Sign in with
Sign up | Sign in
Your question

life expectancy

Last response: in CPUs
a b à CPUs
March 23, 2001 6:12:32 AM

Can anyone tell me an approximate life expectancy on the intel and amd chips?

"If I had a golden retriever, I would name it Silence, because Silence is golden."

More about : life expectancy

a b à CPUs
March 23, 2001 1:08:33 PM

What speeds are you talking about? and what conditions?

a heavily overclocked cpu can sometimes last only 3 or 4 years, but a normal stock CPU (no matter the brand) will usually last indefinately... it can be many many years depending on the conditions it's in and how it's cared for (eg... you gotta keep it cool and not crush it!!!)

No one can really give a good estimate coz no-one really keeps a CPU for as long as it can last, 10 years would make it completely dead... it's only ten years since computers were still room sized... So think of it in that manner and see if it still matters to you... = ]

<b><font color=blue>Note to self: Never buy <i>anything</i> without checking with <font color=red>Tom</font color=red> first...
a b à CPUs
March 23, 2001 2:22:00 PM


CPU in the hands of a moron
its life expectancy is inversely proportionnal of the degree of stupidity of that moron

But in the hands of a sane and intelligent person it can last for a long time in the prime condition ( heatsink correctly in place, casing correctly ventilated, etc.. )

(overclocked cpu aren't lasting necessarily a long time.. depends mostly if the condition are prime or not ).
Related resources
March 23, 2001 2:51:19 PM

"it's only ten years since computers were still room sized"

UMMMMM I had a IBM PCjr back in 85 the size of what I got now.

March 23, 2001 3:14:28 PM

I would say a AMD t-bird with it's fragile/gay/oc'd core would last about 1-2yrs maybe less(some die in < 2secs), Intel cpu's would last 5+

"Amd cpu...Gone in 2 secs flat, it truly is a fast chip!"
March 23, 2001 3:40:44 PM

I seen a few AMD's posts claiming under 10 seconds.

AMD wins!

Keychains anyone?
March 23, 2001 3:46:19 PM

He does have a point you know if you fry your chip from over clocking you are not advanced but a stupid idiot that’s why the product is at that speed by design specifications. You use a product by specifications and keep good upkeep on it will last for almost ever I bet your power supply will go before the chip does. Oh well each to their own.


<font color=blue>Just some advice from your friendly neighborhood blue man </font color=blue> :smile:
March 23, 2001 4:10:59 PM

The question should be "whats the life expectancy of a HSF"

Answer: about a year depending on environment.

AMD is dead without HSF period. Intel will survive if HSF fails.

What makes this topic choice is AMD users will neglect to check/replace the HSF so we can expect more AMD keychains over the next few months to a year =)

<P ID="edit"><FONT SIZE=-1><EM>Edited by fugger on 03/23/01 01:11 PM.</EM></FONT></P>
a b à CPUs
March 23, 2001 4:58:44 PM

>The question should be "whats the life expectancy of a HSF"
>Answer: about a year depending on environment.

Ah, give me a break...I've seen 1 CPU fan failure in 12+ years of working with CPUs requiring fans. And that was on a 486. I've seen far more power supply and drive failures then CPU fan failures. In fact, we've got about 20 PCs in this office alone that are about 5 years old, and the tech sitting next to me says he doesn't remember and HSF failures on any of these boxes. And these are just crappy sleeve-bearing oem fans. As for environment, I've seen some that looked like they had a gray carpet on the blades they were that dusty.

Yea, and AMD CPU is probably toast if the fan dies (if it has a good heatsink, a bios moniter may catch it before it cooks), but that doesn't happen very often. Certainly not nearly as often as you state.

Oh yea, one of the boxes I work on has dual CPUs & 2 fans on each HS. So, I should expect to replace a HSF every 3 months on that box? Really! I'm way overdue then @ 2 years of continuous operation.

In theory, there is no difference between theory and practice.
In practice, there is.
March 23, 2001 5:05:38 PM

could be so, 10 seconds it quite remarkable, working at 100 deg celcius, isnt it?

assuming the chips are well setup, with proper cooling, not overclocked they might live forever, at least they would survive their warranty periods. heavy working on it might reduce the life of the cpu, and on this front Intel chips would live longer than AMD ones I guess.

