Anandtech: The truth of CPU degradation

[maverick7]

nice post, very informative.

 

[Mathos]

I assume that you are referring to electromigration.

ah yes, was having a brain fart. Which lead to the derailed train of thought.

 

[yomamafor1]

He misspoke and for some reason you are having a hard time understanding what he said. You actually posted the proof below that you were asking him for. What he meant to say was "the Hafnium high K gate CPUs". Intel uses Hafnium as the dielectric in the high K gate 45nm CPUs. Don't get me wrong, I'm not one to defend dragonsprayer, and he said stuff, in his post, that I'm pretty sure contradicts previous posts of his, but c'mon let us not split hairs.
Technology@Intel · 45-Nanometer-Hafnium-based-High- k-dielectric-Metal-Gate
And here is the proof you asked for, if you can call it proof. I didn't say it was not accurate just not proof.

These processors are built on a new 45nm High-K process that invariably makes them predisposed to accelerated degradation when subjected to the same voltages used with last-generation's 65nm offerings.

The same "proof" that you presented in the Anandtech quote earlier.

Thanks for the clarification. However, it is not what I was looking for.

He said, "the high k CPU degrades faster", which I find that statement baseless. Just like you posted, Anand simply said 45nm HK/MG CPUs are likely to degrade faster IF they are subjected to the voltage of 65nm CPUs. This means, 1.6V on Conroe is going to have different effect than 1.6V on Wolfdale, and Conroe is likely to less longer due to higher tolerance of electromigration at that voltage.

 

[Zorg]

What that means to me is that the 45nm is more susceptible to electromigration at a given OC/over volt than it's 65nm counterpart. therefore at a high OC/over volt one needs to be more concerned with good cooling etc. Anyway, I don't want to be known as the guy that agreed with dragonsprayer, because he certainly sprays a lot of BS. In this case I don't think he was that far off the mark. I think it was more of a misunderstanding, and a little splitting hairs.

 

[yomamafor1]

I don't think you should couple overvolt and overclock together, although they are related. Anand simply said 45nm CPUs are more susceptible to electromigration when overvolted, compared to the 65nm. However, Wolfdale certainly needs less overvolting to achieve the same overclock.

I agree its was more of a misunderstanding on our part. However, I want DS to know that 45nm CPUs do not degrade faster. They will if you give them the same amount of overvoltage as on the 65nm parts, but in terms of the voltage needed for the same overclock, there's no evidence that would suggest they degrade faster.

 

[Zorg]

I don't couple OC and over volt. I have a Q6600 at a mild OC of 3G and I have reduced the voltage to 1.275 from the VID of 1.3, showing 1.232 at idle in cpuz with C1E enabled. I was saying that I got the feeling that Anandtech was more concerned with the OC and over volt as it related to more aggressive OCing, with regard to the differences between the 65nm and the 45nm CPUs

I do understand the knee-jerk reaction to jump on him, based on his previous posts, I feel it myself. But you need to give him a break on this one.

 

[yomamafor1]

I guess. I agree that we're simply playing with word's definition, while we may virtually arguing on the same side.

 

[Zorg]

Cool, but next time he goes off the wall, which will happen, then..... :lol:

 

[endyen]

So its the mobo not the chip. Love it.

Absolutely wrong. Intel is withholding tjunction info. That is the reason the chips do not show temp.
Intel should incorporate the tjunction in the DTS.

 

[endyen]

I agree with a lot of what you are saying, maybe not so much with how you say it.

