How long will an AMD 1.2ghz OCed to 1.3-1.5 last?
A year? two? three? That is, how substantial is CPU life reduction with overclocking?
-eric
A year? two? three? That is, how substantial is CPU life reduction with overclocking?
-eric
how much CPU life is killed when you OC?
- Thread starter eLiu
- Start date
with cpu speeds changing so quickly.. if it made thechip only last 2 years there will probably be a new core/tech to upgrade too... take my p2 for example.. its only 1.5 years old.. if i oc which i have and take a 1 year off.. lets say its 4 year life.. then i still got 3 years... so 1.5 years left. and look wehre the p2 is now..
You know, everyone says that OCing decreases CPU life. But, I have never seen this documented. People say that it might take a few years off the 20 year life span of a CPU. Of course, overclocking really wasn't going on 20 years ago so the data really hasn't come in. It is my belief that if one doesn't increase voltage, OCing does not shorten the life of a processor. You are in fact only running the processor at the speed for which it was meant. But who cares since "I pitty the fool" who is running one of today's processors 20 years from now.
Also, upping the voltage too shouldn't really decrease life span since AMD and Intel have been doing this whenever they realease a new processor in a series with a higher voltage
.
Also, upping the voltage too shouldn't really decrease life span since AMD and Intel have been doing this whenever they realease a new processor in a series with a higher voltage
.Never seen it documented? You can ask PM....he doesn't come here anymore due to flamers, but he works at Intel, and did a lot of research into overclocking its effects longer term. If the Hightly Technical board comes online, you'll be able to talk to him here, but otherwies, you'll have to email him. Do a user search for him, if you want his email address.
It does indeed lower the life expectancy of chips. The only matter is by how much. One way they can die is from overheating, another electromigration (where the electrons litterally strip the CPU of silicon).
EDIT, also, increasing the voltage is a BAD thing. It increases wattage approximately proportional to the square of the voltage. Increasing voltage can sometimes increase stability, but also remember what voltage is - electrical potential.
AMD and Intel have tolerances and specifications for a reason - sometimes you can go beyond them, (and yes, I do too, I overclock sometimes) and sometimes you can't. Going over the clock rate isn't usually a terribly bad thing, but going over the voltage spec (which, I believe, tends to be +-10% of the spec'ed voltage) tends to be a bad thing. AMD and Intel don't go over their initial spec voltages for their CPU's
For example, if the speced voltage for a chip is 1.65v, the later CPU's don't usually go over ~ 1.75v. Later revisions of the chip might, but a revision is technically a different chip (not by a lot, usually), but it will have different tolerances, and thus have different spec'ed ranges of operation.
My 2 cents.
It does indeed lower the life expectancy of chips. The only matter is by how much. One way they can die is from overheating, another electromigration (where the electrons litterally strip the CPU of silicon).
EDIT, also, increasing the voltage is a BAD thing. It increases wattage approximately proportional to the square of the voltage. Increasing voltage can sometimes increase stability, but also remember what voltage is - electrical potential.
AMD and Intel have tolerances and specifications for a reason - sometimes you can go beyond them, (and yes, I do too, I overclock sometimes) and sometimes you can't. Going over the clock rate isn't usually a terribly bad thing, but going over the voltage spec (which, I believe, tends to be +-10% of the spec'ed voltage) tends to be a bad thing. AMD and Intel don't go over their initial spec voltages for their CPU's
For example, if the speced voltage for a chip is 1.65v, the later CPU's don't usually go over ~ 1.75v. Later revisions of the chip might, but a revision is technically a different chip (not by a lot, usually), but it will have different tolerances, and thus have different spec'ed ranges of operation.
My 2 cents.
hey.. dont get too technicakl here.. who cares if it is bad.. do u still use CPU that u bought 4-5 years ago??? I know a lot of poeple dont... unless they just use it for word processing... come on now.. who wants dx2-66.. or even a pentium 2...
My idea is to juice the life out of the cpu...
My idea is to juice the life out of the cpu...
Duron CPU's are so cheap right now (<$30) that if I burned mine I would order a replacement and wouldn't sweat it. Actually I might just buy a spare CPU just in case...
I agree with the sentiment of this thread: it's just that the effects are true, chips can and will die sooner if overclocked, unlike what was previously said.