As for overclocked chips, life reduces exponentially wrt oc ratio, and that might reduce to a couple of years from 10+ yrs, or even a few hours if you take it too far. fugger would like to do such a feat with AMD chips, and in such case these cpus are totally at the mercy of their users.
isnt it so?


die-hard fans, do not have have heat-sinks!
March 23, 2001 5:14:25 PM

George, take a few mins and read thru the forum, lots of keychain posts =)

And as far as your bios saving your CPU if HSF fails... try it =) better yet, ask grizely 1 if his "BIOS protected mobo" saved his AMD cpu from near instant (2 seconds?)death. Grizely1 was in the same boat as you thinking that BIOS will save your ass.

By the time you hear the beeps its too late. dont believe me?? try it. go on, dont be a chicken.

AMD being so critical on temp, how much debris does it take to cause it to fail? reduction in RPM will definitly increase CPU temp. If your CPU is running at 115F how many more degrees will you start to see anomolies? At what temp does perminant damage take place. maybe not complete core meltdown just traces breaking inside the core(hardest to diagnose).
March 23, 2001 5:38:11 PM

<font color=blue>"The question should be "whats the life expectancy of a HSF"

Answer: about a year depending on environment.

AMD is dead without HSF period. Intel will survive if HSF fails."</font color=blue>

More Fabricated Fugger Facts (FFF for short). Oatmeal technician.

Where are the statistics that show HSF will fail in a year? Which site has posted this? Just another "Fugger Fact", no proof, a made up statement to justify his Intel fanaticism.

Through the years I have had dozens of PC's in my office and have never had a ball-bearing HSF unit fail. The oldest machine is 6 years old, and its HSF, though dust covered, still functions properly.

True, non ball-bearing HSF units will fail, and I have had a couple of those fail (1 on an Intel Pentium system, it fried despite the thermal protection). But all approved HSF units for the socket A are ball-bearing. They will just about last forever.

Fugger, the hottest running chip around is the P4. Without the HSF it will fry. Here are some quotes from Maximum PC, April issue:

<font color=blue>"The Pentium 4's massive power consumption translates into increased heat. To dissipate the excess heat, Intel is pushing larger, copper-based heatsinks. Because copper is heavier then the aluminum used in traditional heatsinks, Intel recommends bolting the heatsink to the case itself. This keeps the heatsink from tugging on the proc--and bending the mobo--or falling off,</font color=blue><font color=red> leaving the Pentium 4 to burn to a crisp.</font color=red>

Later in the article they say:

<font color=red>"LESSON:</font color=red><font color=blue> It's hot! Damn hot! Even the mightiest of Pentium III and Athlon heatsinks can't stand up to the sizzling capabilities of the Pentium 4".</font color=blue>

Sounds like Intel copied something from AMD after all....
March 23, 2001 5:40:02 PM

well, AMD manuals say duron/athlon can handle upto 90 deg. although the typical temp should be less than 50.
i guess temp getting above 90 will do strange things to your pc, might crash, hang, generate GPFs or strange errors. if this is what you are having at any moment, immediately switch off the pc and let it rest for a few hours. in the meanwhile, go to the store and get a good HSF.

higher temps will start wearing the traces on silicon, and they wont break immed, altering their properties permenantly and this might induce anamolies in the normal fuctioning of the cpu. its now liable for replacement. or, for example, a crack developing in the mmx circuits might let you run the cpu with non mmx apps, but it will fail the moment any mmx instruction is run. you wont detect it until you run an mmx app, and will still wonder as to why such a perfectly working chip errs here!

btw i dont have any athlon/duron keychains, but have a lot of Cyrix M2 ones, anybody interested???


die-hard fans, do not have have heat-sinks!
March 23, 2001 5:49:32 PM

Bigguy, test your theory and get back with us.

You will find out the P4 will not hurt itself.

