little instability cause? Say it becomes unstable and I don't know what is wrong, will it do anything bad before I can fix it?

 

[ViRGE]

What will this little instability cause? Say it becomes unstable and I don't know what is wrong, will it do anything bad before I can fix it?

It's exactly as Mushkins says: the system becomes unstable at its overclocked speeds. It's permanent in the sense that you won't get those speeds back (at least not without more voltage), but it's not "broken" in the sense that you can just reduce the clockspeed on the processor and get at least a couple more years out of it.

 

[Mushkins]

What will this little instability cause? Say it becomes unstable and I don't know what is wrong, will it do anything bad before I can fix it?

There's honestly no way to tell how the instability will present. You might start getting bluescreens or WHEA errors, applications may start to crash oddly or things just run "off"/slower, it may decide to throw boot errors and not get to the windows desktop. Its the same gamut of instabilities that tell you your original overclock is not stable, just presented over time instead of more immediately.

The biggest one to worry about is the potential for silent data corruption. Which is why it's good to keep solid backups of known good versions of anything important regularly. As a general rule I only overclock gaming rigs, and be sure to keep any absolutely mission critical data far away from them just in case.

 

[XiandreX]

I know this can't be a define answer, but I have a i5 3570k on stock voltage. I would like to try and get it to 4.5GHZ but I am hoping to keep it for at least another 3-5 years, so just wondering if say I put 1.3v through it would it be ok? I have had it a year so far and it does have a EVO cooler so that should be fine.

I am running my 3570k at 4.4ghz stock voltage. I suggest (if you can) running slightly lower at stock volts instead of pushing voltage for diminishing returns.

 

[BrightCandle]

I haven't seen any reports of dead chips at 1.3V yet. At 1.4 we saw some instantly dead sandy bridges and ivy bridge in theory is not as food with increased voltage. If stability and life trumps speed then you will probably not want to go that high, instead limiting yourself a lot earlier than 4.5@1.3. Because there is a remote chance your CPU will just die.

 

[XiandreX]

I haven't seen any reports of dead chips at 1.3V yet. At 1.4 we saw some instantly dead sandy bridges and ivy bridge in theory is not as food with increased voltage. If stability and life trumps speed then you will probably not want to go that high, instead limiting yourself a lot earlier than 4.5@1.3. Because there is a remote chance your CPU will just die.

Stocks voltage is 1.25 right? (not counting overvoltage)

 

[Ratman6161]

"Overvolting kills your CPU" is a concern, but it's important to remember that a lot of these overclocking forums and resources get a lot into the theoretical and put blinders in the practical.......

Totally agree. One thing that I like to point out is that I have never actually had a CPU fail. Motherboards on the other hand, are a different story. I find other components - typically motherboards will fail much sooner than the CPU. The other day I was using the example of my five year old Core 2 Quad 6600 that's been running OC'd from 2.4 to 3.2 Ghz for the whole five years (fairly mile overclock for that chip). It is at the point now, where if that old socket 775 mother board it's on fails, it won't make any sense to buy a new one. The old CPU would still be perfectly good but would end up being decommissioned because its not worth my while to hunt down a 775 board at this stage of the game.

So the life of the CPU is kind of the least of my worries. As long as you don't get outrageous with it you should be fine.

The other observation I've made over the years is that the CPU will become unstable before you get to the point where you are actually doing it harm (keeping in mind that I am a stickler for what constitutes "stable").

 

[Smartazz]

My 2500K has degraded since I got it. I used to be able to run 4.5GHz at 1.3 volts but it takes 1.32 volts now. I don't run my desktop 24/7 anymore so I raised it to 4.6GHz at 1.36 volts. I would assume that how often the machine is on plays a major role.

 

[XiandreX]

My 2500K has degraded since I got it. I used to be able to run 4.5GHz at 1.3 volts but it takes 1.32 volts now. I don't run my desktop 24/7 anymore so I raised it to 4.6GHz at 1.36 volts. I would assume that how often the machine is on plays a major role.

Seems ( to me anyways) that 4.4ghz at stock is safer than 4.5ghz at 1.3 for the long term anyways..

 

[ShintaiDK]

It depends on how much overvoltage and by the chip lottery.

In short, dont overvolt and dont overclock if you cant afford to lose it. And if your data is important, then dont do any of the two either.

 

[XiandreX]

It depends on how much overvoltage and by the chip lottery.

In short, dont overvolt and dont overclock if you cant afford to lose it. And if your data is important, then dont do any of the two either.

Very carefully chosen words ShintaiDK and I approve 🙂 makes a lot of sense

 

[Jacky60]

CPU's are designed to run for several years (more than 5) if you're an enthusiast I suspect even with relatively extreme overclocking/over volting you'll replace CPU before it dies and it will live until it's obsolete much more often than not. I never read about CPU's failing here and if they did I suspect I'd have read about them by now.

