• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Long term overclocking effects?

L

llmercll

[H]ard|Gawd
Joined
Feb 17, 2010
Messages
1,369
This is something I've been wondering for awhile.

Assuming temps are great (as good or better than stock temps), and stock voltage (or even an under-volt), does overclocking, either the cpu or gpu, lead to earlier hardware failure?

I know that heat does, for obvious reasons
I know that increased voltage can

But does simply working the cpu/gpu harder than it's designed to have any impact? Can someone explain to me why or why not, and explain a little bit?

I did a bit of reading, and at the moment, I'm under the impression it does SLIGHTLY decrease lifespan, but something minor like a few months on a 10 year period. Has to do with the parts wearing a bit faster?
 
nope.. at least not enough that its going to matter.. just stay under the recommended max voltage and dont hover right at the thermal limit of the processor 24/7 and you will be fine..
 
I'd say it's a mixed bag of an answer. It can be done successful if done right with all of the factors taken into account and the RIGHT hardware to do it.

I've had one LGA755 3.0Ghz OCed to 3.4Ghz burn out within a year. The temps were a little high, but the Thermaltake air cooling managed it. Running F@H did it in. You can't run an OC and a heavy load 24/7 on a computer for long. I had to replace the dead 3.0Ghz with a 2.4Ghz and limp along thereafter.

On another machine, i limited the OC and how heavy the load would run and for how long. Well, the machine is still running, but after the first six months, the thermal paste failed. I had to redo the thermal paste. The thermal paste was the OEM to begin with.

When OCing you have to take in all the concerning factors to successfully do the job right. One little slip will cause the hardware to fail prematurely.
 
bmaverick said:
I'd say it's a mixed bag of an answer. It can be done successful if done right with all of the factors taken into account and the RIGHT hardware to do it.

I've had one LGA755 3.0Ghz OCed to 3.4Ghz burn out within a year. The temps were a little high, but the Thermaltake air cooling managed it. Running F@H did it in. You can't run an OC and a heavy load 24/7 on a computer for long. I had to replace the dead 3.0Ghz with a 2.4Ghz and limp along thereafter.

On another machine, i limited the OC and how heavy the load would run and for how long. Well, the machine is still running, but after the first six months, the thermal paste failed. I had to redo the thermal paste. The thermal paste was the OEM to begin with.

When OCing you have to take in all the concerning factors to successfully do the job right. One little slip will cause the hardware to fail prematurely.
Click to expand...


3.0 to a 3.4ghz overclock.. and it failed within a year.. it was most likely going to fail within that year stock.. unless it was a pentium 4 with HT.. 400mhz isnt anything on most processors these days except for the horrid phenom I's..
 
none.

When you overclock to high Tplh or Tphl isn't respected, and the behavior of the switch is undefined. What you get is junk logic. No damage is done to the chip.

edit: yeah this is assuming all the systems that supply everything can keep up. Overclocking increases switching speed which means these little FET caps are charging and discharging at a faster rate, which means they draw more current, if the voltage regulation cant handle that or... I donno, the crystal oscillator cant keep up with your new frequencies (which would be news to me), you could well fry your chip. On a quality board, or rather even just a non-disfuctional board, overclocking is completely safe.

That said, if you bring the chip up to 100C or 200% above its stock voltage your in trouble. All Semiconductors days are numbered, just like everything else by using them your wearing them out. That said, at stock voltage with any overclock and in reasonable thermal conditions your good for at least a decade or two
 
Last edited:
yeah, sometimes I wonder what would happen if all of us [H] computer guys would invest the money we do in computers into something else. I mean, we spend thousands of dollars apiece per year on equipment that literally will be worth half that in one year. just something to think about...
 
would invest the money we do in computers into something else
Click to expand...

Hookers and blow typically lead to drastically reduced lifetime and reliability compared to a mild OC.
 
if you're doing it right, the difference should be negligible. i bought a 2500+ barton when they first came out, immediately turned it up to 2.2ghz and it's still working just fine. it sees typical "average joe" usage.
 