Oatmeal is good for you.
a b à CPUs
March 23, 2001 6:03:12 PM

Ok, I thought long and hard about including the line about the bios saving your CPU, cuz I just knew that would be the part you'd jump on out of the whole post. No, I don't have any data, and no, I'm not gonna try it. But I still think that if you're not overclocking and you monitor fan rpm & cpu temp you should have a reasonable chance of catching it. But yes, a thermally protected CPU would be better.

Now, why don't you address my main point. You are tremendously overstating the failure rates of heat sink fans. By your reckoning, I should be replacing a fan about every 3 months on the dual cpu box with 2 fans per cpu. And yet it's gone 2 years, nonstop.

In an office with 30 PCs, you would expect a fan failure about every 2 weeks.

That 96 CPU Boeing cluster we talked about last week should expect a failure every 4 days on average?

What about those big clusters? ASCI Red, with 9536 CPUs should be killing fans at at rate of about 1 every 55 minutes by your estimate. Lets see, thats about 9500 fans a year @ $20 each, plus a full time employee just to install them. That'll add a chunk to your maintenance.

And that 50000 hour MTBF routinely listed for HSF units? They've overstated that by more then a factor of 5? Wow, somebody call 60 minutes!

My point is not that AMD cpus don't need thermal protection. They do. My point is that you consistently shoot your mouth off without the vaguest hint of a clue. Which doesn't lend itself to credibility.

In theory, there is no difference between theory and practice.
In practice, there is.
a b à CPUs
March 23, 2001 9:00:10 PM

I mean fast computers, not everyone had a desktop..

There were still servers in government buildings that had a computer that took up the whole room... In 89 i had a desktop too, but you said 85, there's 5 years of progression we're talking about...
THink about it...

<b><font color=blue>Note to self: Never buy <i>anything</i> without checking with <font color=red>Tom</font color=red> first...
March 23, 2001 11:47:49 PM

Yes. I believe everyone should buy ntel processors too. And support Rambus and Microsoft. Boycotting those un-wholey Linux users should be part of the licensing agreement.

I have had 3 ntel systems, and all of them were excellent for the first 6 months. Along those lines, the life of a processor will outlast it's ability to outprocess the competitor. The 21st century is awesome like that. Things are obsolete/ancient after 3 months.
March 23, 2001 11:49:41 PM

Finally someone that sees the linux community as evil. Bah everything freeware, that entire idea is a loose loose situation. Programmers have to feed the kids you know.


<font color=blue>Just some advice from your friendly neighborhood blue man </font color=blue> :smile:
March 24, 2001 1:08:55 AM

We will need an Intel owner to do that. I will never own a P4. Perhaps a future Intel product, if it offers the best price/performance. I will be upgrading my Son's Celeron 450 in a few months. If Intel has a chip that performs better then the AMD counterpart, and is less expensive, I will buy it. In the meantime, it looks like I will get him a Duron 800.
March 24, 2001 2:02:02 AM

Hey dip-[-peep-], my asus a7v saved me when I forgot to plug in my fan on my AMD 1Ghz gues what no key chain. You have nothing of use to say. You are a LOSER!

The question is not how long the cpu will last but how long it will be before you want to upgrade.

Opinions are like assholes, everyone has one and most of them stink
a b à CPUs
March 25, 2001 2:00:16 AM

Yo Fugger,
Please tell us where you get your crappy heat I can make sure I NEVER buy from them. Seriously, if you arr only getting 1 year out of your fans, you're doing something very wrong or buying complete [-peep-].
March 26, 2001 4:59:30 AM

get good heat sinks from taisol, coolermaster, alpha etc.

btw, i just had a call from a friend about his AMD K6II 400 crashing often. He tried to scan the disk, try installing win98 and what not. The computer was on all the time until I reached his plalce, for about 5-6 hrs.

And first thing I did was to open the chassis and check whether the HS/F is running. I wasnt!! Just as I touched the heast sink fins my fingers got burnt, leaving the marks of the fins on them. I guess it was about 90-95 deg. celcius.

Now the machine dint even boot, I thought the CPU is dead, and I took it to my place to checkout on my K6II machine.

And would you believe it, this <b>AMD</b> cpu, running at 98 deg.C, for almost 5 hrs, was still working!!!