Will I be using my Athlon in 4 years? Well, considering my current budget, and the fact that I'm in school, I may!! If I had my way, I'd probably do a major upgrade at least every 6 months, but there ARE people who need to use their sole computer, long term. It just doesn't happen to be you guys. I don't get why everyone dismisses what I say. I still know people who use 486's as their main/ONLY machine! Some people have to be a bit more frugal than ya'll....I hope I don't have to be
My point being, have I overclocked? Sure I have. Do I recommend it to some people I know? Sure do. But you just can't dismiss the truth. But, considering that I was getting "too technical," I guess I'll refrain from responding to people. Wooh.
Will I be using my Athlon in 4 years? Well, considering my current budget, and the fact that I'm in school, I may!! If I had my way, I'd probably do a major upgrade at least every 6 months, but there ARE people who need to use their sole computer, long term. It just doesn't happen to be you guys. I don't get why everyone dismisses what I say. I still know people who use 486's as their main/ONLY machine! Some people have to be a bit more frugal than ya'll....I hope I don't have to be
My point being, have I overclocked? Sure I have. Do I recommend it to some people I know? Sure do. But you just can't dismiss the truth. But, considering that I was getting "too technical," I guess I'll refrain from responding to people. Wooh.
I dont' think OC is "necessary" unless you HAVE TO!
Say, you're getting bad performance watching that DivX, or getting unacceptable framerate with your fav. FPS. If you're encountering that, it means your CPU's near its usefulness, at least to whatever you do with it. Then you might as well OC, and if it burns out, you need a new one anyways. But just letting you hang on to it longer let you buy a better CPU at a better price.
And OC just to show some benchmark, or testing to see how high you can go? that's really what a hardcore "hobbist" would do. I only up the FSB when I feel lucky.
Say, you're getting bad performance watching that DivX, or getting unacceptable framerate with your fav. FPS. If you're encountering that, it means your CPU's near its usefulness, at least to whatever you do with it. Then you might as well OC, and if it burns out, you need a new one anyways. But just letting you hang on to it longer let you buy a better CPU at a better price.
And OC just to show some benchmark, or testing to see how high you can go? that's really what a hardcore "hobbist" would do. I only up the FSB when I feel lucky.
I think, taking everything into account, that the sort of people who overclock are pretty unlikely to be the sort of people who will settle for a 4 year old PC. So it's not a major issue, is it?
It really doesn't matter.
The CPU that cost you $200 today, will only cost $30 in 2 years.
Even if you are stuck running the same system in 2 to 5 years, and somehow the CPU dies, a replacement will only be a little more than shipping.
My first real CPU, 486 DX2-80, was a steal at $160.
Duron's have dropped into the $30 range.
1GHz T-birds were around $300, and now around $100.
If my Duron 750 @ 1GHz dies in a year, I would be HAPPY to pop in a 1.4GHz CPU for $40.
BTW, overclocking by 10% to 15% isn't really overclocking, as the chips have that much tolerance built-in. It's when you push 25% to 33% that you really are overclocking the thing, and pushing it past what it was designed to be able to do.
The only time overclocking is NOT nessasary is when you can afford a CPU that does the speed you want. But if you can not afford the CPU, there is nothing wrong with spending 1/2 to 1/4 the price, and getting a slower CPU, which easily overclocks to the intended speed. $35 for a 1GHz is hard to beat. The whole point of overclocking is not speed, but is instead being frugal. Getting the most speed at the lowest end cost.
The CPU that cost you $200 today, will only cost $30 in 2 years.
Even if you are stuck running the same system in 2 to 5 years, and somehow the CPU dies, a replacement will only be a little more than shipping.
My first real CPU, 486 DX2-80, was a steal at $160.
Duron's have dropped into the $30 range.
1GHz T-birds were around $300, and now around $100.
If my Duron 750 @ 1GHz dies in a year, I would be HAPPY to pop in a 1.4GHz CPU for $40.
BTW, overclocking by 10% to 15% isn't really overclocking, as the chips have that much tolerance built-in. It's when you push 25% to 33% that you really are overclocking the thing, and pushing it past what it was designed to be able to do.
The only time overclocking is NOT nessasary is when you can afford a CPU that does the speed you want. But if you can not afford the CPU, there is nothing wrong with spending 1/2 to 1/4 the price, and getting a slower CPU, which easily overclocks to the intended speed. $35 for a 1GHz is hard to beat. The whole point of overclocking is not speed, but is instead being frugal. Getting the most speed at the lowest end cost.
Raising the voltage on a CPU increases heat production. Someone recently (I can't remember who) published a finding that said that for every 10 degrees Celsius your CPU temp increases you lose aprox. 1/2 the life of the CPU.
Personally I don't really care. I usually upgrade my CPU/mobo once per year, and I've never seen a CPU die that soon.
