Can a cpu degrade in performance over time

[Smackaroy]

Even if it only a couple of percent?

[Henry Butt]

No the performance of a CPU never degrades. If it was exposed to high temperatures all the time (80-90 degree celsius) then the performance may degrade.

[Smackaroy]

1 minute ago, Henry Butt said:

No the performance of a CPU never degrades. If it was exposed to high temperatures all the time (80-90 degree celsius) then the performance may degrade.

Ok thanks

[manikyath]

2 minutes ago, Henry Butt said:

If it was exposed to high temperatures all the time (80-90 degree celsius) then the performance may degrade.

the silicon may degrade, it may shorten lifespan.. (but then again, what *is* silicon lifespan to start with?)

but it NEVER makes the chip "go slower" overall. thermal throttling yes, but that doesnt mean the chip's performance has degraded, that's the cooler's problem.

[Henry Butt]

Just now, manikyath said:

the silicon may degrade, it may shorten lifespan.. (but then again, what *is* silicon lifespan to start with?)

but it NEVER makes the chip "go slower" overall. thermal throttling yes, but that doesnt mean the chip's performance has degraded, that's the cooler's problem.

True.

 

[Gamer4714]

The cpu's performance never degrades, it's just the software. The hardware requirements for software is increasing by a small percentage every year. Your cpu may have been the fastest when it was released but after a couple of years it will slow down due to the software. The cpu's old architecture's instructions sets and stuff will become old to the new software. If you run the cpu on software which was there at the time of it's prime, those software would almost the same way it did. And yes cooling is also a factor in cpu slowing down

 

[Smackaroy]

10 minutes ago, manikyath said:

the silicon may degrade, it may shorten lifespan.. (but then again, what *is* silicon lifespan to start with?)

but it NEVER makes the chip "go slower" overall. thermal throttling yes, but that doesnt mean the chip's performance has degraded, that's the cooler's problem.

I'm just wondering why my cpu is performing slightly slower. Could bios up dates do this? 

 

Edit. It could also be slightly hotter in my room now its summer meaning a bit slower clocks too.

[Henry Butt]

12 minutes ago, Smackaroy said:

I'm just wondering why my cpu is performing slightly slower. Could bios up dates do this? 

 

Edit. It could also be slightly hotter in my room now its summer meaning a bit slower clocks too.

You can still check for background tasks, start up apps etc. or the cooler and thermal paste.

 

[Somerandomtechyboi]

Performance never degrades

The cpu itself can degrade so you might notice you need more volts for a certain freq than before but its usually not noticable by the average user

 

Heat can degrade cpus but as long as your cpu doesnt go over 95c should be fine

 

Voltage can degrade and maybe even kill cpus when too high, from what i can tell volts seem to degrade the cpu exponentially after a certain volt but thats just a stupid guess, then again if you follow safe voltage guidelines or manufacturer specs then you should be alright

[xg32]

1 hour ago, Smackaroy said:

I'm just wondering why my cpu is performing slightly slower. Could bios up dates do this? 

 

Edit. It could also be slightly hotter in my room now its summer meaning a bit slower clocks too.

my 2600k did degrade by 100mhz after alot of years at 1.4v, but if you are pushing an optimized oc at a lower voltage...it's alot less likely, if possible at all.

[da na]

I have two LGA 2011 Xeons, they haven't noticeably degraded. They never turbo to their max 3.8 anymore but don't know why that is.

[Henry Butt]

Turboing all the time can degrade performance over time.

[Smackaroy]

1 hour ago, valdyrgramr said:

Heat, if liquid coooling you might need to replace the coolant, coolant degrades in AIOs over times, environment, sw, damage, paste degrading overtime, swapping paste for a worse one, bad spread after a repaste, etc.

The pc is about 6 months old

[accolade]

Check your core voltage

 

I'm noticing my motherboard is not as stable with the VCORE voltage for my 2990WX.  At times it will drop to 0.79V which is not nearly enough and often lead me with cores running at 600Mhz. 

 

 

[Smackaroy]

I'm going to try not to dwell on it considering it's not even a 2 percent drop in only multi core, (single core still performs the same) and I'm not going to notice it unless I do a benchmark which I not normal use.

[Maticks]

Usually find if you overclock at high voltages (1.45V) after several years of operating at that voltage the overclock is not stable and requires more voltage to hold.

Since you are at a limit already of heat your only option is to downclock it 100Mhz and find the next stable overclock point getting a down step of voltage as well.

 

[freeagent]

My 3770K needs a lil more voltage than it did when I first got it.. and I lost 100MHz off the top of my OC. Could be mitigations, could be degradation.. I am not sure.

[Guest]

Interesting topic.

 

Chips can degrade frequency. Typically from long durations on overclocks. I've seen guys with 10 years with 5ghz Intel clock suddenly become unstable and forced to a lower operating frequency. 

 

It's not often you read something like that, but there are some that run rigs that long.