I ran a few tests from BCM diags and SiSoft Sandra, and found no problem with that.

Just read a post of somebody saying a CPU lasing for 10 seconds. This AMD K6II almost went through the hell for 5 hrs and came back absolutely ok.

fugger what do you say? :-) I can send you its pic so that you could put it as your wallpaper.


<font color=blue>die-hard fans don't have heat-sinks!</font color=blue>
a b à CPUs
March 28, 2001 3:47:04 AM

The life expectancy of a processor is an irrelevant number for the simple reason that technological advancement renders them obsolete before they die. That includes overclocked systems (which begs the question.. Why the hell not ?)

This is even more true today, as the rate of change in processor (and system) performance has become a STEEP curve ! YeeHaw !!

The awesome result of this steeper curve is the availability of remarkably fast processors.. CHEAP !

Now, many of these crusty old THG members will remember tragically slow processors of ye' olde tymes .. even 33MHz seemed fast for some of us at the time. Remember those VESA local bus graphics cards ?? oooo, they were cool.

So, maybe I'm easily pleased. Still, I must say that I can't wait for the reality of 266MHz DDR memory systems to compete with RDRAM based systems to force the price of even higher performance systems down to my Happy price zone.

The other consequence of all of these LOW prices is the luxury of being able to choose moderately priced hardware to go fast with STABILTY, and the simplicity of BROAD-BASED COMPATIBILITY. After all, speed is only satisfying if you get there.

So, who's complaining ??
March 28, 2001 6:04:59 PM

It was a pair of "Antec" HSF. They were preinstalled before shipping. (Asus mobo no beeps pre2001 mobo w/bios patched)

I would have to dig around for my keychains. I gave one to one of the receptionists and the other is here somewhere.

I believe Grizely1 has a few he can show off to us.
March 29, 2001 1:26:58 AM

Actually Fugger- ever the fool- is wrong yet again. I have a t-bird 800 clocked to 1000 and I have stopped the fan for 10 minuites (wondering if that burn in idea actually works)and it got noticibly hotter, but did not melt or harm the chip in any way. It got about 74-75C with just a fop32 heatsink and no fan. Anyway if the fan stops on an abit board the board shuts down in plenty of time as long as you have a correctly installed heatsink even without a fan. Also if the temp starts to rise during a windows session, the chip will either lock up or the temp program will kill the computer long before you do any damage. The only real threat for quick distruction is if the heatsink is on wrong.

a b à CPUs
March 29, 2001 1:31:19 AM

Ah, you realize that you are talking about making keychains from chips that were either overclocked or abused (no HSF attached). I don't think those situations can be applied the the question of life expectancy. You can't expect a chip to dump heat without a good thermal conduit to a decent heatsink, and overclockers (like me) are taking the risk. I've yet to burn a chip.. I might add. And ALL of my PIIIs are running well above spec. The worst is 100>120FSB (20% over spec) and the best is 100>140 (40% over spec). And I can replace them all for less than $125. At these prices, I am not worried about it.

I don't have any burnt chips, but I do have an intact 486DX66 "overdrive" processor and a 33DX processor that were retired after a few years of service.
March 29, 2001 2:51:11 AM

hey, with all the keychains that you have to go through, it sure sounds like you've purchased a lot of AMD's and didn't install them right. Learn how to install them right and maybe you won't break the ignition switch in your car because your keys are so heavy with your burnt chips.(schmuck)

:tongue: <font color=green> I LOVE INTEL. It tastes like chicken </font color=green>
March 29, 2001 4:47:03 AM

That seems like it was really smart idea!

“Build your own you will love it more”
March 29, 2001 6:55:07 AM

Well... talking about AMD CPUs, I have one Slot A 750MHz that nearly died because I didn't install the HSF properly. THe computer hanged half way, then I realized that something is seriously wrong. I open up and I can literally feel the heat gushing out. Then, with mutant fingers, I quickly pulled the half-done steak out of the computer and place it outside and use a table fan to blow it. Fortunately, when it cooled down and I install back, it boots normally. That was close.

This site is cool.