Personally I don't really care. I usually upgrade my CPU/mobo once per year, and I've never seen a CPU die that soon.
And what about the others components of a computer? For example, if I raise the frequency of ASUS A7V133 to 145MHz or even 150MHz, can I damage it? And what about hard drives, etc...
Your other components will be just fine.
Up to 133MHz uses 1/3 for PCI and AGP is 2x PCI.
Above 133 uses 1/4 for PCI and still AGP is 2X PCI.
So 145MHz fsb is only running PCI at 36.25MHz, AGP at 72.5MHz.
and 150MHz fsb runs PCI at 37.5MHz with AGP at 75MHz.
Video card and harddrive specs allow 75MHz, (as some PC's ran at 75Mhz without being overclocked), and many allow more.
75MHz was standard and even 83mhz was allowed for years, even when manufactures made crap.
Many Cyrix CPU's required you to run FSB at 75MHz, which back then used 1/2 for PCI, so today's FSB running at 150MHz, is no different on your PC's parts than when it was a factory setting.
Up to 133MHz uses 1/3 for PCI and AGP is 2x PCI.
Above 133 uses 1/4 for PCI and still AGP is 2X PCI.
So 145MHz fsb is only running PCI at 36.25MHz, AGP at 72.5MHz.
and 150MHz fsb runs PCI at 37.5MHz with AGP at 75MHz.
Video card and harddrive specs allow 75MHz, (as some PC's ran at 75Mhz without being overclocked), and many allow more.
75MHz was standard and even 83mhz was allowed for years, even when manufactures made crap.
Many Cyrix CPU's required you to run FSB at 75MHz, which back then used 1/2 for PCI, so today's FSB running at 150MHz, is no different on your PC's parts than when it was a factory setting.
AGP isn't 2x PCI. It's determined by the chipset, and that's 1x @ 66mhz, 2/3x @100mhz, and 1/2x at 133mhz on boards that officially support 133mhz. While the PCI is determined elsewhere on the board, the AGP is determined by the chipset, and no currently available chipset has support for more than 1/2x FSB. The reason I mention this is because while with 1/4 PCI, it is effectively "2x PCI", when you jump up to a 1/5 PCI, it's a different story, and it's not 2x PCI.
I didn't know of a board that uses 1/5.
If one did, at 166MHz, the 1/5 for PCI and AGP at 2X PCI works.
I mention 2x PCI to keep it simple for most, instead of having to know 2/3, 1/2 and 1/1.
If one did, at 166MHz, the 1/5 for PCI and AGP at 2X PCI works.
I mention 2x PCI to keep it simple for most, instead of having to know 2/3, 1/2 and 1/1.
No, my point being, at 166mhz FSB, and a 1/5 PCI divider, you imply the AGP would be 66Mhz. It would, instead, be 166/2, or 83Mhz.
Good point about the cyrix chips though - I had three of 'em.
Good point about the cyrix chips though - I had three of 'em.
ahh...ok thanks for all the info guys
anyway...i want my athy to last until at least the end of next year...it's hard for a 15yr old w/o a job to pop out another 200 (mobo/cpu) every year...hehe (case in point: current cpu: p2 266...dont laugh...it's not funny)
and beyond that...my parents want my cpus to last even longer than that...cuz my used parts get passed on to my sister > hehehe
anyways, another user here told me that OCing works best if you OC early, even if you dont leave it at that speed, i guess it makes the cpu happy lol
soo...i s'pose ill OC it in the beginning...then bring it back down to (near) stock speed...maybe as high as it'll go w/o having to increase temps/voltage much. i mean 1.2ghz w/gf2gts should be plenty good for games
thanks again y'all,
-eric
anyway...i want my athy to last until at least the end of next year...it's hard for a 15yr old w/o a job to pop out another 200 (mobo/cpu) every year...hehe (case in point: current cpu: p2 266...dont laugh...it's not funny
)and beyond that...my parents want my cpus to last even longer than that...cuz my used parts get passed on to my sister >
heheheanyways, another user here told me that OCing works best if you OC early, even if you dont leave it at that speed, i guess it makes the cpu happy lol
soo...i s'pose ill OC it in the beginning...then bring it back down to (near) stock speed...maybe as high as it'll go w/o having to increase temps/voltage much. i mean 1.2ghz w/gf2gts should be plenty good for games
thanks again y'all,
-eric
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 24K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 19K
-
-
-
Intel processors crashing Unreal engine games (and others)- Started by BFG10K
- Replies: 2K
Facebook
Twitter