Does overclocking shorten device life?

[jnelsoninjax]

Me and a buddy were talking about how we like to overclock our CPU's, and he made the off hand remark that he really doesn't push the overclock to say the 5GHz range because he would not want to have to replace the CPU sooner than he had to, which got me thinking: when overclocking it does generate more heat,which theoretically could 'shorten' the life of an electronic equipment. Is this the case, or am I over thinking the whole thing?

[Mindovermaster Moderator]

Well, yes and no...

 

Of course a CPU OC'd does overheat. That's why most OCers use water cooling. It cools the CPU faster than air.

 

Any  CPU, motherboard, RAM, anything can break down rather you use it to its own potential or not. Nothing has a specific date on it.

 

If your CPU is a K model, that thing wants you to overclock it. It is conditioned to do so. A non-K version, I'd be wary about overclocking.

 

That clear your questions?

[Jim K Global Moderator]

Noticeably shorten the lifespan ... no ... theoretically ... maybe ... I guess.  

 

Anyway, you'll typically replace the system loooong before the CPU/GPU fails.

[Mindovermaster Moderator]

1 minute ago, Jim K said:

Anyway, you'll typically replace the system loooong before the CPU/GPU fails.

That, too Jim. Was just going to add that, until I saw your post. lol

[darkgtx]

Personally I like overclocking & undervolting when possible. As long as you have a decent board and stay within operational temps while overclocking everything should be fine. RAM usually starts showing its limits immediately after a bad OC through programs crashing etc. For me OCing is more about finding balance between performance/ heat dissipation.

[Open Minded]

Voltage and heat are the biggest issues for an overclock.  I've had my Sandy Bridge 2600k at 4.0GHz with an undervolt for just over 6 years and it's just as stable as the day I bought it.  Idles at about 30-35C and 50-60C at 100% load.  I run 100% CPU usage for BOINC every night and sometimes all day if I'm out and running around.  Never any problems.

[LaP]

1 hour ago, jnelsoninjax said:

 

Me and a buddy were talking about how we like to overclock our CPU's, and he made the off hand remark that he really doesn't push the overclock to say the 5GHz range because he would not want to have to replace the CPU sooner than he had to, which got me thinking: when overclocking it does generate more heat,which theoretically could 'shorten' the life of an electronic equipment. Is this the case, or am I over thinking the whole thing?

 

Overclocking the cpu no. Heat yes. Increased voltage can be problematic too if you don't know what you are doing. Increasing bus speed over what the ram and mb are rated can be problematic too.

 

My core i5 750 have been overclocked for mostly all of it's life and it's like almost 10 years old i think.

 

But it's a simple bus overclock (the cpu is not unlocked) with ram supporting the frequency of the overclock out of the box and the mb too. So basically the mb and ram can handle the increased bus speed out of the box and only the cpu is overclocked. But the temp are fine. I bought a good cooler and the temp are in the 50ies while gaming which is perfectly fine.

[DevTech]

3 hours ago, jnelsoninjax said:

 

Me and a buddy were talking about how we like to overclock our CPU's, and he made the off hand remark that he really doesn't push the overclock to say the 5GHz range because he would not want to have to replace the CPU sooner than he had to, which got me thinking: when overclocking it does generate more heat,which theoretically could 'shorten' the life of an electronic equipment. Is this the case, or am I over thinking the whole thing?

 

1. Logically it is impossible for heat to be a problem since you will employ the level of cooling you need to keep temperatures "normal" - otherwise the CPU throttles itself to no performance or it crashes - so you will either run to the store and buy cooling or you will stop overclocking.

 

2. Too much voltage is the CPU killer. And current CPUs use a very low voltage level so it's just a bit easier to make a mental error and miscalculate a voltage increase. Many BIOS chips have a "relative voltage multiplier" which can be very easy to forget when you add an absolute increment and then the two settings combine for a dangerous outcome. Never set a relative voltage and never increase voltage more than one smallest increment at a time and always check overclocking forums to research the typical maximum voltage for the chip you have and never exceed that.