1) use the best cooling possible and keep temps as low as possible
2) keep voltage as low as possible

3) the more your run a cpu at 100% output the more damage done to it from high end overclocking

Bolded part cannot be relied on. Many times your overclocked CPU can be suffering just as much damage when idle as when at load. Many of the fried Wolfdales died when near idle. Reasons:

1) Vcore is higher when idle and not using Speedstep; Vdroop occurs only during load
2) Temperature may not drop much when idle because CPU and system fans slow down

So 1) must take into acount the 2) keep voltage as low as possible.
With 2) the fans are controlled by temp, not core speed

When someone says they ran 1.6V for 10 seconds to bench SuperPi-1M, you can infer they probably used 1.6V for several minutes at least.

my perspective - "the sweet spot"

the sweet spot is the speed which is slight lower then where you see a greater increase in temps after a same increase in speed

Here I'd have to disagree and say it depends on the process and die size. A lot of the Wolfdale deaths were at the "sweet spot," which happened to be ~1.4V and 4+ GHz. Unlike with Conroe, the "sweet spot" is not a good indicator of what safe voltages are for Wolfdale because a 107 mm^2 45nm die doesn't produce a lot of heat to manifest diminishing returns on Vcore before you reach heavy electromigration.

Conroes stopped you from frying them because they got hot/unstable and made Vcore bumping futile. Wolfdales don't get hot till it's too late.

Wolfies can get a lot hotter, a lot more quickly.
Everyone should remember though that a slight overclock will at most reduce the life expectancy of a chip from ~ 10 years to about 6.

 

[Zorg]

Absolutely wrong. Intel is withholding tjunction info. That is the reason the chips do not show temp.
Intel should incorporate the tjunction in the DTS.

That's great, Intel the great evil As.h.les. I have a Q6600 G0 which has a Tjunction of 100C, afaik. It was updated from the B3 and ultimately we became aware of that. Let us consider that we have no right to inside information, and that it is given to us out of the goodness of their heart. Everyone has been overclocking the he!! out of the Intel CPUs, which is good, and they have benefited from the publicity. But you can't fault them for making the tjunction unavailable/confusing. The OC has become too easy, now it is not. Get back to basics. If you don't have the time or energy then pay Intel more money. That's better than them throwing away perfectly good chips so that they can bin them by speed, right?

@ All anti Intel Whiners: If you whiners keep pissing, then you will find that you get what you pay for, it's not like AMD is going to pressure them. I can assure you that the whining won't either.

Any responses from the whiners?

Edited to be softer.

 

[Scotteq]

<Reads Article, Reads Thread>

<Applies Common Sense Filter>


So:

(1) If you overclock, your processor will not last as long. Period, End Discussion.

(2) Maintaining low temperatures is certainly as critical, if not MORE critical than an "average" power user may suppose if said user intends on a long-term overclock.

(3) The harder you push for high benchmarks and big clock speeds, your processor will degrade ever faster than a similar setup which isn't pushed as hard. Therefore the trick is to not shorten the life of your Proc to less than your planned build/rebuild/upgrade schedule.

(4) On the newest processors there is less "insulation", and of a different type, being used. This apparently tends to increase said processor's sensitivity to temperature and voltage. Anecdotal evidence from (some?) of the people who have overclocked these indicates this insulation may degrade over time when the processor is run (much?) harder than it's rating.

(5) The lower multipliers on said newer processors force overclockers to push them (relatively) harder in search of their intended clock speeds. See point (4).



Seems to me these are old lessons being re-learned/re-stated.



From this I take:

(A) Members of the (imaginary) "Level~Headed Overclocker's Club" should remain at a pretty good price/effort/performance/lifespan, once the benchmarking guys finish spending their money to show the level-headed ones where that point is.

(B) Going for bigger benchmark numbers got harder with the new process/die size due to the increased sensitivity to overclocking plus the lower CPU Multis and higher FSB speeds.

(C) Going to liquid cooling sooner, rather than later is a Great Good Thing. Thankfully, this is easier and easier due to quality pre~configured and pre~built systems for this (Coolitsystems, etc...). Though your wallet won't be as happy when you're done.



Comments??

 

[zenmaster]

<Reads Article, Reads Thread>

<Applies Common Sense Filter>


So:

(1) If you overclock, your processor will not last as long. Period, End Discussion.