Well in all honesty it's for my kid brother, who would see HEAVY gaming usages. Basically when he's not at school he's gaming on it.
 
well, most games dont utilize the processor like they should. they all rely on the GPU, so he'd be fine on an over clocked core i'd think. get a nice cooler and keep the temps low. it will look cool and keep its life span. i bet that an overclocked core with a proper cooler will last just as long as a core that is all stock.
 
I dont think many people who overclock would care much even if the overclock reduced the cpu lifespan from 10 years to 5 years. Imagine having to use an old athlon xp to game with, that's just torture.
 
demografik said:
if you're doing it right, the difference should be negligible. i bought a 2500+ barton when they first came out, immediately turned it up to 2.2ghz and it's still working just fine. it sees typical "average joe" usage.
Click to expand...

hah, good to see someones still running an athlon XP. barton + Nforce 2 = :D
 
MrWizard6600 said:
hah, good to see someones still running an athlon XP. barton + Nforce 2 = :D
Click to expand...

He's not the only one, I do too. I notice a lot of people still have that rig running (a7n8x deluxe here) I've had it for 8 years or so, and still use it to this day. That's the kind of longevity I'm looking for. I didn't once overclock it though =p

Question: When I overclock, it's on my laptop. I'm not currently overclocking my desktop. I use setFSB, and that overclocks the FSB, which in turn overclocks the cpu. Is that how most people do their overclocking? Is that the "right" method? what else does the fsb overclock?
 
On a laptop or a plain jane dell the set fsb program is about all you can do with the exception of doing a BSEL mod on the pins of the cpu to "trick" the motherboard into thinking it is a cpu with a higher FSB spec. Most of us, for desktops, use enthusiast boards that allow us to directly set the FSB in the computers bios and by far that is the most common and best way to OC.
 
FalconSS said:
yeah, sometimes I wonder what would happen if all of us [H] computer guys would invest the money we do in computers into something else. I mean, we spend thousands of dollars apiece per year on equipment that literally will be worth half that in one year. just something to think about...
Click to expand...


so what drops half the price of what we buy? i can never remember any tech except gfx cards dropping majorly in price. NOthing drops 50% maybe you buy your items the day it comes out from most expensive store? i shop around buy online and never have anything drop in price more than 20 bux.

As far as overclocking if you keep it same volts or small bump in volts and cool better than stock cooling you SHOULD be safe. Ive had one 2.6ghz fail on me when oc'd to 3ghz but it wasnt in my box so not sure why it died. just know it had a good cooler but was in a hot room. I still overclock half a ghz every chip or more on low vid chips.
 
So basically there's always an increased risk of failure, though pretty minor.

I'm using SetFSB to OC my P7450 to 2.4GHz.
 
I've had a q9450@3.6 with volts 1 notch above VID spec (but below max spec) for over a year and it's been Folding for 4 months now practically non-stop (ie 24/7 constant load).

Keep your voltage in safe range at or below max specs and make sure your cooler can handle it and there's almost no reason the oc and the cpu can't live forever. IMHO if you look at the Intel datasheet for your cpu and keep Vcc and Vtt below their max range, you should be fine. That's the line I've always followed, never crossed, and going back that's been 6 cpus, 3+ years I've oc'd that haven't died. (Funny thing is, it wasn't until I got an E2180 that I even tried to oc; back before then on my 1.4ghz T-Bird and my P4 3.0C I didn't know how.)

The other key is know when to say when.
In my case I can run my i7 at 1.2625 which is .015v above stock, hit 3.9, with F@h load temps at 70-71. When I tried going higher, it was taking a Vcc of 1.3+ and temps were hitting 80. 100 mhz+ is not worth more volts and (more importantly in this case) 10c higher temps. My e-peen wants the magic number "4", but really I don't care...I'm happy with where I'm at. Besides that the extra 100mhz probably won't yield me a lower TPF running -bigadv (F@h) anyway, so again, why bother. Just put your own limits down and compare what it takes to go one notch higher and really think about it, see if it's worth it to you to go that extra step.
 
llmercll said:
He's not the only one, I do too. I notice a lot of people still have that rig running (a7n8x deluxe here) I've had it for 8 years or so, and still use it to this day. That's the kind of longevity I'm looking for. I didn't once overclock it though =p

Question: When I overclock, it's on my laptop. I'm not currently overclocking my desktop. I use setFSB, and that overclocks the FSB, which in turn overclocks the cpu. Is that how most people do their overclocking? Is that the "right" method? what else does the fsb overclock?
Click to expand...

on a laptop you need to be very careful and you need to have your hopes very very minimal. Thermal requirements will keep your overclock modest. If you're going to overclock your laptop, make sure you clean its cooling system religiously.