 

3. So the "Shorten the Lifetime" mythology comes from "Electro-Migration" which is a function of increased voltage that exceeds design specs for the chip fabrication process. In theory, if you overvolt the CPU past the process specs for a long time, conductive areas on the chip can "migrate" until it fails. It is hard to prove or disprove in real life since overclockers will trhow out a bad chip and move on to the next greatest thing and without examining the chips with an electron microscope in every case, who knows if it was "Electro-Migration" or they just went one voltage increment too far... In any case if this phenomenon plays any part in the lifetime of an overclocked CPU, it only applies to extreme overvolts and as people have already stated the reduced lifetime would be long after you give in to the latest CPU fashion and upgrade to the new shiney toy.

[jnelsoninjax]

Wow! The responses are amazing here. As I said (or rather implied) before, it was one of those questions that I asked out of sheer curiosity, my friend thinks that he knows everything about everything, so when he starts in on a subject I tend to not believe everything said. As to my system, it's an black edition AMD, so it has the unlocked multiplier,which made it easier to OC without messing with voltages and fsb, and I'll admit that I have no clue about OC,except the little bit that I've picked up here.

[+InsaneNutter MVC]

I remember having the exact same discussion with a friend of mine around 10 years ago, back in 2007. He was adamant overclocking a processor was stupid and would massively reduce its usable lifespan.

 

At the time I had upgraded my PC on a small budget, I went with a Core2Duo E4500 knowing it would overclock really well and I could save a lot of money over the more expensive E6600.

 

At stock the E4500 was 2.2GHz, I overclocked this to 2.93Ghz, to which my friend was adamant with such a massive overclock the processor would fail in a couple of years at most.

 

Well fast forward almost 10 years, it's now 2017 and that PC is still going strong... I added an SSD to it and gave it to my parents a few years ago. The PC is now running Windows 10, used daily by my parents and is still really responsive for everything they do on a computer.

[Mando]

11 hours ago, Open Minded said:

Voltage and heat are the biggest issues for an overclock.  I've had my Sandy Bridge 2600k at 4.0GHz with an undervolt for just over 6 years and it's just as stable as the day I bought it.  Idles at about 30-35C and 50-60C at 100% load.  I run 100% CPU usage for BOINC every night and sometimes all day if I'm out and running around.  Never any problems.

my 2600k has been locked at 4.8ghz since I bought it, 6yr and counting, now with my parents at stock without a single issue. My 7700k will no doubt be the same at its 4.9Ghz current clock.....this time in liquid 115i  ftw!

[Mindovermaster Moderator]

59 minutes ago, Mando said:

my 2600k has been locked at 4.8ghz since I bought it, 6yr and counting, now with my parents at stock without a single issue. My 7700k will no doubt be the same at its 4.9Ghz current clock.....this time in liquid 115i  ftw!

Well, Ks are meant to be overclocked, that is what they are made to do.

[Circaflex]

Overclocking no. Excessive heat and or excessively high voltages yes. 

[Circaflex]

45 minutes ago, Mindovermaster said:

Well, Ks are meant to be overclocked, that is what they are made to do.

Just because it's a k processor doesn't mean it will live any longer or not degrade due to excessive heat or voltages. They are only considered overclockers chips because they have unlocked multipliers. You still face the same risk. 

[Dot Matrix]

Like was said above, it shouldn't really. I ran an OC'd GFX card for nearly 10 years without any issues. 

[Mindovermaster Moderator]

2 hours ago, Circaflex said:

Just because it's a k processor doesn't mean it will live any longer or not degrade due to excessive heat or voltages. They are only considered overclockers chips because they have unlocked multipliers. You still face the same risk. 

I'm not saying it will last any longer. I'm just saying they are overclockable, and made for it.

[Open Minded]

21 hours ago, Circaflex said:

Just because it's a k processor doesn't mean it will live any longer or not degrade due to excessive heat or voltages. They are only considered overclockers chips because they have unlocked multipliers. You still face the same risk. 

Actually, an Intel engineer did an AMA on Reddit and stated the K series CPUs are better binned.  They are the best of the best of that model of chip.  Yes, you do face the same risk and I agree with what you've said, but the K series are the best of what Intel has to offer so it's not just an unlocked multiplier.  I don't know if that will directly translate into longer life for the CPU, however.