(2) Maintaining low temperatures is certainly as critical, if not MORE critical than an "average" power user may suppose if said user intends on a long-term overclock.

(3) The harder you push for high benchmarks and big clock speeds, your processor will degrade ever faster than a similar setup which isn't pushed as hard. Therefore the trick is to not shorten the life of your Proc to less than your planned build/rebuild/upgrade schedule.

(4) On the newest processors there is less "insulation", and of a different type, being used. This apparently tends to increase said processor's sensitivity to temperature and voltage. Anecdotal evidence from (some?) of the people who have overclocked these indicates this insulation may degrade over time when the processor is run (much?) harder than it's rating.

(5) The lower multipliers on said newer processors force overclockers to push them (relatively) harder in search of their intended clock speeds. See point (4).



Seems to me these are old lessons being re-learned/re-stated.



From this I take:

(A) Members of the (imaginary) "Level~Headed Overclocker's Club" should remain at a pretty good price/effort/performance/lifespan, once the benchmarking guys finish spending their money to show the level-headed ones where that point is.

(B) Going for bigger benchmark numbers got harder with the new process/die size due to the increased sensitivity to overclocking plus the lower CPU Multis and higher FSB speeds.

(C) Going to liquid cooling sooner, rather than later is a Great Good Thing. Thankfully, this is easier and easier due to quality pre~configured and pre~built systems for this (Coolitsystems, etc...). Though your wallet won't be as happy when you're done.



Comments??

#1) Agree - OCing Reduces the CPUs life. So does powering your computer on. What we don't know if it may reduce the life of the CPU from 20years to 19years or some other period.

#2) Disagree - This should only be an issue if the CPU is operating outside of the normal operating temperature. For Intel Chips this is about 72c. It's quite rare for even OCers to hit this value. Extra low temps may only help if you plan to use the CPU for more than 10 years. "Life Expectancy" of a CPU is only important if it may be lessened beyond the expected use of the CPU. There is nothing to indicate this is the case.

#3) Agree - But again, So long as you are operating inside of the Thermal and Voltage design specs this should not be an issue with replacement cycle. This is even more true for most OC'ers since those are generalyl the folks who need a bit of power. It's much less likely they will be keeping their PC's 5+ years, since they will need to upgrade to maintain their tasks. (Simple IE, MS Money, users still running Win95 are not your classic OC'ers.)

#4) Disagree - Lack of Evidence. They need to reference techinical documentation from Intel or another manufacturer to state this. However, I would agree that running a CPU outside of it's Temp/Voltage rating could seriously harm the CPU in a relatively short period of time.

#5) Disagree - This is not really a CPU issue as much as a motherboard issue. It's the Mobo more than the CPU that is pushed in these cases. Of course the more you push the FSB of a Mobo, the quicker it will die.


Overall, I agree with you.
My "Disagrees" are not strong disagreements, but more along lines of shades of gray.

The article also seems to fails to really address understanding a CPUs design specs and operating within that.

Example - an E2160 (1.8ghz Default with 800FSB) when it's Clocked to 3.0Ghz (1333 FSB) it is not likely to suffer any more than an E6850 (3.0Ghz Default w/ 1333FSB). The chips are created with identical materials and a nearly identical process. However, If you clock the E6850 to 4.2 Ghz you are more likely to see degradation due to the same 1.2Ghz bump. This is because the E6850 will now likely be forced to run outside of its temperature/voltage guidelines or at least at their max.

 

[homerdog]

Example - an E2160 (1.8ghz Default with 800FSB) when it's Clocked to 3.0Ghz (1333 FSB) it is not likely to suffer any more than an E6850 (3.0Ghz Default w/ 1333FSB). The chips are created with identical materials and a nearly identical process. However, If you clock the E6850 to 4.2 Ghz you are more likely to see degradation due to the same 1.2Ghz bump. This is because the E6850 will now likely be forced to run outside of its temperature/voltage guidelines or at least at their max.