A good number of laptops actually need to be underclocked. The Macbook Air is a good example, that CPU is already putting out way too much heat for the cooling system its running on, bringing it down to 1.2GHz or so would drastically increase its life expectancy.
 
Heats not an issue, my CPU never gets higher than 48c and usually runs around 42c under load, 29c idle.

I'm undervolting AND overclocking at the same time. I wonder if undervolting an overclock could have negative health impact though, moreso than just OCing.
 
llmercll said:
Heats not an issue, my CPU never gets higher than 48c and usually runs around 42c under load, 29c idle.

I'm undervolting AND overclocking at the same time. I wonder if undervolting an overclock could have negative health impact though, moreso than just OCing.
Click to expand...

no it will not.
 
MrWizard6600 said:
hah, good to see someones still running an athlon XP. barton + Nforce 2 = :D
Click to expand...
Another nForce2 rig here (see sig). As you can see it's been OCed by a fair amount, and has been running great for several years. Of course, it helps to have good hardware. :D

sok0 said:
I dont think many people who overclock would care much even if the overclock reduced the cpu lifespan from 10 years to 5 years. Imagine having to use an old athlon xp to game with, that's just torture.
Click to expand...
Hey, easy there bub! :p
 
Last edited:
The hardware you are overclocking will become obsolete LONG before your "sane" overlock has any affect on them.
 
aweusx2 said:
so what drops half the price of what we buy? i can never remember any tech except gfx cards dropping majorly in price. NOthing drops 50% maybe you buy your items the day it comes out from most expensive store? i shop around buy online and never have anything drop in price more than 20 bux.

As far as overclocking if you keep it same volts or small bump in volts and cool better than stock cooling you SHOULD be safe. Ive had one 2.6ghz fail on me when oc'd to 3ghz but it wasnt in my box so not sure why it died. just know it had a good cooler but was in a hot room. I still overclock half a ghz every chip or more on low vid chips.
Click to expand...

Ummm no, they definitly do drop 50% in one year. For example, in late 2008 the intel e2160 was going for about $90....today they are worth about $40.

This is interesting too - showing Intels price cuts per processor: http://www.tomshardware.com/news/Intel-Processor-Price-Cuts,6862.html
 
F@H killed 3 graphics cards I owned. It completely destroyed the core.

OC computers + F@H or any other 24/7 100% Load program = Bad Idea

As long as your temps and voltages are not too high, and your not running 24/7 Full Load programs, you should be fine. I have a Socket 478 Pentium 4 2.53GHZ was running at 3.8GHZ, and it didnt even suffer the SNDS because I had its temps at a max LOAD at 53'C and at 1.74v. It was dismantled months ago, for a 5GHZ project, which I need time to get the POT polished still.
 
Matthew Kane said:
F@H killed 3 graphics cards I owned. It completely destroyed the core.

OC computers + F@H or any other 24/7 100% Load program = Bad Idea

As long as your temps and voltages are not too high, and your not running 24/7 Full Load programs, you should be fine. I have a Socket 478 Pentium 4 2.53GHZ was running at 3.8GHZ, and it didnt even suffer the SNDS because I had its temps at a max LOAD at 53'C and at 1.74v. It was dismantled months ago, for a 5GHZ project, which I need time to get the POT polished still.
Click to expand...

If you fire a signal through an FET that hasn't had a respected setup time what you get out is an analog voltage: it cannot be decoded as either a 1 or a 0. it is, literally, a 0.5. Computers don't do 0.5. The result: 50% chance of bluescreen. The FET that just had this happen to it is indistinguishable from the one that didn't.