 

From my experience, as little as it may be, a CPU fresh from the box will hit its maximum clock on liquid nitrogen. I have seen degradation nearly immediately, while obviously using abnormal cooling scenarios.

 

That said, degradation really comes in different flavors. Abuse being a key factor in shortening the 100,000 hours MTBF (mean time before failure)

 

When it comes to your processors of current time 2020/21, these chips are already overclocked to the maximum current potential before degrading the MTBF rating to a point someone would notice..... maybe 10 years down the road.

 

Fact of the matter is, the counter is ticking while the PC is on. But also know the MTBF is at the rated load of factory specifications, not idle.

 

TLDR - degradation depending on how you use this word in its context, yes a cpu is "degrading" for each hour it has run. But that is a count down on the hours it was built and intended to run for. 

 

But why worry about just the cpu? What's the failure rate on memory? Board, PSU and so forth? 

 

If your electronics use only solid capacitors, like your mainboard, those are 100,000 hours. Theoretically, roughly 10 years. And hopefully without issues. 

[accolade]

Nowadays the motherboard is most likely to be the component that degrades or show signs of defects that will effect every other component especially the CPU.

As it has in my case, which is why I just opened a 3rd RMA on the MSI X399 MEG Creation motherboard

[freeagent]

2 hours ago, accolade said:

Nowadays the motherboard is most likely to be the component that degrades or show signs of defects that will effect every other component especially the CPU.

As it has in my case, which is why I just opened a 3rd RMA on the MSI X399 MEG Creation motherboard

That’s just MSI being MSI. 

[Somerandomtechyboi]

On 7/14/2021 at 10:13 AM, ShrimpBrime said:

That said, degradation really comes in different flavors. Abuse being a key factor in shortening the 100,000 hours MTBF (mean time before failure)

I abuse the shit out of my poor 45nm core cpus and they still havent degraded, ive even run 2v on them and one time even accidentally booted to windows 7 with 2v, still doesnt seem like they have degraded just yet or only minor degradation

[Beat_my_Laptop]

Yes, eventually all electronics break down but usually obsolescence occurs before they naturally degrade noticeably.

 

But, your cpu performance may degrade over time because of because of other factors:

-Dust in your heatsinks, air-coolers, or radiators

-accumulation of programs (from your apps, games, programs, or OS)  that take up background CPU power

-forced obsolescence, I doubt any PC companies do this with their bios updates but Apple does

-accumulation of failed transistors, shorts, heat damage, & dangerous overclocking damage (if you do that). These are small but add up overtime & in conjunction with the above

[Guest]

19 minutes ago, Somerandomtechyboi said:

I abuse the shit out of my poor 45nm core cpus and they still havent degraded, ive even run 2v on them and one time even accidentally booted to windows 7 with 2v, still doesnt seem like they have degraded just yet or only minor degradation

45nm cpu.... what cpu?

 

I've done 2v on processors.

 

Can safely say most motherboards of mid to low quality don't even go up that high.

 

--

 

Then it's "have not degraded"

But 

"Doesn't seem like"

Then 

"Only minor degradation "

 

I'm skeptical you've ever run 2v. The board on ambient cooling probably would have shut off right away, no less that happening at 1.7v.

[freeagent]

OCP is shortly after 1.65v with 32nm, 2v.. iirc..  haven't done that myself since s462.

[Somerandomtechyboi]

12 hours ago, ShrimpBrime said:

45nm cpu.... what cpu?

 

I've done 2v on processors.

 

Can safely say most motherboards of mid to low quality don't even go up that high.

 

--

 

Then it's "have not degraded"

But 

"Doesn't seem like"

Then 

"Only minor degradation "

 

I'm skeptical you've ever run 2v. The board on ambient cooling probably would have shut off right away, no less that happening at 1.7v.

I think the first time this happend i was spamming del to get into the bios so my board cant shutoff, i was quite suprised to see it running at 2v or was it 2.03v or something like that, def panicked cause i know somewhere around 1.9v is the death zone for these 45nm cpus but i think the board shutoff or did i just unplug the thing i forgot

 

Second time was also accudental prob similar to the first time

 

Third time i booted to windows thinking i was running 1.7v but when i checked cpuz it was running at around 2.01v and then it shutoff

 

 

The poor cpu that experienced this was my pentium e6700 which still works btw, just minor degradation i think

 

And yes my board has an auto shutoff at 1.9v and a warning of high voltage at 1.6v though those two can be bypassed by turning off the volt monitoring in the bios

 

Btw its a p5q pro which ive modded with a p5q pro turbo bios

 

And yes since this is completely accidental i was running a cheap chinese tower cooler cause i have nothing better

 

Im currently running a g31m s2c cause my p5q boards seem to murder all of my 500w oem psus that i got locally for 3$, this board only goes to 1.6v so overvolt isnt a concern, im just concerned about the what looks to be 3 phase vrm on it but ive put on heatsinks on them and ive run 4.2ghz 1.4v and seems to be ok