 

[JimmiG]

Keep in mind that temperatures also affect the life span of the CPU. Most who overclock invest in better cooling, so your temps even when overclocked may actually be lower than @ stock speeds with the stock cooler.

I agree with XiandreX that it may make sense to give up some speed in order to get a "safe" overclock. For example, assume that 4.4 GHz = 60 FPS, then 4.5 GHz = 61.4 FPS, provided everything scales perfectly with clock speed. Since nothing scales perfectly, the increase is probably more like 0.5 FPS. If you have to really bump the voltage to achieve those extra 0.5 FPS, it may not be worth it.

 

[Ed1]

Stocks voltage is 1.25 right? (not counting overvoltage)

no, that is not within stock range for 3570k . At 3.6ghz (max clock for all 4 cores) you should see something around 1.05-1.10 .
Thing to remember also is most of the time it will be at low clocks 1600 with voltage below 1.0 (0.950) and only go to max when turbo kicks in ,so amount of time your in max turbo and high voltage is very limited unless you can play games 24/7 and even here most games only use about 50% cpu usage .

If your at all worried still just OC small amount, say 4.2-4.3 that would most likely only take very small bump of voltage or can get away with auto setting (maybe less than 1.2v) your not going to miss 100-200 in respect of 4.0+ speed .
There a sweet spot around 4.2-4.3 where starting from stock and moving up it take hardly any voltage to around 4.2 , then you get to a multiplier were each step up will take more voltage and this produces heat .So you could find this spot and be happy .

 

[XiandreX]

no, that is not within stock range for 3570k . At 3.6ghz (max clock for all 4 cores) you should see something around 1.05-1.10 .
Thing to remember also is most of the time it will be at low clocks 1600 with voltage below 1.0 (0.950) and only go to max when turbo kicks in ,so amount of time your in max turbo and high voltage is very limited unless you can play games 24/7 and even here most games only use about 50% cpu usage .

If your at all worried still just OC small amount, say 4.2-4.3 that would most likely only take very small bump of voltage or can get away with auto setting (maybe less than 1.2v) your not going to miss 100-200 in respect of 4.0+ speed .
There a sweet spot around 4.2-4.3 where starting from stock and moving up it take hardly any voltage to around 4.2 , then you get to a multiplier were each step up will take more voltage and this produces heat .So you could find this spot and be happy .

Ok thanks for clearing that up for me. I didnt increase any voltage when I did my manual overclock. I think I used a template for my bios and board posted from a reliable 4.4Ghz template that a lot of people were hitting on my Motherboard + Cpu. I forget the name offhand.

 

[pmv]

"Overvolting kills your CPU" is a concern, but it's important to remember that a lot of these overclocking forums and resources get a lot into the theoretical and put blinders in the practical.

Yes, pumping more voltage through the chip will reduce its lifespan. Yes, theres a fancy curve and some math that will estimate theoretically by how much under perfect circumstances if its run 24/7/356. Some of those results are big and scary, but need to be taken into context. If a .02v overvolt will reduce the life of your chip by 25%, and that lifespan was say, 20 years (note that these numbers are all made up), the overvolt is knocking five years off the life of the chip, but it'll still last for another 15 years.

As long as your not overvolting to the point of instantly frying the chip, odds are you're going to replace the whole system anyway before the chip itself dies because of the increased stress. There's plenty of cases just on this forum of people with highly overclocked chips of all shapes and sizes eventually hitting a little instability after *years* of use and having to lower the overclock a little to keep stable, but they still work just fine.

Could you equally well _extend_ the lifespan by undervolting and running it slower?

Which would imply there's no more reason to worry about a small over-volt and a slightly reduced lifespan than there is to worry about _not_ under-volting and slightly increasing it?

You could probably extend its lifespan indefinitely by never actually turning the PC on, i.e. running it at 0V and 0Ghz?

 

[Ed1]

Ok thanks for clearing that up for me. I didnt increase any voltage when I did my manual overclock. I think I used a template for my bios and board posted from a reliable 4.4Ghz template that a lot of people were hitting on my Motherboard + Cpu. I forget the name offhand.

Ok, few things to know .
1) auto voltage "will" increase with clock/multiplier as there is a voltage table it looks up per CPU (your VID )
2) be careful with MB auto-clocking apps or profiles, they tend to increase voltage more then needed in many cases .

Put it back to stock (auto everything , Default settings ) and run prime95 and note your load voltage and temps .
This gives you now a baseline to work off of and you'll be able to see how each clock step affects voltage and temps .
Like I said above you might find auto voltage ok with lower OC but above 4.2 I would definitely move to manual/offset values (offset once you know your stable max clock manual voltage) . that way it will idle with low voltage, hence one advantage with offset mode .