[LaP]

On 2017-03-29 at 3:12 AM, DevTech said:

3. So the "Shorten the Lifetime" mythology comes from "Electro-Migration" which is a function of increased voltage

Theoretically higher temp diff can shorten the life of components too. Your cpu might be stable at 80 Celsius while gaming/working but the constant ~50 Celsius temp diff between idle and gaming/working might shorten it's life versus a temp diff of 30 Celsius between load and working/gaming.

[DevTech]

4 hours ago, LaP said:

Theoretically higher temp diff can shorten the life of components too. Your cpu might be stable at 80 Celsius while gaming/working but the constant ~50 Celsius temp diff between idle and gaming/working might shorten it's life versus a temp diff of 30 Celsius between load and working/gaming.

That is cherry picking a situation that is never going to happen in real life - nobody but a complete idiot is going to run their CPU at 80 C 24/7 unless they live in Alaska and their home heating broke down.

 

The subject of how complete idiots can kill a CPU is a different topic perhaps?

 

In the real world there are ZERO heat effects that will harm a modern CPU with it's internal heat sensor and consequent throttling.

[adrynalyne]

5 minutes ago, DevTech said:

That is cherry picking a situation that is never going to happen in real life - nobody but a complete idiot is going to run their CPU at 80 C 24/7 unless they live in Alaska and their home heating broke down.

 

The subject of how complete idiots can kill a CPU is a different topic perhaps?

 

In the real world there are ZERO heat effects that will harm a modern CPU with it's internal heat sensor and consequent throttling.

I guess you missed where he said  components. 

[DevTech]

9 minutes ago, adrynalyne said:

I guess you missed where he said  components. 

I guess you missed both the component he seemed to be talking about was the CPU and that there is no extra heat anyways. Like I said, only idiots overclock without managing cooling and once we allow idiots into the "probability of shortened CPU lifetime" topic then anything is possible.

 

[adrynalyne]

1 minute ago, DevTech said:

I guess you missed both the component he seemed to be talking about was the CPU and that there is no extra heat anyways. Like I said, only idiots overclock without managing cooling and once we allow idiots into the "probability of shortened CPU lifetime" topic then anything is possible.

 

Components is plural bud.

 

Component is not.

 

He said components.

[DevTech]

1 minute ago, adrynalyne said:

Components is plural bud.

 

Component is not.

 

He said components.

I'm sure he spent at least a few hours debating the exact syntax and grammar of his post to enable you to widen the topic from shortened CPU lifetime to "heat damage of unknown random components" but seriously?

 

Heat is simply not a practical real world factor in CPU lifetime. Period.

 

High voltage might be a factor with no real world evidence. Unknown.

 

Conclusion: overclock your CPU and don't worry about lifetime degradation. Small print: forming many possible hypotheses about idiots not included

 

[adrynalyne]

Just now, DevTech said:

I'm sure he spent at least a few hours debating the exact syntax and grammar of his post to enable you to widen the topic from shortened CPU lifetime to "heat damage of unknown random components" but seriously?

 

Heat is simply not a practical real world factor in CPU lifetime. Period.

 

High voltage might be a factor with no real world evidence. Unknown.

 

Conclusion: overclock your CPU and don't worry about lifetime degradation. Small print: forming many possible hypotheses about idiots not included

 

"Heat is simply not a practical real world factor in CPU lifetime. Period."

 

Sorry, but that is just ignorant.

[LaP]

41 minutes ago, DevTech said:

That is cherry picking a situation that is never going to happen in real life - nobody but a complete idiot is going to run their CPU at 80 C 24/7 unless they live in Alaska and their home heating broke down.

 

 

I was not talking about running the cpu at 80 Celsius 24/7. I'm talking about the difference in temp between idle and load. Like if your CPU is running at 80 Celsius while you are gaming but around 35 Celsius idle that's kind of a big difference in temp and theoretically it can shorten the life of the cpu if you are having a lot of gaming sessions every weeks.

 

Theoretically a lot of heating up and cooling down with a big difference in temp should not be good.

 

I prefer to be safe than sorry and never go over ~60 Celsius when i'm overclocking.