I guess that would be assuming that Intel's binning means nothing, which may in fact be the case.

(A) Members of the (imaginary) "Level~Headed Overclocker's Club" should remain at a pretty good price/effort/performance/lifespan, once the benchmarking guys finish spending their money to show the level-headed ones where that point is.

I'm not imagining that my e6750 runs at 3.2GHz undervolted from 1.35V to 1.25V and maybe lower (I've never tried). I don't think I'm significantly shortening the lifespan of my CPU by doing that. The motherboard maybe, but not the CPU.

 

[yomamafor1]

(C) Going to liquid cooling sooner, rather than later is a Great Good Thing. Thankfully, this is easier and easier due to quality pre~configured and pre~built systems for this (Coolitsystems, etc...). Though your wallet won't be as happy when you're done.

Coolitsystems should not be considered as "quality pre-configured water cooling system". They may be sufficient at cooling dual cores, and low end quad cores at stock clock, they are not designed to handle high heat. Ability to handle up to 250W with only 80W of TEC? What a bunch of marketing BS.

Coolitsystem in my dictionary is considered as expensive trash.

 

[FHDelux]

I think its assumed that most people who have been overclocking a long time know when they are pushing the system too hard and accept that lifespan is shortened because of overclocking. Most people overclocking hard, aren't even holding onto their proc's for long enough to see degredation. Granted it is tempting to push 45nm chips harder because they stay cool and use a ton less voltage than the 65nm parts, but all you have to do is look at the spec on intel's website for whatever sSpec your part is to find out what the safe ranges to run your processor in are. If you are riding the edge of acceptable conditions, you processor wont last as long as it would if you were being more conservative. On that same note, if you are pushing a 45nm processor at 1.4+ vcore, and your temps are pushing the edge of acceptable, you are out of spec and can expect a higher rate of decay, thats the life of a computer parts.

http://processorfinder.intel.com/details.aspx?sSpec=SLAPL

 

[Scotteq]

Coolitsystems should not be considered as "quality pre-configured water cooling system". They may be sufficient at cooling dual cores, and low end quad cores at stock clock, they are not designed to handle high heat. Ability to handle up to 250W with only 80W of TEC? What a bunch of marketing BS.

Coolitsystem in my dictionary is considered as expensive trash.

That was simply the name that popped into my head... <shrug>

 

[SpinachEater]

You pay to play.

[:thegreatgrapeape:7]

 

[LukeBird]

An interesting article, and something that will keep me of the same mindset I have now. That is to buy faster stock components than bother to OC.
I was considering having a bit of a mess around, but that his has set my mind that it's just not worth it, for what I would gain (at most, some extra 3DMark bragging points... 😉)
I think the problem is, so many people think that just bumping the FSB on their new E8xxx is doing no harm, although this obviously shows that this isn't the case, hence mucho crying in 6/12 months when they go pop!
There are certainly plenty of OCers who see the danger and take it as an acceptable risk, but there are plenty more who think it's like free money...

 

[SpinachEater]

You pay to play.

[:thegreatgrapeape:7]

😗

 

[zenmaster]

An interesting article, and something that will keep me of the same mindset I have now. That is to buy faster stock components than bother to OC.
I was considering having a bit of a mess around, but that his has set my mind that it's just not worth it, for what I would gain (at most, some extra 3DMark bragging points... 😉)
I think the problem is, so many people think that just bumping the FSB on their new E8xxx is doing no harm, although this obviously shows that this isn't the case, hence mucho crying in 6/12 months when they go pop!
There are certainly plenty of OCers who see the danger and take it as an acceptable risk, but there are plenty more who think it's like free money...

Actually It if Free Money.
And Buying a Faster Stock Chip is not going to change the rate of decay.

The two chips are phyiscally the same.
Simply because Intel sets one to a higher default speed does not mean it will not decay any slower than a different CPU set to the lower defaults but then configured in BIOS to use higher speeds.