Edit: One more thing since you have a Z77 Asus MB , If you find your changing multiplier but when you get to windows it shows a older setting, this is a Asus bios bug .
The fix is reflash with latest or same if your up to date , many pull there hair trying to figure whats going on with this .So if your going up and down with 40x, 41x ,42x, 43x etc you might run into this after a while .

 

[blackened23]

CPU's are designed to run for several years (more than 5) if you're an enthusiast I suspect even with relatively extreme overclocking/over volting you'll replace CPU before it dies and it will live until it's obsolete much more often than not. I never read about CPU's failing here and if they did I suspect I'd have read about them by now.

I wouldn't brush the concerns off so quickly. I have, in the past, had CPUs die or become unstable over time for whatever reason from pushing excessive voltage -- this is less of a concern nowadays due to improved idle states where you can use idle voltages when the CPU isn't in turbo mode. Back in the olden days, you generally were forced to use maximum voltage 24/7. I've had various chips that just mysteriously became unstable at previously stable overclocks - and in each case I knew that no other system component was at fault (since I plunked a new CPU in with no issues). Now, in each case the instability didn't present itself until 2+ years later, in which I generally wanted to replace the system anyway. So I suppose it wasn't a huge deal since I over-volted knowing the risk, and knowing that I would want to replace the system in a couple of years anyway.

It's always a risk you take when you overvoltage by an excessive amount -- But it is less of a concern now since you can take advantage of offset and adaptive voltage, which allows the CPU to dial down to low voltage during idle states. That definitely helps increase longevity of overclocked / over-volted CPUs these days.

 

[Ajay]

I wouldn't brush the concerns off so quickly. I have, in the past, had CPUs die or become unstable over time for whatever reason from pushing excessive voltage -- this is less of a concern nowadays due to improved idle states where you can use idle voltages when the CPU isn't in turbo mode. Back in the olden days, you generally were forced to use maximum voltage 24/7. I've had various chips that just mysteriously became unstable at previously stable overclocks - and in each case I knew that no other system component was at fault (since I plunked a new CPU in with no issues). Now, in each case the instability didn't present itself until 2+ years later, in which I generally wanted to replace the system anyway. So I suppose it wasn't a huge deal since I over-volted knowing the risk, and knowing that I would want to replace the system in a couple of years anyway.

It's always a risk you take when you overvoltage by an excessive amount -- But it is less of a concern now since you can take advantage of offset and adaptive voltage, which allows the CPU to dial down to low voltage during idle states. That definitely helps increase longevity of overclocked / over-volted CPUs these days.

Definitely true and I've seen the same issue. I've over-volted too heavily for six months (some K7 variant), then clocked it down when I built a new machine for my wife. When it was time for another new machine for her (i.e., I was ready for a new one!), I took back her old one to add to my F@H farm and could not overclock it to the same level. So I settled for a mild OC and moved on. That was a good lesson.

 

[spat55]

Looks like I have lots of reading too do!!! Thanks.

 

[spat55]

Very carefully chosen words ShintaiDK and I approve 🙂 makes a lot of sense

Think I am going to keep this at stock, until it becomes obsolete for me. The reason I brought it will overclocking in mind is so when it is getting so I want a upgrade I can push it hard for a year till the thing explodes!

 

[XiandreX]

Ok, few things to know .
1) auto voltage "will" increase with clock/multiplier as there is a voltage table it looks up per CPU (your VID )
2) be careful with MB auto-clocking apps or profiles, they tend to increase voltage more then needed in many cases .

Put it back to stock (auto everything , Default settings ) and run prime95 and note your load voltage and temps .
This gives you now a baseline to work off of and you'll be able to see how each clock step affects voltage and temps .
Like I said above you might find auto voltage ok with lower OC but above 4.2 I would definitely move to manual/offset values (offset once you know your stable max clock manual voltage) . that way it will idle with low voltage, hence one advantage with offset mode .

Edit: One more thing since you have a Z77 Asus MB , If you find your changing multiplier but when you get to windows it shows a older setting, this is a Asus bios bug .
The fix is reflash with latest or same if your up to date , many pull there hair trying to figure whats going on with this .So if your going up and down with 40x, 41x ,42x, 43x etc you might run into this after a while .

Not sure if you were talking about me but I am using the Msi z77a-g45. My offset is 87%. How does that equate to voltage or am I misunderstanding that part?

 

[Machiavegli]

Your CPU is more likely to become obsolete than suffer any ill effects from overclocking. Motherboards are another story. They definitely seem to die more quickly on me if I am overclocking them.