But even that's not happening with a stable overclock (the above is infact the definition of an unstable overclock). I just cant describe how frequency is totally and completely irrelevant to the products life, and I'm scratching my head trying to come up with a good analogy. edit: the thing to remember here is that an overclock doesn't make the transistor to more work. An overclocked CPU isn't working any harder than one that isn't... kinda. Thats not a very good analogy because it is as evidenced by the fact that its power consumption is higher.

I've heard some people mention an effect called "electron migration" used on this forum. I'm not a chemist, so I don't know exactly what that is or how it works, but I was told that if you run a current through an FET for long enough eventually the semiconductor itself will shift in its crystal state such that there are nanoscopic changes in resistances whiten the semiconductor. This can lead to improper switching. Thus, I presume, if you move larger currents through the semiconductor this effect happens faster. By overvolting you are moving larger currents. Its also worth mentioning that even at stock voltage this is happening.

That's from overvolting, not overclocking.

If your argument is that a part running at 70C all the time will die faster than one the same part running at 40C all the time, I'd agree. If your argument is that running a CPU at 1.7V will die faster than one being run at 1.5V, again I'd agree.

But if, with all other circumstances identical, your suggesting that a 3.5GHz CPU will die any faster than the exact same one running at 2.0GHz, your wrong. But it is true that the 3.5GHz CPU will consume more power, thus requires more cooling to make the circumstances identical between the two. Its also worth mentioning that with faster switching comes higher power consumption and with higher power consumption comes a higher stress on the power regulation. "all other circumstances identical" requires solid voltage regulation with a PSU which has a nice low noise and very capacitive 5V rail.
 
Last edited:
Generally, unless you go totally insane with voltage and/or heat, the CPU will outlast your desire to use it.
 
bigddybn said:
The hardware you are overclocking will become obsolete LONG before your "sane" overlock has any affect on them.
Click to expand...
Exactly. I had an E6400 (2.13GHz) @ 3.2GHz in my server folding for three years 24/7 without a single missed unit. It'd be on without a reboot for anywhere from 3-6 months straight too. I have since dismantled the server and have retired the E6400 to my parent's computer, and it's now back at stock speeds and undervolted, still no issues. As long as you start with a properly working chip and overclock responsibly, the life you take off the part won't matter as it'll be worth more as recyclable scrap than for processing by the time it matters.
 
All of this talk about old processors still running makes me want to pull out my old ASUS mobo and Athlon XP 1800 - god that was an epic overclocking CPU. I have really been wanting to make a simple HTPC - perhaps a Boxee setup with my old 9800 pro?
 
MrWizard6600 said:
If you fire a signal through an FET that hasn't had a respected setup time what you get out is an analog voltage: it cannot be decoded as either a 1 or a 0. it is, literally, a 0.5. Computers don't do 0.5. The result: 50% chance of bluescreen. The FET that just had this happen to it is indistinguishable from the one that didn't.

But even that's not happening with a stable overclock (the above is infact the definition of an unstable overclock). I just cant describe how frequency is totally and completely irrelevant to the products life, and I'm scratching my head trying to come up with a good analogy. edit: the thing to remember here is that an overclock doesn't make the transistor to more work. An overclocked CPU isn't working any harder than one that isn't... kinda. Thats not a very good analogy because it is as evidenced by the fact that its power consumption is higher.

I've heard some people mention an effect called "electron migration" used on this forum. I'm not a chemist, so I don't know exactly what that is or how it works, but I was told that if you run a current through an FET for long enough eventually the semiconductor itself will shift in its crystal state such that there are nanoscopic changes in resistances whiten the semiconductor. This can lead to improper switching. Thus, I presume, if you move larger currents through the semiconductor this effect happens faster. By overvolting you are moving larger currents. Its also worth mentioning that even at stock voltage this is happening.

That's from overvolting, not overclocking.

If your argument is that a part running at 70C all the time will die faster than one the same part running at 40C all the time, I'd agree. If your argument is that running a CPU at 1.7V will die faster than one being run at 1.5V, again I'd agree.