This is the major flaw in the entire article.

There is one truth: The more voltage and more heat applied to a processor the faster it will wear out.
However, That decay would be the same between two processor with different default speeds if they were set to the same speed and voltage while under the same heat conditions.

All of the charts involving life span and warranty are complete hogwash.
They have no data to back them up.

It's quite common for Intel CPUs to last 10-15years or more running 24/7.
(Yes, I've seen many many many running like this. Never seen a CPU go Poof under such conditions.)

Rather, The reason for the warranty is many fold.
Longer warranties always increase costs.
A CPU could go bad for many reasons - Including Aging PSUs causing bad voltage etc...
Ergo, the longer you warranty a CPU the more costs you will incur even if your CPUs will run perfectly fine for 20 years.

Additionally, there is more cost involved the longer you warranty a CPU for replacement purposes.
Intel needs to keep stock around on CPUs they stopped selling 3 years ago to fulfill Warrranties.
Often they cannot simply give a newer faster CPU since it will not work in the old system with the failed CPU.
This is cost.

It's an article filled with many pretty graphs and charts, but absolutely no substance.

They could have written a very simple sentence stating that the higher the clock, voltage, and heat you subject a processor the shorter it will last. Conversely, if you reduce the clock, voltage, and heat you will lengthen the life span.

The corollary to "I don't want to OC because it will reduce the lifespan of the CPU" is that you should underclock and undervolt all of your CPUs to increase their lifespan.

What they have done is written an article (Possibly even paid for by Intel/AMD to slow down OCers) that completely lacks any facts and if read carefully you will note actually makes very few real claims. Rather they are very vague statements.

Myself, I don't care if I reduce the lifespan of my CPU from 20 years to 19 years. It will be donated in less than 5.

 

[amdwilliam1985]

Nice article, I think I have the general ideas.

I got a question for the experts, let's say that I have 2 pc running same setups,
pc 1, place it in a hot place, resulting much higher overall system temp.
pc 2, overclock the cpu a little (ex, 500mhz to 1ghz with maybe a little voltage added), place it in a AC cooled place with good cpu cooler.

Let's say that pc 1 has cpu idle to load temps of 35 to 50c,
and pc 2 has cpu idle to load temps of 25 to 40c.

Here comes my question, which one will last longer? Sorry in advance if I didn't explain it more clearer.

 

[zenmaster]

Nice article, I think I have the general ideas.

I got a question for the experts, let's say that I have 2 pc running same setups,
pc 1, place it in a hot place, resulting much higher overall system temp.
pc 2, overclock the cpu a little (ex, 500mhz to 1ghz with maybe a little voltage added), place it in a AC cooled place with good cpu cooler.

Let's say that pc 1 has cpu idle to load temps of 35 to 50c,
and pc 2 has cpu idle to load temps of 25 to 40c.

Here comes my question, which one will last longer? Sorry in advance if I didn't explain it more clearer.

They will both last until you don't want them any longer.
Most likely until about 2023-2025.

The hard part will likely be finding a PSU that still fits the Motherboard power connectors or other system parts to use with your CPU.

 

[amdwilliam1985]

From my understanding of the reading, pc2 will last longer correct?
It got the lower temperatures even though it has higher clock and voltage.

 

[alpine18]

Good article. A couple years ago I was part of the crowd that thought that o/c did nothing to chip lifespan as long as you kept things cool.

I have an E6400 that o/c'd to 3.0 Ghz on air cooling and stock voltage out of the box. It would o/c to 3.2, but I left it at 3.0 24x7. Ran great for six months. One day my system crashed and it wouldn't o/c past 2.8 - I put an older water cooling kit on it and it ran fine for another month until I had to step it down to 2.6 - now I'm lucky to get 2.4 out of it for a few hours. It's now left at stock speed and works fine 24x7.