But if, with all other circumstances identical, your suggesting that a 3.5GHz CPU will die any faster than the exact same one running at 2.0GHz, your wrong. But it is true that the 3.5GHz CPU will consume more power, thus requires more cooling to make the circumstances identical between the two. Its also worth mentioning that with faster switching comes higher power consumption and with higher power consumption comes a higher stress on the power regulation. "all other circumstances identical" requires solid voltage regulation with a PSU which has a nice low noise and very capacitive 5V rail.
Click to expand...

I think you got me wrong for a bit MrWizard6000, as in my previous post, I mentioned [H] overclocks + a program taxing the system at 100% load 24/7 isn't a great idea, I've read countless of articles and threads, where the system will start bluescreen after a whiles of 100% load programs hence one like Folding@Home, once after the bluescreen the system can no longer be used at that overclocked level.

Also I mentioned that voltage and temperature are vital to an overclock as well as using good grade components. And as demonstrated with an example my 6 year old P4 Northwood at 3.8 last year, the chip has been at that speed for 5 years now, since given to me, why didn't it suffer the Sudden Northwood Death Syndrome? Well as you mentioned electro migration can kill your component even if the state of change is just the slightest, and unlike 70% of the Northwoods that suffered this from overclocking, mine hasn't and still going strong. Why? Voltage control and good temperatures. I've experienced that for a overclocked Northwood A or B rev. you need to keep temps below 56'C and your voltage must never exceed 1.75v unless your doing quick burst of extreme overclocking with Dry Ice or Liquid Nitrogen, then things are looked different that way.
 
Matthew Kane said:
F@H killed 3 graphics cards I owned. It completely destroyed the core.
Click to expand...

Were the nVidia Cards? Some of the nvidia stock coolers are able to handle the constant load with additional air flow.

Back to the OP's question, i have had no issues with overclocking and stressing the CPU.
I have ran Folding@Home on all my systems & have overclocked many of them in the past without any issues. Having a CPU at 100% load 24/7 does require good cooling and limiting you overclock. Example would be only going to 3.0Ghz on a Q6600 instead of the higher 3.2-3.8 that some users get here. While you could do it, it is *safer* to bring it down a level due to the nature of the application we use.

I have not seen any evidence to support it will damage the cpu if you keep the temps and voltage in check.
 
MrWizard6600 said:
hah, good to see someones still running an athlon XP. barton + Nforce 2 = :D
Click to expand...

My barton 2500 mobile (best dam o/c chip since the celeron 300a's, with core2 E6600 a close second) on an Abit NF7-S is still going strong at 2.4ghz with barely any voltage bump. That's been like 6 years daily use, not a single problem.
 
Intel Q6600@3.6Ghz Asus P5K 8gb mushkin memory 9x450 Built aug 2006, semi retired Jan 2010. Was only shut off sometimes to clean out dust, a few moves, a few lan parties. 24/7 operation was the norm for it.

Until it was retired, it still successfully ran 24+hrs of prime 95. Though over that span, I changed heatsinks from a TRU to a TRUE to a H50. Video card upgraded to a gtx260 from 8800gtx and it's seen hard drives come and go and come back again. I'll resurrect it one day to media center pc for my home theater, but right now the ps3 is doing a great job at that.

My experience is if done right, most systems are workhorses. I still have a P3 733@1ghz running for my mom's machine, a AMD Athlon XP 1.2@ 1.4 for a sister's box, and a P4 2.8@ 3.4 for my dad's box. All rigs I used in the past and are still being used by immediate family. I shudder at the amount of dust those 3 machines have in them right now.
 
Long term effects = saves you money and increases the length of time you're likely to actually use the hardware in my experience. The oldest pc I still have is a P4 2.8@3.4ghz with a 6800GS OC used for 4 years, but that has sat in the closet for over 3 years now besides using the psu for leak testing.
 
Extreme overclocking kept for a long period of time can severely decrease the lifespan of a CPU, but even then your CPU will outlive how long you will use it for.

Even if you had a Core i7 at 4GHZ for 5 years, the Core i7 would most likely still live for another 10 years.
 
You must log in or register to reply here.
Back
Top